Figure 16.1 Hemolytic uremic syndrome (TMA), D+. The glomerulus is slightly hypocellular, and most of the glomerular capillary lumina are closed as a result of thickening of the capillary walls. Red blood cells and fragmented red blood cells are seen in the mesangial areas. The specimen is from a child with VTEC infection. This type of glomerular change is typical during the acute stage of the disease. (Ã—400.) (Courtesy of Drs. Charles Alpers and Vivette D'Agati.)
Editors: Jennette, J. Charles; Olson, Jean L.; Schwartz, Melvin M.; Silva, Fred G.
Title: Hepinstall's Pathology of the Kidney, 6th Edition
Copyright Â©2007 Lippincott Williams & Wilkins
> Table of Contents > Volume One > 16 - Hemolytic Uremic Syndrome, Thrombotic Thrombocytopenic Purpura, and Other Thrombotic Microangiopathies
Hemolytic Uremic Syndrome, Thrombotic Thrombocytopenic Purpura, and Other Thrombotic Microangiopathies
Zoltan G. Laszik
Fred G. Silva
Hemolytic uremic syndrome (HUS) and thrombotic thrombocytopenic purpura (TTP) are discussed together in one chapter along with related conditions of systemic sclerosis (scleroderma) and radiation nephropathy because of the significant overlap in clinical presentations, pathologic pictures, and certain pathogenetic mechanisms. Indeed, HUS and TTP show so much overlap that they are regarded by some authors as different clinical expressions of the same disease, whereas the form of systemic sclerosis with severe renal involvement has been listed by some investigators as one of the subdivisions of HUS.
Hemolytic Uremic Syndrome and Thrombotic Thrombocytopenic Purpura
Historical Background and Nomenclature
In 1925, Moschcowitz (1) published a case report of a 16-year-old girl who had a sudden onset of fever, anemia, and central nervous system (CNS) involvement. After an episode of pulmonary edema, she lapsed into a coma and died. She did not have renal failure, and clinical findings were limited to traces of albumin, hyalin casts, and granular casts in the urine. Blood urea and creatinine levels were normal. At autopsy, hyalin thrombi were present in the capillaries and terminal arterioles of the heart, liver, and kidneys. This report was followed by the observations that thrombocytopenia was an additional feature in Moschcowitz's disease and that the thrombi were composed predominantly of platelets (2). Singer et al (3) were the first to use the term thrombotic thrombocytopenic purpura. The term was introduced in 1952 by Symmers (4) for the vascular lesions of TTP.
In 1955, Gasser et al of Switzerland (5hemolytic uremic syndrome to describe a syndrome
encountered in five children that consisted of hemolytic anemia, thrombocytopenia, and acute renal failure. The patients had cortical necrosis of the kidneys, and cerebral symptoms were also encountered. HUS was considered a new syndrome that was distinct from TTP for the following reasons: (a) HUS affected mostly infants, whereas TTP was primarily recognized as a disease of adults; (bc) the dominant pathologic finding in HUS (renal cortical necrosis) was different from findings observed previously in TTP. However, renal cortical necrosis is present only in some patients with HUS. The characteristic glomerular and arteriolar changes of HUS were subsequently described by Habib et al in 1958 (6). Shortly thereafter, Habib's group proposed the term thrombotic microangiopathy, which they borrowed from Symmers (4), to consolidate the vascular lesions of HUS and TTP. Since the original description of HUS, it has become apparent that HUS may also be seen in adults and that the severity of renal involvement may vary from patient to patient, depending on the extent of vascular involvement.
The microscopic features of HUS and TTP are similar to those seen in related conditions, such as scleroderma renal crisis, malignant hypertension, eclampsia or preeclampsia, and postpartum renal failure. Therefore, thrombotic microangiopathy is an appropriate term to describe the pathologic picture in HUS, TTP, and related conditions and is in widespread use.
In addition to thrombotic microangiopathy, Symmers (4) also used the term thrombotic microangiopathic hemolytic anemia for the cases he described. The term has now been shortened to microangiopathic hemolytic anemia
Relationship Between Hemolytic Uremic Syndrome and Thrombotic Thrombocytopenic Purpura
Although the morphologic lesions of these two conditions are thought by many to be virtually identical, certain clinical differences have been enumerated. Chief among them are that TTP (a) occurs among older patients, (b) affects the CNS more commonly, (c) exhibits less frequent and less severe involvement of the kidney, and (d) involves multiple organs and has a poorer prognosis. However, these differences are not absolute, and some differences are less consistent than others.
First, with regard to age, TTP is not restricted to adults, and appreciable numbers of cases are seen in infants and children (7,8). Similarly, HUS is not confined to childhood, as was originally thought. Although the classic form of HUS is most common in infancy and childhood, many cases are reported in adults (9,10,11,12,13). Second, severe CNS involvement can also be encountered in HUS. Neurologic involvement was recorded in the original patients of Gasser et al (5), and CNS involvement is the most common cause of death in children who die during the acute phase of HUS (14). Third, clinical and morphologic renal abnormalities can be found in TTP and may occasionally be severe (7,8,15). In three reviews, clinical renal involvement was present in greater than 76% of patients (7,8,); however, renal involvement is usually mild and rarely equals the severity of that seen in HUS. Fourth, extrarenal involvement can be seen in HUS, as in TTP, including the colon, liver, pancreas, heart, and brain (16).
In addition, a review of a series of patients with TTP showed that the classic pentad of symptoms, as defined by Amorosi and Ultmann (7) and consisting of fever, thrombocytopenic purpura, microangiopathic hemolytic anemia, neurologic manifestations, and renal dysfunction, was present in only 40% of patients during the course of their illness (8). The frequent clinical presentation of TTP lacking the features of the classic pentad and the urgency to initiate treatment early in the disease have resulted in a gradual change of the diagnostic criteria required for TTP during the past 20 years. In appropriate clinical context, most investigators consider the two principal clinical featuresâ€”microangiopathic hemolytic anemia and thrombocytopenia (the diagnostic dyad)â€”to be sufficient to establish the diagnosis of TTP (17). This change in the diagnostic criteria of TTP over time has further narrowed the differences between HUS and TTP.
Therefore, to many investigators and clinicians, the distinction between HUS and TTP appears to be less clear than has been generally claimed. Indeed, with the exception of childhood diarrhea-associated HUS, there are no specific diagnostic criteria to distinguish TTP from HUS, and therefore, the use of the inclusive terms thrombotic microangiopathy, TTP/HUS, or HUS/TTP is appropriate (1718). However, for historical reasons and because clinical renal involvement is predominant in HUS but not in TTP, we will discuss the two conditions separately. Because both also have common basic features, their pathogenesis is considered together.
Microangiopathic Hemolytic Anemia
Microangiopathic hemolytic anemia refers to a type of anemia seen in HUS, TTP, and other related thrombotic microangiopathies. Microangiopathic hemolytic anemia is characterized by deformed and fragmented erythrocytes (helmet cells, burr cells, schistocytes) in the peripheral circulation; the Coombs test is usually negative. Thrombocytopenia is commonly present, and the small blood vessels, particularly the renal arterioles and glomerular capillaries, show fibrin deposition in their walls and lumina.
Investigators have suggested that the bizarre morphologic features of the red cells and the hemolytic anemia are caused by mechanical red cell fragmentation by shearing by fibrin strands at the site of vascular lesions (19). However, investigators have also pointed out that in the early stages of HUS (and sometimes later), thrombi in glomeruli and arterioles are either absent or too scant to be responsible for the mechanical destruction of red blood cells (9,20). The increased lipid peroxidation and reduced antioxidant activity of red blood cells described in HUS suggest enhanced erythrocyte susceptibility to oxidative hemolysis (21). Neuraminidase, secreted by bacteria or viruses, has been invoked as a possible cause of the damage to the red blood cells by its action on the cell membrane (22). Although microangiopathic hemolytic anemia is considered to be one of the principal features of TTP and HUS, there are a number of well-documented cases published in the literature of both HUS and TTP without evidence of schistocytes (13,23,24).
Hemolytic Uremic Syndrome
The earliest descriptions of HUS were confined to infancy and childhood, and the typical presentation of HUS had a prodromal episode of diarrhea followed by the onset of hemolytic anemia, thrombocytopenia, and renal failure. In addition to this classic form, which in most cases is caused by verotoxin (VT)â€“producing Escherichia coli (VTEC) infection, is an atypical form that is not associated with diarrhea. In many instances, the atypical form is related to an underlying disease or condition. In other cases, no apparent etiologic factor can be identified, and the disease is considered idiopathic.
Classification and Epidemiology
Also known as diarrhea-positive (D+) or epidemic HUS, classic or typical HUS is associated with prodromal diarrhea. This form occurs mainly in young children, accounts for most cases seen in North America and Europe, and develops in isolated cases or as outbreaks occurring mostly in the summer (25). Although the designation â€œepidemic HUSâ€ is also used for the classic form of HUS, sporadic cases may also present with prodromal diarrhea. In North America and Western Europe, most of the D+ cases are associated with infection from VTEC serotype O157 (26,27,28,29,30). However, many other serotypes of VTEC have also been linked to D+ HUS (30,31,32). Infection with Shiga toxinâ€“producing Shigella dysenteriae serotype 1 has been a common cause of D+ HUS in developing countries in Asia (33) and Africa (34) but not in industrialized countries (35). Thrombotic microangiopathy in the classic form owing to VTEC infection is most often confined to the glomeruli, with a consequently good prognosis.
Also known as diarrhea-negative (D-) HUS, the term â€œsporadic HUSâ€ is also used in this context; however, because some of the D-cases do present with recurrences and D+ cases also occur sporadically, the term â€œsporadic HUSâ€ is not equivalent with D-HUS. This form usually presents without prodromal diarrhea in both children and adults and in various reports accounts for approximately 5% to 12% of all cases of HUS (36,37). The onset is usually insidious, and marked proteinuria and hypertension are characteristic features. While D+ cases are mostly associated with VTEC infection, D-HUS represents a heterogeneous group; cases in the D-group have been linked to a wide variety of triggers, including nonenteric bacterial infections, viruses, drugs, malignancy, pregnancy, and systemic diseases. Cases of the D-form may also be familial (38) and may follow a relapsing or progressive course (39). The familial cases may be associated with genetic or acquired abnormalities of the complement regulatory proteins (40,41,42,43,4445). A few cases of D-HUS have also been shown to be induced by VTEC infection (46,47,48,49); however, most cases in the D-group have unknown origin. Thrombotic lesions involving the renal arterioles appear to be more frequent in the atypical forms. The D-form of HUS is associated with greater morbidity and higher mortality rates than D+ disease.
Based on the clinical presentation, etiologic factors, and underlying diseases, the following subgroups of D-HUS can be distinguished:
These rare forms account for fewer than 3% of all cases of HUS. They occur in more than one member of the same family, may occur at any age, and may follow a recurrent pattern. The transmission is either autosomal recessive or autosomal dominant. Abnormalities in the complement regulatory proteins with underlying hereditary genetic disorder(s) have been identified in some of the affected patients and asymptomatic family members. von Willebrand factor cleaving protease (ADAMTS13) deficiency was also shown in some patients with familial HUS. Changes are common in the renal arteries, hypertension may be present, and the disease is usually severe.
Sporadic (Noninfectious) and Recurrent (Nonfamilial) Forms
Genetic or acquired deficiencies of some of the complement regulatory proteins have been identified in a significant proportion of patients with these forms of HUS. Also, some patients with the recurrent form of HUS and their asymptomatic family members were shown to have
von Willebrand factor cleaving protease (ADAMTS13) deficiency. However, most cases in this group are idiopathic. The morphologic changes are similar to those seen in the familial form, and the disease is, just like in the familial form, usually severe.
Sporadic Forms Related to Infections with Nonâ€“VT-Producing Organisms
These cases occur at all ages and are a sequel to infection with such organisms as Streptococcus pneumoniae, Salmonella typhi, and viruses. Unlike in the D+ group, most of these infections are usually not associated with diarrhea. In the United States, infections with S. pneumoniae37). The prognosis is variable; HUS associated with infection carries an increased risk of mortality and renal morbidity compared with VTEC HUS (,51,52).
Forms Accompanying Systemic Diseases
This category includes those cases found in association with various autoimmune disorders (antiphospholipid syndrome, systemic sclerosis, systemic lupus erythematosus), anti-glomerular basement membrane antibody disease, antineutrophil cytoplasmic autoantibody-associated (â€œpauci-immuneâ€ crescentic) glomerulonephritis, and malignant hypertension.
These forms are related to pregnancy, bone marrow transplantation, irradiation of the kidney, transplant humoral rejection, and use of various drugs such as oral contraceptives, mitomycin C, cyclosporine, tacrolimus (FK-506), interferon, gemcitabine, quinine, ticlopidine, and clopidogrel (49,53,54
55). Women predominate in most categories in adults (17,55). Bloody diarrhea and quinine-associated diseases are more common in White people (56,57).
In its classic form, HUS is a disease of infants and young children. However, when a distinction is made between patients with D+ HUS and those with D-HUS, the children with D+ HUS appear to be younger than those with D-HUS. In a systematic review of the literature on diarrhea-associated HUS that included 49 studies conducted between 1950 and 2001 with 3476 patients enrolled, the mean age of patients was 2.4 years (range 0.1â€“18 years at recruitment into studies) (58). In a series reported by Fitzpatrick and associates (36), the mean age of 21 patients with D-HUS was 4.6 years, and the mean age of 155 patients with D+ HUS was 2.3 years. Renaud et al (59) also noted a relatively high incidence of the atypical (D-) subset of HUS in children greater than 3 years of age.
In classic HUS caused by VTEC infection, the incubation period is usually 3 to 4 days (25). The classic form usually begins with watery diarrhea and may progress rapidly to bloody diarrhea with hemorrhagic colitis (25). Clinically, the colitis may mimic appendicitis, diverticulitis, ulcerative colitis, regional enteritis, or intussusception. In severe cases, pseudomembranous enterocolitis, necrosis, and perforation of the colon with subsequent peritonitis can occur. Some evidence suggests that the colonic changes are part of the thrombotic microangiopathy, with ischemia as the underlying mechanism. Vascular microthrombi have been described in surgically removed segments of colon, as well as in autopsy material (,60,61). However, microthrombotic lesions are seen only in a few small vessels, even when hemorrhagic colitis is extensive; therefore, VTEC infection possibly causes direct endothelial damage in the colon leading to mucosal hemorrhage. It is also possible that the thrombi are secondary to tissue ischemia and/or necrosis. The natural history of the colitis is usually favorable; it can, however, be fatal (14,62).
The second phase of classic HUS occurs abruptly within a few days of the onset of the illness and is characterized by various combinations of acute renal failure, anemia, bleeding abnormalities, CNS disorders (e.g., headache, altered consciousness, paresis, aphasia, syncope, seizures, visual changes, and dysarthria), and cardiovascular changes (e.g., hypertension and congestive heart failure). The clinical manifestations of D-HUS are similar to those of D+ HUS, except that they develop without prior diarrhea and/or hemorrhagic colitis.
The renal manifestations include oliguria or anuria, hematuria, hemoglobinuria, proteinuria, various types of casts in the urine, hyperkalemia, and elevated blood urea nitrogen (BUN) and serum creatinine level. Approximately half of pediatric patients require dialysis (63). Hematuria is usually microscopic, hemoglobinuria is present in a few cases, and proteinuria may be severe. Anuria of more than 4 days' duration was noted in half of the cases of Gianantonio et al (60). Of the 55 infants (younger than 28 months of age) studied by Habib et al (64), 24 were anuric (12 for up to 7 days and 12 for 7 to 18 days), 18 were oliguric (9 for up to 7 days and 9 for 7 to 153 days), and 13 had normal urinary output but a decreased glomerular filtration rate. All 15 children greater than 28 months of age had proteinuria, 7 had oligoanuria, and 8 had normal urinary volume, but with severely impaired renal function.
Manifestations of both anemia and a tendency to bleed include intense pallor, weakness, melena, hematemesis, hematuria, petechiae, and ecchymoses in the skin. Hemoglobin levels are lowâ€”sometimes as low as 2 to
3 g/100 mLâ€”with increased plasma bilirubin levels. The morphologic features of the red blood cells are abnormal, with the presence of helmet cells, burr cells, and fragmented cells (schistocytes) in the peripheral blood; reticulocytes are increased. The Coombs test is almost invariably negative. Leukocytosis is common in the early stages. Platelets are decreased to varying degrees, but megakaryocytes in the bone marrow are usually present in normal numbers. Additional laboratory findings are marked elevation in serum lactic dehydrogenase level and a concomitant reduction in the serum haptoglobin level. The results of coagulation studies in HUS are not strikingly abnormal; a slight prolongation in prothrombin time occurs in approximately 50% of patients. Prolongation of the partial thromboplastin time is less common; plasma fibrinogen levels are variable, and accelerated fibrinolysis with fibrin-split products in the circulation has been recorded. Various tests also indicate functional platelet impairment (65) and destruction (66).
CNS disturbances such as irritability, restlessness, tremors, and ataxia occur. In more severe cases, coma, stupor, and decerebrate rigidity can also be seen. Generalized convulsions were described in about one third of the patients in three series (60,), but the incidence of significant encephalopathy is relatively low (68
Cardiovascular disturbances include hypertension and congestive heart failure. Hypertension is present in many patients, but clinically apparent cardiac involvement is infrequent (69). Pancreatic involvement with the clinical features of acute pancreatitis is relatively frequent in fatal cases and has been recorded in about 20% of cases at autopsy (70). Diabetes mellitus can also be seen (69).
In addition to the classic clinical presentation, incomplete forms of D+ HUS have also been described in children (71). These patients present either with bloody diarrhea, hemolytic anemia and thrombocytopenia without renal failure, or with bloody diarrhea, anemia, hematuria, or proteinuria but without azotemia or thrombocytopenia.
10,11,12,13,72,73). E. coli O157 infection may be the cause of HUS in some adults (12,73); however, many adults do not have the classic type of HUS. Some patients have a history of taking oral contraceptives (10,13,72), whereas other cases may be related to pregnancy, malignancy, chemotherapy, or other conditions. Patients without any of these conditions, described as idiopathic, are more likely to have severe ADAMTS13 deficiency (74). Severe hypertension is more common in adults, and in general, the prognosis is poorer for adults than for children (1273). Adults tend to have more arterial involvement than do infants and children (9).
Verotoxin-Producing Escherichia coli Infection
In the mid-1980s, verotoxin-producing E. coli (VTEC) clearly emerged as the major etiologic factor in the pathogenesis of classic HUS. The term VTEC refers to various strains of E. coli that produce one or two distinct bacteriophage-mediated protein exotoxins, VT1 and VT2. Verotoxin, first described by Konowalchuk et al (75), has an irreversible cytopathic effect on cultured Vero cells, a line of African green monkey kidney cells. Both VT1 and VT2 are composed of a single A subunit of 32 kD and five 7.7-kD B subunits (76 (77), and therefore, VTs are also referred to as Shiga-like toxins. Verotoxin-1 differs from Shiga toxin only by a single amino acid and is 50% homologous with VT2 (78,79).
The clinical significance of VTEC was first recognized by Riley et al (80) in two outbreaks of hemorrhagic colitis and subsequently by Karmali et al in 11 of 15 cases of sporadic HUS (81,82). These early findings have been confirmed by several other investigators in sporadic cases of HUS, as well as in outbreaks of VTEC-associated diarrhea and hemorrhagic colitis complicated by HUS. It is now well established that VTEC is associated with a spectrum of illnesses, including asymptomatic infection, uncomplicated diarrhea, hemorrhagic colitis, and HUS (83). Data from Western Europe and North America indicate that about 90% of children with HUS have some evidence of VTEC infection, and the O157 serogroup is the most commonly involved, seen in up to 83% of cases (26,69). Among various serotypes, the O157:H7 is the most common; however, geographic differences in the occurrence of various serotypes are apparent. Data indicate an emerging role of non-O157 serogroups in some countries in Europe (30,84), and non-O157:H7 VTEC strains predominate in Australia (85). A prospective case-control study from Argentina showed that most children with HUS had some evidence of VTEC infection; however, the O157:H7 serotype of E. coli was uncommon (32). In the United States, serogroup VTEC 0111 is the second most common bacterial cause of HUS after VTEC O157:H7 (86).
The number of outbreaks of E. coli O157 infections reported to the United States Centers for Disease Control and Prevention has increased dramatically, from 4 in 1992 to 46 in 2002 (87). The largest reported outbreak of E. coli O157 in North America affected greater than 500 people, with 45 cases of HUS and 3 deaths (). This outbreak was traced to undercooked hamburgers from a fast-food restaurant chain in Washington State. One of the largest ever outbreaks of E. coli O157 infection occurred during the summer of 1996 in Japan and affected a total of 12,680 patients (89). The probable source of the infection was lunch food supplied to elementary schoolchildren; 121 patients developed HUS and 3 children died.
Out of the total of 350 outbreaks in the United States between 1982 and 2002 there were 8598 cases of E. coli O157 infection, and 354 (4.1%) patients developed HUS (87). A survey report from the United Kingdom showed
that 15% of 1275 patients with E. coli O157 infection developed HUS (90). Young age, prolonged use of antidiarrheal agents, bloody diarrhea, fever, and elevated leukocyte count are reported risk factors for the progression of E. coli O157 enteritis to HUS (,92). Also, the type of VT produced by bacteria seems to play a significant role in the pathogenicity. Escherichia coli strains that produced only VT2 were shown to have higher pathogenicity than those that produced only VT1 (31).
In addition to VTEC, infection with S. dysenteriae type 1 has also been identified as an etiologic factor for D+ HUS. In 1978, Koster et al (93) reported a series of HUS cases from Bangladesh in association with shigellosis, severe colitis, and endotoxemia. Hemolytic uremic syndrome associated with S. dysenteriae type 1 infection is clinically and morphologically similar, but not identical, to the classic (i.e., VETC-associated) form of HUS (94). Morphologically, renal necropsy specimens of eight of nine patients with Shigella-associated HUS showed renal cortical necrosis, extensive glomerular thrombosis, or arterial thrombosis (94). Subsequently, several other reports confirmed the association of S. dysenteriae infection with HUS in both children and adults (34,95,96).
The relationship between VTEC and atypical (D-) HUS is less clearly established. Steele et al (97) reported a case of postpartum HUS with VTEC 06:H12 strain isolated from the stool without preceding diarrhea; in another case, E. coli serotype 0111:NM was found in a woman on oral contraceptives (98). Some cases of atypical HUS were reported following urinary tract infection caused by various serotypes of VTEC (46,47,48). Interestingly, two studies from Italy show evidence of VTEC infection in the majority of their cases of atypical disease (28,99).
The main reservoir of the VTEC is the intestinal tract of healthy cattle. Most outbreaks in the United States have resulted from transmission of the organism through the consumption of undercooked ground beef or dairy products, including raw milk (87). Various other sources and modes of transmission have also been reported, including swimming in infected water, drinking water, and consuming lettuce, apple cider, apple juice, coleslaw, melons, and grapes (87). Outbreaks caused by contaminated drinking water tended to be much larger than all other outbreaks (87). Outbreaks caused by secondary transmission of the organisms from person to person by the fecal-oral route have occurred at child daycare centers, at individual residences, and in communities (87). Outbreaks owing to animal contact are one of the more recently recognized transmission routes; these have occurred at various settings, including farms, county fairs, and petting zoos (87).
Hereditary Aspects: Familial Forms
Familial occurrence of HUS has been well documented, both in siblings and in related family members of different generations. Kaplan et al (38) obtained information on 41 families with 83 affected siblings. Analysis of these data revealed that the patients could be separated into two groups. In the first group, the onset of HUS in the siblings occurred within days to weeks of each other. Most of the families lived in regions where HUS was considered endemic. In the second group, the affected siblings developed their disease greater than 1 year apart. These families were from nonendemic areas. A suggestion was made that the HUS in the first group developed in response to environmental factors, whereas in the second group, an autosomal recessive mode of inheritance was postulated. The prognosis was much better in the first group, with a mortality rate of 19%, versus 68% in the second group. A retrospective review of 373 episodes of D+ HUS occurring in 356 families in Utah and neighboring states between January 1970 and September 2001 revealed that 17 families (4.8%) had more than one family member develop the disease (100). In 12 (3.4%) of these families, HUS episodes occurred within a month (i.e., concurrently); in five families (1.4%), the episodes were separated by intervals of several years. There were no statistically significant differences in demographic, seasonal, laboratory, clinical, or outcome variables between familial subsets (concurrent versus non-concurrent). The authors concluded that concurrent cases suggest a common source of infection or person-to-person transmission, whereas non-concurrent cases suggest common environmental risk factors or perhaps a genetic predisposition. However, genetic predisposition cannot be excluded with the concurrent forms either. Evidence supporting the role of exogenous factors in familial cases has also been provided by reports, including Shiga toxin-associated HUS in five siblings (101) and Campylobacter-associated HUS in a mother and her daughter (102). In addition to autosomal recessive inheritance, apparent autosomal dominant transmission of HUS has also been documented (103).
By April 2005, the International Registry of Recurrent and Familial HUS/TTP had collected data on 70 familial cases of HUS from 30 different families (personal communication, Dr. Giuseppe Remuzzi, April 2005). Forty of the 70 patients had died. Analysis of familial cases by Kaplan and Kaplan (104) shows that patients with the autosomal recessive pattern of inheritance had gradual clinical onset of disease, frequent relapses, and renal disease characterized by vascular lesions; end-stage renal disease (ESRD) and death were common. The clinical and morphologic features of patients with the autosomal dominant pattern of inheritance are similar to those of patients with autosomal recessive inheritance, except that the onset is typically in adults, and death occurs even more frequently in patients with the autosomal dominant
pattern of inheritance (>90%). To complicate the issue of hereditary HUS, there have been several reports of HUS in one sibling and TTP in another (105,106).
Although the pathogenesis of familial HUS is not completely understood, a number of molecular mechanisms contributing to the development of HUS in some of these families have been defined recently. Most of the molecular abnormalities that were identified cause complement activation through the alternative complement pathway, often with resulting complement consumption and HUS in some but not all affected patients. Similar genetic defects have also been described in some sporadic cases of HUS without E. coli or S. dysenteriae infection (40,41,,45,107,108). A detailed account of these abnormalities is given later in this chapter, in the section titled â€œEtiology and Pathogenesis of Hemolytic Uremic Syndrome and Thrombotic Thrombocytopenic Purpura.â€
Recurrent attacks of HUS in the same patient are uncommon. They have been described in families with the autosomal recessive forms of HUS (39), in patients with complement abnormalities (109), in association with anti-factor H autoantibodies (110), in association with oral contraception (109), in prostacyclin abnormalities (111), and following reinfection with VTEC (112). In addition, rare cases of D-recurrences in patients with an initial episode of D+ HUS have also been described (113). The subject was reviewed by Kaplan (39), who noted that, in general, the first episode, as well as recurrences, followed nonspecific prodromal symptoms. Factors responsible for the recurrences were not always apparent, and most patients came from nonendemic areas. The prognosis did not appear to be related to the number of recurrences. The mortality rate was 19%. The issue of recurrent HUS in renal transplantation is addressed in the section titled â€œRenal Transplantation.â€
Association With Other Renal or Systemic Diseases
Morphologic and clinical features of HUS can be seen in association with certain renal or systemic diseases. The renal morphologic changes of systemic sclerosis with scleroderma renal crisis are those of severe HUS with vascular involvement. The term (10) has been used for HUS secondary to severe hypertension, because the renal pathologic features of HUS are virtually identical to those of severe hypertension (i.e., malignant-phase essential hypertension, malignant nephrosclerosis, or malignant hypertension, historically). Some investigators consider the morphologic lesions of malignant nephrosclerosis to be a form of thrombotic microangiopathy (9,). Hypertension is discussed in further detail in Chapter 21.
Anti-glomerular basement membrane antibody disease was complicated by the pathologic changes of thrombotic microangiopathy in 6 of 12 cases in one series (114), and microangiopathic hemolytic anemia was present in 4 of these patients. One of the six patients with thrombotic microangiopathy developed severe hypertension before biopsy that may have contributed to the vascular changes noted. A patient with antineutrophil antibodyâ€“positive crescentic glomerulonephritis and thrombotic microangiopathy has also been described (115). Thrombotic microangiopathy in systemic lupus erythematosus (SLE) is discussed later.
Pregnancy and Oral Contraceptives
Eclampsia and preeclampsia, pregnancy-related TTP, and postpartum renal failure (synonyms for the last include postpartum HUS and postpartum malignant hypertension) are discussed in detail in Chapter 17. Because both the renal morphologic changes and the clinical and laboratory presentation of preeclampsia and eclampsia have some features in common with HUS and TTP, eclampsia or preeclampsia is considered by many investigators to be a form of thrombotic microangiopathy (TMA) (116).
In up to 31% of women with TTP/HUS, the first presentation is during pregnancy or after giving birth (117). During pregnancy, it is most frequent during the second and third trimesters (53). A recent literature review of pregnancy-related case series of TTP/HUS published between 1964 and 2002 indicated that 47% of cases occurred during the postpartum period (53). The typical features of postpartum HUS are sudden deterioration of renal function associated with microangiopathic hemolytic anemia and thrombocytopenia, usually within days, but no more than 3 months after normal pregnancy and delivery (118). The onset of disease is usually preceded by a short symptom-free postpartum interval. Blood pressure is usually normal at the onset of HUS, but severe hypertension frequently develops as the disease progresses. The morphologic changes are those of typical TMA. Although complete renal recovery has been reported in some cases, the mortality rate is high (53,119). The etiology of TTP/HUS during pregnancy and the postpartum period is likely to be multifactorial. Women with various genetic risk factors for thrombosis have a higher frequency of pregnancy-related complications and may also be at increased risk for TTP/HUS (120,121). Decreased von Willebrand cleaving protease (ADAMTS13) activity has been documented during the second and third trimesters of normal pregnancies (122) and has also been postulated as a risk factor for TTP/HUS in this setting. Severe ADAMTS13 deficiency (activity <5% of normal) has been reported in some patients with TTP/HUS occurring during pregnancy or postpartum (122), and pregnancy has also been postulated to trigger the first episode of TTP/HUS in patients with the familial forms of severe ADAMTS13 deficiency (123). One
patient with postpartum TTP/HUS has been described with heterozygous factor H deficiency (45). Additional factors that have been implicated to trigger postpartum TTP/HUS include influenza-like illness, VTEC infection, Shigella infection, streptococcal infection, lupus anticoagulants, and various drugs used during delivery (119,124). HUS has also been recorded in women taking oral contraceptive agents (10,13,72); however, causality between oral contraceptives and HUS has not been clearly established.
The course of many malignant diseases can be complicated by TMA. Most commonly, mucin-producing metastatic adenocarcinomas of stomach, breast, and lung are the underlying diseases (125). Clinically, most of these patients present with hemolytic anemia of abrupt onset. Although mild azotemia may occur, renal failure is not a common feature in this disease. Rarely, the clinical picture may resemble HUS or TTP. In addition to hemolytic anemia, coagulation studies frequently indicate disseminated intravascular coagulation, which suggests a different pathogenetic mechanism from classic HUS. Usual autopsy findings are tumor emboli and fibrin microthrombi, mostly within small pulmonary vessels, including arterioles and capillaries (126).
A number of drugs have been implicated in inducing TMA with the clinical features of HUS or TTP (56,74,127; reviewed in Medina et al ). Among these drugs, quinine is the most common (54,128,129). Patients with quinine-associated TMA are usually older, predominantly white women. The clinical course is usually severe, with high acute mortality. Acute renal failure is common and there is also a high risk of chronic renal failure. Neutropenia, liver toxicity, and relapses after quinine re-exposure are often seen. The disease usually has an explosive onset, and the pathogenesis is drug-dependent antibody-mediated (54,129). The association of TMA with the antiplatelet agents ticlopidine and clopidogrel has also been reported by several authors (54).
Among various chemotherapeutic drugs, mitomycin C is the most commonly reported (130). HUS develops in probably less than 10% of patients treated with mitomycin, and the onset is usually within 4 to 8 weeks of the last dose of chemotherapy (125,131). The onset is usually insidious, and typical clinical presentation includes microangiopathic hemolytic anemia, thrombocytopenia, and renal failure. The course is potentially progressive even after discontinuation of the drug. The prognosis is poor. Most patients die within a few months of diagnosis, either from pulmonary or renal failure or from underlying malignant disease. The pathogenesis is not clear; however, the effects are dose dependent, pointing to direct toxicity (131). Experimental data also suggest that mitomycin also has a direct action on the kidney (132).
Based on biopsy and autopsy reports, the renal changes in mitomycin-associated HUS are those seen in the more severe forms of HUS, with cellular intimal thickening of renal interlobular arteries and thrombotic glomerular lesions (130,133). Mesangiolysis can be severe. Autopsy studies indicate that the microvascular lesions are primarily found in the kidneys; however, fibrin thrombi in the small vessels of the brain and fibrocellular intimal thickening of pulmonary arterioles and small arteries have also been described (133).
Thrombotic microangiopathy has also been described in association with a number of additional drugs, including antibiotics, H2 receptor antagonists, hormones, interferons, nonsteroidal antiinflammatory drugs, vaccines, and various chemotherapy agents, including bleomycin, cisplatin, daunorubicin, cytosine arabinoside, cyclophosphamide, doxorubicin, gemcitabine, and vincristine (54,56, 130). Based on a limited number of detailed pathologic descriptions in these cases, the renal morphologic features seem to be those of typical HUS.
Thrombotic microangiopathy caused by the immunosuppressive agents cyclosporine, tacrolimus, and sirolimus is discussed in the following section.
Thrombotic microangiopathy is a well-known complication of renal transplantation. It may occur in renal transplant recipients either as a recurrent disorder or as a de novo disease (see Chapter 28). De novo TMA in renal transplants has been linked to treatment with various immunosuppressant drugs, viral infections, and anticardiolipin antibodies. Recurrent posttransplant HUS is a complication of the posttransplant course of patients whose original renal disease was HUS. The morphologic features of TMA can also be seen in transplant kidneys in the setting of acute rejection. By virtue of remarkable morphologic similarity between some of the findings in patients with C4d-positive acute humoral rejection and those with HUS, this form of acute humoral rejection is also viewed by some investigators as a subtype of TMA.
In a historical cohort study of 15,870 renal transplant recipients from the United States Renal Data System (USRDS) between January 1, 1998, and July 31, 2000, the incidence of de novo TMA was 0.8% with a 1.26-year mean follow-up (134). The risk of TMA was highest for the first 3 months after transplant. Risk factors for de novo TMA included younger recipient age, older donor age, female recipient, and initial use of sirolimus. Patient survival rate was approximately 50% at 3 years.
Cyclosporine-associated TMA was first recognized in bone marrow transplant patients, followed by descriptions in patients with solid organ transplants, including
kidneys (135,136,137). In renal transplants, clinical recognition of de novo TMA secondary to cyclosporine toxicity can be very difficult, if not impossible. The classic clinical-laboratory features of TMA (microangiopathic hemolytic anemia and thrombocytopenia) are often missing, and rising serum creatinine may be the only abnormal laboratory finding. Zarifian et al (138) described 26 patients who developed biopsy-proven de novo HUS in renal transplants; however, only two of them showed thrombocytopenia and microangiopathic hemolytic anemia with elevated lactate dehydrogenase. These findings indicate a crucial role of renal biopsy in diagnosing de novo TMA in this setting. The morphologic findings of cyclosporine-associated TMA seem to be similar to those described in classic HUS, with both glomerular capillary and arteriolar and arterial lesions present. The precise incidence of cyclosporine-associated TMA is not well established. The frequency varies between 3% and 14% in various reports (138,139). The difference might be attributed in part to clinical variables, differences in biopsy practices, and differences in biopsy interpretation. Tacrolimus (FK506), another calcineurin inhibitor and mainstream immunosuppressant in transplant patients, has also been described in association with de novo TMA in renal transplants (139,140). According to some studies, the incidence of tacrolimus-associated TMA is somewhat lower than that associated with cyclosporine. However, other studies showed no differences in the rates of TMA with the use of cyclosporine versus tacrolimus (134).
Although the precise pathogenetic mechanisms of cyclosporine-induced TMA are not known, there have been a number of observations, both in vitro and in vivo, that link cyclosporine to microvascular thrombosis (reviewed by Ruggenenti ). Direct endothelial injury (142143), suppression of the protein C anticoagulant pathway (144), induced production of thromboplastin by mononuclear cells (145), and increased production and release of high-molecular-weight von Willebrand factor (vWF) multimers from the endothelial cells (143) have all been implicated as potential contributing factors. In addition, cyclosporine but not tacrolimus induces renal hypoperfusion that can further perpetuate thrombosis in renal microcirculation (146). Recently, the role of sirolimus has also been noted as a risk factor for TMA (134,147,148).
Although less often than with calcineurin inhibitors, viral pathogenesis has also been implicated in the development of posttransplant HUS. Viral infections linked to de novo HUS include influenza, cytomegalovirus (CMV), and parvoviruses (149,150151). Both CMV and parvoviruses can cause endothelial injury, which can trigger HUS in this setting (149,151).
Recurrence of HUS in transplant kidneys has been described in both children and in adults (reviewed by Ducloux et al ). The higher overall rate of recurrence in adults is attributed to the higher frequency of the atypical form of HUS (e.g., D-HUS) in adults. In a significant proportion of patients with the atypical form, there is genetic predisposition to the disease, making recurrence more likely in the transplanted kidneys. In children, approximately 80% of HUS cases are secondary to VTEC infection, without any well-defined genetic predisposition. Accordingly, the overall recurrence rate of HUS in renal transplants is much lower in the pediatric population. However, in children who had D-HUS as their original disease, the high recurrence rate (50% to 90%) in transplant kidneys is comparable to that described in adults (153,154). A recent retrospective study from France encountered 16 adult patients who underwent renal transplantation for treatment of ESRD for HUS (153). The recurrence rate was 56% (n = 9), and 25% of patients (n = 4) demonstrated some clinical or pathologic evidence of recurrence that could not be distinguished from acute vascular rejection. The three patients who had no signs of recurrence of the original disease had a previous diagnosis of D+ HUS. A literature review from the same study indicated an overall recurrence rate of 52% for HUS in 71 patients who received a total of 90 grafts. In a recent historical cohort study from the United States, the incidence of recurrent HUS in renal transplants was 29.2% (134).
Human Immunodeficiency Virus
Hemolytic uremic syndrome was first diagnosed in a homosexual man with HIV infection in 1984 (155). Since then, many more patients have been described with HIV infection who had thrombocytopenia, hemolytic anemia, and various other symptoms with the diagnosis of either HUS or TTP (156,157,158,159,160,161,162). HUS has been reported in patients with full-blown AIDS, as well as in patients with asymptomatic HIV infection. Earlier reports on the incidence of HIV-associated TMA varied considerably, ranging from 0.6% to 7% (162,163,164). Some recent data, however, seem to indicate that the incidence of TMA has declined since the advent of highly active antiretroviral (HAART) therapy (160,161). A study from Italy found a 1.4% incidence of TMA in HIV-infected patients from the pre-HAART era (January 1985 to December 1996) and 0% incidence during the HAART period (January 1997 to December 2000) (161). Another study from the United States reported a 0.3% incidence of TMA from a cohort of 6022 HIV-infected
patients during the HAART era (160). The clinical and laboratory features of HIV-related TMA appear to be similar but not necessarily identical to those of idiopathic HUS and TTP (157162). Differences between idiopathic forms and HIV-associated TMAs may include gradual onset of the disease, prolonged survival without plasma therapy, and less severe thrombocytopenia in the HIV-infected patients (162). In contrast to the high complete response rate to plasma therapy in patients with the classic syndrome, patients with HIV-associated TMA may show only partial response to this treatment modality (162).
The clinical diagnosis of HIV-associated TMA might be hampered by frequent comorbidities and multiple drug use in these patients, which can display overlapping clinical and laboratory features with TMA. Although HIV positivity may not adversely affect short-term survival (157), the overall survival of most patients is less than 12 months (156). The mortality of HIV-associated TMA was as high as 100% in some series (161).
The pathogenesis of HIV-associated TMA is not well understood (reviewed by Alpers ). Direct injury of the endothelial cells by HIV has been hypothesized; however, there is no conclusive evidence to support this. A number of other factors have also been implicated as possible triggers or predisposing factors for TMA in HIV-infected patients. These include CMV infection, cryptosporidiosis, AIDS-related cancer, and various drugs (161,162,166). Antiphospholipid antibodies that have been linked as triggers of TMA in various settings seem to be quite common in various viral infections without triggering TMA, and therefore, their pathogenetic role in HIV-associated TMA is in doubt (167,168).
The pathologic features of HUS have been studied in renal biopsies and autopsy material (9,10,13,60,,169,170,171). The basic morphologic changes are similar in most cases, regardless of cause. The severity of clinical disease rests mainly on the extent of involvement and, in particular, the presence of changes in renal arteries. Many data indicate that in D-HUS, the renal arterial involvement is more widespread and severe than in D+ forms. This difference in the renal morphologic features between typical (D+) and atypical (D-) forms may explain the poorer prognosis seen in the D- form. Moreover, certain morphologic features may be more pronounced in some forms of HUS, such as mesangiolysis in HUS associated with mitomycin and bone marrow transplantation. Sometimes, changes of HUS may be superimposed on those of other glomerular or vascular diseases (e.g., lupus nephritis or vascular lesions in chronic hypertension or arteriosclerosis).
The morphologic lesions of the kidney in childhood HUS most likely represent only the most severe end of the spectrum, since the usual course of HUS in children is relatively mild and only those patients with the most severe or lingering clinical symptoms undergo biopsy or autopsy (<5% to 10% of children with HUS). Most children with classic (D+) HUS do not require a biopsy and recover with no or minor residual renal symptoms.
The microscopic features of HUS have been traditionally divided into early (acute) and late (chronic) changes. We follow this approach in the microscopic description; however, some overlap between acute and chronic features may occur.
Renal cortical necrosis is a frequent finding in patients who die of HUS and is variable in its extent. Sometimes, one sees large areas of necrosis, but more often the necrosis is patchy, although widespread, so the swollen kidney has a reddish, mottled appearance. Calcification may be apparent in the previously necrotic areas in patients who have survived for longer periods; this can be seen on radiographs of the abdomen. Other patients who die may have no apparent cortical necrosis, although petechial hemorrhages are seen in an enlarged, swollen kidney. Bilateral nephrectomy specimens from patients who have developed irreversible renal failure with uncontrollable hypertension may be of normal or reduced size if the patient had an extended period of hemodialysis before nephrectomy. Reduction in size is partly a consequence of the damage inflicted by the disease itself, particularly if there has been extensive arterial narrowing; however, for patients who have undergone long-term dialysis, dialysis arteriopathy may also be observed. Focal scarring or areas of calcification may be seen corresponding to previous areas of necrosis.
Glomerular morphologic features vary according to the severity and the duration of the disease and the presence or absence of arterial changes. The percentage of glomeruli with pathologic changes may also vary; in some instances, few glomeruli are involved, whereas in other cases, most of the glomeruli are affected.
In the early stages, glomeruli may show thickening of the capillary walls, caused mainly by expansion (swelling) of a thin layer between the endothelial cells and the underlying basement membrane. Severe swelling of the glomerular endothelial or interposed mesangial cells with subendothelial widening may occlude the capillary lumina (Fig. 16.1). The term bloodless is used to characterize glomeruli with complete closure of the capillary lumina (Fig. 16.2). Separation of the endothelium from the underlying basement membrane and production of new basement membrane-like
material by endothelial or interposed mesangial cells results in the occasional double-contour appearance of the glomerular capillary walls, which is best seen with silver or periodic acidâ€“Schiff (PAS) (Fig. 16.3). The glomerular capillary lumina may have fragmented red blood cells, fibrin, and platelet thrombi (Fig. 16.4). Fibrin is sometimes clearly detectable beneath the glomerular capillary endothelial cells. Larger localized areas of fibrin may be seen in the glomerular capillary tufts, particularly in continuity with thrombus or fibrinoid necrosis in the afferent arteriole as it enters the glomerulus (Figs. 16.5 and 16.6). The glomeruli may also be congested, containing red blood cells in dilated capillary loops, especially in cases with severe vascular involvement (Fig. 16.7). This feature is sometimes designated â€œglomerular paralysis.â€ This change is typical in patients in the early stages of cortical necrosis; frank glomerular necrosis usually develops later. Small crescents may occasionally be present. An increased number of polymorphonuclear leukocytes in the glomerular capillaries can be seen in some cases; this feature may be prominent in patients with D+ HUS (170). Capillary thrombi, endothelial swelling, and congestion were identified as typical glomerular findings in patients with severe D+ HUS (170).
Mesangial abnormalities may also be present during the early stages. A fibrillar appearance of the glomerular mesangium is a characteristic feature, particularly in patients with narrowed arteries and arterioles. The reason for this fibrillar appearance is not obvious; collapsed
glomerular capillary walls and mesangial edema may be contributing factors. Fibrin and fragmented red blood cells can be seen in the mesangium (). Mesangial cells, although often swollen and hypertrophic, are usually not increased in number during the acute phase of the disease. If mesangial cell proliferation occurs, it is usually slight, focal, segmental, and late.
Occasionally, mesangiolysis can also be seen. The term mesangiolysis was first used by Yajima (172) in 1956 in patients with nephritis associated with subacute bacterial endocarditis. However, the glomerular capillary cysts, one of the most typical features of mesangiolysis, were described earlier by Pearce (173) in an experimental model of glomerular lesion induced by Crotalus adamanteus venom. The term refers to partial or complete dissolution of the mesangial matrix and cells. The affected glomerular lobules of mesangiolysis stain poorly because of mesangial edema. The borders of the dissolving mesangium are hazy, and the mesangial matrix is difficult to identify. Usually, no associated inflammatory reaction or fibrin deposition is seen (174). Eventually, glomerular basement membranes become unanchored from the underlying dissolving mesangial mass, leading to markedly dilated and sometimes cystic capillaries (Fig. 16.8). A particularly severe and widespread form of mesangiolysis can occur in HUS after bone marrow transplantation or mitomycin therapy. However, in classic D+ HUS, mesangiolysis has rarely been described. Mesangiolysis can also be seen in diabetes mellitus, various forms of glomerulonephritis, and transplant glomerulopathy (175). â€œHealingâ€ of
mesangiolysis may lead to proliferating or sclerosing glomerular changes as the disease progresses. In the late stage of mesangiolysis, the mesangium may be thickened by pale fibrillary (sclerotic) material. This process of healing and sclerosing may lead to a distinctive pattern of glomerular sclerosis (â€œbland sclerosisâ€) characterized by loss of glomerular cells and capillary lumina, but with at least partial preservation of the lobular architecture.
Ischemic-type glomerular injury, characterized by collapse of the glomerular capillary tuft and thickening and wrinkling of the capillary basement membranes, during the acute stage of HUS usually indicates severe vascular lesions such as (a) arteriolar or arterial thrombi, (b) acute thickening of the arterial or arteriolar intima, or (c) coexistent chronic hypertensive vascular disease (arteriosclerosis). Focal necrotizing glomerular lesions can also be seen in HUS, albeit rarely. If present, they are usually small, affecting only a few capillary loops or a segment of the glomerulus. Sometimes, the necrotizing lesion is associated with arteriolar thrombosis and/or fibrinoid necrosis of the arteriolar wall.
Fig. 16.9). The double contour is composed of new (inner) basement membrane and the original (outer) basement membrane. In the more advanced stage of HUS, the double contours with the mesangial sclerosis and occasional mild hypercellularity may give rise to a pattern of glomerular injury reminiscent of a membranoproliferative glomerular lesion (Fig. 16.10). Double contours of the glomerular capillary walls may apparently persist for several months or years.
The membranoproliferative pattern injury is nonspecific because it can also be seen in immune complexâ€“mediated glomerular diseases such as idiopathic forms of membranoproliferative glomerulonephritis, cryoglobulinemic glomerulonephritis, transplant glomerulopathy, and glomerular lesions with paraprotein deposition (176). Light microscopic differentiation of membranoproliferative glomerulonephritis from membranoproliferative-type glomerular injury of advanced HUS is usually straightforward; in HUS, the double contours of the glomeruli are usually sparse, and no or only minor mesangial hypercellularity is present. In addition, electron microscopy does not disclose discrete, electron-dense, immune-type deposits in advanced HUS.
Segmentally sclerotic glomerular lesions can occasionally be seen in cases with evolving chronicity. The lesions closely resemble those seen in idiopathic focal segmental glomerulosclerosis with focal and segmental collapse of the glomerular capillary lumina, mesangial matrix accumulation, and visceral epithelial cell hyperplasia overlying the segmentally sclerotic segments. These segmentally sclerotic changes may represent healed necrotizing glomerular lesions. Alternatively, chronic sclerosing-type glomerular injury with segmental features can also develop as part of evolving chronicity affecting glomeruli.
Chronic ischemic-type glomerular injury is characterized by thickening and wrinkling of the glomerular capillary basement membranes, simplification of the glomerular tuft, widening of Bowman's space between the collapsed glomerular loops and Bowman's capsule, and collagen accumulation internal to Bowman's capsule replacing Bowman's space. The ischemic changes may affect the glomeruli globally or segmentally. Simplification refers to shrinkage of the glomerular capillary tuft accompanied by an apparent decrease in the number of normal glomerular lobules and by apparent loss of mesangial matrix and cells. Accumulation of PAS-negative collagen material inside Bowman's capsule usually begins at the hilar region of the glomerulus, and eventually the collagen may involve the entire circumference of Bowman's space. If the changes progress to complete glomerular ischemic obsolescence, the glomeruli appear as small, hypocellular, compact, eosinophilic masses (â€œtombstonesâ€ or globally sclerotic glomeruli). However, with PAS, the collapsed PAS-positive glomerular tuft can easily be distinguished from PAS-negative collagenization of Bowman's space. In patients who have TMA superimposed on glomerulonephritis (e.g., lupus glomerulonephritis), the glomeruli may be markedly hypercellular in addition to the typical changes of HUS (Fig. 16.10).
Arteries and Arterioles
Various changes occur in the renal arteries and arterioles. Arteriolar and arterial changes are more common in patients with D-HUS (9,171). In the early stages, renal arterioles show swelling of the endothelial cells and subendothelial space. The arteriolar lumen may be severely narrowed, and sometimes fragmented red blood cells are seen in the thickened arteriolar wall. Infiltration of the arteriolar wall by fibrin may occur, a change often referred to as fibrinoid necrosis). The term fibrinoid necrosis is probably a misnomer; little evidence indicates that cellular necrosis is a constant feature of the lesion. Fibrinoid necrosis is thought to be related to increased vascular permeability and nonspecific trapping of plasma proteins, including fibrin in arteriolar walls. Fibrinoid necrosis tends to occur only at the hilum of the glomerulus, and it may only involve the thickened intima of the arterioles. More often, however, the media of the arteriole is also affected. Unlike in true leukocytoclastic vasculitis, acute inflammatory cell infiltrate is rarely seen in fibrinoid necrosis with HUS. As the disease progresses, the involved arterioles tend to become hyalinized, losing the staining reactions for fibrin. Hyalinized arterioles show homogenous, eosinophilic, refractile, strongly PAS-positive acellular material accumulated in the intima or media. Fibrin thrombi may also be seen in the afferent arterioles, and these may continue into the glomerular capillary tuft (Figs. 16.9glomeruloid.
Interlobular renal arteries may show two major changes. One of the early abnormalities is swelling of the intima, which may be accompanied by a suffusion of red blood cells (some of which may be fragmented) or fibrin. Fibrin may appear deep in the intima and may permeate the wall extensively; as in arterioles, this lesion is also called fibrinoid necrosis. The second lesion is intimal swelling, which is usually sparsely cellular, containing mainly lucent amorphous material with a mucoid appearance (Fig. 16.12mucoid intimal hyperplasia; it may be severe, with marked narrowing of the lumen. Often, one sees a rapid proliferation of cells in the intima that might be regarded as organization of the intimal edema. The proliferating intimal cells are myointimal cells and are responsible for the cellular intimal thickening that occurs later in the course of the disease (Fig. 16.1364) and are responsible for the secondary ischemic glomerular changes such as shrinkage of glomerular capillary tufts and wrinkling and thickening of the capillary walls. Fibrous
replacement of the thickened intima is a later change in the interlobular arteries. Occasionally, thrombosis (Fig. 16.14) and recanalization can also be seen in the interlobular arteries. Changes similar to those seen in the interlobular arteries may also be present, although less frequently, in the larger (arcuate and intralobar) arteries. Especially in older patients with pre-existing chronic hypertension, the acute vascular lesions of HUS may be superimposed on chronic vascular changes such as intimal fibroplasia, medial hypertrophy, or arteriolar or arterial hyalinosis.
The tubules frequently contain hyalin casts and red blood cells. Frank tubular epithelial necrosis may occur, or acute tubular necrosis, as is seen with ischemia, may be present. In later stages, tubular atrophy may be seen. In cases with cortical necrosis, small patchy infarcts or larger necrotic areas are seen. Rarely, calcifications of the cortex can be widespread in chronic cases of cortical necrosis.
The interstitium may be edematous or fibrous, and in some cases it contains mild mononuclear cell infiltration. Large numbers of red blood cells are present in the interstitium in areas of cortical necrosis.
In the glomeruli, the usual finding is the presence of Fibrinogen or fibrin along the capillary loops in a continuous, broken linear or granular pattern (10,13,); fibrinogen or fibrin is found less frequently in the mesangium () (). The deposits along the capillary walls may be accompanied by IgM (10,82), by C3 (10,13,82,177,178), less frequently by IgG (20178), and only rarely by IgA (10) (Fig. 16.16). Intracapillary thrombi also contain fibrinogen or fibrin and fibrin fragments.
Arterioles and small arteries often exhibit fibrinogen or fibrin in their walls, usually in a subendothelial position (1013,177,178) (Fig. 16.15B10,177,178), as were C3 (10,13,177,), C1q (10), and IgG and IgA (10). Intravascular thrombi also show positive fluorescence for fibrinogen or fibrin.
The changes described from electron microscopy have been remarkably constant (9,10,169,178,179
and swelling of the endothelial cells. The acellular subendothelial â€œfluffâ€ is pale and rarefied and contains irregular collections of electron-dense material (Fig. 16.17). This subendothelial material in the lamina rara interna of the glomerular basement membrane is usually granular and has a variable electron density. Sometimes it has a fibrillar or beaded appearance, but it usually lacks the periodicity and electron density of fully developed fibrin. The exact nature of the subendothelial material is unknown, but it is thought to represent the breakdown products of intravascular coagulation or cell debris organized into the capillary wall. Signs of endothelial damage may occur early during the course of HUS, with features such as swelling (20), localized areas of detachment of the endothelial cytoplasm from the basement membrane (179), and cytolysis (,169). One may see intracapillary thrombi composed of amorphous osmiophilic material admixed with fibrin, platelets, and deformed red blood cells (Fig. 16.18). The platelets may be numerous and may fill the capillary lumen. Platelets can sometimes be seen between the endothelial cells; occasionally, the remnants of platelets are visible within the lamina densa of the glomerular capillary basement membrane. However, platelets are evanescent and may not be identified. Electron-dense fibrillar fibrin wisps are sometimes a conspicuous feature in the lumina or in the subendothelial space. The swollen mesangial matrix appears as a meshwork filled with electron-dense, finely granular or fibrillar material similar to the changes seen in the subendothelial areas (179). Actual mesangiolysis may be seen, and this progressive disintegration results in capillary ectasia. One sees frequent visceral epithelial foot process effacement.
In older lesions, wrinkling and collapse of the glomerular basement membranes may occur and are conspicuous in some cases. Multiple layers of material resembling basement membrane are often seen in the capillary wall (10) with cellular (mesangial) interposition (Fig. 16.19
Arterioles and arteries show changes in the endothelial cells similar to those seen in the glomeruli. One sees swelling, cytolysis, and detachment of the endothelium from the underlying structures (9) with a widening of the intima. The intima has a lucent appearance, with strands or granules of greater electron density (Fig. 16.20). Structures consistent with fibrin are found at various depths of the vessel wall; luminal thrombi made up of platelets, Fibrin, and electron-dense material may be present. In the later stages, elongated myointimal cells abound in the thickened intima.
Changes in Other Organs
Arteriolar and arterial changes may be confined to the kidney, but sometimes small thrombi may also be seen in
capillaries and arterioles or small arteries of other organs. An autopsy study of 65 pediatric patients with D+ HUS showed that extrarenal pathologic changes were most common in the large bowel, followed by the CNS, the heart, and the pancreas (180). Argyle et al found extrarenal microthrombi in eight of nine autopsies of children (16). The thrombi were associated with hemorrhagic colonic necrosis in six patients and with pancreatic islet cell necrosis in three patients; microthrombi were also present in the brain, heart, adrenals, and lungs. In the colon, it was not possible to determine whether the capillary thrombi resulted from the primary process of HUS or whether they were secondary to nearby tissue necrosis. In another study, small infarcts of the liver or microscopic areas of cell death were described in association with fibrin thrombi in sinusoids, and small areas of infarction of the brain were recorded.
Outcome and Prognostic Features
Until the 1970s, many children died during the acute anuric or oliguric stage in classic HUS. Since the advent of hemodialysis, the overall mortality rate has declined to 12% in cases of HUS and it is only about 4% with classic (D+) HUS (14). At this time, acute renal failure has practically been eliminated as a cause of death in the acute phase of HUS.
Most studies indicate that pediatric patients with D-HUS have a worse prognosis and renal outcome than those with D+ HUS (49181). However, a study from the western United States showed that children with atypical disease had milder nephropathy and, with the exception of those with recurrent disease, did not experience a worse outcome (182). The prognosis is also much better in infants than in children. However, age by itself at the onset of HUS is not a prognostic feature (59). The difference in outcomes between children and infants is most likely related to the high incidence of the atypical subset (D-) of HUS in children older than 3 years, a subset that is uncommon in infants and carries a worse prognosis.
A number of clinical-laboratory features, such as the severity of gastrointestinal tract symptoms (183), the longer duration of dialysis (184), the duration of anuria (185,186), hypertension (185), initial peripheral leukocytosis (187), and neurologic involvement during the acute phase (63), have been proposed as predictors for poor
long-term renal outcome in pediatric patients with HUS. However, many of these proposed prognostic factors have not been uniformly confirmed by various investigators. A large study from Italy that included 387 children with HUS with a median age of 1.9 years indicated concomitant absence of prodromal diarrhea and VTEC infection as the best indicators of poor renal prognosis (49).
A recent systematic review of the literature addressed the long-term prognosis of patients with D+ HUS (58). Data on 3476 patients (aged 1 month to 18 years at recruitment) from 49 studies published between 1950 and 2001 with a mean follow-up of 4.4 years revealed a 12% incidence of death or permanent ESRD (death: 9%, ESRD, 3%) and a 25% incidence of long-term renal sequelae. However, the long-term renal sequel, as characterized by the incidence of glomerular filtration rate lower than 80 mL/min per 1.73 m2, hypertension, and proteinuria, was extremely variable, ranging from 0% to 64% in various studies. The higher severity of acute illness, particularly CNS symptoms and the need for initial dialysis, was strongly associated with worse long-term prognosis.
In general, adults fare less well than children. A survey of cases of HUS in adults until 1973 revealed that 73% died of renal failure or had terminal renal failure requiring dialysis (72). In more recent studies, the mortality rates in adult HUS/TTP were still relatively high, with an overall mortality of 30% or lower (11,,188,18912). Eleven patients had a prodrome of colitis, and in seven of these patients, VTEC infection was detected. In 10 patients, the HUS was related to underlying diseases, such as SLE or cancer. The remaining 16 patients had neither diarrheal illness nor an underlying disorder. The mortality rate was 30%, 14% developed chronic renal failure, and 24% had recurrent episodes. None of the patients with
colitis died, and no recurrences were noted in the â€œcolitisâ€ group. Those patients who developed HUS without colitis had a less favorable survival and experienced recurrences. The findings of this study prompted the conclusion that the disease or trigger (such as infection) that gives rise to HUS/TTP is the most important determinant of prognosis. These investigators also noted that most of these patients also had symptoms of TTP, and therefore adult HUS cannot be clearly distinguished from TTP. Similarly, a wide variety of causes were encountered for HUS in 55 adult patients seen between 1990 and 1998 in France (189). In patients with the â€œprimary formsâ€ of HUS (i.e., HUS without underlying diseases or conditions such as HIV infection; malignancy; kidney, liver, or bone marrow transplantation; or nephropathies), the relatively low mortality rate (13.3%) and excellent renal outcome (73.3% of patients recovering normal renal function) was attributed to improved therapy (i.e., plasma infusions).
Relation of Pathologic Picture to Clinical Findings and Prognosis
A comprehensive study of HUS in childhood and infancy was carried out by Habib et al (64). Pathologic changes were correlated with the clinical picture and outcome. Three pathologic categories were distinguished: cortical necrosis, predominantly glomerular involvement, and predominantly arterial involvement. The prognosis was best in the group with predominantly glomerular involvement, and it was by far the worst in the group with predominantly arterial involvement. The type with predominantly arterial involvement tended to occur in children, as opposed to infants. This is also the type that is common in adults.
Another study, by Thoenes and John (9), also showed that the histologic changes of HUS could be assorted into one of three categories. In the first group, the glomeruli showed the classic changes of HUS, accompanied by relatively mild arteriolar changes. This form was most frequent in young children. In the second group, severe obliterative changes predominated in the interlobular arteries. Arterioles showed fibrinoid changes, and glomeruli revealed thickening and wrinkling of capillary walls and occasional fibrinoid necrosis. This form was most common in a heterogeneous group of adults with the clinical features of malignant hypertension, postpartum renal failure, acute renal failure, and acute or subacute glomerulonephritis. The third group, which showed mixed glomerular and arterial lesions, was usually seen in older children. A study by Morel-Maroger et al (13) confirmed the importance of vascular lesions in a large series of HUS in adults. The presence of arterial occlusion was a harbinger of poor prognosis. In a more recent clinicopathologic study of 34 pediatric patients with D-HUS, renal arteriolar and/or arterial changes were identified in 15 patients (171); the presence of arterial disease correlated with a poor prognosis. The study emphasized the prognostic value of renal biopsy in D-HUS.
The impact of the renal biopsy findings on the long-term outcome of classic (typical) childhood HUS was addressed in a study from the Necker-Enfants-Malades Hospital in Paris, France (190). Twenty-nine patients were followed for 15 to 25 years, and the long-term outcome showed a close correlation with the original biopsy findings. Ten of 11 patients with cortical necrosis developed either ESRD (n = 4), chronic renal failure (n = 3), or significant residual renal functional abnormalities (n = 3). The nine patients who had TMA involving less than 50% of the glomeruli were symptom-free or had only mild renal sequelae.
However, a report from the Southwest Pediatric Nephrology Study Group (16) raised some doubt about the prognostic impact of renal biopsies from samples taken soon after the onset of the disease. In this study of children, no arterial lesions were identified in renal tissues obtained during the first 16 days of hospitalization, as opposed to those biopsies from samples taken after 16 or more days of hospitalization. The explanation for the apparent absence of arterial lesions during the first 16 days was that arterial changes may evolve slowly over time and may not be recognizable at the light microscopic level in the early stages of the disease. Moreover, sampling error could be a significant factor in the interpretation of renal biopsy specimens from patients with HUS. Furthermore, only the most severely affected patients undergo biopsy, and more than 90% of patients who do not undergo biopsy seem to be cases that spontaneously resolve.
In summary, the presence of renal arterial occlusive lesions determines the outcome in a given patient. As a rule, extensive glomerular changes in themselves may not necessarily signify a poor prognosis, as long as the changes are confined to the glomeruli and are accompanied by only minor arteriolar changes. Exceptions to this rule exist, and extensive cortical necrosis may be seen in the presence of widespread glomerular lesions but with sparing of the arteries. Information on the sequence of changes leading to chronic renal failure is scanty.
Thrombotic Thrombocytopenic Purpura
The disease we call thrombotic thrombocytopenic purpura today was first described in 1925 (1) and was given its present name by Singer et al in 1947 (3). The classic pentad of symptoms includes fever, hemolytic anemia, thrombocytopenia, neurologic involvement, and renal manifestations (7). However, many patients lack one or more of these signs or symptoms. In a series of adults with TTP, the classic pentad of symptoms was present in only 34% of patients (157). In an earlier study, anemia, purpura, and neurologic abnormalities developed in about 75% of patients (8).
The peak incidence of TTP is in the third decade of life, but the disease has been described in infancy and in extreme old age (8). The anemia is of the microangiopathic hemolytic type, and the thrombocytopenia is manifested as petechiae and ecchymoses of the skin and hemorrhages in various viscera. Lactate dehydrogenase elevation in the serum is a common feature. Most of the clinical manifestations of TTP are the consequence of widespread thrombotic lesions in various organs. The symptoms vary according to the number and distribution of thromboses. Neurologic manifestations include headache, altered consciousness, paresis, aphasia, syncope, seizures, visual changes, and dysarthria. Neurologic abnormalities may change rapidly and may be remittent. The same thrombotic involvement of the microvasculature is responsible for renal changes and their clinical sequelae, and these are considered in the next section. Severe abdominal pain is the major presenting symptom in many patients. Thrombotic thrombocytopenic purpura may pursue a rapidly fatal course, with only approximately 10% of patients surviving in an early review (7). In a later review of the disease (8), however, 46% of patients survived. Prognosis has significantly improved in recent years, and reports show survival in up to 80% of patients (188). The reasons for this improvement are multiple and include not only earlier detection and recognition of milder forms, but also the introduction of more effective forms of therapy. Treatment with exchange plasmapheresis and antiplatelet agents has proved beneficial (18). Relapses after initial response to treatment are relatively frequent (chronic or relapsing forms of TTP).
A report from the Centers for Disease Control for Prevention estimated the annual incidence of TTP to be 3.7 per 1,000,000, based on an analysis of U.S. death certificates filed between 1968 and 1991 (191). A more recent study by Miller et al (55) using a computerized database from a large health insurer showed that the standardized incidence rate of TTP in the United States (3.8 per 1,000,000) for the years of 1990 to 2000 was similar to that previously reported by TÃ¶rÃ¶k et al (191). However, a more accurate estimate of the incidence of all patients with clinically suspected TTP/HUS based upon the Oklahoma TTP/HUS Registry Database (17) showed an approximately threefold greater incidence (11.29 per 1,000,000) than the two previous reports on the incidence of TTP (55,191). However, the standardized incidence rate of patients with idiopathic TTP/HUS using the Oklahoma Database was similar (4.46 per 1,000,000) to the earlier estimates (55,191).
A review of 196 published cases of TTP noted a 45% frequency of renal involvement, as defined by serum creatinine levels above 2 mg/100 mL and BUN levels above 30 mg/100 mL; however, only 12% of patients had creatinine levels over 5 mg/100 mL and BUN levels above 50 mg/100 mL (8). In a literature review, Eknoyan and Riggs (15) noted that the clinical outcome in TTP was significantly worse in patients with BUN levels above 60 mg/ 100 mL. In their own 15 patients with TTP, these investigators recorded that all had a positive urine heme test (14 with red blood cells in the urine), but in only 2 patients was renal failure the dominant clinical feature. These investigators also commented that the various clinical renal abnormalities do not necessarily depend on microvascular thrombotic lesions in the kidney; such clinical features as jaundice, hemoglobinuria, and volume depletion may all play a role in renal malfunction. Nonetheless, the important point is that severe renal damage was uncommon in this series of TTP patients. Reports of cases with renal lesions sufficiently severe to cause irrevocable damage to the kidney are extremely sparse (15).
The kidneys may be enlarged, sometimes considerably, and petechial hemorrhages are present under the capsule. The cortex is widened, with petechiae often visible on the cut surface and on the pelvic mucosa.
The microscopic features of TTP in the kidney are those of thrombotic microangiopathy. Glomeruli show microthrombi similar to those seen in arterioles (see next paragraph) (Fig. 16.21); the number of glomeruli affected is usually not great. However, the majority of glomeruli
may be affected. Sometimes, the glomerular capillary walls show focal thickening. Some slight proliferation may also be seen, and sclerotic changes are not uncommon. However, the glomerular changes are usually not as dramatic as in HUS.
One of the most conspicuous features of TTP is the presence of eosinophilic, granular thrombi in terminal renal interlobular arteries, or more commonly, in afferent arterioles. The most common site is the junction of the afferent arteriole and the glomerular tuft, sometimes called the infundibulum or glomerular hilum or hilus (Fig. 16.22). The thrombotic material often merges with the vessel wall, particularly in older lesions. Sometimes both fresh and organizing thrombotic lesions can be recognized. The lumen is seldom completely obliterated. Often the arterioles show severe, almost aneurysmatic dilation (4,192). In addition, the arterioles may show a proliferation of cells assumed to be endothelial cells; this change was first reported by Baehr et al (2) and was later confirmed by several authors. The cellular proliferation in the arterioles is sometimes a prominent feature, and the collections of cells, often concentrically arranged, may attain the size of glomeruli. Capillary channels with or without edematous extracellular matrix can sometimes be recognized. Because these proliferations resemble glomeruli, they are referred to as glomera or glomeruloid structures (193). Glomeruloid structures are not confined to TTP and have been seen in HUS, SLE, and various forms of glomerulonephritis (193). Lipid-containing macrophages may be present in the intima of small renal arteries undergoing proliferative changes. Cellular intimal thickening of interlobular arteries is found in the few cases with severe renal failure. Subendothelial hyalin deposits can be seen in association with the thrombotic lesions.
Tubular changes include iron pigment and hyalin droplets in the cytoplasm of the proximal convoluted segment, with varying degrees of tubular loss. Chronic inflammatory cells may be seen in the interstitium, which shows fibrosis proportional to the amount of tubular loss.
An early study by Craig and Gitlin (194) showed that the thrombi reacted with antibody to fibrinogen, but not with antibody against platelets. The conclusion was that the thrombi were composed of a derivative of fibrinogen or fibrin. Feldman et al (195) performed studies in three patients using renal biopsy material and spleen removed surgically. Fibrin and, to a lesser extent, IgG were deposited in the subendothelial layer of many renal medium-sized and small arteries, arterioles, and veins. Fibrin and IgG also were detected in the walls of the glomerular capillaries and in the mesangium. The thrombi in the lumen of arterioles contained fibrin, but not IgG. No C3 could be demonstrated. In an account of a single renal biopsy, arterioles were described as not showing IgG or C3; no fibrin thrombi were identified (196).
On autopsy material from 23 patients with TTP, Asada et al (197) showed strong immunohistochemical staining in the renal vascular thrombi as well as in the subendothelial hyalin deposits with factor VIIIâ€“related antigen but only weak staining with fibrinogen or fibrin. This staining pattern was in contrast to the immunohistochemical findings of thrombi in patients with disseminated intravascular coagulation, in which strong fibrinogen or fibrin and weak factor VIIIâ€“related antigen staining was observed. Electron microscopic analysis showed numerous platelets in the glomerular capillary thrombi. The authors concluded that thrombi in TTP are composed of platelets. The strong subendothelial factor VIIIâ€“related antigen positivity was interpreted by these investigators as suggesting that the hyalin deposits may not be the result of increased vascular permeability, but they may be platelet thrombi incorporated into the vascular wall.
The electron microscopic features of TTP are similar to those described earlier for HUS.
Changes in Other Organs
One of the features of TTP is the widespread distribution of microthrombi in arterioles throughout the body, particularly in the brain, myocardium, adrenals, pancreas, intestinal tract, and spleen (7,198). Ischemic changes may be seen in relation to these lesions in various tissues. The thrombotic vessels may show remarkably few changes, although endothelial proliferation similar to that seen in the renal arteries may be present.
Etiology and Pathogenesis of Hemolytic Uremic Syndrome and Thrombotic Thrombocytopenic Purpura
The basic pathologic abnormalities of both HUS and TTP are endothelial damage, intimal swelling, coagulation both within the vessel wall and the vascular lumen, and the sequelae of vascular occlusion. Endothelial damage characterized by swelling, cytolysis, and detachment of the endothelium from basement membranes is seen in the glomeruli, arteries, and arterioles. Edema in the intima of small arteries and arterioles and the subendothelial zone of the glomerular capillaries, which suggests increased permeability of the endothelial cells, is one of the most typical morphologic findings. The presence of fibrin in blood vessel walls is probably related to increased permeability, although this may also represent an intravascular fibrin thrombus that has been incorporated into the vessel wall. Coagulation may be a direct consequence of endothelial damage. Platelets and fibrin are found in glomeruli and small arteries and arterioles. Fibrin thrombi can also be seen in the glomerular capillaries and in the lumina of renal arterioles and small arteries. In addition, a particulate or fluffy, electron-dense material is present in the widened subendothelial region of the glomerular capillary wall. The exact nature of the subendothelial material is not known, but it is clearly related to fibrinogen (or fibrin) or other coagulation proteins and probably some matrix proteins (e.g., fibronectin). In TTP, microthrombi are widely disseminated throughout the entire microvasculature.
Endothelial damage appears to be a crucial feature that precedes the development of additional vascular lesions. Injury to endothelium may result in switch of endothelial anticoagulant phenotype to procoagulant phenotype, decreased fibrinolytic activity, exposure of thrombogenic subendothelial surfaces, platelet activation, imbalance between prostacyclin (PGI2) and thromboxane A2 (TXA) production, release of unusually large vWF multimers, and cellular activation with up-regulation of adhesion molecules, chemokines, cytokines, and transcription factors (199,200,201,202,203,204,205). All of these factors may contribute to platelet aggregation, thrombus formation, and impaired removal of fibrin with subsequent severe vascular and organ damage. Although the pathogenic pathways of the endothelial injury for some of the causative agents of HUS/TTP have been well delineated, for most of the proposed causative agents the exact mechanism by which the endothelium is injured is still not fully understood. Some factors thought to play a major role in the endothelial injury in HUS/TTP are discussed.
Structurally, both verotoxins (VT) are composed of an enzymatically active 32-kD A subunit and five 7.7-kD B subunits that allow binding of the toxin to specific globotriaosylceramide (Gb3) cell surface receptors on the endothelial cells (206207). Binding to the receptor is followed by internalization of the toxin, with subsequent retrograde transport from the Golgi complex to the endoplasmic reticulum (95). From there, the A subunit is translocated to the cytosol, where it inhibits protein synthesis through depurination of a specific adenosine in 28S ribosomal RNA (77,208). However, there is also evidence that treatment of cultured endothelial cells with sublethal doses of VT leads to endothelial cell activation with up-regulation of a number of chemokines, cytokines, cell adhesion molecules, and transcription factors (199,200). The proinflammatory microenvironment induced by endothelial cell activation sets the stage for inflammatory changes followed by microvascular thrombosis. Such changes have been documented experimentally using human microvascular endothelial cells pretreated with VT1 and exposed to flowing blood at high shear stress (201). Platelet activation and adhesion were followed by thrombosis, a process dependent on endothelial P-selectin and platelet endothelial cell adhesion molecule-1 (PECAM-1) expression. Recent evidence also indicates that VT stimulates the release of unusually large von Willebrand factor (ULvWF) multimers from cultured endothelial cells (209). In addition, cleavage of ULvWF multimers by vWF-Cleaving Protease (ADAMTS13) is significantly impaired in the presence of VT, suggesting a potential role of VT in the development of microvascular thrombosis (209). Verotoxin can also induce apoptosis in various cell types, including human microvascular endothelial cells (210).
Verotoxin 1 and VT2 bind to different epitopes of the Gb3 receptor, with differences in the binding affinity and kinetics. Binding of VT2 to the receptor is slower; however, dissociation from the receptor is also slower than for VT1, allowing longer time for internalization (211). This might explain the significantly higher in vitro toxicity of VT2 on human endothelial cells (212).
Oral infection with VT-producing E. coli (VTEC) results in bacterial adherence to the epithelial cells of colonic mucosa and local destruction of brush border microvilli (â€œattaching and effacingâ€ lesion) (213). The bacterial adherence to the epithelial cells is mediated by intimin, a 97-kD outer membrane protein (214). Verotoxins gain access to systemic circulation via translocation across polarized intestinal epithelial cells, a process facilitated by neutrophil transmigration (215,216). In vitro data indicate that the toxin may induce interleukin-8 expression in the intestinal epithelium (217); local cytokine production may trigger an inflammatory response, which then contributes to the tissue damage and loss of colonic barrier function. Although erythrocytes, platelets, and monocytes can all
bind VTs, recent in vitro and in vivo data seem to support polymorphonuclear leukocytes as the carriers of the toxin within the circulation (218,). Transfer of VT from the transporter (i.e., polymorphonuclear leukocyte) to the endothelial cells is facilitated by the 100-fold higher affinity of the receptor on the endothelial cells (218).
Experimental data indicate that endothelial cells may be the primary target sites for VTs; human endothelial cells express Gb3 receptors and are susceptible to the cytotoxic action of Shiga toxin in vitro (220). In addition, basal levels of Gb3 receptors were shown to be approximately 50 times higher in renal endothelial cells than in the umbilical endothelial cells (221); differential expression of Gb3 receptors may be responsible for preferential involvement of the kidney in HUS. In vivo studies also suggest that the organ distribution of VT receptors determines the localization of microvascular lesions in rabbits injected with VT1 (222). However, cells other than endothelial cells, such as renal tubular epithelial cells expressing high levels of Gb3, may also be targeted by VTs and may contribute to tissue injury (223).
However, it is clear that factors other than VTs may also play a significant role leading to endothelial injury in the classic form of HUS. In vitro synergism of cytotoxic potential has been described between lipopolysaccharide (LPS, an endotoxin) and VTs and also between tumor necrosis factor-alpha (TNFÎ±) and VTs (224,225). Globotriaosylceramide 3 receptor expression on human endothelial cells can be induced by proinflammatory cytokines such as TNF and interleukin-1 (226). In vivo, Shiga toxin was found to sensitize mice to the lethal effect of both LPS and TNF; Shiga toxin also induced TNF synthesis within the kidney, an effect that may contribute to renal injury (227). Endotoxemia has been described in HUS (93,228), and specific antibodies to the LPS of VT-producing organisms have been detected in patients with D+ HUS (229). In addition, Heyderman et al () described decreased levels of core antibodies to LPS in patients with severe D+ HUS (HUS patients presenting with diarrhea). These investigators proposed that, in severe D+ HUS, enteric inflammation results in the dissemination of LPS into the systemic circulation, consumption of endotoxin core antibodies, activation of inflammatory cells, and disruption of endothelial function.
Hemolytic uremic syndrome may be associated with neuraminidase-producing organisms (22,37,50,51,52,231). Neuraminidase, which is produced by certain bacteria such as Streptococcus pneumoniae and possibly some viruses, has the potential to damage endothelial cells, platelets, and also red blood cells. Neuraminidase acts by removing N-acetylneuraminic acid from membrane surfaces and exposing the hidden T-crypt antigen (Thomsen-Friedenreich antigen) (22). Because most people have preformed circulating IgM antibodies to this antigen, the antigen-antibody reaction damages endothelial, red cell, and platelet surfaces (22) and leads to intravascular thrombosis, hemolysis, and thrombocytopenia. Hemolysis caused by the direct action of neuraminidase on the red blood cells provides an explanation for those cases of HUS in which anemia occurs with scant or no evidence of a TMA (9,20).
Anti-Endothelial Cell Antibodies
Anti-endothelial cell antibodies provide a further mechanism by which damage to endothelial cells can take place. Koenig et al (232) detected autoantibodies to cryptic antigens using Western blot analysis of cellular proteins derived from human renal microvascular endothelial cells in 13 of 14 patients with TTP and in 4 of 5 patients with HUS. A possible explanation of this finding is that endothelial injury may expose cryptic endothelial antigens to immune recognition, which results in subsequent autoantibody production. However, another study found anti-endothelial cell antibodies in only a small proportion of patients with TTP, bringing into question the significance of these antibodies in the pathogenesis (233).
Polymorphonuclear Leukocyte Activation
Activation of polymorphonuclear leukocytes may play a role in the pathogenesis of HUS in childhood. A high polymorphonuclear leukocyte count at presentation in D+ HUS indicates a poor prognosis (). Significantly elevated circulating levels of neutrophil activator interleukin-8 were shown to correlate with both serum Î±1-antitrypsinâ€“complexed elastase (a marker of neutrophil activation) and, in D+ HUS, the circulating polymorphonuclear count (234). Activation of neutrophils can also take place upon binding VTs, as suggested by observations both in vitro and in vivo (218,219).
A possible role of polymorphonuclear leukocytes in the pathogenesis of HUS can be explained by their ability to damage endothelial cells. Activated neutrophils can generate superoxide anions (O2-) that, when combined with endothelial-derived nitric oxide (NO), form the highly cytotoxic hydroxyl radical. Evidence indicates that, in the acute stage of recurrent forms of thrombotic microangiopathy, NO formation is increased in association with signs of lipid peroxidation (235). Increased lipid peroxidation is likely the consequence of the interaction of NO with neutrophil-derived oxygen radicals. NO, a potent vasodilator, can also be toxic to endothelial cells at higher than physiologic concentration (236). The synthesis and release of NO by the endothelial cells are increased by increased laminar fluid shear stress (a traction force of flow on the endothelial cells) (237); this observation is relevant to HUS/TTP because of the increased shear stress in the
narrowed vessels in HUS/TTP. In addition, NO may activate peripheral blood mononuclear cells and may promote release of TNFÎ± (238), further amplifying the inflammatory process. There is strong in vitro evidence indicating that activated polymorphonuclear leukocytes may contribute to endothelial injury in D+ HUS. Polymorphonuclear leukocytes stimulated by VTs induce apoptosis in co-cultured endothelial cells (203). Also, neutrophils isolated from children with D+ HUS are more adherent to human umbilical vein endothelial cells (HUVEC) than control neutrophils (239).
Platelet Activation or Aggregation
Platelet activation or aggregation is considered to be a major feature in the pathogenesis of both HUS and TTP. Thrombocytopenia is an almost universal feature in HUS and TTP, and serum concentrations of markers of platelet activation (Î²-thromboglobulin, platelet factor 4) are high (240). Thrombocytopenia is caused by consumption of activated platelets in microthrombi and mechanical destruction in peripheral, damaged microvessels. Platelet activation is initiated by endothelial damage and thrombin generation, which is followed by platelet aggregation at the site of injury.
2, the main product of platelet cyclooxygenase, is vasoconstrictive and has potent platelet-aggregating properties. An imbalance between PGI2 and TXA2 may result in platelet activation with its various consequences. Some data suggest that decreased production of PGI2 by damaged endothelial cells in HUS may play a role in platelet activation. Low plasma levels of 6-keto-prostaglandin F1Î± (the stable metabolite of PGI2) have been noted in patients with HUS (). Noris et al (205) demonstrated defective renal and normal systemic biosynthesis of PGI2 in children with acute HUS.
Investigators have also identified specific platelet-aggregating factors that could be responsible for the platelet thrombi of HUS and TTP. Such â€œspecificâ€ platelet-aggregating factors have been demonstrated in the blood in both TTP (241) and HUS (242). A cysteine proteinase (calpain) with platelet-aggregating activity that acts in the presence of vWF was shown in the sera of patients with TTP (243,244). Another 37-kD platelet-agglutinating protein known as protein p37 was identified from patients with TTP; p37 forms complexes with platelet membrane glycoprotein IV (CD36) (245). Because CD36 is expressed by both endothelial cells and platelets, these investigators hypothesized that p37â€“CD36 binding on the endothelial cells is the cause of endothelial injury and also mediates platelet/endothelial interaction. Anti-CD36 antibodies were also found in patients with TTP; such antibodies may contribute to vascular damage by a similar mechanism (246). Platelet-activating factor, a phospholipid produced by platelets and endothelial, mesangial, and other cells, promotes platelet aggregation. Significantly elevated urinary concentrations of platelet-activating factor were found in a few patients with acute-phase HUS (247). However, whether platelet-activating factor causes the disease or is a marker of endothelial injury is not known. Platelet-aggregating activity can also be detected when normal plasma is treated with either VTs or neuraminidase (248).
128,). Crossreactivity of these antiplatelet antibodies was also shown with granulocytes, T lymphocytes, and red blood cells (128,129). Thrombocytopenia in these patients is apparently related to antiplatelet antibodies; however, it is not clear how endothelial damage is initiated.
In addition to their role in thrombogenesis, activated or aggregated platelets accumulated in small vessels may trigger intimal proliferation and fibrosis by releasing various growth factors such as platelet-derived growth factor (PDGF) and transforming growth factor-Î² (TGF-Î²) (249).
Although intravascular coagulation is an important feature of HUS and TTP, no apparent clinical or laboratory signs of disseminated intravascular coagulation are usually present. The lack of disseminated intravascular coagulation in HUS and TTP may be caused by an organ-limited activation of coagulation (i.e., kidney in HUS) or by low-grade coagulation. Thrombotic changes in HUS and TTP can be explained on the basis of an imbalance between fibrin formation and its removal. A defect in fibrin removal has been shown in human HUS by the demonstration of a plasma inhibitor of glomerular fibrinolysis. The circulating inhibitor of fibrinolysis in HUS has been identified as plasminogen-activator inhibitor type 1 (PAI-1); normalization of plasma PAI-1 levels in children with HUS correlated with improvement in renal function (250). In a study from Japan, both serum PAI-1 and tissue-type plasminogen activator levels were shown to be elevated in patients with TTP, a finding suggesting the coexistence of hypercoagulable and hyperfibrinolytic states (251). Up-regulation of PAI-1 and a membrane receptor for urokinase-type plasminogen activator were shown in the glomeruli and arterial walls of renal biopsies from patients with HUS/TTP (252). Tissue distribution of tissue-type plasminogen activator was also determined in this study and was similar to that seen in normal kidneys; that is, it was present in the glomeruli and vascular endothelial cells. The results of this study suggest that up-regulation of PAI-1 may play a role in the defect in fibrin removal in thrombotic microangiopathies. In a recent in vivo study of children with D+ HUS, serum levels of PAI-1 were elevated, indicating inhibition of fibrinolysis (253). Additional laboratory features of altered coagulation
that had preceded the clinical onset of HUS included elevations in plasma concentrations of prothrombin fragments 1+2, tissue plasminogen activator (t-PA) antigen, t-PA/PAI-1 complex, and D-dimer levels. The authors concluded that in HUS, thrombin generation and inhibition of fibrinolysis precede and may also be the cause of renal injury.
Additional plasma constituents with a possible role in the abnormal coagulation process were also evaluated. The concentrations of tissue factor pathway inhibitor were significantly decreased, and concentrations of thrombomodulin (cofactor of anticoagulant protein C on the surface of the endothelial cells) were significantly increased in patients with an acute stage of TTP (254). Although increased thrombomodulin levels may simply reflect endothelial injury, the significance of these findings should be further elucidated.
von Willebrand Factor and ADAMTS13 Abnormalities
Endothelial cells are the primary source of circulating vWF multimers. However, vWF multimers that are produced by the endothelial cells are larger than those found in normal plasma. In normal circumstances, vWF-cleaving metalloprotease (ADAMTS13) degrades the secreted multimers by cleaving vWF on the surface of the endothelial cells at the peptide bond between amino acid residues 842Thr and 843Met in the A2 domain of the subunit (255). Genetic or acquired deficiencies of ADAMTS13 activity result in deficient cleavage and persistence of ULvWF multimers (i.e., multimeric forms of vWF, larger than that in normal plasma) in circulation. Genetic deficiencies are linked to homozygous or compound heterozygous mutations of the ADAMTS13 gene, whereas the more common acquired deficiencies are caused by circulating inhibitory autoantibodies against ADAMTS13 (256,257,258,259,260). Thrombotic thrombocytopenic purpura in patients who have hereditary ADAMTS13 deficiency is often referred to as congenital TTP.
Relationship between clinical disease phenotype and deficient ADAMTS13 activity has been addressed in a number of studies in patients with TTP and HUS (257,258,261,262,263,264,). Some studies reported nearly perfect correlation between TTP and severe deficiency in ADAMTS13 activity (258,261,262). However, other studies showed that a significant proportion of patients with the clinical diagnosis of HUS (typical, atypical, and familial) can also have severe ADAMTS13 deficiency (257,263,264,265). Therefore, ADAMTS13 deficiency is not a reliable marker to distinguish TTP from HUS. Still, findings of a recent study suggest that severe ADAMTS13 deficiency defines a subset of patients among TMA patients with distinct laboratory and clinical profiles (266). Patients with severe ADAMTS13 deficiency and TMA showed more relapses and significantly lower platelet count, medium hematocrit levels, and serum creatinine versus those patients who had TMA but did not have severe ADAMTS13 deficiency (266). In contrast, neither the presenting features nor the clinical outcomes were different in another study that compared patients with idiopathic TTP/HUS who had severe ADAMTS13 deficiency with those who did not (74). Many patients in this study in both categories responded to plasma exchange treatment, a clear indication that severe ADAMTS13 deficiency cannot be used as a basis for withholding plasma exchange treatment (74). In addition, contrary to the common assumption that most patients with idiopathic TTP/HUS have severe ADAMTS13 deficiency, only 33% of patients of those with idiopathic TTP/HUS (i.e., TTP/HUS not associated with hematopoietic stem cell transplantation, pregnancy, drugs, or bloody diarrhea) in this study showed severe deficiency (74). Interestingly, the relative incidence rate for TTP/HUS associated with severe acquired ADAMTS13 deficiency is nine times greater in Blacks than in non-Blacks (), and among those with acquired ADAMTS13 deficiency, the frequency of obesity is also significantly increased (74). Some recent data also indicate a possible role of functional impairment of ADAMTS13 in the pathogenesis of D+ (Shiga toxinâ€“mediated) HUS (209).
Complex abnormalities in circulating vWF multimeric patterns have been described in patients with TTP and HUS. In some patients with single-episode TTP, ULvWF multimers have been found in plasma during the TTP episodes. In chronic-relapsing TTP, ULvWF multimers were found in the plasma during clinical remission but not during relapses (267). Decreased number of largest vWF multimers (i.e., largest normal multimers) also occurs in patients with either single-episode TTP or chronic-relapsing TTP. Investigators have also shown in HUS that the largest plasma vWF multimer forms are reduced in the acute phase and return to normal along with normalization of thrombocytopenia as patients recover (268267). Enhanced vWF proteolysis may also contribute to uncontrolled platelet aggregation in HUS and TTP. Von Willebrand factor digested by calpain (calcium-dependent cysteine protease found in the sera of patients with acute TTP) is highly reactive with activated platelets, binds to platelet glycoproteins IIb and IIIa, and participates in formation of the platelet aggregates (269). High shear stress rates favor the interaction of large vWF multimers with platelet glycoprotein Ib-IX-V receptors that, in turn, activates platelets and promotes glycoprotein IIb-IIIa receptorâ€“dependent platelet aggregation and formation of platelet thrombi in the microcirculation (270).
Decreased levels of serum complement in HUS are not uncommon. However, many patients experience only
transient depression of serum complement levels during the course of their disease (271). In other patients, hypocomplementemia may be persistent and may predispose to the development of HUS. Such cases with persistent hypocomplementemia have been described in sporadic as well as familial and recurrent forms of HUS in association with hereditary or acquired disorders of the complement system (45,109,272,273). However, there are also patients with various genetic abnormalities of the complement regulatory proteins that develop atypical HUS, but their C3 levels remain normal (108,274). Features of atypical HUS with complement abnormalities include a severe form of renal disease.
Carreras et al (109) described a familial relapsing type of HUS involving a 32-year-old woman, her daughter, and a nephew. All three patients had persistent hypocomplementemia. Thrombotic microangiopathy was present on renal biopsy. When 17 members of this family were studied, low serum levels of C3, from activation of the alternative pathway, were found. Subsequently, an association between factor H deficiency, a multifunctional regulatory protein of the alternative complement pathway, and HUS was recognized (272,273,275,276,277). Low levels of factor H owing to homozygous or heterozygous deficiency may result in alternative pathwayâ€“mediated complement consumption with low plasma levels of C3. In recent years, a search for genetic changes identified a variety of mutations of the factor H gene (CFH) in approximately 20% of both sporadic and familial cases of HUS (40,,42,43,44,45). In familial cases, both autosomal recessive and autosomal dominant inheritance has been described (107,274). The molecular abnormalities of the factor H deficiencies are polymorphous; however, the mutations seem to cluster in the C terminal region of the molecule. This area of the molecule is known to be important for both binding to C3 and binding to polyanions, suggesting a particular role of the C terminal domain in the pathophysiology of HUS (45,274). Factor H gene mutations may have prognostic ramifications (107). Atypical HUS in patients with CFH mutations was associated with higher mortality, and the disease manifested earlier than in those patients without CFH mutations (107). In addition, all kidney transplants failed owing to recurrences in patients with CFH mutations (107). However, another study showed that the progression of HUS was variable in patients with homozygous or heterozygous CFH mutations (45).
In addition to genetically mediated deficiencies, an acquired functional deficiency of factor H has also been described recently (110). Anti-factor H antibodies were detected in the plasma of three children who presented with the recurrent form of HUS. Although the plasma factor H antigenic levels were normal and no mutations of the gene were identified, the plasma factor H activity was decreased, indicating that the presence of anti-factor H antibodies led to an acquired functional factor H deficiency. Although all three patients had antinuclear antibodies, none of them had any indication of a systemic autoimmune disease such as SLE.
Two additional members of the regulatory proteins of the alternative complement pathway have also been implicated in the pathogenesis of HUS. Deficiencies of these proteins may lead to alternative pathwayâ€“mediated complement consumption similar to that seen in association with factor H deficiencies. Mutations of the membrane cofactor protein (MCP; CD46) were reported in four families with familial forms of HUS (278,279). Factor I gene mutations were identified in a few patients who presented with the sporadic form of atypical HUS with recurrences (,280). In two cases, a non-sense mutation was associated with heterozygous factor I deficiency. In the third case, a heterozygous mutation resulted in functional factor I deficiency. The mutations were also identified in asymptomatic family members of two patients with HUS.
Reduced penetrance and variable inheritance are also characteristic features of atypical HUS with mutations in the CFH or MCP genes (278,281). In two large independent cohorts with 152 patients, the penetrance of disease phenotype was reported to be approximately 50% (281). A possible explanation for the incomplete penetrance is that polymorphic changes in complement regulatory proteins may act as modifiers. Specific single nucleotide polymorphism (SNP) variants of both CFH and MCP genes were shown to be associated with atypical HUS with concurrent mutations of the same genes (107,281). Caprioli et al (107) showed that the association between atypical HUS and specific SNP variants of the CFH gene was also present in patients without CFH mutations. In one of the two cohorts of patients included in the study of Fremeaux-Bacchi et al (281), the association between atypical HUS and specific SNP variants of the CFH gene was only present in those patients without known mutations in CFH, MCP, and factor 1 genes. However, in the other cohort of patients from the same study, an association between atypical HUS and specific SNP variants of the CFH gene was only seen in those patients known to have a mutation.
Some patients may have complex genetic abnormalities. In two siblings described by Noris et al (106) who presented with recurrent familial TMA (one with exclusive neurologic symptoms and the other with severe renal involvement), both patients had severe ADAMTS13 deficiency as a result of two heterozygous mutations. In addition, a heterozygous mutation of factor H was found in the patient who developed chronic renal failure but not in her sister, who presented with exclusive neurologic symptoms.
In summary, decreased levels or functional deficiencies of some of the complement regulatory proteins of the alternative pathway are associated with HUS in a subset of patients with the atypical form of the disease. Currently, these heterogeneous abnormalities, both genetically determined and acquired, resulting primarily from the dysregulation of the alternative complement pathway are considered to be predisposing factors for atypical HUS. Precise molecular
mechanisms leading to HUS in these patients have not been delineated. Steady complement activation on the surface of the endothelial cells may result in endothelial cell activation and/or injury, predisposing them to further injury that would lead to HUS in some patients. However, some individuals with the predisposing genetic mutation(s) remain asymptomatic, emphasizing the role of additional extrinsic and/or intrinsic factors in the development of the disease. The familial forms with complement abnormalities reveal genetic heterogeneity, and a significant proportion of the sporadic forms also have genetic abnormality. However, more than 50% of patients with the familial forms of atypical HUS, including those with decreased complement levels, still remain without identified etiologic factors.
Antiphospholipid antibodies (i.e., anticardiolipin antibodies and lupus anticoagulants), may prolong phospholipid dependent coagulation times, but paradoxically they seem to be associated with an increased risk of venous and arterial thrombosis (282). Antiphospholipid antibodies have also been described in association with microvascular thrombosis in various clinical conditions such as SLE, lupus-like syndrome, primary antiphospholipid antibody syndrome, and systemic sclerosis (283,284â€“286). A recent systematic review of the literature identified 46 patients between 1983 and 2002 with antiphospholipid antibodies and thrombotic microangiopathic hemolytic anemia (286). More than half of the patients had primary antiphospholipid syndrome (61%), one third had SLE (33%), and the remainder had either lupus-like syndrome (4%) or systemic sclerosis (2%). The most common clinical manifestations were those of HUS (26%), catastrophic antiphospholipid syndrome (23%), postpartum and pregnancy-related acute renal failure (13%), malignant hypertension (13%), and TTP (13%). Morphologic findings that were available from 32 patients indicated fibrin thrombi in the glomerular capillaries, arterioles, and interlobular arteries in 75%, 46%, and 33% of cases, respectively.
287) described TMA as one of the characteristic renal manifestations of primary antiphospholipid syndrome. Other significant vascular changes included fibrous intimal hyperplasia and arterial and arteriolar fibrous and fibrocellular occlusions, which along with TMA were named antiphospholipid syndrome nephropathy (APSN). In a subsequent study from the same group of researchers that included 114 patients with SLE, APSN was identified in 32% of patients who had renal biopsies (288). Thrombotic microangiopathy (as characterized by fibrin deposits demonstrated by immunofluorescence) was present in 18% of patients. Antiphospholipid syndrome nephropathy was statistically associated with lupus anticoagulant but not with anticardiolipin antibodies. In addition, APSN was shown to be an independent risk factor for more severely altered renal function and more severe interstitial fibrosis.
The relationship between antiphospholipid antibodies and the renal changes of SLE is incompletely understood. Farrugia et al (289) reported that the renal morphologic changes in most patients who had lupus anticoagulantâ€“positive SLE and renal dysfunction were indistinguishable from those observed in patients with lupus anticoagulantâ€“negative SLE and renal involvement. However, some of the lupus anticoagulantâ€“positive patients had renal TMA, which was accompanied by a worse prognosis. In contrast, Bhandari et al (290291124,292).
Although there are a number of proposed pathogenetic mechanisms that link antiphospholipid antibodies to thrombosis, the precise mechanisms leading to thrombosis are not known. The pathogenic role of antiphospholipid antibodies in TMAs is also uncertain. Although some of the data suggest that antiphospholipid antibody positivity confers an increased risk of TMA, the role of antiphospholipid antibodies in this process is not proven. Further aspects of SLE-related thrombotic events are considered in Chapter 12.
Numerous additional factors, including various drugs (mitomycin, quinine, and cyclosporine), infectious agents, and apoptosis-inducing plasma factor, have been enumerated to cause or to perpetuate endothelial injury in HUS/TTP (54,56,128,129,293).
Animal models that fully recapitulate the entire spectrum of Shiga toxin (Stx)â€“mediated HUS as seen in humans following oral ingestion of bacteria are lacking. In a baboon model developed by Taylor et al (294), the animals showed thrombocytopenia, microangiopathic hemolytic anemia, and glomerular TMA but only with intravenous infusion of Stx 1. It has also been shown that 100 ng/kg of Stx 1 administered as a single bolus dose results in severe HUS, while administering the same total amount of toxin as four 25 ng/kg doses every 12 hours does not (295), unless lipopolysaccharide is administered concurrently (296).
However, intravenous administration of four 25-ng/kg doses of Stx 2 did cause HUS (298). These findings in baboons corroborate epidemiologic observations (31) and cell culture findings (214) regarding higher toxicity of Stx 2.
298), dogs (299), pigs (300), and mice (). Ferrets were shown to develop glomerular lesions mimicking those seen in HUS following oral infection with E. coli O157:H7 (298). Also, oral infection of gnotobiotic pigs with E. coli O157:H7 or O26:H11 resulted in renal TMA with arteriolar and glomerular involvement similar to that seen in humans (300). However, the thrombocytopenia, microangiopathic hemolytic anemia, and renal failure that are typical of human disease were lacking in this model. The role of an extrinsic toxin (i.e., Stx) to trigger TTP in ADAMTS13-deficient mice has also been implicated (301). The findings of this study also indicated that additional genetic susceptibility factors in the setting of ADAMTS13 deficiency are required for the extrinsic toxins to trigger TTP.
The pathogenesis of HUS and TTP is complex, but it appears to be best explained by regarding endothelial damage as the initial event. Predisposing factors for the development of the disease include hereditary and acquired deficiencies of various complement regulatory proteins and the vWF cleaving protease. Endothelial damage may be initiated by various factors, including VT and Stx, neuraminidase, lytic anti-endothelial cell antibodies, apoptosis-inducing factor, cyclosporine, and mitomycin. Endotoxin, proinflammatory cytokines, and polymorphonuclear leukocytes may also contribute to endothelial damage. Changes in the endothelial cell anticoagulant and procoagulant properties, decrease in PGI2 formation, appearance of ULvWF multimers, and exposure of thrombogenic subendothelial surfaces lead to platelet activation and aggregation. In turn, these factors are responsible for the formation of thrombi and cellular proliferation in glomeruli and arterioles. Red blood cell hemolysis is caused by mechanical disruption in traversing fibrin meshwork or by the direct action of neuraminidase and other circulating antibodies. Thrombocytopenia is produced in several ways, which include consumption in thrombi, circulating platelet-aggregating factors, VT, neuraminidase, and vWF.
Systemic Sclerosis (Systemic Scleroderma)
Systemic scleroderma, progressive systemic sclerosis, and systemic sclerosis are three terms used interchangeably to describe this rare systemic connective tissue disease of unknown origin. Systemic sclerosis occurs in two main forms (302). In the diffuse form (also called diffuse cutaneous form), symmetric skin involvement affects both the distal and proximal parts of the extremities and often the trunk and face. Rapid progression of skin changes and early appearance of visceral involvement are characteristic. In the limited form (also called ), a more confined symmetric involvement of the skin affects the distal part of the extremities (often restricted to the fingers) and the face. The progression of the sclerotic process is relatively slow, and visceral manifestations take a much longer time to become manifest in this form. The terms CREST syndrome (calcinosis, Raynaud's phenomenon, esophageal hypomotility, sclerodactyly, and telangiectasia), acrosclerosis, and acroscleroderma were also used for this more limited, and relatively indolent, form of this disease (303). A three-cutaneous-subset classification system distinguishes limited, intermediate, and diffuse forms of the disease (304). According to this classification, the intermediate form is characterized by skin involvement affecting the upper and lower limbs, neck, and face without truncal involvement. The diffuse form is defined by diffuse skin sclerosis. A rare form without skin involvement (â€œsystemic sclerosis sine sclerodermaâ€) also exists (305). â€œOverlapâ€ syndromes, which include sclerodermatomyositis and mixed connective tissue disease, are also included under the general heading of systemic sclerosis (303). The survival of patients with the limited form is much longer, and death is often caused by diseases other than systemic sclerosis. The group with the diffuse form runs the risk of severe renal damage.
Scleroderma renal crisis is the term used to describe the most severe form of renal involvement in systemic sclerosis. This disorder is characterized by rapidly developing acute renal failure that is often accompanied by elevation of the blood pressure, sometimes to â€œmalignantâ€ levels. Proteinuria may be present as part of scleroderma renal crisis, but it may occur independent of the crisis in certain patients. Apart from proteinuria, hypertension and azotemia also occur outside scleroderma renal crisis, and such patients have a more indolent course.
Clinical Presentation and Clinical Course
Systemic sclerosis is a condition that occurs predominantly in women, with an approximate female-to-male ratio of 3 to 1 (303). It may occur at all ages, but most patients are between 30 and 50 years of age at clinical onset (). The disease is rare in children (307). The annual incidence and prevalence of the disease vary largely among different surveys: 0.6 to 19.1 and 126 to 1,500, respectively, per million per year (308). Patients with the diffuse form of systemic sclerosis have thickening and tightening of
the skin as the earliest signs of the disease. The skin involvement starts in the fingers and then spreads to the forearms, upper arms, thighs, abdomen, and upper chest. The involved skin becomes increasingly shiny, taut, and adherent to the underlying subcutaneous tissue, with impairment of mobility of muscles, tendons, and joints. In contrast, in the limited form, skin involvement is restricted to the fingers, hands, and face. Raynaud's phenomenon usually precedes changes in the skin (306). Esophageal hypomotility, pulmonary fibrosis, and Raynaud's phenomenon occur with equal frequency in both diffuse and limited forms, but congestive heart failure and acute renal failure are considerably more common in the diffuse form (306).
A recent study from the European Scleroderma Study Group reported renal involvement (as characterized by proteinuria, hematuria, or increased creatinine) in 52 of 290 patients (18%) with systemic sclerosis at the time of first observation (309). In two large studies published recently, one from the United States and one from Italy, the reported frequencies of scleroderma renal crisis in patients with the diffuse form of the disease were 19% and 12%, respectively (308,310). The mean follow-up was 9.8 years in the study from the United States and 7.1 years in the study from Italy. The frequency of scleroderma renal crisis was 6% in the limited cutaneous form and 7% in the intermediate cutaneous form of the disease in the Italian study (308). Steen et al (311) compared 36 patients with the diffuse form of systemic sclerosis and scleroderma renal crisis with 212 patients who had the diffuse form of disease without renal involvement. The first group had a shorter mean duration of diseaseâ€”2.4 years versus 4.2 yearsâ€”and had digital pitting scars less frequently. No other significant differences in clinical and laboratory findings were noted. Therefore, patients with the diffuse form of systemic sclerosis apparently are most prone to develop scleroderma renal crisis, and acute renal failure occurs in up to 20% of these patients. In addition, scleroderma renal crisis more likely develops in those patients with the diffuse form of systemic sclerosis who have a relatively short history of the disease. Renal damage is rare in the limited form of systemic sclerosis.
In patients with scleroderma renal crisis, headaches, blurring of vision, and dyspnea are characteristic findings and are sometimes accompanied by convulsions. These features are related to the malignant hypertension that often accompanies scleroderma renal crisis. Some patients, however, do not develop hypertension, whereas others may have only modest elevations in blood pressure (312,313,314). Oliguria is a frequent early finding in scleroderma renal crisis. Levels of serum creatinine or BUN are consistently elevated and become relentlessly higher as the renal crisis progresses. Proteinuria and microscopic hematuria may be present, and microangiopathic hemolytic anemia is a feature in some cases (314). Renin levels are elevated in the blood (315), usually in patients with malignant hypertension.
Other patients with systemic sclerosis may have various combinations of proteinuria, azotemia, and modest degrees of hypertension, and some die of nonrenal causes. Patients with isolated proteinuria were shown to have a poorer prognosis than patients without proteinuria (316,317).
Laboratory findings include anemia, hypergammaglobulinemia, positive rheumatoid factor in one fourth, and antinuclear antibodies in up to 90% of patients (303306). The major autoantibody response is composed of anti-topoisomerase I, anticentromere, and anti-RNA polymerase antibodies. Each of these autoantibodies is present in about 25% of patients, and they are mutually exclusive (318). Anticentromeric antibodies are more common in patients with the limited form of systemic sclerosis (CREST syndrome) than in the diffuse form. Anti-topoisomerase I antibody, formerly known as anti-Scl 70 antibody, is much less common in patients with CREST syndrome than in those with the diffuse form of systemic sclerosis (318). An association of anti-topoisomerase I positivity was found with cardiac and renal involvement, as well as with pulmonary fibrosis (319). Neither the anticentromeric antibodies nor the anti-topoisomerase I antibodies are specific for systemic sclerosis and can also occur in patients with primary Raynaud's disease. However, anti-RNA polymerase antibodies appear to be specific for systemic sclerosis and were found in 45% of patients with the diffuse form and in 6% with the limited form of systemic sclerosis (320). In a Japanese study (320), cumulative survival rates for patients with systemic sclerosis at 10 years after diagnosis were 93% in patients with anticentromere antibodies, 66% in those with anti-DNA topoisomerase I, and 30% in those with anti-RNA polymerases. Cardiac and renal involvement was shown to be associated with the occurrence of anti-RNA polymerases (320). However, in spite of their common occurrence in systemic sclerosis, there is no compelling evidence that the antibodies play a direct role in the pathogenesis.
Among various connective tissue diseases, systemic sclerosis shows the poorest prognosis (302,308,321). The mean 10-year survival rate in patients with systemic sclerosis in studies published between 1961 and 1991 was 47.8% Â± 13.3% (308). In more recent studies published between 1991 and 2002, the 10-year survival rate revealed significant improvement, to 72.0% Â± 7.60% (308). In general, the presence of major organ involvement (lung, kidney, heart) is associated with lower survival rates. In a large demographic study from Italy, the worst 10-year
survival rate (34.8%) was observed in patients with scleroderma renal crisis and for a group of patients with simultaneous lung, heart, and kidney involvement (12.6%) (308). Among the patients who died during the mean follow-up of 7.1 years, the most common causes of death were cardiac (36%) and lung involvement (25%), followed by cancer (15%) and renal involvement (12%) (308). The 9-year cumulative survival rate of patients with the diffuse form of systemic sclerosis with severe renal involvement was only 40% in a large study from the United States (310). However, when patients diagnosed with renal crisis before the availability of angiotensin-converting enzyme (ACE) inhibitors were excluded from the analysis, the survival rate was 68% (310308); however, when death occurred within 5 years of disease onset, renal involvement was the most common cause of death (30%) (310).
The increased survival and decreased mortality owing to scleroderma renal crisis is likely related to better treatment(s) available. In a prospective observational cohort study by Steen and Medsger (322), the 5- to 10-year outcomes of 145 patients with scleroderma renal crisis treated with ACE inhibitors were investigated. Sixty-one percent of patients with scleroderma renal crisis had a good outcome that was similar to that of patients who did not have renal crisis. More than half of the patients who initially required replacement dialysis could discontinue dialysis 3 to 18 months later.
Certain patients with systemic sclerosis and combinations of proteinuria, hypertension, and azotemia pursue an indolent course, although they die earlier than those without any evidence of renal disease. Many die of nonrenal causes.
Pathologic changes in the kidney in scleroderma renal crisis represent the acute renal changes of systemic sclerosis, and this form is considered first.
The kidneys are usually of normal or slightly increased size, although some may be slightly or moderately reduced and scarred because of pre-existing arterial narrowing. The subcapsular surface often shows petechial hemorrhages or paler areas (Fig. 16.23). On cut surfaces, the paler areas are seen to be minute, wedge-shaped infarcts, much smaller than the usual infarcts seen in the kidney. Certain kidneys show a mottled, yellowish-red cortex with extensive necrosis. In general, the changes are similar to those found in HUS and, in certain cases, to the malignant phase of essential hypertension.
Glomerular changes may vary considerably (Fig. 16.24). Some glomeruli may show little change, whereas others may be congested and infarcted. In some cases, one sees a thickening of the capillary walls with a double-contour appearance with silver or PAS methods, similar to that seen in HUS. The glomeruli in scleroderma renal crisis may have a fibrillar appearance identical to the changes seen in HUS. Accumulation of glomerular intracapillary eosinophilic material with the staining characteristics of fibrin occurs; these fibrin thrombi are localized to the capillary lumina. In other instances, the eosinophilic areas are larger and are referred to as fibrinoid necrosis. The areas of fibrinoid necrosis may contain nuclear debris and ruptured glomerular capillary basement membranes. Fragmented red blood cells can be identified in glomeruli, mostly when intravascular fibrin is present in glomeruli and small blood vessels. These red blood cell fragments reflect the microangiopathic hemolytic anemia that may be present (314
Arteries and Arterioles
Arteries of interlobular size, smaller arcuate arteries, and afferent arterioles undergo severe changes, but larger arteries
may be normal or show only nonspecific fibrous intimal thickening commensurate with the age of the patient and arteriosclerotic changes elsewhere.
The characteristic changes in interlobular arteries and smaller arcuate arteries consist of mucoid intimal thickening with concentrically arranged myointimal cellular proliferation (â€œonion skinâ€ lesion). The intimal thickening typically produces considerable reduction of the arterial lumen (Figs. 16.25 and 16.26). The mucinous intimal change, which is similar to that seen in severe hypertension, consists of acid mucopolysaccharides of the hyaluronic acid type. It stains with Alcian blue and metachromatically with toluidine blue, but it is only slightly positive with the PAS reaction. Trichrome stain gives either clear reaction or only a weakly blue staining in the thickened mucoid intima, indicating no or little deposition of mature collagen (Fig. 16.26). Often, eosinophilic material is present in the thickened intima of the smaller interlobular arteries (Fig. 16.27). This finding is consistent with fibrin with phosphotungstic acidâ€“hematoxylin and Fraser-Lendrum staining and by immunohistochemical techniques. Fibrin may be found immediately below the endothelial cells or deep in the thickened intima ( and 16.29). It represents infiltration of fibrin across the endothelial barrier, incorporation of an intraluminal thrombus into the vessel wall, or coagulation within the vessel wall. In addition, red blood cells or fragments of red blood cells (schistocytes) may be seen in the vascular lumina and permeating the vessel wall. The endothelial cells may be swollen or focally denuded. Eosinophilic granular thrombi may be present in the lumina of small arteries and arterioles; they were present in all the cases in which microangiopathic hemolytic anemia was described (314). The internal elastic lamina is usually intact, but it may be slightly frayed, with no major breaks or gaps. The media and the adventitia of the affected arteries are usually unremarkable; however, the media may be thinned by stretching around the extended intima, and the adventitia has been noted by some authors as being slightly fibrotic (316).
Fibrinoid necrosis of the afferent arterioles consists of fibrin in the arterial walls, as seen with small arteries, but the extent of the change is much greater, and the entire wall may be involved. The arteriolar wall may have a smudgy appearance because of swelling of endothelial cells and medial myocytes. Inflammatory cells are usually not seen in fibrinoid necrosis, but such cells may occasionally be present. Sometimes, the arteriolar fibrinoid necrosis is accompanied by luminal thrombus, and on occasion, the process continues into the hilar region of the glomerulus.
Using morphometric techniques on autopsy kidneys, Trostle et al () found greater intimal thickening of arterioles and interlobular arteries in patients with diffuse systemic sclerosis and renal crisis when compared with
sex- and age-matched controls. In the group of patients with diffuse systemic sclerosis, those with renal crisis had greater intimal thickening of the vessels than those without renal crisis.
Tubular changes in systemic sclerosis are secondary to vascular changes. Atrophy of the proximal convoluted segment is frequent. Tubules are necrotic in the infarcted areas, and sometimes enclaves of necrotic tubules may be found without necrosis of the nearby glomerulus.
Fig. 16.30). Foci of chronic inflammatory cells may be present in association with scars. In some cases, plasma cells may be conspicuous. Often, the small renal infarcts show dense infiltration with neutrophils. This finding does not imply infection but simply represents the normal cellular reaction at the periphery of an infarct; because the infarcts are small, the neutrophilic infiltration appears extensive.
Various immunofluorescent findings have been reported in systemic sclerosis in the glomeruli, renal interlobular arteries, and arterioles (313,325,326). Fibrin or fibrinogen can be observed along the glomerular capillary walls and sometimes in the mesangium. Glomerular immunostaining with IgM and C3 is more frequently seen than staining with IgG, IgA, C1q, and C4.
Immunostaining of the interlobular arteries and arterioles is often focal; fibrinogen staining is usually
encountered in both arteriolesâ€”when fibrinoid changes are presentâ€”and in the thickened intima of interlobular arteries (Fig. 16.31AFig. 16.31B). Complement factors C3 and C1q are found with the same frequency as IgM, and C4 also has been commonly reported.
Few studies have noted the electron microscopic appearance of the kidney in systemic sclerosis. Glomerular changes similar to those described earlier for HUS may be seen (Fig. 16.32). The broad electron-lucent widening between the endothelium and the lamina densa (i.e., lamina rara interna with electron-lucent â€œfluffâ€) is the most characteristic feature and sometimes contains scattered electron-dense granular or fibrillar material. This widening of the lamina rara interna corresponds to the thickened glomerular capillary walls seen by light microscopy. Thickening and occasional â€œreduplicationâ€ of the glomerular basement membranes and hyperplasia of mesangial cells have been described. In one case, scattered hyalin deposits with a periodicity of 9 to 12 nm were found in a subendothelial position in the glomerular capillaries. Glomerular intracapillary thrombi, if present, are composed of fibrin and platelets frequently admixed with fragmented red blood cells.
A study by Pardo et al (327
stage of its evolution. In the severe, acute form, the thickened intima of interlobular arteries is composed of more or less parallel bands of material resembling basement membrane lying in a translucent, finely granular bed of ground substance (,328,329). Fibrin tactoids and fragmented red blood cells can occasionally be seen embedded in the ground substance. Elongated myointimal or smooth muscle cells are present (328,329), but no discrete â€œimmune-typeâ€ electron-dense deposits corresponding to the positive immunofluorescence findings have been noted (). However, differentiation of â€œimmune-typeâ€ deposits from areas of hyalin insudation may be difficult.
The kidneys of these patients lack distinctive pathologic features and, when seen at autopsy, reveal fibrous intimal arterial thickening with areas of interstitial fibrosis or tubular atrophy. Although vascular changes such as arterial intimal fibroplasia or arteriolar thickening are nonspecific, in some instances the arterial narrowing may be directly related to systemic sclerosis. In an autopsy case-control study of 35 patients with various forms of systemic sclerosis (CREST syndrome [n = 9], diffuse form [n = 26]), Trostle et al (324) showed that even without scleroderma renal crisis, involvement of the renal arteries and arterioles is present. Seventeen patients with the diffuse form of systemic sclerosis who had no acute renal failure had significantly more severe intimal thickening of intrarenal arterioles and small arteries than the controls. A lesser degree of intimal thickening of arteries wider than 325 Âµm was also noted. Therefore, renal vascular changes of systemic sclerosis conceivably could give rise to the renal picture described and could account for most of the clinical features.
In patients with long-standing systemic sclerosis (and also in those patients who survive scleroderma renal crisis), the interlobular and arcuate arteries may show intimal fibrosis with roughly concentric reduplication of the elastic internal lamina, leading to severe narrowing of the lumen. This pattern of chronic vascular change with reduplication of the elastic internal lamina may also be seen in patients with long-standing essential hypertension and in those with the late stage of severe hypertension; this vascular change has been frequently referred to as an onion skin pattern. However, the term onion skin lesion is also used for the earlier stages of arterial changes in TMA
characterized by concentric intimal proliferation of myointimal cells embedded in loose mucoid material (329).
Hypertension and Relation to Vascular Changes
Hypertension is a common feature in systemic sclerosis. Oliver and Cannon (330) found hypertension in 44 (52%) of 84 patients with systemic sclerosis. These investigators pointed out that this percentage was higher than the 25% reported in the earlier paper of Cannon et al (316), and they believe it reflected the high referral rate of patients with renal complications to their unit. Of the 44 hypertensive patients, 11 presented with the abrupt onset of severe hypertension associated with rapid deterioration of renal function; they had a poor prognosisâ€”9 died within 2 months and the other 2 underwent bilateral nephrectomy and long-term dialysis. The remaining 33 patients had mild to moderate hypertension, and only 5 died during the follow-up period, the mean of which was 4.1 months. There was a longer interval between clinical onset of scleroderma and the development of hypertension in these 33 patients. Presumably, the pathologic features of this group would correspond to what was described in the previous section as the â€œchronic form,â€ but knowledge of the natural history and pathology of this disorder is limited.
Other reports have noted an almost invariably high incidence of hypertension in patients with acute renal failure. Traub et al (315) recorded that all but 1 of 65 patients with scleroderma renal crisis had a diastolic blood pressure over 90 mm Hg, and all but 6 had diastolic blood pressure higher than 100 mm Hg. The abrupt onset of severe hypertension also was noted; of 48 patients in whom appropriate data were available, only 14 had preceding hypertension, and this was usually mild.
The temporal relationship between hypertension and morphologic vascular changes (i.e., which comes first) has been the subject of debate since Volhard and Fahr (331). In systemic sclerosis, the relationship of hypertension to vascular disease is certainly an intriguing question. Considerable evidence indicates that the vascular lesions in patients with scleroderma renal crisis may be the cause of the hypertension rather than the consequence. In a study published by Kovalchik et al (325), renal biopsies were performed on nine patients with systemic sclerosis whose blood pressures and levels of serum creatinine were not elevated. Four patients had severe vascular changes, two had mild changes, and three had normal vessels. Plasma renin activity was elevated in three of the four with severe vascular changes, and further elevation of activity took place in response to cold pressor testing in these four patients but not in the others. This study suggests that in systemic sclerosis, the vascular changes may be primary and the cause of the hypertension. Other cases of systemic sclerosis have been described with severe vascular changes in the kidney, usually accompanied by renal failure in the presence of a normal blood pressure (312,313,314).
Hypertension may be explained on the basis of severe arteriolar narrowing, which may stimulate the reninâ€“angiotensin system to produce excessive amounts of renin and initiate hypertension. Those patients with severe hypertension almost invariably have high plasma renin concentrations or high levels of plasma renin activity (279,315316). Although plasma endothelin levels were shown to be significantly elevated in patients with systemic sclerosis, no difference in endothelin levels was found in patients with versus without hypertension (332). It is conceivable that severe hypertension in systemic sclerosis causes further vascular damage, and this concept is supported by the excellent results seen with the use of antihypertensive drugs.
Etiology and Pathogenesis
Altered collagen production, endothelial or vascular abnormalities, and immune factors are considered to be the major players in the pathogenesis of systemic sclerosis (323,333,337,341
Altered Collagen Production
The increased deposition of extracellular matrix is one of the major morphologic features in systemic sclerosis. Changes in the composition of extracellular matrix affect cellular functions such as cell migration, proliferation, and differentiation. Most of the morphologic studies of abnormal matrix deposition have been carried out in skin, as it is easily obtainable and is frequently involved. The list of extracellular matrix components that are overexpressed
in systemic sclerosis includes several collagens (e.g., types I, III, IV, V, VI), glycosaminoglycans, and noncollagenous glycoproteins such as fibronectin, laminin, and tenascin (334). Central to the enhanced extracellular matrix production are abnormalities in fibroblast function.
Increased matrix synthesis is caused largely by increased transcription rates of genes responsible for production of collagen and other matrix components by fibroblasts. Overexpression of type I collagen genes by dermal fibroblasts has been demonstrated by both in situ hybridization and steady-state mRNA determination in skin biopsies of patients with systemic sclerosis (335,336). Although the precise cause of altered fibroblast function is unknown, there is evidence that increases in fibroblast proliferation and production of extracellular matrix comprise a secondary process in systemic sclerosis. Activation of fibroblasts with increased matrix production may be induced by various cytokines and growth factors released by inflammatory cells or aggregated platelets. It has been suggested that activated mononuclear cells, frequently seen early in systemic sclerosis around the small blood vessels, may influence fibrotic and vascular events by direct cellâ€“cell interaction and by the release of soluble cytokines (337). The infiltrating mononuclear cells are predominantly CD4+ T cells and express the activation marker class II MHC antigen DR (338). Retention of these cells in tissues is likely mediated by specific adhesion moleculeâ€“mediated interactions between lymphocytes, fibroblasts, and extracellular matrix components as suggested by immunohistochemical studies. Induced expression of adhesion molecules Î²1- and Î²2-integrins on the infiltrating lymphocytes, and intercellular adhesion molecule-1 (ICAM-1) on fibroblasts have been demonstrated in skin biopsies of patients with systemic sclerosis (339). Mast cells are also known to be present in early scleroderma skin lesions, frequently in close association with fibroblasts and small vessels. Several of the mast cellâ€“derived products are potentially relevant to scleroderma; these include histamine, platelet-activating factor, tryptase, proteoglycans, and certain cytokines (340).
Both in vitro and in vivo evidence supports an important role of multifunctional growth factors, such as PDGF, TGF-Î², and basic fibroblast growth factor in skin lesions of systemic sclerosis (341). Aggregated platelets, a common morphologic feature in early systemic sclerosis (342), can be an important source of these growth factors. TGF-Î² seems to play a major role in the altered collagen production in systemic sclerosis; it induces fibrosis and angiogenesis in vivo, and it stimulates fibronectin and collagen production in vitro (341). Although the abnormalities in collagen deposition may well be substantial in the pathogenesis of skin and lung lesions in scleroderma, the relevance of these abnormalities to the pathogenesis of renal lesions in scleroderma renal crisis is uncertain. Abnormal collagen synthesis may be more important in the more chronic type of renal involvement with the coarser type of intimal arterial thickening.
Various features in systemic sclerosis point to microvascular abnormalities in the pathogenesis. As in HUS and TTP, endothelial damage is thought to be the initiating factor setting up a chain of events that includes increased capillary permeability, platelet activation, coagulation abnormalities, and altered vasomotor activity. These changes may lead to the typical morphologic and functional alterations seen in systemic sclerosis. However, whether damage to vascular endothelium is the primary event or is preceded by other changes such as activation of the immune system is not clear.
Nonspecific markers of endothelial injury in systemic sclerosis include elevated plasma levels of vWF and thrombomodulin in association with reduced plasma levels of ACE (341,342,343,344). Apoptosis of the endothelial cells has been demonstrated in skin biopsies of patients with systemic sclerosis during the early stages of disease (345). It has also been suggested that systemic sclerosis is initiated by endothelial cell apoptosis (346,347348). Morphologic findings in the kidneys in patients with scleroderma renal crisis also indicate severe endothelial injury similar to that seen in HUS. The etiology of endothelial injury in systemic sclerosis may be diverse; the role of a large number of cytokines and growth factors secreted by activated inflammatory cells, serum cytotoxic factors, down-regulation of complement regulatory proteins, viruses, and antibody-dependent cell-mediated cytotoxicity have all been postulated (349,350). The role of inflammatory cells in endothelial damage has also been suggested. Induced expression of endothelial adhesion molecules (E-selectin and ICAM-1), as revealed by immunohistochemistry in skin biopsy of patients with systemic sclerosis, promotes adhesion of activated lymphocytes to the endothelium (339). Endothelial cell injury may be mediated by direct cellâ€“cell interaction with activated lymphocytes and by release of soluble mediators such as TGF-Î², PDGF, TNF, and several interleukins by the inflammatory cells (323346,351). A possible link between circulating anti-endothelial cell antibodies in patients with systemic sclerosis and induced expression of endothelial adhesion molecules has been shown (352). However, no
direct evidence of a causal relationship between adhesion molecule expression and the pathogenesis of systemic sclerosis has been shown.
Down-regulation of endothelial complement regulatory proteins, such as membrane cofactor protein, decay-accelerating factor, and CD59, has also been shown in punch biopsies of skin from patients with systemic sclerosis (353). The complement-regulatory proteins normally protect endothelial cells from damage to autologous complement. Down-regulation of this defensive barrier may contribute to vascular damage in systemic sclerosis. It is not clear, however, whether down-regulation of these endothelial proteins is a sui generis feature of the pathobiology (i.e., cause of the disease) or is secondary to endothelial damage.
Studies on the nail folds using in vivo capillary microscopy in systemic sclerosis patients have shown enlarged and deformed capillary loops surrounded by relatively avascular areas (354). Fluoresceinated tracers have revealed increased permeability of this capillary system (355). One study of skin lesions in systemic sclerosis revealed subendothelial edema as the earliest morphologic abnormality, followed by intravascular platelet aggregation and dermal accumulation of CD4+ and CD8+ subsets of lymphocytes (342). Tissue fibrosis and decreased intensity of inflammation were recognized at later stages of the disease.
Platelet Activation and Intravascular Coagulation
Various thrombotic lesions in the renal microvasculature shown in histologic sections of the kidney indicate platelet activation. Platelet activation in patients with systemic sclerosis was confirmed by demonstration of elevated levels of circulating platelet aggregates and plasma Î²-thromboglobulin (356). Most data seem to support the notion that platelet activation is secondary to endothelial injury; however, the role of a primary platelet abnormality in systemic sclerosis has also been suggested (357). Damaged endothelium may also trigger thrombosis through induced tissue factor expression and decreased anticoagulant activity.
Investigators have long known that Raynaud's phenomenon is a common feature of systemic sclerosis, and it is exacerbated by cold. The exact mechanism causing the vasoconstriction that accompanies Raynaud's phenomenon is not understood. The physiologic control of vascular tone seems to represent a dynamic interplay among neuropeptides, products of the endothelium (e.g., vasoconstrictor endothelin, vasodilator endothelial-dependent relaxation factor (NO), and PGI2), and platelet release products such as serotonin and TXA2. Complex abnormalities of normal vasoregulation were shown in patients with systemic sclerosis and Raynaud's phenomenon. Elevated levels of endothelin-1 in the circulation have been demonstrated (342a,358). Platelet activation and increased levels of vasoconstrictor TXA2 are also well-known features in patients with systemic sclerosis (356); TXA2 levels were shown to be further increased after the patients were subjected to cold. Neuropeptides, which are present in the sympathetic, parasympathetic, and sensory nervous system, may have vasoconstrictor (neuropeptide Y) or vasodilator (calcitonin gene-related peptide, substance P, and vasoactive intestinal peptide) effects. The exact role of neuropeptides in the increased vascular tone in Raynaud's phenomenon is not known; however, the current hypothesis suggests deficiency of vasodilatory neuropeptides in the pathogenesis (359). The reninâ€“angiotensin system is considered to play a significant role only during the scleroderma renal crisis with malignant hypertension (360).
Vasoconstriction may also occur in certain viscera in patients with systemic sclerosis. In a study of hypertension and renal failure in systemic sclerosis, Cannon et al (316
The contribution of functional vasoconstriction to kidney damage is difficult to determine both in patients who have developed acute renal failure (scleroderma renal crisis) and in those who have not. In those patients with systemic sclerosis and acute renal failure, structural changes are invariably present in the renal vasculature, and these changes in themselves are sufficient to explain the renal decompensation. Although activation of the reninâ€“angiotensin system with its vasoconstrictive consequences could well augment the organic changes, this issue is difficult to resolve. In the case of patients without renal failure, it is conceivable that attacks of vasoconstriction
corresponding to episodes of Raynaud's phenomenon could produce tubular atrophy such as that seen at autopsy in patients with the more chronic lesion. These attacks could also stimulate the reninâ€“angiotensin system with the production of even greater vasoconstriction and elevation of blood pressure. Increased vascular resistance indices were demonstrated by color-flow Doppler ultrasonography in patients who had systemic sclerosis without clinical symptoms of renal damage (361). The resistance indices were significantly elevated in all three vascular sites that were explored (main arteries, interlobar arteries, cortical arteries). In addition, a statistically significant reduction of the resistance indices between the interlobar and cortical vessels was found, indicating increased resistance to flow distal to the interlobar arteries. These findings provide evidence of abnormal renal vascular function in patients with systemic sclerosis without clinical evidence of renal damage. Finally, the ability of vasoconstriction to cause structural changes in arteries and arterioles is not known. Possibly, prolonged periods of intense vasoconstriction could cause structural arterial changes, but we consider this unlikely in the absence of more compelling demonstrations of extreme vasoconstriction over long periods.
Various features of activation of the humoral and cellular immune system frequently seen in patients with systemic sclerosis provide the conceptual basis for the immune hypothesis of pathogenesis. The immune mediators are thought to promote both vascular changes and altered collagen production, which are the major features of systemic sclerosis. The list of these immune abnormalities in patients with systemic sclerosis is long and includes features of circulating B-cell and T-cell activation; various serum autoantibodies; increased serum levels of interleukin-2 receptor and other cytokines such as interleukin-4 and interleukin-6; up-regulationâ€“induced expression of various adhesion molecules on the vascular endothelium, lymphocytes, and fibroblasts; and increased expression of growth factors and growth factor receptors (337,362,363). Additional evidence for the important role of immune activation in the pathogenesis of systemic sclerosis comes from chronic graft versus host disease, in which activation of the immune system may result in an illness resembling systemic sclerosis (364).
Activation of the humoral and cellular immune system provides a plausible explanation for some of the major features of systemic sclerosis; however, the initiating factors for the immune activation are not known.
Several animal models of systemic sclerosis have been described; however, no currently available models exhibit all the aspects of systemic sclerosis (reviewed by Christner and Jimenez ). The most extensively studied models are the tight skin mouse (Tsk/+) and the UCDâ€“200 chicken models. Although both these experimental models show remarkable similarities in some of the morphologic, immunologic, and biochemical aspects to those seen in human systemic sclerosis, only the UCDâ€“200 chicken model displays renal abnormalities. However, unlike in humans, swelling and thickening of the muscular layer of renal blood vessels occur; glomerulonephritis is present in approximately 20% of the birds with glomerular deposition of IgG (366). Endothelial cell apoptosis occurs in various organs (skin, esophagus, lung, kidney) of this model during the initial and early inflammatory stage of the disease, followed by mononuclear cell infiltrates and excessive collagen deposition in the skin and esophagus (347).
Summary of Pathogenesis of Scleroderma Renal Crisis
The most plausible explanation for the renal vascular changes in scleroderma renal crisis relates to endothelial damage of the arteries and arterioles in the kidney. Endothelial damage may result in undue permeability of the endothelial barrier, so various constituents of the blood gain access to the intima and, in the case of smaller arteries and arterioles, to the media. Initially, one notes edema of the intima, which then appears to undergo changes leading to the cellular or fine fibrous intimal thickening seen so commonly. Fibrin often penetrates the wall as part of the excessive permeability and is sometimes a prominent feature; in the arterioles, fibrin may pervade the entire thickness, but in interlobular arteries it is usually confined to the intima. Alternatively, coagulation may take place in the vascular lumen, followed by incorporation of fibrin into the wall by the intimal proliferative changes. The resultant picture could be one of fibrin deeply embedded in the intima. Endothelial damage may also initiate a chain of events leading to coagulation and release of factors that could stimulate smooth muscle cell movement and proliferation into the intima. Microthrombi can be found in arterioles and small arteries, which suggests an endothelial defect or a coagulation abnormality. The initiating factors and the precise mechanisms for endothelial damage are unknown. Abnormalities in cellular and humoral immunity with increased cytokine activity and vasoactive agents may be contributing factors.
Inadvertent or unavoidable irradiation of the kidney during radiation treatment of malignant tumors may cause damage to the organ, with impairment of its function and a rise in blood pressure. The advent of total-body
irradiation approximately 30 years ago as treatment for some immunologically mediated diseases and as part of the conditioning treatment regimen utilized in preparation for bone marrow transplantation (BMT) has expanded the settings in which radiation injury to the kidney may occur. More recently, radiation-induced renal toxicity has also been observed in patients who underwent treatment with various radiolabeled substances, such as monoclonal antibodies, antibody fragments, and low-molecular-weight oncophilic peptides. Fortunately, awareness of the dangers of irradiation and the introduction of techniques specifically designed to minimize acute organ toxicity have reduced the number of cases of radiation nephritis, and this disorder is now rare.
Radiation nephropathy is discussed in this chapter because of the significant overlap of renal morphologic changes between radiation nephropathy and various forms of TMA. Clinical findings and renal morphologic changes secondary to external beam irradiation, total body irradiation, and radionuclide treatment are included.
Early writings have been well reviewed by Redd (367) and Mostofi (368), with the latter author recording 120 cases up to 1964. In the first decade of the 20th century, clinicians were aware of the occurrence of albuminuria and nitrogen retention in patients receiving roentgen-ray therapy, and they realized that care must be taken in its use. Domagk (369) appears to have been the first to give an accurate picture of the irradiated human kidney; he described a 9-year-old girl whose abdomen was irradiated for tuberculous mesenteric lymph nodes. Four months after treatment, a raised temperature, oliguria, and albumin and casts in the urine were noted; death occurred 2 months after these symptoms first appeared. At autopsy, the kidneys were reduced in size and showed glomerular hyalinization or thickening of Bowman's capsule, tubular atrophy or necrosis, and hyaline material in the arterial walls. A short time later, Zuelzer et al (370) described an unusual form of glomerulonephritis in three young children who had received heavy irradiation for either Wilms' tumor or neuroblastoma; these investigators considered the condition in all probability to be radiation nephritis. Renal failure and hypertension developed within 3 months in these patients, and death occurred quickly. The glomeruli were severely involved by either capillary wall thickening or necrosis; tubular loss, interstitial fibrosis, and necroses in afferent arterioles and small arteries also were found. Luxton (371) followed this report with a series of 27 patients whose kidneys had been damaged after irradiation of abdominal lymph nodes as part of the treatment for seminoma of the testicle. He recognized various clinical syndromes (discussed in the next section) and described the pathologic changes. His work was supplemented by that of Russell (372) from the same hospital. Since that time, other smaller series and single case reports have appeared that amply confirm these principal observations, although some modifications to the conceptual formulations are needed to take account of the recorded clinical experience and the results of experimental studies performed since that time. In particular, the recognition of a syndrome of TMA/HUS associated with total-body irradiation as used in BMT procedures has come to be recognized as a unique clinicopathologic entity, as discussed later in this chapter.
Clinical Presentation and Clinical Course
The experience of Luxton is unique, and his initial article (371) and subsequent studies describing the long-term follow-up records of 54 patients treated with abdominal irradiation for malignant testicular tumors (49 cases) and ovarian tumors (5 cases) (371,) set out distinct clinical syndromes (acute radiation nephritis, chronic radiation nephritis, asymptomatic proteinuria, benign hypertension, and malignant hypertension) in the following way.
Acute Radiation Nephropathy (Acute Radiation Nephritis)
Twenty of the 54 patients studied were classified as having acute radiation nephritis. Luxton used the term acute radiation nephritis to describe a syndrome that developed over many months and whose pathologic manifestations, described later, lacked the inflammatory changes characteristic of other acute nephritides (371,373). For this last reason, most experts (367,374) prefer the terms acute radiation nephropathy and chronic radiation nephropathy for the syndromes that Luxton delineated.
Patients with acute radiation nephropathy, which occurred following a latent period of 6 to 12 months after irradiation, had a gradual onset of edema, hypertension, dyspnea on exertion, anemia, headaches, and urinary changes that included proteinuria and the presence of casts. The edema was either confined to the legs or was more generalized, with effusion into the pleural space and pericardial sac; it was absent in a few patients. Hypertension was constant and was commonly at its height within 6 months of the appearance of the first symptom. The elevated blood pressure returned to normal levels in a certain proportion of patients; this took place even in some with malignant hypertension, which was present in almost half of these patients. Anemia was present in all patients and was of a severe normochromic, normocytic type. The blood urea level was raised early in some patients and either remained at a high level or fell.
Subsequent accounts confirmed Luxton's description, and one or two additional features were pointed out. Anemia is common, and in several reports (375,376,377), it was
of the microangiopathic hemolytic type; deformed red blood cells were seen in blood smears, and platelets were decreased in numbers. Fibrin degradation products were present in the circulation in two accounts (376,377), and evidence of intravascular coagulation was claimed in one (376). Currently, these findings would be considered indicative of a TMA/hemolytic syndrome as the basis for the patients' renal dysfunction. Rare reports of the nephrotic syndrome associated with abdominal irradiation also have been described (378).
In general, the time required for acute radiation nephropathy to develop is 6 to 12 months, but certain patients have had an earlier onset, such as 3 or 4 months (377,379), 5 weeks (380), and 4 weeks (381). All of these latter patients had been treated with chemotherapeutic agents such as actinomycin D, doxorubicin, bleomycin, vinblastine, dacarbazine, cyclophosphamide, and vincristine. These cases are of interest in view of the short interval between irradiation and the development of radiation nephropathy. Investigators have frequently suggested that chemotherapeutic agents may increase the risk of radiation damage to the kidney, as discussed in subsequent sections on dosage and BMT.
In acute radiation nephropathy, Luxton related the immediate prognosis to the presence of malignant hypertension, although the exact parameters he used to define this entity were not given. Of a total of 20 patients studied, malignant hypertension developed in 8, and 6 of the 8 died within 3 to 12 months. The other two with malignant hypertension recovered spontaneously, although one died later with chronic renal failure, and the other had some degree of chronic renal insufficiency after 12 years of follow-up. Of the total group of 20, another 2 developed chronic renal failure and died from it 7 to 11 years after irradiation. One other patient died of a hypertensive cerebrovascular accident, and another died of widespread metastases of an original ovarian cancer. The remaining nine were alive and active after an average of 10 years, but all had evidence of chronic nephropathy. In this syndrome, signs of a poor prognosis were considered to be generalized edema, hypertensive retinopathy, and a blood urea level above 100 mg/100 mL during the first 3 months.
Chronic Radiation Nephropathy (Chronic Radiation Nephritis)
Patients with chronic radiation â€œnephritisâ€ (again, the term nephropathy is preferred to Luxton's use of the term nephritis) came from two main sources: (a) those who presented initially with acute radiation nephritis and who continued with signs of chronic renal injury (e.g., the 14 patients described in the previous paragraph who did not die of malignant hypertension early in their course); and (b
Death in chronic renal failure could occur in either of the two groups of patients with chronic radiation nephropathy described in the foregoing paragraph. Although noting that the figures were small, Luxton gained the impression that patients whose cases followed an initial acute attack fared better than those without a preceding attack. Of the 14 patients whose chronic disease followed an acute attack, 9 were leading normal lives up to 12 years after irradiation; 5 of the 9 had impairment of renal function or hypertension (or both), whereas the other 4 showed only protein in the urine. The five deaths were caused by chronic renal failure in three patients, hypertensive cerebrovascular accident in one, and spread of tumor in the other. As for the group of 10 patients in whom chronic nephropathy appeared without an antecedent acute attack, 3 died of chronic uremia on average 7.5 years after irradiation, and 1 died of malignant disease. Malignant hypertension developed in one patient after 10 years, and five patients were alive 10 years after irradiation, two with slowly diminishing renal reserve and three with compensated renal failure.
Thirteen of the 54 patients had renal dysfunction consisting only of proteinuria after an average follow-up of 11 years after radiotherapy. The quantitation of proteinuria was not provided, and the only additional information provided was a statement that â€œstandard renal-function tests were normal.â€
Six of Luxton's patients acquired a benign form of hypertension together with some proteinuria 2.5 to 5 years after irradiation. Two died of congestive failure, but the others were alive and without additional evidence of renal dysfunction at follow-up periods ranging from 9 to 13 years.
Of all the patients studied by Luxton, 15 (28%) acquired malignant hypertension. This complication either occurred as part of the syndrome of acute radiation nephropathy or
it developed independently 18 months to 11 years after irradiation. Many deaths occurred in this group.
The kidneys may be normal in size or contracted. Capsular thickening may be found in the more contracted forms, and the thickened renal capsule fuses with sclerotic tissue surrounding the kidney. The normal-sized kidneys, which correspond to those with little tubular loss and no interstitial fibrosis, have smooth subcapsular surfaces, and on cut surfaces, they may show nothing more than vague mottling. The contracted kidneys, corresponding to those with interstitial fibrosis, have capsules that may be impossible to strip from the cortex, which is reduced and firm. When radiation damage is confined to the main renal artery, the kidney would be expected to be reduced in size, as in renal artery stenosis of any cause. The subcapsular surface would be smooth in younger patients with no intrarenal vascular disease, but it could be granular in older patients.
Most descriptions of glomerular alterations in clinical publications come from examination of kidneys with advanced injury. These descriptions have not, in aggregate, resulted in the delineation of a distinctive histopathologic process. In contracted kidneys, one may see considerable sclerosis of glomeruli, but even in this type of kidney, large numbers of glomeruli show little apparent change. In normal-sized kidneys, the glomeruli may show fibrinoid changes of tufts in continuity with similar changes in arterioles, segmental areas of glomerular scarring, and small fibrous crescents. Frequently, swelling of the glomerular capillary wall endothelium may progress to obliterate the capillary lumen (Fig. 16.33A). Thickening of the capillary walls is sometimes seen, and split, double-contoured glomerular basement membranes can be demonstrated with silver impregnation techniques (382Mesangiolysis, a term not in general use at the time some of these reports were written, can be identified in the illustrations of some of the early descriptions (370)
(Fig. 16.33B). Accumulations of fibrin within glomerular capillaries have been demonstrated using special stains (368).
The most consistently observed changesâ€”endothelial swelling and reactive changes, glomerular capillary basement membrane thickening and splitting, and mesangial thickeningâ€”are characteristic of a chronic mesangiocapillary or membranoproliferative form of glomerular injury. These changes may follow from diverse forms of primary injury, including the late manifestations of TMA. Indeed, some reports of radiation-associated glomerular injury clearly delineate changes consistent with the advanced stages of TMA (376,378,383), although this impression results primarily from the electron microscopic findings in these cases. The abnormalities seen by light microscopy have included segmental fibrinoid necrosis of the glomerular tufts, occlusive intracapillary accumulations of periodic acid-Schiffâ€“positive material with enmeshed erythrocytes, mesangiolysis, endothelial swelling, and splitting of glomerular capillary basement membranes, associated at times with correlative fibrinoid changes of the terminal afferent arterial vasculature. Only rarely have clinical laboratory findings been obtained to corroborate conclusively the presence of coagulation disorder consistent with TMA at the time of biopsy (376).
Appreciation of this potential congruence between TMA and radiation nephropathy is likely to be useful in unifying conceptually some of the various descriptions of radiation nephropathy in the older clinical literature with observations in experimental animals and with the special case of radiation nephropathy/TMA that clearly may occur in association with BMT.
No vascular pathologic alterations are unique to radiation nephropathy. The arterioles and small interlobular arteries often show fibrinoid change without a cellular component in both contracted and normal-sized kidneys. These changes occur with more modest rises in blood pressure than is the case with nonirradiated kidneys and severe hypertension. These arteries may also demonstrate an intima expanded by the accumulation of pale-staining, at times mucoid-appearing, material that most likely represents a type of provisional matrix or loose connective tissue. Thrombotic lesions in the lumina may also be seen (Fig. 16.33C). Larger interlobular arteries often show fine collagenous or cellular intimal thickening and sometimes a denser type of sclerosis. This intimal thickening is usually a patchy process. Foam cells are often seen in the thickened intima of interlobular arteries (370,381,384). The origin of such cells remains unknown, but studies of foam cells in the arterial intima in human atherosclerosis have shown them to be predominately of macrophage origin, with a lesser population derived from smooth muscle cells. Arcuate, interlobar, and segmental arteries may show no changes, patchy fibrous intimal thickening, or more severe and extensive intimal sclerosing changes, especially when the kidneys are reduced in size. Evidence of active or healed vasculitis involving small muscular arteries rarely has been demonstrated in other organs with radiation injury (), but vasculitis (i.e., prominent inflammatory cell infiltration of the arterial intima or medial layers with focal destruction of the involved vessel wall) has not been reported in the kidney. Because the microvascular injury may be focally distributed, its extent may not be fully appreciated in routine histologic sections. Microangiography in animal models has been used to demonstrate more widespread vascular injury than could be visualized using conventional pathologic examination techniques alone (386,387).
Tubules and Interstitium
As in the glomeruli and blood vessels, no distinctive tubular or interstitial pattern of injury characterizes any form of radiation nephropathy. Again, because much of the human material reported comes from cases with advanced injury, descriptions of the tubulointerstitial parenchyma emphasize the nonspecific features of interstitial fibrosis and tubular atrophy common to all forms of severe, chronic renal injury.
In contracted kidneys demonstrating severe, chronic injury, considerable tubular loss is paralleled by extensive glomerular solidification. However, some degree of tubular loss and atrophy may occur in the absence of glomerular solidification, especially in laboratory animals subjected to irradiation. Nonspecific accumulations of mononuclear leukocytes are often seen.
In unscarred kidneys, tubular loss may be negligible or may occur focally in those areas where glomerular damage has occurred. The interstitium may show no changes or only a slight apparent increase in interstitial tissue. Edema with dilated peritubular capillaries may occur in patients dying comparatively early (372), comparable with the early stages of experimental irradiation of the canine kidney (388). Proteinaceous casts may be found in the lumina of tubules, and patchy hyaline droplet change occurs, especially in those cases with the greatest proteinuria. Other nonspecific changes, such as tubular cell vacuolization and desquamation, have been observed.
Some studies in the experimental literature, primarily those involving rodent and canine models, suggest an important and perhaps primary toxic effect on the tubules by radiation as the basis for subsequent radiation nephropathy (389390,391,392,393). Evidence indicates an acute tubular injury corresponding to acute tubular necrosis that may occur in the aftermath of high-dose irradiation (378). This injury appears to be followed by a reparative response of tubular regeneration and proliferation.
Two reports noted no immunoglobulins, complement, or fibrinogen in the glomeruli (376,382). A patient with acute radiation nephritis complicated by nephrotic syndrome (378) had segmental and irregular staining for IgM, C3, C4, and C1q in the glomerular capillary loops and mesangium; trace amounts of IgG and IgA were found in a similar distribution. Another report described IgG, IgM, and fibrinogen along the glomerular capillary walls with a focal and segmental distribution (375). IgM and C3 were detected in arteriolar walls. No distinctive pattern was identified.
Several studies have been carried out using transmission electron microscopy in human kidneys (,376,379,383). In the glomeruli, one sees effacement of the visceral epithelial cell foot processes that is variable in its extent. The glomerular capillary walls usually show a widening of the subendothelial zone, which may contain electron-lucent or finely granular flocculent material. The glomerular endothelial cells are often swollen and contain increased numbers of organelles. In some accounts, these glomerular cells are described as detached from the endothelial aspect of the basement membrane. Mesangial cell cytoplasm or possibly entrapped endothelial cell cytoplasm is sometimes present in the capillary walls, and new glomerular basement membraneâ€“like material may appear on the endothelial aspect. These changes are responsible for the double contours seen on silver staining. Mesangial cells may be swollen, and their cytoplasm may extend along the capillary loops. Fibrin has been seen between the endothelial cells and the glomerular basement membranes. Fewer accounts note changes in tubules and arteries, but in arteries, swollen and hypertrophic endothelium has been recorded along with fibrin or hyaline material in the subendothelial zone.
Etiology and Pathogenesis
Virtually all the morphologic changes described in human radiation nephropathy can be reproduced in experimental models. However, the morphologic presentation of radiation nephropathy varies significantly between various species. In experimental animals, just as in humans, the radiation-induced renal functional deterioration is associated with structural changes affecting all four renal compartments. In rats, radiation doses equivalent to those used in treating human subjects cause progressive renal failure in both the local and total-body irradiation models (394). Progressive glomerular changes with dominant endothelial abnormalities similar to those described in human kidneys with radiation nephropathy are the most characteristic findings in rats, mice, pigs, and nonhuman primates (374,389,391,395,396,397,398). The glomerular capillary endothelial changes include swelling, detachment from the underlying basement membrane, and accumulation of subendothelial electron-lucent material on electron microscopy, with subsequent collapse of the capillary loops and eventual glomerular obsolescence. The glomerular endothelial changes are detectable by fewer than 6 weeks after irradiation in all species. In addition to endothelial changes, accumulation of platelet thrombi within the glomerular capillary loops is also observed in rats, mice, and pigs (374389,390,391,395,399,400,). Mesangial sclerosis has also been described in all animal models and mesangiolysis in some. Tubulointerstitial changes in rats include tubular epithelial degeneration, necrosis, and tubular atrophy that precede the morphologic changes in the arteries and arterioles, suggesting direct toxicity of radiation to tubules (389,390,391,393,401,402). Fibrinoid necrosis of small arteries is a feature of radiation injury in dogs (392). Vascular occlusive lesions, both thrombotic and nonthrombotic, seem to be prevalent in only some experimental models of radiation nephropathy (391). Widespread glomerulosclerosis, severe tubular loss, and extreme interstitial fibrosis were the long-term consequences of radiation injury to the kidney, as shown in dogs that were irradiated in the neonatal period and died before the age of 4 years (403,404,405). These various experimental models show involvement of all four renal compartments of the kidney in the evolution of radiation nephropathy. Although some studies seem to support the prominence of glomerular endothelial injury in the pathogenesis, the relative contributions of the glomeruli, tubules, interstitium, and vessels in radiation nephropathy remain controversial. One traditional hypothesis of the pathogenesis of radiation nephropathy emphasizes the significance of vascular injury in both disease development and outcome (). The other hypothesis holds that parenchymal injury, primarily tubular cell injury, is the key event in the pathogenesis (407).
Mature pigs may serve as useful models for human radiation injury. In such models, when pig kidneys are irradiated at doses no higher than 9.8 Gy, a measurable and progressive decline takes place in renal hemodynamic parameters such as glomerular filtration rate and renal plasma flow (396). Serial renal biopsies in such animals have revealed principally injury to the glomerular peripheral capillary loops, initially characterized by leukocyte influx and attachment to the endothelium, and narrowing of the capillary lumina by the combined processes of endothelial cell swelling and intracapillary leukocyte stasis (396). These morphologic and functional changes can be identified as early as 3 to 6 weeks after irradiation. Immunohistochemical studies in this model using bromodeoxyuridine have detected an increase in glomerular cell
408,409), before the appearance of overt morphologic injury. At subsequent periods of 2 months and later, the evolution of glomerular capillary wall abnormalities includes progressive widening of the subendothelial space with accumulations of electron-lucent material, which is apparent with electron microscopy. Features suggestive of mesangiolytic injury can be identified in the illustrations of some of these studies. These morphologic changes most closely resemble those of TMA as encountered in humans, and they correspond to the descriptions of glomerular radiation injury in humans, mice, and some studies of rats given earlier in this section. In this model, detectable increases in tubular cell proliferation follow those of glomeruli by a period of weeks, morphologic evidence of tubular injury is much more focally distributed than that of glomeruli, and this injury appears reversible over time (408).
Although the precise molecular mechanisms responsible for the development of radiation nephropathy are not known, a number of factors playing potentially important roles in this process have been identified. In addition to angiotensin II, which seems to have a central role in the pathogenesis, there is some in vitro and in vivo evidence that factors such as nitric oxide, TGF-Î², and PAI-1 may also be involved (410,411,412). There is also strong evidence for the role of oxidative stress in fibrogenesis (413).
Hypertension often occurs in irradiation injury, and most case reports mention it. The mechanism is not readily understood, but the renal origin seems undoubted. In several cases, removal of a single damaged kidney has resulted in a lowering of blood pressure, the first of these cases being reported by Dean and Abels (414). In this case, a woman received 4600 R (R, or roentgen, is a unit of radiation exposure introduced in 1928 but now replaced by other units of measurement; 1 R = 2.58 Ã—10-4 coulomb/kg of air ) to the left upper quadrant over 25 days, and hypertension developed 7 years later. At nephrectomy, the left kidney was shrunken in its lower third and showed obliterative arteriopathy in this part. Cogan and Ritter (416) described a 14-month-old boy who received a total dose of 5300 R over 36 days to the left renal area for a neuroblastoma; hypertension developed 3 months later. A left nephrectomy caused the blood pressure to fall. The excised kidney showed diffuse interstitial fibrosis, hyalinization and necrosis of the glomerular tufts, and intimal proliferative changes and thickening of arterial blood vessel walls. These investigators suggested that the hypertension was explainable by a â€œGoldblatt mechanism,â€ presumably meaning that the arterial narrowing in the radiated kidney was responsible. A number of additional cases have also been reported with reversal of hypertension after removal of the irradiated kidney (381,384).
Although no doubt exists about the renal origin of the hypertension in such cases, the precise mechanism is not known. Few reports mention intimal thickening of large arteries (). We have seen a number of cases with prominent intimal thickening of the interlobular and arcuate arteries; however, some cases exhibited little change in the large arteries.
Of great interest in this regard are cases in which irradiation is apparently responsible for narrowing of the main renal artery or for what has been called â€œhypoplasiaâ€ of the abdominal aorta and renal artery (418,419,420,421). In these cases, irradiation was usually performed during infancy or early childhood, and an interval of several years elapsed before hypertension became manifest. This interval was as long as 12 or 13 years in some instances. Unfortunately, reports on the pathologic features of the arterial changes are scant, although in one of them (420), intimal and medial fibrosis with thrombosis in the narrowed renal artery was found. In two cases reported in one series (421419,420), a finding that suggests that the changes were more or less confined to the main renal artery or aorta, with sparing of the intrarenal vasculature. Information on the pathologic features of the kidney is unfortunately sparse in this group, but ischemic changes were described in two patients (415), with no evidence of changes ascribable to radiation.
As described in the preceding sections â€œPathologic Findingsâ€ and â€œExperimental Studies,â€ it is increasingly apparent that the principal site of injury in clinically significant radiation nephropathy is the arterial and glomerular vasculature. The interstitial capillaries may also be prominently involved, but this feature is not as well established and is not always observed. As in other situations in which hypertension and lesions in small blood vessels coexist, the relationship between the two is complex. By analogy with the malignant phase of essential hypertension, one may assume that the fibrinoid necrosis of arterioles and small arteries that is frequently encountered in radiation injury is caused by the elevated blood pressure, irrespective of mechanism. In contrast, necrosis in arterioles and small arteries may be found in irradiated kidneys in the presence of only modest levels of blood pressure, which suggests that the vascular changes are more likely related to the effects of radiation. The experimental evidence is confusing. First, investigators have shown that hypertension can develop in the rat after irradiation without any morphologically demonstrable vascular lesions (421). Second, Fisher and Hellstrom (423) showed that irradiation did exert an effect unrelated to any rise in blood pressure in the affected arteries. In this experiment, hypertension was produced in a series of rats by the application of a silver clip to one
renal artery, with the other renal artery left untouched. After irradiation of both kidneys, vascular necroses were present on both sides. The arterial changes in the â€œnonclippedâ€ kidney would be expected as a response to the elevated blood pressure, but such could not have been the case in the â€œclippedâ€ kidney, which would have been protected against the high blood pressure by the clip on the renal artery. The necrotizing lesions in the clipped kidney must therefore have been caused by the radiation. Third, Asscher et al (424) demonstrated that arteriolar and arterial necrosis can occur after irradiation at levels of blood pressure that, by themselves, are unlikely to cause necrosis. In these experiments, the authors irradiated part of a loop of mesentery and, after producing hypertension by renal artery constriction, compared the irradiated mesentery with the nonirradiated mesentery. Necrotizing arterial changes occurred in the irradiated but not in the untreated mesentery. â€œSensitizationâ€ of the mesenteric arteries to the damaging effect of hypertension was not established until 2 months after the irradiation. That radiation alone was unable to produce the necroses was shown by the absence of these lesions in rats that were irradiated but did not have hypertension. Two possible explanations for the vascular damage as a result of combination of irradiation and hypertension were considered. The first was that irradiation could so weaken the arterial vessel wall that mechanical disruption by the increased intravascular pressure could occur. The second stemmed from Byrom's observation in the rat (425) that focal arterial spasm of cerebral and mesenteric vessels followed severe hypertension. It was suggested that this focal excessive vasoconstriction was a myogenic reaction to the increased intravascular pressure and that it led to arterial necrosis.
Asscher et al (424) speculated that irradiation could render arterial walls more susceptible to hypertensive damage by exaggerating the myogenic response to variations in intravascular pressure. In this way, excessive vasoconstriction could occur as a response to what are merely physiologic variations in blood pressure, with consequent necrosis of the wall. Moreover, according to this concept, the increased susceptibility to vasoconstriction could cause ischemia, leading to the production of hypertension in the same way as occurs with organic occlusion of vessels.
These seemingly contradictory experimental studies do not really help us to determine the relationship between vascular lesions in the kidney and hypertension in humans, but they do show at least that radiation has a profound effect on arterial blood vessels, with potentially serious consequences for the kidney. More recent investigations of radiation-associated vascular injury have identified the endothelium as a principal site of injury, but they have not further clarified mechanisms that result in the fibrinoid vessel wall lesions and injury to arterial smooth muscle cells that were a focus of early studies. Studies of the potential activity of specific growth factors and cytokines now known to be important in arterial wall injury, repair, and sclerosis in various settings are noteworthy for their absence in studies of radiation nephropathy at the present time.
Other studies in animal models of radiation injury are beginning to dissect broader physiologic mechanisms for the development of hypertension. Administration of the ACE inhibitor class of antihypertensive drugs or angiotensin II-receptor blockers is effective in the treatment or prophylaxis of experimental radiation nephropathy (426,427,428,429,430,431), whereas lowering of blood pressure to a similar degree with hydrochlorothiazide apparently has no similar beneficial effect (427). Studies by Juncos et al (432) of unilaterally radiated kidneys in rats revealed hypertension associated with elevated plasma renin levels at 12 weeks after radiation exposure. These studies suggest that radiation-associated hypertension may be mediated by alterations of the reninâ€“angiotensin system, thus leading to a hyperreninemic state. However, no activation of the reninâ€“angiotensin system was observed in a rat model of radiation nephropathy during the first 10 weeks postirradiation (
The preceding observations also suggest that, at present, the link between renal radiation injury and hypertension may invoke mechanisms not necessarily directly resulting from the effects of radiation on the target tissue, but they can be the consequence of chronic or fibrosing renal injury that may occur long after the acute injury. The evidence to support this conclusion is at least threefold: (a) in most clinically reported cases, the occurrence of hypertension takes place many years after radiation exposure; (b) in most case studies documenting the reversal of hypertension after excision of an irradiated kidney, the removed kidneys are typically described in whole or in part as shrunken, scarred, fibrotic end-stage organs; (c) several studies implicate changes in the reninâ€“angiotensin system that result from diminished blood flow in the development of postirradiation hypertension (i.e., Goldblatt-type mechanisms of injury), similar to changes that occur in other forms of advanced chronic renal injury.
A hyperreninemic state has been detected in some patients with radiation-associated hypertension, and this has been reversed with correction of blood pressure by removal of the damaged kidney, as published in isolated cases (381,420,434). In a prospective study of patients treated with high-dose abdominal irradiation for malignant diseases, approximately half developed hypertension (435). The patients with hypertension demonstrated elevated peripheral plasma renin activity after oral administration of captopril compared with controls, and captopril renography was abnormal in five of the eight affected patients. Angiography in the five patients with an
abnormal renogram demonstrated severe stenotic and tortuous changes in small intrarenal arterial vessels in irradiated kidneys without stenosis of the main renal artery.
In aggregate, microvascular injury leading to diminished blood flow and a compensatory increase in renin and angiotensin could account for some proportion of cases of radiation-associated hypertension. Hypertension may also contribute to progressive vascular and parenchymal injury and the maintenance of the hypertension.
Dosage of Irradiation
Early reports gave little idea of the level of dosage required to produce renal damage. Not until the investigations of Paterson (436) and Kunkler et al (437) were published were limits of tolerance determined. Different methods of irradiation were considered with regard to the dose delivered to the kidney and the incidence of renal failure. Investigators showed with irradiation of the abdomen that (a) hypertension and renal failure may be caused when a homogeneous dose of 2300 R is delivered to the whole of both kidneys, and () the risk of renal failure may be reduced when one third of the total volume of the kidneys is outside the irradiated field.
More current considerations of organ susceptibility to radiation have stressed the concept of tolerance dose. The minimal tolerance dose and the maximal tissue tolerance dose refer to severe, life-threatening complication rates of 5% and 50%, respectively, occurring within 5 years of therapeutic radiation treatment (438). Few data define the tolerance doses of human kidneys to a single dose of high-level radiation. Guidelines in this setting are frequently drawn from the studies of Glatstein et al (374), who showed that in rats a single dose of radiation of 19 Gy (a gray [Gy] being the SystÃ¨me International unit of absorbed radiation dose, where 1 Gy = 100 rad = 1 J of absorbed energy per kilogram of material ) to both kidneys resulted in death from renal failure, whereas 90% of similarly irradiated rats survived at doses of 11 Gy. Although the perils of drawing exact correlations from rat kidneys to human kidneys are obvious, investigators have long recognized that dividing the total therapeutic radiation dose to be delivered into fractionated doses over a period of time considerably reduces the organ toxicity that would be engendered by a single dose of the same cumulative magnitude. Accordingly, more relevant tolerance doses are those that specify a specific regimen of fractionated dosing.
The renal minimal tolerance dose has been considered to be 20 to 23 Gy when kidneys are radiated bilaterally in fractionated protocols delivered over a 3- to 5-week period (439). Even the most acute radiation injuries of the kidney identified in humans take weeks to months to develop, and some features of chronic nephropathy or hypertension have been identified years and even decades after radiation exposure (438,439). Few clinical studies to ascertain tolerance doses have sufficient extended follow-up. A summary of such studies by Cassady (439) has identified a threshold radiation dose of 15 Gy, when fractionated protocols are used, for eventual development of detectable renal dysfunction. Whereas the pig and monkey can provide particularly good models of renal radiation injury, they have not yet been fully exploited to guide establishment of criteria for tolerance doses using fractionated schedules.
440), several parameters of renal function (but not histopathologic features) were observed for 3 to 5 years in patients following therapeutic radiation for a variety of malignant diseases. In patients receiving the highest amount of radiation (40 Gy in fractionated doses over 5.5 weeks), creatinine clearances decreased by approximately 20% over the 3- to 5-year interval. Patients receiving lower doses of radiation (17 to 18 Gy in fractionated doses over 3.5 weeks) had no comparable detectable decline in renal function over the same follow-up period. Certain other tests of renal function, including concentrating capacity (measured by urine osmolality), demonstrated no significant alterations as a result of radiation in any of the dosage ranges studied.
An additional point stressed by Dewit et al () in summarizing their work and that of others is that an inverse relationship appears to exist between radiation dose and the length of the latency period before overt nephrotoxicity can be detected. Thus, although the patients receiving high-dose radiation had detectable abnormalities within 3 to 5 years after injury, those patients receiving lower doses may not show such effects until some additional years later. For example, renal dysfunction attributed to the consequences of radiation has been identified at periods of up to 19 years after unilateral renal radiation at doses of 15 to 20 Gy occurring in patients receiving radiation therapy for gastric ulcers (441). This point makes it virtually impossible to determine the precise lengths of time needed to establish tolerance doses for the kidney.
A long-standing but unresolved area of concern is the extent to which augmentation of radiation damage in the kidney by concomitant or sequential administration of chemotherapeutic agents may occur (379,438,439, 442,443). The mechanisms underlying such interactions are complex and poorly understood, but the effects are not the result of simple additive cell killing of a common target cell population. Even when specific agents are directed against different cell populations, tissue sensitivity to both radiation (i.e., a lower tolerance dose) and the chemotherapeutic agents utilized may be enhanced. In other cases, chemotherapeutic agents may have direct renal toxicities, such as the tubular toxicities of cis-platinum, which may potentiate or add to the effect of radiation ().
Because modern therapeutic radiation protocols frequently reduce radiation dosage or administer radiation
sequentially when chemotherapy is being utilized in a given patient, serious complications of combined therapy are infrequent. Specific mechanistic interactions between chemotherapeutic agents and radiation in individual case reports are difficult to evaluate because of the uncontrolled, anecdotal nature of such evidence. Actinomycin D, used in conjunction with radiation to treat Wilms' tumor, can potentiate radiation injury in some tissues, although such an effect in the kidney has not been substantiated (439). Nitrosoureas have also been suspected of potentiating radiation injury in humans (444). In rats, cis-platinum and the nitrosourea BCNU have both been shown to cause additional dose-related decreases in renal function beyond those attributable to radiation alone (443).
445) showed no greater incidence of nephritis or urinary abnormality in these patients than in unexposed children. Another report describes patients living in regions where they would have been exposed to radiation resulting from the nuclear accident at Chernobyl who developed chronic pyelonephritis and glomerulonephritis (446). However, the details needed to verify this report and to compare the epidemiologic data with an appropriate control patient population are not available, and caution in accepting these data is advised.
Radiation Nephropathy After Total-Body Irradiation (Bone Marrow Transplant Nephropathy; Posttransplantation TMA)
Bone marrow transplantation continues to be a widely used therapy for hematologic, lymphoid, and some epithelial and mesenchymal malignant diseases, as well as for some nonneoplastic metabolic and genetic disorders. Worldwide, more than 20,000 bone marrow transplants (BMTs) are performed each year. Several syndromes of acute renal insufficiency have been identified in these patients. Most of these syndromes occur early after BMT (i.e., in the first 3 months) and are usually the result of sepsis, administration of nephrotoxic antibiotics or other therapeutic agents, or hepatorenal syndromes (447,448). They are not discussed further in this chapter. Of particular note is a clinical syndrome of late renal insufficiency, generally occurring after 3 months and with a peak incidence of 9 to 12 months after BMT (447,449,450,451,452,453,454,455,456,457,,459). This syndrome, which occurs in long-term survivors of BMT, has long been recognized to demonstrate clinical and morphologic features of HUS (447,450,451,453,454,460). Synonyms for this syndrome include BMT-associated TMA, radiation nephropathy after total-body irradiation, BMT nephropathy, TTP or HUS following BMT, and HUS/TTP or TTP/HUS after BMT. Recently, the toxicity committee of the Blood and Marrow Transplant Clinical Trials Network recommended the term posttransplantation thrombotic microangiopathy for this disorder (459). A systematic review of the literature from 1966 to 2003 revealed an 8.2% incidence of TTP/HUS in 5423 allogeneic BMT patients (458). However, the reported incidence varied by 125-fold, from 0.5% to 63.6%. This great variation in the incidence underlines the difficulties of the clinical diagnosis of BMT-associated TMA. The principal laboratory diagnostic features of TMA (microangiopathic hemolytic anemia and thrombocytopenia) are often seen in patients who undergo BMT. In addition, these patients can be critically ill with numerous comorbidities, such as opportunistic infections, drug toxicity, radiation-related injury, or acute graft versus host disease (GVHD), all of which can mimic the laboratory and clinical features of TMA. The clinical presentation is either a rapid or gradual decline in renal function manifested by increased BUN and serum creatinine concentrations. Hypertension, hematuria and red blood cell casts, proteinuria, hemolytic anemia, thrombocytopenia, consumptive coagulopathy, and congestive heart failure are frequent findings. Anemia is disproportionately worse than would be expected for the degree of azotemia. Individual patients may experience severe consequences as a result of these abnormalities, whereas in others, the disease process may be mild. The morphologic features encountered in renal biopsies of affected patients are typical of a thrombotic microangiopathic process such as HUS. These features are described in â€œPathologic Findingsâ€ and are illustrated in Figure 16.34.
The high incidence of this clinical and pathologic syndrome has led some investigators to identify it as BMT nephropathy450,). This syndrome was suspected to be a likely manifestation of radiation nephropathy at the time of its first description by Bergstein et al (375), but the difficulty in distinguishing it from confounding contributions
of concurrent drug therapy, infections, and other clinical variables present in this population of extremely ill patients made it difficult to establish this relationship conclusively. As summarized by Zager (447) and Paller (460
The lines of evidence to support this conclusion include (a) renal biopsy findings in affected patients that are virtually identical to those identified as characteristic of radiation nephropathy, as described in â€œPathologic Findingsâ€; (b) similar latencies after exposure to radiation between development of clinical radiation nephropathy and BMT nephropathy; (c) overlapping clinical findings between patients with HUS after BMT and patients with radiation nephropathy and concomitant evidence of a coagulation disorder; and (d) an association between reduction of total radiation dose by partial shielding of the kidneys during total-body irradiation and dramatic lowering of the incidence of BMT-associated HUS. In a series published by Lawton et al (453), this method for reducing the renal radiation dose resulted in a decline of BMT-associated HUS from 26% to 6% in otherwise equivalent patient populations. This last observation in particular serves to establish the pathogenic role of renal irradiation in the development of this nephropathy. However, note should be taken that TMA can also develop in BMT patients who have not received total-body irradiation as part of their conditioning regimen (457461,462,463). In a series published by Elliot et al (462), TTP developed in 2 of 13 patients who underwent BMT after a non-myeloablative conditioning regimen. In the series of Siami et al (463) regarding BMT-associated TMA, four of eight patients with histologic evidence of renal TMA had no total-body irradiation as part of the conditioning regimen. The renal morphologic changes of TMA in patients with prior irradiation were indistinguishable from those who developed TMA without prior irradiation (463).
Treatment for radiation nephropathy in patients who have undergone BMT remains supportive. Many patients exhibit only mild renal insufficiency, and such patients may recover renal function with improvement in concurrent hematologic abnormalities (). More severe cases may require dialysis; in such patients, mortality may be high (457). Plasma exchange, the principal treatment for the idiopathic forms of HUS/TTP in adults, seems to be ineffective in patients with BMT-associated TMA (464). The ineffectiveness of plasma exchange treatment in this setting has been attributed to the absence of ADAMTS13 deficiency in these patients (462,465). However, Vesely et al (74) showed that severe ADAMTS13 deficiency does not identify all patients diagnosed with HUS/TTP who may respond to plasma exchange treatment.
466). Unusually large vWF multimers, indicative of large-scale disruption of endothelial cells, have also been demonstrated in some patients with BMT-associated TMA (462). Endothelial injury may result from multiple pathogenetic factors, including cytomegalovirus infection, GVHD, and cyclosporine treatment potentiated by conditioning regimens (467,468,469).
The dosage of radiation employed, which is most often in the range of 7.5 to 14.0 Gy, is below that usually identified as nephrotoxic when direct renal radiation is administered, and this difference remains puzzling. It may be the consequence of the lesser extent of fractionation employed in total-body irradiation protocols, or it may be the result of some other aspect of the preparatory regimens utilized for BMT.
Cyclosporine, a drug with known potential to cause HUS (135,136,137), is often used to treat the GVHD that frequently occurs in patients who have undergone BMT. Accordingly, this drug, rather than total-body irradiation, has been suspected to contribute to or even to be the principal agent of BMT nephropathy in some patients. As considered in detail by other investigators, careful review of the clinical data provided in reports describing this entity of late renal injury in patients who have undergone BMT reveals that approximately half of affected patients have had no exposure to cyclosporine, a finding making it most unlikely that this drug is the principal nephrotoxin in this disorder (447457). When calcineurin inhibitors were combined with sirolimus following allogeneic hematopoietic stem cell transplantation, the incidence of TMA (10.8%) was significantly higher than in those patients treated without sirolimus (470). In this study, only the use of sirolimus and grade II to IV acute GVHD were associated with TMA in regression analyses.
Descriptions of the renal pathologic features of BMT nephropathy are surprisingly uniform, despite variations in the conditioning regimen used to prepare patients for BMT in different centers and in concurrent laboratory parameters of affected patients (447,451,,457,,472). Whether or not patients have clinical evidence of hemolytic anemia or thrombosis, the renal lesions typically involve glomeruli and small arterial and arteriolar vessels. Glomeruli usually demonstrate increased lobulation, with expansion of the mesangial areas. The expanded mesangium loses the normal appearance of a tightly compacted matrix reactive with silver methenamine stains and assumes a loose, spongiform appearance traversed by thin bands of silver-staining matrix. This loosely organized
mesangium frequently extends into the peripheral capillary walls, where it is often interposed between split basement membranes. The basement membranes, in turn, often reveal a double-contoured appearance with silver methenamine and periodic acid-Schiff. The dissolution of the mesangium, or mesangiolysis, is often accompanied by disruption of the regions where the capillary walls are anchored to the mesangial stalks. The result is a ballooning of the capillary walls into the urinary space, alternately referred to as microaneurysm formation. The capillary lumina often appear narrow or even occluded, apparently as a result of endothelial swelling. Overt glomerular intracapillary thrombi can occasionally be identified, and sometimes red blood cell fragments can be identified in affected capillaries. Less severely injured glomerular capillaries may show only features of widening of the subendothelial space and adjacent mesangial expansion. Acute inflammatory changes, such as influx of neutrophils, segmental necrosis, and crescent formation, typically are not present. Some glomeruli, with their blood supply interrupted by occlusive thromboses in the afferent arterioles and/or interlobular arteries/arcuate arteries, exhibit features of ischemic change characterized by wrinkling and thickening of peripheral capillary walls, foci of capillary collapse, shrinkage of the glomerular tufts, and eventual obsolescence.
Progression of the acute glomerular lesions may be inferred to be similar to that of other TMAs, although biopsy studies that rigorously establish this sequence in patients have not been widely recorded. Areas of mesangial dissolution may be replaced by accumulations of matrix, leading to an appearance of mesangial sclerosis. This sclerosing mesangial process may encroach on capillary lumina, or it may combine with a sclerosing reparative response in adjacent, injured capillary loops to cause focal and segmental glomerulosclerosis. Capillary microaneurysms may become occluded by proteinaceous hyaline material, or alternately they may rupture and collapse, again leading to segmental foci of sclerosis in the glomerular tuft. In some cases, the only sequelae may be persistence of extra layers of basement membrane matrix in the capillary loops. Alternatively, full morphologic resolution of the microthrombotic process in affected BMT-treated patients may also occur within months of the initial injury. In the series of Antignac et al (473), in which two patients underwent repeat biopsy, the nephropathy was markedly progressive, with features of global glomerular sclerosis, glomerular ischemia, and extensive tubular atrophy present in the second biopsy, but in which resolution of mesangiolytic changes and improvement in arteriolar lesions were also reported.
The other principal parenchymal compartment affected is that of the arterial vasculature, especially interlobular arteries and arterioles. The characteristic lesions are again not distinctive in the patients who have undergone BMT but are those common to other forms of TMA. Such changes include mucinous swelling of the subendothelial space, with occasional entrapment of blood cells and red cell fragments in the expanded intima, endothelial swelling, focal intraluminal thromboses, and occasional foci of necrosis involving smooth muscle cells of the arterial vessel walls.
Immunofluorescence findings in this setting are no different from those found in other TMAs and consist of variable but at times prominent deposition of IgM and complement components in both glomerular capillary walls and some affected arterial vessels. Such deposits are not, however, a constant feature in either glomeruli or vessels. Fibrin or fibrinogen is readily detected in almost all cases reported and localizes to intravascular thrombi and the luminal surface of glomerular capillaries and extraglomerular blood vessels. Electron microscopy also reveals features typical of TMAs in general, as described earlier in this chapter, as well as features of radiation nephropathy unassociated with BMT, described in â€œPathologic Findings.â€ In the glomerulus, these features include mesangiolysis, endothelial swelling, focal endothelial denudation from basement membranes, accumulations of finely fibrillar material typical of fibrin tactoids, and overt fibrin- and platelet-rich thromboses. The ultrastructural appearance of arteries includes expansion of the subendothelial zone, which may contain fibrin, elements of blood cells, and variable accumulations of electron-lucent and finely granular, electron-dense material, some of which is likely to represent cellular and thrombotic debris.
Some animal models, principally using rats and mice, have been designed to simulate the process of total-body irradiation used to prepare patients for BMT (428,442, 443,474,475). Several of these protocols further recapitulate the clinical protocols used in patients by subsequent infusion of syngeneic bone marrow cells to avoid the confounding effects of hematopoietic failure that typically occur in this form of cytoreduction therapy. These studies have shown that such models can reproduce the glomerular and, at times, the hematologic features of glomerular injury and TMA described in human patients. These models have been used to dissect the additional effects on the kidney of several of the drugs commonly used in the course of BMT. The results to date stand in some contrast to the human situation. Lawton et al (442) concluded from studies in rodents that no additional nephrotoxic effects on renal function were noted beyond those of total-body irradiation alone for gentamicin, amphotericin, or cyclosporine for periods of up to 6 months after BMT. In this study, busulfan was the only drug tested that exerted an additional nephrotoxic effect. This finding contrasts with the situation in human patients undergoing BMT, in whom the use of gentamicin and amphotericin has been strongly implicated in the acute renal failure that may occur in the first few months after BMT (447), although in humans this effect may be mediated through mechanisms of a hepatorenal syndrome rather than by direct cytotoxicity to renal cells. These
findings indicate possibly important limitations in the use of these rodent protocols to model the human situation.
One interesting and promising result that has arisen from these models is the demonstration of the renoprotective effects resulting from prophylactic administration of ACE inhibitors, particularly captopril (426,,428,476). Some of the beneficial effect observed is almost certainly a consequence of the lowering of blood pressure in irradiated animals. However, the lack of similar efficacy of other antihypertensive agents, such as the calcium channel blocker verapamil, in the same models points to the role of other physiologic pathways in mediating this effect (428). Such pathways may include the ability of ACE inhibitors to lower intraglomerular pressures through their vasodilatory effects on the glomerular arterioles, as well as through potential antioxidant properties present in captopril (460). Individual studies suggest that administration of the antiplatelet and anti-inflammatory agent acetylsalicylic acid (435) and the anti-inflammatory agent dexamethasone (426) also may diminish the nephrotoxic effect of radiation in experimental systems, but such findings have not yet been translated into clinical trials in human patients.
Radiation Nephropathy Caused by Radionuclide Therapy
Radionuclide therapy targets tumors by delivering radioactive particles directly to tumor cells. This is achieved by attaching (chelating) Î²-, Î±-, or Î³-emitting radionuclides to monoclonal antibodies, antibody fragments, or low-molecular-weight oncophilic peptides. Delivery of radioactivity to the target tissue is governed by the specificity of the antibodies or oncophilic peptides to antigenic epitopes or receptors expressed by tumor cells. Radiation injury to the kidney has been observed in association with some of these treatment modalities (reviewed by Lambert et al ).
111I)-diethylenetriamine-penta-acetic acid (DTPA) octreotide and Î² and/or Î³ particle emitters such as yttrium 90 (903)-octreotide (90Y-DOTATOC) or to lanreotide (90Y-DOTALAN) and of lutetium 177 (177Lu) to Tyr3177Lu-DOTATATE) are in use (reviewed by de Jong et al ). The binding activity of various somatostatin octapeptides to various subtypes of somatostatin receptors varies significantly, affecting the efficacy of the treatment. For example, the newer analogue octreotate (DOTA, Tyr3-octreotate) has a ninefold higher activity for the somatostatin receptor subtype 2 as compared with octreotide (DOTA, Tyr3-octreotide) (479). The renal uptake of radioactivity following intravenous administration of radiolabeled octapeptides is one of the highest among various organs, making the kidney a potential target for radiation injury (480).
Nephrotoxicity secondary to treatment with the Î² emitter 90Y-DOTATOC has been reported by various groups (481,482,483). Five of seven patients reported by Moll et al (482) who were administered a normalized cumulative dose of greater than 200 mCi/m2 of 90Y-DOTATOC for advanced neuroendocrine tumors developed chronic renal failure 6 to 16 months after the end of treatment. None of the 22 patients who received a cumulative dose of 200 mCi/m2 developed side effects. Renal biopsies performed in three patients showed typical signs of TMA with glomerular, arteriolar, and small arterial involvement. By light microscopy, the glomeruli revealed capillary loop collapse, focal segmental fibrin thrombi, mesangiolysis, and reduplication of the capillary basement membranes. The arterioles and small arteries showed transmural hyaline deposits and were partially or completely occluded by fibrin and/or protein thrombi and foam cells. Tubular atrophy and interstitial fibrosis were prominent in all cases. Immunofluorescence showed fibrin positivity in the glomeruli and arterial and arteriolar walls and lumen. Electron microscopy exhibited massive thickening of the glomerular capillary walls with widening of the lamina rara interna, new formation of lamina densa material, and activation of the endothelial cells. Another study addressing renal toxicity in patients treated with 90Y-DOTATOC showed that all five patients who experienced more than 20% loss of renal function per year received a biologic effective dose more than 45 Gy (calculated according to the linear quadratic model) (484). However, no renal toxicity was observed in the MAURITIUS trial, which included 154 patients receiving cumulative doses up to 8.58 GBq (5.0 GBq/m2) of 90Y-DOTALAN (485). Likewise, no renal toxicity has been observed with 111In-DTPA-octreotide with cumulative doses up to 100 GBq and renal dose of 45 Gy (486).
No renal toxicity has been reported from clinical trials using various monoclonal antibodies, such as 131I-tositumomab or 90Y-ibritumomab (487,488,489). However, when rhenium 188â€“labeled monoclonal antibodies (188Re-anti-CD66) were applied as part of combined chemotherapy and external-beam total-body irradiation of 12 Gy prior to stem cell transplantation, 6 of 36 patients developed an increase in serum creatinine 6 to 12 months after therapy (490). Radiation nephropathy was diagnosed morphologically in one of six cases.
The higher rate of nephrotoxicity (e.g., radiation nephropathy) in patients treated with radionuclides chelated to small-molecular-weight peptides versus monoclonal
antibodies is likely a result of the high rate of renal catabolism of low-molecular-weight proteins (491). The low-molecular-weight proteins such as octreotide are filtered through the glomerular capillary basement membranes, followed by tubular reabsorption. Proteolytic digestion of the peptides by lysosomal enzymes results in breakdown products, including radionuclide-chelated amino acids. Entrapment of these chelated amino acids within the kidney occurs by binding to intracellular metal binding proteins.
111In, have high ionizing properties over a short range that result in a high relative biologic effectiveness as compared to a similar dose of radiation (). The mean particle range of Auger electrons emitted by 111In is less than one cell diameter, whereas Î² particles emitted by 90Y have a much deeper penetration range (R95 5.7 mm). It has been hypothesized that the absence of significant radiation nephrotoxicity in patients treated with 111In-DTPA compared with 90Y-DOTATOC is a result of the limited range of damage inflicted by 111In.
Because of the negative charge of the glomerular filtration barrier, renal uptake of small-molecular-weight peptides can be suppressed by cationic compounds (491). In addition, competition for the negative charges on the tubular basement membrane between the radiolabeled compound and the cationic compounds may also inhibit renal tubular reabsorption and retention of radiolabeled substances. Experimental data in mice indicate that renal uptake of radiolabeled monoclonal antibodies could be inhibited significantly by cationic amino acids (493). In rats, co-injection of 400 mg/kg L-lysine inhibited renal uptake of 111In-DTPA octreotide by approximately 40% (494). There are several human studies indicating that renoprotection during radionuclide therapy can be achieved by co-infusions of cationic, basic amino acids. In a phase I clinical trial, patients who received two cycles of 90Y-DOTATOC up to 5.55 GBq/cycle, along with lysine with or without arginine infusion immediately before and after therapy, showed no renal toxicity (495). Another study by Waldherr et al (496) showed normal serum creatinine levels during a median follow-up of 6 months in 38 of 39 patients who received 7.4 GBq/m2 of 90Y-DOTATOC in four sessions, along with Hartmann-Hepa 8%, a commercially available mixture containing amino acids. Grade II renal toxicity was observed in only one patient. In a large cohort study of 256 patients, the maximum cumulative dose of 90Y-DOTATOC was defined to be 7.4 to 11 GBq (4.3 to 6.4 GBq/m2) when no renal protection was used and no significant renal functional changes were observed (497498), 35 patients with neuroendocrine gastroenteropancreatic tumors were treated with 177Lu-DOTA-Tyr3-octreotate with cumulative doses of up to 27.8 to 29.6 GBq (750 to 800 mCi; 16.1 to 17.1 GBq/m2). Dosimetric calculations were used to prevent kidney doses exceeding 23 Gy, and renoprotective amino acids were also co-administered. There was no significant increase in creatinine clearance or serum creatinine 3 to 6 months after completion of therapy. However, there are some data indicating that the long-term outcome of renal function in patients treated with peptide receptor radiation therapy may be less favorable, even when renoprotective regimens are applied. A sustained decline in creatinine clearance with an average of 7.3% in patients treated with 90Y-DOTATOC and 3.8% in those treated with 177Lu-DOTATATE was observed with median follow-ups of 2.9 years and 2.3 years, respectively (499). Eleven of 65 patients in this study had more than 15% decline in creatinine clearance per year.
In summary, radiation nephropathy occurs in patients with radionuclide therapy, especially those with peptide receptor radionuclide therapy. Toxicity depends on multiple factors, including the dose of radiation, type of radionuclide, and the physicochemical properties of the carrier molecules. Renal toxicity can be limited by co-administration of positively charged amino acids during therapy; however, gradual renal functional deterioration over the long term may still result.