Causes of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults
Autore: Burton D Rose, MD; James N George, MD
INTRODUCTION — Thrombotic thrombocytopenic purpura-hemolytic uremic syndrome (TTP-HUS) describes acute syndromes with abnormalities in multiple systems. Although some studies appear to distinguish between TTP and HUS, the presenting features are essentially the same in most patients. In a few patients, neurologic abnormalities are dominant and acute renal failure is minimal or not present; these patients are considered by some to represent TTP. When acute renal failure is dominant and neurologic abnormalities are minimal or absent, the disorder is considered by some to represent HUS . However, many patients present with severe neurologic abnormalities, such as seizures and coma, together with acute renal failure; these patients can only be described by the comprehensive term TTP-HUS (see “Evolution of disease definitions” below).
However, all patients are diagnosed primarily by the presence of thrombocytopenia and microangiopathic hemolytic anemia without another clinically apparent cause, and most have some neurologic and renal abnormalities [1-4]. Furthermore, the pathologic changes in patients described as TTP or HUS are identical and the initial treatment of all adult patients is the same — plasma exchange [1,2,5,6]. As a result, these syndromes in adults will be considered together and referred to as TTP-HUS; the one exception is childhood HUS following an episode of severe bloody diarrhea due most often to enterohemorrhagic Escherichia coli (show table 1) [5,6].
The causes of TTP-HUS will be discussed here. The clinical manifestations, diagnosis, and treatment of TTP-HUS are discussed separately. (See “Diagnosis of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults” and see “Treatment of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults”).
EVOLUTION OF DISEASE DEFINITIONS — It is important to understand how these syndromes have changed since their original description.
Thrombotic thrombocytopenic purpura — TTP was first described by Moschowitz in 1925 as a new disease characterized by unique pathological findings of hyaline thrombi in many organs . Amorosi and Ultmann reviewed all 271 published cases up to 1964 and defined the classic pentad of clinical features :
Microangiopathic hemolytic anemia
Neurologic symptoms and signs
Renal function abnormalities
The diagnosis of TTP was supported by the pathologic demonstration of hyaline thrombi in 93 percent of patients; 90 percent of the patients died. However, with the routine use of therapeutic plasma exchange, TTP has changed from a fatal to a curable illness. (See “Treatment of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults”). The availability of curative treatment has created an urgency for diagnosis and therefore the stringency of diagnostic criteria has decreased. Now only the dyad of otherwise unexplained thrombocytopenia and microangiopathic hemolytic anemia (show blood smear 1) are sufficient to establish the diagnosis and initiate treatment; this syndrome has also been called thrombotic microangiopathy [9,10]. This clinical evolution has led to a broader spectrum of TTP .
Hemolytic-uremic syndrome — HUS was first described in 1955 in a report of five children with acute renal failure who died with renal cortical necrosis . However, current childhood HUS is a different disorder, which began in 1980 with the appearance of Shiga toxin-producing Escherichia coli (usually E. coli O157:H7) [13,14]. This disorder is also called typical or diarrhea associated HUS. Children comparable to those in the original report are considered to have atypical or non-diarrhea associated HUS [15,16]. (See “Clinical manifestations and diagnosis of typical hemolytic uremic syndrome in children”). Our current understanding of the pathogenesis and causes of HUS and TTP will be reviewed here. The diagnosis and treatment of these disorders are discussed separately. (See “Diagnosis of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults” and see “Treatment of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults”).
EPIDEMIOLOGY — Based on information collected between 1996 and 2004 from the Oklahoma TTP-HUS Registry, we have estimated the incidence of TTP-HUS as follows :
Suspected TTP-HUS — 11 cases/million population per year
Idiopathic TTP-HUS — 4.5 cases/million per year
Severe ADAMTS13 deficiency — 1.7 cases/million per year
For all three categories, incidence rates were greater for women and for blacks. As an example, for those with severe ADAMTS13 deficiency (ie, <5 percent of normal activity), the age-sex standardized incidence rate ratio of blacks to non-blacks was 9.3 (95% CI 4.3-20). For the 206 patients with suspected TTP-HUS (ie, all patients with an initial episode of clinically suspected TTP-HUS in whom the decision to initiate plasma exchange treatment was made), the clinical setting in which these patients presented included the following:
Idiopathic — 37 percent
Drug-associated — 13 percent
Autoimmune disease — 13 percent
Sepsis — 9 percent
Pregnancy/postpartum — 7 percent
Bloody diarrhea prodrome — 6 percent
Hematopoietic cell transplantation — 4 percent
Of the 27 patients with severe ADAMTS13 deficiency, 24 (89 percent) were deemed to be idiopathic. Of the three remaining patients, two were postpartum and one had HIV infection. PATHOLOGY — TTP-HUS is associated with thrombi composed primarily of platelets in affected organs. In the kidney, for example, the primary histologic changes are thrombi in the glomeruli and arterioles, subendothelial widening of the glomerular capillary wall on electron microscopy due to the deposition of fibrin-like material, and, during healing, mucoid intimal thickening and then onion-skin hypertrophy of the interlobular arteries (show histology 1A-1E) . (See “Mechanisms of immune injury of the glomerulus”).
These histologic changes are consequences or manifestations of endothelial cell injury. The term glomerular capillary endotheliosis that describes this injury was first used to describe the injury in preeclampsia . The pathologic and some of the clinical findings of TTP-HUS are similar to those seen in several other renal disorders, including malignant nephrosclerosis, preeclampsia, scleroderma, chronic transplant rejection, the antiphospholipid syndrome, and rarely cyclosporine and tacrolimus nephrotoxicity and radiation nephritis (see below).
ADAMTS13 deficiency — An underlying cause of and the mechanism for platelet consumption in some patients with TTP-HUS has been elucidated, which may help to distinguish TTP from HUS (see below) [19-23]. Von Willebrand factor (VWf) is synthesized in endothelial cells and assembled in larger multimers that are present in normal plasma. The larger multimers, called unusually large von Willebrand factor (ULVWf), are rapidly degraded in the circulation into the normal size range of VWf multimers by a specific von Willebrand factor-cleaving protease (or cleaving metalloproteinase, now called ADAMTS13: A Disintegrin-like And Metalloprotease with ThromboSpondin type 1 repeats) [24-27].
Employing a genome-wide linkage analysis, the responsible locus among four pedigrees with congenital TTP was identified on chromosome 9q34, an area subsequently found to contain a gene that encodes for a metalloproteinase that is thought to be the cleaving protease . This study identified 12 mutations in the ADAMTS13 gene, accounting for 14 of the 15 disease alleles studied in individuals with very low (2 to 7 percent of normal activity) protease activity. Numerous gene mutations in the ADAMTS13 gene have been described .
ADAMTS13 deficiency could lead sequentially to the accumulation of ULVWf multimers, platelet aggregation, and the platelet clumping that is characteristic of the disease . ULVWf multimers (including unique forms arising from proteolytic digestion ) accumulate in patients with TTP, being found in the platelet thrombi and serum [3,32,33]. These ULVWf multimers can attach to activated platelets, thereby promoting platelet aggregation. (See “Biology and normal function of von Willebrand factor”).
ULVWf multimer accumulation in TTP has been associated with absent or markedly diminished ADAMTS13 activity due to an inherited or acquired deficiency [1,3,19,20,28,34-37]. However, absent ADAMTS13 activity without manifestations of a thrombotic microangiopathy has been reported in several families with congenital TTP [38,39]. It is clear from these observations that absence of ADAMTS13 activity is a necessary but not sufficient condition in order to bring about the clinical syndrome of TTP in these family members.
Inhibitory antibody — An inhibitory autoantibody to the ADAMTS13 metalloproteinase has been found at varying titers among a high percentage of patients with the idiopathic form of this disease [40,41]. The inhibitory IgG is postulated to be directed against varying elements of the protease . It is unclear why the antibodies recognize this complex and why they are produced, although it has been suggested that levels of ADAMTS13 less than five percent with or without an inhibitor antibody may be part of a larger autoimmune response (see “Autoimmunity in idiopathic TTP” below) .
TTP versus HUS — It is not clear whether ADAMTS13 levels can be used to distinguish TTP from HUS. In one series, for example, ADAMTS13 activity was evaluated in 30 and 23 patients with TTP and HUS, respectively . The following results were noted:
All patients with TTP displayed some deficiency in ADAMTS13 activity. The six patients with familial TTP completely lacked activity; among the 24 patients with nonfamilial disease, 20 had severe (<5 percent of normal activity) and four had moderate (5 to 25 percent) ADAMTS13 deficiency when assessed during the acute event.
An inhibitor to ADAMTS13 was found in 20 of the 24 patients with nonfamilial TTP (83 percent). This inhibitor was found to be an IgG molecule in five of five samples tested; by comparison, none of those with familial TTP had a circulating inhibitor.
ADAMTS13 activity was normal in 21 of 23 patients with HUS (93 percent). A mild decrease was observed in two patients with nonfamilial disease.
In a separate study of 29 children with diarrhea-associated HUS, ADAMTS13 activity was normal in 26, slightly reduced in two, and severely reduced in one . The last patient, while testing positive for infection with E. coli O157, had clinical features more consistent with TTP (eg, fluctuating neurologic signs, fever, and abnormal liver function). In a second study in 83 children with hemolytic or thrombocytopenic episodes with or without neurologic symptoms or renal failure, none of the 24 children with a clinical diagnosis of HUS had an ADAMTS13 level <62 percent . In contrast, all eight children with a clinical diagnosis of TTP had levels 5 percent. Two of fifty children with a prior diagnosis of idiopathic thrombocytopenic purpura and one child with a prior diagnosis of Evans’ syndrome were also found to have ADAMTS13 deficiency, confirming that patients with TTP may have only thrombocytopenia and hemolytic anemia, without neurologic or renal abnormalities.
Additional findings in support of the role of ADAMTS13 activity were provided in a second study of 37 patients with TTP . In this report, 39 plasma samples obtained during an acute event displayed a severe deficiency of protease activity. By comparison, 16 plasma samples from those in remission and 76 samples from normal subjects had normal activity. Twenty six of the 39 samples with decreased activity were found to have IgG antibodies with inhibitor properties.
Based upon the preceding observations, the underlying mechanism of HUS does not appear to be related to a deficiency in ADAMTS13 activity [19,34,45,46]. However, these data require confirmation in a prospective case series to validate the apparent distinction between TTP and HUS, in which the initial diagnosis was made clinically by the physicians providing the patient samples.
As noted above, when neurologic abnormalities are dominant and acute renal failure is minimal or absent, some consider TTP to be present. When acute renal failure is dominant and neurologic abnormalities are minimal or absent, the disorder is considered by some to represent HUS (show table 2) . Current information clearly supports initial data that protease deficiency is most commonly present in patients with neurologic findings and no renal failure (ie, clinical TTP) [19,21].
ADAMTS13 activity in other conditions — ADAMTS13 activity is typically absent or severely deficient (ie, less than five percent of normal) in patients with TTP [40,47]. Less pronounced reductions in protease levels have been noted in a number of circumstances, most of which are not confused clinically with TTP. These include [4,21,40,47-52]:
Disseminated intravascular coagulation
Idiopathic thrombocytopenic purpura
Hematopoietic cell transplantation and cyclosporine
Infection, severe sepsis
Systemic lupus erythematosus
Heparin induced thrombocytopenia
Leukemia, other neoplasms
Acute inflammatory states
Some studies have noted an inverse relationship between plasma levels of the protease and von Willebrand factor (vWF) antigen and collagen-binding activity [50,51,53], suggesting that ADAMTS13 activity levels may fall as a response to the liberation of ultralarge vWF multimers into the circulation in these settings.
An intriguing observation is that protease levels appear to decrease during the last two trimesters of pregnancy, being lowest at 36 to 40 weeks of gestation and the early puerperium [50,54]. This observation may be relevant to the increased frequency of TTP-HUS during pregnancy, especially at the time of delivery. (See “Acute renal failure in pregnancy”).
ADAMTS13 levels can also be influenced by previous transfusions. These can mask the presence of severe deficiency, since the half-life of ADAMTS13 in plasma is 2 to 3 days . Other factors — Several other hypotheses have been proposed for the pathogenesis of TTP-HUS in patients lacking ADAMTS13 deficiency [2-4]. These include endothelial injury, increased platelet aggregation, and genetic factors.
Endothelial injury — Platelet activation in TTP-HUS may be a secondary response to endothelial injury [56,57]. The endothelial damage could be directly induced by a drug (see “Cancer, cancer chemotherapy, and HCT” below and see “Ticlopidine” below and see “Quinine” below) or indirectly via neutrophil or platelet activation [58,59].
A related question is the propensity of only certain vascular endothelial beds to be damaged in TTP-HUS. The renal and cerebral vessels are commonly involved, while the pulmonary and hepatic microvasculature are usually spared. This pattern was reproduced in one study which evaluated the ability of plasma from adult patients with TTP-HUS to induce apoptosis in microvascular endothelial cells of diverse tissue origin . When incubated with patient plasma, cells of dermal, renal, and cerebral origin were susceptible to apoptotic death in vitro, while those of pulmonary and hepatic origin were not.
These differences in vascular endothelial damage might be explained by the anatomic restriction of CD36, the thrombospondin receptor (also known as glycoprotein (GP) IV when found on the platelet surface). CD36 is found on human microvascular endothelial cells (HMVEC), but not endothelial cells of large (umbilical) vessels (HUVEC). In a series of patients with classical TTP-HUS, antibodies were found to HMVEC and HUVEC in 73 to 100 percent and 36 to 55 percent, respectively . Antibody reactivity in these patients was significantly greater against HMVEC than HUVEC, was inhibited by pretreatment with a heterologous anti-CD36 antibody, and was reduced following successful plasma exchange therapy.
Shiga toxin — Most cases of childhood HUS and occasional cases in adults follow an episode of bloody diarrhea; a Shiga toxin released from E. coli (particularly type O157:H7) or other bacteria is associated with the development of HUS in most of these patients (see “Shiga toxin-producing E. coli” below and see “Microbiology, pathogenesis, and epidemiology of enterohemorrhagic Escherichia coli”, section on Shiga toxins) .
As noted above, ADAMTS13 activity is typically normal in affected children . The comprehensive term TTP-HUS is appropriate in the sporadic cases in adults, since some adults present with acute thrombocytopenia and microangiopathic hemolytic anemia with no renal function abnormalities following hemorrhagic colitis caused by E. coli O157 .
Multiple factors may contribute to the development of TTP-HUS in this setting. Endothelial injury has been attributed to multiple factors including direct toxin damage , neutrophil accumulation, and increased release of endothelins , and other cytokines and chemokines . Neutrophil accumulation may be induced by one or more of the following mechanisms: toxin-induced activation of leukocyte adhesion molecules (such as E-selectin) , toxin-induced secretion of proinflammatory substances (including interleukin-8 and platelet activating factor) [66,67], or inhibition of apoptosis [68,69]. Leukocytosis is common in diarrhea-induced HUS and the prognosis tends to be proportional to the degree of elevation in the white cell count; these observations are compatible with a pathogenetic role for neutrophil activation .
Shiga toxin may also have a variety of other effects:
Colonic vascular injury, allowing endotoxin and other inflammatory mediators to gain access to the systemic circulation [62,70]. The amount and rate of endotoxin absorbed may modulate disease expression .
Directly promote platelet aggregation, which is characteristic of TTP-HUS 
Upregulate tissue factor activity in proximal tubular cells, which may initiate activation of the coagulation pathway .
Indirectly sensitize renal epithelial cells to increased heme toxicity .
The prominent involvement of the kidney in postdiarrheal HUS suggests that Shiga toxins have a predilection for the renal circulation. Two different observations provide evidence in support of this hypothesis:
The glycolipid receptors (Gb3) in the endothelial cell membrane that bind toxin appear to be preferentially expressed in the kidney and the glomeruli [74,75]. As an example, expression in renal endothelial cells is 50 times greater than in umbilical vein endothelial cells . Differential localization of this receptor in infant versus adult kidneys may be partly responsible for the enhanced risk of the disease among children [76,77].
The bacterial toxin appears to selectively stimulate the release of tumor necrosis factor within the kidney . How this occurs is not clear, but TNF may promote vascular injury in part by increasing expression of the glycolipid receptors .
The Shiga toxin can also cause apoptosis of and/or significant damage to renal tubular epithelial cells [79-81]. The degree to which tubular injury contributes to renal failure in postdiarrheal HUS are unknown.
Plasminogen activator inhibitor — Another factor that may underlie platelet aggregation in TTP-HUS is the presence of plasminogen-activator type 1. Increased levels of plasminogen activator inhibitor type 1 (PAI-1), the primary inhibitor of the fibrinolytic compounds tissue-type plasminogen activator and urokinase, has been described in children with postdiarrheal HUS [82,83]. More severely affected patients had a higher plasma PAI-1 concentration, and normalization of plasma levels spontaneously or by peritoneal dialysis was correlated with an improvement in renal function. It is not clear, however, if the enhanced release of PAI-1 was a primary response or was secondary to endothelial injury. A primary elevation in this factor might be expected to lead to fibrin thrombi rather than the platelet thrombi characteristic of TTP-HUS. (See “Thrombotic and hemorrhagic disorders due to abnormal fibrinolysis”, section on Plasminogen activator inhibitor).
Lack of a platelet inhibitor — Decreased circulating levels of a normal inhibitor of platelet aggregation, similar to the protease previously described, could also explain both the formation of platelet thrombi and the occasional reversal of the disease following plasma infusions. (See “Treatment of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults”).
Genetic factors — Autosomal recessive and dominant forms of familial HUS have been rarely described and are predominantly observed in children. A number of studies have identified mutations in complement regulatory proteins, with abnormalities in other proteins not related to the complement system also appearing to be causative. This is discussed in detail separately. (See “Clinical manifestations and diagnosis of atypical hemolytic uremic syndrome in children”)
In adults, the possible predisposing effect of factor V Leiden was suggested by the observation in one report that the prevalence of this abnormality was significantly increased among white patients with thrombotic microangiopathy (TM) and normal levels of ADAMTS13 (4 of 11) compared with white controls (6 of 186, odds ratio: 17.1, 95 percent confidence interval 5.4 to 54) . Factor V Leiden was not found in 19 white patients with TM and intermediate or low levels of the protease, or in any of 12 African-American patients. The prevalence of a number of other inherited thrombophilias did not differ between subjects with TM and controls.
Autoimmunity in idiopathic TTP — The frequent occurrence of a TTP-like picture in patients with systemic lupus erythematosus or other autoimmune disorders has suggested that idiopathic TTP with undetectable levels of ADAMTS13 may represent an autoimmune disorder (see “Antiphospholipid antibodies” below).
As an example, one study evaluated clinical and laboratory manifestations in 46 patients with idiopathic TTP-HUS . When compared to the 15 patients with detectable ADAMTS13 activity (ie, >25 percent of normal), the 31 patients with undetectable levels (ie, <5 percent of normal) had a higher incidence of autoimmune manifestations, including presence of an ADAMTS13 inhibitor in 55 percent and antinuclear antibodies in 71 percent.
In contrast, those with detectable ADAMTS13 activity had mainly severe renal failure without findings suggestive of an autoimmune disorder, more consistent with a diagnosis of HUS. However, as noted in our own studies, the level of ADAMTS13 in such patients appears to have no prognostic value in terms of response to treatment with plasma exchange, relapse, or survival. Presence of an antibody to ADAMTS13 may, however, suggest the benefit of adding corticosteroids to the treatment program. (See “Diagnosis of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults”, section on vWF cleaving protease (ADAMTS13) deficiency and see “Treatment of thrombotic thrombocytopenic purpura-hemolytic uremic syndrome in adults”, section on Corticosteroids).
ETIOLOGY — Although many cases of TTP-HUS are idiopathic, a variety of underlying causes have been identified (show table 1) [2-4,85,86]:
Shiga toxin-producing E. coli — Most cases of childhood HUS and occasional cases in adults follow an episode of bloody diarrhea (called typical HUS in children) . (See “Clinical manifestations and diagnosis of typical hemolytic uremic syndrome in children”, section on Typical HUS).
A Shiga toxin released from enterohemorrhagic E. coli (EHEC, particularly type O157:H7) or other bacteria appears to be associated with the development of HUS in most of these patients [62,87]. Both endothelial injury and platelet activation may contribute to Shiga toxin-induced HUS. The prominent involvement of the kidney in postdiarrheal HUS suggests that Shiga toxins have a predilection for the renal circulation. (See “Shiga toxin” above). It has been estimated that EHEC cause at least 70 percent of cases of postdiarrheal HUS in the United States, and that 80 percent of these cases are caused E. coli O157:H7 . In contrast, the majority of cases of postdiarrheal HUS in Australia are due to non-O157 Shiga toxin producing E. coli; O157 infections are rare.
In a large and severe outbreak of E. coli O157:H7 infection in the United States, 9 percent of cases developed HUS . Because of the clinical significance of this strain of E. coli, efforts to develop a vaccine against this organism are currently underway . (See “Microbiology, pathogenesis, and epidemiology of enterohemorrhagic Escherichia coli”). It is not clear if early antibiotic administration for diarrhea caused by E. coli O157:H7 increases the risk of developing HUS, perhaps due to enhanced toxin release as the bacteria are killed . (See “Diarrheagenic Escherichia coli”).
Extrarenal involvement is common in children with postdiarrheal HUS . As an example, neurologic symptoms, including seizures, are seen in up to one-third of affected children. In addition to local thrombus formation, some neurologic abnormalities may reflect other factors such as cerebral edema or excess drug accumulation due to the impaired renal function . One such class of drugs may be antimotility agents given to minimize the diarrhea; it has been proposed, although not proven, that these drugs may also exacerbate the HUS by decreasing fecal excretion of E. coli, thereby increasing exposure to the toxin. Pancreatitis is another extrarenal manifestation, occurring in approximately 20 percent of cases . Although pancreatic involvement is often subclinical, permanent insulin-dependent diabetes mellitus can occur.
Quinine — Quinine, used in over-the-counter preparations to treat muscle cramps, has been associated with recurrent episodes of TTP-HUS [93-95]. Quinine-dependent antiplatelet antibodies appear to be responsible for the platelet consumption; both antierythrocyte and antineutrophil antibodies also may be present.
In our experience, quinine is the most common cause of drug-induced TTP-HUS, being responsible in 14 of 20 patients in whom a specific drug could be implicated . A careful history is often required to establish the presence of quinine-induced disease. In some cases, for example, initial or recurrent episodes can be triggered by quinine contained in beverages or nutrition health products (eg, tonic water, quinine water, or listed in other products as cinchona, show table 3) [94-96]. In one series, cessation of quinine, institution of plasma exchange, and hemodialysis (required in seven patients) led to recovery with no residua in all nine cases .
In our series of 17 patients, all were females who had been taking quinine intermittently for nocturnal leg cramps for many years . Most presented with the acute onset of anuria, nausea and vomiting, chills and fever, abdominal pain, and diarrhea. Neurologic symptoms, most commonly confusion and lethargy, were present in 12. The frequency and severity of neurologic abnormalities and the occasional absence of renal failure among these patients suggest that the disorder should be termed TTP-HUS, rather than HUS as in previous reports .
Hemodialysis was required in 14, chronic renal failure developed in eight, and there were four deaths, three of which occurred during the initial episode. This experience contrasts with a previous report which suggested that the prognosis for complete recovery was good . Two of the 13 survivors have had a total of three relapses after repeat use of quinine. If recurrences are to be prevented, clinicians must be alert for quinine ingestion as a possible cause of TTP-HUS.
Cancer, cancer chemotherapy, and HCT — TTP-HUS has been associated with adenocarcinomas of the breast, gastrointestinal tract, pancreas, or prostate [97,98] and has also been described as a complication of cancer therapy. This primarily occurs with one of four regimens: mitomycin C; cisplatin with or without bleomycin; gemcitabine; and the use of radiation and high dose chemotherapy prior to hematopoietic cell transplantation (HCT) [99-107].
It is unclear whether TTP-HUS is a complication of allogeneic HCT per se, since the thrombotic microangiopathy seen in these patients can often be attributed to prior chemotherapy, drug toxicity, total body irradiation, grade III-IV acute graft-versus-host disease, hepatic venoocclusive disease, and/or disseminated fungal, bacterial, and/or viral infections [105,106]. (See “Renal failure following hematopoietic cell transplantation”). It is presumed that direct endothelial injury is the initiating event in these settings . However, the onset of clinically evident disease is often delayed, frequently occurring months after chemotherapy has been discontinued or HCT has been performed [101,108,109]. Affected patients typically present with slowly progressive renal failure, new or exacerbated hypertension , and a relatively bland urine sediment, often occurring in the absence of clinically apparent tumor. The syndromes resembling TTP-HUS following mitomycin C and other chemotherapeutic agents such as cyclophosphamide appear to be a direct toxicity related to the cumulative dose of the drug.
There is no evidence that plasma exchange, which is used in most patients with TTP-HUS, is beneficial in disease that is induced by chemotherapy or is associated with hematopoietic cell transplantation. (See “Renal failure following hematopoietic cell transplantation”, section on TTP-HUS).
Antiphospholipid antibodies — Antiphospholipid antibodies (such as the lupus anticoagulant) are autoantibodies directed against anionic phospholipid antigens, such as cardiolipin. (See “Clinical manifestations and diagnosis of antiphospholipid syndrome”). Most patients present with venous and arterial thromboses or frequent abortions, but a thrombotic microangiopathy similar to TTP can occur, with thrombocytopenia, hemolytic anemia, and fragmented red cells [110-113]. Although other symptoms and signs of lupus are often present, the lupus anticoagulant can occur as an isolated event.
In some cases, a TTP-HUS like syndrome occurs in systemic lupus erythematosus in the absence of antiphospholipid antibodies . These cases may be pathogenetically similar to idiopathic TTP-HUS. In patients with the antiphospholipid syndrome or systemic lupus erythematosus, the distinction from TTP-HUS may not be possible and plasma exchange treatment is appropriate . Autopsy evidence for TTP is common in patients who have died with systemic lupus erythematosus and neurologic abnormalities . (See “Types of renal disease in systemic lupus erythematosus”, section on Vascular disease).
Pregnancy and oral contraceptives — TTP-HUS may be related to pregnancy or the postpartum period  and can also be induced by the use of estrogen-containing oral contraceptives . In addition to de novo disease in pregnant women, TTP-HUS that initially occurred in nonpregnant women has relapsed during a subsequent pregnancy and recurrent TTP-HUS has developed during successive pregnancies [117-119].
In many case series of TTP-HUS, 10 to 25 percent of patients were pregnant or in the postpartum period [120,121]. However, the incidence of TTP-HUS among all pregnancies is only 1 in 25,000 . The time of onset is variable. In one report of 13 pregnancies complicated by TTP-HUS, three developed before midpregnancy, eight peripartum, and two several weeks postpartum .
When developing during or after pregnancy, TTP-HUS must be distinguished from severe preeclampsia accompanied by acute thrombocytopenia, a disorder which tends to resolve spontaneously within several days after delivery. (See “Acute renal failure in pregnancy” section on Thrombotic microangiopathy for a brief review on how to differentiate between these disorders).
Immunosuppressive agents — TTP- HUS may be induced by cyclosporine therapy [122-124]. Although direct endothelial injury probably plays an important role in this setting , cyclosporine may also increase platelet aggregation . This form of TTP-HUS is often, but not always, reversible with discontinuation of cyclosporine [123,125].
Reports suggest that the immunosuppressive agent tacrolimus (FK506) can also produce this complication [126,127]. Other studies, however, have found that, among renal allograft recipients with TTP-HUS who are taking cyclosporine, switching to tacrolimus is associated with a significantly high rate of graft salvage .
HUS has also been reported with the combination cyclosporine/sirolimus immunosuppressive regimen. (See “Sirolimus in renal transplantation”).
Two other factors can cause an HUS-like picture following renal transplantation; in each case, thrombus formation is limited to the kidney, rather than being a systemic phenomenon. One is hyperacute humoral rejection that is now rarely seen because of more sensitive typing procedures. The other is the administration of the anti-T cell monoclonal antibody muromonab-CD3 (OKT3) [129,130]. There is a first-dose effect with OKT3 to stimulate coagulation, a response that may be mediated by the release of tumor necrosis factor-alpha (and perhaps other cytokines) from circulating mononuclear cells . The risk appears to be greatest in patients receiving high doses (10 mg/day for 2 weeks versus the more standard dose of 5 mg/day), occurring in 10 percent of patients in one study . The development of HUS has also been limited thus far to patients receiving OKT3 as prophylaxis against rejection; in this early posttransplant period, endothelial injury from ischemia and/or surgery may contribute to the thrombotic process . (See “Major side effects associated with OKT3”).
Antiplatelet agents — Two antiplatelet agents, ticlopidine and clopidogrel, used in the treatment of a variety of cardiovascular disorders have been associated with the development of TTP-HUS [133-137].
Ticlopidine — The reported incidence of TTP-HUS following ticlopidine use for cardiac stenting has ranged from 1 case in 1600 to 1 in 4800 [136,138]. In a series of 98 cases of TTP-HUS associated with ticlopidine from the Food and Drug Administration’s MedWatch program; all cases occurred within twelve weeks; and 75 percent between three and twelve weeks .
Uncontrolled observations suggests that treatment with plasmapheresis may be associated with decreased mortality [135,136,138]. In two series consisting of 158 cases of TTP-HUS associated with ticlopidine, the mortality rate in patients receiving plasmapheresis was significantly lower than in those who did not undergo this procedure (24 versus 50 percent and 18 versus 57 percent, respectively) [135,136].
The cause of TTP-HUS following the use of ticlopidine appears to be unrelated to the antiplatelet activity of this agent [46,134], and multiple factors may contribute. In a study of seven patients who developed TTP-HUS two to seven weeks after initiation of ticlopidine therapy, the following abnormalities were found :
Binding of von Willebrand factor (vWF) to single platelets was increased in the three patients tested during the most thrombocytopenic phase of their episodes of TTP-HUS Initial plasma samples from all seven patients lacked the high molecular weight VWF multimers and were severely deficient in the VWF cleaving protease (ADAMTS13)
IgG molecules isolated from plasma in five of these patients inhibited VWF metalloproteinase in control plasma
The above abnormalities resolved upon the clinical remission that accompanied plasma exchange and discontinuation of ticlopidine
In a second report, plasma from two of five ticlopidine-linked TTP-HUS patients demonstrated inhibition of the VWF cleaving metalloproteinase in vitro . In addition, all five plasmas induced apoptosis in primary cultures of human dermal, glomerular, and hepatic microvascular endothelial cells (see “Endothelial injury” above).
Clopidogrel — Although phase III clinical trials reported no incidence of TTP-HUS with clopidogrel, its subsequent widespread use has been associated with the infrequent development of TTP-HUS [137,141,142]. A review utilizing multiple sources uncovered a total of 37 probable or possible cases of clopidogrel-associated TTP-HUS . Clopidogrel had been prescribed for 2 weeks in 65 percent of the cases. Those receiving plasma exchange within three days of TTP-HUS onset had a significantly better survival than those treated after three days (100 versus 27 percent, respectively). Four patients required 20 exchanges, and relapses occurred in three.
Since clopidogrel and ticlopidine are similar molecules, it is possible that the cause of TTP-HUS with both agents is similar (see “Ticlopidine” above). In fact, immunoglobulin inhibitors of ADAMTS13 activity were detectable in two patients with TTP-HUS following use of clopidogrel, a finding that disappeared with remission in one individual .
HIV infection — Thrombotic microangiopathy has been reported in patients with HIV infection, although its incidence in the post-HAART era appears to be declining [143,144]. The mechanism by which thrombotic microangiopathy occurs in this population is not understood, but direct endothelial injury may be involved .
Thrombotic microangiopathy in HIV-infected patients is most commonly observed among young men (mean age of 35 to 40 years), particularly Caucasians [144,146], and appears to be associated with advanced disease. As an example, in the CHORUS study of 6022 patients with HIV infection, the 17 patients (0.3 percent) with thrombotic microangiopathy had the following significant characteristics in comparison with those who did not have thrombotic microangiopathy :
Lower mean CD4+ cell counts
Higher HIV-1 RNA levels
A higher incidence of AIDS
Higher incidences of infection with Mycobacterium avium complex and hepatitis C
Despite treatment, one-year survival is very low due to the severity of the underlying disease [143,144,146,147]. However, in our own experience, TM in HIV-infected patients has most often been associated with the presence of opportunistic infection or other illness .
Valacyclovir — A hemolytic-uremic syndrome-like syndrome has been reported with high dose valacyclovir prophylaxis (8 g/day) against cytomegalovirus infection in patients with advanced HIV infection [149,150]. In a large trial, thrombotic microangiopathy occurred in 14 of 523 patients receiving valacyclovir compared to 4 of 704 patients randomized to acyclovir (2.7 versus 0.6 percent) . The precise role of valacyclovir, if any, remains unclear since the patients were receiving many other drugs.
Pneumococcal infection — Although most cases of postinfectious HUS occur after an episode of bloody diarrhea, rare patients have developed this disorder after a pneumococcal infection . How this might occur is not known. Plasma exchange or infusion should be avoided in the HUS associated with pneumococcal infection. In this setting, the plasma may contain antibodies against the Thomsen-Friedenreich antigen, which can accelerate polyagglutination and hemolysis .
ADDITIONAL INFORMATION — Additional information concerning thrombocytopenic conditions (ie, drug-induced, ITP, TTP-HUS, and thrombocytopenia in pregnancy) can be found on a website maintained and updated by Dr. James N. George at the University of Oklahoma Health Sciences Center: moon.ouhsc.edu/jgeorge.