Understanding Hypercoagulability with Nephrotic Syndrome: How the Clot Thickens

Thromboembolism is a life-threatening complication of nephrotic syndrome with variable reported incidence for deep venous thrombosis (approximately 15%), pulmonary embolus (10%–30%), and renal vein thrombosis (25%–37%). Complications after the initial event include recurrent thromboembolism, post-thrombotic syndrome, and chronic thromboembolic pulmonary hypertension. Thromboembolism is more common in adults (25%) compared with children (3%), with higher childhood incidence in congenital (10%) and secondary nephrotic syndrome (17.1%).1 Patients with pulmonary emboli or renal vein thrombosis may be asymptomatic or present with the acute onset of symptoms. Younger adult patients with renal vein thrombosis often develop sudden flank pain and gross hematuria, whereas older adults tend to be asymptomatic. In adults, thromboembolism usually occurs within 6 months after nephrotic syndrome is initially diagnosed.2 The incidence of arterial thromboembolism is less than venous thromboembolism, but it remains significant. Mahmoodi and colleagues found an eight-fold higher risk of arterial thrombi in 298 patients with primary and secondary nephrotic syndrome compared with the general population over a 10-year time frame. Traditional risk factors of atherosclerotic disease were predictors of arterial thrombi.2 Membranous nephropathy, minimal change disease, focal segmental glomerulosclerosis, membranoproliferative glomerulonephritis, membranous lupus nephritis with antiphospholipid antibody, and amyloidosis are glomerular diseases with particularly high rates of venous thromboemboli. Studies have shown hypoalbuminemia to be the strongest predictor of risk for venous thromboembolism. A serum albumin level of <2.8 g/dl was associated with a 2.5-fold higher risk of venous thromboembolism compared with lower values in a cohort of 898 patients with membranous nephropathy.3 Each 1.0 g/dl reduction in serum albumin resulted in doubling the risk of venous thromboembolism. A ratio of proteinuria to serum albumin has been found to be more predictive of venous thromboembolic events than serum albumin alone.2 The hypercoagulable state in nephrotic syndrome is postulated to stem from multiple factors, including genetic predisposition; increased number, activation, and aggregability of platelets; and localized clotting activation in the kidney. It is generally accepted that the increased hypercoagulability in nephrotic syndrome is largely due to glomerular loss of anticoagulant and profibrinolytic proteins coupled with increased liver synthesis of procoagulant proteins (fibrinogen, factor V, factor VIII). Marked urinary loss of antithrombin III has been postulated to be an important cause of the imbalance of anticoagulants and procoagulants1. However, studies have shown both decreased and normal levels of antithrombin III in patients with nephrotic syndrome. In this issue of CJASN, Abdelghani and colleagues4 examine the contribution of antithrombin III deficiency to the hypercoagulable state of nephrotic syndrome in 208 patient samples. Patient samples were collected from three incident nephrotic syndrome cohorts: the Nephrotic Syndrome Study Network (NEPTUNE [n=147]), the Pediatric Nephrology Research Consortium (PNRC [n=38]), and the Columbus cohort (n=23). The authors measured both antithrombin III antigen levels using two immunoassays and antithrombin activity by a functional assay. Clinically relevant antithrombin deficiency has been previously defined as <70% of antigen level or activity, and this was used as the threshold for deficiency.5 These authors previously demonstrated in rat models that antithrombin activity, but not antigen levels, correlated with proteinuria and thrombin generation,6 indicating thrombotic risk may correlate with decreased qualitative antithrombin activity rather than increased urinary loss. Hypercoagulopathy was measured by endogenous thrombin potential (ETP), a thrombin generation assay. ETP was previously used by these authors to estimate nephrotic syndrome hypercoagulopathy, and prior results demonstrated correlation with nephrotic syndrome severity.7 In addition, this study also included meta-analyses of 27 studies of nephrotic syndrome that measured antithrombin III antigen levels, activity, or both. These meta-analyses assessed whether antithrombin III deficiency (<70% antigen or activity) was associated with active nephrotic syndrome in adult and pediatric cohorts. The authors found that antithrombin antigen and activity levels were not consistently related to plasma albumin or proteinuria, which are established clinical markers of higher risk of thromboembolism, as mentioned above. Antithrombin antigen level, but not activity, was related to the hypercoagulopathy in adult nephrotic syndrome. Antithrombin activity was not consistently associated with hypercoagulability in childhood nephrotic syndrome. No difference in hypercoagulopathy, measured by ETP, was seen in the plasma of those with normal versus deficient antithrombin III levels. Even lower antithrombin III levels divided by quartile failed to show correlation with hypercoagulopathy. In addition, ex vivo supplementation of antithrombin to severely antithrombin III antigen-deficient (33.8±0.78%) plasma samples did not change nephrotic syndrome hypercoagulopathy. The meta-analyses of studies showed that antithrombin III deficiency using antigen or activity of <70% was not a uniform feature for nephrotic syndrome. It was more common in children than adults. The authors concluded that antithrombin does not play a significant role in the hypercoagulopathy of nephrotic syndrome. It is worth mentioning that samples from the pediatric cohort PNRC did show correlations between antithrombin activity and plasma albumin and ETP, but this was not demonstrated in the pediatric NEPTUNE subcohort. One possible reason was a difference in storage of plasma in the PNRC cohort that could have resulted in lower antithrombin III levels in this cohort. Of note, most of the patient samples (70.7%) in this study were from the NEPTUNE cohort, and 37% of these patients were on immunosuppression compared with none of the patients in the PNRC or Columbus cohorts. This might have been one reason for the large difference in median baseline proteinuria of 1.9 (0.3–3.6) g/g in NEPTUNE versus 9.84 (5.43–18.49) g/g in the PNRC cohort and 5.2 (1.2–13.9) g/g in the Columbus cohort. The median serum albumin levels were 3.2 (2.4–3.7) g/dl, 2.21 (1.85–2.84) g/dl, and 3.8 (3.5–4.4) g/dl in the NEPTUNE, PNRC, and Columbus cohorts, respectively. It is possible there would have been a stronger correlation between antithrombin deficiency with hypercoagulopathy if samples from NEPTUNE patients treated with immunosuppression were excluded. Importantly, the study uses ETP to estimate the risk of hypercoagulopathy rather than measuring antithrombin III levels in actual thromboembolic events. Antithrombin III deficiency, measured by either antigen levels or activity, does not appear to be the major cause of higher risk of thromboembolism in nephrotic syndrome. Studies are inconsistent when measuring levels of procoagulants and anticoagulants in nephrotic syndrome.1 Protein S deficiency has been postulated to play a role, but measurements of levels are not straight forward. Protein S is bound to C4b-binding protein, a large protein (570 kd) that is part of the complement system. Free protein S is a cofactor for protein C. Loss of free protein S in the urine coupled with preserved C4b-binding protein may lead to low free protein S levels, although studies do not consistently demonstrate this finding. Platelet number, activation, and aggregability through elevations of von Willebrand factor, hyperfibrinogenemia, hypercholesterolemia, and hypoalbuminemia in nephrotic syndrome may be increased and likely contribute to the hypercoagulable state.8 Genetic predisposition to venous thromboembolism may play a role in higher risk of patients with nephrotic syndrome. Nephrotic patients have been identified with mutations in Factor V Leiden, prothrombin G20210A, and methylenetetrahydrofolate reductase, and other genes that increase thrombotic risk. Other factors leading to increased thromboembolic events include inflammation and the presence of central venous catheters. Finally, alterations in the fibrinolytic system in nephrotic syndrome may lead to a reduced capacity to dissolve thrombi. Plasminogen and tissue plasminogen activator may be reduced, whereas alpha 2 macroglobulin and lipoprotein A, which inhibit fibrinolytic activity, are increased.1 In addition, experiments using plasma from animal models of nephrotic syndrome and nephrotic patients demonstrated thrombi with a denser fibrin network that had increased resistance to fibrinolysis compared with typical thrombi. The density of the fibrin network was proportional to disease severity being increased by higher levels of proteinuria and lower levels of serum albumin.9 One unanswered question is why membranous nephropathy is associated with the highest risk for developing venous thromboembolism.1,9 In 1313 patients with nephrotic syndrome, the adjusted hazard ratio for venous thromboembolism was 10.8 for membranous nephropathy compared with IgA nephropathy.10 There is presently no identifiable role for anti–phospholipase A2 receptor antibody in the development of thromboemboli. The authors point to the potential role of endothelial damage from antithrombin deficiency.4 Cytokine profiles of some membranous nephropathy patients in China exposed to fine air particulate matter demonstrated high IL-17A levels after in vitro stimulation.11 Elevated IL-17 A levels have been associated with thromboemboli in animal models, and membranous patients with higher levels IL-17A in this study had more thromboembolic events (P=0.03) compared with those with lower levels.11 One could speculate anti-PLA2R, or antibodies to other membranous antigens such thrombospondin type-1 domain-containing 7A, neural epidermal growth factor-like 1 protein, or Exostosin 1/Exostosin 2 disrupt vascular quiescence, leading to a prothrombogenic state via intermediary pathways. Hopefully, future research will help further clarify the pathophysiology of the hypercoagulable state in nephrotic syndrome to increase the accuracy of predicting thromboembolic risk of prophylactic anticoagulation.