D-dimer levels are markedly raised in HIV-related thrombotic thrombocytopenic purpura

Abstract

The association between HIV infection and thrombotic thrombocytopenic purpura (TTP) has been apparent since the late 1980s [1]. The incidence of TTP in HIV-infected individuals has been estimated to be 15–40 times that in the general population [2]. In South African hospitals it is now by far the commonest cause of TTP. Successful treatment depends on the early institution of plasma infusion or plasmapheresis, and it is often necessary to make the diagnosis in the presence of microangiopathic haemolytic anaemia and thrombocytopenia alone before the full diagnostic pentad of features (including renal failure, neurological symptoms and fever) has emerged [3]. This requires te exclusion of other possible causes for the clinical picture, including disseminated intravascular coagulopathy (DIC). The absence of abnormalities on a DIC screen is therefore quoted as an important diagnostic criterion. However, although other coagulation parameters appear normal, we have frequently observed extremely high D-dimer levels in patients with HIV-related TTP. We therefore analysed the results of coagulation screening in patients with HIV-related TTP and compared these with the findings in HIV-negative patients with TTP. The HIV status and results of DIC screens performed on consecutive patients diagnosed with TTP at the Johannesburg hospital and associated teaching hospitals between August 2002 and January 2005 were reviewed. Patients were identified from records of haematology consults received during this period. The diagnosis of TTP required at least the presence of microangiopathic haemolytic anaemia and thrombocytopenia in the absence of other possible causes. Laboratory data were retrieved from the hospital computer system. The HIV status was confirmed on enzyme-linked immunosorbent assay testing. Twenty-two patients with HIV-related TTP and three patients with idiopathic TTP were identified. Two patients with pregnancy-related TTP who suffered spontaneous abortions before assessment were excluded, as alterations in coagulation parameters in these patients were not solely dependent on the TTP process. The median haemoglobin and platelet levels at diagnosis were 5.2 g/dl and 11 × 109/l, respectively, for patients with HIV-related TTP and 8.0 g/dl and 19 × 109/l, respectively, for patients with idiopathic TTP. The median international normalized ratio, partial thromboplastin time and antithrombin levels were normal and were not significantly different between the two groups (international normalized ratio 1.13 versus 1.12, partial thromboplastin time 31.5 versus 31.5 s, antithrombin 103 versus 108 IU). However, the median D-dimer level was significantly higher in the HIV-related TTP group (3.73 mg/l, range 0.95–6.95, versus 0.28 mg/l, range 0.22–0.3, P = 0.0058; Fig. 1).Fig. 1: Median D dimer levels in HIV-negative ( n = 3) compared with HIV-positive ( n = 22) patients with thrombotic thrombocytopenic purpura.This study confirmed our clinical impression that high D-dimer levels, in the absence of any other abnormalities in coagulation parameters, appear to be a consistent feature of HIV-related TTP. In our experience this finding is often helpful in suggesting the diagnosis. The D-dimer levels appear to be significantly higher than those in patients with idiopathic TTP although the small numbers of HIV-negative patients in this study limits the reliability of the comparison. It is possible that the very high D-dimer levels noted in patients with HIV-related TTP reflect differences in the pathogenesis. It is becoming increasingly clear that a deficiency of ADAMTS13, a von Willebrand factor (VWF)-cleaving protease, although contributory is not sufficient in itself to precipitate an episode of TTP [4]. Endothelial cell injury is likely to play a key role in initiating and sustaining HIV-related TTP. Direct infection of endothelial cells by the virus itself has been documented [5], and the release of cytokines such as TNF-α and IL-1β or proteins secreted by the HIV-infected cells has been shown to increase monocyte–endothelial cell interactions and cause the detachment of endothelial cells [6]. VWF release is taken as a marker of endothelial cell injury, and VWF levels have been shown to be markedly elevated in HIV infection, with higher levels correlating significantly with disease progression and increased viral load [7]. Florid VWF release as well as loss of the natural antithrombotic properties of the endothelium as a result of endothelial cell injury could certainly cause platelet deposition and microvascular thrombosis. Activation of coagulation pathways and subsequent fibrinolysis as a result of this process would give rise to the raised D-dimer levels noted. If extensive enough, the platelet deposition and microvascular thrombosis may be sufficient in isolation to result in the clinical picture of TTP. Concomitant ADAMTS13 deficiency (acquired either through the presence of autoantibodies, as is typically described [8], or as a result of the consumption of the enzyme through the overwhelming release of VWF, as has been demonstrated in association with the use of 1-deamino-8-D-arginine vasopressin; desmopressin [9]) would result in the accumulation of ultralarge VWF multimers, which may be important in sustaining the process.