Venous thromboembolism associated with the management of acute thrombotic thrombocytopenic purpura


Dr Helen Yarranton, Haemostasis Research Unit, Department of Haematology, University College London, 3rd Floor, Jules Thorn Building, 48 Riding House Street, London, W1N 8AA, UK. E-mail:


Summary. Venous thromboembolism (VTE) is not a feature of thrombotic thrombocytopenic purpura (TTP), but there has been a recent report of VTE in association with plasma exchange (PEX) treatment for TTP using the solvent detergent (SD) plasma, PLAS+®SD. We reviewed the occurrence of VTE in 68 consecutive patients with TTP (25 men, 43 women). Eight documented VTE events [six deep venous thromboses (DVTs), three pulmonary emboli] were identified in seven patients (all female) during PEX therapy. All six DVTs were associated with central lines at the site of thrombosis. Other known precipitating factors included pregnancy, immobility, obesity and factor V Leiden heterozygosity. VTE occurred at a mean of 53 d following the first PEX. The European SD plasma, Octaplas® was the last plasma to be used in PEX prior to the VTE in 7/8 events. This is the first report of VTE following Octaplas® infusion. VTE is a multifactorial disease and, although several known precipitating factors were present in all patients in this study, the use of large volumes of SD plasma in PEX may be an additional risk factor. We recommend prevention of VTE with graduated elastic compression stockings (class I) at diagnosis and prophylactic low-molecular-weight heparin once the platelet count rises above 50 × 109/l.

Thrombotic thrombocytopenic purpura (TTP) is a life threatening disorder characterized by thrombi in the microvasculature, resulting in blood vessel occlusion that gives rise to tissue ischaemia and end organ damage. The clinical features of TTP are diverse and may manifest as fluctuating sensorimotor signs, renal impairment, cardiac ischaemia or arrythmias and abdominal pain, depending on the organs affected. The diagnosis of TTP is initially suspected clinically upon the presence of microangiopathic haemolytic anaemia and thrombocytopenia in the absence of any other identifiable cause (Thompson et al, 1992).

Urgent plasma exchange (PEX) is the only proven effective therapy for TTP, and its introduction in the 1970s was associated with a dramatic improvement in the survival rate, from around 10% up to 70–90% (Rock et al, 1991). The conventional replacement fluid used in PEX is fresh-frozen plasma (FFP). Cryosupernatant (CSP) is at least as effective as FFP (Rock et al, 1996; Zeigler et al, 2001). CSP may have theoretical advantages in the management of TTP because of the lack of the largest von Willebrand factor (VWF) multimers that have been implicated in the pathogenesis of TTP (Moake, 1995). TTP patients treated with PEX are exposed to a significant risk of viral transmission because of the large volumes of plasma routinely required in their treatment. In England and Wales, the current estimated risks of viral transmission from a single blood donor component are 1·9 per million donations for human immunodeficiency virus (HIV) 1 + 2, 0·6 per million donations for hepatitis C and 5–20 per million donations for hepatitis B (Pamphilon, 2000). Recently, pathogen inactivation procedures for plasma, such as solvent detergent (SD), methylene blue/light (MB) and psoralen S-59 treatment processes, have been developed to minimize the risk of transfusion-transmitted infections. SD-treated pooled plasma (SD-plasma) has been used successfully in the treatment of acute TTP and in patients refractory to standard exchange therapy (Harrison et al, 1996; Evans et al, 1999). SD-plasma, like CSP, is deficient in the largest VWF multimers and may be of benefit in PEX for TTP. Although MB-treated plasma (MB-plasma) has the advantage of being prepared from single donors, it may be less efficacious in the setting of TTP (de la Rubria et al, 2001). Psoralen S59-treated plasma is being assessed in an ongoing efficacy trial in North America (L. Corash, Cerus Corporation, personal communication).

The classical thrombotic lesions in TTP occur in the arterioles and capillaries, and are mainly composed of platelets and VWF (Asada et al, 1985). Converse findings are found in venous thrombi, which have plentiful fibrinogen and a relative lack of platelets and VWF. Venous thromboembolism (VTE) is not a feature of TTP or its treatment with PEX, but there are a few recent publications documenting the occurrence of small numbers of VTE during acute TTP (Rizvi et al, 2000), including hepatic vein thrombosis (Hsu et al, 1995) and central retinal vein and artery occlusion (Schwartz et al, 2000). Flamholz et al (2000) reported the occurrence of deep venous thrombosis (DVT) following PEX therapy for TTP with PLAS+®SD (manufactured by V.I. Technologies, USA), the form of SD-plasma used in North America. These DVTs were associated with low protein S activity levels in the patients at the time of thrombosis. The only SD-plasma licensed in Europe is Octaplas® (Octapharma, Vienna, Austria) and, to our knowledge, there have been no reports of VTE following Octaplas® infusion.

Given the recent concerns of VTE occurring with PLAS+®SD in North America, we have retrospectively reviewed our patients with TTP, with specific reference to the occurrence of VTE and any possible association with specific types of plasma infusions.

Patients and methods

Patients.  Between May 1997 and May 2002, 68 patients with acute TTP were referred for primary management or a second opinion. There were 43 women and 25 men, and the mean age was 41 years (range 19–70 years). From 1999, it was routine practise to measure VWF cleaving protease and inhibitor levels at diagnosis of TTP (Allford et al, 2000). Central lines were inserted in > 95% of TTP patients receiving plasma exchange. During the 5-year study period, seven patients suffered eight documented VTE events during PEX therapy for the treatment of acute TTP. For each of these seven patients, the case history notes were examined retrospectively and blood results were obtained from the hospital laboratory computer. Possible precipitants of VTE were analysed. The patients were all treated initially with daily 1·0 plasma volume exchange (except patient 2 who received 0·5 plasma volume exchange). The components used as the replacement fluid included FFP [National Blood Service (NBS)], CSP (NBS) and SD-plasma (Octaplas®, Octapharma, Austria). Confirmation of all VTE events was with Doppler ultrasound scanning, intravenous lineogram or computerized tomo-graphy for DVT, and ventilation-perfusion scanning for pulmonary emboli (PE). Thrombophilia screens were performed in all seven patients following their VTE and when in remission after their acute TTP. The screen included Factor V Leiden and the G20210A prothrombin gene mutation, functional protein S, protein S free antigen, functional protein C, antithrombin, fibrinogen, lupus anticoagulant (by dilute Russell's viper venom time test), and IgG and IgM anticardiolipin antibodies. All assays were performed by routine standardized techniques.

Measurement of protein S activity and free antigen levels in FFP, CSP and Octaplas®.  Functional protein S activity levels were measured by the STA®-Staclot® protein S clotting assay (Diagnostica Stago, Asnieres-Sur-Seine, France), and free protein S antigen levels were measured by immuno-turbidimetric method using STA®-Liatest® free protein S kit (Diagnostica Stago). Functional and free antigen protein S levels were measured in 10 single donor units of FFP (NBS), nine single donor units of CSP (NBS) and in 10 single production batches of Octaplas®. Normal ranges for protein S activity (male, 0·65–1·45 IU/ml; female, 0·50–1·20 IU/ml) and free protein S antigen (male, 0·70–1·48 IU/ml; female, 0·50–1·34 IU/ml) were obtained from 20 normal adult men and 20 normal adult women. The interassay coefficient of variation (CV) and intra-assay CV for measuring functional protein S activity were 3·5% and 3·2%, respectively [normal quality control (QC)], and 12·5% and 5·1% respectively (pathological QC). The interassay CV and intra-assay CV for measuring free protein S antigen were 2·7% and 2·6%, respectively (normal QC), and 2·8% and 5·4% respectively (pathological QC).

Protein S levels were not routinely measured during PEX prior to the VTE event in the eight patients in this study, but a series of archived plasma samples were available for patient 3 from the initiation of PEX therapy. Functional and free antigen protein S levels were measured in a total of 19 samples taken between d 1 and d 35 of PEX treatment which included 2 weeks using CSP initially followed by a 3-week period using SD-plasma.


Patients and acute TTP episode

Eight proven VTE events were identified retrospectively in seven patients during treatment for an acute episode of TTP. All seven patients fulfilled the criteria for a diagnosis of acute TTP (Rock et al, 1991). The characteristics of the patients at the time of diagnosis of TTP are given in Table I. All seven patients were women and their mean age at the time of their VTE was 31 years (range 21–44 years). Five of the VTE events occurred during the first episode of TTP and three occurred during a relapse of TTP. At the onset of acute TTP, three patients were taking a combined oral contraceptive pill, one was pregnant (15 weeks gestation), one who relapsed had a bacteraemia and in one patient a central venous catheter had been in-situ for a prolonged period (> 150 d). No trigger factors for TTP were identified for the other two acute episodes of TTP. Seven of the eight (88%) TTP episodes were associated with neurological abnormalities and four (50%) had abdominal symptoms (abdominal pain or vomiting). None had significant renal impairment (defined as creatinine > 177 µmol/l).

Table I.  Patient characteristics at the time of diagnosis of the acute episode of TTP associated with the venous thromboembolic event.
FeaturesPatient 1Patient 2Patient 3Patient 4Patient 5Patient 6Patient 7
(event a)(event b)
  1. F, female; TTP, thrombotic thrombocytopenic purpura; OCP, oral contraceptive pill; Hb, haemoglobin; LDH, lactate dehydrogenase; NR, no result; RBC, red blood cell.

Age/sex26 F26 F24 F32 F35 F4021 F44 F
TTP episode1st episoderelapse1st episode1st episoderelapserelapse1st episode1st episode
Possible trigger factorsOCPpregnancyprolonged central venous accessOCPinfectionnone identifiedInfection (OCP)none identified
Hb (g/dl)6·214·357·67·68·35·87·4
Platelet (× 109/l)1627954395816
LDH (U/l) (normal range 240–480)NR40614963000224NR16521625
RBC fragmentationpresentpresentpresentpresentpresentpresentpresentpresent
Neurological abnormalitiespresentabsentpresentpresentpresentpresentpresentpresent
Renal impairmentabsentabsentabsentabsentabsentabsentabsentabsent
Abdominal symptomspresentpresentabsentabsentabsentpresentpresentabsent

Venous thromboemboli and precipitating factors

The details of the eight VTE events are given in Table II. These VTE events included six documented DVTs, three PEs (DVT and PE were present simultaneously at one event in patient 5) and also one pulmonary artery thrombosis. All the DVTs were associated with central venous catheters. The DVTs occurring in the iliac and femoral veins (four in total) were related to catheters in the femoral veins, and the brachiocephalic (n = 1) and subclavian vein (n = 1) thromboses were related to internal jugular and subclavian vein catheters respectively. The one pulmonary artery thromboses was related to a Swan–Ganz catheter in the pulmonary artery. Other acquired precipitating factors of VTE for the eight events included pregnancy (n = 1), immobility (n = 8) and obesity (defined as body mass index > 30) (n = 3). One patient had a family history of VTE; her mother had a DVT associated with pregnancy. Thrombophilia was only detected in one patient who was heterozygous for the factor V Leiden mutation. The mean platelet count at the time of occurrence of VTE was 161 × 109/l (range 12–232 × 109/l). The platelet count had been in the normal range for a mean of 18 d (range 2–77 d) prior to the occurrence of VTE in those with a normal platelet count (6/8) at the time of VTE. In two cases, the platelet count had fallen below 150 × 109/l on the day of the VTE, after previously achieving a normal platelet count.

Table II.  Characteristics of venous thromboemboli and precipitating factors.
FeaturesPatient 1Patient 2Patient 3Patient 4Patient 5Patient 6Patient 7
(event a)(event b)
  1. VTE, venous thromboembolism; PE, pulmonary embolism; DVT, deep venous thrombosis; d, day; PEX, plasma exchange; NA, not applicable; NAD, no abnormality detected.

VTEPEDVTDVTDVTPE and arterial thrombosisPE and DVTDVTDVT
Position of VTEmultiple PEs in both lung fieldsexternal illiacilliacBrachio -cephalicmultiple PEs and pulmonary arterymultiple PEs and common femoral veinsubclavianileofemoral
Central line positioninternal jugular + subclavianfemoralfemoralinternal jugularSwan–Ganz in pulmonary arteryfemoral + subclaviansubclavianfemoral
Day of VTE (d 1 = 1st day of PEX)142116110623252450
Platelet count (× 109/l) at diagnosis of VTE2021223256193225188181
Number of days since platelet count > 150 × 109/l2NA77NA145414
Other acquired precipitating factors for VTEimmobility obesityimmobility pregnancy obesityimmobilityimmobilityimmobilityimmobility obesityimmobilityimmobility
Thrombophilia screenNADNADNADFactor V Leiden heterozygoteNADNADNADNAD

Plasma exchange and adjuvant therapies

VTE occurred at a mean of 53 d (range 14–161 d) following the first therapeutic PEX for the acute episode of TTP (Table III). Octaplas® was the last plasma to be used in PEX prior to the VTE in seven of the eight events. Patient 1 did not receive Octaplas® prior to her VTE event. Her first exposure to Octaplas® was after the diagnosis of her VTE. Octaplas® was used as the sole plasma in PEX for treatment of three of the acute episodes. Two of these had chronic intermittent TTP and during previous TTP episodes had responded to PEX with Octaplas® after failure with FFP and CSP. The remaining six acute episodes of TTP were initially treated with PEX using either FFP or CSP. In five of these cases, SD-plasma was substituted prior to the development of VTE, because of a poor response (failure to achieve a normal platelet count after at least 6 d of PEX with FFP or CSP) in four cases and following a severe allergic reaction in one case. The average number of days of PEX prior to the VTE event was 35 d (range 13–87 d) using Octaplas® as the replacement fluid. The average total volume of component used for PEX to sustain haematological remission (normal haemoglobin and platelet count) in these patients was 21, 20 and 90 L for FFP, CSP and Octaplas® respectively.

Table III.  Plasma exchange and adjuvant therapies for acute TTP.
FeaturesPatient 1Patient 2Patient 3Patient 4Patient 5Patient 6Patient 7
(event a)(event b)
  1. VTE, venous thromboembolism; d, day; PEX, plasma exchange; No., number; SD, solvent/detergent-treated plasma; FFP, fresh-frozen plasma; CSP, cryosupernatant; TTP, thrombotic thrombocytopenic purpura.

Day of VTE (d 1 = 1st day of PEX)142116110623252450
No. of days of SD PEX (days over which SD PEX occurred) 013 (7–17) 87 (15–161) 59 (17–112)11 (7–23)24 (1–24)23 (1–23)30 (11–49)
No. of days of FFP PEX (days over which FFP PEX occurred) 6 (1–6) 0  0 16 (1–16) 3 (1–3) 0 0 0
No. of days of CSP PEX (days over which CSP PEX occurred) 7 (7–13) 0 14 (1–14)  0 3 (4–6) 0 010 (1–10)
Aspirin prescribedyesyesyesyesyesyesnoyes
Heparin or warfarin prophylaxis (days given)noneheparin (1–21)nonewarfarin (97–106)nonenonenonenone
Adjuvant therapies for TTPcorticosteroids vincristine defibrotidecorticosteroids defibrotidecorticosteroids vincristine defibrotide cyclosporinecorticosteroids vincristine cyclosporinecorticosteroids vincristinevincristinevincristine defibrotidecorticosteroids vincristine

One patient (patient 2) received VTE prophylaxis [low-molecular-weight heparin (LMWH)] from the time of TTP diagnosis, and another (patient 4) received warfarin (1 mg) as central line prophylaxis 9 d before her VTE. Low-dose aspirin (75 mg), as prophylaxis against microvascular arterial thrombosis, was administered during seven of the eight episodes prior to the diagnosis of VTE when the platelet count became greater than 50 × 109/l, according to local protocol (Yarranton & Machin, 2002).

Corticosteroids were administered in six of the eight TTP episodes. Other adjuvant therapies were also employed following a poor response to PEX alone. These included vincristine (7/8), defibrotide (4/8) and cyclosporine (1/8).

Protein S levels

The median values for functional protein S and free protein S antigen for the units of FFP and CSP and the batches of Octaplas® tested are given in Table IV. The protein S activity (normal male range 0·65–1·45 IU/ml) was reduced in all 10 units of Octaplas® (range 0·55–0·64 IU/ml), but within the normal range for the other components. Octaplas® showed less variability between individual units, because it is prepared from pooled plasma.

Table IV.  Functional protein S (PS) and free protein S antigen levels in FFP, CSP and Octaplas®.
 Functional PS IU/mlFree PS:Ag IU/ml
  1. Results are given as the medians (25th to 75th percentiles).

Reference values0·65–1·45 (male)0·70–1·48 (male)
0·50–1·20 (female)0·50–1·34 (female)
FFP0·91 (0·81–1·05)0·81 (0·65–1·11)
CSP1·02 (0·98–1·09)1·1 (0·98–1·24)
Octaplas®0·58 (0·56–0·60)0·65 (0·63–0·67)

In patient 3, the average functional protein S level was 0·68 IU/ml (SD 5·54) (normal female range 0·50–1·20 IU/ml) and average free protein S antigen level was 0·78 IU/ml (SD 8·07) (normal female range 0·50–1·38 IU/ml) during PEX using CSP (n = 9), and 0·56 IU/ml (SD 6·20) and 0·70 IU/ml (SD 7·79), respectively, during PEX using Octaplas® (n = 10). These samples were taken immediately before PEX.


We identified eight episodes of VTE occurring in seven patients from a total of 68 consecutive adult patients (i.e. in 12%) treated with PEX for TTP. Seven of the VTE events followed PEX using SD-plasma (Octaplas®) and one followed PEX using CSP, as the last plasma infused. The patients who suffered VTE were relatively young [mean age 31 years (range 21–44 years), c.f. mean of 41 years (range 19–70 years) in the group as a whole], female and 75% had normal platelet counts (mean 161 × 109/l, range 12–232 × 109/l) at the time of occurrence of VTE. Notably for 7/8 VTE episodes, patients were on low-dose aspirin, which is known to provide some thromboprophylaxis against VTE (Rodgers, 2000).

VTEs are not a feature of TTP, but recently there have been a small number of reports of their occurrence. A large prospective study from North America, describing adverse effects associated with PEX in 71 consecutive patients with TTP/haemolytic–uraemic syndrome, identified two venous thromboses requiring anticoagulation (Rizvi et al, 2000). In this American population, the thromboembolic complication rate was 3%. In our study, the thromboembolic rate was 12%. The replacement fluid for PEX in the large prospective American study was either FFP or CSP. SD-plasma was not used. A further report from the USA described three patients with TTP who developed DVTs following PEX using SD-plasma (PLAS+®SD) (Flamholz et al, 2000). In this study, two patients had low levels of functional protein S activity: 24% and 26% (measured by STA®-Staclot® protein S clotting assay; Stago Diagnostica) on the day of DVT diagnosis. The third patient did not have functional protein S measured on the day of DVT diagnosis, but recorded a nadir of 28% following 2 d of PEX using PLAS+®SD. PLAS+®SD is known to have reduced protein S activity as compared with FFP. Functional protein S levels from 2 units of PLAS+®SD were reported as 24·8% and 15% (using STA®-Staclot® protein S clotting assay) (Flamholz et al, 2000).

Occasional deaths due to thromboembolic complications or severe bleeding have also been reported in patients with severe liver disease who received PLAS+®SD ( pdf). To our knowledge, there have been no reports of VTE following Octaplas®, the only SD-plasma licensed in Europe. No thrombotic adverse events were reported in a randomized trial of SD-plasma (Octaplas®) and standard FFP in 49 patients with coagulation defects due to liver disease, who required invasive procedures or liver transplantation (Williamson et al, 1999). The manufacturing process of PLAS+®SD differs from that of Octaplas® in that it includes an ultra/diafiltration step that leads to an increase in concentration of coagulation factors (Horowitz et al, 1992) (PLAS+®SD summary of basis for FDA approval: Consequently PLAS+®SD has a greater imbalance of activators and inhibitors of coagulation, which potentially could lead to thrombotic complications when used in large volumes or in susceptible patients.

Seven of the eight VTE events in this study followed PEX using Octaplas® SD-plasma. This is the first report of VTE occurring following Octaplas® infusion. Octaplas®, like PLAS+®SD, is known to have reduced protein S activity (median 0·45 IU/ml, range 0·42–0·53) (Leebeek et al, 1999). These levels are confirmed by the findings of the functional protein S levels in 10 batches of Octoplas® tested in this study: median 0·58 IU/ml (range 0·55–0·63). Protein S activity levels were not consistently measured in our patients during their PEX using Octaplas®. Archived samples were available in patient 3, in whom there was a trend to lower functional protein S levels during PEX using Octaplas® (mean 0·56 IU/ml) compared with CSP (mean 0·68 IU/ml) as the replacement fluid. A study in patients following open heart surgery found that functional protein S level (STA®-Staclot®; Stago Diagnostica) did not increase after SD-plasma infusion, but showed a significant elevation after infusion with FFP (Haubelt et al, 2002). The median functional protein S level prior to infusion was 0·44 IU/ml in the SD-plasma group and 0·31 IU/ml for the FFP group, and 60 min post infusion the levels were 0·42 IU/ml and 0·38 IU/ml respectively (P < 0·0001). Theoretically, the infusion of Octaplas®, which has lower levels of protein S, may contribute to an increased risk of venous thromboembolism when transfused in large volumes, such as repetitive PEX therapy for TTP patients.

VTE is a multifactorial disease and several known precipitating factors for VTE were undoubtedly present in all the cases included in this study. Venous thrombosis is a well-recognized complication of central venous catheter insertion, and an increased rate is seen in those who possess the FV Leiden mutation (Fijnheer et al, 2002). Other known risk factors identified in the patients in this study were pregnancy, obesity, the FV Leiden mutation and family history. In addition, all cases had complicated TTP that responded slowly to PEX alone and required adjuvant therapies. Furthermore, these patients usually had recovering platelet counts and under these conditions platelets are often more susceptible to activation. These are possible further risk factors. Limitations of this study are that it was retrospective and the number of events were insufficient to assess whether Octaplas® was a risk factor for VTE when used in PEX in this group of patients. It should not be forgotten that Octaplas® is potentially a very useful plasma in TTP patients. Octaplas®, like CSP, lacks the largest VWF multimers, which may be of theoretical benefit in TTP, as the presence of the highest-molecular-weight multimers are thought to be implicated in the pathogenesis of TTP (Moake et al, 1994). A reduction in the number of allergic reactions is another potential advantage of SD-plasma, given that the plasma pooling process results in extreme dilution of those antibodies responsible for an immune reaction. One patient in this study received Octaplas® for this reason. Patients with TTP have responded to Octaplas® PEX after previous failure with FFP and CSP (Harrison et al, 1996). It would, therefore, be inadvisable to prohibit the use of Octaplas® in PEX for TTP treatment. The optimal plasma would be pathogen inactivated, efficacious, retaining all haemostatic activity and without other adverse effects.

In conclusion, patients receiving PEX therapy for acute TTP are at risk of developing VTE. Central venous catheter insertion and immobility are known risk factors for VTE that are present in most patients with acute TTP. The use of large volumes of SD-plasma in therapeutic PEX may be an additional risk factor. The significant morbidity and mortality associated with VTE makes their prevention very important. This is especially relevant in patients with TTP because the requirement of therapeutic anticoagulation is likely to be complicated by low platelet counts. Following the recent reports of VTE in TTP patients (Flamholz et al, 2000), and the events reported here, we recommend that VTE prophylaxis should be used in all acute TTP patients. Fitted graduated elastic compression stockings (class I) should be worn from the time of hospital admission, and prophylactic doses of low-molecular-weight heparin, in addition to low-dose aspirin, should be administered when the platelet count rises above 50 × 109/l. Central venous catheters, usually necessary for PEX, should be removed as soon as possible when no longer required. Physicians need to be vigilant in monitoring for signs of VTE and be aware that SD-plasma, when transfused in large volumes, may be associated with an increased risk of developing VTE.


We thank Mike Murphy for critical review of the manuscript, and Andrew Lawrie and Gordon Purdy for performing the protein S assays. The Department of Haematology at University College London Hospitals NHS Trust has previously received an unrestricted educational grant from Octapharma Limited.