Abnormally short activated partial thromboplastin time values are associated with increased risk of recurrence of venous thromboembolism after oral anticoagulation withdrawal


Dr Cristina Legnani, Department of Angiology and Blood Coagulation ‘Marino Golinelli’, University Hospital S. Orsola-Malpighi, Via Albertoni, 15, 40138 Bologna, Italy. E-mail: legnani@aosp.bo.it


This study prospectively evaluated the relationship between activated partial thromboplastin time (aPTT) and risk of venous thromboembolism (VTE) recurrence after oral anticoagulant (OA) withdrawal in patients with a previous unprovoked VTE event. Six hundred twenty-eight patients (331 males; median age: 67 years) were followed after OA interruption (mean follow-up = 22 months). Three to four weeks from OA discontinuation patients were given a complete thrombophilic work-out, including aPTT (automated aPTT). Recurrent symptomatic VTE events (objectively documented) occurred in 71/628 (11·3%, 6·8/100 person-years) patients. The VTE recurrence rate was 17·5% and 7·5% in patients with aPTT in the lower (ratio ≤0·90) and in the upper (ratio >1·05) quartiles. The recurrence risk was more than twofold higher in patients with ratio ≤0·90 versus those of the reference category [Relative risk (RR): 2·38 (95% confidence interval (CI): 1·18–4·78)]. As expected, the increase in recurrence risk disappeared after adjustment for factor VIII, IX and XI levels [RR: 1·74 (95%CI: 0·43–2·76)]. In contrast, the risk was persistently increased in patients with a ratio ≤0·90 [RR: 2·07 (95%CI: 1·02–4·18)] after adjustment for age, gender and d-dimer level. The aPTT predictive value was independent of the presence of inherited thrombophilic alterations. In conclusion, abnormally short aPTT values are associated with a significantly increased risk of VTE recurrence.


There is strong evidence that high plasma levels of some coagulation factors, such as factor VIII (FVIII), IX (FIX) and XI (FXI), are independent risk factors for a first venous thromboembolism (VTE) event (Koster et al, 1995; O'Donnell et al, 1997; Kraaijenhagen et al, 2000; Meijers et al, 2000; Vlieg et al, 2000). More recently, it has been shown that patients with high levels of FVIII (Kyrle et al, 2000; Legnani et al, 2004) or FIX (Weltermann et al, 2003) are also at higher risk of VTE recurrence after a first unprovoked VTE episode.

The activated partial thromboplastin time (aPTT) is a global clotting test that explores the classical intrinsic pathway of blood coagulation and it is sensitive to the levels of FVIII, FIX and FXI, and other clotting factors (i.e. factor II, fibrinogen). The aPTT is a simple, inexpensive test that is widely used for screening in clinical laboratories. It was recently reported that a shortened aPTT was associated with an increased risk of a first episode of VTE (Tripodi et al, 2004). The risk of VTE, though reduced, was increased also after adjustment for FVIII levels, indicating that the risk was only partially mediated by the high FVIII levels. However, whether an abnormally short aPTT is a risk factor for VTE recurrence among patients with a previous unprovoked VTE event is still unknown.

The aim of this multicentre study was to prospectively evaluate the relevance of the aPTT, measured after the discontinuation of oral anticoagulant (OA) therapy, on VTE recurrence in patients with a first unprovoked VTE event.

Patients and methods

Patients and study design

Patients who had received OAs for at least 3 months after a first objectively documented unprovoked episode of deep vein thrombosis (DVT) of the lower limbs and/or pulmonary embolism (PE) were eligible for the study. The VTE events were classified as unprovoked if not associated with recent surgery, trauma, immobilisation, fracture, pregnancy/puerperium (all within 3 months before the VTE episode) or active cancer. Patients with liver or renal failure, antithrombin deficiency, presence of lupus anticoagulant (LAC), other indications for OA, limited life expectation or those who lived too far from the study centre were excluded. Informed consent was obtained from all enrolled patients.

Patients were enrolled in 13 Italian anticoagulation clinics between February 1995 and December 2004 and were prospectively investigated after OA therapy was stopped. The day of OA discontinuation was the day of enrolment and venous blood was sampled to perform thrombophilic tests 3–4 weeks later. On the day of study entry, patients were instructed to wear 40 mmHg elastic stockings for no <2 years and to immediately refer to the treating clinical centre in case of the onset of symptoms attributable to a new or recurrent VTE episode. Physical examinations were scheduled 3 months after OA withdrawal and every 6 months thereafter. Follow-up ended on the 24th month from study entry or earlier in case of death, VTE recurrence, or OA resumption for any other indication. The patient and/or the family doctor were contacted if a periodic examination was missed.

Blood sampling and thrombophilia work-out

Blood was collected from the antecubital vein into 0·129 mmol/l trisodium citrate; plasma was prepared by centrifugation for 20 min at 2000 g at room temperature; plasma and blood for DNA extraction aliquots were snap-frozen and stored at −70°C.

The screening for thrombophilic alterations was performed locally and included the following tests: antithrombin, protein C and protein S level measurement, DNA analysis for R506Q Factor V Leiden mutation and G20210A prothrombin mutation, LAC.

All other tests were performed centrally in the laboratory of the Co-ordinating Centre by technicians unaware of the clinical characteristics of the patients. The aPTT was determined using the Automated aPTT reagent (bioMerieux, Rome, Italy) on a STA Compact (Roche Diagnostics, Monza, Italy); a locally prepared pooled normal plasma was used to calculate the ratios. d-dimer levels were measured by the VIDAS assay (bioMerieux). FVIII levels were measured by a chromogenic method using a commercial assay (Coamate factor VIII; Chromogenix by Instrumentation Laboratory, Milan, Italy) as described elsewhere (Legnani et al, 2004). Clotting FIX and FXI levels were measured by a one-stage clotting method, using factor-deficient plasmas (Dade Behring, Marburg, Germany) and the Automated aPTT (bioMerieux) on a STA Compact (Roche Diagnostics). Calibration curves were performed using locally prepared pooled normal plasma; the results were expressed as IU/ml.


Objectively documented DVT recurrence and/or fatal or non-fatal PE (first event or recurrence) were considered as outcomes. Patients with DVT symptoms in the ipsilateral or contralateral leg as the first event were tested for d-dimer levels (for its high-negative predictive value) and compression ultrasonography (CUS) to be compared with the CUS examination customarily performed in all patients at the time of OA interruption. In cases with suspected DVT recurrence in the ipsilateral lower limb, unequivocal non-compressibility of a previously compressible venous segment or an increase of at least 4 mm in the residual diameter were the criteria considered as diagnostic indication for DVT recurrence. If the CUS result was non-diagnostic (and the d-dimer was altered), an echocolordoppler or contrast venography was performed. In patients presenting with symptoms compatible with PE (either first event or recurrence), diagnosis was based on the results of objective algorithms using clinical probability, ventilation-perfusion or helical computer tomography lung scan, CUS if indicated and d-dimer test. All VTE events were evaluated by an adjudication board of physicians who were unaware of the laboratory results. All deaths were recorded and coded as attributable to a venous event or any other cause.

Statistical analysis

Continuous variables are presented as median (range); the Mann–Whitney U-test was used for group comparison. Differences between groups were assessed by the chi-squared test (Yates’ correction). Times of recurrence (uncensored observations) or follow-up times in patients without recurrence (censored observations) were analysed according to survival-time methods (Kalbfleish & Prentice, 1980). All variables with a univariate level of significance of <0·2 were included in the multivariate analysis. The aPTT values, FVIII, FIX and FXI levels were analysed in a Cox proportional-hazard model as a categorised variable to compare the relative risks (RRs) associated with different values; 95% CI were calculated with an approximate method and a two-sided P-value ≤0·05 was considered as statistically significant. The Statistical Package for the Social Sciences (spss) software, version 11·0 (SPSS, Chicago, IL, USA) was used for data processing.


The demographic and clinical characteristics of the 628 patients included in the study are presented in Table I. The total follow-up period was 1041·0 years. The median observational period for the complete cohort was 22·4 months. Twenty patients were withdrawn from the study because of a new requirement for antithrombotic treatment (nine cases of superficial thrombophlebitis, five cases of atrial fibrillation, four cases of peripheral arterial disease and two cases of stroke) and 11 cases were withdrawn because they died; the cause of death was attributed to VTE in two cases and to reasons other than VTE recurrence in the remaining patients. These patients were withdrawn from the study and were censored at the time of withdrawal. One patient was censored when he moved to a different town and five patients were lost to follow-up after the first 3-month visit.

Table I.   Baseline characteristics of the 628 enrolled patients.
  1. VTE, venous thromboembolism; DVT, deep vein thrombosis; PE, pulmonary embolism; OA, oral anticoagulant.

Sex (M/F)331/297
Age, years [median (range)]67 (20–89)
 ≥65 years, n (%)354 (56·4)
Site of the first VTE
 Proximal DVT, n (%)448 (71·3)
 Proximal DVT + PE, n (%)109 (17·4)
 Isolated PE, n (%)71 (11·3)
Thrombophilic alteration
 Protein C deficiency, n (%)9 (1·4)
 Protein S deficiency, n (%)6 (1·0)
 R506Q Factor V Leiden mutation, n (%)76 (12·1)
 G20210A prothrombin mutation, n (%)49 (7·8)
 Combined alterations, n (%)12 (1·9)
Duration of OA therapy, months [median (range)]7 (3–119)
 ≤6 months, n (%)236 (37·6)
Total duration of follow-up (years)1041
Median follow-up (months)22·4

Of the 628 patients in the study population, 71 had recurrent VTE (11·3%; 95%CI: 8·9–14·0) [DVT in 57 (30 contralateral) and isolated PE in 14 cases; 43 males]. The incidence rate was 6·8 per 100 person-years (95%CI: 5·3–8·5). The VTE recurrence was unprovoked in 61 cases, while trauma/immobilisation was present in three cases, surgery in two cases, active tubercolosis in one case and deep venous extension of superficial phlebitis in four cases. The median time between OA interruption and recurrence was 7·4 months (95%CI: 0·1–19·6). There was no significant difference in the duration of OA treatment between patients with [7·1 months (95%CI: 3·0–33·0)] and without recurrence [7·2 months (95%CI: 3·0–119·0)].

Table II shows the RR of VTE recurrence after stratification of the aPTT values into quartiles. The rate of recurrence was significantly higher in subjects whose aPTT values were below the 25th percentile (ratio ≤0·90) than in those of the reference category (17·5% vs. 7·5%, P = 0·016). Univariate analysis showed that the risk of recurrence was more than twofold higher in patients with an aPTT ratio below 0·91 [RR: 2·38 (95%CI: 1·18–4·78)]. Kaplan–Meier analysis identified a clear divergence between the rate of VTE recurrence in patients with an aPTT value above or below the 25th percentile (Fig 1). An aPTT ratio ≤0·90 conferred a RR of recurrence of about two times [hazard ratio: 1·93 (95% CI: 1·24–3·62), P = 0·006]. At the end of the follow-up period, the cumulative incidence of recurrence was 20·0% in patients with an aPTT below the 25th percentile and 11·0% in those whose aPTT was above the 25th percentile.

Table II.   Relative risk of venous thromboembolism (VTE) recurrence according to the activated partial thromboplastin time (aPTT) values.
aPTT ratioPercentileVTE recurrence (%)Univariate relative risk (95% CI)Multivariate relative risk (95% CI)*Multivariate relative risk (95%CI)†
  1. *Multivariate relative risks were calculated with adjustment for factor VIII, IX and XI (in quartiles).

  2. †Multivariate relative risks were calculated with adjustment for age, sex and altered/normal d-dimer levels.

≤0·90Below 25th28/160 (17·5)2·38 (1·18–4·78)1·74 (0·43–2·76)2·07 (1·02–4·18)
0·91–0·9725th–50th14/169 (8·3)1·09 (0·50–2·41)0·89 (0·33–2·38)1·10 (0·50–2·44)
0·98–1·0551st–75th18/152 (11·8)1·63 (0·77–3·46)1·08 (0·65–4·65)1·21 (0·55–2·64)
>1·05Above 75th11/147 (7·5)1 (reference)1 (reference)1 (reference)
Figure 1.

 Kaplan–Meier estimates of the risk of venous thromboembolism (VTE) recurrence according to the activated partial thromboplastin time (aPTT) values.

Univariate analysis (see Table III) showed that age ≥65 years, altered d-dimer (>500 ng/ml), FVIII, FIX and FXI levels in the upper quartiles (>2·33 IU/ml, >1·51 IU/ml and >1·40 IU/ml respectively) were associated with a significantly increased risk of VTE recurrence [RR: 2·45 (95%CI: 1·44–4·19); 2·65 (95%CI: 1·57–4·46); 3·01 (95%CI: 1·47–6·19); 3·06 (95%CI: 1·29–7·28); 2·14 (95%CI: 1·01–4·58) respectively]. The risk was slightly higher in males [RR: 1·42 (95%CI: 0·88–2·28)]. In contrast, the risk of VTE recurrence was not increased by the duration of OA therapy, the presence of thrombophilic alterations other than increased levels of FVIII, FIX or FXI, and was not related to the type of index event (with or without PE). All variables with a univariate level of significance of <0·2 were included in the multivariate analysis, which analysed the risk of VTE recurrence associated with aPTT values.

Table III.   Relative risk of venous thromboembolism recurrence according to different patient characteristics.
 Univariate relative risk (95% CI) P-value
  1. *Thrombophilic alterations = protein C or protein S deficiencies; R506Q Factor V Leiden and G20210A prothrombin mutations.

  2. †Reference categories = FVIII, FIX and FXI levels below the 25th percentile.

  3. PE, pulmonary embolism; DVT; deep vein thrombosis.

Age ≥65 years2·45 (1·44–4·19)0·001
Gender = male1·42 (0·88–2·28)0·150
Duration of OA therapy ≤6 months0·94 (0·58–1·52)0·941
Presence of thrombophilic alterations*0·87 (0·50–1·52)0·872
Index event = isolated PE or DVT + EP 0·95 (0·56–1·60)0·838
Altered d-dimer levels (>500 ng/ml)2·65 (1·57–4·46)0·000
Factor VIII levels above the 75th percentile (>2·33 IU/ml)†3·01 (1·47–6·19)0·003
Factor IX levels above the 75th percentile (>1·51 IU/ml)†3·06 (1·29–7·28)0·011
Factor XI levels above the 75th percentile (>1·40 IU/ml)†2·14 (1·01–4·58)0·049

The levels of FVIII, FIX and FXI were significantly higher in patients with an aPTT ratio ≤0·90 compared with those of patients with a longer aPTT [median and (range); FVIII: 2·14 (0·87–5·41) vs. 1·66 (0·50–4·64) IU/ml, P < 0·0001; FIX: 1·50 (1·02–3·08) vs. 1·25 (0·67–2·01) IU/ml, P < 0·0001; FXI: 1·41 (0·72–2·37) vs. 1·14 (0·41–2·28) IU/ml, P < 0·0001]. As expected (see Table II), the increase in the risk of VTE recurrence disappeared after adjustment for FVIII, FIX and FXI levels [RR: 1·74 (95%CI: 0·43–2·76)]. In contrast, after adjustment for age, sex and d-dimer levels the risk of VTE recurrence was significantly increased in patients with an aPTT value below the 25th percentile [RR: 2·07 (95%CI: 1·02–4·18)].


The aPTT is a standard screening test used in coagulation laboratories throughout the world, from the most specialised laboratories to the smallest clinics. This test explores the intrinsic pathway of the coagulation cascade and is also sensitive to unfractionated heparin. As the aPTT is abnormally prolonged in patients with inherited or acquired coagulopathies that involve defects within the intrinsic pathway, such as haemophilia and disseminated intravascular coagulation, it is extensively used for the detection of clotting factor defects in patients with abnormal bleeding (White, 2003).

More than 30 years ago it was suggested that an abnormally short aPTT value might be a predictor of thrombosis (Gallus et al, 1973; Pilgeram, 1974; McKenna et al, 1977) but the use of this test has continued to focus on the relevance of prolonged aPTT clotting times and little attention has been paid to abnormally short aPTT values. It has previously been reported that increased levels of biochemical markers of thrombin generation and fibrin deposition are associated with shortened aPTT values (Korte et al, 2000; Ten Boekel & Bartels, 2002). Abnormally short aPTT clotting times were demonstrated to predict adverse outcome in terms of general mortality in a general hospital setting (Reddy et al, 1999) and, more recently, have been associated with an increased risk of a first VTE event (Tripodi et al, 2004). Data on the relationship between an abnormally short aPTT and the risk of recurrence after a first VTE event are scarce. It has been previously demonstrated that patients with high levels of FVIII and FIX are at higher risk for VTE recurrence after OA discontinuation (Kyrle et al, 2000; Weltermann et al, 2003; Legnani et al, 2004). As the aPTT is sensitive to the clotting factor levels of the intrinsic pathway, it can be expected that high FVIII, FIX or FXI levels will be associated with shorter aPTT values.

This multicentre prospective study showed that abnormally short aPTT values are a risk factor of VTE recurrence in patients with a first unprovoked VTE episode. Patients with an aPTT value in the lower quartile (ratio ≤0·90) at OA discontinuation have a twofold increased risk of VTE recurrence compared with those with higher aPTT values. The risk of recurrence associated with shortened aPTT values was significantly increased at univariate analysis and also after adjustment for all those variables with a univariate level of significance <0·2 at the univariate analysis (age, sex and altered d-dimer). It should be noted that abnormally short aPTT values are associated with an increased VTE recurrence risk, independently of the presence of inherited thrombophilic alterations (protein C or proteins S deficiencies; R506Q factor V Leiden and G20210A prothrombin mutations). The increase in VTE recurrence risk has been shown to be most probably due to the higher FVIII, FIX and FXI found in patients with shorter aPTT, as the risk was no more significantly increased after adjustment for these clotting factor levels. The aPTT might be a convenient alternative option for the measurement of all procoagulant clotting factors. Indeed, while FVIII, FIX and FXI measurements are complex, costly tests and are performed only in specialised laboratories, the aPTT is a very simple, easy and inexpensive test that is carried out in all coagulation laboratories.

Some potential limitations of our study should be taken into account. We have evaluated only one commercial aPTT reagent and it might be that different results would be obtained using different aPTT systems. There are a wide variety of aPTT test systems; they differ in chemical composition of reagents, activator and the method of endpoint detection. As the aPTT test detects a wide range of procoagulant and anticoagulant effects it has been already reported that different aPTT systems have a variable responsiveness to low clotting factor levels (Marlar et al, 1984; Turi & Peerschke, 1986). Data are lacking but it could be speculated that the responsiveness of the aPTT to high-clotting factor levels would suffer from the same problems. Furthermore, it should be considered that clotting activation during/after blood sampling may lead to abnormally short aPTT values. Unfortunately, repeated blood sampling and testing, to reduce the risk of artefacts, was not possible in the present study.

In conclusion, our study shows that abnormally short aPTT values are associated with a significantly increased risk of VTE recurrence after OA withdrawal in patients with a first unprovoked VTE episode. As the role of shortened aPTT in predicting VTE, as well as VTE recurrence, is certainly influenced by the setting in which the aPTT test is performed, confirmation of our results with other aPTT systems is warranted.


Study sites and principal investigators (all the participating centres are affiliated to the Italian Federation of Anticoagulation Clinics – FCSA; the number of enrolled patients are shown in parenthesis):

U.O. Angiologia e Malattie della Coagulazione ‘Marino Golinelli’, Policlinico S.Orsola-Malpighi, Bologna – G. Palareti (438, Co-ordinating Centre); I ° Div. di Medicina Interna - Centro Emostasi e Trombosi, Arcispedale Santa Maria Nuova, Reggio-Emilia – A. Ghirarduzzi (51); A.O. Istituti Ospitalieri, Cremona – S. Testa (44); Centro Regionale Malattie Emorragiche e Trombotiche - Div. Ematologia, Ospedale S. Bortolo, Vicenza – A. Tosetto (18); Centro Sorveglianza Anticoagulati - Malattie della Coagulazione e Angiologia Medica - Div. di Med. Interna, Ospedale “S. Cuore di Gesu’”, Gallipoli – L. Ria (12); Sezione Trasfusionale, Ospedale San Leopoldo Mandic, Merate – N. Erba (11); Centro Emostasi, Ospedale Regionale, Parma – C. Pattacini (10); Servizio di Prevenzione e Terapia della Trombosi, Ospedale ‘Ex Busonera’, Padova – V. Pengo (10); Laboratorio Analisi - Ambulatorio per il Controllo della Terapia Anticoagulante Orale, Presidio Ospedaliero di Faenza – E. Bucherini (9); Laboratorio Analisi - Divisione di Cardiologia, Ospedale di Bentivoglio – E. Cerè (9); Laboratorio di Patologia Clinica, Presidio Ospedaliero S.Maria Incoronata dell'Olmo, Cava dei Tirreni – C. Villani (8); Centro Trombosi, A.O. di Careggi Universita’ di Firenze – D. Prisco (4); U.O. Medicina Interna, Policlinico Universitario Messina – A. Trifiletti (4).