Sigurd Liestøl, Department of Haematology, Ullevål University Hospital, N-0407 Oslo, Norway. Tel.: +47 22 11 91 01/+47 22 11 92 40; fax: +47 22 11 91 81; e-mail: firstname.lastname@example.org
Summary. Background: Several studies suggest that antiphospholipid antibodies interfere with the activity of activated protein C (APC). This acquired form of APC resistance has been proposed as a possible pathogenic mechanism underlying hypercoagulability associated with the antiphospholipid syndrome (APS).Objectives: We wanted to investigate the inhibitory effect of recombinant APC (rAPC) on ex vivo thrombin generation in plasma and the modification of this effect by the presence of lupus anticoagulants (LA).Patients/Methods: We analyzed plasmas from 81 patients with LA (52 patients fulfilling the criteria for the APS) and 91 controls. Percent inhibition of the endogenous thrombin potential (ETP) as a parameter of APC sensitivity was determined in plasmas using a thrombin generation-based APC resistance test probed with rAPC. All results were normalized using pooled normal plasma (PNP) as a reference.Results: Normalized percent inhibition of ETP by APC was lower in patients with LA [61.4%, 95% confidence interval (CI) 45.8–74.5%] compared to controls (107.8%, 95% CI: 107.1–109.3%). In patients with LA and APS, median inhibition was lower than in patients with LA without APS (44.6%, 95% CI: 30.1–55.7% vs. 78.8%, 95% CI: 73.9–95.8%). This difference also persisted when patients on warfarin therapy were excluded from the APS subgroup.Conclusions: APC resistance can be demonstrated with a thrombin generation-based test in a majority of patients with the LA laboratory phenotype. A history of thrombotic events in patients with LA is associated with a stronger resistance to the anticoagulant effect of APC.
An activated protein C (APC) resistant phenotype in the absence of the factor (F) V Leiden genotype is seen in individuals using hormonal replacement therapy or oral contraceptives, and in acquired conditions such as antiphospholipid antibodies (APA). Such APC resistance has consistently been reported to be a risk factor for venous thrombosis [1–5], and phenotypic evaluation of APC resistance has therefore been suggested as an alternative screening strategy .
Several studies suggest that APA may interfere with the anticoagulant activity of APC to induce acquired APC resistance. This effect of APA has been proposed to be a pathogenic mechanism underlying the hypercoagulability associated with the antiphospholipid syndrome (APS) [6–13]. However, circumstantial evidence for an association between an APA-dependent APC-resistant phenotype and thrombotic events is still limited [14,15].
The calibrated automated thrombography (CAT) assay determines the time integral of thrombin generation after triggering with tissue factor, which is called the endogenous thrombin potential (ETP). According to previous reports, the CAT assay allows study of the protein C (PC) system in patients with lupus anticoagulants (LA), as it demonstrates severe resistance to APC even when there is a considerable prolongation of the coagulation lag time [11,12,16]. Although ETP values in most patients with LA have been reported to be within the range seen in healthy controls [11,16], large inter-individual variabilities have also been demonstrated , and ex vivo thrombin generation assays do not usually reveal a hypercoagulable state in these patients. However, a comparison of ex vivo thrombin generation in the presence and absence of APC enables assessment of the function of the PC system even if the in vitro coagulation process is impaired by LA or therapeutic anticoagulants, such as warfarin [16,18].
The aim of this study was to investigate the inhibitory effect of recombinant APC (rAPC) on ex vivo thrombin generation in plasma measured with the CAT assay and the modification of this effect by the presence of LA. We hypothesized that the presence of LA may induce a resistance to the anticoagulant activity of APC in some individuals and that such resistance may be related to a history of thrombosis. If so, the presence of APC resistance might constitute a clinically useful variable in patients with the LA phenotype. As LA may be responsible for both the clinical prothrombotic state and the paradoxical in vitro prolongation of coagulation times, we also wanted to see if the ‘anticoagulant’ phenotype seen with ordinary coagulation tests or prolongation of the CAT-assay lag time (time to thrombin burst) might be correlated with the APC resistant phenotype.
Materials and methods
The patient material comprised 81 patients (56 females and 25 males, median age 38 years, age range 20–63 years) who were classified as LA positive by routine laboratory investigation. Of these patients, 52 (64%) met the criteria for APS  [venous thrombosis only (n =29), arterial thrombosis only (n =8), obstetric complication only (n =8), and combined events (n =7)]. 34 of the patients with LA and APS used warfarin, and their International Normalized Ratio (INR) values were 1.6–4.3 at blood collection. None of the included patients was a carrier of the FV Leiden or prothrombin gene mutations. Levels of antithrombin, PC, and protein S were within reference limits, and none of the included women was using oral contraceptives or hormonal replacement therapy. This study was approved by the institutional review board.
The presence of LA was confirmed following the guidelines proposed by the APA Standardization Subcommittee of the International Society on Thrombosis and Haemostasis  using the diluted activated partial thromboplastin time (dAPTT) and the diluted Russel Viper Venom time (dRVVT) based lupus ratio tests performed on 1:1 mixtures of patient plasma and pooled normal plasma (PNP) . Samples positive in at least one of these two tests were defined as LA positive.
The control material comprised 53 presumably healthy hospital employees (33 women and 20 men, median age 40 years, range 23–61 years) not on any medication, who tested normal in routine thrombophilia screening including FV Leiden and prothrombin gene mutations, the coagulation inhibitors antithrombin, PC, and protein S, and LA and anticardiolipin antibodies.
As control subjects for patients using warfarin, we included 35 individuals on long-term warfarin therapy for non-valvular atrial fibrillation without LA and who tested normal in routine thrombophilia screening. Their INR values were 1.5–4.1 at blood collection.
Informed consent was obtained from all patients and controls. Blood samples were collected in 5 mL siliconated glass tubes (Vacutainer®, Belliver Industrial Estate, Plymouth, UK) containing one-tenth volume 0.129 m buffered sodium citrate. Platelet-poor plasma (PPP) was obtained by centrifugation at 2000 × g for 15 min at 20 °C followed by filtration (Millex-GV 0.22 μm Filter Unit, Millipore, Molsheim, France) and stored in aliquots at −70 °C. The majority of the blood samples from patients were collected by referring laboratories, centrifuged at 2000 × g for 15 min at 20 °C followed by transportation of plasma at ambient temperature within 1–2 days to our laboratory for further filtration and freezing, whereas blood samples from controls were collected and fully processed at our laboratory.
For all patients, blood sampling was performed at least 3 months after a thromboembolic or obstetrical adverse event, as acquired APC resistance due to an acute thrombotic episode has previously been reported .
PNP was prepared by pooling equal volumes of PPP from each 16 of the 53 subjects of the control group, and aliquots were stored at −70 °C. An INR assay (Nycotest PT, Axis-Shield PoC/Medinor ASA, Oslo, Norway) was performed on all plasmas from patients and controls on oral anticoagulant therapy with warfarin.
dAPTT-based lupus ratio test
An integrated approach was applied to LA testing, based on the dAPTT using 1:1 mixture of test and normal plasma with low (screening) and high (confirmatory) phospholipid concentrations . Results were calculated as the ratio between the clotting times (i.e. the clotting time at low phospholipid concentration divided by the clotting time at high phospholipid concentration), and this ratio was normalized against a corresponding ratio obtained with normal plasma in the same run. This normalized ratio was defined as the lupus ratio.
Sensitivity to APC measured by the CAT assay
The CAT assay was performed essentially as described by Hemker et al.  and according to the manual provided by Thrombinoscope B.V. (Maastricht, the Netherlands). Coagulation was triggered by recalcification in the presence of 5 pm recombinant human tissue factor (Innovin®, Dade Behring, Marburg, Germany), 4 μm sonicated, synthetic, procoagulant phospholipids containing phosphatidylserine (20 mol%), phosphatidylethanolamine (20 mol%), and phosphatidylcholine (60 mol%) (Avanti Polar Lipids, Alabaster, AL, USA ), and 417 μm fluorogenic substrate Z-Gly-Gly-Arg-AMC (Bachem, Bubendorf, Switzerland). Fluorescence was monitored with a fluorometer (Fluoroscan Ascent, ThermoLabsystems, Helsinki, Finland), and the ETP was calculated using the Thrombinoscope® software (Thrombinoscope B.V.).
The sensitivity to rAPC (American Diagnostica, Stamford, CT, USA) was investigated in all plasma samples mixed 1:1 with fresh-frozen PNP in order to diminish the influence of reduced coagulation factor activity as a result of sample deterioration or warfarin use. The ETP of each plasma sample was determined in both the presence and absence of 5 nm APC (f.c.). Because of loss of functional activity and batch-to-batch variability, minor day-to-day adjustments of the rAPC concentration were necessary in order to maintain a stable inhibition of ETP of about 80–90% in PNP with rAPC. The inter-assay coefficient of variation was 5.6% (mean=1300 nm min, n =8) for the ETP in PNP and 6.8% (mean=86%, n =8) for inhibition of ETP in PNP with rAPC.
PNP was run in parallel on each plate. Data were expressed as normalized ETP (ETPTest Plasma/ETPPNP), and as normalized percent inhibition of ETP by dividing the percent inhibition of ETP in test plasma caused by APC by the percent inhibition of ETP determined in PNP alone (Fig. 1):
Although most of the data were normally distributed, the comparison groups did not have similar variances (Fig. 2), and non-parametric methods were therefore applied. Results were reported as medians with 95% confidence intervals (CIs). The two-sided Mann–Whitney test was used for non-parametric comparison of test results derived from patient and control groups. Spearman’s rank correlation coefficient rs(=ρ) was calculated in order to evaluate the statistical significance of a possible association between the continuous variables. When evaluating the observed differences in medians between groups, P-values less than 0.05 were considered statistically significant. Statistical analyses were performed with SPSS (v. 11.0) software (SPSS Inc., Chicago, IL, USA), and CIs for medians were obtained from tabulated ranks according to Altman .
Ex vivo thrombin generation in plasmas from patients with LA and controls
In patients with LA not using warfarin, the normalized ETP was lower than in controls, with a median of 0.85 (95% CI: 0.81–0.93) for patients without APS (n =29) and 0.94 (95% CI: 0.83–1.00) for patients with APS (n =18) as compared with 1.01 (95% CI: 0.96–1.05) for controls (n =53; P <0.001, both LA subgroups when compared with controls). In patients with LA and APS on warfarin therapy (n =34) the median normalized ETP was 0.63 (95% CI: 0.59–0.66) as compared with 0.73 (95% CI: 0.70–0.77) in patients without LA on warfarin therapy (n = 38, P < 0.001; Table 1).
Table 1. Normalized endogenous thrombin potential (ETP) in the absence and presence of recombinant activated protein C (rAPC) and percent inhibition of ETP in plasmas from patients with lupus anticoagulants (LA) and controls
Normalized percent inhibition of ETP
Values are given as medians and 95% confidence intervals. Data are normalized against corresponding values for pooled normal plasma (PNP). The comparisons between patients with LA and controls are performed by using the two-sided Mann–Whitney test. * indicate a significant (P-value < 0.001) difference in median normalized percent inhibition of ETP between patients with LA and the corresponding control group according to warfarin use (‘All APS’ is compared with ‘All controls’).
Patients with LA
No APS, no warfarin
APS, no warfarin
Effects of LA on APC resistance and association with thrombotic events
The median normalized percent inhibition of ETP in all plasmas from patients with LA (61.4%, 95% CI: 45.8–74.5%, n =81) was lower than in plasmas from controls (107.8%, 95% CI: 107.1–109.3%, n =91), and the difference was statistically highly significant (P <0.001), consistent with increased APC resistance in the patients. Subgroup analysis of patients with LA with respect to the presence of APS and ongoing warfarin therapy showed a substantially lower median normalized percent inhibition of ETP in patients with LA and APS (n =52) compared to those with LA, but no APS (n =29; 44.6%, 95% CI: 30.1–55.7% vs. 78.8%, 95% CI: 73.9–95.8%; P =0.003). This difference persisted even when patients on warfarin therapy (n =34) were excluded from the APS subgroup (52.0%, 95% CI: 41.0–81.2% vs. 78.8%, 95% CI: 73.9–95.8%; P = 0.03). All patient subgroups independently showed lower median normalized percent inhibition of ETP than corresponding control subgroups, and all differences were statistically significant (all P <0.001; Table 1, Fig. 2). The above observations may alternatively be presented as a statistically significant difference in normalized ETP measured with APC between patients and controls (Table 1). However, differences in ETP measured without APC would then not be taken into account. Median normalized percent inhibition of ETP in plasmas from APS patients with venous thrombosis (32.6%, 95% CI: 26.9–57.0%, n =34) tended to be lower than in plasmas from APS patients with arterial thrombosis and/or obstetrical complications (52.0%, 95% CI: 35.8–81.2%, n =18), but the difference was not statistically significant (P =0.2).
Association between APC sensitivity, coagulation lag time, and dAPTT-based lupus ratio
APC sensitivity, expressed as the normalized percent inhibition of ETP, in plasmas from all patients with LA (n =81) was inversely correlated with the prolongation of normalized lag time (time to thrombin burst) measured with the CAT assay (rs=0.78, P < 0.001). The strength of the correlation was maintained even if the groups were divided according to warfarin use: patients not on warfarin therapy (n =47), rs=−0.81, P <0.001; patients on warfarin therapy (n =34), rs=−0.80, P <0.001 (Fig. 3A). In accordance with this finding, there was an inverse correlation between the normalized percent inhibition of ETP and another parameter of the in vitro‘anticoagulant’ effect of LA, the dAPTT-based lupus ratio: all patients with LA (n =81), rs=0.62, P < 0.001; patients not on warfarin therapy (n =47), rs=−0.58, P < 0.001; patients on warfarin therapy (n =34), rs=−0.69, P < 0.001 (Fig. 3B).
APC resistance and warfarin use
In patients with LA and APS on warfarin therapy (n =34) median normalized ETP in the presence of APC tended to be higher than in corresponding patients not on warfarin therapy (n = 18; 3.69, 95% CI: 2.38–4.50 vs. 2.61, 95% CI: 2.02–5.04; P =1.0) as opposed to the normalized ETP in the absence of APC (0.63, 95% CI: 0.59–0.66 vs. 0.94, 95% CI: 0.83–1.00; P < 0.001). Accordingly, the normalized percent inhibition of ETP by rAPC tended to be lower in patients with LA and APS on warfarin therapy than in corresponding patients not on warfarin therapy (33.8, 95% CI: 28.8–55.7 vs. 52.0, 95% CI: 41.0–81.2; P =0.3; Table 1).
In controls on warfarin therapy (n = 38), the median normalized ETP in the absence of APC was lower than in controls not on warfarin therapy (n = 53; 0.73, 95% CI: 0.70–0.77 vs. 1.01, 95% CI: 0.96–1.05; P <0.001). This is in accordance with the effect seen in patients with LA and APS. However, a markedly reduced median normalized ETP was seen in the presence of APC in controls on warfarin therapy as compared to controls not on warfarin therapy (0.17, 95% CI: 0.13–0.28 vs. 0.58, 95% CI: 0.54–0.72; P <0.001). The normalized percent inhibition of ETP was slightly higher in controls on warfarin therapy than in corresponding controls not on warfarin therapy (114.9, 95% CI: 111.0–121.6 vs. 106.7, 95% CI: 105.7–107.5; P <0.001; Table 1).
There was statistically no significant difference in median INR between LA patients and controls on warfarin, and there were equal proportions of warfarin users among included patients and controls. No correlation was seen between INR levels and normalized percent inhibition of ETP in patients with LA on warfarin therapy (rs=0.02, P =0.9).
Most of the plasmas from patients with LA in our study showed resistance to the anticoagulant effect of APC expressed as a reduced ability of exogenous rAPC to reduce ex vivo thrombin generation measured with the CAT assay. A wide range of inhibitory effects was evident, and may partly be regarded as a reflection of the heterogeneous nature of LA. Interestingly, subgroup analysis revealed that a history of thrombotic events in patients with LA was associated with a stronger APC resistance.
Data on ETP with and without APC have also been presented (Table 1), but we have mainly focused on percent inhibition of ETP, for the following reason. With LA present, ETP without APC may be substantially reduced because of in vitro inhibition of procoagulant reactions. In this case, ETP with APC may underestimate the degree of APC resistance. As clearly seen in controls, warfarin therapy substantially reduces the ETP both in the absence and presence of APC. In this setting, the percent inhibition of ETP by APC still reflects APC sensitivity, whereas ETP with APC reflects the combined effects of warfarin and APC.
An in vitro inhibition of procoagulant reactions was evident in most of the plasmas with LA, expressed as a median reduction in ex vivo thrombin generation compared to controls in the absence of APC, However, a net procoagulant effect in the presence of APC can nevertheless be demonstrated in most patients, as APC anticoagulant activity seems to be more potently inhibited than prothrombin activation. Actually, the potent inhibition of APC present in most patient plasmas results in a 2- to 3-fold increase in ex vivo thrombin generation compared to controls in the presence of APC, and this difference between patients with LA and controls can be demonstrated independent of warfarin use.
This selective inhibitory potency, which might provide an explanation for an increased thrombotic risk in patients with LA, has previously been shown to be highly dependent on the composition of phospholipids involved [24,25]. The membrane requirements for the APC complex differ from those of procoagulant complexes, especially with respect to the presence of phosphatidyletanolamine (PE), and PE dependence for anti-APC activity of LA plasmas has been documented .
LA probably interfere with both procoagulant and anticoagulant complexes on membrane surfaces in vivo. Therefore, if the CAT assay was performed in the absence of a natural anticoagulant (e.g. APC), the assay would be insensitive to an overall procoagulant state caused by the LA mediated inhibition of the anticoagulant. This may be illustrated by the fact that thrombin generation assayed with a chromogenic assay has been reported to be substantially reduced in patients with LA . With the CAT assay, the ETP has previously been reported to be within the limits of normal controls [11,12]. However, a wide range of inter-individual differences in ETP has also been reported, and in some plasmas with LA the ETP is substantially reduced . Previous studies report on the dissociated effects of LA on prothrombinase and APC activity . In contrast, we found a correlation between APC resistance and CAT lag time or dAPTT lupus ratio. This finding could indicate that strong LA associated with a more potent inhibition of procoagulant complexes in the absence of APC, as seen in clotting assays, exert proportionally even stronger inhibition of APC and thus induce a net procoagulant state in the presence of APC. If the effects of APA on APC activity correlate with the effects on procoagulant mechanisms, this may suggest that both inhibitory processes are mediated by the same mechanism: blocking the access of other proteins to the anionic phospholipids. This could be an important conclusion and could help us to understand the mechanism of APA in vitro and in vivo.
The APS is associated with both arterial and venous thrombosis, and different pathogenetic mechanisms may be involved. In our study, we did not find any statistical evidence in support of an association between APC resistance and the type of thrombotic event. However, as APC resistance generally is associated with venous thrombosis only, and a lower median normalized percent inhibition of ETP in patients with venous thrombosis compared to arterial and/or obstetrical events was seen, one might hypothesize that APC resistance in patients with APS behaves the same way.
The sensitivity of the CAT assay to the anticoagulant effect of warfarin therapy is reduced by 1:1 mixing with PNP. However, median ex vivo thrombin generation is still reduced to about 70% in control plasmas from warfarin users compared to non-users. As the APC resistance test is performed with a fixed amount of exogenous APC not affected by warfarin and the baseline procoagulant potential is reduced in plasmas from warfarin users, one would expect a more marked relative reduction of thrombin generation. In accordance with this, we found a median reduction of ex vivo thrombin generation in warfarin users to 30% of that seen in non-users in the presence of APC (Table 1). The normalized percent inhibition of ETP is thus actually stronger in controls on warfarin therapy compared to controls not on warfarin therapy.
In the LA positive patient group, this confounding effect of warfarin is by far counteracted by resistance to the anticoagulant effect of APC. Plasmas from APS patients on warfarin therapy actually show stronger resistance to APC than plasmas from APS patients not on warfarin therapy. This difference may reflect the predominance of venous thrombosis seen in APS patients on warfarin or generally a more serious clinical condition in patients maintained on therapy after their thrombotic event.
Acquired APC resistance expressed as a clotting time or thrombin generation ratio has become a well-documented independent risk factor for venous thrombotic events [1–3,5,14,15]. It has been postulated that the net amount of ex vivo thrombin generation expressed as the ETP in the presence of APC may be a relevant indicator of thrombotic risk in patients with LA. However, exogenous rAPC at a fixed and arbitrarily chosen concentration may not properly reflect the endogenous PC system in vivo, and the conditions that best represent the in vivo situation still remain to be determined .
The present ETP assay probed with exogenous rAPC is insensitive to thrombomodulin (TM) dependent APC activation, which has been reported to be inhibited by antibodies to TM in some patients with LA and thrombosis . However, using the CAT assay, a discrepancy between resistance to TM and APC has not yet been demonstrated .
In conclusion, the CAT assay with APC demonstrates increased ex vivo thrombin generation in most plasmas from patients with LA compared to plasmas from controls independent of warfarin use, and the degree of resistance to APC seems to be associated with thrombotic events.
Disclosure of Conflict of Interests
The authors state that they have no conflict of interest.