HIV-infected patients appear to be at increased risk for venous thrombotic events (VTEs), such as deep vein thrombosis or pulmonary embolism, with numerous case reports of idiopathic (unprovoked or primary) and/or atypical VTEs.1-3 An increased risk for VTE in HIV disease has recently been noted by Sullivan et al,4 who performed a retrospective longitudinal medical record review of 42,000 HIV-infected adults observed for an average of 2.4 years. The incidence of thrombosis in this HIV-infected population was statistically increased over that expected. On multivariate analysis, the factors associated with clinical thrombosis included older age (≥45 years), history of cytomegalovirus or other opportunistic infections, use of megestrol acetate or indinivir, or history of hospitalization. Of importance, however, 53% of the thrombotic events occurred in individuals without a history of recent hospitalization, suggesting that factors other than immobility may be operative.
A second study has recently confirmed a statistically significant increase in VTE among HIV-infected individuals.5 This study evaluated medical records from approximately 37,000 HIV-infected veterans and a similar number of age-, race-, and site-matched, HIV-negative controls from the Veterans Affairs system. In both the pre-highly active antiretroviral therapy (HAART) and HAART eras, the incidence of VTE was statistically increased among HIV-infected individuals, independent of underlying history of malignancy, opportunistic infection, or use of central venous catheters.5 Thrombosis was also found to be associated with a significant decrease in survival.
It is thus apparent that venous thrombotic disease may be increased in the setting of HIV infection, independent of many of the more commonly recognized risk factors. However, the precise biologic mechanisms for this propensity have not yet been well defined. Several studies have shown a decrease in protein S activity in HIV-infected patients, with and without thrombosis.1,6-9 Other possible abnormalities of coagulation reported in association with HIV have included deficiencies of protein C9 and heparin cofactor II10 and abnormalities in fibrinolysis. Furthermore, anticardiolipin antibodies (ACAs) have been found in approximately 45% to 50% of HIV-infected patients,11 and their prevalence may be higher in patients with clinical AIDS.12 Anticardiolipin antibodies are also associated with a hypercoagulable state.13
There is a well-documented association between acute and chronic inflammation and activation of the hemostatic system; the same cytokines responsible for this activation are also up-regulated during the course of HIV infection.14,15 Thus, inflammatory cytokines, such as tumor necrosis factor α, interleukin 1, and interleukin 6, activate the coagulation pathway16 and may also down-regulate the expression of numerous proteins necessary for fibrinolysis.17 Together, these changes could result in a propensity for venous thrombotic disease. Inflammation can also result in a decrease in functional protein S, the cofactor necessary for protein C activity.18 Elevated factor VIII levels are also a component of the inflammatory state,19 and elevated factor VIII activity has been independently associated with an increased prevalence of deep vein thrombosis in HIV-negative individuals.20
HIV-infected patients appear to have a number of underlying pathogenic mechanisms that could predispose to VTE. In an attempt to ascertain if HIV is associated with the development of changes in coagulation predisposing to thrombosis, we studied a group of 94 HIV-infected women and 50 HIV-negative comparators, enrolled in the Women's Interagency HIV Study (WIHS). Women with conditions known to be associated with an increased risk for VTE were excluded, including those who were pregnant or using hormones for contraception or replacement, those with history of cancer or autoimmune disease, and those with recent surgery or acute infection. We hypothesized that HIV infection would be associated with elevated factor VIII activity and decreased levels of functional protein S activity and that the severity of these changes would directly correlate with progressive and advancing HIV infection.
The WIHS is a prospective study of HIV-1 infection in women in the United States. From October 1994 to November 1995, 2056 HIV-1-infected women were enrolled at 6 sites (Los Angeles, CA; Washington, DC; San Francisco, CA; New York City/Bronx, NY; Brooklyn, NY; and Chicago, IL), recruited from HIV clinics, street outreach, referral from other studies, and word of mouth. The WIHS methods and cohort characteristics have been published.21 Study participants are seen every 6 months and undergo a standardized history, physical examination, and blood tests at each visit.
Written informed consent was obtained from all participants; the study was approved by the institutional review board.
The current study was a cross-sectional analysis, using data and specimens collected prospectively from participants enrolled in the Los Angeles site of the WIHS, obtained at their second study visit, between April 1995 and May 1996. The sample was selected to represent each of the following 4 groups: (1) HIV-positive women with clinical AIDS at baseline and CD4 count of less than 200 cells/μL; (2) HIV-positive women without clinical AIDS at baseline, but with CD4 count of less than 200 cells/μL (immunologic AIDS); (3) HIV-positive women without clinical AIDS at baseline and with CD4 count of more than 200 cells/μL (asymptomatic HIV), and (4) HIV-negative women. Twenty participants with insufficient or refrozen plasma samples from visit 2 were excluded. Patients were also excluded if they had medical conditions or were taking medications that might influence the study measures, including those who were (1) on oral or other contraceptive hormones or hormonal therapy, (2) were pregnant or within 6 months of pregnancy, (3) had history of cancer or underlying autoimmune disease, (4) had history of anticoagulant therapy, (5) had any acute, active, HIV-related infection at study visit 2, or (6) had history of surgery within 6 months of study visit 2. Although we excluded women with acute infection occurring within 6 months of the study, we did not exclude women with more distant history of such infection (>6 months from study entry), as was present in our group 1 women, with history of clinical AIDS at baseline and CD4 count of less than 200 cells/μL.
We attempted to accrue 50 participants into each group, all of equivalent age. In groups 1 and 2, all qualifying participants were accrued, because fewer than 50 eligible women were available for study in each group. In groups 3 and 4, the first 50 participants satisfying all requirements were used. One participant in group 3 was later excluded because of a history of autoimmune disease, leaving 49 participants in this group.
Data from 144 women were used, including 34 with clinical AIDS and CD4 count of less than 200 cells/μL, 11 without clinical AIDS but with CD4 count of less than 200 cells/μL, 49 HIV-positive without clinical AIDS and with CD4 count of more than 200 cells/μL, and 50 HIV-negative women.
Definition of Clinical AIDS
Clinical AIDS was defined as history of any of the following conditions, associated with underlying HIV infection: opportunistic infections, AIDS-related cancers (Kaposi sarcoma; high- or intermediate-grade, B-cell non-Hodgkin lymphoma; or invasive cervical cancer), wasting syndrome, AIDS dementia complex, Mycobacterium tuberculosis, or recurrent bacterial pneumonia in an HIV-infected individual.
Antiretroviral Medication Use
This cross-sectional analysis was conducted using blood specimens collected between April 1995 and May 1996, before the widespread availability or use of HAART. Antiretroviral therapy, if used at all, consisted of either single-agent use (monotherapy) or combination therapy, including 2 antiretroviral agents.
CD4 Cell Counts and HIV-1 RNA Levels
Laboratory analyses performed as part of the WIHS visit 2 were used. Lymphocyte subsets were analyzed within 24 hours of collection, using standard flow cytometry techniques in our laboratory, which participates in the National Institutes of Health (NIH), Division of AIDS Quality Assurance Program. Plasma HIV RNA levels were measured using freshly frozen specimens separated within 6 hours of phlebotomy using the isothermal NASBA method (Organon-Teknika, Durham, NC), with a minimum value of 4000 copies/mL. These assays were performed in laboratories participating in the NIH-National Institute of Allergy and Infectious Diseases (NIAID) virology quality assurance program.
Plasma Collection and Storage
Whole blood was collected in CPT tubes, kept at room temperature, and centrifuged within 6 hours. Centrifugation was performed at room temperature in a horizontal rotor (swing-out head), at 1500g for a minimum of 20 minutes. Plasma and mononuclear cell layers were then mixed, by inverting the CPT once, and were then shipped to the central WIHS laboratory, where the layer above the gel was removed. Peripheral blood mononuclear cells and plasma were then separated by centrifugation at 400g spin for 10 minutes, followed by removal of the plasma layer and centrifugation at 1000g for 10 minutes, to remove any contaminating peripheral blood mononuclear cells and platelets. The plasma was then stored at −150°C or liquid nitrogen until tested, using methods in which stability over time has been shown.22 Aliquots of plasma used for this study had not previously been thawed and refrozen.
All coagulation assays were preformed on STA analyzer (Diagnostica Stago, Asniers, France). Factor VIII activity was measured in a modified aPTT assay using immune factor VIII-depleted plasma (STA-deficient VIII; Diagnostica Stago). Functional protein S was determined by measuring the cofactor activity of protein S, which enhances the anticoagulant activity of activated protein C, resulting in the prolongation of a modified aPTT assay with protein S-deficient plasma enriched with factor Va (STACLOT Protein S; Diagnostica Stago).23 The lupus anticoagulant (LA) was detected using a modified aPTT with hexagonal phase phosphatidylethanolamine correction of LA prolonged assay (STACLOT LA 20; Diagnostical Stago).24 Further confirmation of the presence or absence of LA was obtained using a dilute Russell viper venom assay, performed on all subjects.25
C-Reactive Protein in Plasma and in Serum
High-sensitivity CRP analyses were performed using N high-sensitivity CRP (Dade Behring, Newark, DE), run on a BNII nephelometer (Dade Behring).
Medians and interquartile ranges (IQRs) for the variables, age, factor VIII activity, protein S activity, plasma CRP, serum CRP, CD4 count, HIV-1 RNA, and the frequency distributions of negative and positive antiphospholipid antibody, were calculated by group. Interquartile range includes the 25th and 75th percentile values, representing the extent of variability in the sample without undue emphasis on extremes, which can occur when data are highly skewed. Factor VIII activity levels were categorized as low (<57% activity), normal (57%-158%), or high (>158%). Protein S activity levels were categorized as low (<58%), normal (58%-137%), or high (>137%), based on the standard laboratory cutoff values in our laboratory.
A rank-based general linear model was used to compare factor VIII, protein S, and serum CRP levels between groups. All comparisons were adjusted for age. Scheffé multiple-comparison procedure was used to evaluate pairwise group comparisons.
All data were analyzed using SAS, version 9.0 (SAS Institute, Cary, NC).
As demonstrated in Table 1, all participants were female, with a median age of 34.2 years (range, 20-59 years). Latinas comprised 50%, 25.0% were African Americans, 18.8% were white, and 6.2% were from other ethnic groups. By design, median CD4 cell counts at the time of coagulation testing differed among the groups (P < 0.001). Likewise, HIV-1 viral load varied, ranging from a median value of nondetectable (<4000 copies/mL) in HIV-positive women with higher CD4 cells (>200 cells/μL), to a median of 61,735 copies/mL in HIV-infected women with immunologic AIDS and a median of 278,143 copies/mL in women with a clinical AIDS diagnosis (P < 0.001 among groups). No antiretroviral therapy was used in 57.1% of women with asymptomatic HIV infection, 9.1% of those with immunologic AIDS, and 29.4% of those with clinical AIDS (Table 1).
Factor VIII Levels
Factor VIII levels varied significantly among the groups (P < 0.001), increasing progressively with more advanced stages of HIV infection (Table 1). The median factor VIII level in HIV-negative women was 116.5% (IQR, 97%-154%), within the expected range of normal (57%-158%). In contrast, asymptomatic, HIV-infected women with CD4 count of more than 200 cells/μL had significantly higher levels of factor VIII (median, 149%; IQR, 111%-202%) (P < 0.05 when compared with HIV-negative women). The factor VIII level was higher still in women with immunologic AIDS (median, 196%; IQR, 150%-234%) (P < 0.001) and reached a median of 211% (IQR, 174%-250%) in women with history of full-blown clinical AIDS (P < 0.001).
Protein S Levels
The median protein S activity levels significantly varied among the groups (P < 0.001), with progressive decreases in protein S levels with advancing HIV disease. The median protein S level in the HIV-negative women was 75.5% (IQR, 66%-85%), well within the expected range of normal (58%-137%). The median protein S level in asymptomatic HIV-infected women with CD4 count of more than 200 cells/μL was not significantly different from that of HIV-negative controls (median, 67%). In contrast, the median protein S was 62% (IQR, 55%-67%) in women with immunologic AIDS (P < 0.05 vs HIV-negative women) and 46% (IQR, 40%-65%) among women with clinical AIDS (P < 0.001 vs HIV-negative women; P < 0.001 vs HIV-positive women with CD4 count >200 cells/μL).
An inverse relationship between rising factor VIII and declining protein S levels was apparent in these women (Fig. 1A), with factor VIII levels increasing (Fig. 1B) and protein S activity levels decreasing (Fig. 1C) as the CD4 counts declined and as the HIV-1 viral load increased (data not shown).
C-reactive protein levels in plasma did not significantly differ among groups. These values were then retested, using serum samples (Table 1), obtained on the same day of testing as the CRP in plasma. Again, no significant differences in serum CRP levels were apparent among the groups, whereas values from plasma and serum correlated almost perfectly with each other (correlation coefficient = 0.99). Median serum and plasma CRP levels were analyzed in terms of their relationships to factor VIII, protein S, body mass index (BMI), current smoking status, history of injection drug use, and evidence of hepatitis C infection. No significant relationship was apparent between CRP levels and these factors, with the exception of BMI, in which CRP values increased with increasing BMI (correlation coefficient r = 0.33, P < 0.001). Group differences in CRP remained nonsignificant with adjustment for age and BMI (data not shown). Nonetheless, CRP and viral load were significantly correlated (r = 0.26, P = 0.01), and this relationship remained when adjusted for age and BMI.
Lupus anticoagulant was tested in all patients and was negative in all.
This study is the first to document a clear relationship between increasing parameters of HIV disease, including lower CD4 cells, higher HIV-RNA levels, and history of clinical AIDS, with laboratory parameters of a hypercoagulable state, including progressive increases in factor VIII activity and decreasing levels of functional protein S activity. Both elevated factor VIII and decreased protein S activity levels have been associated with an increased risk for VTE and would provide a biologic mechanism for the increased prevalence of VTE now reported among HIV-infected individuals.4,5 It is important to note that women with known risk factors for VTE or those with disorders associated with markers of inflammation, including acute AIDS-related infection, pregnancy or surgery within the past 6 months, history of cancer or autoimmune disease, or use of hormonal contraception or replacement therapy, were excluded from this study. The increasing factor VIII and decreasing protein S activity levels documented herein thus represent those changes associated with HIV disease itself, as opposed to other acute or ongoing inflammatory conditions.
The reasons for elevated factor VIII levels and decreased protein S may relate to underlying inflammation of HIV disease itself, as has been well described.18,19 Indeed, the same inflammatory cytokines linked to activation of the coagulation system have also been well described in the setting of progressive HIV.14,15
Of interest, we did not find a relationship between elevated CRP levels, known to be associated with inflammation, and the abnormal values of factor VIII or protein S. In an attempt to elucidate this point more carefully, we repeated the CRP levels in plasma and in serum, finding a high correlation between the two and confirming the lack of association between CRP and elevated factor VIII or decreased protein S. Nonetheless, when evaluating the correlation between CRP and viral load, a significant correlation was found, with increasing CRP correlating with increasing viral load (P = 0.01). It is possible that the relatively weak association between CRP, factor VIII and protein S is simply a consequence of the relatively small numbers of subjects under study. In contrast, it is possible that CRP values and their prognostic significance may vary between men and women, as has recently been demonstrated26 or that the prothrombotic abnormalities we found do not relate to inflammation per se. This area will require further study.
A number of case reports and smaller clinical studies have reported low functional free protein S in patients with HIV infection.1,6-9 In agreement with our findings, this acquired protein S deficiency was most pronounced in patients with advanced HIV diagnosed with immunologic or clinical AIDS. However, in contrast to our own findings, two of these studies failed to show a significant correlation between the levels of free protein S antigen and the CD4 lymphocyte counts.7,8 In these reports, the decrease in functional protein S appeared to result from a decrease in both free and total protein S antigen.6-9 Thus, the decreased free protein S was probably not due to increased complex formation with C4b-binding protein. Although most studies confirm that a decrease in free protein S antigen is commonly observed with advanced HIV disease, there are no published reports of longitudinal measurement of serial protein S over time or studies to determine if immunologic reconstitution with HAART can result in an improvement in protein S levels. These questions should be addressed in future studies.
Individuals with antiphospholipid syndrome are at risk for clinical thrombosis, in addition to persistent laboratory abnormalities, including presence of LA, ACAs, or other antiphospholipid antibodies. Although ACAs may be associated with a hypercoagulable state,13 most studies have found no correlation between high levels of ACAs and clinical thrombosis in the setting of HIV infection.12 Furthermore, LA is a stronger predictor of thrombosis than ACAs in patients with antiphospholipid syndrome.27 It is for this reason that we chose to evaluate LA in this preliminary study. However, our analysis found no case of LA among study participants.
The current study comprised a relatively small number of subjects, and no information regarding clinical evidence of VTE over time was available in this single point-in-time cross-sectional design. However, we were careful to exclude patients with known risk factors for VTE. Despite exclusion of these higher-risk groups, the current data have clearly demonstrated the presence of an abnormal coagulation profile which progresses in magnitude with more advanced HIV disease.
In summary, we have shown that progressive stage of HIV disease, from asymptomatic to history of clinical AIDS, is associated with progressive abnormalities of factor VIII and protein S activity; both of which have been associated with development of VTE. Lupus anticoagulant does not appear to play a role in this advancing hypercoagulable state. Additional studies are currently under way, to elucidate the relationship between these coagulation abnormalities, use of HAART, and the development of clinical VTE over time in the national WIHS cohort.
The authors thank the female participants in the WIHS, who provided all samples and historical information for conduct of this study.
1. Iranzo A, Domingo P, Cadafalch J, Sambeat MA. Intracranial venous and dural sinus thrombosis due to protein S deficiency in a patient with AIDS. J Neurol Neurosurg Psychiatry
2. Park KL, Marx JL, Lopez PF, et al. Noninfectious branch retinal vein occlusion in HIV-positive patients. Retina
3. George SL, Swindells S, Knudson R, et al. Unexplained thrombosis in HIV-infected patients receiving protease inhibitors: report of seven cases. Am J Med
4. Sullivan PS, Dworkin MS, Jones JL, et al. Epidemiology of thrombosis in HIV-infected individuals. The Adult/Adolescent Spectrum of HIV Disease Project. AIDS
5. Fultz SL, McGinnis KA, Skanderson M, et al. Association of venous thromboembolism with human immunodeficiency virus and mortality in veterans. Am J Med
6. Bissuel F, Berruyer M, Causse X, et al. Acquired protein S deficiency: correlation with advanced disease in HIV-1-infected patients. J Acquir Immune Defic Syndr
7. Stahl CP, Wideman CS, Spira TJ, et al. Protein S deficiency in men with long-term human immunodeficiency virus infection. Blood
8. Sorice M, Griggi T, Arcieri P, et al. Protein S and HIV infection. The role of anticardiolipin and anti-protein S antibodies. Thromb Res
9. Erbe M, Rickerts V, Bauersachs RM, et al. Acquired protein C and protein S deficiency in HIV-infected patients. Clin Appl Thromb Hemost
10. Toulon P, Lamine M, Ledjev I, et al. Heparin cofactor II deficiency in patients infected with the human immunodeficiency virus. Thromb Haemost
11. Bloom EJ, Abrams DI, Rodgers G. Lupus anticoagulant in the acquired immunodeficiency syndrome. JAMA
12. Stimmler MM, Quismorio FP Jr, McGehee WG, et al. Anticardiolipin antibodies in acquired immunodeficiency syndrome. Arch Intern Med
13. Ginsburg KS, Liang MH, Newcomer L, et al. Anticardiolipin antibodies and the risk for ischemic stroke and venous thrombosis. Ann Intern Med
14. Fauci AS, Schnittman SM, Poli G, et al. NIH conference. Immunopathogenic mechanisms in human immunodeficiency virus (HIV) infection. Ann Intern Med
15. Emilie D, Peuchmaur M, Maillot MC, et al. Production of interleukins in human immunodeficiency virus-1-replicating lymph nodes. J Clin Invest
16. Hack CE. Tissue factor pathway of coagulation in sepsis. Crit Care Med
17. Nawroth PP, Handley DA, Esmon CT, et al. Interleukin 1 induces endothelial cell procoagulant while suppressing cell-surface anticoagulant activity. Proc Natl Acad Sci U S A
18. Taylor F, Chang A, Ferrell G, et al. C4b-binding protein exacerbates the host response to Escherichia coli
19. O'Donnell J, Tuddenham EG, Manning R, et al. High prevalence of elevated factor VIII levels in patients referred for thrombophilia screening: role of increased synthesis and relationship to the acute phase reaction. Thromb Haemost
20. Kraaijenhagen RA, Ankerint PS, Koopman MMW, et al. High plasma concentration of factor VIIIc is a major risk factor for venous thromboembolism. Thromb Haemost
21. Barkan SE, Melnick SL, Preston-Martin S, et al. The Women's Interagency HIV Study. WIHS Collaborative Study Group. Epidemiology
22. Vaara I, Jonsson M. Preservation of factor VIII activity in plasma by different freezing techniques, as determined with a chromogenic microtray assay. Scand J Clin Lab Invest
23. Wolf M, Boyer-Neumann C, Martinoli JL, et al. A new functional assay for human protein S activity using activated factor V as substrate. Thromb Haemost
24. Triplett DA, Barna LK, Unger GA. A hexagonal (II) phase phospholipid neutralization assay for lupus anticoagulant identification. Thromb Haemost
25. Guidelines on testing for the lupus anticoagulant. Lupus Anticoagulant Working Party on behalf of the BCSH Haemostasis and Thrombosis Task Force. J Clin Pathol
26. Khor LLC, Muhlestein JB, Carlquist JF, et al. Sex- and age-related differences in the prognostic value of C-reactive protein in patients with angiographic coronary artery disease. Am J Med
27. Galli M, Luciani D, Bertolini G, et al. Lupus anticoagulants are stronger risk factors for thrombosis than anticardiolipin antibodies in the antiphospholipid syndrome: a systematic review of the literature. Blood