Persons with HIV infection or AIDS may be at increased risk of thrombosis (venous thrombosis or pulmonary thromboembolism) because of decreased plasma concentrations of protein S [1,2] high levels of anticardiolipin antibodies  heparin cofactor II deficiency  venous damage related to injection drug use  tumors such as Kaposi's sarcoma  cytomegalovirus disease  therapy with megestrol acetate  or indinavir  or the chronic use of venous catheters  The early recognition of venous thrombosis is important because it often precedes pulmonary embolism, a preventable and common cause of death when hospitalized  The incidence of thrombosis has not been described for persons with HIV infection.
Data from January 1990 to December 1998 were analysed from patients in the Adult/Adolescent Spectrum of HIV Disease Surveillance (ASD) Project, a multicenter medical record review in 102 clinics in nine US cities, the methods of which have been described  A standardized instrument was used to collect information from medical records on the patients' clinical conditions and treatments at baseline and every 6 months until the patient's death or the last contact with the patient.
Individuals who had no prospective follow-up (i.e. for whom we had only an abstraction from the baseline record) were excluded. Because the definition of thrombosis was based on ICD-9 codes and one project site did not collect ICD-9 diagnosis codes after January 1995, records abstracted at that site after that date were excluded.
Thrombosis was defined by the ICD-9 codes 415.1 (pulmonary embolism and infarction), 434.0 (cerebral thrombosis), 452 (portal vein thrombosis), 453.2 (vena cava thrombosis), 453.3 (renal vein thrombosis), 453.8 (thrombosis of other specified veins), and 453.9 (thrombosis of unspecified site). The incidence of thrombosis per 1000 person-years (PY) of follow-up was calculated, with 95% confidence intervals (CI). The distribution of anatomical sites of thrombosis was determined according to ICD-9 codes. The cause of death was determined for patients who died during follow-up from the death certificate or from medical record abstraction; the proportion of patients with an antemortem diagnosis of thrombosis, for whom thrombosis was indicated as a cause of death, was calculated.
For logistic regression analysis, PROC GENMOD (SAS Institute, Cary, NC, USA) and generalized equation estimates were used, controlling for repeated observations of individuals over time, to determine factors independently associated with thrombosis among individuals with HIV infection. Regressors were demographic characteristics and factors reported in the medical literature as being associated with thrombosis (Table 1). Results are expressed as adjusted odds ratios with 95% CI.
A total of 335 instances of thrombosis were documented in 273 patients during 103 263 PY among 42 935 individuals. The incidence of thrombosis was 2.6/1000 PY (95% CI, 2.3–3.0), considering only the first thrombosis diagnosis for each patient. The incidence was lower for individuals with immunological AIDS (CD4 T lymphocyte count of < 200 cells/μl; 1.8/1000 PY; CI, 1.2–2.5) or HIV infection and no AIDS (1.3/1000 PY; CI, 1.1–1.6) than for persons with clinical AIDS (6.2/1000 PY; CI, 5.2–7.2). For 230 persons, only one thrombotic event was recorded; for 43 persons, more than one thrombotic event was recorded. The first recorded thrombotic events were most commonly: venous thrombosis, unspecified vein (111, 41%); venous thrombosis, specified vein (104, 38%); and pulmonary embolism (50, 18%). Among patients with an antemortem diagnosis of thrombosis, 167 (61%) died during follow-up. For 22 (13%) of these, thrombosis was indicated as a cause of death on the death certificate or in medical records.
In regression analysis, complete regressor information was available for 85 903 PY (171 805 6 month abstraction intervals) of follow-up; 324 thrombotic events were documented. Thrombosis was significantly associated with an age of 45 or more years, hospitalization, the prescription of megestrol acetate or indinavir in the past 6 months, and a diagnosis of cytomegalovirus (disease or retinitis) or other AIDS-defining opportunistic illness (Table 1). Of the thrombotic events in this analysis, 172 (53%) occurred in individuals without a history of recent hospitalization.
The medical literature contains case reports of thrombosis in persons with HIV infection: pulmonary embolism  retinal vein [14–17] peripheral vein [13,18] cerebral vein  and portal vein [9,20] thrombosis have been reported. The results of this analysis are consistent with earlier studies in that an association was found between thrombosis and the prescription of megestrol acetate [8,10] or indinavir  and cytomegalovirus disease  No increased risk of thrombosis was found among people who had injected drugs  probably because of the small number of current drug users in the analysis.
Thrombosis in individuals not known to be infected with HIV is associated with venous stasis, impaired anticoagulant response, decreased fibrinolysis, and vascular wall damage. For individuals with HIV infection, these predisposing factors may be present or exacerbated because of HIV disease. However, there are other, biologically plausible reasons why people with HIV infection may have an added risk of thrombosis. Acquired protein S deficiency may result in an hypercoagulable state, and certain opportunistic illnesses (such as cytomegalovirus disease  may alter the vascular endothelium. Increased levels of proinflammatory cytokines may also increase the risk of thrombosis by inducing a procoagulant state in the endothelium. Cytomegalovirus infection has also been associated with thrombosis after liver transplantation  this association is thus not unique to individuals with HIV infection, and may be attributable to local procoagulant changes in the vascular endothelium, or to associated clinical interventions such as use of intravenous devices. These data indicate that factors other than a high rate of hospitalization are associated with thrombosis among HIV-infected individuals; over half of the diagnoses in this analysis occurred in individuals with no history of hospitalization.
The ASD project includes diverse study sites and populations  but these data are not necessarily representative of all HIV-infected individuals in the United States. Furthermore, the under-ascertainment of thrombosis among ASD participants is likely because the project was not designed to document this condition and because many occurrences of thrombosis are asymptomatic and undiagnosed  Because of ambiguity in the ICD-9 coding scheme, some ICD-9 codes could represent occlusion of a vessel by a thrombus or by another mechanism. Also, physician diagnoses of thrombosis as documented in the medical record may include presumptive and confirmed cases.
We considered the possibility that the receipt of megestrol acetate and indinavir might be associated with thrombosis because of mutual association with other factors, such as dehydration and debilitation or advanced stage of disease. We attempted to address this concern by controlling for the diagnosis of wasting syndrome, for which megestrol acetate may be prescribed, and by including other protease inhibitors in the statistical model.
The relative importance of all predisposing factors for thrombosis could not be evaluated because we did not collect information on laboratory variables such as protein S levels, anticardiolipin antibodies, and heparin II cofactor. It is possible that these factors explain the association of thrombosis and clinical AIDS diagnoses (e.g. advanced stage of disease) in this analysis.
From these findings, physicians should be alert for thrombosis in recently hospitalized HIV-infected patients, those with AIDS-defining opportunistic illnesses, and those receiving indinavir or megestrol acetate. Because age greater than 45 years is associated with thrombosis, thrombosis may occur more commonly as the HIV-infected population ages. Further study is needed to determine the importance and the mechanism of the association between the prescription of indinavir and thrombosis.
1. Stahl CP, Wideman CS, Spira TJ, Haff EC, Hixon GJ, Evatt BL Protein S deficiency in men with long-term human immunodeficiency virus infection.
Blood 1993, 7:1801–1807.
2. Bissuel F, Berruyer M, Causse X, Dechavanne M, Trepo C Acquired protein S deficiency: correlation with advanced disease in HIV-1 infected patients.
J Acquired Immune Defic Syndr 1992, 5:484–489.
3. Stimmler MM, Quismorio FP, McGehee WG, Boylen T, Sharma OP Anticardiolipin antibodies in acquired immunodeficiency syndrome.
Arch Intern Med 1989, 149:1833–1835.
4. Toulon P, Lamine M, Ledjev I, Guez T, Hollerman MF Heparin cofactor II deficiency in patients infected with the human immunodeficiency virus.
Thromb Haemost 1993, 70:730–735.
5. Laing RBS, Brettle RP, Leen CLS Venous thrombosis in HIV infection.
Int J STD AIDS 1996, 7:82–85.
6. Kaufmann TN, Lourdes Z, Metroka C Thromboembolism in AIDS-related Kaposi's sarcoma.
JAMA 1991, 266:2834.2834.
7. Jenkins RE, Peters BS, Pinching AJ Thromboembolic disease in AIDS is associated with cytomegalovirus disease.
AIDS 1991, 5:1540–1542.
8. Prescribing information for megestrol acetate tablets.
In: Physicians' desk reference.
Montvale, NJ: Medical Economics Company; 1988.
9. Carr A, Brown D, Cooper DA Portal vein thrombosis in patients receiving indinavir, an HIV protease inhibitor.
AIDS 1997, 11:1657–1658.
10. Hirsh J, Hull RD, Raskob GE Epidemiology and pathogenesis of venous thrombosis.
J Am Coll Cardiol 1986, 8:104B–133B.
11. Farizo KM, Buehler JW, Chamberland ME et al
. Spectrum of disease in persons with human immunodeficiency virus infection in the United States.
JAMA 1992, 267:1798–1805.
12. Centers for Disease Control 1993 Revised classification system for HIV infection and expanded surveillance case definition for AIDS among adolescents and adults.
MMWR 1992, 41(No. RR-17):1–19.
13. Becker DM, Saunders TJ, Wispelwey B, Schain DC Case report: venous thromboembolism in AIDS.
Am J Med Sci 1992, 303:395–397.
14. Roberts SP, Haefs TMP Central retinal vein occlusion in a middle-aged adult with HIV infection.
Optom Vis Sci 1992, 69:567–569.
15. Mansour AM, Li H, Segal EI Picture resembling hemicentral retinal vein occlusion in acquired immunodeficiency syndrome: is it related to cytomegalovirus?
Opthalmologica 1996, 210:108–111.
16. Park KL, Marx JL, Lopez PF, Rao NA Noninfectious branch retinal vein occlusion in HIV-positive patients.
Retina 1997, 17:162–164.
17. Freidman SM, Margo CE Bilateral central retinal vein occlusions in a patient with acquired immunodeficiency syndrome.
Arch Opthalmol 1995, 113:1184–1188.
18. Tanimowo M Deep vein thrombosis as a manifestation of acquired immunodeficiency syndrome? A case report.
Cent Afr J Med 1996, 42:327–328.
19. Meyohas MC, Roullet E, Rouzioux C et al
. Cerebral venous thrombosis and dual primary infection with human immunodeficiency virus and cytomegalovirus.
J Neurol Neurosurg Psychiatry 1998, 52:1010–1011.
20. Narayanan TS, Narawane NM, Phadke AY, Abraham P Multiple abdominal venous thrombosis in HIV-seropositive patient.
Indian J Gastroenterol 1998, 17:105–106.
21. Ho DD, Rota TR, Andrews CA, Hirsh MS Replication of human cytomegalovirus in endothelial cells.
J Infect Dis 1984, 150:956–957.
22. Madalosso C, deSouza NF, Ilstrup DM, Wiesner RH, Krom RAF Cytomegalovirus and its association with hepatic artery thrombosis after liver transplantation.
Transplantation 1998, 66:294–297.
Adult/Adolescent Spectrum of HIV Disease Project Investigators
Susan Buskin, PhD, Seattle-King County Department of Public Health, Seattle, Washington; Arthur Davidson, MD, Denver Department of Health and Hospitals, Denver, Colorado; William McCarthy, PhD, AIDS Research Consortium of Atlanta, Atlanta, Georgia; Sharon Melville, MD, MPH, Texas Department of Health, Austin, Texas; Kaye Reynolds, Department of Health and Human Services, Houston, Texas; Judy Sackoff, PhD, New York City Department of Health, New York, New York; Frank Sorvillo, PhD, Los Angeles County Department of Health Services, Los Angeles, California; Susan Troxler, RN, MPH, Louisiana Department of Health, New Orleans, Louisiana; and Linda Wotring, PhD, Michigan Department of Community Health, Detroit, Michigan.