We studied the relationship of serum cholesterol, measured prior to HIV infection, with risk of death from AIDS in a large cohort. A major limitation of our study was the absence of information on HIV risk factors.
Between 1973 and 1975, 332 547 men aged 35-57 years were screened for a clinical trial on heart disease . The average serum cholesterol for those screened was 214 mg/dl, and blood pressure averaged 130/84 mmHg. These men were followed for mortality until the end of 1990, an average of 16 years. Death from AIDS was defined according to the International Classification of Diseases, ninth revision (ICD-9) codes 042—044 . In addition, two other outcomes were defined: death from 'possible AIDS' , and death from 'other infections' (ICD-9 codes 001-139, 460-466 and 480-487) excluding those defining AIDS or possible AIDS.
Two hundred and fifteen men died from AIDS during the 16 years of follow-up. In regression analysis, age, income, blood pressure and serum cholesterol were inversely related to risk of death from AIDS (P < 0.001 for each). Adjusted relative risk estimates for death from AIDS corresponding to cholesterol and blood pressure levels are shown in Table 1. For each measurement, there was no evidence of a threshold effect. From models which included cholesterol and systolic blood pressure (SBP) as continuous variables, a 40 mg/dl (approximately 1 mmol/l) lower serum cholesterol was associated with a 40% increased risk of death from AIDS (P = 0.0001), and a 10 mmHg lower SBP was associated with a 30% increased risk of death from AIDS (P = 0.0001).
Data on risk behaviors (e.g., homosexual activity or injecting drug use) were not collected; however, marital status was obtained from death certificates for those dying from AIDS. Age-adjusted serum cholesterol levels for men who were never married (151 deaths) and those who were married (64 deaths) were 197 mg/dl and 202 mg/dl, respectively (P = 0.40). Age-adjusted SBP and diastolic blood pressure (DBP) levels averaged 123/79 mmHg for men who were never married and 123/80 mmHg for men who had been married (P = 0.97 for SBP; P = 0.59 for DBP).
The cholesterol and blood pressure associations were specific to HIV. In multivariate analyses, serum cholesterol was not significantly associated with death from possible AIDS (P = 0.43; 50 deaths) or death due to other infections (P = 0.78; 595 deaths). SBP was positively associated with death from other infections (P = 0.0001).
An inverse association between cholesterol and blood pressure with death from AIDS could arise if those at high risk of becoming infected with HIV (e.g., homosexuals or bisexuals) had lower cholesterol and blood pressure than other men. Data on this are scant. Evans  compared 44 homosexual and 111 heterosexual men and found that cholesterol averaged 210 mg/dl for homosexuals and 248 mg/dl for heterosexuals; in addition, homosexuals weighed an average of 6.25 kg less than heterosexuals. Other studies have also found that homosexual men weigh less than heterosexual men [5-8]. Body weight is positively correlated with blood pressure and serum cholesterol [9,10]; a weight difference between homo-/bisexual men could explain, in part, both the cholesterol and blood pressure associations.
The presence of other infections at screening may also offer a partial explanation for our findings. Before HIV was documented, an excess prevalence of antigen or antibody to hepatitis B surface antigen among homosexuals was reported . Similarly, early in the AIDS epidemic it was recognized that homosexual men who developed AIDS were more likely to have had sexual contact with a large number of partners, and as a consequence were more likely to be infected with cytomegalovirus and hepatitis B and have a history of other sexually transmitted diseases . In a Chinese cohort, cholesterol levels were 4.2% lower among apparently healthy Chinese men with chronic hepatitis B virus infection compared with those who were not . Thus, it is possible that the men screened, who were at high risk for later becoming infected with HIV, had other infections at that time that may have lowered their cholesterol and blood pressure. Ettinger and Harris  have reviewed a number of mechanisms by which persons with acute and chronic illnesses may develop hypocholesterolemia.
In summary, low cholesterol and blood pressure are strong predictors of death from AIDS. Whether these associations relate to host defense, other infections, or unmeasured confounding variables is unclear.
We acknowledge the many Multiple Risk Factor Intervention Trial investigators for collection of the screening data and helpful comments from G. Bartsch, J. Cohen, M. Kjelsberg, L. Kuller, G. Davey Smith, J. Stamler and R. Stamler.
1. Neaton JD, Grimm RH, Cutler J: Recruitment of participants for the Multiple Risk Factor Intervention Trial (MRFIT).Controlled Clin Trials
1998, 8 (suppl 4):41S-53S.
2. National Center for Health Statistics: Instruction Manual Part 2e. Nonindexed Terms, Standard Abbreviations, and State Geographic Codes Used in Mortality Classification, 1990. Including WHO Amendments to ICD-9, Volume 2.
Hyattsville: US Department of Health and Human Services; 1990.
3. Chang HH, Morse DL, Noonan C, et al
.: Survival and mortality patterns of an acquired immunodeficiency syndrome (AIDS) cohort in New York State.Am J Epidemiol
4. Evans RB: Physical and biochemical characteristics of homosexual men.J Consult Clin Psychol
5. Tourney G, Petrilli AJ, Hatfield LM: Hormonal relationships in homosexual men.Am J Psychiatry
6. Silberstein LR, Mishkind M, Striegel-Moore RH, Tímko C, Rodin J: Men and their bodies: a comparison of homosexual and heterosexual men.Psychosom Med
7. Herzog DB, Newman KL, Warshaw M: Body image dissatisfaction in homosexual and heterosexual males.J Nerv Ment Dis
8. Gettleman TE, Thompson JK: Actual differences and stereotypical perceptions in body image and eating disturbance: a comparison of male and female heterosexual and homosexual samples.Sex Roles
9. Kannel WB, Brand N, Skinner Jr JJ, Dawber TR, McNamara PM: The relation of adiposity to blood pressure and development of hypertension. The Framingham Study.Ann Intern Med
10. Kannel WB, Gordon T (Eds): The Framingham Study: An Epidemiological Investigation of Cardiovascular disease, Sections 4 and 5.
Washington, DC: US Government Printing Office; 1968.
11. Szmuness W, Much MI, Prince AM, et al
.: On the role of sexual behaviour in the spread of hepatitis B infection.Ann Intern Med
12. Sonnabend JA: The etiology of AIDS.AIDS Res
13. Chen Z, Keech A, Collins R, et al
.: Prolonged infection with hepatitis B virus and association between low blood cholesterol concentration and liver cancer.BMJ
14. Ettinger WH, Harris T: Causes of hypacholesterolemia.Coron Artery Dis