The number of new infections occurring in women and persons who inject drugs in the United States is now declining and the human immunodeficiency virus (HIV) epidemic is becoming more concentrated, with almost 70% of all infections occurring in men who have sex with men (MSM) in 2014.1 Overall, the number of new HIV diagnoses among MSM in the United States is stable, but that stability masks substantial regional and racial heterogeneity. New HIV diagnoses among MSM in places like Seattle and San Francisco are declining,2,3 but many areas, particularly in the southern United States, continue to report high but stable numbers of new diagnoses in MSM.4,5 Nationally, the proportion of all infections occurring in black MSM is increasing,1 and several reports have documented extremely high rates of incident infection in young black MSM.6,7
The magnitude of racial and regional disparities in HIV and how they are changing over time remain ill-defined. Life table analyses provide an estimate of a person's cumulative risk of experiencing a health outcome as they age.8–10 Although the exact proportion of men who are MSM remains uncertain, recent studies provide estimates of the proportion of the male population comprised of MSM,11–13 making it possible to create life tables estimating the age-specific risk of HIV among MSM. In this article, we present a life table analysis estimating trends in the incidence of HIV in successive birth cohorts of white and black MSM in King County, Washington and the state of Mississippi. We focus on these areas because of their demographic and cultural dissimilarity and prior evidence that the HIV epidemic in these areas may be moving in opposite directions.
We constructed life tables for the period 1982 to 2012 to estimate the cumulative risk of HIV diagnosis among MSM born between 1940 and 1994 and residing in King County, Washington and Mississippi.
MSM Population Size Estimates
We used US Census data to define the number of white and black males residing in the 2 areas in each year between 1982 and 2012. For each year, we estimated the number of male residents born in each year between 1940 and 1994. Because the census provides estimates for the number of men in 5-year age groups, we interpolated data to smooth changes in population size over time.
Our model required us to estimate the percentage of men who are MSM. Based on a meta-analysis of national, population-based studies conducted 1993 to 2008, Purcell estimated that 2.9% of men had sex with another man in the prior year, 3.9% did so in the prior 5 years and that 6.9% had ever had sex with another man.11 Surveys asking U.S. men about their sexuality consistently find that men in specific areas are more likely to report same sex sexual behavior, suggesting that MSM often migrate to urban centers.12,14 For example, the Behavioral Risk Factor Surveillance Survey (BRFSS) 2013 to 2014 found that 5.7% of adult men in King County report being gay or bisexual, while a Gallup poll found that 4.8% of Seattle metropolitan statistical area residents self-identify as lesbian, gay, bisexual or transgender15; both of these estimates exceed Gate’s national estimate that 3.6% of U.S. men identify as being gay, bisexual or transgender.16 Based on census data on the proportion of households with same-sex couples, Gates estimated that 6.5% of adults in the Seattle metropolitan statistical area, and 2.3% of adults in Mississippi were lesbian, gay, bisexual, or transgender.16 More recently, Grey used a variant of this approach and estimated that 8% and 1.8% of men in King County, WA and Mississippi, respectively, were MSM.13 However, other analyses suggest that survey findings showing that MSM concentrate in specific areas reflect a measure of how “out” the population is as much as true differences in proportion of the population that are MSM.17,18 For our base-case model we assumed that 4% of all US men are MSM and, based on King County BRFSS data, that 5.7% of men age of 30 years or older in King County are MSM, respectively. We assumed that 3% of Mississippi age older than 30 years are MSM, an estimate that assumes substantial net out-migration of MSM from the state.
In developing our model, we assumed that the likelihood of being an MSM did not vary among 18 year olds, that variance in the fraction of men who are MSM in different areas reflects migration. This assumption is supported by studies of twins, which suggest that being an MSM is substantially heritable,19 meaning that the number of children and adolescents who are MSM is not likely to vary greatly by geography. We did not have MSM-specific migration data. However, census data demonstrate that the number of men in each birth cohort residing in King County increases between the age of 18 and 29 and then plateaus. Thus, we assumed that the growth in the proportion of men within each birth cohort in King County who are MSM increases linearly from 4% to 5.7% between the ages of 18 and 29 years, and that the number of men who are MSM in Mississippi decreases from 4% to 3% between these same ages. We assumed that the proportion of men who are MSM does not vary by race.
We conducted sensitivity analyses examining cumulative lifetime risk of HIV by varying the percentage of men 30 years or older who are MSM from 2% to 4% in Mississippi and from 4% to 8% in King County. The ranges include estimates based on national Census data.13,16
New HIV Diagnoses
We used surveillance data from the enhanced HIV/AIDS Reporting System in King County and Mississippi to define the number of HIV diagnoses in MSM 18 years or older occurring each year by race and year of birth (Less than 1% of HIV diagnoses among MSM in both jurisdictions occurred in persons younger than 18 years). Washington State has legally required diagnosing clinicians to report cases of AIDS since 1985 and to report newly diagnosed HIV infections since 1999. The law has required laboratories to report positive HIV diagnostic tests since 1985, and to report all CD4 lymphocyte and HIV RNA test results since 2006. Mississippi has legally required medical providers and laboratories to report cases of AIDS and HIV since 1983 and 1988, respectively. Viral load and CD4 lymphocyte count reporting was not mandatory in Mississippi during the study period. HIV case reports in both areas include information about gender, age, race, and HIV transmission risk. In both areas, MSM behavior is ascertained through a combination of information on the case report, investigations of new cases that include review of medical records, and partner notification interviews.
Assessment of Potential Bias in Ascertainment of MSM Behavior
Changing estimates in the rate of HIV diagnoses in MSM could result from true changes in the rate of diagnosis or from changes in the ascertainment of MSM behavior among newly diagnosed cases. To assess for ascertainment bias, we compared the number of women, MSM, and male non-MSM diagnosed with HIV each year in King County, WA and in Mississippi, reasoning that increasing ascertainment of MSM risk would lead to a decline in the ratio of diagnoses in non-MSM men compared to women.
We constructed life tables for the period 1982 to 2012, including lines of data for men born each year 1940 to 1994, and columns for each age from 18 to 65 years. Equations used in the lifetables are presented in the Supplementary Appendix.
The study did not use any individual level data and was not human subjects research.
In 2012, an estimated 28,682 white and 2646 black MSM age 18 years or older resided in King County, WA, and 21,019 white and 11,324 black MSM were living in Mississippi (Table 1). As of the end of 2012, 7699 white and 753 black MSM had been diagnosed with HIV infection in King County, and 2030 white and 3555 black MSM had been diagnosed with HIV infection in Mississippi (Table 2).
The ratio of non-MSM male to female cases was elevated in Mississippi in the late 1980s and early 1990s, but since 1992 has varied little, from 0.89 to 1.2. Variance has been greater in King County, though the number of female cases is very low, from 1 to 62 each year. These findings suggest that ascertainment of MSM risk has been relatively stable over the most recent 20 years of the analysis period (Supplemental Fig. 1, http://links.lww.com/OLQ/A157).
Cumulative Risk of HIV Diagnosis Among Different Birth Cohorts
Figure 1 presents the lifetime risk of being diagnosed with HIV by age in 10-year birth cohorts among MSM born 1940 to 1994. Looking at sequential birth cohorts, the cumulative risk of HIV diagnosis progressed in 3 phases. In the first phase, age-specific risk increased in MSM born in successive birth cohorts between the 1940s and the mid-1960s. The highest risk occurred among MSM born in 1955 to 1964, a group that were young adults when the HIV/AIDS epidemic began in the 1980s. In King County, an estimated 45% of white and 65% of black MSM born 1955 to 1964 had been diagnosed with HIV by age 55 years. In Mississippi, approximately 22% of white MSM and 51% of black MSM in this cohort had been diagnosed with HIV by age 55 years.
The pattern of risk of new HIV diagnosis observed in second phase birth cohorts—men born 1965 to 1984—varied somewhat among the four groups defined by race and geography. In King County, risk among both white and black MSM declined, with an estimated 12% of white and 18% of black MSM born 1975 to 1984 being diagnosed with HIV by age 35 years, an almost 60% reduction in risk compared to men born 1955 to 1964. The pattern of HIV diagnoses among white MSM in Mississippi was similar to that observed in King County; the risk of HIV diagnosis declined in successive birth cohorts between 1965 and 1984. By age 35 years, 5% among white MSM born 1975 to 1984 had been diagnosed with HIV infection compared with 14% of white MSM born 1955 to 1964, a 64% decline. The pattern among black MSM in Mississippi was notably different. The age of diagnosis in men born 1965 to 1974 and 1975 to 1984 was earlier than in men born 1955 to 1964. However, among men born 1965 to 1974 the risk curve is essentially superimposed on that from the cohort born 1955 to 1965 at ages in the mid-30s forward, suggesting a similar level of cumulative risk despite earlier age of diagnosis. Risk of diagnosis declined somewhat in the cohort born 1975 to 1984, though the magnitude of that decline was lower than in King County and among white MSM in Mississippi.
The pattern of risk diverges dramatically in the most recent birth cohort, men born 1985 to 1994. Among both whites and blacks in King County and among whites in Mississippi, HIV risk is this youngest cohort is slightly higher than that seen in men born 1975 to 1984. In contrast, the risk among black MSM born 1985 to 1994 skyrockets. By age 27 years, approximately 35% of men in this cohort have been diagnosed with HIV. We did not observe a similarly elevated level of risk in any other studied birth cohort in either area.
Racial Disparities in Cumulative HIV Risk
In both King County and Mississippi, black MSM consistently experienced a higher risk of HIV infection than white MSM. However, the absolute and relative magnitude of these disparities, and how they changed over time differed between King County and Mississippi. Figure 2 highlights the comparison of cumulative HIV risk between black and white MSM in three selected 10-year birth cohorts. Among men born 1945 to 1954 in King County, we observed almost no disparity in risk during the initial decade of the epidemic, though the risk of HIV diagnosis among black MSM in this group exceeds that of white MSM at older ages by approximately 9% (44% vs 35%), a relative disparity of 21%. The size of both the absolute and relative disparity between blacks and whites is similar among King County MSM born 1965 to 1974. In contrast, the relative disparity is greater while the absolute disparity has remained stable among MSM born 1985 to 1994. Black MSM born in 1965 to 1974 were 33% more likely to have an HIV diagnosis by age 35 years than white MSM, whereas black MSM born in 1985 to 1994 were twice as likely to have been diagnosed with HIV by age 27 years compared with white MSM. At the same time, the absolute disparity between these final two cohorts was similar (7.2% and 7%) because the risk of HIV diagnoses declined among both blacks and whites.
Racial disparities in Mississippi were more pronounced than in King County (Fig. 2B) and displayed a pattern of consistently widening absolute and relative disparity. Comparing black and white MSM born 1945 to 1954 and 1965 to 1974, the absolute disparity in HIV diagnoses by age 35 years increased from 0.2%% to 26%, whereas the relative disparity increased from 11% to 75%. For the youngest birth cohort of men born 1985 to 1994, by age 27 years, black MSM were more than 10 times as likely to have been diagnosed with HIV than white MSM (35% vs 3%); the absolute disparity in the risk of HIV diagnosis increased to 32%.
Assuming that 8% of men 30 years or older in King County are MSM (rather than 5.7%), 33% of white and 47% of black MSM born 1955 to 1964 (peak risk cohort) would have been diagnosed by age 55 years (Table 3). Among men born 1985 to 1994 (contemporary young MSM cohort), 6% of white and 12% of black MSM would be diagnosed by age 27 years. If only 2% of Mississippi men 30 years or older are MSM, then 34% and 78% of white and black MSM in the peak birth cohort would have been diagnosed by age 55 years; 4% of white and 40% of black MSM in the contemporary young MSM cohort would be infected by age 27 years. The supplemental figure presents how varying assumptions on the percentage of men who are MSM affects our estimates of cumulative HIV diagnosis risk among persons in the 1955 to 1964 and 1975 to 1984 birth cohorts. As shown, varying this assumption does not substantially affect our estimates among young persons because our model assumes variance in the percentage of men who are MSM results from migration between the ages of 18 and 29 years.
We used life tables to estimate the cumulative, age-specific risk of HIV diagnosis among successive birth cohorts of MSM in King County, WA, and Mississippi. Our findings highlight the disparate trajectory of the HIV epidemic in these 2 areas. In King County, the risk of HIV diagnosis peaked among MSM born in the late 1950s and early 1960s declined in subsequent birth cohorts and is now relatively stable. Although important racial disparities persist, the absolute size of these disparities is stable and the risk of HIV has decreased in all groups. The epidemic in Mississippi is different. As in King County, HIV risk among white MSM has declined and stabilized. In contrast, the epidemic among black MSM is explosive, with approximately 35% of young black MSM diagnosed with HIV infection by age 27 years. To our knowledge, this level of risk in unprecedented in the United States and represents a public health crisis.
Our findings highlight the alarming risk of HIV experienced by black MSM and are consistent with some prior reports. Assuming that 4% of men are MSM, the estimated overall annual incidence of HIV among MSM in the United States in 2010 was under 1%.20 HIV Prevention Trials Network 061, a cohort study of black MSM in 6 US cities conducted 2009 to 2011, observed a 5.9% annual risk of HIV infection in men 30 years or younger, whereas smaller cohort studies of young African American MSM in Atlanta and Mississippi have observed incidences of 10% and higher.6,7,21 At the same time, HIV risk among black MSM is King County is much lower, highlighting that risk among black MSM is heterogeneous, and that high levels of infection in that population are not inevitable.
Our findings vary somewhat from other analyses related to the lifetime risk of HIV among MSM and highlight areas of continued uncertainty related to the demography of MSM. To our knowledge, the only published estimate of the lifetime HIV risk among U.S. MSM suggested that 40% of MSM will be HIV infected by age 40 years.22 That estimate was derived by applying a single annual incidence estimate to all MSM regardless of age, birth cohorts, race/ethnicity or geography. However, as illustrated by our data, incidence is not stable over the life course. More recently, CDC estimated the risk of HIV among MSM using recent age and race specific diagnosis rates and assuming that 6.9% of men are MSM.23 They estimated that half of all black MSM and 9% of white MSM would be diagnosed with HIV over their lifetime. Our findings highlight that these risks vary substantially in different parts of the United States and have changed over time.
Our analysis has a number of limitations. First, our estimate of the percentage of men who are MSM is imprecise. An overestimate of the proportion of men who are MSM would lead to an underestimate of the cumulative risk of HIV diagnoses, while an underestimate would lead to an overestimate of that risk. To address this issue, we conducted sensitivity analyses allowing readers to see how different estimates of the percentage of men who are MSM might affect the estimated risk of HIV among white and black MSM in Seattle and MS. Of note, imprecision in our estimate of the proportion of men who are MSM would not affect the trends we observed, nor would such an error affect our estimates of racial disparities unless there is variability in the proportion of men who are MSM by race. Second, we utilized surveillance data to define the number of HIV diagnoses in MSM. Incomplete ascertainment of MSM as a risk factor would lead to underestimation of the HIV incidence among MSM. Our analysis assessing the ratio of non-MSM male cases to female cases suggests that ascertainment has been stable over the most recent 20 years of the analysis period, though the risk of HIV among MSM in both King County and MS was likely underestimated early in the epidemic leading to a potential underestimate of cumulative HIV risk in the oldest birth cohorts. Third, because surveillance data provides the date of diagnosis, we estimated the risk of HIV diagnoses, not the risk of HIV infection. The risk of infection at younger ages is almost certainly higher than our estimate. Also, insofar as HIV testing is increasing, the time from HIV infection to diagnosis is declining24 and some of the trend toward higher rates of diagnosis in young MSM we observed may reflect changing testing patterns, not a longer term trend toward increasing lifetime risk. Fourth, there may have been differences in ascertainment of race over time given changes in the number of racial categories included in the US census, though we did not see significant anomalies in the proportions of men in different groups over time. Finally, recent studies suggest that national HIV surveillance data is likely incompletely deduplicated resulting in an overestimate of the number of persons living with HIV.25 This would lead to an overestimate of HIV risk among MSM which could vary by region. However, it seems likely that this limitation in surveillance data would affect men of all races within a single area, and thus is unlikely to explain the disparities we observed.
In conclusion, our findings illustrate the evolving heterogeneity in the US HIV epidemic. In places like King County, the epidemic peaked years ago in a population that is now in its 50s. The epidemic is relatively stable, and progress has affected both white and black MSM, though important racial disparities persist. In contrast, in places like Mississippi, we are facing a crisis, with rising racial disparities and an unprecedented incidence of HIV diagnoses occurring in young, black MSM. There is an urgent need for new, more aggressive and innovative prevention interventions focused on black MSM, particularly in the southern United States.
1. Birrell PJ, Gill ON, Delpech VC, et al. HIV incidence in men who have sex with men in England and Wales 2001–10: a nationwide population study. Lancet Infect Dis 2013; 13:313–318.
2. Das M, Chu PL, Santos GM, et al. Decreases in community viral load are accompanied by reductions in new HIV infections in San Francisco. PLoS One 2010; 5:e11068.
3. Golden MR, Bennett AB, Dombrowski JC, et al. Achieving the Goals of the National HIV/AIDS Strategy: Declining HIV Diagnoses, Improving Clinical Outcomes, and Diminishing Racial/Ethnic Disparities in King County, WA (2004–2013). Sex Transm Dis 2016; 43:269–276.
4. Louisiana Department of Health and Hospitals. Office of Public Health. 2013 STD/HIV Surveillance Report 2014.
5. Texas Department of State Health Services. Texas HIV Surveillance Report, 2014 Annual Report 2015.
6. Centers for Disease Control and Prevention (CDC) HIV infection among young black men who have sex with men—Jackson, Mississippi, 2006–2008. MMWR Morb Mortal Wkly Rep 2009; 58:77–81.
7. Sullivan PS, Rosenberg ES, Sanchez TH, et al. Explaining racial disparities in HIV incidence in black and white men who have sex with men in Atlanta, GA: a prospective observational cohort study. Ann Epidemiol 2015; 25:445–454.
8. Lee IM, Shiroma EJ, Lobelo F, et al. Effect of physical inactivity on major non-communicable diseases worldwide: an analysis of burden of disease and life expectancy. Lancet 2012; 380:219–229.
9. Pharoah PD, Sewell B, Fitzsimmons D, et al. Cost effectiveness of the NHS breast screening programme: life table model. BMJ 2013; 346:f2618.
10. Soneji S, Beltran-Sanchez H, Sox HC. Assessing progress in reducing the burden of cancer mortality, 1985–2005. J Clin Oncol 2014; 32:444–448.
11. Purcell DW, Johnson CH, Lansky A, et al. Estimating the population size of men who have sex with men in the United States to obtain HIV and syphilis rates. Open AIDS J 2012; 6:98–107.
12. Oster AM, Sternberg M, Lansky A, et al. Population size estimates for men who have sex with men and persons who inject drugs. J Urban Health 2015; 92:733–743.
13. Grey JA, Bernstein KT, Sullivan PS, et al. Estimating the population sizes of men who have sex with men in US states and counties using data from the American Community Survey. JMIR Public Health Surveill 2016; 2:e14.
14. Glick SN, Golden MR. Persistence of racial differences in attitudes toward homosexuality in the United States. J Acquir Immune Defic Syndr 2010; 55:516–523.
15. Newport F, Gate GJ. San Francisco Metro area ranks highest in LGBT percentage. Gallup 2015.
16. Gates GJ. Same-sex Couples and the Gay, Lesbian, Bisexual Populations: New Estimates from the American Community Survey. Los Angeles, CA, USA: Williams Instiute on Sexual Orientation Law and Public Policy, UCLA School of Law, 2006 October, 2006.
17. Stephens-Davidowitz S. How Many American Men Are Gay? New York Times 2013; 7:2013.
18. Stephens-Davidowitz S. Estimating the Closeted Gay Male Population. New York Times 2014; 10:2014.
19. Mustanski BS, Chivers ML, Bailey JM. A critical review of recent biological research on human sexual orientation. Annu Rev Sex Res 2002; 13:89–140.
20. Centers for Disease Control and Prevention. Estimated HIV incidence in the United States 2007–2010, 2012.
21. Koblin BA, Mayer KH, Eshleman SH, et al. Correlates of HIV acquisition in a cohort of Black men who have sex with men in the United States: HIV prevention trials network (HPTN) 061. PLoS One 2013; 8:e70413.
22. Stall R, Duran L, Wisniewski SR, et al. Running in place: implications of HIV incidence estimates among urban men who have sex with men in the United States and other industrialized countries. AIDS Behav 2009; 13:615–629.
23. Hess KHX, Lansky A, Mermin J, et al. Estimating the lifetime risk of a diagnosis of HIV infection in the United States. Boston, MA: Conference on Retroviruses and Opportunistic Infections, 2016.
24. Helms DJ, Weinstock HS, Mahle KC, et al. HIV testing frequency among men who have sex with men attending sexually transmitted disease clinics: implications for HIV prevention and surveillance. J Acquir Immune Defic Syndr 2009; 50:320–326.
25. Xia Q, Braunstein SL, Wiewel EW, et al. Persons living with hiv in the united states: fewer than we thought. J Acquir Immune Defic Syndr 2016; 72:552–7.