The past decade has seen progress towards getting to zero HIV infections in San Francisco overall. New HIV diagnoses dropped from 521 in 2008 to 197 in 2018, a 62% decrease [1]. HIV incidence estimated by a CD4 model [2] corroborates a parallel, commensurate drop in incident infections, from a projected 330 in 2013 to 190 in 2017 [1]. Unfortunately, there is insufficient sample size to estimate HIV incidence separately for smaller populations who may have elevated risk. Directly measured rates of HIV seroconversion observed in longitudinal studies are logistically difficult and costly, and therefore rare [3]. Limitations of sample size and resources are substantial for estimating or measuring HIV incidence among transgender women (hereafter “trans women”), the population with the highest burden of HIV in San Francisco [1,4] and in many parts of the world [5].
We report results from the TransNational Study in San Francisco, the primary aims of which are to measure the rate of HIV seroconversion and identify predictors of HIV acquisition among trans women. The present analysis focuses on identifying demographic disparities in HIV risk, particularly testing hypotheses on whether HIV incidence is higher among trans women who are young, members of racial/ethnic minority groups, living in poverty, and homeless based on differences in HIV prevalence and recent trends in new diagnoses in San Francisco [1]. Trans women were defined by being assigned male sex at birth and currently identify as other than male gender. Trans women were enrolled into the cohort using a long-chain peer referral method previously used to accrue a cohort of young trans women in San Francisco [6] and MSM in Nanjing, China [7]. In brief, initial seeds were identified from diverse social networks and instructed to refer other eligible trans women to the study. Participants were interviewed face-to-face on demographics and risk behaviors and tested for HIV. Trans women who tested negative for HIV and were 18 years or older were eligible. Participants were re-tested and interviewed at 6, 12, and 18 months. Participants who seroconverted were linked to HIV care. We used an incidence density approach (i.e., the number of events divided by the person-time of follow-up) to calculate rates of HIV seroconversion. Incidence rate ratios (IRR) were used to compare differences in rates among subgroups of trans women assuming a Poisson distribution. When a subgroup had zero seroconversions, we calculated 97.5% one-sided confidence intervals and compared groups using the Z-test. Referrals to HIV prevention programs, including PrEP, were offered at each visit. Participants were given $55 for completion of the survey and HIV testing at their initial visit, increasing to $70 for the 18-month visit. Participants were given $20 for each eligible referral to the study. The protocol was approved by the Human Research Protection Program of the University of California San Francisco. Participants provided written informed consent.
Of 415 who were HIV-negative at enrollment and agreed to follow-up, 377 were seen at their 18-month visit and 8 seroconverted by the time of their 18-month visit (92.6% retention). The 8 seroconversions occurred over 604 person-years (py) of follow-up for an incidence rate of 1.3 per 100 py (95% confidence interval [CI] 0.7–2.7) (Table 1).
Table 1.
HIV incidence among trans women in a longitudinal cohort, San Francisco, 2017–2019
| Characteristics | Trans women HIV-negative at baseline | Trans women seroconverting to HIV+ during follow-up | Rate per 100 person-years (95% CI) | Bivariate Comparison | p-value | |||
|---|---|---|---|---|---|---|---|---|
| N | % | N | Person-years | IRR | 95% CI | |||
| Total | 415 | 100 | 8 | 604 | 1.3 (0.7, 2.7) | -- | -- | -- |
| Age (range 18–75 years) | ||||||||
| 18–24 | 57 | 13.7 | 3 | 80 | 3.7 (1.2, 11.6) | 3.9 | 0.9–16.0 | 0.04c |
| 25+ | 358 | 86.3 | 5 | 524 | 1.0 (0.4, 2.3) | Ref | -- | -- |
| Race/ethnicity | ||||||||
| White | 158 | 38.1 | 0 | 233 | 0 (0, 1.6)a | Ref | -- | -- |
| Black/African American | 29 | 70 | 1 | 402 | 2.5 (0.4, 17.7) | -- | -- | 0.29b |
| Hispanic/Latina/x | 133 | 32.1 | 5 | 196 | 2.6 (1.1, 6.1) | -- | -- | 0.03b |
| Other | 95 | 22.9 | 2 | 135 | 1.5 (0.4, 5.9) | - | -- | 0.15b |
| Race/ethnicity | ||||||||
| White | 158 | 38.1 | 0 | 233 | 0 (0, 1.6)a | Ref | -- | -- |
| Trans women of color | 257 | 61.9 | 8 | 371 | 2.2 (1.1, 4.3) | -- | -- | 0.01b |
| Education | ||||||||
| HS/GED or less | 171 | 41.2 | 5 | 244 | 2.1 (0.9, 4.9) | Ref | -- | -- |
| Some College/Technical | 138 | 33.3 | 2 | 201 | 1.0 (0.3, 4.0) | 0.5 | 0.1–2.5 | 0.38c |
| College degree or above | 106 | 25.5 | 1 | 158 | 0.6 (0.1, 4.5) | 0.3 | 0.04–2.6 | 0.26c |
| Annual income | ||||||||
| Above poverty limit | 0 | 0 | 0 | 0 | -- | -- | -- | |
| At or below poverty limit | 410 | 98.1 | 8 | 596 | 1.3 (0.7, 2.7) | -- | -- | 0.45 |
| Employed | ||||||||
| No | 230 | 56.1 | 2 | 327 | 0.6 (0.2, 2.5) | Ref | -- | -- |
| Yes | 180 | 43.9 | 6 | 269 | 2.2 (1.0, 5.0) | 3.7 | 0.7–18.1 | 0.09c |
| Current living situation | ||||||||
| Stably housedd | 201 | 48.4 | 2 | 307 | 0.7 (0.2, 2.6) | Ref | -- | -- |
| Unstably housede | 214 | 51.6 | 6 | 297 | 2.0 (0.9, 4.5) | 3.10 | 0.63–15.33 | 0.15c |
| Ever incarcerated | ||||||||
| No | 196 | 47.3 | 1 | 298 | 0.3 (0.1, 2.4) | Ref | -- | -- |
| Yes | 218 | 52.7 | 7 | 305 | 2.3 (1.1, 4.8) | 6.8 | 0.8–55.6 | 0.04c |
| Health care | ||||||||
| Currently insured | ||||||||
| No | 24 | 5.8 | 2 | 35 | 5.8 (1.5, 23.2) | Ref | -- | -- |
| Yes | 388 | 94.2 | 6 | 564 | 1.1 (0.5, 2.4) | 0.2 | 0.04–0.9 | 0.02c |
| Insurance type | ||||||||
| None | 24 | 6.1 | 2 | 35 | 5.8 (1.5, 23.2) | -- | -- | 0.15b |
| Public | 277 | 70.1 | 6 | 395 | 1.5 (0.7, 3.4) | -- | -- | 0.06b |
| Private | 88 | 22.3 | 0 | 136 | 0 (0, 2.7)a | Ref | -- | -- |
| Public and private | 6 | 1.5 | 0 | 10 | 0 (0, 38.6)a | -- | -- | 0.50b |
Confidence intervals are 97.5% one-sided assuming a Poisson distribution
Because the referent category had 0 seroconversions, the test for differences in HIV incidence was done by the Z-test
p-value on the chi-squared test using a Mantel-Haenszel method to calculate stratified rate ratios
Includes: own your own house, rent a house/apartment/room
Includes: couch surfing, homeless/shelter, single room occupancy hotel, residential treatment facility, transitional/supportive housing, other
Several disparities in HIV incidence rates were noted. Trans women age 18–24 years had a significantly higher HIV incidence (3.7 per 100 py, 95% CI 1.2–11.6) compared to those 25 and older (1.0 per 100 py, 95% CI 0.4–2.3, p=0.04). Age categories above 25 years were collapsed as they were similar in magnitude. HIV incidence was significantly higher among Latina/x trans women (2.6 per 100 py, 95% CI 1.1–6.1, p=0.03) and trans women of color (2.2 per 100 py, 95% CI 1.1–4.3, p=0.01) compared to white trans women. Trans women who had been incarcerated (2.3 per 100 py, 95% CI 1.1–4.8, p=0.04) and those without health insurance (5.8 per 100 py, 95% CI 1.5–23.2, p=0.02) also had significantly elevated HIV incidence.
Incidence estimates among trans women for comparison are rare. Twenty years ago, HIV incidence was calculated among trans women in San Francisco in a retrospective cohort, arriving at a rate of 7.8 per 100 py (95% CI 4.6–12.3) [8]. The estimate suggests a substantial decline to the current level. Serial cross-sectional surveys of trans women in San Francisco show sustained at high levels over the last several years [4].
We recognize the risk in over-interpreting data based on 8 seroconversions. Small numbers may miss true associations in the population. We also acknowledge potential Hawthorne effects. Multiple risk assessments, HIV testing, and referral to prevention programs such as PrEP are likely to dampen HIV incidence, underestimating true rates in the population. Other studies note challenges in measuring HIV incidence among trans women. The iPrEx trial of PrEP efficacy was able to enroll 339 trans women among 2,499 total participants across 11 sites [9]. The researchers cite difficulties in identifying trans women and determining the preventive effects of PrEP specifically for them. Other studies combine MSM with trans women without being able to separate them for analysis [5]. For example, a recent analysis of HIV incidence among key populations in Bangkok across 10 years was unable to distinguish between MSM and trans women [10]. We concur with these researchers in the strong need for trans-specific longitudinal studies, including benchmark measures of HIV incidence and randomized controlled trials with incident endpoints.
On the other hand, the fact that we were able to find significantly elevated HIV incidence for some groups speaks to the potential magnitude of these effects. Significant correlates of HIV incidence found in our cohort echo concerns emerging from citywide surveillance data [1]. Despite decreases overall, new HIV diagnoses have increased for Black/African Americans and Latina/x people. Prior studies also found significantly higher incidence and prevalence of HIV among trans women of color [4,8]. History of incarceration and lack of health insurance point to structural drivers of continued HIV acquisition among trans women that must be addressed. Structural risks of housing instability, low income and education are risk factors for sex work and incarceration and are more prevalent among trans women of color [11–13]. Instability may also preclude trans women from completing the steps required to enroll in health insurance and establish care in the public health system where biomedical HIV prevention is available. Structural factors disproportionately affecting trans women of color are also tied to poor HIV care outcomes and suboptimal access to HIV prevention use [14,15]. Such risks are exacerbated in our city, which has wide disparities in wealth, housing, and employment opportunities [16]. Perhaps the most disheartening finding is the elevated HIV incidence among young trans women. The nearly four-fold higher incidence among transgender youth predicts a continuing high burden of infection for years to come. In an era of intensifying efforts to get to zero HIV infections by 2030 [17], any incidence rate above 1 per 100 py is a worrisome reminder that the endgame of eliminating new HIV infections may see diminishing returns as the remaining cases occur among our most marginalized communities.
Acknowledgements
This study was supported by the National Institute on Minority and Health Disparities (R01MD010678).
Conflicts of Interest and Source of Funding: This study was supported by the National Institute on Minority and Health Disparities (R01MD010678). The authors declare they have no conflicts of interest.
Footnotes
Protection of Human Subjects: Procedures followed were in accordance with the ethical standards of the Institutional Review Board of the University of California San Francisco and with the Helsinki Declaration of 1975, as revised in 2000. Participants provided written informed consent.
References
- 1.San Francisco Department of Public Health. The HIV Epidemiology Annual Report 2018. Accessed on 13 November, 2019 at: https://www.sfdph.org/dph/comupg/oprograms/HIVepiSec/HIVepiSecReports.asp.
- 2.Satcher Johnson A, Song R, Hall HI. Estimated HIV Incidence, Prevalence, and Undiagnosed Infections in US States and Washington, DC, 2010–2014. J Acquir Immune Defic Syndr 2017;76:116–122. [DOI] [PubMed] [Google Scholar]
- 3.Brookmeyer R Measuring the HIV/AIDS epidemic: approaches and challenges. Epidemiol Rev 2010;32:26–37. [DOI] [PubMed] [Google Scholar]
- 4.Raymond HF, Wilson EC, Packer T, et al. High and Stable Human Immunodeficiency Virus Prevalence Among Transwomen With Low Income Recruited With Respondent-driven Sampling, San Francisco, 2010–2016. Sex Transm Dis 2019; 46:118–124. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Poteat T, Scheim A, Xavier J, et al. Global epidemiology of HIV infection and related syndemics affecting transgender people. J Acquir Immune Defic Syndr 2016;72:S210–19. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Wilson EC, Chen Y-H, Arayasirikul S, et al. The impact of discrimination on the mental health of trans*female youth and the protective effect of parental support. AIDS and Behav 2016;20:2203–2011. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Yan H, Yang H, Zhao J, et al. Long-chain peer referral of men who have sex with men: a novel approach to establish and maintain a cohort to measure HIV incidence, Nanjing, China. J Acquir Immune Defic Syndr 2012;59:177–184. [DOI] [PubMed] [Google Scholar]
- 8.Kellogg TA, McFarland W, Perlman J, et al. HIV incidence among repeat testers at a county hospital, San Francisco, USA. J Acquir Immune Defic Syndr 2001;28:59–64. [DOI] [PubMed] [Google Scholar]
- 9.Deutsch MD, Glidden DV, Sevelius J, et al. HIV pre-exposure prophylaxis in transgender women: A subgroup analysis of the iPrEx trial. Lancet HIV 2015;2: e512–e519. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Holtz TH, Wimonsate W, Mock PA, et al. Why we need pre-exposure prophylaxis: incident HIV and syphilis among men, and transgender women, who have sex with men, Bangkok, Thailand, 2005–2015. Int J STD AIDS 2019;30:430–439. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Wilson EC, Chen YH, Arayasirikul S, et al. (2015). Differential HIV risk for racial/ethnic minority trans*female youths and socioeconomic disparities in housing, residential stability, and education. Am J Public Health 2015;105(Suppl 3):e41–47. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Beltran T, Allen AM, Lin J, Turner C, Ozer EJ, Wilson EC. Intersectional discrimination is associated with housing instability among trans women living in the San Francisco Bay Area. Int J Environ Res Public Health 2019;16(22),pii: E4521. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Adamczyk A This map shows the US cities with the greatest income disparities. Accessed on 14 November 2019 at: https://www.cnbc.com/2019/10/09/this-map-shows-the-us-cities-with-the-greatest-income-inequality.html.
- 14.Wilson EC, Turner C, Arayasirikul S, et al. Housing and income effects on HIV-related health outcomes in the San Francisco Bay Area - findings from the SPNS transwomen of color initiative. AIDS Care 2018;30(11):1356–1359. [DOI] [PubMed] [Google Scholar]
- 15.Liu A, Turner C, Arayasirikul S, et al. (2018). Substantial Gaps in the PrEP Continuum Among Transwomen Compared With MSM in San Francisco. HIVR4P, 2018, Madrid, Spain. [Google Scholar]
- 16.Wilson EC, Garofalo R, Harris RD, et al. (2009). Transgender female youth and sex work: HIV risk and a comparison of life factors related to engagement in sex work. AIDS Behav 2009;13(5):902–913. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 17.Katz I, Jha AK. HIV in the United States: Getting to Zero Transmissions by 2030. JAMA 2019;321:1153–1154. [DOI] [PubMed] [Google Scholar]