Objective: To gain insight in the ongoing HIV transmission, we compared sexual risk behaviour pre-HIV and post-HIV seroconversion in 206 MSM participating in the Amsterdam Cohort Studies (1984–2008) before and after the introduction of combination antiretroviral therapy (cART).
Design and methods: MSM completed behavioural questionnaires and were tested for HIV antibodies every 6 months. Trends in anal intercourse and number of sex partners from 4 years before HIV seroconversion until 4 years after diagnosis were analysed with latent class random effects logistic regression models.
Results: The risk of having unprotected anal intercourse (UAI) 1 year after HIV diagnosis decreased significantly when compared with 1 year before diagnosis in both the pre-cART era [difference, 30%; 95% confidence interval (CI), 22–36%] and cART era (difference, 19%; 95% CI, 9–30%). In contrast to a continuing decrease of UAI in the pre-cART era, the probability of UAI in the cART era increased again to preseroconversion levels (61%; 95% CI, 48–74%)) 4 years after diagnosis.
Conclusion: This study provides evidence that recently seroconverted MSM reduce their sexual risk behaviour following HIV diagnosis both in the pre-cART as well as the cART period. However, in the cART period this reduction in sexual risk behaviour is less and returns to pre-cART levels within 4 years. These findings not only confirm the need for early HIV testing but also make it clear that much more effort should go into identifying, counselling, and possibly treating recently seroconverted MSM who have been found to be one of the most important drivers of HIV transmission among MSM.
aDepartment of Research, Cluster of Infectious Diseases, Public Health Service Amsterdam
bDepartment of Clinical Epidemiology, Biostatistics and Bioinformatics, Academic Medical Center, Amsterdam
cCenter for Infectious Disease Control, National Institute for Public Health and the Environment, Bilthoven
dDepartment of Internal Medicine, Division of Infectious Diseases, Tropical Medicine and AIDS (CINIMA), Academic Medical Center, Amsterdam
eJulius Center for Health Services and Primary Care, University Medical Center Utrecht, Utrecht, The Netherlands.
Correspondence to Titia Heijman, MA, Public Health Service Amsterdam, P.O. Box 2200, 1000 CE, Nieuwe Achtergracht 100, 1018WT Amsterdam, The Netherlands. Tel: +31 20 555 5363; fax: +31 20 555 5533; e-mail: firstname.lastname@example.org
Received 7 September, 2011
Revised 7 November, 2011
Accepted 23 November, 2011
Preliminary data of this study was presented as poster at the CROI 2011; 27 March to 3 April 2011; Boston, Massachusetts, USA and orally presented at the AIDS impact Conference 2011; 12–15 September; Santa Fe, California, USA.
In developed countries, MSM receive most of the new HIV diagnoses. Despite a reduced level of infectivity due to effective combination antiretroviral therapy (cART), the HIV-1 incidence within the group of MSM did not decrease, but remained fairly stable or even increased throughout the last decade. [1–4] This ongoing transmission of HIV among MSM has been partly explained by the rise in sexual risk behaviour, especially among potentially discordant sex partners [5–8], after the success of cART established HIV as a chronic disease. In addition, several studies have shown increases in sexually transmitted infections (STIs) among both HIV-negative and HIV-positive MSM [9–11], and new or re-emerging STIs such as lymphogranuloma venereum and hepatitis C are on the rise in HIV-positive MSM. This suggests that a substantial group of HIV-positive MSM who are aware of their diagnosis are not effectively reducing their sexual risk behaviour [12–14]. Some studies have indicated that MSM diagnosed with HIV decrease their sexual risk behaviour immediately after becoming aware of their HIV-positive status [15–17]. However, these studies either did not compare trends in sexual behaviour before and after HIV diagnosis over a long calendar period or relied on qualitative data only. Moreover, the effect of HIV diagnosis on sexual risk behaviour in a time of successful cART may differ from that in the pre-cART era. It is important for prevention to know how HIV diagnosis affects sexual risk behaviour and whether these behaviour patterns have changed since the availability of cART.
To investigate this, we used data from the Amsterdam Cohort Studies (ACS) on HIV infection and AIDS among homosexual men . This cohort provides a unique setting to investigate such research questions longitudinally, as epidemiological, biological, and behavioural data have been collected on an individual basis since 1984. These data enabled us to investigate changes in sexual risk behaviour after HIV diagnosis among MSM who were seronegative at cohort entry and to compare the changes in the pre-cART era with those in the cART era.
Procedures and sample selection
The ACS is an ongoing, open prospective cohort study started in 1984 to investigate the epidemiology, pathogenesis, and prevention of HIV/AIDS. ACS participation is voluntary, and written informed consent is obtained for every participant at intake. The ACS received ethical approval from the Medical Ethical Committee of the Academic Medical Centre Amsterdam, the Netherlands.
Participants visit the Public Health Service of Amsterdam every 3–6 months to complete self-administered questionnaires and to give blood for HIV-1 testing and storage. HIV-1 antibodies have been prospectively detected with consecutive generations of commercially available screening assays (obtained from Abbott Laboratories, Abbott Park, Illinois, USA; Organon International, Oss, the Netherlands, and bioMérieux, Lyon, France) and confirmed by western blot analyses (Genelabs Diagnostics, Singapore). After seroconversion, information on clinical markers (CD4 cell counts and viral load) is collected every 3 months, and behavioural data are collected every 6 months. CD4 cell counts are performed prospectively by cytofluorometry. Until 2000, HIV-1 RNA loads were determined by the NucliSens test (Organon Teknika) which was replaced by a more sensitive DNA test (Organon Teknika).
For the present study, we included all MSM who seroconverted within the ACS between 1984 and 2008. All individuals were HIV-negative at cohort entry, enabling us to use collected behavioural data from preseroconversion to postseroconversion. This resulted in a study population of 206 MSM.
Baseline characteristics included nationality, educational level, and age. Nationality was categorized as Dutch and non-Dutch. Educational level was divided into three categories: low (no education or primary school), middle (secondary school or lower vocational school), and high (higher vocational school or university). Age at HIV diagnosis was calculated as age at the first HIV-positive visit. Sexual behaviour variables referred to sexual behaviour in the 6 months preceding a study visit. Sexual behaviour variables included total number of male sex partners (reporting any type of sexual contact), anal intercourse (both insertive and receptive), and frequency of condom use during anal intercourse on a five-point scale ranging from ‘always’ to ‘never’. We defined unprotected anal intercourse (UAI) as not always having used a condom while engaging in this practise.
Calendar time at first HIV-positive visit was categorized into two periods: the pre-cART era (1984–1995) and the cART era (1996–2008). Number of sex partners was categorized as ‘0–1 partner’ and ‘2 or more’. Individual plasma HIV-1 RNA levels (PVL) were transformed into logarithmic copies per millilitre and presented as median levels. CD4 cell counts were given in cells per microliter. ART-naive visits were defined as visits of participants before the start of any type of antiretroviral treatment (clinical information). The moment of HIV awareness, referred to as HIV diagnosis, was defined as the first cohort visit during which the participants were informed that their status had become HIV-positive.
Characteristics of the seroconverted individuals in each of the two periods were described, and differences between the two periods were tested with the χ2-test for independence for categorical variables and Student's t-test or Mann–Whitney U test for continuous variables using SPSS version 17.0 (SPSS Inc., Chicago, Illinois, USA). To investigate individual changes in risk behaviour before and after HIV diagnosis, the analyses were restricted to the time span from 4 years before until 4 years after HIV diagnosis. Trends in risk behaviour were analysed by a latent class random effects logistic regression model. Every individual had a probability of belonging to one of three classes: consistently safe, that is, individuals without risk behaviour; consistently unsafe, that is, individuals who engaged in high-risk sexual behaviour at least once every 6 months; or switchers, that is, individuals who were inconsistent with respect to risk behaviour over the 6 months periods.
For the switchers, we used a random effects logistic regression model. For both pre-cART and cART era, the intercept and the time trend were allowed to differ per individual via the random effects. Before HIV diagnosis, the average level of risk behaviour was allowed to depend on the time period (pre-cART era versus cART era). Also, average risk behaviour was allowed to differ in the 9 months before the first HIV-positive test, as sexual behaviour shortly before seroconversion might differ from earlier behaviour. After HIV diagnosis, both the average level of risk behaviour and the time trend were allowed to differ in the cART era from the pre-cART era. Furthermore, the average level was allowed to differ in the first 6 months after the first HIV-positive test, as the 6 months in the questionnaire could overlap with the period before awareness of HIV-positive status. Results per individual were transformed to the risk scale via the inverse of the logit function.
These probabilities were combined with the probabilities of belonging to the other classes to obtain an overall individual probability of engaging in specific sexual risk behaviour which were averaged to obtain an overall probability. Trends are presented graphically as probabilities of engaging in a specific sexual behaviour over time. The risk difference between the time before and after HIV diagnosis was calculated as well. We compared the probabilities at 1 year before HIV diagnosis to 1 and 4 years after HIV diagnosis.
A Bayesian approach was used for estimation of the parameters, starting with noninformative priors. Posterior distributions were obtained via Markov Chain Monte Carlo methods, with the OpenBUGS program via the R2WinBUGS interface in the R statistical programming language (R2WinBUGS-package, R) [19,20]. Three chains were generated on the basis of different sets of starting values. Parameter estimates were the medians of the posterior distributions. The range from the 2.5% quartile to the 97.5% quartile was used to quantify the uncertainty in the parameter estimates. This range can be interpreted as a 95% confidence interval (CI) and will be referred to as such.
Of the 206 MSM in the study, 125 (61%) seroconverted for HIV in the pre-cART era and 81 (39%) in the cART era. Participants’ characteristics are shown in Table 1. Seroconverted individuals in the cART era were somewhat younger (mean 33.5 years) than those in the pre-cART era (mean 36.5 years). In both periods, seroconverted individuals were predominately Dutch (93%) and had attained a middle-level or high-level of education (82%). The median CD4 lymphocyte count at the first positive visit was 650 cells/μl [interquartile range (IQR) 500–810)] and the median HIV-1 viral load was 4.5 log copies/ml (IQR 3.9–5.2), and no significant differences between the two periods were observed. Of the 81 seroconverted individuals in the cART era, 66 remained ART-naive up to 4 years after seroconversion.
Trends in risk behaviour before and after the first HIV-positive visit in both the pre-combination antiretroviral therapy and combination antiretroviral therapy eras
Number of sex partners
The median number of sexual partners in the preceding 6 months during both the pre-cART and cART eras decreased significantly from 12 (IQR 4.5–24.4) at 1 year before HIV diagnosis to 6 (IQR 1.6–15.4) at 1 year after HIV diagnosis (P < 0.001). There was no significant difference between the eras in the probability of MSM engaging in any type of sexual contact with two or more different male partners during any 6 months period (Fig. 1a and 2a). Overall, in both eras there was a 92% probability of the MSM having had sex with two or more partners 1 year before HIV diagnosis. One year after HIV diagnosis, this probability decreased significantly by 20% (95% CI 13–28%) to 72% and remained approximately the same up to 4 years.
Trends in anal intercourse are presented in Fig. 1b and 2b. In the pre-cART era, the probability of the MSM engaging in anal intercourse decreased from 88% at 1 year before HIV diagnosis to 68% at 1 year after diagnosis (difference, 20%; 95% CI 14–25%) and remained on this level the following years. In contrast, the probability of their engaging in anal intercourse in the cART era decreased by only 5.3% (95% CI 1.3–10%) from 98.5% at 1 year before diagnosis to 93.2% 1 year after, with a slow decline to 90.5% at 4 years following diagnosis. MSM who seroconverted in the cART period had an 11% (95% CI 7.4–14%) higher probability of practising anal intercourse 1 year before diagnosis than MSM who seroconverted in the pre-cART period. This difference in anal intercourse increased to a maximum of 25% (95% CI 19–31%) 1 year after HIV diagnosis and was 23% (95% CI 12–30%) 4 years after diagnosis.
Unprotected anal intercourse
In (Fig. 1c and 2c, 1d and 2d, 1e and 2e) the pre-cART era, the probability of engaging in UAI decreased from 68% at 1 year before HIV diagnosis to 38% 1 year after diagnosis [a 30% decrease (95% CI 22–36%)] and to 32% (95% CI 27–38%) 4 years after diagnosis. In the cART era, the probability of MSM engaging in UAI also significantly decreased from 72% at 1 year prior to HIV diagnosis to its lowest point of 53% (difference, 20%; 95% CI 9–30%) 1 year after HIV diagnosis. However, by 4 years after diagnosis, the probability increased to 61% (95% CI 48–74%). The risk difference of engaging in UAI in the cART versus the pre-cART era at 1 year before HIV diagnosis was not statistically significant [5% lower pre-cART (95% CI 11% lower to 2% higher)], but after HIV diagnosis the risk difference increased and reached statistical significance 1 year after diagnosis (difference, 14%; 95% CI 4–24%) and increased to 30% 4 years after HIV diagnosis(95% CI 15–42%).
To exclude the possible influence of treatment with cART on the probability of engaging in UAI in the cART period, we conducted a sensitivity analysis including only the visits by ART-naive individuals (N = 208 per 248 visits in the cART period). Results of this analysis were comparable: at 4 years after HIV diagnosis in the cART period, the probability of UAI was 62% (95% CI 44–77%).
Finally, when we looked more specifically at insertive (IUAI) and receptive UAI (RUAI), we found trends similar to UAI, with the overall probabilities of RUAI somewhat higher than the probabilities of IUAI (Fig. 1d and 2d, 1e and 2e). Interestingly, the risk difference between the pre-cART and cART eras at 1 year before HIV diagnosis was larger for IUAI (20%) than for RUAI (10%), although both differences were statistically significant.
Among HIV-negative MSM who seroconverted for HIV, and became aware of their positive HIV status, we found after HIV awareness a significant reduction in both UAI and number of sex partners. This agrees with findings from other studies [15–17]. However, we found that this reduction in UAI was significantly smaller and appeared more temporary during the cART era compared with the pre-cART era. Importantly, the decreasing trend in UAI after HIV diagnosis lasted only 1 year and then it tended to increase again; after 4 years, the probability of engaging in UAI among seroconverted individuals in the cART era reached almost the same levels as it had 1 year before the HIV diagnosis. As most of these men were still untreated, it is likely that they might contribute to the ongoing HIV transmission among MSM. One could argue that the observed difference in age at the moment of diagnosis between seroconverted individuals in the pre-cART versus the cART eras partly explains our findings. Although there has been some evidence for higher risk taking in young MSM (less than 25 years of age) , other studies did not find such a difference . Moreover, the age difference of 3 years in our study is relatively small, and trends in sexual risk behaviour and HIV incidence in our cohort  were similar for different ages. Therefore, we think this age difference will not explain the changes in probability of UAI before and after HIV diagnosis in our study.
The long-term rebound of sexual risk behaviour among HIV-positive MSM during the cART era has important implications for HIV transmission to the negative population. However, in the Netherlands, the uptake of HIV testing and its promotion significantly increased after the introduction of cART. This opened opportunities to engage in risk reduction strategies other than condom use, especially serosorting and viral sorting . It could be that the increased reported rates of UAI are partially due to UAI with seroconcordant partners. These strategies might limit forward HIV transmission in varying degrees of efficiency [24–26], although their effectiveness is still under discussion. Moreover, they will not sufficiently protect HIV-positive MSM from health threats, like co-infection with hepatitis C virus, lymphogranuloma venerum, and other STIs . Unfortunately, such strategies have been measured in the cohort only in recent years and, therefore, could not be incorporated in this study. Future research should take these strategies into account when investigating UAI in relation to HIV transmission.
Interestingly, we found that MSM in the cART era reported, next to RUAI, a higher level of IUAI before HIV seroconversion compared with the pre-cART era. We cannot readily explain this change in risk pattern nor can we estimate its impact on HIV acquisition.
The decrease in UAI directly following HIV diagnosis underlines the usefulness of early detection of HIV infection. The impact of sexual risk reduction will be most effective when MSM are diagnosed during the primary infection when PVL levels are still high. So, early diagnosis is not only important in preventing disease progression  but it will also give MSM the opportunity to reduce their forward transmission risk during the highly infectious period of primary infection .
Shortly after receiving the diagnosis of HIV, MSM begin to shift their thoughts and behaviour toward the implications of their new status. After HIV diagnosis, men have to shift from using a condom to protect only themselves to also protect the sexual partner. We believe that MSM immediately after HIV diagnosis can benefit from supportive interventions aimed at assisting them in making optimal choices for a healthy sex life for themselves and their future partners . This is now even more important, as the rebound of sexual risk behaviour seems eminent in the cART era and the period of condom use during anal sex after HIV diagnosis is shorter now than that in the pre-cART period.
In conclusion, not only have HIV testing and treatment possibilities in MSM evolved but also the dynamics of sexual risk behaviour immediately after diagnosis. In this cohort of MSM who seroconverted for HIV, risk reduction after HIV diagnosis is less than that in the pre-cART era and appears only temporary. To reduce the ongoing HIV transmission among MSM, preventive interventions should not only be directed at HIV-positive MSM generally, but they should especially focus on behavioural support of men shortly after HIV diagnosis. Moreover, given the reduction in sexual risk behaviour after HIV diagnosis combined with higher viral load levels early in the infection, increased efforts to identify primary HIV infection are essential in preventing forward transmission.
T.H. interpreted the data, and wrote the draft manuscript; R.B.G. did the statistical analyses and helped with interpretation of the data. U.D., R.A.C., and M.P. contributed to the idea and interpretation of data. I.S. supervised the overall study, and contributed to the analyses and interpretation of data. All authors contributed to the final version of this manuscript.
The authors would like to thank all the personnel of the ACS (currently, Marc van Wijk and Marjolein Martens) for their contribution in data collection, Iralice Jansen for data preparation, and Sally Ebeling for editing the final manuscript. Furthermore, the authors gratefully acknowledge all the study participants for their cooperation and participation which made this study possible.
All the authors have read and approved the text as submitted to AIDS.
This study is funded by the research and development fund of the Public Health Service of Amsterdam, the Netherlands.
The Amsterdam Cohort Studies on HIV infection and AIDS, a collaboration between the Public Health Service of Amsterdam, the Academic Medical Center of the University of Amsterdam, Sanquin Blood Supply Foundation and the University Medical Center Utrecht, are part of the Netherlands HIV Monitoring Foundation and financially supported by the Netherlands National Institute for Public Health and the Environment. (http://www.amsterdamcohortstudies.org/).
Conflicts of interest
All authors declare that there are no conflicts of interest.
1. van Griensven F, de Lind van Wijngaarden JW, Baral S, Grulich A. The global epidemic of HIV infection among men who have sex with men
. Curr Opin HIV AIDS
2. Sullivan PS, Hamouda O, Delpech V, Geduld JE, Prejean J, Semaille C, et al. Reemergence of the HIV epidemic among men who have sex with men in North America, Western Europe, and Australia, 1996–2005
. Ann Epidemiol
3. Scheer S, Kellogg T, Klausner JD, Schwarcz S, Colfax G, Bernstein K, et al. HIV is hyperendemic among men who have sex with men in San Francisco: 10-year trends in HIV incidence, HIV prevalence, sexually transmitted infections and sexual risk behaviour
. Sex Transm Infect
4. Jansen IA, Geskus RB, Davidovich U, Jurriaans S, Coutinho RA, Prins M, et al. Ongoing HIV-1 transmission among men who have sex with men in Amsterdam: a 25-year prospective cohort study
5. Bezemer D, de Wolf F, Boerlijst MC, van Sighem A, Hollingsworth TD, Prins M, et al. A resurgent HIV-1 epidemic among men who have sex with men in the era of potent antiretroviral therapy
6. Dilley JW, Woods WJ, McFarland W. Are advances in treatment changing views about high-risk sex?
. N Engl J Med
7. Stolte IG, Dukers NH, Geskus RB, Coutinho RA, de Wit JB. Homosexual men change to risky sex when perceiving less threat of HIV/AIDS since availability of highly active antiretroviral therapy: a longitudinal study
8. Zablotska IB, Crawford J, Imrie J, Prestage G, Jin F, Grulich A, et al. Increases in unprotected anal intercourse with serodiscordant casual partners among HIV-negative gay men in Sydney
. AIDS Behav
9. Chen SY, Gibson S, Katz MH, Klausner JD, Dilley JW, Schwarcz SK, et al. Continuing increases in sexual risk behavior and sexually transmitted diseases among men who have sex with men: San Francisco, California, 1999–2001, USA
. Am J Public Health
10. Rietmeijer CA, Patnaik JL, Judson FN, Douglas JM Jr. Increases in gonorrhea and sexual risk behaviors among men who have sex with men: a 12-year trend analysis at the Denver Metro Health Clinic
. Sex Transm Dis
11. Stolte IG, Dukers NH, de Wit JB, Fennema JS, Coutinho RA. Increase in sexually transmitted infections among homosexual men in Amsterdam in relation to HAART
. Sex Transm Infect
12. Crepaz N, Marks G, Liau A, Mullins MM, Aupont LW, Marshall KJ, et al. Prevalence of unprotected anal intercourse among HIV-diagnosed MSM in the United States: a meta-analysis
13. Urbanus AT, van de Laar TJ, Stolte IG, Schinkel J, Heijman T, Coutinho RA, et al. Hepatitis C virus infections among HIV-infected men who have sex with men: an expanding epidemic
14. de Vries HJ, Van der Bij AK, Fennema JS, Smit C, de Wolf F, Prins M, et al. Lymphogranuloma venereum proctitis in men who have sex with men is associated with anal enema use and high-risk behavior
. Sex Transm Dis
15. Rachinger A, Stolte IG, van de Ven TD, Burger JA, Prins M, Schuitemaker H, et al. Absence of HIV-1 superinfection 1 year after infection between 1985 and 1997 coincides with a reduction in sexual risk behavior in the seroincident Amsterdam cohort of homosexual men
. Clin Infect Dis
16. Fox J, White PJ, Macdonald N, Weber J, McClure M, Fidler S, et al. Reductions in HIV transmission risk behaviour following diagnosis of primary HIV infection: a cohort of high-risk men who have sex with men
. HIV Med
17. van Kesteren NM, Hospers HJ, van Empelen P, van Breukelen G, Kok G. Sexual decision-making in HIV-positive men who have sex with men: how moral concerns and sexual motives guide intended condom use with steady and casual sex partners
. Arch Sex Behav
18. van Griensven GJ, de Vroome EM, Goudsmit J, Coutinho RA. Changes in sexual behaviour and the fall in incidence of HIV infection among homosexual men
19. R Development Core team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; 2011.
20. Sturtz S, Ligges U, Gelman A. R2WinBUGS: a package for running WinBUGS from R
. J Stat Software
21. Crepaz N, Marks G, Mansergh G, Murphy S, Miller LC, Appleby PR. Age-related risk for HIV infection in men who have sex with men: examination of behavioral, relationship, and serostatus variables
. AIDS Educ Prev
22. Dufour A, Alary M, Otis J, Noel R, Remis RS, Masse B, et al. Correlates of risky behaviors among young and older men having sexual relations with men in Montreal, Quebec, Canada. Omega Study Group
. J Acquir Immune Defic Syndr
23. Gorbach PM, Weiss RE, Jeffries R, Javanbakht M, Drumright LN, Daar ES, et al. Behaviors of recently HIV-infected men who have sex with men in the year postdiagnosis: effects of drug use and partner types
. J Acquir Immune Defic Syndr
24. Brennan DJ, Welles SL, Miner MH, Ross MW, Rosser BR. HIV treatment optimism and unsafe anal intercourse among HIV-positive men who have sex with men: findings from the positive connections study
. AIDS Educ Prev
25. Brewer DD, Golden MR, Handsfield HH. Unsafe sexual behavior and correlates of risk in a probability sample of men who have sex with men in the era of highly active antiretroviral therapy
. Sex Transm Dis
26. Vallabhaneni S, Loeb L, Bragg L, McConnell J, Hartogensis W, Grant R, et al
. Seroadaptive tactics adopted by HIV+ MSM can contribute to profound and sustained reductions in HIV transmission risk following HIV diagnosis
. Conference on Retroviral and Opportunistic Infections (CROI)
2011; Boston, Massachusetts, USA. abstract 1038.
27. Truong HM, Kellogg T, Klausner JD, Katz MH, Dilley J, Knapper K, et al. Increases in sexually transmitted infections and sexual risk behaviour without a concurrent increase in HIV incidence among men who have sex with men in San Francisco: a suggestion of HIV serosorting?
. Sex Transm Infect
28. Grijsen M, Steingrover R, Wit F, de Wolf F, Lange J, Verbon A, et al
. An RCT comparing no treatment with 24 or 60 weeks of temporary ART during primary HIV infection
. In: Conference on Retroviral and Opportunistic Infections (CROI)
2011; Boston, Massachusetts, USA. abstract 161.
29. Cambiano V, Rodger AJ, Phillips AN. ’Test-and-treat’: the end of the HIV epidemic?
. Curr Opin Infect Dis
30. Elford J, Bolding G, Davis M, Sherr L, Hart G. Trends in sexual behaviour among London homosexual men 1998–2003: implications for HIV prevention and sexual health promotion
. Sex Transm Infect