Since July 2005, individuals at high risk for HIV-1 acquisition have been recruited for an open cohort study in preparation for a HIV-1 vaccine efficacy trial in a Kenya Medical Research Institute (KEMRI) clinic in Mtwapa town, coastal Kenya. This town, approximately 20 kilometers north of Mombasa, is known for its busy night life and many bars and nightclubs, which are frequented by sex workers . Participants were identified for recruitment into the study by 10–15 trained peer mobilizers who approached individuals through personal networks and at venues where sex workers meet to establish contact with clients . Adults aged 18–49 years were eligible if they met any of the following criteria: HIV-1-negative and reporting any of transactional sex work, a recent sexually transmitted infection (STI), multiple sexual partners, sex with an HIV-1-infected partner, or anal sex during the 3 months before enrolment .
Detailed cohort procedures have been described elsewhere [9, 10]. In brief, during enrollment and monthly follow-up visits, a face-to-face interview using a standardized risk behaviour questionnaire, HIV-1 testing and counseling using rapid point of care antibody tests, risk-reduction counselling, medical history and physical examination were performed. During monthly follow-up visits, participants were re-assessed for HIV-1 acquisition risks, treated for genital symptoms suggestive of STIs, offered hepatitis B vaccination and provided with risk reduction counselling. For this analysis, only MSM were retained.
At each study visit, two rapid antibody test kits (Determine, Abbott Laboratories; Unigold, Trinity Biotech) were used in parallel for HIV-1 testing. Discordant rapid HIV-1 test results were resolved using HIV-1 RNA (Xpert® HIV-1 Qual, Cepheid). Pre- and post-seroconversion samples were tested for HIV-1 RNA using Amplicor Monitor 1.5 (Roche) through 2015, then Xpert® HIV-1 Qual (Cepheid) starting in 2016. Gonococcal infection was diagnosed among participants who reported urethral or rectal symptoms by the detection of Gram-negative, intracellular diplococci consistent with Neisseria gonorrhoeae in urethral or rectal secretions . Prevalent syphilis infection was diagnosed by a positive rapid plasma reagin (RPR, tested annually) titre confirmed by Treponema pallidum haemagglutination assay (TPHA). Incident syphilis was defined as a four-fold increase in RPR titre confirmed by TPHA .
Preparing MSM for PrEP uptake
Between January-June 2017, we offered standardized educational messages to cohort participants on benefits, risks, eligibility and upcoming availability of daily PrEP during individual discussions with clinicians at follow-up visits. In addition, weekly group educational sessions led by counselors were provided to cohort participants who had expressed interest in learning about PrEP. In both individual and group sessions, we educated participants about known predictors of HIV-1 acquisition among MSM in our cohort, including RAI, group sex, any recent unprotected sex, having sex with men only and gonorrhea infection within the past six months . Because younger age (18-24 years) had become an additional independent predictor in our cohort , we explained to individuals in this age group their higher risk. Using these identified risk factors in the CDHRS and additional MoH PrEP eligibility criteria , we designed an individualized PrEP eligibility score sheet based on risks reported during the previous three months to target PrEP counseling (see Additional file 1).
Since June 2017, PrEP has been offered to eligible cohort participants in follow-up. Participants were evaluated for PrEP eligibility as described above. Those who were eligible by either CDHRS or MOH criteria according to the PrEP Eligibility Score Sheet were offered PrEP, and their renal function (i.e. creatinine), hepatitis B surface antigen (HBsAg), and symptoms of acute HIV-1 infection (AHI) assessed according to the Kenyan MoH PrEP guidelines . Participants who tested positive for HBsAg were offered PrEP with close monitoring of liver function while those who tested negative were vaccinated against hepatitis B infection . Participants who had symptoms compatible with AHI, or those meeting specific risk criteria that increased their risk of HIV-1 acquisition (e.g. RAI, or group sex) were tested for HIV-1 RNA (Xpert® HIV-1 Qual, Cepheid) to rule out AHI prior to PrEP initiation. Participants with no contraindication to PrEP were counselled about the risks, benefits and limitations of PrEP, educated about recognizing AHI symptoms, and provided with a 30-day PrEP supply. Individuals who had previously taken PrEP through other organisations were invited to transfer to KEMRI PrEP programme if they so desired.
During monthly follow-up visits, participants not taking PrEP were reassessed for eligibility and offered PrEP if eligible. Participants taking PrEP completed a computer-assisted self-interview to assess PrEP adherence and motivation to continue PrEP and were monitored for adverse effects, offered syndromic STI treatment as clinically indicated, and tested for HIV-1. PrEP adherence and sexual risk reduction counseling were provided prior to PrEP refill. Participants with symptoms or signs compatible with AHI were tested for HIV-1 RNA as described above. Participants who tested HIV-1-positive (either on RNA or rapid antibodies) had PrEP discontinued and were counselled and linked to HIV-1 care and treatment.
CDHRS eligibility: This variable was defined as having any of the following risk factors at any visit, categorized as either yes or no: age 18-24 years, having only male sex partners, RAI, any recent unprotected sex, and group sex. Individuals who had any of these risk factors in the 3 months before screening were considered eligible for PrEP by CDHRS criteria.
MoH eligibility: This variable was defined as having any of the following characteristics per MoH PrEP guidelines  at any visit, categorized as either yes or no: sex with a regular partner of known HIV-1-positive or unknown HIV-1 status in the past week, sex with any partner of known HIV-1-positive or unknown HIV-1 status in the past month, transactional sex (defined as receiving payment for sex with cash, living expenses, or goods) in the past 3 months, sharing needles among intravenous drug users in the past 3 months, sex after alcohol use in the past month, recurrent use of post-exposure prophylaxis (PEP, defined as PEP use more than once in the past 6 months), inconsistent condom use in the past week and recent STI (defined as a positive gram stain of urethral or rectal secretions or a new syphilis diagnosis within 6 months). Individuals who had any of these characteristics were considered eligible for PrEP by MOH criteria.
Other variables evaluated as potential predictors of PrEP uptake included the number of reported sexual partners in the past week; paying for sex with cash, living expenses, or goods in the past 3 months; and demographic data collected at enrollment (e.g., education, religion, marital and employment status).
Data analysis and statistical methods
Historic cohort 2005-2016
Predicting HIV-1 acquisition:
We censored data for each participant at the end of 2016, at the last visit (for those lost to follow-up) or at the last seronegative and HIV-1-RNA-negative visit (for those who acquired HIV-1 infection during follow-up). We obtained total observation time for all participants in the study by adding up separate observation times and expressing these in terms of pre-PrEP person-years. To assess the performance of the MoH criteria to identify MSM at risk of HIV-1 acquisition in the historic cohort (2005-2016), we assigned a score of one point to each characteristic (described above) reported and summed these scores to generate a total MoH score for each participant visit. A total CDHRS score was also calculated for each visit, following published methodology . We assessed sensitivity, specificity and area under the receiver operator characteristic (ROC) curve (AUC) for the MoH and CDHRS eligibility criteria using a non-parametric ROC analysis. We compared the AUC for the CDHRS eligibility score to the AUC for the MoH eligibility score using a test of equality of ROC areas.
PrEP cohort June-December 2017
PrEP eligibility and uptake:
PrEP provision in limited programmes targeting key populations began in the area around January 2017. Because we did not have reliable data on PrEP use from outside programs, we excluded data collected in the period between January–May 2017. PrEP became available to the KEMRI cohort in June 2017. PrEP baseline was defined as the first study visit by a given participant during June–December 2017. PrEP uptake was defined as acceptance of PrEP by an eligible participant. We censored data for each participant at the end of 2017, at the last visit (for those lost to follow-up) or at the last seronegative and HIV-1-RNA-negative visit (for those who acquired HIV-1 infection during follow-up). Nine MSM who had started PrEP through another program were excluded. We calculated the number and proportion of MSM eligible by MoH vs. CDHRS criteria at PrEP baseline and presented the results using a Venn diagram. We then compared the proportion of MSM eligible for PrEP by each criterion at baseline and at the last visit in 2017 using McNemar’s test for paired proportions, to determine consistency of PrEP eligibility over time.
We used descriptive statistics to compare baseline demographic and behavioural characteristics of eligible men who accepted PrEP at baseline to eligible men who did not accept PrEP at first offer. We then used generalized linear modeling with log link Poisson regression and robust error variance to identify factors independently associated with PrEP uptake at baseline. Potential predictors of PrEP uptake significant in bivariable analysis at P ≤ 0.2 were included in multivariable modeling. P values were 2-sided, and significance was set at P ≤ 0.05. Data were cleaned, recoded and analyzed using Stata 15.0 (StataCorp LP, College Station, TX).
The KEMRI Ethics Review Committee approved the study. All participants provided written informed consent.