Does participation in an HIV vaccine efficacy trial affect risk behaviour in South Africa?

BACKGROUND: Increased sexual risk behaviour in participants enrolled in HIV prevention trials has been a concern. The HVTN 503/Phambili study, a phase 2B study of the Merck Ad-5 multiclade HIV vaccine in South Africa, suspended enrollment and vaccinations following the results of the Step study. Participants were notified of their treatment allocation and continue to be followed. We investigated changes in risk behaviour over time and assessed the impact of study unblinding. METHODS: 801 participants were enrolled. Risk behaviours were assessed with an interviewer-administered questionnaire at 6-month intervals. We assessed change from enrolment to the first 6-month assessment pre-unblinding and between enrolment and at least 6 months post-unblinding on all participants with comparable data. A one-time unblinding risk perception questionnaire was administered post-unblinding. RESULTS: A decrease in participants reporting unprotected sex was observed in both measured time periods for men and women, with no differences by treatment arm. At 6 months (pre-unblinding), 29.6% of men and 35.8% of women reported changing from unprotected to protected sex (p<0.0001 for each). Men (22%) were more likely than women (14%) to report behaviour change after unblinding (p=0.009). Post-enrolment, 142 (45%) of 313 previously uncircumcised men underwent medical circumcision. 663 participants completed the unblinding questionnaire. More vaccine (24.6%) as compared to placebo recipients (12.0%) agreed that they were more likely to get HIV than most people (p<0.0001), and attributed this to receiving the vaccine. CONCLUSION: We did not find evidence of risk compensation during this clinical trial. Some risk behaviour reductions including male circumcision were noted irrespective of treatment allocation.

Response definition criteria for ELISPOT assays revisited.

No consensus has been reached on how to determine if an immune response has been detected based on raw data from an ELISPOT assay. The goal of this paper is to enable investigators to understand and readily implement currently available methods for response determination. We describe empirical and statistical approaches, identifying the strengths and limitations of each approach to allow readers to rationally select and apply a scientifically sound method appropriate to their specific laboratory setting. Five representative approaches were applied to data sets from the CIMT Immunoguiding Program and the response detection and false positive rates were compared. Simulation studies were also performed to compare empirical and statistical approaches. Based on these, we recommend the use of a non-parametric statistical test. Further, we recommend that six medium control wells or four wells each for both medium control and experimental conditions be performed to increase the sensitivity in detecting a response, that replicates with large variation in spot counts be filtered out, and that positive responses arising from experimental spot counts below the estimated limit of detection be interpreted with caution. Moreover, a web-based user interface was developed to allow easy access to the recommended statistical methods. This interface allows the user to upload data from an ELISPOT assay and obtain an output file of the binary responses.