The efficacy of candidate HIV vaccines in non-human primates and humans: a systematic review

Over the last four decades, Acquired Immunodeficiency Syndrome (AIDS) and Human Immunodeficiency Virus (HIV) have been studied intensively in animal models and humans, hoping to find an effective and safe vaccine to prevent this debilitating disease. However, no vaccine has become available, and the translatability of the animal models to humans has not been assessed systematically. Our systematic and comprehensive review of published HIV trials in non-human primates (NHPs) and humans aimed to answer the question: "What is the concordance of candidate HIV vaccine efficacy in NHPs and humans?". A review protocol was posted on Prospero on 20 December 2023. We searched PubMed and EmBase on 25 and 26 December 2023 with comprehensive search strings, and used a combination of AI-assisted and manual screening (two independent screeners). Data extraction followed a custom-made template, and risk of bias was evaluated following standard tools (both by two independent reviewers). From 50 624 records retrieved by the searches, 216 were included, comprising data from 40 750 human subjects and 3 326 NHPs. Differences between NHP and human studies were observed for e.g. the number and type of vaccines administered, the control conditions and the level of infection, further discussed in the manuscript. Our explorative frequentist random-effects model meta-analysis shows that overall, the NHP and human results were not significantly different (Q=1.75; df=1; p=0.19). However, risk of bias was mainly unclear and confidence intervals were large. To conclude, limited animal-to-human translation cannot be held responsible for the current lack of an effective preventive HIV vaccine, but the results do not support the value of NHP models either. All extracted data are available from a searchable online spreadsheet, to aid those working on HIV vaccine development in more easily retrieving relevant studies.