BACKGROUND: Poor retention in care and non-adherence to antiretroviral therapy (ART) continue to undermine the success of HIV treatment and care programmes across the world. There is a growing recognition that multifaceted interventions - application of two or more adherence-enhancing strategies - may be useful to improve ART adherence and retention in care among people living with HIV/AIDS. Empirical evidence shows that multifaceted interventions produce better results than interventions based on a singular perspective. Nevertheless, the bundle of mechanisms by which multifaceted interventions promote ART adherence are poorly understood. In this paper, we reviewed theories on ART adherence to identify candidate/potential mechanisms by which the adherence club intervention works. METHODS: We searched five electronic databases (PubMed, EBSCOhost, CINAHL, PsycARTICLES and Google Scholar) using Medical Subject Headings (MeSH) terms. A manual search of citations from the reference list of the studies identified from the electronic databases was also done. Twenty-six articles that adopted a theory-guided inquiry of antiretroviral adherence behaviour were included for the review. Eleven cognitive and behavioural theories underpinning these studies were explored. We examined each theory for possible 'generative causality' using the realist evaluation heuristic (Context-Mechanism-Outcome) configuration, then, we selected candidate mechanisms thematically. RESULTS: We identified three major sets of theories: Information-Motivation-Behaviour, Social Action Theory and Health Behaviour Model, which explain ART adherence. Although they show potential in explaining adherence bebahiours, they fall short in explaining exactly why and how the various elements they outline combine to explain positive or negative outcomes. Candidate mechanisms indentified were motivation, self-efficacy, perceived social support, empowerment, perceived threat, perceived benefits and perceived barriers. Although these candidate mechanisms have been distilled from theories employed to explore adherence to ART in various studies, the theories by themselves do not provide an explanatory model of adherence based on the realist logic. CONCLUSIONS: The identified theories and candidate mechanisms offer possible generative mechanisms to explain how and why patients adhere (or not) to antiretroviral therapy. The study provides crucial insights to understanding how and why multifaceted adherence-enhancing interventions work (or not). These findings have implications for eliciting programme theories of group-based adherence interventions such as the adherence club intervention.