Whereas penile vaginal intercourse (VI) is thought to be the dominant mode of HIV transmission in sub-Saharan Africa, cross-sectional studies in the region indicate the preponderance of heterosexual anal intercourse (AI) among high activity groups. A dynamic, heterosexual core group model with risk of infection through both vaginal and anal pathways is formulated and comprehensively analysed. The model is coupled to a general population model and fitted to HIV prevalence data for Zimbabwe in order to explore the parameter space related to heterosexual AI. The model fit supports a core group size ranging between 5-20% and exposure risk to AI in excess of 50%. The control effort quantified by the reproductive number (RA) at commencement of the epidemic corresponds to R0=4.40. With the contribution of heterosexual AI to Zimbabwe's (and that of the sub-Saharan African region) epidemic unknown, the study apportioned infections between the two infection pathways. New infections due to VI ranged from 2-4.5% and 0.5-2.7% from heterosexual AI. The study estimates infection probabilities ranging from 0.15 to 0.35 for both receptive and infective AI. By quantifying the incidence due to VI and AI risks, we put emphasis on the necessity for targeted interventions. To project the potential impact of heterosexual AI in high HIV prevalence settings, we raised the core group size to 20% in the year 2010 coinciding with reported heterosexual AI prevalence outside of Africa and allowed the proportion of infection risk associated with AI to vary. Prevalence and incidence projections were made up to the year 2020 starting from a baseline value of zero exposure to AI transmission risk per sex act, progressively increasing exposure to 50% and 70% respectively. A 50% exposure to AI would result in HIV prevalence scaling up by 23% from the baseline values in year 2020. Increasing exposure to 70% was projected to increase HIV prevalence by 38% in year 2020. The HIV infection risks associated with AI are recognised and inform HIV policy for men who have sex with men, yet the same risks are ignored in HIV intervention programmes for heterosexuals. This study highlights the potential danger of increasing prevalence of heterosexual AI in settings with high HIV prevalence. Evolving and globally cross pollinating sexual behaviors compel for dovetailing HIV policy making with sexology.