INTRODUCTION: Understanding factors influencing bone accrual in childhood and bone loss in early adulthood may inform approaches to improve skeletal health and reduce fracture risk in adulthood. This thesis aimed to determine the effect of chronic HIV infection on bone architecture, including bone density, bone size, and bone strength in children, adolescents and premenopausal women. METHODS: I designed the peripheral quantitative computed tomography (pQCT) protocols, performed pQCT scans, trained a team of pQCT radiographers and worked with them to collect pQCT data in Zimbabwean children and adolescents living with HIV (CWH) and children and adolescents living without HIV (CWOH). Participants included 8 – 16 year old children from Harare, recruited from clinics at Parirenyatwa and Sally Mugabe hospitals and schools within the same catchment area. In a second previously conducted study premenopausal women living with HIV (WLWH) and without (WLWOH), aged 18 years or older, had been recruited from clinics in Soweto, Johannesburg. In both studies questionnaires captured clinical and demographic information. Outcomes were tibial pQCT measured 4% trabecular and 38% cortical volumetric bone mineral density (vBMD), 4% and 38% cross-sectional area (CSA), and stress-strain index (SSI) in both children and women. Additional outcomes included radial pQCT measured 4% trabecular and 66% cortical vBMD, 4% and 66% CSA and SSI in women only. I graded each of the pQCT scan slices for image quality. I compared differences in means by HIV status in both children and women. In children, I used linear regression, incorporating an interaction term for pubertal status to test for differences in mean and in change (∆) in bone outcomes over one year. I further used structural equation modelling (SEM) to evaluate whether impaired height mediates the effect of HIV on ∆ bone outcomes per year. In women, I used piecewise regression analysis to estimate trajectories of ∆ bone outcomes between women with and without HIV, exploring the change in trajectory upon starting anti-retroviral therapy (ART). RESULTS: In total, 247 (149 (60%)) WLWH and 609 children (303 CWH; 50% female) were included in these studies. More CWH were pre-pubertal (i.e. Tanner 1; 41% vs. 23%) than CWOH, yet they were similar in age. Median age at ART initiation was 4 (IQR 2–7) years, whilst median ART duration was 8 (IQR 6–10) years. Male and female CWH in later puberty had lower trabecular vBMD, CSA (at 4% and 38%) and SSI than those without HIV at baseline and also at follow-up, whilst cortical density was similar. Puberty modified the effect of HIV on pQCT bone outcomes. Height mediated the effect of HIV on ∆bone outcomes in female CWH but not in males. WLWH had lower vBMD than women living without HIV at both the radial and tibial trabecular rich sites but not at the cortical sites. The longitudinal analysis showed that WLWH who were initiated on ART lost total vBMD at the distal tibia but gained total vBMD at the distal radius, over the 24 months, suggesting that South African women who are in their thirties may still be gaining bone at the radius while bone accrual at the tibia is complete. Furthermore, the skeletal response to HIV infection and ART initiation may vary depending on bone compartment (i.e., trabecular or cortical bone) and skeletal site (i.e. weight-bearing tibia or non-weight-bearing radius). This compartment and site-specificity may suggest that declines in bone in WLWH is more likely due to ART initiation than due to HIV infection. CONCLUSIONS: Despite long-term ART, I identified deficits in bone density, size and predicted strength in children, adolescents and women living with HIV in Zimbabwe and South Africa. These findings are concerning as this may translate into higher fracture risk later in life. With the growing numbers people on ART, Southern Africa might be expected to see an increase in fracture incidence and it is a question of concern, whether the healthcare services are ill prepared for that.