OBJECTIVES: Health-related quality of life (HRQL) is modulated by the type and frequency of symptoms experienced in daily life. Such a relationship is likely determined in part by treatment regimen and health care system. We studied how symptoms reported by HIV patients under PI- vs. NNTI-based HAART relate to different HRQL dimensions on the PROQOL-HIV questionnaire. METHODS: N=424 HAART patients (41±10 yrs., 36% females, 8 countries), without comorbidity, completed a 31-symptom HIV checklist (presence/absence) and the PROQOL-HIV questionnaire (43 Likert items, 8 dimensions). Hierarchical cluster analysis uncovered the structure of symptoms experienced in patients under PI (N=242) or NNTI (N=182) regimen. Canonical correlation analysis (CCA) showed the relationships between symptoms and HRQL dimensions, and between-country variations. RESULTS: A common group of symptoms related to body fat (lipodystrophy) and weight changes was shared by patients across treatment regimens. However, frequency of some side-effects–sleep disturbance, headache, diarrhea, nausea, fatigue and pain–were more frequent with PI regimens. The CCA showed that a) body fat change, sleep disturbance, skin problems and abdominal pain symptoms were related to the ‘body change’ and ‘physical health and symptoms’ HRQL dimensions (canonical correlation, 0.78), and b) symptoms concerning physical appearance were related to the ‘stigma’ dimension, whereas other side-effects did not. It also replicated a common finding whereby patients from Western countries (France, Australia, USA) tend to report better HRQL compared to Chinese and Cambodian patients, the latter also experiencing more symptoms due to IP-based treatment. CONCLUSIONS: Treatment-related symptoms aligned with PROQOL-HIV dimensions and known differences between countries. The use of CCA as an exploratory multiple endpoints model helped to unravel complex relationships between symptoms and HRQL facets. The choice of treatment strategies should rely not only on symptoms experience, but also account for their relations to HRQL in light of varying access to care.