Context: Patients with HIV may have increased risk of atherosclerotic cardiovascular disease owing to multiple biological mechanisms. Objective: To evaluate the evidence for subclinical atherosclerosis among patients with HIV. Design: Systematic review of observational studies. Data sources: We searched Medline, Cochrane DSR, ACP Journal Club, DARE, CMR, HTA, NHSEED, Embase and the Cochrane Controlled Trials Register for studies published before November 2008. Study selection: Eligible studies were cross-sectional, cohort, or case-control studies reporting carotid ultrasound intima-media thickness (CIMT), focal plaque incidence, or coronary artery calcium (CAC), as determined by HIV positivity or protease inhibitor (PI) exposure. Data extraction: Two independent reviewers abstracted data using a standardised form. The primary outcome was weighted mean difference (WMD) for CIMT comparing HIV positive versus negative patients. Other outcomes included WMD by PI exposure and the odds ratio (OR) for a focal carotid plaque or CAC. Data from six cross-sectional, seven case-control and 13 cohort studies were included, involving 5456 HIV positive and 3600 HIV negative patients. Results: The weighted mean CIMT was 0.04 mm thicker among patients with HIV than among non-HIV patients (95% CI 0.02 to 0.06; p<0.001). HIV positivity was not associated with carotid plaque or CAC. PI exposure did not significantly affect CIMT, carotid plaque, or CAC. There was evidence of publication bias and stratified analysis and meta-regression showed outcomes were influenced by study design, age, gender and smoking. However, HIV positivity slightly increased CIMT even after sensitivity analyses. Conclusions: HIV infection and PI exposure are not strong independent risk factors for subclinical atherosclerosis. Confounding may contribute to overestimation of the risk associated with HIV and PI exposure.