Since 1996, widespread availability of combination antiretroviral therapy (ART) has significantly improved survival of HIV-infected persons in industrialized countries. This has prompted researchers in Europe and North America to investigate whether mortality among HIV-infected persons receiving ART might reach levels similar to those in the general population. In this paper by Lewden et al., more than 80 000 patients from 31 European countries are included in an analysis to estimate crude ART patient mortality rates for persons 518 years of age, who initiated ART during 1998–2008, had a known gender and date of birth, known baseline CD4þ T-cell (CD4) count, and 51 follow-up visit. The standardized mortality ratio (SMR) is used to compare mortality between ART patients and the general population. The SMR is the ratio of the observed number of deaths in the ART cohorts to the expected number of deaths if ART cohorts had experienced general population country-, calendar-year-, gender-, and age-specific mortality. SMRs were estimated using random effects Poisson models, adjusted for age, gender, HIV transmission group and history of AIDS at ART start. The overall ART patient mortality rate of 1.2 per 100 person years (PY) is similar to that recently reported for 13 cohorts from nine countries in Europe and North America (0.95/100 PY for the period 2001– 2009) and 23 cohorts from 10 European countries, Australia and Canada (0.86/100 PY for the period 2004–2006), supporting previous estimates that average annual risk of death for ART patients has declined since 1996–2001, when estimates of about 3/100 PY were reported. The authors were specifically interested in assessing whether ART patient mortality could reach levels similar to those experienced in the general population, if ART patients achieved and maintained CD4 counts 5500/mm. This article builds on previous work by Lewden et al. in 2007, which reported that ART patients in a French cohort, who maintained CD4 counts 5500/mm, had mortality rates similar to those of the general population 6 years after ART start. In this analysis, only certain subgroups of patients who achieved CD4 counts 5500/mm reached mortality rates similar to those observed in the general population. First, among all male ART patients, who achieved and maintained a CD4 count 5500/mm for 53 years, mortality was similar to that of males in the general population [SMR 1.0, 95% confidence interval (CI) 0.8–1.4]. In contrast, mortality among all female ART patients, who achieved a CD4 count 5500/mm, was always higher than that of females in the general population, even after 5 years with a CD4 count 5500/mm. Higher SMRs among females compared with males has been documented in other studies. Authors suggest that differences in the prevalence of low socio-economic status (SES) and/ or smoking between HIV-infected and uninfected persons may be more pronounced for females than males and this might explain the higher SMRs for females. Adjustment of SMRs for SES indicators and smoking burden, variables not available to authors for this analysis, would be needed to further investigate SMR differences by sex. Secondly, among the subgroup of non-injection drug users (non-IDUs) with a CD4 count 5500/mm, males immediately on achieving this CD4 threshold (SMR 0.9, 95% CI 0.7–1.2), and females after 3 years in this CD4 stratum (SMR 1.1, 95% CI 0.7–1.7), had similar mortality rates to those in the Published by Oxford University Press on behalf of the International Epidemiological Association 2012. International Journal of Epidemiology 2012;1–2