Background:  Numerous definitions of elite control (EC) exist in the literature varying on follow-up time and viral load restrictions. An evaluation of EC definitions in a large cohort is needed to provide insight on how best to define EC status.

Methods:  Using CASCADE data from 25,967 individuals with well-estimated dates of HIV seroconversion (SC), we considered 2 leading definitions of EC while ART-naïve and AIDS free: HIV-1 infected for ≥1 year, ≥3 consecutive viral load measurements <75 copies/mL, and all previous viral load <1000 copies/mL (definition A) and HIV-1 infected for ≥10 years, with ≥2 viral load measurements, ≥90% of which had to be <400 copies/mL (definition B). For each definition, individuals classified as EC were assumed EC since SC. We estimated proportion EC of all eligible, time to EC status loss, % time since SC spent as EC, overall ART-naïve CD4 slope (using linear mixed models on the square-root scale), and % EC with ≥1 CD4 <500 cells/mm³ while ART-naïve. To simulate what happens in practice whereby individuals present at clinics without complete past viral load history, we defined EC gold standard to be those with all viral load undetectable for ≥5 years with ≥2 viral loads having detection limit ≤200 copies/mL. We then sampled 100 random periods ≥1 year with ≥3 viral load measures from each individual with ≥5 years follow up. Finally we estimated proportion of EC fulfilling definitions A and B and proportion of our EC gold standard classified as EC within each definition A and B.

Results: Definition A: Of 17,582 HIV-1 infected ≥1 year with ≥3 viral load measurements, 98 (0.56%, 95%CI 0.45 to 0.67) were EC; 59 lost EC after median (95%CI) 8.7 (6.9 to 13.7) years. Median (IQR) % time spent as EC was 90% (66 to 100) with mean (95%CI) CD4 slope 0.06 (0.00 to 0.12) √CD4/year; 45% of EC had ≥1 CD4 <500 cells/mm³. From the simulation we estimated 0.52% (0.48 to 0.55) to be EC of whom 14.3% (12.7 to 16.0) were EC gold standard. Definition B: Of 6739 HIV-infected ≥10 years with ≥2 viral load measurements, 49 (0.72%, 0.52 to 0.93) were EC; 17 lost EC after median 22.6 (15.2,-) years. Median (IQR) % time spent as EC was 100% (92 to 100) with mean CD4 slope –0.03 (–0.08 to 0.02) √CD4/year; 53% of EC had ≥1 CD4 <500 cells/mm³. From the simulation we estimated 1.09% (1.02 to 1.16) to be EC of whom 16.1% (14.3 to 17.9) were EC gold standard. Of 120 individuals classified as EC by either definition, 27 (23%) were classified as such by both.

Conclusions:  Current definitions of EC provide an indefinite characterization of elite status.