Variation in estimated viral suppression associated with the definition of viral suppression used

Objective:The proportion of people living with HIV with suppressed viral load is a key indicator of the HIV care continuum. We explored how this proportion varied depending on how it was calculated.Design:Observational cohort study.Methods:We calculated the proportion of the Johns Hopkins HIV Clinical Cohort who were virally suppressed each year, 2010-2018, based on different denominators; thresholds for suppression (≤20, ≤50, ≤200, or ≤400 copies/ml); and strategies for summarizing multiple viral load measurements (we classified persons as suppressed if they had any lab, ≥50% of labs, last lab, or all labs below the threshold). We also calculated 5-year risk of all-cause mortality associated with each classification of viral suppression.Results:Three thousand eleven persons contributed 60 858 viral load values to this analysis. Proportion classified as virally suppressed ranged from 51.8 to 92.5%, depending on the definition used and persons included in the calculation. Requiring more labs below the threshold; using a lower threshold; and assuming persons lost to follow-up were not suppressed (stricter definitions) resulted in a lower proportion estimated to be suppressed. Suppression by stricter definitions were associated with better 5-year survival.Discussion:The proportion suppressed varied greatly as a function of the subset of persons in whom it was calculated, the threshold used for suppression, and the way multiple viral loads per person per year were summarized. Measures of durable viral suppression, and low-level viremia (20-400 copies/ml), should be considered in describing the health of people with HIV, in addition to the standard estimates of suppression.