Study population and setting
Persons with HIV (PWH) and laboratory-confirmed SARS-CoV-2 from 36 institutions (21 states and 3 international locations) were consecutively enrolled into this cohort study from April 1 to July 1, 2020. The study included 286 PWH; mean age was 51.4 years, 26% were female, 78% of patients had HIV for more than 5 years, mean CD4+ T cell count was 531 cells/mm^3, and 89% had suppressed HIV virus to undetectable levels, indicating successful ongoing treatment. The study explored clinical characteristics associated with severe COVID-19 outcomes (composite of ICU admission, mechanical ventilation, or death) or hospitalization.
Summary of Main Findings
Among the 286 PWH included, 57% (n=164) were hospitalized and 18% (n=50) had a severe clinical outcome. After backward selection (p-value for inclusion <0.2), older age, CD4 cell count <200, chronic kidney disease, chronic lung disease, and presence of at least 3 additional comorbidities were associated with hospitalization; older age, CD4 cell count <200, hypertension, chronic lung disease, and presence of more at least 3 additional comorbidities were associated with severe clinical outcome. HIV viremia, HIV treatment regimen, race/ethnicity, and sex were not associated with either outcome.
This study aggregated data from multiple clinical sites and includes a population broadly representative of PWH in the US. The analysis used generalized estimating equations to adjust for within-region differences.
The study does not provide detail on the selection process within participating sites, but given that more than half of the sample was hospitalized with COVID-19, it appears to substantially overrepresent those with severe or critical disease. Model-based variable selection was used and the sample size was relatively small in relation to the number of variables that were controlled for, which can lead to overfitting and inaccurate estimates of association. Patients received a variety of therapies for COVID-19, but this was not adjusted for in the analysis, and considering receipt of therapy is related to baseline factors and clinical outcomes, this would likely result in residual confounding. Missing data were excluded, and the paper did not describe the presence of data entry and quality control protocols (beyond duplicates) to ensure the validity of the data from multiple sites. The study lacks a control group to enable the comparison of COVID-19 outcomes between PWH and persons without HIV.
This is one of the largest studies to characterize the influence of HIV-related factors on COVID-19 clinical outcomes.
This review was posted on: 29 September 2020