Skip to main content

Outcomes of COVID-19 related hospitalization among people with HIV in the ISARIC WHO Clinical Characterization Protocol (UK): a prospective observational study

Our take —

This was a prospective cohort study that included 47592 participants who were hospitalized between Jan 18 and June 4,2020 with COVID-19 in 207 hospitals in London, Wales and Scotland. 122 of the total participants were HIV positive (PWH). PWH had significantly higher mortality at 28 days post-hospitalization compared to HIV negative individuals after adjusting for age, gender, ethnicity, time of symptoms onset, comorbidities and disease severity at presentation (adjusted HR 2.87). The participants who were excluded in the study due to missing HIV status had similar demographic and disease severity status to those with negative HIV status, therefore, their exclusion likely did not bias the results. However, the study did not include information regarding CD4 cell count or HIV viral load that could explain the higher mortality among PWH. Additionally, all explanatory variables were modeled as binary which could have resulted in inadequate adjustment due to residual confounding.

Study design

Retrospective Cohort

Study population and setting

This is a prospective cohort study including 47592 adult participants (47470 were HIV negative, 122 were HIV positive) who were hospitalized due to SARS-CoV-2 infection in 207 hospitals in London, Scotland and Wales between Jan 18 and Jun 4,2020 (The data were extracted from the International Severe Acute Respiratory and Emerging Infection Consortium (ISARIC) World health organization (WHO) Clinical Characterization Protocol (CCP) prospective cohort database). The study compared the mortality at 28 days post-hospitalization between participants with HIV (PWH) and those without HIV. Data were censored at the last day of follow up if the patient remained alive at 28 days or was transferred into another hospital. Data pertaining to symptoms onset, comorbidities (10 comorbidities were collected as binary variables), age, gender and ethnicity were ascertained at the day of admission, while data pertaining to disease severity and laboratory parameters were collected upon admission and at 3,6, and 9 days post-hospitalization. COVID-19 infection status was divided into three categories (community acquired (symptoms onset 14 days post hospitalization) and intermediate (symptoms onset 3 to 7 days post hospitalization). Survival analysis was used to compare the outcome (28 days mortality) between PWH and patients without HIV. Adjustment for potential confounders was done using Cox regression modelling.

Summary of Main Findings

Participants with HIV (PWH) were younger (median age of 56 years VS 74 years) and had fewer comorbidities (3 or more comorbidities were 10.7% VS 25.7%), but had higher rate of obesity (17% VS 11.4%) and moderate/severe liver disease (5% VS 2%). On day 28, 30 out of 122 (24.6%) PWH had died compared to 13969 out of 47470 (26.7%) HIV-negative participants. After adjusting for age, higher mortality was observed among PWH (adjusted hazards ratio (aHR) 1.47, 95CI :1.01-2.14). After adjusting for age, gender, ethnicity, comorbidities, disease severity at baseline (presence of absence of hypoxia [oxygen saturation less than 94%]), COVID-19 infection status, and duration of symptoms onset at baseline, PWH were at increased risk of dying within 28 days post hospitalization (aHR 1.69, 95CI:1.15-2.48) compared to HIV negative participants. The strength of the association increased in analyses limited to participants <60 years old (aHR 2.67, 95 CI: 1.70-4.86).

Study Strengths

This is the largest prospective cohort study to date that evaluated 28-day post-hospitalization mortality among PWH diagnosed with SARS-CoV-2 infection compared to HIV negative participants. Only 11.8% (6401/53993) of the initially identified participants through the database were excluded due to missing data (date of admission, HIV status, having an admission date after June 4,202 or unable to confirm their HIV status). Also, the clinical and demographic characteristics of the participants who had missing HIV status were highly similar to those without HIV, which decreases the risk of bias due to missing data.


The duration of follow up was only 28 days, therefore subjects who died after 28 days post-hospitalization were considered censored at their last day of follow up. This could have affected the study results if more participants with negative HIV status had died after 28 days post-hospitalization due to COVID-19. The study lacks important information regarding CD4 counts and HIV viral load which could explain the higher mortality among PWH. All explanatory variables in the data were modeled as binary variables, which could have resulted in residual confounding due to inadequate adjustment.

Value added

This is the largest prospective cohort study to date that compared mortality among patients with HIV who were hospitalized with COVID-19 to HIV negative individuals.

This review was posted on: 12 February 2021