There is conflicting information regarding the severity and outcomes of COVID-19 in people living with human immunodeficiency virus (PLHIV). Data from previously published matched case–control studies showed no differences in severity or outcome between HIV-infected and HIV-negative patients.1–3 Also, a cohort study of 2988 HIV-infected patients evaluating the association between HIV and COVID-19 diagnoses, hospitalisation, and in-hospital death in New York State reported poor outcomes in HIV-infected patients compared with HIV-negative patients.4 There are no specific studies investigating the severity and outcomes of critically ill patients with COVID-19 and HIV-infection. We aim to investigate whether the clinical presentation, severity and outcomes of COVID-19 in critically ill HIV-infected patients were comparable to those seen in non-HIV-infected patients.

The CIBERESUCICOVID is a multicentre, observational, prospective/retrospective cohort study (NCT04457505) of 6512 consecutive patients with SARS-CoV-2 infection who were admitted to 69 ICUs in Spain between February 2020 and July 2022. Approved by the ethical committee – HCB/2020/0370. Data was collected as previously described.5 The primary outcome was all-cause 90-day mortality. Secondary outcomes included all-cause in-hospital, 15-day, 30-day and 1-year mortality, length of ICU and hospital stay, and ventilator and ICU-free days. Propensity score matching (PSM) method6,7 was used to obtain the balance among baseline variables between patients with and without HIV. We used a 1:2 nearest-neighbour matching with age, sex, number of comorbidities and SOFA as covariates and an exact matching constraint on wave of COVID-19 and centre, without replacement and within a calliper width of 0.6. We estimated the 15-, 30- and 90-day, and 1-year mortality with 95% confidence intervals (CI).8 We analysed the association between HIV infection and mortality by means of Cox regression analyses. Survival curves of patients with and without HIV were obtained using the Kaplan–Meier method and compared using the Gehan–Breslow–Wilcoxon test.9

Fig. 1 depicts patients flow-chart and Table 1 describes the full cohort (N=6406: 6372 HIV-negative and 34 PLHIV) and the propensity score matching (N=93, 62 HIV-negative and 31 PLHIV) cohort. In the full cohort, patients with HIV infection were younger, presented higher numbers of comorbidities, were more frequently current smokers, alcohol abusers, had higher APACHE II scores at ICU admission than patients without HIV infection. After PSM, no significant differences were found in age, number of comorbidities, proportion of smokers or alcohol abusers. However, the APACHE II score was higher in PLHIV (Table 1). Sex and SOFA did not differ between groups before and after PSM. In the full cohort, cardiac ischaemia was the only complication that differed between groups, showing a higher rate in the HIV infection group compared to non-HIV infection group. After PSM, no significant differences were found in any complication. Finally, all outcomes did not differ between groups before and after PSM. In the full cohort, the 15-, 30-, 90-day, and 1-year mortality rates were 12% (95% CI, 12–13%), 25% (95% CI, 23–26%), 34% (95% CI, 33–35%) and 39% (95% CI, 38–40%) in the non-HIV infection group, compared with 16% (95% CI, 3–28%), 22% (95% CI, 8–36%), 22% (95% CI, 8–36%) and 27% (95% CI, 11–42%) in the HIV infection group, respectively. After PSM, the 15-, 30-, 90-day, and 1-year mortality rates were 10% (95% CI, 2–17%), 21% (95% CI, 11–31%), 26% (95% CI, 15–37%) and 30% (95% CI, 18–42%) in the non-HIV infection group, compared with 14% (95% CI, 1–26%), 21% (95% CI, 6–35%), 21% (95% CI, 6–35%) and 26% (95% CI, 9–42%) in the HIV infection group, respectively. In the full cohort, Cox regression analyses for 15-, 30-, 90-day, and 1-year mortality did not find significant differences between groups, giving HRs of 1.24 (95% CI, 0.52–2.99), 0.92 (95% CI, 0.44–1.94), 0.66 (95% CI, 0.32–1.40) and 0.73 (95% CI, 0.36–1.46), respectively. After PSM, Cox regression analyses for 15-, 30-, 90-day, and 1-year mortality did not find significant differences between groups, giving HRs of 1.40 (95% CI, 0.39–4.95), 1.01 (95% CI, 0.38–2.65), 0.81 (95% CI, 0.32–2.07) and 0.89 (95% CI, 0.37–2.15), respectively.

Fig. 1.

Flow chart of the study population.

(0.42MB).

Of the 34 PLHIV included in the study, 26 (76%) were males, with a median (IQR) age of 51 (47; 65) years. Data on the probable route of exposure were available in 47% of cases. HIV infection was acquired by men who have sex with men and intravenous drug use in 7 (44%) and 5 (31%) cases, respectively. Data on available ART therapy was available in 29 cases (85%). Twenty-five out of 29 cases (86%) were on antiretroviral therapy (ART) at the time of hospital admission. ART regimens were based on integrase inhibitors, non-nucleoside reverse transcriptase inhibitors, protease inhibitors, and other regimens in 11 (44%), 9 (36%), 3 (12%), and 2 (8%) cases, respectively. Data on HIV-RNA viral load in plasma was available in 27 patients (79%). Twenty-one out of 27 patients (84%) had an undetectable HIV-RNA viral load in plasma (<50copies/mL). Data on CD4+ T cell count was available in 21 patients (62%). The median (IQR) last CD4+ T cell count was 518 (421; 823)/mm3. The percentage of PLHIV with less than 200 CD4cells/mm3 was 4.7%. Three patients (10%) presented co-infection with HCV, and three patients (10%) had HBV co-infection. There were also no differences regarding outcomes between HIV-infected patients with undetectable and detectable HIV viral load.

This is the first national multicentre study that describes the clinical characteristics and outcomes of critical COVID-19 in PLHIV admitted to the ICU and compares them with the HIV-negative general population, concluding that the short- and medium-term prognosis is similar in PLHIV who are virological suppressed on ART and are not immunosuppressed (CD4≥200cells/mm3). In our study PLHIV accounted for approximately 0.5% of total hospitalised, critically ill COVID-19 cases in the ICU. This data is in line with previous studies.2–4 Despite PLHIV presenting higher APACHE-II scores at ICU admission, outcomes did not differ between groups before and after PSM (in-hospital, 30-, 90-day and 1-year mortality). Similar results were reported by previous studies in hospitalised patients.10–12 Data from a recent systemic review and meta-analysis that included data from 28 studies found no difference in the risk of death between HIV-infected patients and HIV-negative (OR, 1.09; 95% CI, 0.93–1.26; P>0.001).12

The major strengths of this study include its multicentre nature, the consecutive inclusion of all patients from each unit, thorough checking of data quality, and the high number of patients analysed and long-term follow-up. On the other hand, despite exhaustive propensity score analysis for underlying conditions, a possible limitation of the propensity score methods is their inability to control for unmeasured confounding. A main limitation is the small number of cases, precluding any robust conclusions. In particular, the analysis of outcomes using this size sample may have led to a large type-II error that prevents us from generalising our results. However, there are no similar studies investigating the severity and outcomes of critically ill patients with COVID-19 and HIV-infection, so this information is of value. Also, these results do not apply to PLHIV who are off ART and/or are immunosuppressed, because they are underrepresented in our study. Finally, as we examined real-world data, limitations associated to their observational nature and missing data should be considered.

Although limited by the small sample size, our main conclusion is that critically ill COVID-19 in non-immunosuppressed and virologically suppressed HIV-infected individuals seems to present neither a more severe disease nor a worse clinical outcome than HIV-negative patients. Further research in large cohorts should be encouraged to improve our knowledge on the impact of SARS-CoV-2 in critically ill HIV-infected patients.