COVID-19 vaccine effectiveness by HIV status and history of injection drug use: a test-negative analysis.

Introduction: People living with HIV (PLWH) and/or who inject drugs may experience lower vaccine effectiveness (VE) against SARS-CoV-2 infection.

Methods: A validated algorithm was applied to population-based, linked administrative datasets in the British Columbia COVID-19 Cohort (BCC19C) to ascertain HIV status and create a population of PLWH and matched HIV-negative individuals. The study population was limited to individuals who received an RT-PCR laboratory test for SARS-CoV-2 between 15 December 2020 and 21 November 2021 in BC, Canada. Any history of injection drug use (IDU) was ascertained using a validated administrative algorithm. We used a test-negative study design (modified case-control analysis) and multivariable logistic regression to estimate adjusted VE by HIV status and history of IDU.

Results: Our analysis included 2700 PLWH and a matched population of 375,043 HIV-negative individuals, among whom there were 351 and 103,049 SARS-CoV-2 cases, respectively. The proportion of people with IDU history was much higher among PLWH compared to HIV-negative individuals (40.7% vs. 4.3%). Overall VE during the first 6 months after second dose was lower among PLWH with IDU history (65.8%, 95% CI = 43.5-79.3) than PLWH with no IDU history (80.3%, 95% CI = 62.7-89.6), and VE was particularly low at 4-6 months (42.4%, 95% CI = -17.8 to 71.8 with IDU history vs. 64.0%; 95% CI = 15.7-84.7 without), although confidence intervals were wide. In contrast, overall VE was 88.6% (95% CI = 88.2-89.0) in the matched HIV-negative population with no history of IDU and remained relatively high at 4-6 months after second dose (84.6%, 95% CI = 83.8-85.4). Despite different patterns of vaccine protection by HIV status and IDU history, peak estimates were similar (≥88%) across all populations.

Conclusions: PLWH with a history of IDU may experience lower VE against COVID-19 infection, although findings were limited by a small sample size. The lower VE at 4-6 months may have implications for booster dose prioritization for PLWH and people who inject drugs. The immunocompromising effect of HIV, substance use and/or co-occurring comorbidities may partly explain these findings.