Substance-specific Mortality Risk Stratification Among Adults With Substance Use Disorders During Acute Respiratory Infection: A 3.4-Million-patient N3C Cohort Study.

Abstract

Objectives: To quantify 30-day all-cause mortality associated with pre-existing substance use disorders (SUDs) during acute SARS-CoV-2 infection, to compare mortality risk across major SUD subtypes (tobacco, alcohol, opioid, cannabis, cocaine, amphetamine, inhalant, and unspecified), and to evaluate the extent to which documented comorbidity burden accounts for observed SUD-associated mortality differences.

Methods: Retrospective cohort study of adults aged 18-65 years or older with confirmed SARS-CoV-2 infection (March 2020-June 2023) from 63 US healthcare organizations contributing to the National COVID Cohort Collaborative (N3C). Confirmed infection required any 1 of an International Classification of Diseases, 10th Revision (ICD-10) diagnosis code of U07.1, a positive SARS-CoV-2 laboratory test (polymerase chain reaction or antigen), or a nirmatrelvir-ritonavir (Paxlovid) prescription. Multivariable logistic regression estimated adjusted odds ratios (aORs) for 30-day mortality associated with any SUD and with each SUD subtype, adjusting for age, sex, race and ethnicity, body mass index, and tobacco smoking status. Sensitivity analyses added the Elixhauser Comorbidity Index (ECI) and used an unrestricted cohort.

Results: Among 3,435,480 adults with confirmed SARS-CoV-2 infection, 406,064 (11.8%) had a pre-existing SUD documented before the index date, and 14,866 (0.43%) died within 30 days of the index date. Thirty-day mortality was higher among individuals with SUDs than those without (1.05% vs. 0.35%). SUDs were associated with more than doubled adjusted odds of mortality (aOR 2.38; 95% CI: 2.28-2.49). Additional adjustment for comorbidity burden attenuated but did not eliminate the association (aOR 1.53; 95% CI: 1.46-1.61). Alcohol use disorder (AUD) (aOR 2.63; 95% CI: 2.48-2.78) and opioid use disorder (OUD) (aOR 2.53; 95% CI: 2.35-2.74) conferred the highest risks. Mortality differentials between individuals with and without SUD persisted throughout the study period despite overall declines in COVID-19 mortality.

Conclusions: Pre-existing SUDs, particularly alcohol (AUD) and opioid use disorders (OUD), identified a population at substantially increased 30-day all-cause mortality after COVID-19 diagnosis. Because unmeasured structural, behavioral, and treatment-related factors likely contribute to the residual excess risk beyond measured comorbidity, SUD is best interpreted as a clinically useful marker of heightened vulnerability rather than as an independent biological cause of mortality. SUD status, AUD, and OUD in particular, warrant explicit incorporation into short-term risk stratification for respiratory infections in both routine clinical care and pandemic preparedness planning.