Among 167,000 ICU Pneumonia Patients, Cannabis Misuse Was Not Independently Linked to Higher Mortality | Research

The takeaway

In a large US database study of 167,095 ICU patients with pneumonia, alcohol misuse was associated with increased mortality while opioid misuse was linked to lower mortality, but cannabis, stimulant, and polysubstance misuse showed no consistent mortality association after adjustment.

Intensivists, pulmonologists, addiction medicine specialists, hospital epidemiologists.

Cannabis misuse showed no independent mortality association in 167,095 ICU pneumonia cases

What the researchers found

Alcohol misuse (5.0% of patients) was associated with increased in-hospital mortality (adjusted OR 1.12, 95% CI: 1.06-1.19). Opioid misuse (1.5%) was associated with decreased mortality (OR 0.46, 95% CI: 0.39-0.53). Cannabis misuse (0.6%), stimulant misuse (0.6%), and polysubstance misuse (1.2%) showed lower mortality in unadjusted analyses but no consistent association after adjustment.

Why it matters

This is one of the largest studies to examine how different types of substance misuse relate to pneumonia outcomes in ICU. The finding that cannabis misuse was not independently associated with worse mortality provides some reassurance, though the study captured only diagnosed misuse disorders.

The numbers in context

N=167,095 ICU pneumonia patients. Alcohol misuse: 5.0%. Cannabis: 0.6%. Opioid: 1.5%. Stimulant: 0.6%. Polysubstance: 1.2%. No substance misuse: 91.1%. Alcohol mortality OR=1.12. Opioid mortality OR=0.46.

How the study worked

Retrospective cohort study of 167,095 ICU patients with pneumonia in the Premier Healthcare Database (2010-2017). Multivariable mixed-effects logistic regression adjusted for age, comorbidities, and hospital characteristics.

What this study cannot tell us

Administrative data likely underdiagnose substance misuse, especially cannabis. Only 0.6% had cannabis misuse documented, far below population prevalence. Detection bias (younger, healthier patients may be more likely to have substance use documented). Cannot distinguish between active and historical use.

How to read the evidence

Large retrospective database study with multivariable adjustment, but administrative data likely underdiagnose substance misuse, especially cannabis.

When this study was published

Published in 2023 using data from 2010-2017.

The bigger picture

The protective association with opioid misuse is likely explained by younger age and fewer comorbidities in that group, plus possible detection bias. The lack of independent cannabis association should be interpreted cautiously given the low prevalence (0.6%) and the likelihood that cannabis use is underdiagnosed in administrative data.

Questions still open

Is the low cannabis misuse prevalence (0.6%) due to underdiagnosis, and would higher detection rates change the findings?
Does the opioid misuse mortality benefit persist when controlling for age and healthy user bias?

Common questions

Does cannabis use worsen pneumonia outcomes?

In this large study, cannabis misuse was not independently associated with increased mortality in ICU pneumonia patients after adjusting for age, comorbidities, and hospital factors.

Which substance was most harmful for pneumonia patients?

Alcohol misuse was the only substance type consistently associated with increased ICU mortality. Cannabis, stimulant, and polysubstance misuse showed no consistent independent association.

Read the original research

Association between Substance Misuse and Outcomes in Critically III Patients with Pneumonia.

Annals of the American Thoracic Society, 20(4), 556-565

View on PubMed ↗Publisher’s paper ↗

Citation

Reynolds, Paul M; Afshar, Majid; Wright, Garth C; Ho, P Michael; Kiser, Tyree H; Sottile, Peter D; Althoff, Meghan D; Moss, Marc; Jolley, Sarah E; Vandivier, R William; Burnham, Ellen L. (2023). Association between Substance Misuse and Outcomes in Critically III Patients with Pneumonia.. Annals of the American Thoracic Society, 20(4), 556-565. https://doi.org/10.1513/AnnalsATS.202206-532OC