676: The Association of Smoking Status and Mortality Among Critically Ill Influenza Patients

Introduction: Among critically ill influenza patients, the association of smoking status with short-term mortality remains unclear with some studies suggesting a “smoker’s paradox.” We studied ICU patients to evaluate the association of smoking status with short-term mortality and hypothesized that current smokers would have higher mortality than never smokers. Methods: We used statewide data to study adults admitted to intensive care units of acute care hospitals in Texas during 2016 Q1 through Q2 2024 with a principal diagnosis of influenza. Hospitalizations with a principal diagnosis of influenza were identified using International Classification of Diseases, Tenth Revisions, Clinical Modification (ICD-10-CM) codes selected from Clinical Classification Software Refined (CCSR) category RSP003: Influenza. The primary exposure was smoking status identified using CCSR categories MBD024: Tobacco related disorders and FAC016: Personal history of nicotine dependence. Overlap propensity score weighting was applied with results expressed as adjusted risk ratio and 95% confidence interval (aRR [95% CI]). Results: A total of 15,887 hospitalizations were included in the study of which 62% were never smokers, 21% former smokers, and 17% current smokers. Current smokers were younger than never smokers (40% vs 64% aged ≥65 years) but had higher Deyo comorbidity index (mean 2.3 vs 2.0); p < 0.0001 for each comparison. Unadjusted short-term mortality was lower in current smokers (3.0% vs 4.9%). After overlap weighting, all risk-adjustment covariates were well balanced and each smoking status had approximately equal weight. Within the overlap population, current smoking remained associated with lower short-term mortality compared to never smoking, 3.1% vs 4.5% (aRR 0.69 [95% CI 0.53 to 0.91]). The association of former smoking status and mortality compared to current and never smokers were not statistically significant. Conclusions: Contrary to our hypothesis, results showed current smokers had lower short-term mortality than never smokers, supporting a possible “smoker’s paradox.” Former smokers had similar outcomes to never smokers. These findings should be interpreted with caution. Further research is needed to understand if biological or clinical factors explain this association.