Effects of Smoking on Neurocognitive Outcomes in Patients with Carbon Monoxide Poisoning

TL;DR

This study found no evidence that smoking protects against brain damage from carbon monoxide poisoning after adjusting for other factors.

Contribution

The study used propensity score matching to reduce bias and test the 'smoker’s paradox' hypothesis in CO poisoning.

Findings

Before matching, smokers had better outcomes, but this disappeared after adjusting for confounders.

After PSM, no significant difference in neurocognitive outcomes was found between smokers and non-smokers.

The relative risk of poor outcomes in smokers was not statistically significant after controlling for covariates.

Abstract

Background/Objectives: The concept of the “smoker’s paradox” in prior research posits that smoking could potentially offer neuroprotective effects in cases of acute carbon monoxide (CO) poisoning. This study aimed to determine the validity of this hypothesis by minimizing selection bias and confounding variables in a comparison of neurocognitive outcomes between smokers and non-smokers following acute CO poisoning. Methods: A total of 1150 patients were included in this retrospective study. Propensity Score Matching (PSM) was used to control for variables such as age, initial Glasgow Coma Scale (GCS) score, co-morbidities, and HBO2 therapy application. Neurocognitive outcomes were assessed and compared between smokers and non-smokers. Results: In the initial analysis, 1150 patients were divided into non-smoking (61.7%) and smoking (38.3%) groups. Before PSM, smokers had a lower rate of poor outcomes (12.1% vs. 6.8%, p = 0.004). However, baseline differences emerged, with non-smokers being older with longer CO exposure, while smokers had more males and higher rates of intentional CO poisoning and drug ingestion. Smokers had a 0.56 times lower relative risk of poor outcomes compared to non-smokers (95% CI: 0.38–0.84, p = 0.004). After meticulous 1:1 PSM with 15 covariates, 317 patient pairs were matched, creating balanced cohorts. Neurocognitive outcomes at 1 month post-CO exposure showed no significant differences between matched non-smoking and smoking groups (9.8% vs. 7.9%, p = 0.461). Post-PSM, the relative risk for poor outcomes remained approximately 0.81 times lower in smokers compared to non-smokers, but with no statistical significance [0.81 (95% CI: 0.49–1.33, p = 0.402)]. Conclusions: Our findings do not support the idea of a protective effect from smoking in the context of acute CO poisoning. After accounting for potential confounders through PSM, we found that smoking status was not significantly associated with more favorable neurocognitive outcomes.

Figures