Carbon monoxide detector effectiveness in reducing poisoning, Wisconsin 2014–2016

Abstract

Introduction: Carbon monoxide (CO) is a colorless, odorless, and nonirritating gas. The most common exposures are from gas powered appliances such as furnaces, water heaters, stoves, and vehicles. To prevent poisoning, CO detectors with audible alarms were developed. This study aims to evaluate the effectiveness of CO detectors in reducing poisoning in Wisconsin. Methods: Records were queried from National Poison Data System for unintentional CO exposures that occurred in residences in Wisconsin during 2014–2016 (N = 703). After applying sample exclusion criteria, notes were abstracted for cases where CO detector use was mentioned (n = 408). Logistic regression analyses were used to assess the association between having a CO detector alarm and CO poisoning. Linear regression analyses were used to assess the relationship between having a CO detector alarm and poisoning severity. Results: In logistic models, odds of CO poisoning were 3.2 times higher (95% CI: 1.5, 6.9) among those who had no CO detector compared to those who had a CO detector that alarmed. In linear models, not having a CO detector was associated with a 0.34 point (95% CI: 0.17, 0.54) change in outcome severity score compared to having a CO detector that alarmed. Discussion: Individuals who were exposed to CO in the absence of a CO detector were more likely to be poisoned and to have more severe medical outcomes than those that had a CO detector that alarmed.