Clinical Markers Associated With Risk of Suicide or Drug Overdose Among Individuals With Smoking Exposure: A Longitudinal Follow-up Study of the COPDGene Cohort

Abstract

Background: Studies have shown that COPD and smoking are associated with increased suicide risk. To date, there are no prospective studies examining suicide risk among individuals with smoking exposure along a spectrum of pulmonary diseases ranging from normal spirometry to severe COPD.

Research question: Which clinical variables predict death by suicide or overdose of indeterminate intent in a large cohort of individuals with smoking exposure within the Genetic Epidemiology of COPD (COPDGene) study?

Study design and methods: We studied data from 9,930 participants involved in COPDGene, a multisite, prospective cohort study of individuals with smoking exposure. Primary cause of adjudicated deaths was identified by using death certificates, family reports, and medical records. Time to death by suicide/overdose was examined as the primary outcome in Cox regression models including age, sex, race, BMI, pack-years, current smoking status, airflow limitation (FEV₁ % predicted), dyspnea (modified Medical Research Council scale score ≥ 2), 6-min walk distance, supplemental oxygen use, and severe exacerbations in the prior year with time-varying covariates and other causes of death as a competing risk.

Results: The cohort was 47% female and 33% Black (67% White); they had a mean ± SD age of 59.6 ± 9.0 years and a mean FEV₁ % predicted of 76.1 ± 25.5. Sixty-three individuals died by suicide/overdose. Factors associated with risk of suicide/overdose were current smoking (hazard ratio [HR], 6.44; 95% CI, 2.64-15.67), use of sedative/hypnotics (HR, 2.33; 95% CI, 1.24-4.38), and dyspnea (HR, 2.23; 95% CI, 1.34-3.70). Lower risk was associated with older age (per-decade HR, 0.45; 95% CI, 0.31-0.67), higher BMI (HR, 0.95; 95% CI, 0.91-0.99), and African-American race (HR, 0.41; 95% CI, 0.23-0.74). Severity of airflow limitation (FEV % predicted) was not associated with suicide risk.

Interpretation: In this well-characterized cohort of individuals with smoking exposure with and without COPD, risk factors for suicide/overdose were identified that emphasize the subjective experience of illness over objective assessments of lung function.

Keywords: COPD; overdose; prospective cohort study; suicide deaths; tobacco smoking.

Figure 1

Kaplan-Meier survival curves for clinical variables of interest. Figure shows the Kaplan-Meier survival curves for time to death by suicide/overdose for groups stratified according to clinical variables of interest at baseline. These variables included current tobacco use, dyspnea (defined by a modified Medical Research Council Dyspnea scale score ≥ 2) and use of sedative/hypnotic medications. Unlike the primary Cox regression models, which model these variables as time-varying covariates, groups here are based on baseline variables only, and without adjustment, to illustrate differences in risk between groups on these variables of interest at baseline.