We used logistic regression to assess our outcome (resumption of firefighting duties) after adjustment for age, smoking status, and the presence or
absence of airflow obstruction; results are expressed as odds ratios and 95 percent confidence intervals.