Identifying COPD patients at increased risk of mortality: Predictive value of clinical study baseline data

Background: Identifying chronic obstructive pulmonary disease (COPD) patients at increased risk of mortality is an important component of effective disease management. Methods: A pooled analysis of patients with severe COPD, from two well-controlled 1-year studies, was conducted using Cox regression and spline analysis to evaluate predictability of baseline demographic data and in-study variables for mortality risk, and to evaluate the effect of treatment allocation to budesonide and formoterol, versus their respective control groups, on these outcomes. Results: In the pooled analysis, a Cox regression model. reported a higher baseline St George's Respiratory Questionnaire (SGRQ) total score as a significant predictor of mortality (hazard ratio 1.037 [95% confidence interval. 1.021-1.054]; p < 0.0001). The 36-item short-form health survey (SF-36) mental and physical. component scores were also predictive of an increased mortality risk (p < 0.05). Age, forced expiratory volume in 1 s (FEV1), body mass index and smoking status were not significant predictors. Spline analysis of baseline variables revealed a linear association between SGRQ total score and mortality risk over 1 year (logarithmic scale). Other baseline variables, including FEV1, showed different bimodal patterns in the spline analysis. There was no difference in mortality in the formoterol. versus the non-formoterol treatment group while budesonide-containing treatment was associated with reduced 1-year, all-cause, in-study mortality compared with non-budesonide therapy. Conclusion: Health status measured by SGRQ and SF-36 may be important for predicting COPD patients at increased mortality risk, with SGRQ total score emerging as the strongest predictor compared with other baseline covariates. (C) 2008 Elsevier Ltd. All rights reserved.