Exertional hypoxemia in stable COPD is common and predicted by circulating proadrenomedullin.

ABSTRACT

BACKGROUND: The prevalence of exertional hypoxemia in unselected patients with COPD is unknown. Intermittent hypoxia leads to adrenomedullin (ADM) upregulation through the hypoxia-inducible factor-1 pathway. We aimed to assess the prevalence and the annual probability to develop exertional hypoxemia in stable COPD. We also hypothesized that increased ADM might be associated with exertional hypoxemia and envisioned that adding ADM to clinical variables might improve its prediction in COPD. METHODS: A total of 1,233 6-min walk tests and circulating proadrenomedullin (proADM) levels from 574 patients with clinically stable, moderate to very severe COPD enrolled in a multinational cohort study and followed up for 2 years were concomitantly analyzed. RESULTS: The prevalence of exertional hypoxemia was 29.1%. In a matrix derived from a fitted-multistate model, the annual probability to develop exertional hypoxemia was 21.6%. Exertional hypoxemia was associated with greater deterioration of specific domains of health-related quality of life, higher severe exacerbation, and death annual rates. In the logistic linear and conditional Cox regression multivariable analyses, both FEV1% predicted and proADM proved independent predictors of exertional hypoxemia (P < .001 for both). Adjustment for comorbidities, including cardiovascular disorders, and exacerbation rate did not influence results. Relative to using FEV1% predicted alone, adding proADM resulted in a significant improvement of the predictive properties (P = .018). Based on the suggested nonlinear nomogram, patients with moderate COPD (FEV1% predicted = 50%) but high proADM levels (> 2 nmol/L) presented increased risk (> 30%) for exertional desaturation. CONCLUSIONS: Exertional desaturation is common and associated with poorer clinical outcomes in COPD. ADM improves prediction of exertional desaturation as compared with the use of FEV1% predicted alone. TRIAL REGISTRY ISRCTN Register; No. ISRCTN99586989; URL www.controlled-trials.com.