Background and objective: Although sputum culture in patients with an acute exacerbation of COPD is of uncertain value, it is routinely done. The ability to clinically identify patients likely or unlikely to yield bacterial sputum isolates would potentially reduce unnecessary tests. The objective of this study was to identify the clinical predictors of positive sputum cultures in this patient population.
Methods: Consecutive patients with a COPD exacerbation requiring an emergency visit were prospectively enrolled. Quantitative sputum culture was performed on-site. Data on current smoking, sputum purulence, FEV(1), Medical Research Council chronic dyspnoea scale, BMI, severe exacerbations in the preceding year requiring hospitalization, PaO(2), PaCO(2), Acute Physiology and Chronic Health Evaluation (APACHE) II score, and oral and inhaled steroid use were recorded.
Results: Of the 94 patients enrolled, sputum from 36 yielded bacterial pathogens. These patients were characterized by a higher frequency of purulent sputum, lower FEV(1), BMI and PaO(2,) higher APACHE II score and more frequent use of inhaled steroids (P < 0.05). On multivariate regression, purulent sputum, FEV(1) and BMI were independent determinants of a positive sputum culture. Using receiver-operator-optimized thresholds for these variables (purulent sputum, FEV(1) < 35% predicted and BMI < or = 22 kg/m(2)), we proposed a regression coefficient-weighted prediction model that accurately determined the likelihood of sputum bacterial isolation.
Conclusions: A prediction model based on the variables of purulent sputum, FEV(1) and BMI predicted sputum culture result with about 90% accuracy. Pending further validation, this model may save valuable healthcare resources.