Gurgling breath sounds may predict hospital-acquired pneumonia

Objectives: To determine whether gurgling sounds heard during speech or quiet breathing, with or without a stethoscope over the glottis, predict hospital-acquired pneumonia(HAP). Methods: All patients admitted to the respiratory or general medicine ward of a 350-bed community teaching hospital were eligible. Patients were examined each day, and those who had upper airway gurgling, heard with or without the stethoscope, during breathing or speech at any point during admission were noted. Assuming an overall incidence of HAP ( > 48 h after admission) of 5% to 10% and estimated incidence of 30% to 50% in patients with gurgle, 20 patients with gurgle and 60 patients without gurgle, matched on the same day and ward of admission, were included in the study. Demographic, physiologic, and outcome variables were compared using univariate and multivariate techniques to ascertain whether gurgling is independently associated with HAP, rate of transfer to ICU, and inhospital mortality. Results: Twenty patients with gurgle were compared with 60 patients without gurgle. Patients with gurgle were older (78.5 vs 65.2 y; P < .001), more likely to reside in nursing homes(75% vs 6%; P < .001), and were more likely to have dementia(70% vs 13%; P < .001). In multivariate analysis, dementia (odds ratio [OR] = 23.4; 95% CI, 4.2-131.9) and recent(within 24 h) treatment with opiates (OR = 14.7; 95% CI, 2.2-97.5) emerged as the only statistically significant independent predictors of gurgling. HAP occurred in 55% of patients with gurgle compared with 1.7% of patients without gurgle (P < .001), and 50% of patients with vs 3.3% of patients without gurgle required transfer to ICU (P,.001). After adjustment for age, Charlson score, dementia, opiate administration, and stroke, gurgling emerged as the sole independent predictor of HAP(OR = 140.1; 95% CI, 5.6-3,529.4) and ICU transfer(OR = 35.1; 95% CI, 4.1-303.7). Gurgling did not predict mortality; the Charlson comorbidity index was the only significant predictor of inhospital death. Conclusions: Gurgling sounds heard during quiet breathing or speech are independently associated with HAP.