• Keith S Kaye, Malini Stalam, Wendy E Shershen, and Donald Kaye.
• Department of Medicine, Duke University Medical Center, Durham, North Carolina 27710, USA. kaye0001@mc.duke.edu
• Am. J. Med. Sci. 2002 Nov 1; 324 (5): 237242237-42.

BackgroundThe differential diagnosis of acute infection in elderly nursing home patients is often difficult. This study evaluated pulse oximetry in pneumonia in this population.MethodsA case-control study was performed in a veteran's nursing home involving 2 analyses: (1) pneumonia patients (case subjects) were compared with patients with nonpulmonary infections (control subjects) at time of acute infection; (2) differences in paired values measured at time of infection versus a noninfected baseline were compared for pneumonia patients and control subjects. Vital signs including pulse oximetry were obtained routinely (at least monthly) and with acute illness.ResultsOxygen saturations were lower in 45 pneumonia patients than in 22 patients with acute nonpulmonary infections (P < 0.001). An oxygen saturation < 94 gave a sensitivity for pneumonia of 80%, specificity of 91%, and positive predictive value of 95%. The drop in oxygen saturation from the last baseline value was greater in pneumonia patients than in control subjects (P < 0.001). The sensitivity of an oxygen saturation drop >3% from baseline for pneumonia was 73% with specificity and positive predictive values of 100%.ConclusionsPulse oximetry may be very helpful in evaluating acutely infected nursing home residents. The present study suggests that in acutely infected nursing home patients, a decrease in oxygen saturation of >3% from baseline, as well as a single oxygen saturation of <94, should suggest pneumonia. A decrease from baseline of <4% or a single oxygen saturation of 94 or higher suggests that pneumonia is unlikely.

28 keywords...

hide…