To assess the association of selected clinical factors and specific medication use (proton pump inhibitors, H2 receptor antagonists [H2 blockers], and angiotensin-converting enzyme inhibitors) with presence of pneumonia in patients with stroke undergoing acute inpatient rehabilitation. Matched case-control study in a freestanding urban academic inpatient acute rehabilitation hospital. Participants were 72 stroke survivors, consisting of 36 patients who developed pneumonia during rehabilitation hospitalization individually matched in order of decreasing priority on age, sex, stroke side, depth, and severity with 36 patients with stroke not developing pneumonia. Potential risk factors, including severe dysphagia, dietary interventions, presence of tracheostomy or feeding tube, and specific medications, were assessed for association with pneumonia during rehabilitation using separate univariate and multivariate analyses. Functional change was assessed using the functional independence measure. Although pneumonia was associated with proton pump inhibitors or H2 blockers (odds ratio, 3.3; 95% confidence interval, 1.0-13.7), any feeding tube (odds ratio: 5.0; 95% confidence interval, 1.4-27.0), severe dysphagia (odds ratio: 15.0; 95% confidence interval, 2.3-631), and tracheostomy (odds ratio: 10; 95% confidence interval, 1.4-434.0) on univariate evaluation, none of these individual factors was significantly associated with pneumonia in a multivariate model. Risk factors were found to be highly related to each other. Odds of pneumonia did not significantly decrease with angiotensin-converting enzyme inhibitors (odds ratio: 0.9; 95% confidence interval, 0.2-3.0). Patients with pneumonia had a significantly lower functional independence measure score at discharge. A reduction in pneumonia was not found with the use of angiotensin-converting enzyme inhibitors. Although tracheostomies, feeding tubes, proton pump inhibitor or H2 blocker use, and the presence of dysphagia were identified as risk factors for pneumonia on univariate analyses, none of these factors demonstrated an independent association with pneumonia on multivariate analyses. It may be more that the underlying impairment, rather than the assessed interventions, may confer greater risk of pneumonia in the poststroke patient.
Read full abstract