Corticosteroids, Antitussives, and Inhalers for Lower Respiratory Tract Infections in US Primary Care: A Prospective Cohort Study.

Abstract

Physicians often prescribe medications other than antibiotics to address symptoms in patients with lower respiratory tract infection (LRTI), but their frequency and effectiveness have been poorly studied. To describe use of non-antibiotic medications in adult outpatients with LRTI, and whether their use is associated with reduced duration or severity of illness. Prospective cohort study. Adult outpatients 18 to 75 years with acute cough and at least one lower or systemic respiratory symptom suspected of having a LRTI. The percentage of patients given non-antibiotic medications and propensity score matched analysis of association of systemic corticosteroids, benzonatate, and albuterol inhaler on duration of cough, severity of cough, receipt of an antibiotic, and need for an unscheduled follow-up visit. Of 718 patients, 690 had baseline severity data, 517 had diary data for duration, and 499 had data for severity. Benzonatate (23.2%), an albuterol inhaler (19.1%), and a systemic corticosteroid (18.6%) were commonly prescribed. In univariate analysis, receipt of benzonatate, an albuterol inhaler, or any antitussive was associated with greater baseline severity of cough. Receipt of benzonatate was also associated with greater overall duration and severity of cough. Receipt of a systemic corticosteroid was associated with a greater likelihood of receiving an antibiotic (54.7% vs. 24.2%, p < 0.001). In the propensity score matched analysis, there was no association between receiving a systemic corticosteroid, an albuterol inhaler, or benzonatate and the duration of cough, severity of cough, or need for an unscheduled follow-up visit. Receiving a systemic corticosteroid was associated with an increased likelihood of also getting an antibiotic (adjusted odds ratio 1.37). Despite having no evidence of benefit and well-known harms, systemic corticosteroids and benzonatate were commonly prescribed. Adequately powered pragmatic trials in primary care settings are badly needed to provide appropriate data to guide practice.