Strong and consistent associations of precedent chronic rhinosinusitis with risk of non-cystic fibrosis bronchiectasis

Abstract

Background: Chronic rhinosinusitis (CRS) and bronchiectasis commonly co-occur, but most prior studies were not designed to evaluate temporality and causality.

Objectives: In a sample representing the general population in 37 counties in Pennsylvania, and thus the full spectrum of sinonasal and relevant lung diseases, we aimed to evaluate the temporality and strength of associations of CRS with non-cystic fibrosis bronchiectasis.

Methods: We completed case-control analyses for each of 3 primary bronchiectasis case finding methods. We used electronic health records to identify CRS and bronchiectasis with diagnoses, procedure orders, and/or specific text in sinus or chest computerized tomography scan radiology reports. The controls never had any indication of bronchiectasis and were frequency-matched to the 3 bronchiectasis groups on the basis of age, sex, and encounter year. There were 5,329 unique persons with bronchiectasis and 33,363 without bronchiectasis in the 3 analyses. Important co-occurring conditions were identified with diagnoses, medication orders, and encounter types. Logistic regression was used to evaluate associations (odds ratios [ORs] and 95% CIs) of CRS with bronchiectasis while adjusting for confounding variables.

Results: In adjusted analyses, CRS was consistently and strongly associated with all 3 bronchiectasis definitions. The strongest associations for CRS (ORs and 95% CIs) were those that were based on the text of sinus computerized tomography scan reports; the associations were generally stronger for CRS without nasal polyps (eg, OR = 4.46 [95% CI = 2.09-9.51] for diagnosis-based bronchiectasis). On average, CRS was identified more than 6 years before bronchiectasis.

Conclusion: Precedent CRS was strongly and consistently associated with increased risk of bronchiectasis.

Keywords: Bronchiectasis etiology; epidemiology; risk factors; sinusitis.

Overview of persons in the three case-control analyses, identifying unique persons and total sample sizes, Pennsylvania, 2010–2019. Concerning the overlap in case groups, from largest to smallest, there were 3,786 (71.1%) persons in the chest CT scan group only; 581 (10.9%) in all three case groups; 306 (5.7%) in diagnosis and order groups; 237 (4.5%) in the order only group; 189 (3.6%) in the order and text groups; 159 (3.0%) in the diagnosis only group; and 71 (1.3%) in the diagnosis and text groups. EHR = electronic health records.