Study population and selection criteria
Medical records of patients diagnosed with idiopathic bronchiectasis between April 2014 and December 2019 at a single hospital were retrospectively reviewed. (1) Bronchial dilatation, visibility of peripheral airways, bronchial wall thickening, and small airway abnormality confirmed by multiple radiologists and pulmonologists on HRCT and (2) patient who has received regular follow-up for over 1 year were required for inclusion. (1) Secondary bronchiectasis due to rheumatoid arthritis, ulcerative colitis, DPB, nontuberculous mycobacteriosis, and cystic fibrosis ruled out by clinical symptoms, serological test, bacteriological examination, and/or endoscopy; (2) patients with a history of AE; (3) patients who were unable to continue macrolide therapy due to side effects; and (4) patients with interstitial pneumonia and malignant tumors were the exclusion criteria. This study was approved by the Atsugi City Hospital Ethics Committee. All patient records were anonymized prior to analysis, and informed consent was received from all patients (approval number, R2-02).
Data analysis and comparison factors
All patients who met the study criteria were divided into two groups, severe and non-severe, according to a severity assessment performed by HRCT. The severe group included patients with bronchial dilatation, visibility of peripheral airways, bronchial wall thickening, and small airway abnormality in three or more lobes and in both lungs. The non-severe group had the same abnormalities in two or fewer lobes or in one side of lung. Patients in each group were also divided into subgroups based on treatment with and without long-term, low-dose macrolide therapy (shown in Fig. 1).
First, all patients were divided into subgroups, and clinical characteristics and AE incidence were compared across the groups. The patients were then divided into the non-severe and severe groups, and the same comparison was performed. Eventually, the differences in clinical characteristics between the subgroups with and without macrolide therapy were compared in the non-severe and severe groups, respectively, and the relationship between macrolide therapy and AE incidence in each subgroup was evaluated. Sex, age, body mass index (BMI), smoking status, clinical symptoms, forced vital capacity and forced expiratory volume in the first second of expiration percent predicted, bacterial colonization, treatment, and comorbidities were the variables set for comparison.
Review of radiology
All patients underwent HRCT, and radiologists’ interpretation reports were made available for review. Patterns of abnormal findings were classified as bronchial dilatation, visibility of peripheral airway, bronchial wall thickening, and small airway abnormality by multiple radiologists and pulmonologists (shown in Fig. 2). Bronchial dilatation is the ratio between the diameter of the bronchus and pulmonary artery running parallel (bronchoarterial ratio) of one or more. Visibility of the peripheral airway is the bronchus found within 1 cm of the pleura. Small airway abnormality is the centrilobular granular shadow, tree-in-bud appearance, and mosaic attenuation.
Definition of long-term, low-dose macrolide therapy and AE
Low-dose macrolide therapy in all patients consisted of erythromycin or clarithromycin administration. Each dose for erythromycin or clarithromycin was set at 400–600 or 200–400 mg, respectively. All patients administered macrolide therapy continued receiving it for at least 1 year. AE presented with acute deterioration (usually over several days) with worsening local symptoms (cough, increased sputum volume or change of viscosity, increased sputum purulence with or without increasing wheeze, breathlessness, and hemoptysis) and/or systemic upset.
Statistical analysis
The average, standard deviation, median, 25th and 75th percentile points, and range were calculated for continuous variables. Frequency and ratio were calculated for discrete variables. For intergroup comparisons, Fisher’s exact test was used for discrete variables alongside the Bonferroni correction for pairwise comparisons. All groups were compared using analysis of variance and pair comparisons by t-test in the case of the parametric method for comparison of continuous variables. For the nonparametric method, all groups were compared using the Kruskal–Wallis test, and pair comparisons were made using the Mann–Whitney U test. A logistic regression analysis was performed to investigate the relationship between macrolide therapy and AE. P < 0.05 indicated statistical significance. Statistical analyses were performed with SPSS software, version 23.0 (IBM Japan, Ltd., Tokyo, Japan).