Study design
We performed an observational study of historical cohorts from 4 Spanish teaching hospitals with multidisciplinary and standardized non-CF bronchiectasis outpatient clinics.
Study population
The study population comprised 296 consecutive patients aged ≥ 18 years who had been diagnosed with non-CF bronchiectasis of widely varying causes and for whom radiological extension and clinical and functional impairment were confirmed. Patients had to have been followed for at least 5 years during the period 2002–2010 before they could be considered for inclusion in analysis.
Patients had to have at least 2 sputum samples cultured for mycobacteria while (in a clinically stable phase) during the 5 years after the diagnosis. According to the recommendations of the Spanish Society of Pulmonology and Thoracic Surgery, the causes ruled out in idiopathic bronchiectasis were as follows: immunodeficiency with evidence of defective antibody production, gastroesophageal reflux disease, allergic bronchopulmonary aspergillosis, mycobacterial infections prior to development or diagnosis of bronchiectasis, cystic fibrosis, primary ciliary dyskinesia, and α1-antitrypsin deficiency [8]. CF was ruled out by 2 negative sweat test results in patients with bronchiectasis of unknown cause or a clinical presentation compatible with CF [9]. The study was approved by the Ethics and Research Committee of each center (registration number of the coordinating center: 0088-89-2011).
Diagnosis of bronchiectasis
Bronchiectasis was diagnosed based on a high-resolution computed tomography scan of the chest that was interpreted by radiologists experienced in respiratory disorders. Images were obtained in full inspiration (1-mm collimation and 10-mm intervals from the apex to the base of the lungs). The presence of bronchiectasis was confirmed based on the criteria published by Naidich et al. [10]. The extent of bronchiectasis was evaluated according to the number of lobes affected, with the lingula and middle lobe considered as independent lobes.
Data collection
Data were collected from all clinically stable patients and included the following: age, gender, body mass index (BMI, kg/m2), etiology, smoking habit (pack-years), dyspnea according to the modified Medical Research Council scale, macroscopic appearance of sputum (percentage of patients with purulent sputum), type of bronchiectasis (cystic or noncystic), radiological findings (number of lobes affected by bronchiectasis), and spirometry findings (forced vital capacity [FVC] and forced expiratory volume in the first second [FEV1] as both absolute values and percent predicted). We also recorded hospitalizations secondary to severe exacerbations and the number of exacerbations. All variables were obtained within 6 months after the radiological diagnosis of bronchiectasis, except for hospitalizations and the number of exacerbations, which were obtained prospectively during the year after the radiological diagnosis. Long-term treatments (antibiotics, oral macrolides, and oral corticosteroids) taken for at least 1 year after the radiological diagnosis of bronchiectasis were also recorded.
One sputum sample was recovered every 6 months during a clinically stable phase and cultured for mycobacteria, bacteria, and fungi. Additional sputum cultures were obtained whenever considered necessary by the clinician.
A stable clinical situation was defined as the absence of clinical criteria of exacerbation and no antibiotics or corticosteroids in the preceding 4 weeks [11]. Exacerbation was defined as the acute onset and persistence of changes in sputum characteristics (increased volume, thicker consistency, greater purulence, and hemoptysis) and/or increased breathlessness unrelated to other causes [12].
Microbiology
Each respiratory sample was collected under sterile conditions and processed immediately or conserved at 4 °C. Smears were stained using the auramine acid fast method and scanned with microscopy fluorescence light at × 200 and × 400. All respiratory samples were processed according to the Tacquet-Tison technique. Processed samples were inoculated on Lowenstein-Jensen, Coletsos, and liquid medium (VersaTREK, formerly ESPII, Difco, Detroit, Michigan, USA). Cultures were incubated at 37 °C for at least 8 weeks.
Classic phenotyping assays were used. A specific gene probe for detection of M. avium complex was first used in 1990 (Accuprobe, GenProbe Inc., San Diego, California, USA). In 2003, species identification was complemented by 16S rRNA gene sequence analysis.
Sputum samples underwent auramine acid-fast staining and examination for routine bacteriological examination simultaneously. Only sputum samples with more than 25 polymorphonuclear leukocytes and fewer than 10 squamous cells on Gram stain were considered valid samples and processed for bacterial culture. Bacterial chronic lung infection was defined as isolation of the same potentially pathogenic microorganism (PPM) after the diagnosis of bronchiectasis in >50 % of respiratory cultures during the 5-year study period [13].
Statistical analysis
The statistical analysis was performed using SPSS, version 15.0 (SPSS, Chicago, Illinois, USA). All data were expressed as mean (SD) or median (IQR) for quantitative variables and as absolute values and percentages for qualitative variables. The normality of the distribution was assessed using the Kolmogorov-Smirnov test. Prevalence of isolation of NTM was defined as the percentage of patients with at least 1 positive culture for NTM during the study. Patients were divided into 2 groups: NTM-positive patients (those with at least 1 positive sputum culture that showed growth of mycobacteria at any visit) and NTM-negative patients. Depending on the distribution of the variables, the t test or Mann-Whitney test was used to compare 2 means, and the chi-square test (with Fisher exact test if necessary) was used to compare qualitative or dichotomous variables. Variables of clinical interest (according to the current literature or the researcher’s opinion) and those that presented statistically significant differences (p <0.1) in the univariate analysis were included as independent variables in a logistic regression model based on the backward stepwise technique (Wald test). In order to determine the probability of having at least 1 isolation of NTM in different patient profiles, ROC curves were used to dichotomize continuous variables of special clinical interest according to the researchers (age, FVC, and BMI). The cut-off points were those with the greatest capacity to discriminate between patients with and without NTM, as follows: FVC (%predicted, with a cut-off of 75 %), age (with a cut-off of 50 years), and BMI (with a cut-off point of 23 kg/m2).
The independent variables that were finally included were age ≥ 50 years, BMI ≤ 23 kg/m2, macroscopic appearance of the sputum (mucoid vs mucopurulent or purulent), FVC ≥ 75 %, and isolation of PPMs (apart from NTM) from the sputum. The odds ratio (OR) and confidence intervals (95 % CI) for the independent variables were also calculated. P values < 0.05 were considered statistically significant in the logistic regression analysis.
A figure was constructed to show the different patient profiles with their probabilities of having at least 1 isolation of NTM depending of the presence or absence of the dichotomized variables independently associated with the isolation of NTM.