Retrospective Observational Study of Pulmonary Non-Tuberculous Mycobacterial Infections (NTMI): Assessing the Role of Surgery and Trends in Macrolide Resistance

Abstract

Purpose: Pulmonary non-tuberculous mycobacterial infections (NTMI) are persistent infections that are associated with significant health care utilization, morbidity and mortality. Current medical treatment for NTMI has limited efficacy, especially among macrolide resistant isolates. Although macrolide antibiotics are now recommended for prophylaxis among patients with chronic lung disease, there is limited surveillance of macrolide resistance among NTM isolates. For some patients, surgery may be an adjunct to medical treatment, but there is a paucity of controlled outcome data for surgery compared with medical therapy alone especially for the most common NTMI infection with M. avium complex (MAC). Here we perform a retrospective cohort study of MAC infection to assess the role of surgery and in addition assess trends in macrolide resistance amongst all NTM species cultured from patients receiving care in multi-institutional health care system in Boston. Methods: Patients meeting American Thoracic Society (ATS) microbiologic criteria for the diagnosis of NTMI between 2001 and 2016 were identified through a retrospective electronic chart review and separated according to treatment received. We included only patients with radiologically focal MAC disease who received an adequate antimycobacterial regimen. We assessed the effect of surgery on 2-year all-cause mortality, microbiologic response, body mass index, pulmonary function, and radiographic disease through univariate and multivariate logistic regression. The yearly prevalence of macrolide resistance was calculated from positive non-tuberculous mycobacteria (NTM) cultures sent for drug susceptibility testing and analyzed for significant trends using a chi-squared trend test. Results: Of 745 patients meeting microbiologic diagnostic criteria for NTMI, a total of 98 patients met the inclusion criteria and 16/98 (16%) of patients underwent surgical resection for NTMI treatment. Surgically treated and medically treated patients were similar in their baseline demographics, comorbidities, and FEV1. Although statistical power was limited, univariate analysis showed no statistically significant improvement in 2-year all-cause mortality (OR=0.87, 0.18-4.32), lack of microbiologic response (OR=0.49, 0.1-2.41), BMI outcomes (OR=0.35, 0.04-3.45), pulmonary function outcomes (OR=1.44, 0.19-11.1), radiographic outcomes (OR=0.88, 0.24-3.28) or composite outcome (OR=0.47, 0.12-1.79). Multivariate logistic regression for the effect of surgery on composite outcome indicated no statistically significant benefit to surgery (OR=0.45, 0.09-1.57). A total of 460 patients had microbiologic samples sent for macrolide susceptibility testing between 2001 and 2016 with MAC as an increasing proportion of this sample (p<.001) and M. abscessus (MAB) decreasing as proportion of this sample over this time period (p<0.001). The rate of macrolide resistance significantly increased over that time period (p=.03) from 0% in 2001 to 14% (95% CI 7.5-20.8%) in 2016. Conclusion: This is the first case-control study assessing the causal effect of surgery on outcome for patients with MAC pulmonary infection. As surgery for MAC disease is rare, sample size and power were limited, despite this surgery was measured to have a favorable albeit not statistically significant effect on all assessed outcomes except pulmonary function. We found a significant increase in the prevalence of macrolide resistance in NTM, though still rare, surveillance should be continued over time for this emerging issue.