PREDICTORS OF DISEASE PROGRESSION IN PULMONARY MYCOBACTERIUM AVIUM INFECTION

Abstract

SESSION TITLE: Controversies in Nontuberculous Mycobacteria Treatment SESSION TYPE: Original Investigations PRESENTED ON: 10/21/2019 1:30 PM - 2:30 PM PURPOSE: Pulmonary mycobacterium avium infection (MAI) is associated with substantial morbidity and mortality. Retrospective studies suggest 5-year mortality may be as high as 27-28%. However, the indolent nature of the disease makes decisions to treat difficult. Furthermore, treatment is prolonged and often poorly tolerated, with success rate of only 29% over a 5 year period with a treatment discontinuation rate of 29% due to intolerance. A prediction model is needed to assess probability of disease progression METHODS: We retrospectively reviewed all positive mycobacterial cultures between 2015 and 2017 in three hospitals in New York City and identified cases that met ATS/IDSA criteria for pulmonary MAI. Patients with available imaging were selected and images were reviewed for progression. When results were unclear, a pulmonologist and another researcher reviewed the images. Prior risk factors for progression were chosen based on prior studies, including body mass index (BMI), age, gender, phenotype, tobacco use, emphysema, asthma, prior lung surgery, and prior tuberculosis infection. Likelihood ratios were calculated for categorical variables and p-values were calculated using the chi squared statistic. P-values were calculated for continuous variables using the student’s t test. Logistic regression analysis was performed using variables with statistically significant correlation with disease progression. RESULTS: Mean BMI was lower in progressive disease 20 vs 24 (p=0.0045). Positive likelihood ratios (LR) were higher in the following variables: tuberculosis (2.78, p=0.041), cavitary phenotype (2.8, p=0.0015), and smoking (1.60, p=0.012). Variables not found to be significant included age (mean 70 vs 70, p=0.79), attempted treatment (LR 1.36, p=0.26), male gender (LR 1.3, p=0.20), COPD (LR 1.39, p=0.24), and asthma (LR 0.32, p=0.083). Using logistic regression, a formula was constructed to calculate the probability of disease progression as follows: P(progression)=1/[1+eˆ-(0.59-0.11*BMI+1.23*TB+1.15*smoking+1.32*cavitary)] CONCLUSIONS: A number of variables were found to have statistically significant interaction with disease progression. Treatment was likely not associated with progression because higher-risk cases were treated in this retrospective study. CLINICAL IMPLICATIONS: This equation allows for calculation of probability of disease progression using easily-available clinical characteristics. It serves as an adjunct in decision-making as to whether to treat pulmonary MAI. DISCLOSURES: No relevant relationships by Young Im Lee, source=Web Response No relevant relationships by Jeffrey Lio, source=Web Response No relevant relationships by Reiichiro Obata, source=Web Response