1. Volume 152, Issue 4, Pages 800-809 (October 2017)
Safety and Effectiveness of Clofazimine for Primary and Refractory Nontuberculous Mycobacterial Infection 
Stacey L. Martiniano, MD, Brandie D. Wagner, PhD, Adrah Levin, MPH, Jerry A. Nick, MD, Scott D. Sagel, MD, PhD, Charles L. Daley, MD 
CHEST 
Volume 152, Issue 4, Pages (October 2017)
DOI: /j.chest
Copyright © 2017 American College of Chest Physicians Terms and Conditions

2. Figure 1
Kaplan-Meier curve showing proportion of patients who discontinued clofazimine over time due to all causes: ADRs, culture conversion (cleared), death/lost to follow-up, treatment failure, or who are still taking clofazimine. Twenty-two (20%) were still taking clofazimine at the time of analysis. Overall, 16 patients stopped due to a certain (n = 3), probable (n = 9), or possible (n = 4) ADR after a median of 101 days (95% CI, ). ADRs = adverse drug reactions.
CHEST  , DOI: ( /j.chest )
Copyright © 2017 American College of Chest Physicians Terms and Conditions

3. Figure 2
Kaplan-Meier curve showing proportion of patients who discontinued clofazimine over time according to each category of ADRs. Six patients stopped due to GI ADRs only (median, 86 days [95% CI, ]), five stopped due to GI ADRs plus either skin or neurologic/constitutional ADRs (94 days [95% CI, 3-289]), and two stopped due to skin-related ADRs (167 days [95% CI, ]). One patient each stopped due to allergy, dizziness, and laboratory test abnormality. See Figure 1 legend for expansion of abbreviation.
CHEST  , DOI: ( /j.chest )
Copyright © 2017 American College of Chest Physicians Terms and Conditions

4. Figure 3
Change in FEV1 % predicted (n = 15) with clofazimine for responders (red) and nonresponders (blue). Mixed model analysis of FEV1 in the 2 years prior to and 2 years following clofazimine initiation is shown. Nonresponders had a significant worsening of decline following the addition of clofazimine (mean ± SE, –5.3% ± 1.3% per year [P < .01]; change in slope, –3.3 ± 1.8 [P = .04]). Responders showed an improvement in FEV1 % predicted decline posttreatment (–1.3% ± 1.5% per year [P = .27]; change in slope, 2.3% ± 2.1% [P = .29]). Post-clofazimine slope and change in slope with initiation of clofazimine in responders was significantly different than nonresponders (–1.3% ± 1.5% vs –5.3% ± 1.3% per year [P = .04]; 2.3% ± 2.1% vs –3.7% ± 1.8% [P = .03]), respectively).
CHEST  , DOI: ( /j.chest )
Copyright © 2017 American College of Chest Physicians Terms and Conditions