Background: Pellagra is caused by niacin (vitamin B3) deficiency and manifested by a distinctive dermatitis. Isoniazid is critical for treating tuberculosis globally and is a component of most regimens to prevent tuberculosis. Isoniazid may contribute to pellagra by disrupting intracellular niacin synthesis. In 2017, Malawian clinicians recognized a high incidence of pellagra-like rashes after scale-up of isoniazid preventive treatment (IPT) to people living with HIV (PLHIV). This increase in pellagra incidence among PLHIV coincided with a seasonal period of sustained food insecurity in the region, which obscured epidemiological interpretations. Although isoniazid has been implicated as a secondary cause of pellagra for decades, no hypothesis-driven epidemiological study has assessed this relationship in a population exposed to isoniazid. We developed this case-control protocol to assess the association between large-scale isoniazid distribution and pellagra in Malawi. Methods: We measure the relative odds of having pellagra among isoniazid-exposed people compared to those without exposure while controlling for other pellagra risk factors. Secondary aims include measuring time from isoniazid initiation to onset of dermatitis, comparing niacin metabolites 1-methylnicotinamide (1-MN), and l-methyl-2-pyridone-5-carboxamide (2-PYR) in urine as a proxy for total body niacin status among subpopulations, and describing clinical outcomes after 30-days multi-B vitamin (containing 300 mg nicotinamide daily) therapy and isoniazid cessation (if exposed). We aim to enroll 197 participants with pellagra and 788 age- and sex-matched controls (1:4 ratio) presenting at three dermatology clinics. Four randomly selected community clinics within 3-25 km of designated dermatology clinics will refer persons with pellagra-like symptoms to one of the study enrollment sites for diagnosis. Trained study dermatologists will conduct a detailed exposure questionnaire and perform anthropometric measurements. A subset of enrollees will provide a casual urine specimen for niacin metabolites quantification and/or point-of-care isoniazid detection to confirm whether participants recently ingested isoniazid. We will use conditional logistic regression, matching age and sex, to estimate odds ratios for the primary study aim. Discussion: The results of this study will inform the programmatic scale-up of isoniazid-containing regimens to prevent tuberculosis.

Setting: Tuberculosis (TB) treatment facilities in Haiti. Objective: To assess factors associated with loss to follow-up (LTFU) among patients receiving treatment for tuberculosis (TB) in Haiti. Design: We analyzed Haiti's national surveillance data for patients started on anti-tuberculosis treatment from 2011 to 2015 to determine factors associated with LTFU using multivariable logistic regression and describe LTFU in terms of subnational units to target future intervention strategies. We also conducted a survival analysis to estimate hazard ratios of factors associated with time to LTFU. Results: Of 81 490 TB cases reported, 7423 (9.1%) were LTFU during anti-tuberculosis treatment, increasing from 7.1% in 2011 to 10.3% in 2015. Six high-volume facilities had significantly higher rates of LTFU (14.3-31.9%) than the rest of the country, accounting for 18.8% of all TB cases reported, but 41.7% of all LTFU patients. Male sex, previous treatment history, and human immunodeficiency virus infection were associated with higher rates of LTFU. The median time to LTFU was 94 days. Conclusion: A small number of facilities accounted for disproportionately high rates of LTFU. These results identify characteristics of facilities and individuals leading to concentrated interventions to reduce LTFU and improve treatment success.