Determinants of mortality among patients with drug-resistant tuberculosis in northern Nigeria

BackgroundDrug-Resistant tuberculosis (DR-TB) is estimated to cause about 10% of all TB related deaths. There is dearth of data on determinants of DR-TB mortality in Nigeria. Death among DR-TB treated cohorts in Nigeria from 2010 to 2013 was 30%, 29%, 15% and 13% respectively. Our objective was to identify factors affecting survival among DR-TB patients in northern Nigeria.MethodsDemographic and clinical data of all DR-TB patients enrolled in Kano, Katsina and Bauchi states of Nigeria between 1st February 2015 and 30th November 2016 was used. Survival analysis was done using Kaplan-Meier and multiple regression with Cox proportional hazard modeling.ResultsMean time to death during treatment is 19.2 weeks and 3.9 weeks among those awaiting treatment. Death was recorded among 38 of the 147 DR-TB patients assessed. HIV co-infection significantly increased probability of mortality, with an adjusted hazard ratio (aHR) of 2.35, 95% CI: 1.05–5.29, p = 0.038. Treatment delay showed significant negative association with survival (p = 0.000), not starting treatment significantly reduced probability of survival with an aHR of 7.98, 95% CI: 2.83–22.51, p = 0.000. Adjusted hazard ratios for patients started on treatment more than eight weeks after detection or within two to four weeks after detection, was beneficial though not statistically significant with respective p-values of 0.056 and 0.092. The model of care (facility vs. community-based) did not significantly influence survival.ConclusionBoth HIV co-infected DR-TB patients and DR-TB patients that fail to start treatment immediately after diagnosis are at significant risk of mortality. Our study showed no significant difference in mortality based on models of care. The study highlights the need to address programmatic and operational issues pertaining to treatment delays and strengthening DR-TB/HIV co-management as key strategies to reduce mortality.**背景** 耐多药结核病（Drug-Resistant tuberculosis, DR-TB）据估计约占所有结核病相关死亡病例的10%。目前尼日利亚境内关于耐多药结核病死亡影响因素的相关数据仍较为匮乏。2010年至2013年，尼日利亚接受治疗的耐多药结核病队列患者的死亡率分别为30%、29%、15%与13%。本研究旨在明确尼日利亚北部地区耐多药结核病患者的生存影响因素。 **方法** 本研究纳入了2015年2月1日至2016年11月30日期间，在尼日利亚卡诺州、卡齐纳州与包奇州登记入组的所有耐多药结核病患者的人口统计学与临床资料。采用Kaplan-Meier法与Cox比例风险回归模型开展多元回归生存分析。 **结果** 治疗期间死亡患者的平均死亡时间为19.2周，等待治疗患者的平均死亡时间则为3.9周。本次评估的147例耐多药结核病患者中，共有38例发生死亡。人类免疫缺陷病毒（Human Immunodeficiency Virus, HIV）合并感染可显著升高患者的死亡风险，校正后风险比（adjusted hazard ratio, aHR）为2.35，95%置信区间（CI）为1.05~5.29，P=0.038。治疗延迟与患者生存状况呈显著负相关（P=0.000）；未及时启动治疗可显著降低患者的生存概率，其校正后风险比为7.98，95%置信区间为2.83~22.51，P=0.000。对于在检测后8周以上启动治疗，或检测后2~4周内启动治疗的患者，其校正后风险比显示存在获益，但该结果未达到统计学显著性，对应的P值分别为0.056与0.092。照护模式（医疗机构照护 vs. 社区照护）对患者生存状况无显著影响。 **结论** 人类免疫缺陷病毒（HIV）合并感染的耐多药结核病患者，以及确诊后未及时启动治疗的耐多药结核病患者，均面临显著的死亡风险。本研究未发现不同照护模式下患者死亡率存在显著差异。本研究强调，需针对性解决治疗延迟相关的项目与运营问题，并强化耐多药结核病与人类免疫缺陷病毒的联合管理，以此作为降低死亡率的核心策略。