Time and risk factors for death among smear-positive pulmonary tuberculosis patients in the Health District of commune VI of Bamako, Mali, 2016

The Global Tuberculosis (TB) report estimated that 10.0 million people developed TB disease in 2017, with 1.3 million deaths among HIV-negative people and with an additional 300,000 deaths from TB among HIV-positive people [1]. The new End TB Strategy aims to reduce by 95% the deaths due to tuberculosis and by 90% the tuberculosis incidence rate by 2035 compared with 2015 [2]. Identification of the factors associated with dying from TB is useful for evidence-based interventions to achieve these objectives. In Mali, TB mortality was 13 per 100,000 population in 2014, and 11 per 100,000 in 2017 [1, 3], a relative decrease of 15%. To this evolution, the objectives of the End TB strategy may not be achieved in Mali in 2035.

In 2009, a study conducted in Commune VI of Bamako found a lethality of 14% among cases of smear positive pulmonary tuberculosis (SPPTB) [4]. However, to our knowledge, no study has been conducted to identify the risk factors for death from SPPTB in this district. In the literature, these deaths occurred often in the first 2 months of treatment [5‐8]. Many factors could explain this.

Older age [5, 9, 10], male sex [7, 8, 11, 12], unemployment [13], imprisonment [7] have been associated with death in several studies. In a study in South Korea, alcohol consumption several times a week was identified as a protective factor [14]. However, in a systematic review alcohol abuse was a risk factor for death from SPPTB [15]. In other studies, low weight [5, 16], TB / HIV coinfection [5, 6, 12, 13, 17], cases of tuberculosis retirees were at higher risk of dying during treatment [9, 12, 14]. Our aim was to determine the time and the risk factors for death in positive microscopy pulmonary tuberculosis in the health district of Commune VI of Bamako for treatment cohorts from 2011 to 2015.

Our study found that the majority of deaths (53%) occurred during the intensive phase of TB treatment. This is consistent with other studies in different countries [5, 7, 18].

The factor most strongly associated with death in our study was positive HIV status. SPPTB / HIV positive patients were 7 times more likely to die during TB treatment than SPPTB / HIV negative. This is concordant with a systematic review that found that in the context of high incidence of tuberculosis and HIV seroprevalence, HIV was found to be a risk factor, but not in areas with low tuberculosis incidence and HIV seroprevalence [17]. However, early initiation of TB / HIV patients on ARV and cotrimoxazole therapy may decrease this lethality. Data on ARV and cotrimoxazole use were not included in our study. These data would probably improve the current analyses. In Nigeria, TB/HIV coinfection, not receiving ARVs, not receiving cotrimoxazole were independently associated with death of TB patients [6]. Other studies have also found an association between HIV and death among TB patients [5, 8, 9, 12, 14, 19, 20]. Vasankari in Finland found immunosuppression, virtually all due to causes other than HIV infection, and concomitant diseases and their medical treatment to be a risk factor for death of TB patients [11].. Only two patients in his cohort had HIV-coinfection. The second most important factor was the weight < 40 kg at the initiation of treatment. We found that SPPTB patients whose initial weight was below 40 kg were 3 times more likely to die and this finding is similar to those found in studies conducted in different countries [5, 14, 16]. Malnutrition, which can lead to underweight, has been one of the risk factors for death among tuberculosis patients [17]. The third most important factor was age ≥ 45 years. Patients with aged 45 years or older had twice the instant risk of dying during tuberculosis treatment.

This is also supported by many other studies conducted in Africa [5, 8, 9, 12], in Europe [11] and in Asia [10, 13‐15]. The last factor associated was unknown HIV status. SPPTB patients with unknown HIV status had 2 times more instant risk of dying on TB treatment than those with negative / HIV status. Patients with unknown HIV status may contain both HIV positive and HIV negative cases. Given the difference in risk of death compared to HIV negative, there may be enough HIV positive cases to achieve this result. Testing for HIV would give us more edification. Similarly, a study conducted in South Africa, has found unknown HIV status had twice as much risk of dying compared to negative HIV status [9].

This study has some limitations. As our study was based on surveillance data, some informations were not collected which did not allow us to test and adjust on other factors such as unemployment status, alcohol abuse, anemia, rural dwelling and comorbidities (heart disease, end-stage renal disease, diabetes, cancer) that have been considered as risk factors in the literature. However, this work has some strengths. To our knowledge, this study is the first publication to study the time and risk factors for death in patients with positive microscopy pulmonary tuberculosis in the District of Commune VI of Bamako. The study covered a period of 5 years and concerned all patients in the district; providing a good representativeness for the estimation of the factors sought.

The majority of deaths among SPPTB patients on treatment occurred during the first two months of treatment which is the intensive phase of the TB treatment course. Older age (≥ 45 years), low weight at treatment initiation (< 40 kg), unknown and positive HIV status were the independent risk factors associated with SPPTB deaths. We recommend that the National Tuberculosis Program establish a specific system to closely monitor SPPTB patients aged 45 years or older and patients with positive or unknown HIV status, also screen for HIV in patients with unknown HIV status. We also recommend nutritional support for malnourished SPPTB patients.