A risk score for prediction of poor treatment outcomes among tuberculosis patients with diagnosed diabetes mellitus from eastern China

Abstract

Persons living with diabetes (PLWD) with newly diagnosed tuberculosis are at greater risk of poor treatment outcomes. Identifying and prioritizing high-risk subgroups of PLWD and tuberculosis for tuberculosis programs to target has been rarely performed. We investigated risk factors for poor tuberculosis treatment outcomes among PLWD and developed a predictive risk score for tuberculosis control prioritization. Among PLWD diagnosed with tuberculosis, demographic, clinical, and tuberculosis treatment outcome data were collected. Poor treatment outcomes included treatment failure, death, default, and transfer. Multivariable logistic regression modeling was used to analyze risk factors of poor treatment outcomes. Risk scores were derived based on regression coefficients to classify participants at low-, intermediate-, and high-risk of poor treatment outcomes. Among 335 PLWD newly diagnosed with tuberculosis, 109 were cured and 172 completed treatment. Multivariable logistic regression found that risk factors of poor treatment outcomes included bacteriologically-positivity, low body mass index, no physical activity, and pulmonary cavitation. Rates of poor treatment outcomes in low- (0–2), intermediate- (3–4), and high-risk (5–8) groups were 4.2%, 10.5%, and 55.4% (Ptrend < 0.0001), respectively. The risk score accurately discriminated poor and successful treatment outcomes (C-statistic, 0.85, 95% CI 0.78–0.91). We derived a simple predictive risk score that accurately distinguished those at high- and low-risk of treatment failure. This score provides a potentially useful tool for tuberculosis control programs in settings with a double burden of both tuberculosis and diabetes.

Tuberculosis (TB) treatment challenges in TB-diabetes comorbid patients: a systematic review and meta-analysis

“…Out of 49 studies included in this meta-analysis, nine studies were from India [ 24 , 27 , 30 , 32 , 36 , 44 , 47 , 57 , 58 ], four from South Korea [ 1 , 29 , 31 , 49 ], three from Taiwan [ 7 , 23 , 40 ], two from Ethiopia [ 2 , 50 ], four from Mexico [ 28 , 34 , 42 , 43 ], six from China [ 17 , 51 , 52 , 54–56 ], one from Thailand [ 25 ], two from Malaysia [ 46 , 62 ], two from Nigeria [ 35 , 59 ], two from Brazil [ 37 , 45 ], two from Georgia [ 38 , 41 ] and one each from Indonesia [ 26 ], Canada [ 33 ], Pakistan [ 39 ], London [ 61 ], Armenia [ 48 ], Nepal [ 53 ] and Albania [ 60 ]. Among these, the study design of 23 studies was retrospective cohort, 19 studies were prospective, one study was a cross-sectional study and one study was a case-control study.…”

Machine learning algorithms using national registry data to predict loss to follow-up during tuberculosis treatment

“…Among the variables selected, prior TB, consumption habits (alcohol, tobacco, or drug use), age (adult and elderly), biological sex, HIV infection, and schooling level were the risk factors that most contributed to an LTFU during TB treatment. Some of these characteristics have been explored and linked to unfavorable TB treatment outcomes through the relationship with poor therapy adherence, LTFU, and treatment discontinuation [ 20 – 27 ]. It is important to highlight that our study identified history of prior TB as the variable with the most significant impact on the model’s ability to predict LTFU.…”

“…In a previous study, a similar score was developed to predict unfavorable anti-TB treatment outcomes in people living with diabetes from China, however using clinical and radiologic data [ 23 ]. Another study from Mexico developed an algorithm to predict mortality, failure, and drug resistance in newly diagnosed TB patients with clinical features and laboratory tests [ 27 ]. In contrast, our score could be applied in patients with or without diabetes, by utilizing only clinical information, without the necessity of laboratory data or radiographic exams.…”

Machine learning algorithms using national registry data to predict loss to follow- up during tuberculosis treatment

“…Among the variables selected, prior TB, consumption habits (alcohol, tobacco, or drug use), age (adult and elderly), biological sex, HIV infection, and schooling level were the risk factors that most contributed to an LTFU during TB treatment. Some of these characteristics have been explored and linked to unfavorable TB treatment outcomes through the relationship with poor therapy adherence, LTFU, and treatment discontinuation[ 11 , 22 – 28 ].…”

“…In a previous study, a similar score was developed to predict unfavorable anti-TB treatment outcomes in people living with diabetes from China, however using clinical and radiologic data[ 24 ]. Another study from Mexico developed an algorithm to predict mortality, failure, and drug resistance in newly diagnosed TB patients with clinical features and laboratory tests[ 28 ]. In contrast, our score could be applied in patients with or without diabetes, by utilizing only clinical information, without the necessity of laboratory data or radiographic exams.…”