Preplanned Studies: Predictors for Treatment Outcomes in Patients with Multi-drug Resistant Tuberculosis — China, 2018–2020

Multi-drug resistant tuberculosis (MDR-TB) is caused by the Mycobacterium tuberculosis that is resistant to rifampicin and isoniazid (1). MDR-TB has become a global public health concern, which seriously threatens the realization of the goal of “stopping the tuberculosis (TB) epidemic” by 2035 (1-3). China is a country with a high burden of MDR-TB (1). The coronavirus disease 2019 (COVID-19) pandemic poses significant challenges to control MDR-TB in China (1,4-5). Therefore, a model based on multi-modal data may predict the treatment outcome in the early treatment phase. The results of the study suggest that physicians should pay special attention to dynamic changes of the sputum bacterium and lung images at the intensive stage, and it may help physicians adapt the therapeutic schedule to improve the cure rate.

A multi-center retrospective study was conducted in twenty-three sentinel hospitals selected from sixteen provincial level administration divisions (PLADs) in China. There were 6 hospitals in the east, 3 hospitals in the west, 5 hospitals in the south, 4 hospitals in the north and 5 hospitals in the middle region of China. Among them, there were 8 tertiary hospitals, 10 second-grade hospitals, and 5 community hospitals.

These treatment outcomes were based on World Health Organization (WHO) recommendations, defined as follows: 1) ‘Cured’ meant the course of treatment was completed according to the national guidelines without evidence of treatment failure, and the sputum culture was negative for 3 consecutive times or more at least 30 days after the enhancement period. 2) ‘Treatment failure’ was defined as treatment terminated due to the following reasons: at least two anti-TB drugs in the treatment protocol needed to be permanently changed, no negative conversion after the enhancement period, reversed bacteriological test results in the continuing period after negative conversion, evidence indicating acquired drug resistance to fluoroquinolones or second-line injections, adverse drug reactions, or death directly associated with MDR-TB (2).

All patients with MDR-TB who visited one of the 23 sentinel hospitals between January 2018 and December 2020 were considered as possible subjects. Patients who met the following inclusion criteria were considered: 1) MDR-TB patients, with the definition of MDR-TB as presented by the WHO was used (2). 2) Signed informed consent. Exclusion criteria included the following: 1) Patients with other serious diseases (mental diseases, various cancers, hepatitis patients, serious skin diseases, severe metabolic disease, human immunodeficiency virus/acquired immunodeficiency syndrome, etc). 2) Pregnant or lactating patients and patients with pneumonia, pneumoconiosis or other lung diseases.

MDR-TB treatment was defined as receiving recommended regimen in the national MDR-TB control program (2-3). For each MDR-TB case, sputum smear, sputum culture, drug sensitivity test (DST), physical examination and routine blood counts, biochemical tests, and urinalyses were recorded through monthly examinations. Chest-computed tomography (CT) was performed during the intensive phase and continuation phase to all patients according to the National TB Program (NTP) of China. Sputum specimens were collected and examined through direct smear microscopy for the presence of acid-fast bacilli (AFB) using Ziehl-Neelsen staining. A conventional DST was performed using the agar proportion method on enriched Middle-Brook 7H10 medium against first-line anti-TB drugs (FLDs). The rapid DST (Xpert MTB/RIF^®) assay was performed to detect rifampicin resistance (RR) in smear-positive sputum samples. Second-line anti-TB drugs (SLDs) (1), such as Ethambutol, Streptomycin, Kanamycin, Ofloxacin, P-aminosalicylic acid, and Pyrazinamide, were also included. All positive cultures were submitted to the upper-level laboratory in the prefectural Center for Disease Control and Prevention (CDC).

As to all-round utilization of our clinical data, electronic medical record systems and questionnaires with telephone surveys, and underlying prognostic variables were collected with structured questionnaire using REDCap (version 10.0, Vanderbilt University, Nashville, USA). The questionnaire involved socio-demographic characteristics, risk factors, MDR-TB diagnosis and a history of TB, comorbidities, clinical laboratory indicators, and drug resistance at month zero. Data were cleaned in Microsoft Office Excel (version 2020, Microsoft Corp, Washington, USA) and analyzed with SAS (version 9.4, SAS Institute Inc, Cary, USA). Univariable analysis was done for computing odds ratios (ORs) and their 95% confidence intervals (CIs). The level of significance was assessed by the Wald χ² test. Variables with P<0.05 were entered into a multivariable logistic regression model to examine their independent effects through stepwise deletion of variables. P<0.05 was considered significant using a two-tailed test.

A total of 556 patients completed the therapeutic process, including cured patients (n=389) and patients with treatment failure (n=167). The average age of cured patients was lower than that of treatment failure patients (37.1, 40.2, Z=−3.844, P<0.001). The number of lobes involved in the pulmonary lobe in month 0 of the cured patients were less than that of non-cured patients (3, 5, Z=−6.695, P<0.001). Furthermore, the cavity count in the initial month in cured patients was lower compared to that in failure patients (1, 2, Z=−4.689, P<0.001).

Single factor analysis showed that variables related to the negative prognosis of MDR-TB patients included age, marriage, irregular treatment, lesions involving lung lobes in month zero, cavities counts in month zero, time from TB diagnosis to MDR-TB diagnosis, time from MDR-TB diagnosis to treatment, time from TB diagnosis to MDR-TB treatment, leukocytes, platelets, erythrocytes, sedimentation rate, streptomycin resistance, Ofloxacin resistance, Para-aminosalicylic acid treatment, Pyrazinamide treatment, Moxifloxacin treatment, and Kanamycin treatment (Table 1). In addition, sputum culture results and lung images (lesion absorption, cavity closure) were associated with unsuccessful treatment outcomes, including a sputum culture at month 1, a sputum culture at month 2, a sputum culture at month 3, a sputum culture at month 6, lesion absorption at month 6, and cavity closure at month 6 (Table 1). Finally, age, irregular treatment, time from MDR-TB diagnosis to treatment, erythrocyte sedimentation rate, Ofloxacin resistance, sputum culture month 3, lesions in the pulmonary lobe month zero, cavity count month zero, and cavity closure month 6 were associated with treatment outcome (Table 2).