In this study, we identified predictors and derived prognosis nomogram of 9-months MDR-TB treatment success.
During the treatment period, the success rate was 73%, which was slightly below than the success rate recommended by the WHO (75%). The high rate of mortality (15%) and lost to follow-up (7%) due to the several factors such as delay diagnosis or the HIV infection could be explained the low success rate in our sample. Compared to the recently published studies [4–10], our success rate was higher than those reported in Morocco (53.5%) [5], Brazil (60%) [4], and in Bashkortostan region of Russia (67%) [11], comparable to those reported in Baluchistan province of Pakistan (71.6%) [8] and Tanzania (75.7%) [6], and lower than those reported, in Yemen (77.4%) [7], Eastern Taiwan (78.4%) [9], and Rwanda (87.3%) [10]. Possible reasons for these discrepancies are selection bias, differences in MDR-TB regimens, genetic background, and differences between health-care systems.
After 9-months MDR-TB treatment, 7% of our patients were lost to follow-up, which was higher to those reported in Rwanda (0.6%) [10], comparable to those reported in Baluchistan province of Pakistan (7.5%) [8], and lower than those reported recently in several studies [4,6,9] as in Morocco where 34.6% of patients lost to follow-up [5]. Poor adherence to treatment due to the adverse drug reactions was an important factor associated with loss-to-follow-up in our report. Moreover, the higher treatment costs, long duration of treatment, poor knowledge/understanding of MDR-TB, high level of poverty, a low economic status, low family support, and failure health services were previously reported as an important predictors associated with loss-to-follow-up [18,19]. These high rates of loss of follow-up are alarming for health systems, causing a danger to the population, possibly increasing the incidence of extensively drug-resistant tuberculosis. The failure rate was 2% in our study, which was higher to those reported in Rwanda (0.6%) [10], Tanzania (0.6%) [6], and Baluchistan province of Pakistan (1.1%) [8], but lower than those found in Yemen (3.5%) [7], Eastern Taiwan (5.4%) [9], Morocco (6.9%) [5], and Brazil (9%) [4].
We identified the higher BMI, good adherence to treatment, absence of depression and chest pain, and normal platelets count as independent predictors associated with successful MDR-TB treatment. Similar results [7,8] have been found that the lower BMI (£18.5 kg/m2) is an independent factor for unsuccessful treatment in MDR-TB patients. It has been suggested that the lower immunity, poor absorption from gastrointestinal tract or inadequate dosing drugs in underweight patients are possible explanation of association between MDR-TB treatment outcomes and BMI [20].
Not surprisingly, good adherence to treatment was associated with higher success rate with an adjusted odds ratio (aOR) of 34. In our sample only 17% of patients were poor adherents during treatment, among them 80% are classified as unsuccessful treatment outcomes. The poor adherence rate reported in this study was lower than those found in recent meta-analysis (20%) conducted in migrant population [21] where the social risk factors were the main reason. In addition, although drug adverse reactions may explain non-adherence to treatment, we are convinced that the multiplication of awareness campaigns during treatment can help reduce these rates.
The absence of anxious and/or depression was found as important factor to MDR-TB treatment outcomes, no-depressed patients had an increase chance of being cured or complete treatment (aOR = 8.9). Although we have captured depression by asking patients whether they are depressed or not, it would be important to confirm the association between depression and treatment success by measuring it through validated scales such as the PHQ-9 scale [22]. This association suggests the conducting of prospective studies to evaluate the efficacy of psychotherapy management during the treatment of MDR-TB patients.
Chest pain and normal plaquette count were found to be associated with treatment outcomes, patients with no chest pain and patients who had normal platelets count at baseline had an increase chance of being successful treatment (aOR = 3.2 and 1.004 respectively). Although more frequently rifampicin and rarely pyrazinamide induce thrombocytopenia (<150,000 mm3), at inclusion only 8% of our patients had platelets <150,000 mm3.
In univariate analysis, we found that younger patients and negative HIV patients had an increase chance of being successful treatment (OR = 1.03 and 1.93 respectively). Similar results have been reported by studies conducted elsewhere [8,10,12]. Nevertheless, others factors including non-use alcohol misuse [12], non-smoker [5], acid fast bacilli negative sputum [4,7,9], first episode of MDR-TB [4,9], cavitary lesions on X-ray [6,7], resistance to streptomycin and Ethambutol [6] had been reported as independent predictors to successful treatment. While, we not found any association between cavitary lesions and treatment outcomes. These discordant finding suggest the conduct of an international meta-analysis on individual data to definitively identify factors associated with successful treatment.
Based on the number of easily accessible predictors identified in the final logistic model, we constructed a nomogram in a rigorous methodology framework that allowed prediction of individual treatment success with high precision. The nomogram had good discriminatory capacity, and there was good agreement between the nomogram prediction and actual rates of success. Moreover, using the ROC curve methodology, the threshold of 135 points was identified to be the optimal cut-off to select patients with high chance of being successful treatment. However, the nomogram needs to be externally validated on independent samples including non-Guinean patients to establish the generalisability of the model. While, some missing parameters such as diabetes status, smoking and alcohol use, and biomarkers were limits of this study, they might be further used to improve nomogram.