Predictors of unsuccessful interim treatment outcomes of multidrug resistant tuberculosis patients

A major obstacle in the successful control of tuberculosis (TB) is multidrug-resistant TB (MDR-TB), defined as mycobacterium strains resistant to both isoniazid and rifampicin. The management and treatment of MDR-TB is complex and it is difficult to achieve favorable treatment outcomes as compared to drug-sensitive TB, even under optimal circumstances. In-part, this is attributed to the lengthy treatment of MDR-TB patients with comparatively less effective, more toxic and costly regimens that contain combinations of first line (FLDs) and second line anti-TB drugs (SLDs) [1‐3]. Moreover, scarcity of an evidence base from randomized controlled trials, inadequate number of SLDs, the absence of political commitment, the limited number of experts and laboratories and the sale of anti-TB drugs in the private sector of high-TB burden countries, have also contributed toward the unsuccessful treatment outcomes seen among MDR-TB patients [2‐5].

MDR-TB treatment is usually in the duration of 18 to 20 months and at times longer. Therefore, final treatment outcomes can only be evaluated after two to three years of patient registration on the MDR-TB treatment register. During the course of treatment the National Tuberculosis Programme (NTP) managers and healthcare authorities/providers require an indication of the treatment effectiveness among the enrolled patients [6]. This is more important when a Programmatic Management Unit for Dug-resistance TB (PMDT) site is newly established. Therefore, the WHO has set various interim benchmarks for monitoring drug-resistant TB programmes in order to take appropriate and timely action to control the burden of TB [6]. Of these, sputum culture negativity during the early months of MDR-TB treatment is widely used as a reliable interim indicator of non-infectiousness (i.e. microbiological endpoint) and for tracking clinical progress and the effectiveness of treatment [7, 8]. Based on the assumed predictive value of sputum culture conversion for end TB treatment outcome, this is used as an early microbiological end point in phase II clinical trials of TB treatment [9]. Yew et al. have reported that sputum culture conversion within the first three months of treatment was predictive of complete cure in 100% of MDR-TB patients. Sputum culture conversion at two months of treatment has also been reported as a predictor of successful treatment outcomes among MDR-TB patients [10]. Similarly, those patients who did not achieve sputum culture conversion after two months of treatment were 2.7 times more likely to develop unsuccessful treatment outcomes than those who converted within the first two months of treatment [11]. A recently conducted study which included 1712 MDR-TB patients from two separate cohorts, also found statistically significant positive associations between sputum culture conversion at two months of treatment and successful outcomes among patients with unknown HIV status. However, the study reports that overall association of sputum culture conversion with a successful treatment outcome was significantly greater at six months [9]. In addition to sputum culture conversion, the proportion of patients who are deceased at six months is also commonly used as an interim indicator of early treatment response. Figures on loss to follow-up at six months due to a variety of reasons such as: symptoms relief, adverse drug reactions, patient’s lack of awareness about the total treatment duration [12] is also helpful in the assessment of adherence to treatment regimens, and overall performance of the programme [6] (refer to Additional file 1: Appendix S1 section for study schema).

Pakistan has the fourth highest MDR-TB burden of any country in the world and is the highest in the Eastern Mediterranean Region of the WHO (EMRO) [13]. The first National Drug Resistance Survey (DRS) of Pakistan was completed in September 2013. According to the results of that DRS, the MDR-TB prevalence amongst new patients of pulmonary TB (PTB) was 3.7% and 18.1% in retreatment PTB patients [14]. The comparatively high prevalence of MDR-TB in retreatment TB cases in Pakistan is suggestive of defective anti-TB treatment in the past, and could be attributed to divergent guidelines and practices, poor patient adherence to anti-TB treatment, over-the-counter (OTC) sales of medicines and the poor quality of anti-TB drugs in Pakistan [15]. This also generates concerns about the suitability of the WHO Category II regimen in Category I failures. According to the WHO guidelines, in a country with a high burden of drug-resistant TB such as Pakistan, it is recommended that patients who fail Category I treatment should be evaluated for drug resistance, rather than putting them on the Category II regimen [15, 16]. It is a grim reality that despite the current epidemiological picture of MDR-TB in Pakistan, the projected government health expenditures (GHE) for MDR−/TB remains at less than 1.5% of Gross Domestic Product (GDP) per capita. Further, this is the lowest amongst the high-TB burden countries [13]. To date, several studies from different areas of Pakistan have documented the number of incident cases and treatment outcomes of MDR-TB patients, with no evidence of uniformity in the results [3, 17‐20]. Moreover, no study from Pakistan has reported interim treatment outcomes of MDR-TB patients. The authors are not aware of any study that has reported the factors associated with unsuccessful interim treatment outcomes. Besides this, previous studies have not evaluated treatment outcomes in a cohort of patients registered at the newly established PMDT site of the Bahawal Victoria Hospital; that serves a large population of the Southern part of the Punjab province of Pakistan. Considering the recommendations of the WHO and the gaps in the existing literature, it is important to conduct cohort analyses of all MDR-TB patients in the local setting. Therefore, this study was conducted to evaluate the interim treatment outcomes of MDR-TB patients’ registered at a PMDT site of the NTP (Pakistan) established at Bahawalpur, Punjab, Pakistan. A vigilant analysis of the management and treatment status of MDR-TB patients, with an emphasis on the predictors of unsuccessful interim treatment outcomes, provides some understanding of progress and the way forward towards achieving the goal of a treatment success rate ≥ 90% for TB.

This study set out to evaluate the interim treatment outcomes of MDR-TB patients in in South Punjab province of Pakistan. By the end of the study period, interim treatment outcomes were available for 80 MDR-TB cases, of whom 32 (40%) were classified under the unsuccessful interim treatment outcome category. In terms of unsuccessful interim treatment outcome, 12 (15%) patients died, while nine (11.3%) were lost to follow-up by six months. Based upon these results, it is evident that the study site has not achieved the treatment success rate set out in the goals of the Stop TB Strategy (75%), End TB strategy (90%) and the United Nations Sustainable Development Goals (80%) [13].

This failure is contributed by over one-quarter of the patients who died or were lost to follow-up. In accordance with the findings of previous reports [2, 25‐27] there was a high loss to follow-up rate (11.3%) in the current study. This is alarming because treatment was free and there was a consistent and reliable supply of drugs, monthly food and transport allowances for both patients and treatment supporters, health education and counselling by the pharmacist and psychiatrist, home visits by the treatment coordinator and weekly checkups by the medical officer at the nearest health center. This support was provided to all patients, and still the “lost to follow up” rate is sub-optimal. Additionally, in the case that a patient defaulted for more than four weeks, patient tracing was also performed. Despite these efforts the literature points to the following reasons for loss of patients to follow-up; toxicity of the MDR-TB treatment regimen, poor disease knowledge, rural domicile, lower socioeconomic status, lower education levels, previous history of anti-tuberculosis treatment, lack of family support and dissatisfaction with health care worker attitudes [12, 26, 28, 29]. Strengthening of the Pakistan healthcare system with proper training of treatment supporters and comprehensive counseling of patients with “patient friendly” follow up services may help to reduce the rate of loss to follow-up in Bahawalpur.

The mortality rate (15%) observed in this study is comparable to previous studies from Pakistan, Uzbekistan and the Russian Federation, [3, 27, 30]. In contrast, lower mortality rates have been reported elsewhere [2, 25, 26, 31]. In Bulgaria, during 2009–2010, mortality rates of 38% and 53.2% were observed among patients who were lost to follow-up [32]; therefore, masking of deaths due to high rates of loss to follow up of patients could be a reason for the lower mortality rates in the aforementioned studies. The higher mortality rate in this study, based upon the literature, could be attributed to delayed detection of the disease or treatment initiation, lower education, greater number of previous TB episodes, history of diabetes, poor bacteriologic response and inadequate treatment [32]. Therefore, new strategies for appropriate MDR-TB detection and management should be implemented at the study site.

An alarming finding of our study was positive association between ofloxacin resistance and the unsuccessful interim treatment outcome. This finding is of major concern and highlights the importance of fluoroquinolones in MDR-TB treatment regimens and this study aligns with previous reports from Pakistan [3, 12, 17, 18] and other countries [26, 33] supporting an increase in fluoroquinolone resistant M. tuberculosis strains. This high-level of resistance to ofloxacin in the Pakistani population is clearly concerning, and may be directly attributed to over the counter (OTC) sale of antibiotics and extensive, unregulated and irrational use of fluoroquinolones in particular; especially in pneumonia and uncomplicated respiratory-tract infections [34].

A very high rate of ofloxacin resistance was observed in the cohort of patients in this study. The previously reported high cross-resistance between ofloxacin and levofloxacin (87%) among MDR-TB patients in five MDR-TB high burden countries including Pakistan, calls into question the national guidelines’ recommendation of indiscriminate use of levofloxacin in treating MDR-TB patients [23]. In order to improve MDR-TB patients’ adherence and treatment outcomes, the WHO has recently recommended a new, shorter regimen for the treatment of MDR-TB patients. This regimen comprised of 4–6 months of kanamycin, moxifloxacin, prothionamide, clofazimine, pyrazinamide, ethambutol and high-dose isoniazid followed by 5 months of moxifloxacin, clofazimine, pyrazinamide and ethambutol [35] has produced excellent results under research conditions, in various settings [36, 37]. As the new regimen contains moxifloxacin instead of levofloxacin, the low prevalence of moxifloxacin resistance among rifampicin resistant TB patients in Pakistan (13.8% at 0.5 μg/ml which reduced further to 1.4% at 2 μg/ml), and very low rate of cross resistance between moxifloxacin and ofloxacin (7%) [33] makes the new regimen the most suitable option for treating ofloxacin resistant MDR-TB patients, and should be adopted without further delay [38]. Furthermore, the rational use of fluoroquinolones must be enacted immediately, to assist in preserving its effectiveness and the ongoing use in MDR-TB treatment regimens.

Sputum culture negativity during the early months of MDR-TB treatment (especially after two months) - is widely adopted as a reliable interim indicator of non-infectiousness (i.e. microbiological endpoint) and for monitoring and predicting ongoing progress, effectiveness and ultimately success of treatment [7, 8]. Successful treatment outcomes have been reported in MDR-TB patients who achieved sputum culture negativity after two months of treatment [10, 39]. In our study, sputum culture positivity even after two months of MDR-TB treatment was positively associated with unsuccessful interim treatment outcomes. Similarly unsuccessful treatment outcomes have been reported in MDR-TB patients with positive sputum culture results after two months of treatment [40]. According to a study from Hong Kong, culture negativity after two months of MDR-TB treatment was 100% and 52.3% prognostic of treatment success and failure, respectively [10]. In addition, a secondary analysis of data from two observational cohort studies and a study from the Dominican Republic has reported better treatment outcomes in MDR-TB patients with negative sputum culture results after two months of treatment [11, 41]. The results from our study also advocate that early negativity of sputum culture is a positive predictor of successful interim treatment outcomes, and vice versa. This may help clinicians in the early detection of patients who are at greater risk of unsuccessful treatment outcomes. Moreover, our study results also advance credibility for the clinical utility of checking sputum cultures as a method of evaluating the treatment effectiveness.

In the current study, at baseline visit, a notable proportion of patients had above normal serum creatinine levels. Upon cross tabulation, the authors found a statistically significant positive association between previous use of aminoglycosides and the above normal serum creatinine level at baseline visit (p-value = 0.038). This finding supports the previously reported nephrotoxic effects of these drugs [42, 43]. An important and unique finding of our study was the positive relationship between the baseline above normal serum creatinine level and unsuccessful interim treatment outcomes. Injectable SLDs play a central role in the management of MDR-TB [44, 45]. Resistance to injectable SLDs has been widely reported as a predictor of unsuccessful treatment outcomes among MDR-TB patients [45, 46]. Nephrotoxicity in MDR-TB patients is usually managed by reducing the dose or frequency of injectable SLD administration or suspending the causative agent [6, 47, 48]. The authors suspect that for preventing further damage to the kidney in patients with nephrotoxicity at the baseline visit, initiating treatment with low doses of injectable SLDs in this group of patients might result in the suboptimal concentrations and in turn, interim unsuccessful treatment outcomes. These patients require more detailed tailoring and significant alterations in their treatment regimens and this may complicate therapy and reduce adherence to their MDR-TB treatment. Therapeutic drug monitoring in this group of patients would provide an opportunity for clinicians to individually manage each patient’s anti-TB therapy.

As with any study, this research has raised questions in the process of answering others. From a clinical viewpoint studies are needed to confirm the finding of positive association between above normal baseline serum creatinine levels and unsuccessful interim treatment outcomes. From a health system viewpoint, there are deficiencies in Pakistan with a lack of clinical hospital pharmacists [49] and there is a need to develop dose modification protocols and incorporate the role of the hospital pharmacist into the adjustment of anti-TB drug dosing; especially in patients with previously compromised renal function. This could be evaluated as an intervention study. Likewise, the high level of loss to follow-up needs to be further explored, based on the fact that comprehensive support was provided to the participants in this study and it is important to understand in a qualitative way why this follow-up system has failed so dramatically [12].

Regardless of the fact that this study was performed with a robust and standardized methodology adopted from the WHO, our study still has a number of limitations. First, findings of this study may not be generalized for the whole of Pakistan because the sample was drawn from a single PMDT site in the southern part of the Punjab province of Pakistan. Nevertheless, the site is well established with all healthcare facilities available, and policies and practices at this site are expected to be comparable to other PMDT sites in Punjab specifically, but also Pakistan in general. Second, due to the retrospective nature of the study, (some important data was missing in the patient notification forms) it was not possible to include findings of baseline lung cavitation which has been previously reported as an important predictor of unsuccessful treatment outcome in TB [3, 32].

Despite free treatment and programmatic efforts to improve patients’ adherence with MDR-TB treatments, the high rate of unsuccessful interim treatment outcomes in the current cohort of Pakistani patients is concerning. Educational and psychosocial support interventions and decentralizing of treatment services may help to decrease the loss to follow-up rate at the study site. Having a focus on enhanced clinical management to patients with ofloxacin resistance and above normal serum creatinine levels at the baseline visit may improve the treatment outcomes at the center. In future, wide scale studies are needed to confirm the finding of a positive association between above normal baseline serum creatinine levels and unsuccessful interim treatment outcomes in MDR-TB patients.