Introduction

India has one of the highest burdens of multidrug-resistant tuberculosis (MDR-TB) globally, with an estimated 64,000 patients having MDR strains among notified pulmonary TB cases in 2013, which would account for more than 20 percent of the estimated global burden of drug-resistant tuberculosis [1]. The increase in MDR-TB cases within the country has been acknowledged by the Revised National Tuberculosis Control Program (RNTCP) [2] and a multi-faceted response plan has been prepared to combat this challenge [3].

A number of studies have described the prevalence of MDR-TB in a number of different geographical locations across the country [4–11]. The overall magnitude of the MDR-TB epidemic in Mumbai, one of the most known Indian mega-cities, is increasingly being documented. D’Souza et al in 2009 reported epidemic levels of drug resistance amongst new TB cases in central Mumbai [5]. In 2011 a cases series of totally-drug resistant TB (TDR-TB) in the city captured the attention of the medical community and the international media [12,13]. In late 2014, a survey among HIV-infected patients in the greater metropolitan Mumbai reported both alarming levels of drug-resistant tuberculosis among new and previously treated TB cases, as well as complex resistance patterns, especially high levels of resistance to fluoroquinolones and other second line anti-TB drugs [14]. However, as past studies have tended to emphasize the diagnosis, treatment and care of MDR-TB, there is a dearth of literature highlighting variations in resistance patterns over the years among selected populations.

The aim of this study was to document the patterns of drug resistance observed in MDR-TB patients receiving anti-tuberculosis treatment in Mumbai, India and to identify the trends in these patterns over time.

Methods

Results

Drug resistance patterns

Throughout the nine years of the study period, the proportion of study participants having MDR-TB strains was 29.4%. Pre-XDR-TB was the most common form of drug-resistant TB in this Mumbai population at approximately 57% of the entire cohort. The proportions of XDR- and XXDR-TB were 9.7% and 4.1% respectively (Table 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 1. Resistance patterns among MDR-TB patients diagnosed among 8 facilities in greater metropolitan Mumbai, 2005–2013.

https://doi.org/10.1371/journal.pone.0116798.t001

Table 2 details the resistance patterns as recorded for each first and second line anti-TB drug over time. Of note was that 235/340 (69.1%) and 179/287(62.45) of patients tested for Ofloxacin and Moxifloxacin respectively were found to be resistant. Table 3 provides data on the additional resistance recorded among the MDR-TB patients; 35 (10.3%) of patients had developed resistance to 8 or more drugs, beyond the rifampicin and isoniazid while 55 (16.2%) patients were resistant to all fluoroquinolones and injectable drugs.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 2. Resistance profile (first and second-line anti-TB drugs) for MDR-TB patients in greater metropolitan Mumbai, 2005–2013.

https://doi.org/10.1371/journal.pone.0116798.t002

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Table 3. Additional resistance to first and second-line anti-TB drugs among MDR-TB patients in greater metropolitan Mumbai, 2005–2013.

https://doi.org/10.1371/journal.pone.0116798.t003

Trends over time

The trends of the different forms of MDR-TB over time were not found to be statistically significant using a Mantel-Haenszel linear-by-linear association chi-square test (for preXDR; x² = 0.012, p = 0.91, for XXDR; x² = 2.96, p = 0.08) and Poisson regression models. It was observed that the proportion of patients with MDR-TB decreased from 39.4% in the period 2005–2007 to 27.8% in 2011–2013, while the proportion of patients with XDR-TB increased from 6% in 2005–2007 to 10.6% in 2011–2013 and the proportion of patients with XXDR-TB increased from 0% in 2005–2007 to 5.5% in 2011–2013 (Fig. 1).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Trends over time of resistance patterns among DR-TB patients diagnosed in 8 facilities in greater metropolitan Mumbai, 2005–2013

https://doi.org/10.1371/journal.pone.0116798.g001

Moreover, the proportions of patients with additional resistance to ofloxacin, moxifloxacin and ethionamide significantly increased over time. The linear-by-linear association chi-square-values and p-values were x² = 7.87, p = 0.005 for ofloxacin, x² = 9.19, p = 0.002 for moxifloxacin and x² = 4.14, p = 0.04 for ethionamide (Table 2).

Discussion

To our knowledge, this is the first study of multidrug-resistant tuberculosis (MDR-TB) trends over an extended period of time (2005–2013) in greater metropolitan Mumbai.

The six-fold increase in the absolute number of MDR-TB cases (Table 1) identified among the eight study facilities between the first 3-year period (2005–2007) and the last 3-year period (2011–2013), as well as the trend towards increasingly advanced patterns of resistance (i.e. XDR- and XXDR-TB), were both most likely due in large part to improving access to 1^st and 2^nd-line DST. Other possible reasons include increasing awareness amongst clinicians of the possibility of MDR-TB and the need to send specimens for DST. Regardless, these data showing increasingly advanced resistance patterns represents a ‘wake up call’ to all of us working to combat TB in Mumbai and in other parts of the country.

Just as concerning is the finding that the most common form of drug-resistant TB in this study population was not MDR-TB but pre-XDR-TB, which is mainly due to the high levels of baseline fluoroquinolone resistance that are present in Mumbai [14,15]. This predominance of pre-XDR-TB has a number of important implications: firstly, the current treatment strategy that relies on standardized regimens to treat a traditional classification of MDR-TB carries a high risk of being suboptimal treatment. In many cases, a standardized MDR-TB regimen will not contain ‘four drugs likely to be effective’, as recommended by the WHO, if prescribed to a person with disease due to a pre-XDR-TB strain [16]. Not only does a suboptimal treatment regimen reduce the likelihood of a cure, but it can also lead to amplification of drug resistance, such that the pre-XDR-TB strain becomes an XDR-TB strain, which in turn can be transmitted within the household and community.

Secondly, all patients at risk of suffering from active TB disease due to a drug-resistant strain need to have drug susceptibility testing (DST) performed not just against first-line anti-TB drugs, but also against flouroquinolones and other second-line anti-TB drugs during the initial work-up. The current trend in India is to move towards routine use of a molecular TB diagnostic, for example Xpert MTB/RIF (also known as GeneXpert), which can rapidly detect rifampicin resistant strains of Mycobacterium tuberculosis within 2 hours. Although this can be viewed as an important step forward, it should be acknowledged that use of GeneXpert in settings like Mumbai may contribute to suboptimal prescribing practices, since detection of rifampicin resistance alone may mask a diagnosis of pre-XDR or XDR-TB (or worse). In settings like Mumbai, GeneXpert alone will not control the MDR-TB epidemic: TB culture and full DST must accompany its implementation. This will require an increasing state and government investment in laboratory capacity in order to meet the needs of TB patients and clinicians.

Thirdly, inappropriate prescribing patterns related to active TB disease in the private health sector and easy/over-the-counter access to fluoroquinolones both need to be halted as soon as possible. Although high levels of amplification of drug resistance have been demonstrated under well-established DOTS programme conditions [17], additional factors are contributing to the MDR-TB situation in Mumbai, namely a chaotic and unregulated private health sector that operates in parallel with the public DOTS programme, in which suboptimal TB treatment regimens are commonly being prescribed [18–20].

The results we have observed are in a sample population spread out over the cities of Mumbai, Navi Mumbai and Thane. Given the high population density in these settings, where a large proportion of the population lives in extreme poverty in slums, these data are unlikely to be representative of the state of Maharashtra and even less likely of the whole of India. However, the living conditions in Mumbai and the common practices in the private health sector (i.e. the prescribing of inappropriate TB treatment regimens and over-the-counter availability of fluoroquinolones and other drugs having anti-TB properties) are similar to those of other large metropolitan centres in the country.

This study is subject to the usual limitations inherent with retrospective design and data collection. Our sample size is small and may be not fully representative of the population of DR-TB patients within the setting. The extremely limited laboratory capacity in Mumbai prevented us from having a larger sample; until 2011 there was only one accredited laboratory in Mumbai for DST to first-line anti-TB drugs (which was private) and until December 2013 there was no accredited laboratory for DST to second-line drugs. Moreover we believe that there were several cases of MDR-TB cases that did not undergo 2^nd line DST, as the cost had to be covered by the patients themselves, and therefore the representativeness of our study may be further limited. We feel that the numbers of cultures and tests performed during the study period were abysmal, considering the size of the population and the magnitude of the MDR-TB epidemic in Mumbai. As the study was limited to bacteriologically confirmed patients, specific populations, especially small children and HIV-infected patients with culture negative tuberculosis might have been underrepresented in this study. The lack of statistical significance in the trends analysis can be attributed to the small sample size; we believe that the descriptive statistics illustrate the trends over time in drug resistance patterns but they should be interpreted with caution. We were not able to analyze patient characteristics even though we had access to individual patient data rather than aggregated data. However, we had access to limited demographic and clinical data of individual patients. We were also not able to ascertain whether some patients were tested more than once during the study period. However, considering the small size of the cohort in each facility, individual patients were easily identifiable by the treating physicians who kept their records up-to-date and therefore double entries were rather uncommon. Last, we had no access to fingerprinting data for this population, which could have complemented our description of the circulating strains. Despite these limitations we believe that our study adds important data and insights into the description of the drug-resistant tuberculosis epidemic in metropolitan Mumbai.

We emphasize the importance of developing a robust surveillance system for drug resistance in the city and the country rather than depending on infrequent national and limited sub-national surveys. National surveys albeit scientifically acceptable, they may mask significant variance in the magnitude of the epidemic in various localities areas or specific populations. This is especially relevant for vast countries like India and mega-cities like Mumbai where several drug-resistant TB epidemics may exist.

Ongoing research is necessary to provide an evidence base for appropriate public health action. Use of targeted molecular-guided cluster analyses should be considered in order to identify patterns of transmission at the population level. In the face of growing proportions of pre-XDR-TB, XDR- and XXDR-TB, these efforts must be initiated promptly. Official recognition and categorization of TB strains that are resistant to three or more anti-TB drugs (e.g. pre-XDR-TB) are needed, not only to standardize research efforts, but also to help ensure that these patients receive appropriate treatment regimens.

Conclusion

The observed trends in TB drug resistance patterns in metropolitan Mumbai highlight the need for individualized drug regimens, designed on the basis of full drug susceptibility results to first- and second-line anti-TB drugs. A drug-resistant TB case-finding strategy, based solely on rapid molecular techniques such as GeneXpert that identify only TB with Rifampicin resistance, may contribute to suboptimal treatment regimens for the many patients suffering from disease due to pre-XDR-TB strains, and may lead to amplification resistance and spread of further worse forms of drug-resistant TB within the community.

Acknowledgments

The authors wish to acknowledge the strong collaboration of hospital authorities and all staff from the eight facilities involved in this study, especially Rajam Iyer and Aparna Iyer. Finally, we acknowledge the patients suffering from drug-resistant tuberculosis in Mumbai and their struggle to access diagnosis and treatment.