Adverse drug reactions and risk factors for discontinuation of multidrug-resistant tuberculosis regimens in Gujarat, western India

Introduction

In India in 2015, an estimated 2.5% of new cases and 16% of previously treated cases had multidrug-resistant tuberculosis (MDR-TB).^[1] A study reported that the percentage of adverse drug reactions (ADRs) among patients with MDR-TB stands at 58%.^[2] About 18 adverse events have been reported in various other studies.^[3] Not much difference has been observed in previous studies with respect to ADRs as far as the comparison between patients with human immunodeficiency virus (HIV) coinfected with TB and those with TB but HIV-non-infected is concerned.^[3] Hepatotoxicity due to pyrazinamide and ethambutol has been reported previously.^[4] However, another study reported that hepatotoxicity while on an MDR-TB regimen did not result in poor treatment outcomes.^[5] We studied the ADRs and predictors for discontinuation of the MDR-TB regimen.

Methods

Study design and setting

This cross-sectional study was conducted in Bhavnagar district of Gujarat, western India. Bhavnagar district caters to a large number of patients with TB of the same and the surrounding districts of Amreli and Botad. It is situated 255 km southwest of Ahmedabad. The study was conducted through the District TB Centre (DTC) with patients with MDR-TB registered over 3 months (September to November 2016). All 135 patients with MDR-TB registered with the DTC of Bhavnagar district were included in the study.

Inclusion criteria

All patients with MDR-TB above 10 years of age, of either gender, who were on drugs for MDR-TB for at least 2 months (including defaulters), who gave written informed consent were included in the study. The patients were diagnosed as MDR-TB using Cartridge Based-Nucleic Acid Amplification Test (CB-NAAT) at the designated microscopy centre of Sir Takhtsinhji Hospital (government district hospital) of Bhavnagar.

Data collection

Data were collected by a personal interview through home visits of the patients. The home visits in Bhavnagar corporation area were done along with a tuberculosis health visitor (TBHV). The home visits in Bhavnagar rural area were done along with senior treatment supervisors (STSs).

The investigators filled up a one-page questionnaire on ADRs through home visits. The patients’ case papers were also checked for additional data such as laboratory investigations. The questionnaire consisted of sociodemographic variables, addiction, HIV status, duration of treatment, drugs currently on and ADRs, the reason for stopping the particular drug, previous history of default, the reason for default, etc. Tobacco consumption and alcohol drinking were asked for through a single question but were not quantified. The questionnaire was translated into Gujarati for ease of data collection. The translated questionnaire was validated with concurrent validity. The face validity and content validity were checked by a group of experts and the suggested changes were incorporated. The questionnaire was then pilot tested and necessary corrections done.

Variables

The primary outcome variable was descriptive in nature, wherein the ADRs among patients with MDR-TB have been described. The dichotomous variable of having or not having an ADR was compared with the input variables. The secondary outcome variable was whether patients with MDR-TB were compelled to discontinue treatment due to ADRs, which was expressed as a percentage. Another secondary outcome variable was a dichotomous variable of whether or not patients with MDR-TB discontinued their regimen, which was compared with the input variables. The input variables were sociodemographic variables, weight, duration of treatment, duration of ADRs, HIV status, comorbid conditions and addiction.

Statistical analysis

Independent samples t-test was applied for comparison of a continuous variable among the two groups. Chi-square test was applied for comparison of a pair of categorical variables. Mid-P exact p-values were estimated where the assumptions of chi-square tests were not valid (80% of expected cell values were <5 or 1 of the expected cell values was <1). Open Epi software version 2.2 was used to do all the statistical analyses.^[6] Multiple logistic regression was done by the ‘enter’ method, considering discontinuation of the MDR-TB regimen as the dependent variable and entering the following variables as the independent variables (predictors predicting discontinuation) in Step 1: age in years, years of schooling, being single (unmarried), duration of treatment, being in the intensive phase of regimen, presence of ADR and per capita income. Adjusted odds ratios with 95% CI were calculated. A p value <0.05 was considered significant.

Ethical considerations

Approval for carrying out the study was obtained from the Institutional Review Board of the Government Medical College Bhavnagar, Gujarat. Permission for carrying out the study was obtained from the District TB Officer of Bhavnagar district.

Socioeconomic classification

For socioeconomic status, Modified Prasad's classification was used taking All India Consumer Price Index for Industrial Workers value of 268 for August 2016 for Bhavnagar district.^[7],[8]

Results

Of 135 patients, 18 patients had taken MDR-TB drugs for <2 months; 9 patients were <10 years of age and 14 patients did not give written informed consent to participate in the study. Therefore, 94 patients with MDR-TB were included with a mean (SD) age of 35.9 (14) years. The mean (SD) years of schooling were 4.9 (4) years. The mean (SD) weight of the patients was 46.9 (11) kg. The mean (SD) per capita income of the patients was ₹1927 (1330).

Among the 94 patients with MDR-TB, 56.4% were below 35 years of age; 68.1% were males; 60.6% were literate; 79.8% were married and 59.6% belonged to the upper socioeconomic status (class IV and V) of Modified Prasad's classification. Of the 94 respondents, 34% were homemakers; 28.7% were daily wage labourers and 17% were not doing any work due to the physical weakness caused by MDR-TB [Table - 1].

Table 1: Sociodemographic data of the patients (n=94)

On average, the time since diagnosis of MDR-TB was 14.2 months and patients were on MDR-TB drugs for 17.2 months. A majority of them (95%) had taken Cat-I/Cat-II drugs previously, while only 2% had taken MDR-TB drugs before being put on the current regimen. Among the 94 patients, 49% were in the intensive phase; 43% were in the continuation phase and 8% were taking the Modified Cat-IV regimen. Regarding addictions, 20.2% were tobacco chewers, 5.3% were tobacco smokers and 2.1% were addicted to alcohol. Of the 5 patients presenting with comorbid conditions, 3 had diabetes, 1 had HIV and 1 had chronic renal failure. Of the 3 patients with abnormal blood investigations, 1 had a serum creatinine level of 4 g/dl and the other two had serum uric acid level of 10.9 g/dl.

The prevalence of ADRs among patients with MDR-TB was 73% with 69 patients having at least one ADR. [Table - 2] lists the ADRs of MDR-TB drugs among the respondents. Only 1 of the 94 patients was coinfected with HIV and was on antiretroviral therapy (ART; tenofovir, lamivudine and efavirenz). The CD4 count of the patient was 94 cells/cmm. The patient had a rash, tingling, numbness, vision changes, fatigue and dizziness as ADRs. Since only one patient had HIV coinfection, no comment could be made on the difference in ADRs due to ART.

Table 2: Adverse drug reactions in patients (n=94) with multidrug-resistant tuberculosis (multiple answers)

The patients had ADRs for a mean (SD) duration of 15.3 (10) months. Among the 94 patients with MDR-TB, 7.4% were compelled to think of discontinuing the treatment due to ADRs; 8.5% had discontinued their Cat-I/Cat-II regimen in the past; 11.7% had discontinued MDR-TB drugs in the past; 13.8% had their drug regimen changed due to ADRs but 94.7% had good adherence to their present regimen (took at least 80% of their doses till date). ADRs were the reason for 75% of patients discontinuing their Cat-I/Cat-II regimen in the past and 64% of patients discontinuing the MDR-TB regimen in the past [Table - 3].

Table 3: Details of adverse drug reactions (ADRs; n=94)

The presence/absence of ADRs was not statistically significantly associated with age group, gender, literacy status, marital status, socioeconomic status, drug regimen, presence/ absence of a comorbid condition, adherence, tobacco chewing, tobacco smoking and alcohol addiction.

Patients with MDR-TB who had an ADR had a mean (SD) age of 36.7 (13.9) years compared with those who did not have an ADR whose age was 33.6 (14.6) years. This difference was not statistically significant (p=0.35).

Similarly, there was no statistically significant difference among those having ADR(s) and those not having ADR(s) for years of schooling, per capita income, weight, time since diagnosis of MDR-TB and time since starting treatment of MDR-TB.

There was no statistically significant difference among those discontinuing and those not discontinuing their regimen for age, years of schooling, per capita income, weight, time since diagnosis of MDR-TB, time since starting treatment of MDR-TB and time since experiencing ADR(s).

Adherence to an MDR-TB regimen was associated with a 97.7% reduction in risk of discontinuation than those who did not adhere to their regimen. Tobacco chewers were four times more likely to discontinue their MDR-TB regimen than those not chewing tobacco. Those who were compelled to think of discontinuing their regimen due to ADR(s) were seven times more likely to actually discontinue than those who were not compelled to think so. Discontinuation of the MDR-TB regimen was not statistically significantly associated with age group, gender, literacy status, marital status, socioeconomic status, drug regimen, presence/absence of comorbid conditions, tobacco smoking, alcohol addiction, presence/absence of ADRs and change in drug regimen due to ADRs.

On multiple logistic regression, none of the predictors were found to be statistically significantly associated with discontinuation of the MDR-TB regimen [Table - 4].

Table 4: Multiple logistic regression for predictors of discontinuation of multidrug-resistant tuberculosis regimen* (n=94)

Discussion

Each drug has a pharmacological action and has some side-effect or ADR attached to it and the treatment regimen of TB is no exception. We assessed the ADRs of patients with MDR-TB to highlight the difference in those with HIV coinfection. We found that 73% of patients with MDR-TB had ADRs. A study in Tamil Nadu found the incidence of ADRs to be 58%.^[1] A study done in Lucknow found this to be 47%.^[2] A study conducted at the Tuberculosis Chest Institute of Bengaluru reported that the majority of their patients had one or more ADRs.^[3] A study in Aurangabad (Maharashtra) reported that nearly 34% of patients received second-line therapy with ADRs.^[4] A study in Nagpur, Maharashtra, found 50% of patients with MDR-TB had ADRs.^[5] Another study from Maharashtra (Mumbai) reported that 28.8% of patients had ADRs,^[6] whereas another study in Mumbai reported 45% ADRs in patients with drug-resistant TB.^[7] A study in Ahmedabad conducted at a tertiary care hospital found it in 57.1% of patients.^[8] Another study from Ahmedabad reported that 93.8% of patients with MDR-TB had ADRs.^[9] A study from New Delhi revealed that all patients with MDR-TB reported ADRs, from as minor as loss of taste to as severe as a psychotic reaction.^[10] Bhushan et al. in 2014 reported 81.2% of ADRs among patients with drug-resistant TB.^[11] Patel et al. in 2015 found that in seven districts of Gujarat 54.9% of patients with MDR-TB patients developed ADRs in 2 years of follow-up.^[12] The average duration of ADRs was 60–90 days.^[12] We found the mean duration of ADRs to be 15.3 months. Thus, various studies in India have reported the prevalence of ADRs to be in the range of 29%–94%.

Only one patient was reactive for HIV in our study, which was also the case in another study done in Ahmedabad.^[9] A study from Lucknow mentioned that none of their 98 patients with MDR-TB had HIV coinfection.^[2] Hence, to study the difference in ADRs among HIV-infected and non-infected patients with MDR-TB, we need a larger sample size.

A review published by the Indian Council of Medical Research mentions overlapping ADRs in patients put on both anti-TB drugs and ART.^[13] A similar finding was reported by a study from Mumbai, where they did not find any statistically significant difference in the occurrence of ADRs between patients with MDR-TB taking either first-line or second-line ART.^[14] However, they highlighted the concern of co-administration of tenofovir with aminoglycosides and capreomycin increasing the risk of renal toxicity; co-administration of efavirenz and cycloserine increasing the risk of psychiatric adverse effects; and co-administration of stavudine and ethionamide, cycloserine or high-dose isoniazid increasing the risk of peripheral neuropathy.^[14]

Gastrointestinal symptoms were the most commonly reported ADRs in many published studies.^{[2],[3],[4],[5],[8],[9],[10],[11],[12],[14]} Peripheral neuropathy (tingling) was the second most common in a study in Mumbai.^[14] Headache was the second most common ADR reported in the research conducted in Ahmedabad,^[9] which was also the case in our study. We found tingling, headache, numbness, dizziness and nausea to be the most common ADRs.

ADRs were the main reason for discontinuation of treatment of TB in our study. A qualitative study conducted to assess the outcomes of patients with MDR-TB in Nagpur found that 30% of patients felt adverse drug effects as an important barrier to treatment adherence.^[15] A study from Bengaluru reported 1.7% of patients discontinuing treatment and 10.5% of patients requiring a change in drug regimen due to ADRs.^[3] Another study from Ahmedabad stated that 33% of patients required a change in drug regimen due to ADRs.^[8] A study from Punjab reported that 12.1% of patients had to change their drug regimen due to ADRs.^[11] The study from Lucknow highlighted that 17.4% of patients required a change in their regimen due to ADRs.^[2] The study in Aurangabad stated that 9.4% of patients required a change in regimen due to psychiatric and ototoxic ADRs.^[4] The study in Mumbai found that 20% of patients required a change in drug regimen due to moderate-to-severe ADRs.^[7] We found that 13.8% of patients with MDR-TB required to change their drug regimen due to ADRs.

A study from Punjab found almost half the male patients (42%) had a history of drug and/or alcohol abuse, 11% were tobacco smokers and 28% were tobacco chewers.^[11] We found that 20.2% of patients were tobacco chewers, 5.3% were tobacco smokers and 2.1% were addicted to alcohol. This difference was probably due to the different addictions prevalent in these two states, with alcohol being banned in Gujarat and tobacco chewing being the prevalent practice in Bhavnagar district.^[16] Prasad et al. reported no statistically significant difference in patients suffering from ADRs with respect to gender, drug addiction, alcohol addiction and age group.^[2] Our study also did not find any difference between patients with or without ADRs with respect to their age group, gender, tobacco chewing, tobacco smoking and alcohol addiction.

Limitations

As our study was done in one district, its generalizability to the entire state of Gujarat is questionable. As the sample size was small, our study was not able to compare the ADRs among patients with MDR-TB with or without HIV coinfection.

Conclusions

The frequency of ADRs among patients with MDR-TB is high. Twenty-three patients reported ADRs in our study; tinnitus, headache, numbness, dizziness and nausea were the most common ones. ADRs often compel patients to think of discontinuing treatment and their regimen needs to be changed due to it. ADRs were the main reason for discontinuation of treatment of TB.

A close watch on the occurrence of ADRs and appropriate alteration in treatment of such patients with MDR-TB may allow for a higher compliance to treatment.

Acknowledgements

We would like to thank Dr Dixit Kapadia (State TB-HIV coordinator), Dr K.R. Solanki (DTO Bhavnagar), Dr Vikas Sinha (Dean, Government Medical College Bhavnagar), Mr Mahendra Solanki (District TB-HIV supervisor of Bhavnagar), TBHVs of Bhavnagar Corporation and taluka-level STSs of Bhavnagar rural for their guidance and all respondents for their support in the study.

This study was financially supported by The Tuberculosis Association of India, New Delhi.

Conflicts of interest. Nil