You are here

  • Home
  • Archive
  • Volume 14, Issue 2
  • Risk factors for tuberculosis-related death among adults with drug-sensitive pulmonary tuberculosis in Selangor, Malaysia from 2013 to 2019: a retrospective cohort study using surveillance data

Article Text

Article menu

Download PDFPDF

Public health
Original research
Risk factors for tuberculosis-related death among adults with drug-sensitive pulmonary tuberculosis in Selangor, Malaysia from 2013 to 2019: a retrospective cohort study using surveillance data
  1. Mohammad Haikal Suhairi1,2,
  2. Mariam Mohamad1,
  3. Mohamad Rodi Isa1,
  4. Mas Ahmad Sherzkawee Mohd Yusoff3,
  5. Nurhuda Ismail1
  1. 1Department of Public Health Medicine, Universiti Teknologi MARA, Sungai Buloh, Malaysia
  2. 2Bahagian Pengurusan Latihan Kementerian Kesihatan Malaysia, Putrajaya, Malaysia
  3. 3Jabatan Kesihatan Negeri Selangor, Shah Alam, Malaysia
  1. Correspondence to Dr Mariam Mohamad; mariammd{at}uitm.edu.my

Abstract

Objectives Due to the paucity of literature on risk factors for tuberculosis (TB)-related death, we determine the sociodemographic and clinical risk factors associated with TB-related deaths among adult pulmonary TB (PTB) patients on treatment in Selangor, Malaysia.

Design Retrospective cohort study.

Setting Routinely collected primary care data from all government TB clinics in Selangor.

Participants Data of 24 570 eligible adult PTB patients from 2013 to 2019 were obtained from Selangor’s State Health Department surveillance records. We included PTB patients aged at least 15 years old at the time of diagnosis with complete documentation of the dates of diagnosis, treatment initiation, end of treatment/follow-up and treatment outcomes. We excluded patients whose diagnoses were changed to non-TB, post-mortem TB diagnosis and multidrug-resistant TB (MDR-TB) patients.

Primary and secondary outcome measures TB-related death, determined from the recorded physicians’ consensus during the TB mortality meeting.

Results TB-related death was significantly associated with far (adjusted HR (aHR) 9.98, 95% CI 4.28 to 23.28) and moderately advanced (aHR 3.23, 95% CI 1.43 to 7.31) radiological findings at diagnosis; concurrent TB meningitis (aHR 7.67, 95% CI 4.53 to 12.98) and miliary TB (aHR 6.32, 95% CI 4.10 to 9.74) involvement; HIV positive at diagnosis (aHR 2.81, 95% CI 2.21 to 3.57); Hulu Selangor (aHR 1.95, 95% CI 1.29 to 2.93), Klang (aHR 1.53, 95% CI 1.18 to 1.98) and Hulu Langat (aHR 1.31, 95% CI 1.03 to 1.68) residing districts; no formal education (aHR 1.70, 95% CI 1.23 to 2.35); unemployment (aHR 1.54, 95% CI 1.29 to 1.84), positive sputum smear acid-fast bacilli (AFB) at diagnosis (aHR 1.51, 95% CI 1.22 to 1.85); rural residency (aHR 1.39, 95% CI 1.13 to 1.72) and advancing age (aHR 1.03, 95% CI 1.02 to 1.03).

Conclusions Far and moderately advanced radiological findings, concurrent TB meningitis and miliary TB involvement, HIV positive, Hulu Selangor, Klang and Hulu Langat residing districts, no formal education, unemployment, positive sputum smear AFB, rural residency and advancing age are risk factors of TB-related death. Our findings should assist in identifying high-risk patients requiring interventions against TB-related death.

  • Public health
  • Mortality
  • Epidemiology
  • Tuberculosis

Data availability statement

Data may be obtained from a third party and are not publicly available. The data that produced the findings of this study are available in the MOH’s MyTB, but restrictions may apply in requesting the data. Data are, however, available from the authors on reasonable request and with permission from the Selangor’s State Health Department.

http://creativecommons.org/licenses/by-nc/4.0/

This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.

https://doi.org/10.1136/bmjopen-2023-080144

Statistics from Altmetric.com

    Request Permissions

    If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

    STRENGTHS AND LIMITATIONS OF THIS STUDY

    • The large sample size obtained from the national tuberculosis (TB) surveillance registry enables population-level inferences to be made.

    • Multiple individual-level sociodemographic and clinical information were considered for analysis, allowing better adjustment for confounding and measurement of the degree of influence of each independent variable on the outcome.

    • TB-related death was chosen as the outcome measure instead of all-cause mortality during treatment for cause-specific risk factor analysis.

    • Residual confounding may be present due to the absence of modifiable information throughout the treatment period, including TB bacillary load, HIV viral load, glycaemic control and smoking behaviour.

    Introduction

    Tuberculosis (TB) is an infectious disease that threatens public health globally. The WHO declared the TB epidemic a global health emergency in 1993.1 One of the prominent concerns of TB is its association with lethality. Despite concerted global efforts to improve the accessibility to curative TB treatment, TB remains one of the leading causes of death in recent years. It has emerged to be the leading cause of death attributable to a single infectious agent since 2007.2–4 Furthermore, the recent COVID-19 pandemic has disrupted essential TB services across many countries, resulting in a rise in global TB mortality rates. In 2020, the predicted number of deaths attributed to TB was 1.5 million, which subsequently rose to 1.6 million in 2021.5 Thus, it is imperative to decrease TB mortality.

    Existing surveillance data suggested that the outcome of an active TB patient is influenced by factors beyond the administration of an effective treatment regimen. Moreover, a multicentred study and a systematic review uncovered substantial disparities in the risk factors of TB deaths across various situations and geographical locations.6 7 This knowledge gap has spurred research into the potential factors that might contribute to TB mortality across diverse global contexts. Sociodemographic and clinical features have been frequent candidates for analysis, as these factors have been recognised as essential determinants of TB outcome.3 Nevertheless, many researchers opted for all-cause mortality as their outcome measure, possibly motivated by the WHO’s TB treatment outcome definition of ‘died’, which encompassed all TB patients who passed away during treatment irrespective of the underlying cause.8 Nonetheless, focusing on deaths directly due to TB will yield a better understanding of the progression of TB disease within its host and provide crucial empirical data for devising strategies to avert TB-related fatalities.

    Based on recent data, TB mortality also poses a significant public health issue in Malaysia. TB is the country’s primary cause of mortality from communicable diseases, and the annual trend is steadily rising.9–13 Among the Malaysian states, Selangor is the leading contributor to the nation’s TB mortality every year. From 2019 to 2021, it was observed that Selangor accounted for over 20% of the total TB deaths in Malaysia.11–13 In light of the state’s high burden of TB deaths, our research focuses on the population of TB patients in Selangor as a means to gain insight into the survival outcome of TB patients within the local context. We also intended to evaluate the impact of the risk variables identified in the existing literature on TB mortality among TB patients in Selangor. Specifically, our focus was solely on death directly attributed to TB, also known as TB-related death. We limited our analysis to adult pulmonary TB (PTB) cases, as there was a lack of bacteriological confirmation for the majority of extrapulmonary (EPTB) and paediatric TB cases. Additionally, most TB-related deaths were seen in patients with PTB. Thus, we aimed to determine the risk factors associated with TB-related deaths among adult PTB patients receiving treatment in Selangor from 2013 to 2019.

    Materials and methods

    Study design and setting

    We conducted a retrospective cohort study encompassing all adult PTB patients registered in Selangor from 1 January 2013 to 31 December 2019. Selangor is a constituent state of Malaysia situated on the western coast of Peninsular Malaysia, covering a land area of 7951 km2. According to available data, Selangor recorded an approximate population of 6.5 million individuals in the year 2019, 76% of whom were aged 15 years or older.14 The state is divided into nine administrative districts, namely Petaling (the capital district), Hulu Langat, Klang, Gombak, Sepang, Kuala Langat, Kuala Selangor, Hulu Selangor and Sabak Bernam (listed in descending order based on their respective population and urbanisation level).14

    Data source

    All registered TB patients in Malaysia were offered the WHO-recommended directly observed therapy for their treatment administration approach.15 In this approach, another individual is assigned to monitor a patient’s medication adherence, who can be a healthcare worker or a family member.15 In Malaysia, a TB treatment centre is also assigned for each new TB patient for the continution of care.

    We used the surveillance records provided by the TB/Leprosy Unit of Selangor’s State Health Department. The personal information, sociodemographic and clinical characteristics, as well as treatment details of TB patients in Malaysia, were documented and stored in MyTB, which is the online TB surveillance registry established by the Ministry of Health (MOH). The TB treatment centre assigned for each TB patient is required to input the patients’ relevant information and treatment progress in the MyTB.16 The TB surveillance units in each district carried out the data collection and recording process at district level, with oversight provided by the TB surveillance unit at the state level. The State Health Department has access to the data of all TB patients registered within their respective state. The department is responsible for submitting a comprehensive, state-level report to the national TB surveillance unit.16

    In order to maintain patient anonymity, we requested removing any identifying information, including the patient’s name, contact number, identification card (IC) number and residential address.

    Participants

    For our study participants, we included patients whose diagnosis indicated pulmonary parenchyma involvement of TB. Patients must also be at least 15 years old at diagnosis to be included. The dates of diagnosis, treatment initiation and end of treatment/follow-up must be documented appropriately to quantify the treatment duration. We excluded patients whose diagnoses were changed to non-TB or diagnosed with TB at postmortem. Postmortem TB cases were excluded due to diagnosis misclassification risk, as the diagnosis of these types of cases cannot be confirmed either via gold standard diagnostic modality (sputum culture) or clinical response to treatment. We also excluded MDR-TB patients due to substantial case definition, reporting protocol differences and inability to determine whether an MDR-TB patient received the benefit of the second-line MDR-TB regimen.

    Variables

    We retrieved sociodemographic information at diagnosis, including the year of diagnosis, age, sex, nationality, residing district (respective to administrative districts in Selangor), residing location (urban or rural), highest education status and employment status. Whereas for clinical information, sputum smear for acid-fast bacilli stain (ssAFB) and chest radiography results, treatment history, concurrent EPTB involvement, HIV status, diabetes mellitus (DM) status and smoking status. SsAFB and chest radiography are necessary diagnostic modalities for the diagnosis of PTB.15 All potential PTB cases are required to submit minimum of two sputum samples, with at least one early morning sample to a quality-assured laboratory for SsAFB testing.15 Chest radiography results are categorised into four groups. No lesion indicates no radiological lesion consistent with TB; minimal indicates slight lesion, confined to small parts of one or both lungs but not exceeding the upper zone with no cavitation. Moderately advanced indicates either a dense confluent lesion confined in one-third of one lung area or a slight to moderate density lesion disseminated in one or both lungs. If cavitation is present, the maximum diameter is 4 cm. Far advanced indicates lesions that are more extensive than moderately advanced findings.15 A TB patient who has never received anti-TB treatment or was treated for less than 4 weeks is considered a new case. In contrast, those previously diagnosed with TB and received at least 4 weeks of anti-TB treatment were deemed previously treated.15 16 PTB patients may also exhibit concurrent EPTB involvement, typically involving lymph nodes, bone, joints and pleura. TB meningitis and miliary TB, which by themselves are two EPTB manifestations frequently linked to mortality.17 18

    For outcome variables, the survival time was measured from the date of TB treatment commencement to the date of the end of treatment/follow-up or the time of censoring. We applied administrative censoring 1 year (365 days) after the treatment initiation. The patient’s status at the end of survival time indicated whether the patient experienced the event of interest (TB-related death, based on the consensus from the mortality audit meeting) or was censored. We censored all other treatment outcomes, including non-TB-related deaths and deaths with undetermined causes. This outcome variable is based on the TB treament outcome applied by MOH Malaysia, based on recommendation by WHO.16 For TB patients’ treatment outcome, any patient with ssAFB positive at diagnosis (a smear-positive patient) whose ssAFB has been negative in the final month of treatment and at least one prior instance is considered cured.16 On the other hand, a smear-positive patient whose ssAFB is still positive after 5 months of therapy or a smear-negative patient whose ssAFB became positive after 2 months of therapy is considered a treatment failure.16 A smear-negative patient who finished the prescribed treatment and was clinically evaluated as cleared of TB is considered as completed treatment.16 A patient is documented as defaulted treatment if the person’s therapy is interrupted for at least two consecutive months.16 Patients who emigrated from Malaysia during treatment (including non-Malaysians who returned to their countries and Malaysians who continued their treatment overseas) without information of their treatment outcome are considered as transferred out.16 A patient who died for any reason during treatment is documented as died.16 Subsequently, a mortality meeting is held among the involved physicians to determine whether the patient’s death was from TB, based on expert discussion and available documentation, including the death certificate and the death notification registry.16

    Bias

    All of the authors were not involved in the treatment of the study participants and, therefore, did not attend the mortality audit meeting, which determined our study outcome. The retrospective nature of our study design also prevented any influence on the consensus of the meeting that might introduce information bias. We also requested the TB/Leprosy Unit of Selangor’s State Health Department to remove any identifying information, including the patient’s name, contact number, IC number and residential address from the dataset to maintain patient anonymity and avoid selection bias.

    Patient and public involvement

    None.

    Statistical analyses

    We first described our participants’ sociodemographic and clinical information (binary and categorical variables) in numbers and percentages except for age (a continuous variable), in which mean and standard deviation (mean±SD) were used instead. We then tabulated the participants for each year of diagnosis against their respective status at the end of survival time to illustrate the trend TB death throughout the study period. We proceed with Mantel-Haenszel test of trend to detect any trend of TB death throughout the study period.

    We then performed Cox proportional hazard regression for univariable and multivariable analyses to identify risk factors associated with TB-related death. For univariable analysis, we determined calculated HRs and corresponding 95% CIs with p values for each variable. For multivariate Cox analysis, we only considered variables with univariable p values of less than 0.05. We employed backward stepwise regression based on the likelihood ratio as the variable selection procedure for our multivariable model. We calculated HRs and corresponding CIs while adjusting for all available covariates. We considered interactions between two variables based on subject matter knowledge and its significance in the full model. We assessed the proportional hazard assumption and found no violation. Consequently, we detect neither outliers nor significant multicollinearity. We repeated the Cox proportional hazard univariable and multivariable analyses using the complete cases, and compared both results to assess the effect of missing data on the outcome. We usedd IBM’s SPSS V.27 for the analysis.

    Missing data

    We encountered missing data in our dataset and were unable to trace all the data from the primary data source. We then performed missing data evaluation based on the methods suggested by Ridout to determine the mechanism and the strongest predictors of our missing data.19 The methods involved creating a dummy variable for each variable with missing data to indicate whether a participant’s data are present or missing. Each dummy variable was then used as the dependent variable, and multivariable logistic regression analysis was performed using backward stepwise method to determine the significant predictors of missing data for each dummy variable, based on the corresponding p<0.05. We determined that our data were missing at random based on the presence of significant predictors for each dummy variable and opted for single imputation (Bayesian stochastic regression imputation). We used fully conditional specification method with 50 iterations to impute our missing data.

    Results

    TB/Leprosy Unit of Selangor’s State Health Department registered 33 157 TB patients between 2013 and 2019. After applying our inclusion and exclusion criteria, 24 570 patients were included in our analysis (figure 1).

    Figure 1
    Figure 1

    Flow diagram of participant selection and exclusion for the study. TB, tuberculosis; PTB, pulmonary TB; MDR-TB, multidrug-resistant TB.

    There were 1526 participants (6.2% out of all included patients) with missing data, with four independent variables involved. The HIV status variable had the highest missing data, with the information missing in 1086 participants (4.4%). The information on chest radiography at diagnosis was missing in 410 participants (1.7%), ssAFB at diagnosis was missing in 63 participants (0.3%) and residing location was missing in 36 participants (0.1%). All missing variables were determined to be missing at random and subsequently imputed for missing data.

    Table 1 summarises the baseline characteristics of included patients after single imputation of missing data. Our oldest participant was 117 years old. The mean±SD for age was 41.95±16.17 years old. Sixty-six per cent of the participants were male, while more than 50% were Malays. More than two-thirds of the participants resided in the districts of Petaling, Hulu Langat and Gombak districts (30.7%, 23.9% and 15.5%, respectively). The plurality (83.1%) were urban residents. Fifty-four per cent had attained secondary education as the highest education level and 58.1% were employed. Most participants were positive for ssAFB (73.5%) and had minimal chest radiological findings (66.3%). The majority were new cases (91.0%), with no EPTB involvement (96.4%), negative for HIV (93.9%), no DM (79.0%) and non-smokers (69.6%). For treatment outcome, almost half of the participants (49.4%) were cured. Less than 10% of the participants died during treatment, with 595 participants (2.4% of all participants) experienced TB-related death.

    View this table:
    Table 1

    Sociodemographic and clinical information on adult pulmonary tuberculosis (PTB) patients in Selangor (2013–2019) (n=24 570)

    TB death and its trend throughout the years

    Table 2 tabulates the participants for each year of diagnosis according to their respective status at the end of survival time. The percentage of participants who experienced TB death showed an increasing annual trend. Mantel-Haenszel trend test showed a significant result (p=0.018, χ2=5.642), indicating significant presence of TB death trend among the participants throughout the study period.

    View this table:
    Table 2

    Participants by year of diagnosis

    Factors associated with TB-related death within 1 year of treatment

    Table 3 presents the covariates associated with TB-related death within 1 year of treatment from the imputed dataset and table 4 presents the same analysis using the complete dataset. The variables selected for the multivariable analysis, as well as the HR and CI, were almost similar. We decided to report the result based on the imputed dataset.

    View this table:
    Table 3

    Univariable and multivariable analysis of risk factors of TB-related death within 1 year of treatment among adult PTB patients in Selangor, Malaysia (2013–2019) on imputed dataset (n=24 570)

    View this table:
    Table 4

    Univariable and multivariable analysis of risk factors of TB-related death within 1 year of treatment among adult PTB patients in Selangor, Malaysia (2013–2019) on complete dataset (N=23044)

    After considering the confounding and interaction effect in the multivariable model, we found that advancing age was a significant sociodemographic risk factor (adjusted HR (aHR) 1.03, 95% CI 1.02 to 1.03). Residence of Hulu Selangor (aHR 1.95, 95% CI 1.29 to 2.93), Klang (aHR 1.53, 95% CI 1.18 to 1.98) and Hulu Langat (aHR 1.31, 95% CI 1.03 to 1.68) districts had significantly increased risk of TB-related death compared with their counterpart in Petaling. Rural residency (aHR 1.39, 95% CI 1.13 to 1.72), no formal education (aHR 1.70, 95% CI 1.23 to 2.35) and unemployment (aHR 1.54, 95% CI 1.29 to 1.84) were also significant sociodemographic risk factors for TB-related death. For clinical risk factors, positive ssAFB at diagnosis (aHR 1.51, 95% CI 1.22 to 1.85), moderately advanced (aHR 3.23, 95% CI 1.43 to 7.31) and far advanced (aHR 9.98, 95% CI 4.28 to 23.28) radiological findings at diagnosis, concurrent TB meningitis (aHR 7.67, 95% CI 4.53 to 12.98) and miliary TB involvement (aHR 6.32, 95% CI 4.10 to 9.74) and HIV positive at diagnosis (aHR 2.81, 95% CI 2.21 to 3.57) were significantly associated with TB-related death.

    Discussion

    We found a significant increasing trend of TB death annually among our participants, based on our Mantel-Haenszel trend test result. Further investigation is, therefore, warranted to verify our finding, perhaps in a larger study sample of local population, and subsequently explore the underlying mechanism if necessary. To note, the head of the TB/Leprosy Sector of MOH, Dr Mohamed Naim Abdul Kadir attributed the increasing annual trend of TB-related deaths in Malaysia to delayed presentation, which was fueled by low public awareness of TB. Focusing on TB-related deaths in Selangor, he postulated that high population density as well as a significant proportion of urban poor contributed to the state’s worsening annual trend.20 Currently, available local studies found that knowledge and awareness regarding TB infection were mixed among Malaysian subpopulations, where those who are involved in managing TB patients (healthcare workers, TB contacts) had higher TB awareness than those who are not involved. However, the studies identified the social stigma of TB as a significant barrier to seeking medical attention for TB among Malaysians.21–23

    Multivariate analysis revealed that each 1-year increase in patient age is associated with a 1.03-fold increase in hazards of TB-related death. Our results were comparable to those of earlier studies conducted in China, Ethiopia and Iran, where their aHR for age ranged between 1.052 and 1.10.24–26 Considering our findings, we proposed that interventions to reduce TB-related deaths should target the older population for two reasons. First, older adults are more susceptible to poor TB outcomes due to biological vulnerability (progressive decrease in lung function, immunosenescence and malnutrition), poor heath-seeking behaviour, a tendency for atypical presentation, financial limitation and stigma.27 28 Second, the rapid ageing of the global population implies that TB among older adults will become a more significant burden in the future.29 30

    Furthermore, we found that residents of Hulu Selangor, Hulu Langat and Klang districts had a significantly higher risk of TB-related death than those in Petaling. We hypothesised a disparity in the distribution of essential TB services and healthcare resources between Petaling (the capital district) and the non-capital districts. We also speculate that poor air quality in the districts, particularly in Klang, may contribute to poor respiratory health and disease outcomes among their residents.31 32 Recent epidemiological studies in China also supported our speculation, providing evidence that poor air quality is associated with mortality among TB patients.33 34 However, we recommended further investigation and verification for future research.

    Living in rural areas was a significant risk factor for TB-related death, consistent with the findings of rural–urban TB disparity studies, which demonstrated favourable outcomes among urban TB patients. The authors attributed their results to greater accessibility to healthcare services and more efficient follow-up systems in urban areas compared with their rural counterparts.35–37 Subsequently, no formal education was a significant risk factor for TB-related death. Our result echoed previous studies’ findings from developing and developed countries.38–40 Individuals with a higher level of formal education are associated with positive health-seeking behaviour and compliance with treatment. Furthermore, the education level is a good indicator of a person’s economic situation, which determines one’s capacity to practice healthy living and access health resources if needed.41

    We also discovered that unemployment was a risk factor for TB-related death. Despite the availability of government-subsidised TB treatment services in most countries, a TB-infected individual had to bear the cost of seeking medical attention (including expenditures on transportation and medical investigation). A TB-infected person may become sick and unfit for work, leading to income loss.42 43 Such costs may prevent a financially deprived patient from accessing healthcare and maintaining treatment adherence, affecting the treatment outcome. Moreover, the catastrophic cost associated with TB treatment is an adverse outcome recognised and addressed in one of the milestones for the WHO’s End TB Strategy.43 44 To mitigate this problem, Selangor’s state government has begun providing financial assistance to all Malaysian TB patients receiving treatment in Selangor since 2020.

    For clinical risk factors, we found that positive results on sputum AFB and advanced pathology on chest radiography were both significant risk factors for TB-related death, similar to earlier studies.45 46 The infection-to-disease TB spectrum dictates that positive ssAFB result and severe chest radiological findings indicate a high bacillary load of Mycobacterium tuberculosis (the pathogen for TB), which leads to a more severe TB manifestation and lower survival compared with a TB patient with negative ssAFB result or less severe radiological finding.47 Thus, early detection of TB is vital to prevent poor outcomes and transmission to others. In Malaysia, enhancing case detection is one of the strategies under the current national strategic TB control plan, with two primary interventions: improving accessibility and quality of TB screening and diagnosis for high-risk populations and integrating latent TB cases’ diagnosis and management into a systematic screening of active TB.48

    Concurrent TB meningitis and miliary TB involvement were significant risk factors for TB-related death, supported by recent findings.49 50 On their own, TB meningitis and miliary TB are both devastating manifestations of TB, as their non-specific symptoms make early diagnosis difficult. Additionally, they were associated with severe complications, including permanent neurological complications for TB meningitis and acute respiratory distress syndrome for miliary TB.51 52 Next, we confirmed that HIV was a significant risk factor for TB-related death, consistent with findings from other recent studies elsewhere.24 25 HIV has a synergistic effect on TB, accelerating the disease’s progression and complicating its management.41

    Despite DM status not considered in the multivariable analysis, literature suggested that the comorbidity was associated with TB death. DM has long been recognised to complicate the treatment of TB patients by making the host more susceptible to infection and debilitating healing process of the host. DM also frequently associated with renal and cardiovascular complications which further complicate TB treatment.53 In terms of direct effect of DM towards TB death per se, an updated systematic review and meta-analysis on the effects of DM on TB treatment outcomes found that TB-DM patients had almost two times higher odds of death than the non-DM TB patients. However, the same review noticed the lack of information of the cause of death among the included studies. Thus, the result cannot be taken as the association of DM towards death due to TB, since DM itself is associated with mortality, particularly cardiovascular death.54

    We recommend that managing adult PTB patients with significant risk factors of TB-related death should be prioritised and optimised. Our findings can guide TB programme managers at the ground level to identify these high-risk patients for closer monitoring and provide strong evidence for policy-makers to develop intervention programmes to reduce the TB-related death toll. Furthermore, we recommend that TB patients with significant clinical risk factors of TB-related death should be prioritised in intervention programmes as these risk factors are associated with higher HRs than their sociodemographic counterparts.

    For future studies, we recommend conducting a similar in a different geographical setting, such as impoverished communities, residents of industrial areas with high air pollution level and regional areas to test the generalisability and ability to extrapolate our result to a different setting and provide a general understanding of how living in a particular area affects TB mortality.

    Strengths and limitations

    Our study findings drew strength from our large sample size, allowing us to generalise to all adult PTB patients registered in Selangor. We exclusively considered TB-related deaths, excluding deaths from other causes, including accidental death and death due to causes more associated with DM and HIV. Hence, our findings offer a more accurate, cause-specific risk factor analysis during treatment. Moreover, our findings provide new knowledge on local adult PTB patients, contributing crucial information on the geographical variation of TB-related death risk factors mentioned in the literature.

    However, this study was not without limitations. We only used our participants’ baseline information obtained during treatment preinitiation. Bacillary load (reflected in ssAFB and chest radiography findings), viral load in HIV patients, glycaemic control in DM patients and smoking behaviour in smokers, which may change throughout the treatment period, were unavailable from our data source. Therefore, the analysis may be subjected to residual confounding due to the potential influence of these factors. The presence of missing data limits our ability to retrieve full information in the pursuit to provide an accurate description of our study population. We resorted to single imputation method to estimate the missing information.

    Furthermore, despite using a large dataset, our result only represented the adult PTB population in Selangor. Therefore, our result may have limited generalisation to other populations, for example, all TB patients in Malaysia.

    Another limitation is that since the secondary data source was used for the study, important information such as patient’s medication, side effect and toxicity of TB medication, body mass index and health-related risk factors such as malnutrition and poverty were unable to be included since this information was not captured in the surveillance records. Regarding malnutrition, its relative importance and contribution towards TB incidence varies significantly among countries. For Malaysia, the estimate percentage of TB incidence attributable to undernourishment were less than 10%.5 Based on the the Sustainable Development Goals prevalence of undernourishment (PoU), Malaysia ranged between 3.8% and 6.1% during the 2013–2019 period. For comparison, the global PoU was between 7.5%and 7.9% for the same time period.55

    We concluded that far advanced and moderately advanced radiological findings at diagnosis, concurrent TB meningitis and miliary TB involvement, HIV positive at diagnosis, Hulu Selangor, Klang and Hulu Langat residing districts, no formal education, unemployment, positive SsAFB at diagnosis; rural, male and advancing age were risk factors associated with TB-related death within 1 year of treatment were significantly associated with adult PTB patients on treatment in Selangor from 2013 to 2019. Our findings may guide TB programme managers in prioritising the management of high-risk patients and assist policy-makers in formulating intervention programmes to reduce the death toll from TB among adult PTB patients.

    Data availability statement

    Data may be obtained from a third party and are not publicly available. The data that produced the findings of this study are available in the MOH’s MyTB, but restrictions may apply in requesting the data. Data are, however, available from the authors on reasonable request and with permission from the Selangor’s State Health Department.

    Ethics statements

    Patient consent for publication

    Ethics approval

    This study had been registered in the National Medical Research Register (registration number NMRR-21-358-58741(IIR)). Ethical approval for this study had been granted by the Medical Research Ethics Committee (MREC) of MOH. Since the research was based on secondary data and patient identifying information had been stripped from the dataset for anonymity, no consent for publication was taken.

    Acknowledgments

    We thank the Director-General of Health Malaysia for the permission to publish this paper. We also thank Dato' Indera Dr. Sha'ari bin Ngadiman of Selangor's State Health Department for the data provision. We acknowledge the hard work of all healthcare staff involved in the Malaysian TB surveillance system, particularly in Selangor, without whom our study would not be a reality.

    References

    1. ↵
      1. Nathavitharana RR,
      2. Friedland JS
      . A tale of two global emergencies: tuberculosis control efforts can learn from the Ebola outbreak. Eur Respir J 2015;46:293–6. doi:10.1183/13993003.00436-2015
      OpenUrlAbstract/FREE Full Text
    2. ↵
      1. Furin J,
      2. Cox H,
      3. Pai M
      . Tuberculosis. Lancet 2019;393:1642–56. doi:10.1016/S0140-6736(19)30308-3
      OpenUrlCrossRef
    3. ↵
      1. World Health Organization
      . Global tuberculosis report 2021; 2021.
    4. ↵
      1. World Health Organization
      . Global tuberculosis report 2020; 2020.
    5. ↵
      1. World Health Organization
      . Global tuberculosis report 2022; 2022.
    6. ↵
      1. Waitt CJ,
      2. Squire SB
      . A systematic review of risk factors for death in adults during and after tuberculosis treatment. Int J Tuberc Lung Dis 2011;15:871–85. doi:10.5588/ijtld.10.0352
      OpenUrlCrossRefPubMed
    7. ↵
      1. Lefebvre N,
      2. Falzon D
      . Risk factors for death among tuberculosis cases: analysis of European surveillance data. Eur Respir J 2008;31:1256–60. doi:10.1183/09031936.00131107
      OpenUrlAbstract/FREE Full Text
    8. ↵
      1. World Health Organization
      . Definitions and reporting framework for tuberculosis – 2013 revision (updated December 2014 and January 2020); 2013.
    9. ↵
      1. Ministry of Health Malaysia
      . Petunjuk Kesihatan 2018; 2018.
    10. ↵
      1. Ministry of Health Malaysia
      . Petunjuk Kesihatan 2019; 2019.
    11. ↵
      1. Ministry of Health Malaysia
      . Petunjuk Kesihatan 2020; 2020.
    12. ↵
      1. Ministry of Health Malaysia
      . Petunjuk Kesihatan 2021; 2021.
    13. ↵
      1. Ministry of Health Malaysia
      . Petunjuk Kesihatan 2022; 2022.
    14. ↵
      1. Department of Statistics Malaysia
      . Household income and basic amenities survey report by state and administrative district Selangor; 2020.
    15. ↵
      1. Ministry of Health Malaysia
      . Management of tuberculosis, 3rd ed. 2012.
    16. ↵
      1. Ministry of Health Malaysia
      . Manual Sistem Maklumat Tibi Kesihatan (Semakan 2018); 2019.
    17. ↵
      1. Pang Y,
      2. An J,
      3. Shu W, et al
      . Epidemiology of extrapulmonary tuberculosis among Inpatients, China, 2008-2017. Emerg Infect Dis 2019;25:457–64. doi:10.3201/eid2503.180572
      OpenUrlCrossRefPubMed
    18. ↵
      1. Khan FY
      . Review of literature on disseminated tuberculosis with emphasis on the focused diagnostic workup. J Family Community Med 2019;26:83–91. doi:10.4103/jfcm.JFCM_106_18
      OpenUrl
    19. ↵
      1. Ridout MS
      . Testing for random dropouts in repeated measurement data. Biometrics 1991;47:1617–9;
      OpenUrlCrossRefPubMedWeb of Science
    20. ↵
      1. Bernama
      . TB deaths high due to late treatment [Malaysiakini]. 2019. Available: https://www.malaysiakini.com/news/503769
    21. ↵
      1. Idris NA,
      2. Zakaria R,
      3. Muhamad R, et al
      . The effectiveness of tuberculosis education programme in Kelantan, Malaysia on knowledge, attitude, practice and stigma towards tuberculosis among adolescents. Malays J Med Sci 2020;27:102–14. doi:10.21315/mjms2020.27.6.10
      OpenUrl
    22. ↵
      1. Jamaludin TSS,
      2. Ismail N,
      3. Saidi S
      . Knowledge, awareness, and perception towards tuberculosis disease among international Islamic University Malaysia Kuantan students. Enfermería Clínica 2019;29:771–5. doi:10.1016/j.enfcli.2019.04.116
      OpenUrl
    23. ↵
      1. Loh SY,
      2. Zakaria R,
      3. Mohamad N
      . Knowledge, attitude, and stigma on tuberculosis and the associated factors for attitude among tuberculosis contacts in Malaysia. Medeni Med J 2023;38:45–53. doi:10.4274/MMJ.galenos.2023.14478
      OpenUrl
    24. ↵
      1. Birlie A,
      2. Tesfaw G,
      3. Dejene T, et al
      . Time to death and associated factors among tuberculosis patients in Dangila Woreda, Northwest Ethiopia. PLoS One 2015;10:e0144244. doi:10.1371/journal.pone.0144244
    25. ↵
      1. Xie Y,
      2. Han J,
      3. Yu W, et al
      . Survival analysis of risk factors for mortality in a cohort of patients with tuberculosis. Can Respir J 2020;2020:1654653. doi:10.1155/2020/1654653
    26. ↵
      1. Rahmanian V,
      2. Rahmanian K,
      3. Rahmanian N, et al
      . Survival rate among tuberculosis patients identified in south of Iran, 2005-2016. J Acute Dis 2018;7:207. doi:10.4103/2221-6189.244172
      OpenUrl
    27. ↵
      1. Caraux-Paz P,
      2. Diamantis S,
      3. de Wazières B, et al
      . Tuberculosis in the elderly. JCM 2021;10:5888. doi:10.3390/jcm10245888
      OpenUrl
    28. ↵
      1. Raghu S
      . Challenges in treating tuberculosis in the elderly population in tertiary Institute. IJTB 2022;69:S225–31. doi:10.1016/j.ijtb.2022.10.008
      OpenUrl
    29. ↵
      1. United Nations
      . World population prospects. 2022. Available: https://population.un.org/wpp/Graphs/DemographicProfiles/Line/900
    30. ↵
      1. Teo AKJ,
      2. Rahevar K,
      3. Morishita F, et al
      . Tuberculosis in older adults: case studies from four countries with rapidly ageing populations in the Western Pacific region. BMC Public Health 2023;23. doi:10.1186/s12889-023-15197-7
    31. ↵
      1. Mabahwi NA,
      2. Ling HL,
      3. Mohamed Musthafa SNA, et al
      . Air quality-related human health in an urban region. case study: state of Selangor, Malaysia. EnvironmentAsia 2018;11:194–216. doi:10.14456/EA.2018.15
      OpenUrl
    32. ↵
      1. Ling OHL,
      2. Marzukhi MA,
      3. Qi JK, et al
      . Impact of urban land uses and activities on the ambient air quality in Klang valley, Malaysia from 2014 to 2020. PM 2020;18:239–58. doi:10.21837/pm.v18i14.829
      OpenUrl
    33. ↵
      1. Liu Y,
      2. Zhao S,
      3. Li Y, et al
      . Effect of ambient air pollution on tuberculosis risks and mortality in Shandong, China: a multi-city modeling study of the short- and long-term effects of pollutants. Environ Sci Pollut Res Int 2021;28:27757–68. doi:10.1007/s11356-021-12621-6
      OpenUrl
    34. ↵
      1. Wang X-Q,
      2. Zhang K-D,
      3. Yu W-J, et al
      . Associations of exposures to air pollution and greenness with mortality in a newly treated tuberculosis cohort. Environ Sci Pollut Res Int 2023;30:34229–42. doi:10.1007/s11356-022-24433-3
      OpenUrl
    35. ↵
      1. Balaky STJ,
      2. Mawlood AH,
      3. Shabila NP
      . Survival analysis of patients with tuberculosis in Erbil, Iraqi Kurdistan region. BMC Infect Dis 2019;19:865. doi:10.1186/s12879-019-4544-8
    36. ↵
      1. Mutembo S,
      2. Mutanga JN,
      3. Musokotwane K, et al
      . Urban-rural disparities in treatment outcomes among recurrent TB cases in Southern province, Zambia. BMC Infect Dis 2019;19:1087. doi:10.1186/s12879-019-4709-5
    37. ↵
      1. Wu Y-C,
      2. Lo H-Y,
      3. Yang S-L, et al
      . Comparing the factors correlated with tuberculosis-specific and non-tuberculosis-specific deaths in different age groups among tuberculosis-related deaths in Taiwan. PLoS ONE 2015;10:e0118929. doi:10.1371/journal.pone.0118929
    38. ↵
      1. Montes K,
      2. Atluri H,
      3. Silvestre Tuch H, et al
      . Risk factors for mortality and multidrug resistance in pulmonary tuberculosis in Guatemala: a retrospective analysis of mandatory reporting. J Clin Tuberc Other Mycobact Dis 2021;25:100287. doi:10.1016/j.jctube.2021.100287
    39. ↵
      1. Valencia-Aguirre S,
      2. Arroyave I,
      3. García-Basteiro AL
      . Educational level and tuberculosis mortality in Colombia: growing inequalities and stagnation in reduction. Cad Saude Publica 2022;38. doi:10.1590/0102-311X00031721
    40. ↵
      1. Álvarez JL,
      2. Kunst AE,
      3. Leinsalu M, et al
      . Educational inequalities in tuberculosis mortality in sixteen European populations. Int J Tuberc Lung Dis 2011;15:1461–7, doi:10.5588/ijtld.10.0252
      OpenUrlCrossRefPubMed
    41. ↵
      1. Kayembe J-M
      . Determinants of survival of patients with tuberculosis in developing countries. In: Kayembe J-M, ed. Tuberculosis. IntechOpen, 2018.
    42. ↵
      1. World Health Organization
      . Implementing the end TB strategy: the essentials. 2015.
    43. ↵
      1. Ghazy RM,
      2. El Saeh HM,
      3. Abdulaziz S, et al
      . A systematic review and meta-analysis of the catastrophic costs incurred by tuberculosis patients. Sci Rep 2022;12:558. doi:10.1038/s41598-021-04345-x
    44. ↵
      1. World Health Organization
      . The end TB strategy global strategy and targets for tuberculosis prevention, care and control after 2015; 2014.
    45. ↵
      1. Liu K,
      2. Ai L,
      3. Pan J, et al
      . Survival analysis and associated factors for pulmonary tuberculosis death: evidence from the information system of tuberculosis disease and mortality surveillance in China. RMHP 2022;15:1167–78. doi:10.2147/RMHP.S368593
      OpenUrl
    46. ↵
      1. Alemu A,
      2. Bitew ZW,
      3. Worku T, et al
      . Predictors of mortality in patients with drug-resistant tuberculosis: a systematic review and meta-analysis. PLoS One 2021;16:e0253848. doi:10.1371/journal.pone.0253848
    47. ↵
      1. Pai M,
      2. Behr MA,
      3. Dowdy D, et al
      . Tuberculosis. Nat Rev Dis Primers 2016;2:16076. doi:10.1038/nrdp.2016.76
      OpenUrl
    48. ↵
      1. Ministry of Health Malaysia
      . National strategic plan for tuberculosis control (2016-2020); 2016.
    49. ↵
      1. Bukundi EM,
      2. Mhimbira F,
      3. Kishimba R, et al
      . Mortality and associated factors among adult patients on tuberculosis treatment in Tanzania: a retrospective cohort study. J Clin Tuberc Other Mycobact Dis 2021;24:100263. doi:10.1016/j.jctube.2021.100263
    50. ↵
      1. Hammami F,
      2. Koubaa M,
      3. Chakroun A, et al
      . Comparative analysis between tuberculous meningitis and other forms of extrapulmonary tuberculosis. Germs 2021;11:23–31. doi:10.18683/germs.2021.1237
      OpenUrl
    51. ↵
      1. Davis AG,
      2. Rohlwink UK,
      3. Proust A, et al
      . The pathogenesis of tuberculous meningitis. J Leukoc Biol 2019;105:267–80. doi:10.1002/JLB.MR0318-102R
      OpenUrlCrossRef
    52. ↵
      1. Wakamatsu K,
      2. Nagata N,
      3. Kumazoe H, et al
      . Prognostic factors in patients with miliary tuberculosis. J Clin Tuberc Other Mycobact Dis 2018;12:66–72. doi:10.1016/j.jctube.2018.07.001
      OpenUrl
    53. ↵
      1. Pearson-Stuttard J,
      2. Blundell S,
      3. Harris T, et al
      . Diabetes and infection: assessing the association with glycaemic control in population-based studies. Lancet Diabetes Endocrinol2016;4:148–58. doi:10.1016/S2213-8587(15)00379-4
      OpenUrl
    54. ↵
      1. Huangfu P,
      2. Ugarte-Gil C,
      3. Golub J, et al
      . The effects of diabetes on tuberculosis treatment outcomes: an updated systematic review and meta-analysis. Int J Tuberc Lung Dis 2019;23:783–96. doi:10.5588/ijtld.18.0433
      OpenUrlPubMed
    55. ↵
      1. Food and Agricultural Organization of the United Nations (FAO)
      . SDG indicators data portal. 2023. Available: https://www.fao.org/sustainable-development-goals-data-portal/data/indicators/2.1.1-prevalence-of-undernourishment/en [Accessed 27 Oct 2023].

    Footnotes

    • Contributors MHS contributed to the conceptualisation, project administration, data curation, methodology, formal analysis, draft writing and editing. MM contributed to the methodology, formal analysis, project administration, draft review and responsible for the overall content as guarantor. MRI contributed to the methodology and draft review. MASMY and NI both contributed to the draft review. All authors have read and approved the final manuscript.

    • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

    • Competing interests None declared.

    • Patient and public involvement Patients and/or the public were not involved in the design, or conduct, or reporting, or dissemination plans of this research.

    • Provenance and peer review Not commissioned; externally peer reviewed.