Abstract
Tuberculosis (TB) is a leading killer from a single infectious agent globally. In 2019, Malaysia’s TB incidence rate was 92 per 100,000 population, and the TB mortality rate was estimated at 4 cases per 100,000 population per year. However, the state of Sabah had a higher burden of TB with a notification rate of 128 per 100,000 population and a TB case fatality rate of 8% compared to the national figure. This study aims to provide a comprehensive report on TB deaths epidemiology and its associated factors at a sub-national level. This nested case-control study used Sabah State Health Department TB surveillance data from the Malaysia national case-based TB registry (MyTB) between 2014 and 2018. Cases were defined as all-cause TB deaths that occurred before anti-TB treatment completion from the time of TB diagnosis. Controls were randomly selected from TB patients who completed anti-TB treatment. The TB mortality rate had increased significantly from 9.0/100,000 population in 2014 to 11.4/100,000 population in 2018. The majority of TB deaths occurred in the first two months of treatment. TB-related deaths were primarily due to advanced disease or disseminated TB, whereas non-TB-related deaths were primarily due to existing comorbidities. Many important independent risk factors for TB deaths were identified which are useful to address the increasing TB mortality rate.
Keywords: tuberculosis, case fatality, mortality, Sabah, Malaysia
1. Introduction
Tuberculosis (TB) is caused by the bacterium Mycobacterium tuberculosis and is a leading killer from a single infectious agent globally, ranks higher than HIV/AIDS, and is one of the top 10 causes of death worldwide [1]. In 2019, an estimated 10 million people fell ill with TB; of these, 1.2 million TB deaths were among human immunodeficiency virus (HIV)-negative people, and 208,000 deaths among HIV-positive people [1]. The World Health Organization (WHO) has set the target of a 35% reduction in TB deaths by 2020 compared with 2015 as one of the targets to end the TB epidemic [2]. However, the WHO reports [1] that the annual number of TB deaths is not falling fast enough to reach the 2020 milestone, whereby the cumulative reduction between 2015 and 2019 was only 14%.
TB incidence and TB mortality rate are essential indicators of TB burden. In 2019, the estimate for TB incidence rate in Malaysia was 92 cases per 100,000 population, and the TB mortality rate was estimated at 4 cases per 100,000 population per year [1]. However, Sabah state had a higher burden of TB with a notification rate of 128 cases per 100,000 population and a TB case fatality rate of 8% compared to the national figure [3].
The TB control program in Malaysia has been strengthened to be aligned with the WHO End TB Strategy. The strengthening of the program is important to ensure that the targets will be achieved by 2035. Under the strategy, one of the targets is to reduce deaths due to TB by 35% in the year 2020, 75% in 2025, 90% in 2030, and finally reach the target of reduction by 95% in the year 2035 compared with 2015 [2]. TB control strategies and activities need to be tailored to different settings or states in Malaysia. To come up with a comprehensive national strategic plan for TB, an in-depth analysis of the TB disease epidemiology for each state is crucial.
Studies looking into factors associated with TB deaths have been carried out worldwide in high and low TB incidence countries, and have found important factors which include patients’ sociodemographic characteristics, TB disease characteristics, TB treatment characteristics, and underlying comorbidities [4,5]. However, the findings are different between countries and regions. In Malaysia, studies investigating specifically TB death and its associated factors are limited. A recent study [6] reporting factors associated with all-cause mortality among TB patients was conducted at the national level; however, similarities or differences between states were not explored. TB death situation in the state of Sabah will be explored in this study. Sabah is located in East Malaysia and is economically less developed compared to the states in peninsular Malaysia. Sabah also has a large number of immigrants from the Philippines and Indonesia living in overcrowded settlements. Therefore, this study aims to provide a comprehensive report on TB deaths epidemiology and its associated factors specifically for the state of Sabah. Sabah state contributes the highest TB burden, i.e., 19.7% of Malaysia’s total reported TB cases [3].
2. Methods
2.1. Study Population and Source of Data
This nested case-control study used Sabah State Health Department TB surveillance data taken from the Malaysia national case-based TB registry (MyTB) between 2014 and 2018. Cases were defined as all-cause TB deaths that occurred before anti-TB treatment completion from the time of TB diagnosis. Controls were randomly selected from TB patients registered between 2014 and 2018 who completed anti-TB treatment. Sociodemographic, comorbid, disease, and treatment characteristics of cases and controls were extracted from MyTB. Sabah population data were obtained from the Department of Statistics Malaysia.
2.2. Definitions
2.2.1. Outcome Variables
The main outcome variable was all-cause TB death, which was categorized into two groups: TB-related death and non-TB-related death. TB death in this study was defined as all-cause mortality that occurs before completing anti-TB treatment from the date of TB diagnosis. All TB death cases in MyTB will be classified into TB-related or non-TB-related death by the state and district level TB death audit committee. The members of the committee are a Chest physician, Family Medicine Specialist, TB program manager, doctors, and health staff involved in the management of TB patient. In this study, TB-related death was defined as the underlying cause of death being due to TB directly, whereas non-TB-related death was defined as any underlying cause of death not directly due to TB as decided by the TB audit committee.
2.2.2. Explanatory Variables
The independent variables that may be associated with TB death were assessed, including sociodemographic variables (age, gender, place of residence, nationality, education level, and income status), comorbid variables (smoking status, diabetes mellitus, and human immunodeficiency virus infection), disease characteristics (category of TB, type of case detection, BCG scar, TB anatomical site, initial sputum acid-fast bacilli status, chest radiography findings at diagnosis, multi-drug resistant TB status, TB meningitis, and TB military), and treatment characteristics (treatment regime during intensive phase, directly observed treatment achievement during intensive phase and treatment delay, and treatment delay).
The lesion on chest radiography (CXR) findings at diagnosis was categorized into three groups, i.e., minimal, moderate, and advanced. Based on the Malaysia TB clinical practice guidelines, minimal lesion is defined as having slight lesions with no cavity and lesions confined to small parts of one or both lungs but the total extent not exceeding the upper zone; moderate lesion is defined as having dense lesions not exceeding one-third of one lung or both lungs not exceeding the volume of one lung, or presence of cavity not exceeding 4 cm in diameter; advanced lesion defined as lesions are more extensive than the moderate lesion.
Directly observed therapy (DOT) was defined as anti-TB medication ingestion that was directly supervised by health staff or trained family members. Under the TB control program, TB patients will take their medication under the supervision of either the health staff or a trained family member. The DOT supervisor will sign the TB treatment booklet after observing the patient ingesting the correct dosage of anti-TB drugs. The information on the achievement of DOT supervision was obtained from the TB treatment booklet. DOT supervision is achieved during the intensive phase when 80% of the prescribed daily doses are taken and supervised. Treatment delay in days was calculated by subtracting the two dates (i.e., date of treatment started minus date of diagnosis). An interval of 14 days or longer between the two dates was defined as a delay in starting anti-TB treatment.
2.3. Statistical Analysis
Data were expressed as count and percentage for categorical variables and the mean ± standard deviation (SD) or median with interquartile range (IQR) for continuous variables. The sociodemographic, comorbid, disease, and treatment characteristics between all-cause TB deaths (cases) and controls were compared using Pearson’s chi-square test or Fisher’s exact test and the independent-samples t-test where appropriate.
All variables found to be significant (p < 0.05) from the univariable analysis were considered for inclusion in multivariable analysis. Multivariable binary logistic regression was used to assess the associations of selected independent variables with all-cause TB deaths. A simultaneous forced-entry method was employed in multivariable analysis. The model fit was assessed using Hosmer–Lemeshow goodness-of-fit test, where a p-value greater than 0.05 means that the model is a good fit. A similar approach was also used to identify risk factors for all-cause TB deaths before anti-TB treatment started compared to after anti-TB treatment started. Time to death in days was calculated by subtracting the two dates (i.e., date of death minus date of treatment started). Death before treatment is initiated is defined when the difference between the two dates is zero.
Multinominal logistic regression analysis was used for comparison between cases of TB-related death and non-TB-related death with the controls to identify the factors associated with TB-related and non-TB-related deaths. Crude odds ratio (cOR) and adjusted odds ratio (aOR) with a 95% confidence interval (95%CI) and Wald p-value were reported in all the regression analysis results to show the strength of association between independent and outcome variables. All data preparation, calculations, and analyses were performed using Microsoft Excel and IBM Statistical Package for the Social Science (SPSS) software version 26. The level of significance was set at a p-value less than 0.05.
2.4. Ethical Consideration
Ethical approval for this study was obtained from the Medical Ethics Committee of the University Sabah Malaysia on 25 September 2020 JKEtika 4/20 (8), and from the Medical Research and Ethics committee (MREC), Ministry of Health Malaysia on 14 September 2020 (approval code NMRR-20-1289-55461). The Sabah State Health Director granted permission to access the de-identified case-listing of TB patients from the MyTB database, and patient consent was not obtained because this study only uses patient’s information from the MyTB database.
3. Results
3.1. Sample Characteristics
Between 2014 and 2018, there were 24,277 TB patients registered in MyTB from Sabah. Of these, 1794 were identified as TB death, and 2000 controls were randomly selected from 20,557 TB patients who completed anti-TB treatment. A total of 1926 from the cohort were excluded, i.e., loss to follow-up (1533), changed diagnosis (181), failed treatment (36), not evaluated (17), and incomplete information of causes of TB death (159) (Figure 1). Of the 1794 TB deaths, 668 (37.2%) categorized as TB-related and 1126 (62.8%) as non-TB-related deaths. Generally, TB patients who died during treatment were older with a mean age of 52.0 years (sd ± 19.8) compared with the controls with a mean age of 39.0 years (SD ± 17.6). Sociodemographic, comorbid, disease, and treatment characteristics of cases and controls are presented in Table 1.
Figure 1.
Selection of cases and controls from TB registry 2014–2018.
Table 1.
Characteristics of cases and controls.
| Variables | TB Death | Control n (%) |
p-Value | ||
|---|---|---|---|---|---|
| TB-Related n (%) |
Non-TB Related n (%) |
All TB Death n (%) |
|||
| Sociodemographic Characteristics | |||||
| Age (mean ± SD), years | 44.6 ± 20.8 | 56.4 ± 17.8 | 52.0 ± 19.8 | 39.0 ± 17.6 | <0.001 ** |
| Age groups | |||||
| 0–14 | 34 (5.1) | 15 (1.3) | 49 (2.7) | 78 (3.9) | <0.001 |
| 15–44 | 303 (45.4) | 250 (22.2) | 553 (30.8) | 1156 (57.8) | |
| 45–64 | 202 (30.2) | 460 (40.9) | 662 (36.9) | 585 (29.3) | |
| ≥65 | 129 (19.3) | 401 (35.6) | 530 (29.5) | 181 (9.1) | |
| Gender | |||||
| Female | 265 (39.7) | 353 (31.3) | 618 (34.4) | 817 (40.9) | <0.001 |
| Male | 403 (60.3) | 773 (68.7) | 1176 (65.6) | 1183 (59.2) | |
| Place of residence | |||||
| Rural | 462 (70.0) | 778 (69.6) | 1240 (69.7) | 1462 (73.1) | 0.010 |
| Urban | 198 (30.0) | 340 (30.4) | 538 (30.3) | 527 (26.5) | |
| Nationality | |||||
| Malaysian | 379 (56.7) | 937 (83.2) | 1316 (73.4) | 1484 (74.2) | 0.555 |
| Non-Malaysian | 289 (43.3) | 189 (16.8) | 478 (26.6) | 516 (25.8) | |
| Education level | |||||
| Tertiary | 8 (1.2) | 16 (1.4) | 24 (1.3) | 49 (2.5) | <0.001 |
| Secondary | 136 (20.4) | 350 (31.1) | 486 (27.1) | 841 (42.1) | |
| Primary | 128 (19.2) | 291 (25.8) | 419 (23.4) | 462 (23.1) | |
| No formal education | 396 (59.3) | 469 (41.7) | 865 (48.2) | 648 (32.4) | |
| Income regular | |||||
| No | 504 (75.4) | 780 (69.3) | 1284 (71.6) | 1182 (59.1) | <0.001 |
| Yes | 164 (24.6) | 346 (30.7) | 510 (28.4) | 818 (40.9) | |
| Comorbid characteristics | |||||
| Smoking status | |||||
| No | 497 (74.4) | 777 (69.0) | 1274 (71.0) | 1405 (70.3) | 0.606 |
| Yes | 171 (25.6) | 349 (31.0) | 520 (29.0) | 595 (29.8) | |
| Diabetes mellitus status | |||||
| No | 620 (92.8) | 953 (84.6) | 1573 (87.7) | 1834 (91.7) | <0.001 |
| Yes | 48 (7.2) | 173 (15.4) | 221 (12.3) | 166 (8.3) | |
| HIV status | |||||
| Negative | 643 (97.4) | 1022 (92.2) | 1665 (94.1) | 1967 (98.4) | <0.001 |
| Positive | 17 (2.6) | 87 (7.8) | 104 (5.9) | 32 (1.6) | |
| Disease characteristics | |||||
| Category of TB | |||||
| New case | 613 (91.8) | 1003 (89.1) | 1616 (90.1) | 1884 (94.2) | <0.001 |
| Retreatment case | 55 (8.2) | 123 (10.9) | 178 (9.9) | 116 (5.8) | |
| Type of case detection | |||||
| Passive | 625 (93.6) | 1015 (90.1) | 1640 (91.4) | 1801 (90.1) | 0.148 |
| Active | 39 (5.8) | 111 (9.9) | 154 (8.6) | 199 (10.0) | |
| BCG scar | |||||
| No | 363 (54.3) | 482 (42.8) | 845 (47.1) | 604 (30.2) | <0.001 |
| Yes | 305 (45.7) | 644 (57.2) | 949 (52.9) | 1396 (69.8) | |
| TB anatomical site | |||||
| PTB | 602 (90.1) | 932 (82.8) | 1534 (85.5) | 1779 (89.0) | 0.001 |
| Extra PTB | 66 (9.9) | 194 (17.2) | 260 (14.5) | 221 (11.1) | |
| Initial sputum AFB | |||||
| Positive | 541 (82.7) | 681 (63.1) | 1222 (70.5) | 1434 (71.7) | 0.032 |
| Negative | 113 (17.3) | 398 (36.9) | 511 (29.5) | 512 (25.6) | |
| CXR status at diagnosis | |||||
| Minimal/normal | 181 (28.1) | 510 (46.8) | 691 (39.9) | 1103 (57.1) | <0.001 |
| Moderate | 336 (52.2) | 474 (43.5) | 810 (46.7) | 741 (38.4) | |
| Advanced | 127 (19.7) | 105 (9.6) | 232 (13.4) | 87 (4.5) | |
| MDRTB status | |||||
| No | 480 (99.6) | 839 (99.9) | 1319 (99.8) | 1691 (100.0) | 0.084 * |
| Yes | 2 (0.4) | 1 (0.1) | 3 (0.2) | 0 | |
| TB meningitis | |||||
| No | 67 (58.8) | 205 (77.9) | 272 (72.1) | 255 (93.1) | <0.001 |
| Yes | 47 (41.2) | 58 (22.1) | 105 (27.9) | 19 (6.9) | |
| TB miliary | |||||
| No | 93 (81.6) | 232 (88.2) | 325 (86.2) | 268 (97.8) | <0.001 |
| Yes | 21 (18.4) | 31 (11.8) | 52 (13.8) | 6 (2.2) | |
| Treatment characteristics | |||||
| Treatment regime (intensive phase) | |||||
| FDC | 341 (51.0) | 563 (50.0) | 904 (50.4) | 1428 (71.4) | <0.001 |
| Loose tablet | 327 (49.0) | 563 (50.0) | 890 (49.6) | 572 (28.6) | |
| Treatment delay (≥14 days) | |||||
| No | 659 (98.7) | 1105 (98.1) | 1764 (98.3) | 1975 (98.8) | 0.277 |
| Yes | 9 (1.3) | 21 (1.9) | 30 (1.7) | 25 (1.3) | |
| DOT achieved (intensive phase) | |||||
| No | 268 (41.4) | 329 (29.8) | 597 (34.1) | 7 (0.4) | <0.001 |
| Yes | 380 (58.6) | 774 (70.2) | 1154 (65.9) | 1993 (99.7) | |
p-value to compare all TB death cases with controls using Chi-square test. * Fisher’s exact test, ** Independent t-test; BCG = Bacillus Calmette–Guérin; CXR = Chest radiograph; FDC = Fixed-dose combination; DOT = Directly observed therapy.
3.2. Trend of TB Death
The TB death rate increased from 9.0/100,000 population in 2014 to 11.4/100,000 population in 2018 while the TB notification rate showed a small reduction over the same period. The proportion of TB-related deaths from total TB deaths remained around 37% each year during the study period (Table 2). All-cause TB deaths were mainly not directly related to the TB disease itself.
Table 2.
Trend of TB death, 2014–2018.
| Year | 2014 | 2015 | 2016 | 2017 | 2018 | Total |
|---|---|---|---|---|---|---|
| Total TB cases registered | 4747 | 4464 | 4953 | 5105 | 5008 | 24,277 |
| Total TB death cases | 312 | 327 | 415 | 473 | 426 | 1953 |
| Non-TB-related death | 198 | 213 | 255 | 241 | 219 | 1126 |
| TB-related death | 114 | 110 | 148 | 168 | 128 | 668 |
| TB-related death (%) | 36.5% | 34.1% | 36.7% | 41.1% | 36.9% | 37.2% |
| TB notification rate (per 100,000 population) |
136.2 | 126.0 | 137.5 | 139.4 | 134.6 | 134.7 * |
| TB death rate (per 100,000 population) ** |
9.0 | 9.2 | 11.5 | 12.9 | 11.4 | 10.8 * |
| Case fatality rate (%) | 6.6% | 7.3% | 8.4% | 9.3% | 8.5% | 8.0% |
* Average. ** Chi-square for trend = 21.83, p < 0.001.
3.3. Timing of TB Death
Of the 1794 TB deaths, 142 (7.9%) occurred before anti-TB treatment could be started. These deaths happened soon after TB diagnosis was made or during TB work-up before TB diagnosis was confirmed. Most (74.9%) of all TB deaths occurred in the first two months of TB diagnosis (Figure 2). The median time to death for all patients who died during anti-TB treatment was 21.5 days (IQR 5.0–61.0). The median time to death for patients who died of TB-related and non-TB-related causes was 7.0 days (IQR 2.0–30.0) and 34.0 days (IQR 11.0–81.0), respectively. Being a non-citizen (aOR = 2.32, 95% CI 1.44–3.72) and having far-advanced lung lesion on chest radiography (CXR) (aOR = 3.43, 95% CI 2.02–5.84) increased the odds of dying before anti-TB treatment started. The risk factors for TB death before treatment started are shown in Table 3.
Figure 2.
Frequency of TB-related, non-TB-related, and all TB deaths by timing of death.
Table 3.
Binary logistic regression analysis for risk factors of all-cause TB death before treatment started.
| Variables | Time of TB Death | Univariable Analysis | Multivariable Analysis | |
|---|---|---|---|---|
| After Treatment Started, n (%) | Before Treatment Started, n (%) | cOR (95% CI) | aOR (95% CI) | |
| Age (mean ± SD), years | 45.2 ± 19.8 | 44.1 ± 19.8 | 0.99 (0.98–1.01) | |
| Age groups | ||||
| 0–14 | 43 (87.8) | 6 (12.2) | 1 | 1 |
| 15–44 | 487 (88.1) | 66 (11.9) | 0.95 (0.39–2.37) | 0.71 (0.27–1.84) |
| 45–64 | 618 (93.4) | 44 (6.6) | 0.51 (0.21–1.26) | 0.83 (0.31–2.21) |
| ≥65 | 504 (95.1) | 26 (4.9) | 0.37 (0.14–0.95) | 0.87 (0.30–2.48) |
| Gender | ||||
| Female | 562 (90.9) | 56 (9.1) | 1 | 1 |
| Male | 1090 (92.7) | 86 (7.3) | 0.79 (0.56–1.13) | 1.01 (0.68–1.50) |
| Place of residence | ||||
| Rural | 1147 (92.5) | 93 (7.5) | 1 | 1 |
| Urban | 493 (91.6) | 45 (8.4) | 1.13 (0.78–1.63) | 1.19 (0.79–1.78) |
| Nationality | ||||
| Malaysian | 1242 (94.4) | 74 (5.6) | 1 | 1 |
| Non-Malaysian | 410 (85.8) | 68 (14.2) | 2.78 (1.97–3.94) | 2.32 (1.44–3.72) |
| Diabetes mellitus status | ||||
| No | 1443 (91.7) | 130 (8.3) | 1 | 1 |
| Yes | 209 (94.6) | 12 (5.4) | 0.64 (0.35–1.17) | 0.76 (0.36–1.62) |
| HIV status | ||||
| Negative | 1543 (92.7) | 122 (7.3) | 1 | 1 |
| Positive | 93 (89.4) | 11 (10.6) | 1.50 (0.78–2.87) | 3.66 (1.80–7.43) |
| Type of case detection | ||||
| Passive | 1507 (91.9) | 133 (8.1) | 1 | 1 |
| Active | 145 (94.2) | 9 (5.8) | 0.70 (0.35–1.41) | 0.36 (0.13–0.99) |
| BCG scar | ||||
| No | 765 (90.5) | 80 (9.5) | 1 | 1 |
| Yes | 887 (93.5) | 62 (6.5) | 0.67 (0.47–0.94) | 0.78 (0.49–1.23) |
| CXR status at diagnosis | ||||
| Minimal/normal | 650 (94.1) | 41 (5.9) | 1 | 1 |
| Moderately advanced | 753 (93.0) | 57 (7.0) | 1.20 (0.79–1.82) | 1.53 (0.99–2.36) |
| Far-advanced | 203 (87.5) | 29 (12.5) | 2.27 (1.37–3.74) | 3.43 (2.02–5.84) |
3.4. Causes of TB Death
Sepsis is both responsible for TB-related (22.3%) and non-TB-related (17.9%) death. In TB-related death, sepsis occurred because of underlying advanced TB or disseminated TB. In contrast, in non-TB-related death, it was primarily because of the patient also having multiple comorbidities. Advanced TB or having disseminated TB at presentation is the main cause of TB-related death, accounting for 62.9% of all causes of TB-related death. Pneumonia, existing chronic lung disease, cancer, AIDS, cardiac events, and other comorbidities are common specific causes of non-TB-related death. The details on specific causes of TB-related and non-TB-related death are shown in Table 4.
Table 4.
Specific causes of TB-related and non-TB-related death.
| Cause of TB Death | Frequency | % |
|---|---|---|
| TB-related death | ||
| Advanced/disseminated | 420 | 62.9% |
| Sepsis | 149 | 22.3% |
| TB meningitis | 42 | 6.3% |
| TB miliary | 10 | 1.5% |
| Acute respiratory distress syndrome | 10 | 1.5% |
| Underlying HIV | 17 | 2.5% |
| Severe pneumonia | 6 | 0.9% |
| MDRTB | 2 | 0.3% |
| Others | 12 | 1.8% |
| Sub-total | 668 | 100.0% |
| Non-TB-related death | ||
| Sepsis | 201 | 17.9% |
| Pneumonia-HAP | 117 | 10.4% |
| Pneumonia-CAP | 48 | 4.3% |
| Pneumonia-other | 95 | 8.4% |
| Lung-other | 52 | 4.6% |
| Lung-COAD | 20 | 1.8% |
| Carcinoma-lung | 48 | 4.3% |
| Carcinoma-other | 93 | 8.3% |
| AIDS | 94 | 8.3% |
| Heart-acute coronary syndrome | 91 | 8.1% |
| Heart-congestive heart failure | 23 | 2.0% |
| Heart-other | 9 | 0.8% |
| Stroke | 29 | 2.6% |
| Meningitis-bacterial, other | 20 | 1.8% |
| Upper gastrointestinal bleeding | 25 | 2.2% |
| Liver failure–drug induced | 15 | 1.3% |
| End stage renal failure | 15 | 1.3% |
| Injury/accident | 9 | 0.8% |
| Other | 41 | 3.6% |
| Indetermined | 81 | 7.2% |
| Subtotal | 1126 | 100.0% |
3.5. Risk Factors of TB Death
In multivariable binary logistic regression analysis, in which all-cause TB deaths were compared to the controls, older age, male, socioeconomically disadvantaged, human immunodeficiency virus (HIV) positive, and advanced TB disease were independent risk factors for all-cause TB death. Patients on anti-TB treatment with FDC and with their treatment supervision under DOT achieved, are less likely to die from all-cause TB death. The details of risk factors for all-cause TB death are shown in Table 5.
Table 5.
Binary logistic regression analysis of risk factors for all-cause TB death.
| Variables | Univariable Analysis | Multivariable Analysis |
|---|---|---|
| cOR (95% CI) | aOR (95% CI) | |
| Age groups | ||
| 0–14 | 1 | 1 |
| 15–44 | 0.76 (0.53–1.10) | 0.97 (0.58–1.62) |
| 45–64 | 1.80 (1.24–2.62) | 2.61 (1.56–4.35) |
| ≥65 | 4.66 (3.14–6.92) | 7.46 (4.39–12.67) |
| Gender | ||
| Female | 1 | 1 |
| Male | 1.31 (1.15–1.50) | 1.17 (0.97–1.38) |
| Place of residence | ||
| Rural | 1 | 1 |
| Urban | 1.20 (1.04–1.39) | 1.47 (1.22–1.77) |
| Income regular | ||
| No | 1.74 (1.52–1.99) | 1.46 (1.22–1.77) |
| Yes | 1 | 1 |
| HIV status | ||
| Negative | 1 | 1 |
| Positive | 3.84 (2.57–5.74) | 5.61 (3.51–8.97) |
| Category of TB | ||
| New case | 1 | 1 |
| Retreatment case | 1.79 (1.40–2.28) | 1.63 (1.21–2.21) |
| TB anatomical site | ||
| PTB | 1 | 1 |
| Extra PTB | 1.36 (1.13–1.65) | 1.86 (1.43–2.42) |
| CXR status at diagnosis | ||
| Minimal/normal | 1 | 1 |
| Moderately advanced | 1.75 (1.52–2.00) | 1.71 (1.43–2.05) |
| Far-advanced | 4.26 (3.27–5.54) | 4.56 (3.31–6.28) |
| Treatment regime (intensive phase) | ||
| FDC | 1 | 1 |
| Loose tablet | 2.46 (2.15–2.81) | 2.11 (1.78–2.52) |
| DOT (intensive phase) | ||
| No | 1 | 1 |
| Yes | 0.007 (0.003–0.014) | 0.007 (0.003–0.014) |
Hosmer–Lemeshow test, p = 0.156; Nagelkerke pseudo R2 = 0.481.
In multinominal logistic regression analysis in which TB-related and non-TB-related deaths were compared to the controls, no formal education, smoking, and passive case detection were found to be independent risk factors for TB-related deaths but not for non-TB-related deaths. Non-citizen was found to be an independent risk factor for non-TB-related death but not for TB-related death. Older age was found to be an independent risk factor for both TB-related and non-TB-related deaths, but the effect is larger in non-TB-related deaths. Those aged 65 years and above experienced increased odds (aOR = 14.86, 95% CI 7.55–29.27) of dying compared to those aged 14 years and below from non-TB-related causes. However, the same age group experienced less odds of dying (aOR = 3.02, 95% CI 1.62–5.65) from TB-related causes. The other risk factors for both TB-related and non-TB related deaths are shown in Table 6.
Table 6.
Multinominal logistic regression analysis for association between sociodemographic, disease, and treatment characteristics with TB-related and non-TB-related death.
| Variables | TB-Related Death | Non-TB Related Death | ||
|---|---|---|---|---|
| Univariable Analysis | Multivariable Analysis | Univariable Analysis |
Multivariable Analysis |
|
| cOR (95% CI) | aOR (95% CI) | cOR (95% CI) | aOR (95% CI) | |
| Age groups | ||||
| 0–14 | 1 | 1 | 1 | 1 |
| 15–44 | 0.60 (0.39–0.92) | 0.87 (0.48–1.56) | 1.13 (0.64–1.99) | 1.53 (0.79–2.97) |
| 45–64 | 0.79 (0.51–1.22) | 1.31 (0.72–2.38) | 4.09 (2.32–7.20) | 5.26 (2.71–10.18) |
| ≥65 | 1.64 (1.03–2.59) | 3.18 (1.70–5.94) | 11.52 (6.45–20.57) | 14.89 (7.57–29.32) |
| Nationality | ||||
| Citizen | 1 | 1 | 1 | 1 |
| Non-citizen | 2.19 (1.83–2.63) | 1.86 (1.46–2.37) | 0.58 (0.48–0.70) | 0.71 (0.56–0.89) |
| Place of residence | ||||
| Rural | 1 | 1 | 1 | 1 |
| Urban | 1.19 (0.98–1.44) | 1.54 (1.21–1.94) | 1.21 (1.03–1.43) | 1.47 (1.20–1.79) |
| Income regular | ||||
| No | 2.13 (1.75–2.59) | 1.63 (1.28–2.07) | 1.56 (1.34–1.82) | 1.32 (1.08–1.60) |
| Yes | 1 | 1 | 1 | 1 |
| Smoking | ||||
| No | 1 | 1 | 1 | 1 |
| Yes | 0.81 (0.67–0.99) | 0.73 (0.57–0.93) | 1.06 (0.91–1.24) | 0.96 (0.78–1.17) |
| Category of TB | ||||
| New case | 1 | 1 | 1 | 1 |
| Retreatment case | 1.46 (1.04–2.03) | 1.59 (1.08–2.36) | 1.99 (1.53–2.59) | 1.80 (1.31–2.47) |
| Type of case detection | ||||
| Active | 1 | 1 | 1 | 1 |
| Passive | 1.61 (1.14–2.26) | 1.57 (1.02–2.40) | 0.99 (0.78–1.26) | 1.02 (0.75–1.40) |
| CXR status at diagnosis | ||||
| Minimal/normal | 1 | 1 | 1 | 1 |
| Moderately advanced | 2.76 (2.26–3.39) | 2.27 (1.79–2.88) | 1.38 (1.18–1.62) | 1.23 (1.02–1.49) |
| Far-advanced | 8.89 (6.49–12.19) | 6.58 (4.56–9.50) | 2.61 (1.93–3.53) | 2.47 (1.74–3.53) |
| Treatment regime (intensive phase) | ||||
| FDC | 1 | 1 | 1 | 1 |
| Loose tablet | 2.39 (2.00–2.87) | 2.15 (1.72–2.68) | 2.49 (2.15–2.91) | 2.41 (2.00–2.90) |
| DOT (intensive phase) | ||||
| No | 1 | 1 | 1 | 1 |
| Yes | 0.005 (0.002–0.011) | 0.006 (0.003–0.013) | 0.008 (0.004–0.018) | 0.008 (0.004–0.018) |
Pearson Chi-square goodness of fit, p = 0.364; Nagelkerke pseudo R2 = 0.458.
4. Discussion
Our analysis of TB surveillance data collected over five years from 2014 to 2018 showed an increasing trend of case fatality rate each year. The case fatality rate had increased from 6.6% in 2014 to 8.5% in 2018. The increasing fatality rate is not in line with the WHO End TB strategy milestone target to reduce TB deaths by 35% in 2020 compared to 2015. A majority (62.8%) of all TB deaths were due to non-TB-related causes, consistent with a study conducted among hospitalized TB patients in Taiwan [7]. Generally, TB deaths mainly involved older male patients who came from a low socioeconomic background.
Starting anti-TB treatment immediately after TB diagnosis may improve treatment outcomes. However, this study found that 7.9% of all TB deaths occurred before treatment started, occurring during TB work-up, or soon after TB diagnosis was made. In some instances, treatment cannot be started immediately at the time of diagnosis. It usually takes a few days for the patients to come back to the clinic after their TB work-up confirms their diagnosis. TB deaths before treatment started were associated with non-citizen and those with advanced lung lesions from their initial CXR. Advanced lung lesions at presentation indicate a delay in seeking treatment. As non-citizens contribute about 30% of all TB deaths, special attention is needed for this sub-group population. To detect TB early among this population, active case detection needs to be considered.
In this study, the first two months of treatment is a critical time as a majority of all-cause mortality among TB patients (74.9%) occurred during this period. This finding is similar to studies conducted in Ethiopia [8] and South Korea [9]. To address the high risk of death during the intensive phase of treatment, close monitoring and effective clinical management of TB patients need to be emphasized. In this study, the median time to death for all TB deaths of 21.5 days (IQR 5.0–61.0) is slightly shorter duration compared with 23 days (IQR 8–48) among TB patients in Taiwan [7]. The median time to death for patients who died of TB directly (7.0 days IQR: 2.0–30.0) was shorter than patients who died of non-TB-related causes (34.0 days IQR: 11.0–81.0). These findings are similar to the median survival time among TB patients in Taiwan [7] and the United States of America [10]. TB-related deaths were mainly due to advanced or disseminated TB at presentation, whereas non-TB-related deaths were due to existing comorbidities and sepsis.
In the multivariable binary logistic analysis, being male and older age were significantly associated with all-cause TB deaths, consistent with findings reported in Taiwan [11] and Nigeria [12]. Those staying in an urban area were at higher odds of dying while on anti-TB treatment. This observation could be due to the presence of a large migrant population in major towns in Sabah. Many of them with undocumented status face difficulties accessing standard healthcare, thus resulting in delaying their TB diagnosis and treatment. Studies elsewhere have also reported that marginalized populations often living in poverty experience a higher rate of TB mortality [12,13,14]. Disease characteristics such as retreatment cases, extra pulmonary tuberculosis, and advanced lesion on CXR were found to be independent risk factors for all-cause TB deaths. These TB disease characteristics at diagnosis indicate that patients seek healthcare late. Special TB control activities are needed for these marginalized populations to improve TB case detection and subsequently reduce TB deaths. Besides starting anti-TB treatment as early as possible, the choice of delivery and type of drugs given need special consideration to improve TB treatment outcomes. In this study, we found that patients on the fixed-dose combination (FDC) compared to loose tablets of anti-TB, and when DOT supervision was achieved were associated with reduced odds of dying. However, patients with TB–HIV co-infection were six times higher odds of dying compared to those who were HIV negative. Strengthening TB/HIV control program integration is crucial to improve case management of TB–HIV co-infected patients. Studies conducted elsewhere that showed favorable treatment outcomes of between 79.8% [5] and 89.8% [15] can be achieved with the targeted clinical management of TB–HIV co-infected patients, specifically for those with other risks.
The multinominal logistic regression analysis found similarities and differences in factors associated with TB-related and non-TB-related deaths. Older age is significantly associated with TB-related and non-TB-related deaths, but the effect was higher in non-TB-related deaths. This finding suggests that TB patients who died of non-TB-related causes are much older with multiple comorbidities compared with much younger TB patients who died of TB-related causes. Non-citizens had a 31% increased odds of dying from TB-related causes but not from non-TB-related causes. The high odds of dying from TB-related causes could be due to non-citizens presenting with more advanced disease. We also found that TB diagnosed through passive case detection was a significantly independent risk factor for TB-related death but not for non-TB-related death. Thus, to reduce deaths due to TB-related causes, more emphasis should be given to active case findings in addition to existing passive case findings to ensure TB diagnosis can be made in the early stage of the disease.
Some limitations in our study include the possibility of misclassification of the causes of TB deaths into either TB-related or non-TB-related. However, this misclassification is minimized as all included TB deaths in this study had been audited by a special committee at the district and state level. The actual number of TB deaths could be underestimated as we excluded deaths without information on the specific cause of death and those lost to follow-up. The strength of our study is that we analyze state-wide population-based surveillance data; thus, the findings can be generalized to the Sabah population and other states in Malaysia with similar population characteristics and TB burden. We also found important factors associated with TB deaths that could be used to develop a predictive model for TB mortality.
5. Conclusions
The yearly TB notification rate was relatively constant, but the TB mortality rate increased each year between 2014 and 2018. The majority of TB deaths occurred in the first two months of treatment and were attributed equally to TB-related and non-TB-related causes. TB-related deaths were primarily due to advanced disease or disseminated TB, whereas non-TB-related deaths were primarily due to existing comorbidities and sepsis resulting in septic shock. Many important independent risk factors for TB deaths were identified, which are useful to re-strategize TB control programs to address the increasing TB mortality rate. These factors can also be used to develop a predictive model for TB mortality. A predictive model is useful to identify TB patients at higher risk of mortality early so that specific interventions can be rendered accordingly.
Acknowledgments
The authors would like to thank the Director General of Health Malaysia for his permission to publish this article.
Author Contributions
Conceptualization and methodology, R.A., Y.C.L.; data collection and processing, Y.C.L.; data analysis, R.A.; writing the manuscript, R.A.; reviewing the manuscript, R.A., Y.C.L. All authors have read and agreed to the published version of the manuscript.
Funding
This research received no external funding.
Institutional Review Board Statement
Ethical approval for this study was obtained from the Medical Ethics Committee of the University Sabah Malaysia on 25 September 2020 JKEtika 4/20 (8), and from the Medical Research and Ethics committee (MREC), Ministry of Health Malaysia on 14 September 2020 (approval code NMRR-20-1289-55461).
Informed Consent Statement
The Sabah State Health Director granted permission to access the de-identified case-listing of TB patients from the MyTB database, and patient consent was not obtained because this study only uses patient’s information from the MyTB database.
Data Availability Statement
The data used to produce the findings of this study are available from the author, R.A., but will not be made available publicly as the authors have no permission to do so from the Ministry of Health Malaysia.
Conflicts of Interest
The authors declare no conflict of interest.
Footnotes
Publisher’s Note: MDPI stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.
References
- 1.World Health Organization . Global Tuberculosis Report 2020. WHO; Geneva, Switzerland: 2020. [Google Scholar]
- 2.World Health Organisation . The End TB Strategy. WHO; Geneva, Switzerland: 2015. [Google Scholar]
- 3.Goroh M.M.D., Rajahram G.S., Avoi R., Boogaard C.H.A.V.D., William T., Ralph A.P., Lowbridge C. Epidemiology of tuberculosis in Sabah, Malaysia, 2012–2018. Infect. Dis. Poverty. 2020;9:119. doi: 10.1186/s40249-020-00739-7. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Waitt C.J., Squire S. A systematic review of risk factors for death in adults during and after tuberculosis treatment [Review article] Int. J. Tuberc. Lung Dis. 2011;15:871–885. doi: 10.5588/ijtld.10.0352. [DOI] [PubMed] [Google Scholar]
- 5.Massavirov S., Akopyan K., Abdugapparov F., Ciobanu A., Hovhanessyan A., Khodjaeva M., Gadoev J., Parpieva N. Risk Factors for Unfavorable Treatment Outcomes among the Human Immunodeficiency Virus-Associated Tuberculosis Population in Tashkent City, Uzbekistan: 2013–2017. Int. J. Environ. Res. Public Health. 2021;18:4623. doi: 10.3390/ijerph18094623. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 6.Tok P.S.K., Liew S.M., Wong L.P., Razali A., Loganathan T., Chinna K., Ismail N., Kadir N.A. Determinants of unsuccessful treatment outcomes and mortality among tuberculosis patients in Malaysia: A registry-based cohort study. PLoS ONE. 2020;15:e0231986. doi: 10.1371/journal.pone.0231986. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Lin C.-H., Kuo Y.-W., Wang J.-Y., Hsu C.-L., Chen J.-M., Cheng W.-C., Lee L.-N. Tuberculosis mortality: Patient characteristics and causes. BMC Infect. Dis. 2014;14:5. doi: 10.1186/1471-2334-14-5. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Birlie A., Tesfaw G., Dejene T., Woldemichael K. 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. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Min J., Kim J.S., Kim H.W., Shin A.Y., Koo H.-K., Lee S.-S., Kim Y.-K., Shin K.-C., Chang J.H., Chun G., et al. Clinical profiles of early and tuberculosis-related mortality in South Korea between 2015 and 2017: A cross-sectional study. BMC Infect Dis. 2019;19:735. doi: 10.1186/s12879-019-4365-9. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Beavers S.F., Pascopella L., Davidow A., Mangan J.M., Hirsch-Moverman Y.R., Golub J.E., Blumberg H.M., Webb R.M., Royce R.A., Buskin S.E., et al. Tuberculosis Mortality in the United States: Epidemiology and Prevention Opportunities. Ann. Am. Thorac. Soc. 2018;15:683–692. doi: 10.1513/AnnalsATS.201705-405OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Yen Y.-F., Feng J.-Y., Pan S.-W., Chuang P.-H., Su V.Y.-F., Su W.-J. Determinants of mortality in elderly patients with tuberculosis: A population-based follow-up study. Epidemiol. Infect. 2017;145:1374–1381. doi: 10.1017/S0950268817000152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Adamu A.L., Gadanya M.A., Abubakar I.S., Jibo A.M., Bello M.M., Gajida A.U., Babashani M.M., Abubakar I. High mortality among tuberculosis patients on treatment in Nigeria: A retrospective cohort study. BMC Infect. Dis. 2017;17:170. doi: 10.1186/s12879-017-2249-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 13.Arcoverde M.A.M., Berra T.Z., Alves L.S., Dos Santos D.T., Belchior A.D.S., Ramos A.C.V., Arroyo L.H., De Assis I.S., Alves J.D., De Queiroz A.A.R., et al. How do social-economic differences in urban areas affect tuberculosis mortality in a city in the tri-border region of Brazil, Paraguay and Argentina. BMC Public Health. 2018;18:795. doi: 10.1186/s12889-018-5623-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Bhargava A., Bhargava M. Tuberculosis deaths are predictable and preventable: Comprehensive assessment and clinical care is the key. J. Clin. Tuberc. Other Mycobact. Dis. 2020;19:100155. doi: 10.1016/j.jctube.2020.100155. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Pinsai S. 1342. Impact of HIV Infection on Treatment Outcome of New Tuberculosis Patients Attending Tuberculosis and Antiretroviral Treatment Services in the Community-Based Hospital, Thailand: A Retrospective Cohort Study. Open Forum Infect. Dis. 2019;6:S485–S486. doi: 10.1093/ofid/ofz360.1206. [DOI] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Data Availability Statement
The data used to produce the findings of this study are available from the author, R.A., but will not be made available publicly as the authors have no permission to do so from the Ministry of Health Malaysia.