INTRODUCTION
Tuberculosis (TB) is a leading cause of death by an infectious disease worldwide and among people living with HIV (PLWH).1–3 PLWH are 15–21 times more likely to develop active TB and more likely to die from it when they do, compared with people without HIV.1 In 2020, a total of 1.5 million people died of TB, including 214,000 PLWH.1,4 As of 2020, the estimated TB incidence in sub-Saharan Africa was 220 per 100,000 population per year, and the estimated TB incidence among PLWH was significantly higher at 2279 per 100,000 population per year.1
Given that TB disproportionately affects PLWH, proper diagnosis, treatment, and prevention of TB is critical. While 85% of those newly initiated on TB treatment in 2018 achieved treatment success, a large gap remains between the number of diagnosed TB cases and the number of estimated cases.1 Despite advances in TB diagnostics, TB among PLWH is still underdiagnosed because of atypical signs, symptoms, and the paucibacillary nature of the disease in this population.1,5 Diagnosing TB in PLWH is particularly challenging because one of the main diagnostic methods for TB until recently, GeneXpert MTB/RIF, was only 82%–88% sensitive in PLWH.5–7 Although the WHO has recommended Xpert MTB/RIF Ultra replace Xpert MTB/RIF in 2017, utilization of this more sensitive assay has been slow in resource-limited settings.8 As such, a comprehensive approach to TB diagnosis using bacteriological testing in conjunction with clinical assessments are paramount to the identification and treatment of TB, particularly in lower resource settings.
Continual programmatic assessments have highlighted gaps in HIV/TB service delivery in a variety of settings that have informed resource allocation and prioritization of activites.9–11 This has allowed HIV/TB programs to evolve from increasing TB case detection among PLWH, to scale-up of TB preventive treatment.9–12 The African Cohort Study (AFRICOS) is uniquely positioned to characterize progress in TB diagnosis and prevention in PLWH across 4 high HIV-TB burden African countries (Uganda, Kenya, Nigeria, and Tanzania), given participants are enrolled from the President's Emergency Plan for AIDS Relief (PEPFAR)–supported programs and can serve as a proxy to assess strengths and gaps in current PEPFAR HIV-TB programmatic strategies. We assessed the prevalence and incidence of HIV-TB co-occurring in AFRICOS and identified factors associated with prevalent and incident TB to better understand the current drivers of HIV-TB co-occurrence and gaps in TB screening and diagnosis. In addition, we characterized the current methods of diagnosis in prevalent cases to inform strategies to optimize active case finding to treat and subsequently prevent TB transmission.
METHODS
Study Design and Setting
AFRICOS is a prospective, observational cohort study, enrolling PLWH and HIV-uninfected participants at 12 PEFPAR supported sites across 5 programs in Uganda, Kenya (South Rift Valley and Kisumu), Tanzania, and Nigeria, as previously described.13 Individuals were eligible for enrollment if they were aged 15 years or older and consented to data and specimen collection (see Figure S1, Supplemental Digital Content, https://links.lww.com/QAI/C12).
Laboratory Methods
Laboratory assessments included quantification of CD4 T-lymphocyte count and viral load (VL) (copies/mL). Sputum samples were collected from all participants annually, regardless of symptoms, or at any visit that a participant presented with any of the cardinal TB symptoms of cough, fever, night sweats, or weight loss. Samples from study initiation in 2013 to December 2021 were evaluated for active TB and for rifampicin resistance using the Cepheid Gene Xpert MTB/RIF platform. Additional clinical diagnostics included mycobacterial culture and molecular or culture-based drug resistance testing.
Data Collection and Definitions
On enrollment and at subsequent visits every 6 months, PLWH completed a physical examination, medical history, sociodemographic questionnaire, TB symptom screening, and phlebotomy. Participants were classified as having a history of TB if they had a TB diagnostic WHO code abstracted from their medical records before enrollment. Demographic variables collected include sex, age, marital status, education, employment status, number of residents in household, year of enrollment (dichotomized into before vs after 2017 to reflect the time of PEPFAR program wide scale-up of isoniazid preventive therapy), and clinical site. HIV-specific variables included antiretroviral therapy (ART) use (yes, no) and regimen abstracted from medical records, self-reported ART adherence in the past month (no missed ART doses, missed 1 doses), duration on ART, length of time in HIV clinical care, length of time since HIV diagnosis, CD4 count (<200 cells/mm3, 200 cells/mm3), VL (on ART for less than 6 months, on ART for 6 or more months and VL <1000 copies/mL and on ART for 6 or more months and VL ≥1000 copies/mL), TB diagnosis method (bacteriological, clinical), hyperglycemia, and body mass index (BMI). Additional variables included in the analysis were substance use and incarceration status. Definitions and categorizations of analytic variables not specified here have been previously described and summarized in Table S1, Supplemental Digital Content, https://links.lww.com/QAI/C13.13
Active TB was defined as meeting one of the following criteria: (1) bacteriologically confirmed through smear microscopy, culture, or WHO-approved rapid diagnostics (including GeneXpert MTB/RIF), (2) clinically indicated and having initiated combination therapy for active TB in the absence of bacteriological confirmation, or (3) identified by medical record abstraction within 3 months of enrollment. Participants were considered to be on combination therapy for active TB at enrollment if they were receiving (1) rifampicin (RIF), isoniazid (INH), ethambutol, and pyrazinamide or (2) INH and RIF for the final 4 months of treatment for active TB. Participants solely prescribed INH-based TB regimens were considered to be on preventative therapy.
We determined TB prevalence at entry or within 3 months of enrollment into AFRICOS, counting (1) previous diagnoses (those receiving continued combination TB therapy); (2) diagnoses made because of initial testing on entry into the cohort and within 3 months of enrollment; and (3) diagnosis based on WHO or ICD-10 codes in medical records at entry or within 3 months of enrollment.
Statistical Methods
Descriptive statistics using Pearson χ2 tests were used to determine significant differences in clinical and sociodemographic variables among participants with prevalent TB, compared with those without TB, at or within 3 months of enrollment. Logistic regression was used to estimate unadjusted and adjusted odds ratios (aORs) and 95% confidence intervals (CIs) for associations between clinical and sociobehavioral factors and prevalent TB disease.
Incidence rates (IRs) were calculated for participants living with HIV without TB at or within 3 months of enrollment, as the number of new TB diagnoses divided by person-years (PY) of follow-up. CIs were calculated using the quadratic approximation to the Poisson log likelihood for the log–rate parameter. Cox proportional hazard models were used to assess unadjusted and adjusted hazard ratios and 95% CIs for associations between time to incident TB and clinical and sociobehavioral predictor variables. Time-varying covariates were accounted for in the model. Participants were censored at competing events (either loss to follow-up or death). Time at risk of TB started at entry into the cohort, and failure or censor date was observed over the duration of the study at follow-up visits.
All variables that were associated with prevalent or incident TB (P < 0.2) in the bivariate analysis were included in the respective multivariable analysis in addition to the following variables identified a priori: age, sex, and clinical site. Additionally, we tested for multicollinearity; variables with variance inflation factors greater than 10 were removed from the multivariable model. Missing data were folded into the reference category for modeling. All Analyses were performed in SAS software, version 9.3 (SAS Institute, Cary, NC) and Stata software, version 16.0 (StataCorp, College Station, TX)
Ethical Clearances
The study was approved by Institutional Review Boards of the Walter Reed Army Institute of Research, Makerere University School of Public Health, Kenya Medical Research Institute, Tanzania National Institute of Medical Research, and Nigerian Ministry of Defense.
RESULTS
TB Prevalence and Diagnosis Methods
From 21 January 2013 to 1 December 2021, 3171 PLWH were enrolled, and 93 (2.9%) had TB at or within 3 months of the enrollment visit. The median age of our analytic sample was 37.4 (Interquartile range: 29.3–45.5) years, and 1858 individuals (58.6%) were female (Table 1). A higher proportion of PLWH with prevalent TB was aged 25–39 years, male, and had primary or some secondary education. A greater proportion of PLWH with TB was significantly underweight, had CD4 counts <200 cells/mm3, was on ART for <6 months, and was incarcerated compared with their TB-free counterparts. Of the 93 cases of prevalent HIV-TB co-occurrence identified, most of them (65.6%) were bacteriologically confirmed (Table 1). Of the 132 participants who underwent RIF resistance testing, 3% (n = 4) had resistance, of which 2 were cases with prevalent TB.
TABLE 1. -
Enrollment Characteristics of PLWH by TB Co-occurrence Status
|
Total n = 3171 |
No TB n = 3078 |
TB Case n = 93 |
P
*
|
| Age (years)†
|
|
|
|
<0.001
|
| 15–24 |
539 (17.0) |
528 (17.2%) |
11 (11.8%) |
|
| 25–39 |
1346 (42.4%) |
1288 (41.8%) |
58 (62.4%) |
|
| 40–49 |
808 (25.5%) |
790 (25.7%) |
18 (19.4%) |
|
| 50+ |
477 (15.0%) |
471 (15.3%) |
6 (6.5%) |
|
| Sex |
|
|
|
0.014
|
| Male |
1313 (41.4%) |
1263 (41.0%) |
50 (53.8%) |
|
| Female |
1858 (58.6%) |
1815 (59.0%) |
43 (46.2%) |
|
| Clinical site |
|
|
|
0.32 |
| Kayunga, Uganda |
553 (17.4%) |
539 (17.5%) |
14 (15.1%) |
|
| South Rift Valley, Kenya |
1095 (34.5%) |
1059 (34.4%) |
36 (38.7%) |
|
| Kisumu West, Kenya |
563 (17.8%) |
547 (17.8%) |
16 (17.2%) |
|
| Mbeya, Tanzania |
607 (19.1%) |
585 (19.0%) |
22 (23.7%) |
|
| Abuja & Lagos, Nigeria |
353 (11.1%) |
348 (11.3%) |
5 (5.4%) |
|
| Marital status†
|
|
|
|
0.33 |
| Not married |
1503 (47.4%) |
1464 (47.6%) |
39 (41.9%) |
|
| Married |
1666 (52.5%) |
1613 (52.4%) |
53 (57.0%) |
|
| Education†
|
|
|
|
0.010
|
| None or some primary |
1014 (32.0%) |
996 (32.4%) |
18 (19.4%) |
|
| Primary or some secondary |
1281 (40.4%) |
1231 (40.0%) |
50 (53.8%) |
|
| Secondary and above |
874 (27.6%) |
850 (27.6%) |
24 (25.8%) |
|
| Currently employed†
|
|
|
|
0.59 |
| No |
1981 (62.5%) |
1921 (62.4%) |
60 (64.5%) |
|
| Yes |
1188 (37.5%) |
1156 (37.6%) |
32 (34.4%) |
|
| Year enrolled |
|
|
|
0.24 |
| 2013–2017 |
2773 (87.4%) |
2688 (87.3%) |
85 (91.4%) |
|
| 2018–2020 |
398 (12.6%) |
390 (12.7%) |
8 (8.6%) |
|
| Total # people in household†
|
|
|
|
0.99 |
|
3 |
1005 (31.7%) |
976 (31.7%) |
29 (31.2%) |
|
| >3-6 |
1489 (47.0%) |
1445 (46.9%) |
44 (47.3%) |
|
| >6 |
669 (21.1%) |
650 (21.1%) |
19 (20.4%) |
|
| Consume alcohol†
|
|
|
|
0.78 |
| No |
2582 (81.4%) |
2506 (81.4%) |
76 (81.7%) |
|
| Yes |
587 (18.5%) |
571 (18.6%) |
16 (17.2%) |
|
| Smoker†
|
|
|
|
0.95 |
| No |
3026 (95.4%) |
2938 (95.5%) |
88 (94.6%) |
|
| Yes |
142 (4.5%) |
138 (4.5%) |
4 (4.3%) |
|
| Ever been incarcerated†
|
|
|
|
0.007
|
| No |
2846 (89.8%) |
2771 (90.0%) |
75 (80.6%) |
|
| Yes |
322 (10.2%) |
305 (9.9%) |
17 (18.3%) |
|
| Hyperglycemia†,‡
|
|
|
|
0.89 |
| No |
2826 (89.1%) |
2744 (89.1%) |
82 (88.2%) |
|
| Yes |
296 (9.3%) |
287 (9.3%) |
9 (9.7%) |
|
| Time since HIV diagnosis†
|
|
|
|
<0.001
|
| <1 yr |
1177 (37.1%) |
1118 (36.3%) |
59 (63.4%) |
|
| 1-5 yrs |
794 (25.0%) |
771 (25.0%) |
23 (24.7%) |
|
| >5 yrs |
1157 (36.5%) |
1146 (37.2%) |
11 (11.8%) |
|
| Duration in HIV care†
|
|
|
|
<0.001
|
| <6 mo |
1056 (33.3%) |
997 (32.4%) |
59 (63.4%) |
|
| 6 months to <2 yrs |
424 (13.4%) |
413 (13.4%) |
11 (11.8%) |
|
| ≥2 yrs |
1665 (52.5%) |
1642 (53.3%) |
23 (24.7%) |
|
| Duration on ART†
|
|
|
|
<0.001
|
| ART-naïve |
900 (28.4%) |
862 (28.0%) |
38 (40.9%) |
|
| <6 mo |
437 (13.8%) |
411 (13.4%) |
26 (28.0%) |
|
| 6 months to <2 yrs |
433 (13.7%) |
423 (13.7%) |
10 (10.8%) |
|
| 2 years to <4 yrs |
354 (11.2%) |
347 (11.3%) |
7 (7.5%) |
|
| ≥4 yrs |
1042 (32.9%) |
1030 (33.5%) |
12 (12.9%) |
|
| ART regimen |
|
|
|
0.002
|
| AZT/NVP/3 TC |
509 (16.1%) |
504 (16.4%) |
5 (5.4%) |
|
| AZT/EFV/3 TC |
168 (5.3%) |
165 (5.4%) |
3 (3.2%) |
|
| TDF/NVP/3 TC |
176 (5.6%) |
174 (5.7%) |
2 (2.2%) |
|
| PI-based |
191 (6.0%) |
187 (6.1%) |
4 (4.3%) |
|
| TLE |
974 (30.7%) |
936 (30.4%) |
38 (40.9%) |
|
| TLD |
204 (6.4%) |
203 (6.6%) |
1 (1.1%) |
|
| Other |
49 (1.5%) |
47 (1.5%) |
2 (2.2%) |
|
| ART-naïve |
900 (28.4%) |
862 (28.0%) |
38 (40.9%) |
|
| Missed doses ART (past month)†
|
|
|
|
0.002
|
| Not on ART |
900 (28.4%) |
862 (28.0%) |
38 (40.9%) |
|
| No missed doses ART |
1937 (61.1%) |
1885 (61.2%) |
52 (55.9%) |
|
| Missed 1+ doses ART |
333 (10.5%) |
331 (10.8%) |
2 (2.2%) |
|
| Body mass index (BMI)†
|
|
|
|
0.002
|
| Underweight |
372 (11.7%) |
344 (11.2%) |
28 (30.1%) |
|
| Normal |
2022 (63.8%) |
1967 (63.9%) |
55 (59.1%) |
|
| Overweight/obese |
770 (24.3%) |
761 (24.7%) |
9 (9.7%) |
|
| CD4 count†
|
|
|
|
<0.001
|
| <200 cells/mm3
|
586 (18.5%) |
547 (17.8%) |
39 (41.9%) |
|
| ≥200 cells/mm3
|
2546 (80.3%) |
2492 (81.0%) |
54 (58.1%) |
|
| Viral load†
|
|
|
|
|
| <6 months on ART |
1333 (42.0%) |
1269 (41.2%) |
64 (68.8%) |
<0.001
|
| <1000 copies/mL |
1581 (49.9%) |
1559 (50.6%) |
22 (23.7%) |
|
| ≥1000 copies/mL |
216 (6.8%) |
210 (6.8%) |
6 (6.5%) |
|
| History of pulmonary TB
§
|
|
|
|
<0.001
|
| No |
2957 (93.3%) |
2893 (94.0%) |
64 (68.8%) |
|
| Yes |
214 (6.7%) |
185 (6.0%) |
29 (31.2%) |
|
| TB diagnosis method¦
|
|
|
|
— |
| Bacteriological |
|
|
61 (65.6%) |
|
| Clinical |
|
|
32 (34.4%) |
|
Data are presented as n (column %).
Abbreviations: PLWH, people living with HIV; TB, tuberculosis; ABC, abacavir; 3 TC, lamivudine; AZT, azidothymidine (zidovudine); TDF, tenofovir; TLE, tenofovir/lamivudine/efavirenz; TLD, tenofovir/lamivudine/dolutegravir.
*P values were calculated using Pearson χ2 tests; Bold indicates significance at P < 0.05.
†Percentages do not add up to 100 due to missing values.
‡Hyperglycemia was defined as a fasting glucose >99 mg/dL, nonfasting glucose >199 mg/dL, or receipt of hypoglycemic medications.
§History of pulmonary TB was defined as any of the following WHO codes abstracted from the participant's medical record: W38.1 (Positive Gene Expert), W36.1 (Smear +), or W50.0.
¦Bacteriological confirmation of TB was diagnosed by GeneXpert MTB/RIF; clinical diagnosis of TB was indicated if the participant was receiving combination therapy for active TB in the absence of bacteriological confirmation.
Factors Associated With Prevalent TB
After adjustment, the odds of prevalent TB were significantly higher among those who had completed primary school/had some secondary-level education and secondary-level education and above, compared with those with no/some primary education (Table 2). The adjusted odds of TB were 2.59 times greater among PLWH who had been diagnosed with HIV for 1–5 years compared with PLWH diagnosed with HIV for >5 years (95% CI: 1.17–5.71). Compared with PLWH with a normal BMI, there were greater adjusted odds of TB among those who were underweight (aOR 2.51, 95% CI: 1.52–4.15). Those with a CD4 count <200 cells/mm3 had higher adjusted odds of TB compared with those with higher CD4 counts (aOR 1.89, 95% CI: 1.18–3.03). Associations between prevalent TB and age, sex, incarceration status, ART regimen, and VL did not persist after adjustment (Table 2).
TABLE 2. -
Factors Associated With Prevalent TB at Enrollment Among PLWH
|
OR (95% CI) |
P
|
aOR (95% CI)†
|
P
|
| Age (years) |
|
|
|
|
| 15–24*
|
Ref |
|
— |
|
| 25–39 |
2.17 (1.23–4.16)
|
0.02
|
1.50 (0.73–3.08) |
0.27 |
| 40–49 |
1.10 (0.51–2.34) |
0.81 |
0.86 (0.37–1.97) |
0.71 |
| 50+ |
0.61 (0.23–1.67) |
0.34 |
0.46 (0.16–1.35) |
0.15 |
| Sex |
|
|
|
|
| Female |
Ref |
|
— |
|
| Male |
1.67 (1.11–2.53)
|
0.02
|
1.35 (0.85–2.15) |
0.20 |
| Clinical site |
|
|
|
|
| Kayunga, Uganda |
Ref |
|
— |
|
| South Rift Valley, Kenya |
1.31 (0.70–2.45) |
0.40 |
1.52 (0.74–3.11) |
0.25 |
| Kisumu West, Kenya |
1.13 (0.54–2.33) |
0.74 |
1.49 (0.66–3.36) |
0.34 |
| Mbeya, Tanzania |
1.45 (0.73–2.86) |
0.28 |
1.52 (0.71–3.28) |
0.28 |
| Abuja & Lagos, Nigeria |
0.55 (0.20–1.55) |
0.26 |
0.62 (0.19–1.97) |
0.41 |
| Education |
|
|
|
|
| None or some primary*
|
Ref |
|
— |
|
| Primary or some secondary |
2.13 (1.25–3.64)
|
0.006
|
2.08 (1.17–3.69)
|
0.01
|
| Secondary and above |
1.48 (0.81–2.72) |
0.20 |
2.03 (1.02–4.05)
|
0.04
|
| Ever been incarcerated |
|
|
|
|
| No*
|
Ref |
|
— |
|
| Yes |
2.03 (1.19–3.49)
|
0.01
|
1.78 (0.95–3.34) |
0.07 |
| Time since HIV diagnosis |
|
|
|
|
| >5 yrs*
|
Ref |
|
— |
|
| 1–5 yrs |
3.22 (1.56–6.65)
|
0.002
|
2.59 (1.17–5.71)
|
0.01
|
| <1 yr |
5.70 (2.98–10.91)
|
<0.001
|
2.48 (0.99–6.21) |
0.05 |
| Duration in HIV care |
|
|
|
|
| ≥2 yrs*
|
Ref |
|
— |
|
| 6 months to <2 yrs |
1.09 (0.44–2.70)
|
<0.07
|
|
|
| <6 mo |
4.29 (2.63–6.99)
|
<0.001
|
|
|
| ART regimen |
|
|
|
|
| TLE |
Ref |
|
— |
|
| TLD |
0.12 (0.02–0.89)
|
0.03
|
0.23 (0.03–1.87) |
0.17 |
| AZT/NVP/3 TC |
0.24 (0.10–0.63)
|
0.003
|
0.66 (0.23–1.89) |
0.43 |
| AZT/EFV/3 TC |
0.45 (0.14–1.47) |
0.18 |
0.91 (0.25–3.27) |
0.88 |
| TDF/NVP/3 TC |
0.28 (0.07–1.18) |
0.08 |
0.57 (0.12–2.62) |
0.46 |
| PI-Based |
0.53 (0.19–1.49) |
0.22 |
1.08 (0.33–3.57) |
0.89 |
| Other |
1.05 (0.25–4.48) |
0.94 |
1.34 (0.26–6.89) |
0.72 |
| ART naïve |
1.09 (0.69–1.72) |
0.72 |
0.70 (0.40–1.22) |
0.20 |
| Body mass index (BMI) |
|
|
|
|
| Normal*
|
Ref |
|
— |
|
| Underweight |
2.87 (1.80–4.58)
|
<0.001
|
2.51 (1.52–4.15)
|
<0.001
|
| Overweight/obese |
0.42 (0.21–0.85)
|
0.01
|
0.48 (0.23–0.99)
|
0.049
|
| CD4 count |
|
|
|
|
| ≥200 cells/mm3
*
|
Ref |
|
— |
|
| <200 cells/mm3
|
3.34 (2.19–5.10)
|
<0.001
|
1.89 (1.18–3.03)
|
0.009
|
| Viral load |
|
|
|
|
| <1000 copies/mL*
|
Ref |
|
— |
|
| <6 months on ART |
3.51 (2.17–5.68)
|
<0.001
|
1.88 (0.85–4.16) |
0.11 |
| ≥1000 copies/mL |
1.99 (0.80–4.93) |
0.13 |
1.45 (0.55–3.83) |
0.45 |
Bold indicates significance at P < 0.05. Abbreviations: TB, tuberculosis; PLWH, people living with HIV; OR, odds ratio; aOR, adjusted odds ratio; CI: confidence interval; ART, antiretroviral therapy, ABC, abacavir; 3 TC, lamivudine; AZT, azidothymidine (zidovudine); TDF, tenofovir; TLE, tenofovir/lamivudine/efavirenz; TLD, tenofovir/lamivudine/dolutegravir.
*Missing data were folded into reference category.
†Model building: includes age, sex, site plus all predictors with a P value of <0.2 in the bivariate analyses; dropped duration in HIV care because of collinearity; removed missed doses and duration on ART because analysis was not restricted to PLWH on ART.
TB Incidence
There were 79 incident cases of TB and 13,161 PY postenrollment over the course of the study, only 3 of whom were ART-naïve during entry into AFRICOS. The median observation time per person was 2.98 years (Interquartile range: 1.51–4.50). The overall TB IR among PLWH was 600 per 100,000 PY (95% CI: 481–748) (Table 3). When stratified by program, South Rift Valley, Kenya, had the highest TB IR with 793 cases per 100,000 PY (95% CI: 574–1094), followed by Kisumu, Kenya (IR: 703, 95% CI: 443–1116) and Uganda (IR: 518, 95% CI: 301–892). Comparatively, Tanzania (IR: 339, 95% CI: 162–711) and Nigeria (IR: 295, 95% CI: 111–786) had the lowest TB IRs.
TABLE 3. -
Crude Incidence Rate of TB Among PLWH per 100,000 Person-Years
|
TB Incidence Rate (95% CI) |
| Overall |
600 (481–748) |
| Program |
|
| Kayunga, Uganda |
518 (301–892) |
| South Rift Valley, Kenya |
793 (574–1094) |
| Kisumu West, Kenya |
703 (443–1116) |
| Mbeya, Tanzania |
339 (162–711) |
| Abuja & Lagos, Nigeria |
295 (111–786) |
Abbreviations: TB, tuberculosis; PLWH, people living with HIV; CI: confidence interval.
Factors Associated With Incident TB
After adjustment, the risk of incident TB among PLWH with <5 years since their HIV diagnosis was nearly 2 times higher than PLWH ≥5 years since their HIV diagnosis (95% CI: 1.06–3.35; Table 4). PLWH on dolutegravir/lamivudine/tenofovir (TLD) had a 78% lower risk of incident TB compared with those on tenofovir/lamivudine/efavirenz (TLE) (95% CI: 0.08–0.63). Compared with PLWH who had a normal BMI, those classified as underweight had 2.28 times the risk of experiencing TB (95% CI: 1.27–4.10). PLWH with a CD4 count of <200 cells/mm3 had 2.85 times the risk of TB compared with those with higher CD4 counts (95% CI: 1.58–5.13). In addition, PLWH on ART for <6 months had 5 times the risk of TB disease (95% CI: 2.09–11.93) compared with those who were virally suppressed. None of the variables in the adjusted model interacted significantly with time (global test for PH assumption: P = 0.96).
TABLE 4. -
Factors Associated With Incident TB Among PLWH (n = 79)
*
|
HR (95% CI) |
P
|
aHR (95% CI)‡
|
P
|
| Age (years)¦,
§
|
|
|
|
|
| 15–39†
|
Ref |
|
— |
|
| 40–49 |
1.38 (0.84–2.27) |
0.20 |
1.60 (0.96–2.67) |
0.07 |
| 50+ |
0.87 (0.47–1.62) |
0.66 |
1.00 (0.52–1.92) |
0.99 |
| Sex |
|
|
|
|
| Female |
Ref |
|
— |
|
| Male |
1.41 (0.91–2.19) |
0.13 |
1.24 (0.77–1.98) |
0.38 |
| Clinical site |
|
|
|
|
| Kayunga, Uganda |
Ref |
|
— |
|
| South Rift Valley, Kenya |
1.53 (0.81–2.89) |
0.19 |
1.92 (0.99–3.72) |
0.054 |
| Kisumu West, Kenya |
1.36 (0.67–2.79) |
0.40 |
1.65 (0.78–3.50) |
0.19 |
| Mbeya, Tanzania |
0.62 (0.25–1.57) |
0.31 |
0.82 (0.32–2.12) |
0.68 |
| Abuja & Lagos, Nigeria |
0.57 (0.19–1.76) |
0.33 |
0.73 (0.23–2.32) |
0.59 |
| Marital status
§
|
|
|
|
|
| Married†
|
Ref |
|
|
|
| Not married |
0.80 (0.51–1.27) |
0.34 |
|
|
| Highest level of education§
|
|
|
|
|
| None or some primary†
|
Ref |
|
|
|
| Primary or some secondary |
1.11 (0.67–1.84) |
0.68 |
|
|
| Secondary and above |
0.81 (0.44–1.47) |
0.48 |
|
|
| Employment status
§
|
|
|
|
|
| Unemployed |
Ref |
|
|
|
| Employed |
0.84 (0.53–1.33) |
0.47 |
|
|
| Total no. of people in household
§
|
|
|
|
|
| ≤3†
|
Ref |
|
|
|
| 4–6 |
1.08 (0.65–1.80) |
0.78 |
|
|
| >6 |
1.13 (0.61–2.10) |
0.71 |
|
|
| Alcohol use
§
|
|
|
|
|
| No†
|
Ref |
|
|
|
| Yes |
0.75 (0.36–1.57) |
0.45 |
|
|
| Cigarette use
§
|
|
|
|
|
| No†
|
Ref |
|
|
|
| Yes |
0.95 (0.23–3.86) |
0.94 |
|
|
| Time since HIV diagnosis¦,§
|
|
|
|
|
| ≥5 years†
|
Ref |
|
— |
|
| <5 yrs |
2.08 (1.26–3.41)
|
0.004
|
1.89 (1.06–3.35)
|
0.03
|
| Duration in HIV care¦,§
|
|
|
|
|
| ≥2 years†
|
Ref |
|
|
|
| <2 yrs |
1.21 (0.64–2.27) |
0.55 |
|
|
| ART regimen
§
|
|
|
|
|
| TLE |
Ref |
|
— |
|
| TLD |
0.18 (0.06–0.51)
|
0.001
|
0.22 (0.08–0.63)
|
0.005
|
| AZT/NVP/3 TC |
0.43 (0.18–1.00) |
0.049 |
0.65 (0.26–1.63) |
0.36 |
| AZT/EFV/3 TC |
0.20 (0.03–1.45) |
0.11 |
0.28 (0.04–2.09) |
0.22 |
| TDF/NVP/3 TC |
0.87 (0.35–2.20) |
0.77 |
0.99 (0.36–2.68) |
0.98 |
| PI-based |
0.98 (0.50–1.89) |
0.94 |
1.00 (0.49–2.06) |
0.99 |
| Other |
0.80 (0.19–3.47) |
0.77 |
0.84 (0.19–3.69) |
0.82 |
| ART-naïve |
1.26 (0.45–3.58) |
0.66 |
0.47 (0.14–1.59) |
0.22 |
| Body mass index (BMI)
§
|
|
|
|
|
| Normal†
|
Ref |
|
— |
|
| Underweight |
2.37 (1.33–4.22)
|
0.003
|
2.28 (1.27–4.10)
|
0.006
|
| Overweight/obese |
0.58 (0.33–1.04) |
0.069 |
0.68 (0.37–1.25) |
0.21 |
| CD4 count
§
|
|
|
|
|
| ≥200 cells/mm3
†
|
Ref |
|
— |
|
| <200 cells/mm3
|
3.77 (2.23–6.38)
|
<0.001
|
2.85 (1.58–5.13)
|
<0.001
|
| Viral load
§
|
|
|
|
|
| <1000 copies/mL†
|
Ref |
|
— |
|
| <6 months on ART |
3.48 (1.73–7.00)
|
<0.001
|
5.00 (2.09–11.93)
|
<0.001
|
| ≥1000 copies/mL |
2.65 (1.38–5.09)
|
0.003
|
1.62 (0.80–3.30) |
0.18 |
Bold indicates significance at P < 0.05. Abbreviations: TB, tuberculosis; PLWH, people living with HIV; HR, hazard ratio; aHR, adjusted hazard ratio; CI: confidence interval; ABC, abacavir; 3TC, lamivudine; AZT, azidothymidine (zidovudine); TDF, tenofovir; TLE, tenofovir/lamivudine/efavirenz; TLD, tenofovir/lamivudine/dolutegravir.
*Participants with TB at enrollment visit did not contribute to time-to-event models.
†Missing data were folded into reference category.
‡Model building: includes age, sex, site plus all predictors with a P value of <0.2 in the bivariate analyses; removed missed doses and duration on ART because analysis was not restricted to PLWH on ART; history of TB was dropped because we did not collect information on the outcome that would distinguish between latent reinfection or a new episode of postprimary infection, which is beyond the scope of this study.
§Time-varying covariates included in the model.
¦Categories collapsed because of small cell sizes.
Sensitivity Analysis
When restricting to participants with complete case data (n = 3000), we found similar results, except for the following differences: in the primary prevalence model, an overweight/obese BMI was associated with lower odds of TB, while this was not significant in the complete case version of the model. In addition, age was not significant in the primary prevalence model; however, the 25–39 year age group showed significantly higher odds of prevalent TB in the complete case model (see Table S3, Supplemental Digital Content, https://links.lww.com/QAI/C13). In the primary time-to-event analysis, <5 years since HIV diagnosis, underweight BMI, a low CD4 count, and being on ART for <6 months were associated with incident TB, while these factors were not significant in the complete case analysis (see Table S5, Supplemental Digital Content, https://links.lww.com/QAI/C13). In addition, we observed a higher hazard for TB among the 40–49 year age group and a lower hazard for TB in Nigeria in the complete case analysis, although these variables were not significant in the primary time-to-event analysis. Given the number of cases with incident TB decreased substantially from 79 to 45 cases in the complete case analysis, thereby reducing the sample size and subsequent statistical power, we chose to report the primary analyses and included the complete case analyses (see Tables S2–5, Supplemental Digital Content, https://links.lww.com/QAI/C13).
DISCUSSION
We found that crude TB incidence varied widely between AFRICOS clinical sites, with the lowest TB IR in Nigeria and the highest in South Rift Valley, Kenya, although multivariable analysis showed no significant differences between clinical sites. Our data show substantially lower crude TB IRs compared with corresponding WHO 2020 country-level data; however, our rates are not directly comparable with the WHO estimates, given AFRICOS IRs are only representative of the PEPFAR clinic population, and not standardized across the general PLWH population, as seen with the WHO data.1 In addition, AFRICOS identified most TB diagnoses at entry or within 3 months of enrollment. This is likely due to active diagnostic testing, regardless of symptoms, and a broad and conservative case definition of TB. Contrary to most studies, we did not observe significant age or sex differences in cases with prevalent or incident TB after adjusting for sociodemographic and clinical factors.14–23 Regarding sex, this finding could be because AFRICOS enrolled PLWH from HIV care and treatment facilities and men, who are more likely to have TB disease, are less likely to be enrolled in HIV care when HIV-infected, and are comparatively under-represented in the AFRICOS cohort.20,21 Focused interventions and differentiated service delivery models targeting men may support more consistent follow-up that could help minimize persistent gender gaps in TB diagnosis and treatment among PLWH.
Higher education was associated with prevalent TB. Contrary to the literature, we found that those with higher levels of educational attainment had twice the odds of prevalent TB compared with those with little to no education.24–26 This association is likely confounded by occupation, which may be vastly different in rural vs urban settings. For example, those with lower education in rural areas may be more likely to work outdoors in occupations such as farming, with a lower risk of TB exposure.27 By contrast, in urban settings, lower education may be associated with higher TB risk occupations such as factory work.24 As such, more research on the interaction between occupation and settings is required to understand this finding.
A shorter length of time since HIV diagnosis among PLWH was associated with higher odds of prevalent TB. This is likely because the shorter the length of time an individual has been aware of their positive HIV status, the greater the opportunity for TB co-ocurrence due to suboptimal immune function from delayed HIV treatment. Our results further support this explanation, given that PLWH with CD4 counts <200 cells/mm3 were at greater risk and higher odds of incident and prevalent TB, respectively, compared with those with higher CD4 counts, which is well established in the literature.28 A low CD4 count is often a result of delayed HIV diagnosis and late treatment initiation, which increases the risk of opportunistic infections such as TB. Furthermore, our findings also show that PLWH on ART for less than 6 months had 5 times the risk of TB acquisition compared with those who were on sustained ART and virally suppressed. Thus, there is a need to promote timely HIV testing and immediate ART initiation to further reduce the risk of TB and associated morbidity and mortality. Strategies improving linkage to sustained HIV care and treatment to facilitate long-term viral suppression are critical to reduce immunodeficiency and prevent TB.29
The currently preferred first-line TLD regimen was associated with lower incident TB compared with the standard TLE regimen, which most study participants were on when these data were analyzed, although the effect of TLD itself is difficult to isolate given national TB incident rates were dropping, primarily in Kenya, where most patients were enrolled. However, the efficacy on dolutegravir-based regimens have been well characterized in randomized controlled trials, and systematic reviews have shown that TLD was better tolerated, had fewer adverse events and side effects, and higher treatment adherence compared with TLE.30–34 As such, PLWH on TLD may be more likely to have sustained viral suppression due to improved treatment adherence, which enhances immune activity and could reduce TB acquisition. In addition, the scale-up of multimonth dispensing (MMD) of ART and the practices of mask wearing and social distancing may have resulted in less exposure to TB. Specific longitudinal analyses that account for changes in local TB incidence are required to assess the effect of TLD vs TLE on decreased TB incidence, and more data will be required to understand the effect of COVID-19, mitigation practices, and MMD on TB diagnosis.
PLWH in our cohort who were underweight had a higher risk of incident TB and odds of prevalent TB compared with those with a normal BMI. These findings are well established in the literature and consistent with other studies from low- and middle-income countries (LMIC) that found a bidirectional association between low BMI and TB.35–40 TB disease is associated with a catabolic state and alters appetite mediators, which can result in substantial weight loss, while malnutrition is also associated with an elevated risk of developing HIV-associated TB.35–40 While nutrition-specific interventions have been implemented in LMIC as part of TB treatment since 2013, there seem to be several patient-, facility-, and health system-level barriers that influence the uptake of these interventions given we see low BMI persist among those with TB in our study.41,42 Long-term, sustainable, systemic changes affecting social determinants of health, such as food assistance in food insecure settings, access to social services, and health systems strengthening, are required to reduce barriers to nutrition-based interventions.41 While nutritional supplementation may be helpful in TB risk reduction, more empirical evidence is needed to assess the efficacy of nutrition on TB acquisition and recovery.43–45
Among all cases with prevalent TB identified at enrollment, 66% were bacterially confirmed and 34% identified by their receipt of combination therapy for active TB. Given the limited sensitivity of GeneXpert and the high mortality of untreated TB in PLWH, methodologies using clinical and therapeutic metrics should be considered to augment diagnostics.46 As such, clinicians should be encouraged to treat TB when clinical suspicion is high, even when test results are negative. Further evaluation of the cost effectiveness of new diagnostic algorithms (including symptom screening, point-of-care C-reactive protein, Determine TB-LAM, Xpert MTB/RIF Ultra, and culture) to improve TB case detection while controlling costs at PEFPAR clinical sites is warranted.47
We acknowledge several limitations of this study. First, there may be selection bias because PLWH who volunteer to participate may be fundamentally different from those who do not. Additionally, participants in the study have better access to TB screening and diagnostic services, given rigorous study follow-up, and may not be representative of the general population of clients receiving care at PEPFAR facilities. Regarding diagnosis methods, participants identified as having active TB through their receipt of combination therapy may have been misclassified, potentially inflating the number of prevalent TB cases observed. Given that this is an observational study, there is potential for unmeasured confounders that distort associations between the factors we observed to be associated with prevalent TB. Additionally, due to the cross-sectional analysis of prevalent TB, causality of exposure on outcome may be difficult to discern. We were also unable to assess the effect of pre/post Treat All on TB incidence. Given Treat All implementation was rolled out over several years from 2015, delineating participants into pre/post Treat All was difficult because participants enrolled into AFRICOS at different points in time. However, we did use CD4 count and time to ART initiation as a proxy measure for Treat All in our analysis. Finally, we were unable to account for latent TB disease and isoniazid prophylaxis in these analyses.
CONCLUSIONS
This study characterizes progress in TB diagnosis and treatment among PLWH in 4 high HIV/TB burden countries in sub-Saharan Africa. Insight from this PEPFAR-supported study can be used to identify gaps in current programmatic activities and inform programmatic priorities, which will aid in reducing HIV-TB co-occurrence and subsequent morbidity and mortality among PLWH. Given the programmatic emphasis by PEPFAR and the WHO on HIV testing and counseling of patients with TB, linkages to HIV care and treatment, and intensified TB case finding and infection control, assessment of the state of the HIV-TB epidemic and its drivers are critical to evaluating the success of these efforts.1
Early HIV diagnosis and immediate ART initiation among PLWH can reduce the risk of acquiring TB, mitigating the impact and burden of TB disease, particularly among those in their most productive years. Transitioning to TLD and sustaining viral suppression should continue to be prioritized, given their marked impact on reducing the risk of HIV-TB co-occurrence in this cohort. There is still a need to scale-up and implement nutrition-based interventions to prevent and manage TB among PLWH. Efforts to intensify screening, identification, and notification could promote early TB diagnosis and responsive management and treatment.
DISCLAIMER
The investigators have adhered to the policies for the protection of human subjects as prescribed in AR 70–25. The views expressed are those of the authors and should not be construed to represent the positions of the US Army or the Department of Defense. This work was supported by the PEPFAR through a cooperative agreement between the Henry M. Jackson Foundation for the Advancement of Military Medicine, Inc., and the US Department of Defense [W81XWH-11-2-0174, W81XWH-18-2-0040]. Research reported in this publication was also supported by the National Institute of Allergy and Infectious Diseases of the National Institutes of Health under Award Number T32AI114398. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
ACKNOWLEDGMENTS
The authors thank the study participants, local implementing partners, and hospital leadership at Kayunga District Hospital, Kericho District Hospital, AC Litein Mission Hospital, Kapkatet District Hospital, Tenwek Mission Hospital, Kapsabet District Hospital, Nandi Hills District Hospital, Kisumu West District Hospital, Mbeya Zonal Referral Hospital, Mbeya Regional Referral Hospital, Defence Headquarters Medical Center, and the 68th Nigerian Army Reference Hospital.
The authors also thank the AFRICOS Group—from the US Military HIV Research Program Headquarters Group: Danielle Bartolanzo, Alexus Reynolds, Katherine Song, Mark Milazzo, Leilani Francisco, Steven Schech, Badryah Omar, Tsedal Mebrahtu, Elizabeth Lee, Kimberly Bohince, Ajay Parikh, Jaclyn Hern, Emma Duff, Kara Lombardi, Michelle Imbach, and Leigh Anne Eller; from the AFRICOS Uganda Group: Hannah Kibuuka, Michael Semwogerere, Prossy Naluyima, Godfrey Zziwa, Allan Tindikahwa, Claire Nakazzi Bagenda, Hilda Mutebe, Cate Kafeero, Enos Baghendaghe, William Lwebuge, Freddie Ssentogo, Hellen Birungi, Josephine Tegamanyi, Paul Wangiri, Christine Nabanoba, Phiona Namulondo, Richard Tumusiime, Ezra Musingye, Christina Nanteza, Joseph Wandege, Michael Waiswa, Evelyn Najjuma, Olive Maggaga, Isaac Kato Kenoly, and Barbara Mukanza; from the AFRICOS South Rift Valley, Kenya Group: Jonah Maswai, Rither Langat, Aaron Ngeno, Lucy Korir, Raphael Langat, Francis Opiyo, Alex Kasembeli, Christopher Ochieng, Japhet Towett, Jane Kimetto, Brighton Omondi, Mary Leelgo, Michael Obonyo, Linner Rotich, Enock Tonui, Ella Chelangat, Joan Kapkiai, Salome Wangare, Zeddy Bett Kesi, Janet Ngeno, Edwin Langat, Kennedy Labosso, Joshua Rotich, Leonard Cheruiyot, Enock Changwony, Mike Bii, Ezekiel Chumba, Susan Ontango, Danson Gitonga, Samuel Kiprotich, Bornes Ngtech, Grace Engoke, Irene Metet, Alice Airo, and Ignatius Kiptoo; from the AFRICOS Kisumu, Kenya Group: John Owuoth, Valentine Sing'oei, Winne Rehema, Solomon Otieno, Celine Ogari, Elkanah Modi, Oscar Adimo, Charles Okwaro, Christine Lando, Margaret Onyango, Iddah Aoko, Kennedy Obambo, Joseph Meyo, and George Suja; from the AFRICOS Abuja, Nigeria Group: Michael Iroezindu, Yakubu Adamu, Nnamdi Azuakola, Mfreke Asuquo, Abdulwasiu Bolaji Tiamiyu, Afoke Kokogho, Samirah Sani Mohammed, Ifeanyi Okoye, Sunday Odeyemi, Aminu Suleiman, Lawrence C. Umeji, Onome Enas, Miriam Ayogu, Ijeoma Chigbu-Ukaegbu, Wilson Adai, Felicia Anayochukwu Odo, Rabi Abdu, Roseline Akiga, Helen Nwandu, Chisara Sylvestina Okolo, Ogundele Taiwo, Otene Oche Ben, Nicholas Innocent Eigege, Tony Ibrahim Musa, Juliet Chibuzor Joseph, Ndubuisi C. Okeke; from the AFRICOS Lagos, Nigeria Group: Zahra Parker, Nkechinyere Elizabeth Harrison, Uzoamaka Concilia Agbaim, Olutunde Ademola Adegbite, Ugochukwu Linus Asogwa, Adewale Adelakun, Chioma Ekeocha, Victoria Idi, Rachel Eluwa, Jumoke Titilayo Nwalozie, Igiri Faith, Blessing Irekpitan Wilson, Jacinta Elemere, Nkiru Nnadi, Francis Falaju Idowu, Ndubuisi Rosemary, Amaka Natalie Uzeogwu, Theresa Owanza Obende, Ifeoma Lauretta Obilor, Doris Emekaili, Edward Akinwale, and Inalegwu Ochai; from the AFRICOS Mbeya, Tanzania Group: Lucas Maganga, Emmanuel Bahemana, Samoel Khamadi, John Njegite, Connie Lueer, Abisai Kisinda, Jaquiline Mwamwaja, Faraja Mbwayu, Gloria David, Mtasi Mwaipopo, Reginald Gervas, Dorothy Mkondoo, Nancy Somi, Paschal Kiliba, Ephrasia Mwalongo, Gwamaka Mwaisanga, Johnisius Msigwa, Hawa Mfumbulwa, Peter Edwin, and Willyhelmina Olomi.
REFERENCES
1. World Health Organization. Global Tuberculosis Reports 2021. Geneva, Switzerland: World Health Organization; 2021.
https://www.who.int/teams/global-tuberculosis-programme/tb-reports/global-tuberculosis-report-2021. Accessed November 28, 2021.
2. Lozano R, Naghavi M, Foreman K, et al. Global and regional mortality from 235 causes of death for 20 age groups in 1990 and 2010: a systematic analysis for the Global Burden of Disease Study 2010. Lancet. 2012;380:2095–2128.
3. Ortblad KF, Lozano R, Murray CJ. The burden of HIV: insights from the global burden of disease study 2010. AIDS. 2013;27:2003–2017.
4. World Health Organization. Tuberculosis. Geneva, Switzerland: World Health Organization; 2021. Available at:
https://www.who.int/news-room/fact-sheets/detail/tuberculosis Accessed Decemeber 1, 2021. Published.
5. MacGregor-Fairlie M, Wilkinson S, Besra GS, Goldberg Oppenheimer P. Tuberculosis diagnostics: overcoming ancient challenges with modern solutions. Emerging Top Life Sci. 2020;4:435–448.
6. WHO Guidelines Approved by the Guidelines Review Committee. Automated real-time nucleic acid amplification technology for rapid and simultaneous detection of tuberculosis and rifampicin resistance: Xpert MTB/RIF assay for the diagnosis of pulmonary and extrapulmonary TB in adults and children: policy update. Geneva, Switzerland: World Health Organization; 2013. Available at:
https://apps.who.int/iris/handle/10665/112472.
7. Zifodya JS, Kreniske JS, Schiller I, et al. Xpert Ultra versus Xpert MTB/RIF for pulmonary tuberculosis and rifampicin resistance in adults with presumptive pulmonary tuberculosis. Cochrane Database Syst Rev. 2021;2:CD009593.
8. World Health Organization. Next-generation Xpert® MTB/RIF Ultra assay recommended by WHO; 2017. Available at:
https://www.who.int/news/item/25-03-2017-next-generation-xpert-mtb-rif-ultra-assay-recommended-by-who. Accessed February 1, 2022.
9. Hamada Y, Getahun H, Tadesse BT, Ford N. HIV-associated tuberculosis. Int J STD AIDS. 2021;32:780–790.
10. Shah HD, Nazli Khatib M, Syed ZQ, et al. Gaps and interventions across the diagnostic care cascade of TB patients at the level of patient, community and health system: a qualitative review of the literature. Trop Med Infect Dis. 2022;7:136.
11. Rewari BB, Kumar A, Mandal PP, Puri AK. HIV TB coinfection - perspectives from India. Expert Rev Respir Med. 2021;15:911–930.
12. Scott V, Chopra M, Azevedo V, et al. Scaling up integration: development and results of a participatory assessment of HIV/TB services, South Africa. Health Res Pol Syst. 2010;8:23.
13. Ake JA, Polyak CS, Crowell TA, et al. Noninfectious comorbidity in the african cohort study. Clin Infect Dis. 2019;69:639–647.
14. Gupta A, Wood R, Kaplan R, et al. Tuberculosis incidence rates during 8 years of follow-up of an antiretroviral treatment cohort in South Africa: comparison with rates in the community. PLoS One. 2012;7:e34156.
15. Liu E, Makubi A, Drain P, et al. Tuberculosis incidence rate and risk factors among HIV-infected adults with access to antiretroviral therapy. AIDS. 2015;29:1391–1399.
16. Gunda DW, Maganga SC, Nkandala I, et al.
Prevalence and risk factors of active TB among adult HIV patients receiving ART in northwestern Tanzania: a retrospective cohort study. Can J Infect Dis Med Microbiol. 2018;2018:1–7.
17. Enos M, Sitienei J, Ong'ang'o J, et al. Kenya tuberculosis
prevalence survey 2016: challenges and opportunities of ending TB in Kenya. PLoS One. 2018;13:e0209098.
18. Dodd PJ, Looker C, Plumb ID, et al. Age- and sex-specific social contact patterns and incidence of Mycobacterium tuberculosis infection. Am J Epidemiol. 2016;183:156–166.
19. Fu H, Lin H-H, Hallett TB, Arinaminpathy N. Explaining age disparities in tuberculosis burden in Taiwan: a modelling study. BMC Infect Dis. 2020;20:191.
20. Hertz D, Schneider B. Sex differences in tuberculosis. Semin Immunopathol. 2019;41:225–237.
21. Horton KC, MacPherson P, Houben RMGJ, et al. Sex differences in tuberculosis burden and notifications in low- and middle-income countries: a systematic review and meta-analysis. PLoS Med. 2016;13:e1002119.
22. Nyamogoba H, Mbuthia G. Gender-age distribution of tuberculosis among suspected tuberculosis cases in western Kenya. Med Sci Inl Mel .l. 2018;7:1.
23. Zhu M, Han G, Takiff HE, et al. Times series analysis of age-specific tuberculosis at a rapid developing region in China, 2011–2016. Sci Rep. 2018;8:8727.
24. Bang KM, Weissman DN, Wood JM, Attfield MD. Tuberculosis mortality by industry in the United States, 1990-1999. Int J Tuberc Lung Dis. 2005;9:437–442.
25. Chenciner L, Annerstedt KS, Pescarini JM, Wingfield T. Social and health factors associated with unfavourable treatment outcome in adolescents and young adults with tuberculosis in Brazil: a national retrospective cohort study. Lancet Glob Health. 2021;9:e1380–e1390.
26. Clarke M, Dick J, Zwarenstein M, Diwan V. DOTS for temporary workers in the agricultural sector. An exploratory study in tuberculosis case detection. Curationis. 2003;26:66–71.
27. Semilan HM, Abugad HA, Mashat HM, Abdel Wahab MM. Epidemiology of tuberculosis among different occupational groups in Makkah region, Saudi Arabia. Sci Rep. 2021;11:12764.
28. Ellis PK, Martin WJ, Dodd PJ. CD4 count and tuberculosis risk in HIV-positive adults not on ART: a systematic review and meta-analysis. PeerJ. 2017;5:e4165.
29. Meehan S-A, Sloot R, Draper HR, et al. Factors associated with linkage to HIV care and TB treatment at community-based HIV testing services in Cape Town, South Africa. PloS one. 2018;13:e0195208.
30. Stellbrink HJ, Reynes J, Lazzarin A, et al. Dolutegravir in antiretroviral-naive adults with HIV-1: 96-week results from a randomized dose-ranging study. AIDS. 2013;27:1771–1778.
31. Clotet B, Feinberg J, van Lunzen J, et al. Once-daily dolutegravir versus darunavir plus ritonavir in antiretroviral-naive adults with HIV-1 infection (FLAMINGO): 48 week results from the randomised open-label phase 3b study. The Lancet. 2014;383:2222–2231.
32. Trottier B, Lake JE, Logue K, et al. Dolutegravir/abacavir/lamivudine versus current ART in virally suppressed patients (STRIIVING): a 48-week, randomized, non-inferiority, open-label, Phase IIIb study. Antivir Ther. 2017;22:295–305.
33. Walmsley SL, Antela A, Clumeck N, et al. Dolutegravir plus abacavir-lamivudine for the treatment of HIV-1 infection. N Engl J Med. 2013;369:1807–1818.
34. Kanters S, Vitoria M, Doherty M, et al. Comparative efficacy and safety of first-line antiretroviral therapy for the treatment of HIV infection: a systematic review and network meta-analysis. Lancet HIV. 2016;3:e510–e520.
35. Maro I, Lahey T, MacKenzie T, et al. Low BMI and falling BMI predict HIV-associated tuberculosis: a prospective study in Tanzania. Int J Tuberc Lung Dis. 2010;14:1447–1453.
36. Niyongabo T, Henzel D, Idi M, et al. Tuberculosis, human immunodeficiency virus infection, and malnutrition in Burundi. Nutrition. 1999;15:289–293.
37. Paton NI, Castello-Branco LR, Jennings G, et al. Impact of tuberculosis on the body composition of HIV-infected men in Brazil. J Acquired Immune Deficiency Syndromes Hum Retrovirology. 1999;20:265–271.
38. Swaminathan S, Padmapriyadarsini C, Sukumar B, et al. Nutritional status of persons with HIV infection, persons with HIV infection and tuberculosis, and HIV-negative individuals from southern India. Clin Infect Dis. 2008;46:946–949.
39. Ding Y, Raterink R-J, Marín-Juez R, et al. Tuberculosis causes highly conserved metabolic changes in human patients, mycobacteria-infected mice and zebrafish larvae. Sci Rep. 2020;10:11635.
40. Kennedy N, Ramsay A, Uiso L, et al. Nutritional status and weight gain in patients with pulmonary tuberculosis in Tanzania. Trans R Soc Trop Med Hyg. 1996;90:162–166.
41. Guideline WHO. Nutritional Care and Support for Patients with Tuberculosis. Geneva: World Health Organization; 2013.
42. Degefa MG, Bezabih AM, Kahsay ZH, Belachew AB. Barriers and facilitators of nutrition assessment, counseling, and support for tuberculosis patients: a qualitative study. BMC Nutr. 2021;7:58.
43. Si ZL, Kang LL, Shen XB, Zhou YZ. Adjuvant efficacy of nutrition support during pulmonary tuberculosis treating course: systematic review and meta-analysis. Chin Med J. 2015;128:3219–3230.
44. Zheng Y, Ma A, Wang Q, et al. Relation of leptin, ghrelin and inflammatory cytokines with body mass index in pulmonary tuberculosis patients with and without type 2 diabetes mellitus. PLoS One. 2013;8:e80122.
45. Haiqing C, Chen L, Yin C, et al. The effect of micro-nutrients on malnutrition, immunity and therapeutic effect in patients with pulmonary tuberculosis: a systematic review and meta-analysis of randomised controlled trials. Tuberculosis (Edinb). 2020;125:101994.
46. Arend SM, van Soolingen D. Performance of Xpert MTB/RIF Ultra: a matter of dead or alive. Lancet Infect Dis. 2018;18:8–10.
47. Martinez L, Andrews JR. Improving tuberculosis case finding in persons living with advanced HIV through new diagnostic algorithms. Am J Respir Crit Care Med. 2019;199:559–560.
