Abstract
Background:
Diabetes mellitus (DM) increases the risk of tuberculosis (TB) progression and poor treatment outcomes. Rising global DM prevalence presents an emerging threat to TB control. We sought to determine whether DM affects TB transmissibility.
Methods:
From 2009 to 2012, we enrolled 3,109 microbiologically-confirmed pulmonary TB patients and their 12,767 household contacts (HHCs) whom we followed for one year for the occurrence of TB infection as measured by a tuberculin skin test and for TB disease. We assessed the association between index patient DM and TB infection and disease occurrence in HHCs.
Results:
The DM status of index patients was not associated with TB infection among child HHCs (adjusted prevalence rate ratio [aPRR], 1.05; 95% CI: 0.78–1.42) or with incident TB infection at 6 months among HHCs uninfected at baseline (adjusted cumulative rate ratio [aCRR], 0.85; 95% CI: 0.66–1.09). Among the 12,442 HHCs without TB disease at baseline, 368 (3.0%) developed TB during the 12-month follow-up. HHCs exposed to an index TB patient with DM had a reduced incidence of TB disease (aCRR = 0.33, 95% CI: 0.13–0.85).
Conclusion:
In this cohort study, diabetes mellitus in index patients did not increase the risk of TB infection among HHCs and was instead associated with a substantially lower risk of incident TB disease. These findings challenge the prevailing assumption that DM uniformly amplifies transmission due to its association with smear positivity and cavitary disease, suggesting that its influence on TB dynamics may be more complex than previously understood.
Keywords: tuberculosis, diabetes mellitus, transmission dynamics
Summary:
We evaluated whether diabetes in TB index patients influences transmission to household contacts. Despite higher smear positivity, diabetic patients did not increase TB infection risk and were associated with significantly lower TB disease incidence among exposed contacts, suggesting complex host-pathogen interactions.
Introduction
Multiple studies have demonstrated that diabetes mellitus (DM) affects the clinical progression of tuberculosis (TB) in individuals suffering from both conditions. People with diabetes mellitus are more likely to develop TB disease after infection[1–3], to present with more severe illness[4] and to experience poorer treatment outcomes[4,5]. Recent research has focused on optimizing the clinical management of TB-DM patients[6].
In addition, TB-DM patients are more likely to be sputum smear-positive[7–9], have cavitary lesions on chest X-rays[10], and experience prolonged cough compared to those without diabetes[11]. Since these factors are closely linked to TB transmission[12], it is commonly assumed that TB-DM patients are more likely to transmit TB to their contacts than those with TB alone. Recent mathematical modeling studies have proposed that TB-DM patients are 50% more likely to transmit TB than those without DM due to increased disease susceptibility and infectiousness, predicting that DM could account for one-third of TB cases in Indonesia by 2050[13].
Despite these associations, there is little empirical evidence that TB-DM patients are more likely to infect their direct contacts or that these infections consistently lead to TB disease. Mtb exhibits considerable phenotypic heterogeneity that is influenced by the host metabolic environment[14]. For example, Mtb growth at acidic pH relies on host-derived carbon sources like pyruvate, acetate, and cholesterol, while growth is inhibited by other sources such as glucose and glycerol[15]. Although it is not clear if or how DM might alter the carbon sources available to Mtb within relevant host cells, the profound metabolic changes that accompany DM are likely to lead to phenotypic variation in Mtb that may be clinically significant.
Given these findings, it is important to directly measure the impact of pre-existing DM on TB infectiousness, rather than relying solely on bacillary load or cavitary lesions as proxies for TB transmissibility. To date, the only study directly examining this association yielded inconclusive results[16]. In this study, we make use of a longitudinal study of TB patients and their household contacts (HHCs) to assess the impact of DM on TB transmissibility and the likelihood of disease development in exposed contacts.
Methods
Study setting and population
Between September 2009 and August 2012, we conducted a prospective cohort study across 20 districts in Lima, Peru, as described previously[17]. In brief, rather than using convenience sampling, we systematically invited all newly diagnosed TB patients aged ≥16 years presenting to participating health clinics to enroll as index patients. We asked for permission to invite their HHCs to participate in the study, and study staff visited the households of consenting index patients within two weeks of diagnosis.
TB diagnoses of index patients were confirmed by sputum smear microscopy and mycobacterial culture. At enrollment, we collected clinical and socio-demographic information from index patients and HHCs, including gender, height, weight, socioeconomic status (SES)[18], alcohol use, smoking, TB symptoms, Bacillus Calmette-Guérin (BCG) vaccination history, and comorbidities such as HIV and DM. HIV testing was offered to participants with unknown status. All TB patients underwent chest X-rays.
HHCs with TB symptoms were clinically evaluated. Baseline TB infection was assessed using a tuberculin skin test (TST) in all HHCs, except those with co-prevalent TB, prior TB disease, or a previous positive TST. HHCs were then re-evaluated for TB infection and disease at six and 12 months of follow-up.
Index patients were classified as having DM if they reported a prior diagnosis or were on hypoglycemic medication. We also measured serum fructosamine levels at TB diagnosis for 1,523 randomly selected index patients with positive smear cultures. Fructosamine, like HbA1c, is a glycated protein, but reflects glucose control over 2–3 weeks rather than several months.
Risk of TB infection among HHCs
We assessed TB infection risk in HHCs using two outcomes: baseline TB infection and incident infection at the 6-month follow-up among those uninfected at enrollment. Because adults in Lima were likely previously exposed to Mtb—and thus could have a positive TST unrelated to recent infection—we restricted the baseline analysis to children aged 15 and under.
Risk of active TB disease among HHCs
We assessed the risk of active TB disease among HHCs using two outcomes. The primary outcome was the cumulative risk of newly-developed TB, defined as diagnosis ≥14 days after the index patient’s diagnosis during 12-month follow-up. We hypothesized that if DM increases infectiousness, HHCs might develop disease earlier. To test this, we created a composite outcome of co-prevalent (diagnosed 30 days before to 14 days after the index case) and incident TB together as a secondary measure of TB disease risk.
Data analyses
We used modified Poisson generalized estimating equations to assess associations between risk factors and HHC TB outcomes. Univariable analyses were conducted for all index and HHC predictors, followed by backward stepwise selection (p<0.2) to build multivariable models. Retained variables were reported with effect estimates and 95% CIs. Kaplan–Meier curves were generated to compare TB incidence over time among HHCs exposed to index patients with versus without DM.
Sensitivity analyses
We conducted several sensitivity analyses to address potential misclassification and confounding. First, to reduce misclassification of undiagnosed diabetes, we excluded index patients without known DM but with fructosamine levels above the 85th percentile. This threshold was informed by a prior study in Peru, which reported that 13% of TB patients had an HbA1c ≥6.5% at diagnosis[19]. Among the 82 index patients with known DM and available fructosamine measurements, 73 (89%) had levels above this threshold, supporting its validity for detecting possible DM. Second, we repeated the analyses but defined index patients not reporting DM but with a serum fructosamine level >340 μmol/L as having undiagnosed uncontrolled DM[20]. Third, we considered that metformin use may modify the effect of DM on TB transmission and further stratified diabetic index patients by metformin use. Fourth, as we previously reported that DM status was associated with Mtb lineage, we repeated analyses further adjusting for Mtb strain lineage. Fifth, to address potential age-related confounding, we restricted analyses to index patients aged ≥40 years. For analyses using incident TB disease as the outcome, we further adjusted for use of isoniazid preventive therapy.
Transmission dynamics among HHCs with DM
We hypothesized that if certain Mtb strains were adapted to the metabolic environment of DM, HHCs who also have DM might face higher TB risk than HHCs without DM following exposure to a DM index patient. To explore this, we assessed TB outcomes among HHCs exposed to diabetic index patients, stratified by HHC DM status. Given the low prevalence of DM among HHCs, these exploratory analyses were limited to frequency tables and Fisher’s exact tests.
Results
We enrolled 12,767 HHCs of 3,109 individuals with microbiologically confirmed TB. Among the 3,083 (99.2%) index patients from whom self-reported DM status or the use of hypoglycemic drugs was obtained at baseline, 173 (5.6%) had DM (Table S1). Compared to index TB patients without known DM, those classified as having DM were more likely to be positive by sputum smear microscopy (80.2% versus 72.5%, p=0.03).
Risk of TB infection among HHCs
Baseline TB infection status of child contacts
Among 4,259 child HHCs whose baseline TB infection status was known, 1,124 (26.4 %) were infected at baseline (Tables 1 and S2). We found no increased risk of TB infection at baseline in the child HHCs of DM index patients (adjusted prevalence risk ratio [aPRR] = 1.05, 95% CI = 0.78–1.42) (Table 2). We observed minimal changes in effect size when we excluded index patients with undiagnosed DM (aPRR = 1.10, 95% CI=0.77–1.57), classified all index patients with a serum fructosamine level ≥ 340 μmol/L as having DM (aPRR = 1.03 (0.78–1.37), adjusted for the Mtb lineage of index patients (aPRR = 1.14 (0.79–1.63), restricted the analysis to index patients aged 40 or older (aPRR = 1.03, 95% CI = 0.72–1.47), or stratified DM index patients by metformin use (not on metformin: aPRR = 0.79, 95% CI: 0.43–1.45; on metformin: aPRR = 1.16, 95% CI = 0.81–1.67) (Table S3).
Table 1.
Child contacts stratified by the DM status of their index patients.
| Characteristic | Total (N = 4,259) | Exposed to index patients without DM (N = 4,060) | Exposed to index patients with DM (N = 199) |
|---|---|---|---|
| Baseline TB infection | |||
| No | 3,135 (73.6) | 2,987 (73.6) | 148 (74.4) |
| Yes | 1,124 (26.4) | 1,073 (26.4) | 51 (25.6) |
| Index age, years | |||
| 15–30 | 2,455 (57.6) | 2,449 (60.3) | 6 (3.0) |
| 31–45 | 1,056 (24.8) | 1,011 (24.9) | 45 (22.6) |
| 46–60 | 404 (9.5) | 322 (7.9) | 82 (41.2) |
| >60 | 344 (8.1) | 278 (6.8) | 66 (33.2) |
| Index sex | |||
| Female | 1,958 (46) | 1,877 (46.2) | 81 (40.7) |
| Male | 2,301 (54) | 2,183 (53.8) | 118 (59.3) |
| Index smoking status | |||
| Nonsmoker | 4,072 (97.2) | 3,875 (97) | 197 (99.5) |
| Light smoker | 59 (1.4) | 59 (1.5) | 0 (0.0) |
| Heavy smoker | 60 (1.4) | 59 (1.5) | 1 (0.5) |
| Index drinking status | |||
| Non-drinker | 2,365 (57.8) | 2,251 (57.7) | 114 (59.1) |
| Drinker | 1,728 (42.2) | 1,649 (42.3) | 79 (40.9) |
| Index smear status | |||
| Negative | 1,141 (26.9) | 1,106 (27.3) | 35 (17.7) |
| + | 1,173 (27.6) | 1,129 (27.9) | 44 (22.2) |
| ++ | 779 (18.4) | 729 (18.0) | 50 (25.3) |
| +++ | 1,152 (27.1) | 1,083 (26.8) | 69 (34.8) |
| Index coughing duration, days | |||
| 0–13 | 821 (19.6) | 787 (19.7) | 34 (17.4) |
| 14–28 | 1,401 (33.4) | 1,342 (33.6) | 59 (30.3) |
| 29–55 | 1,041 (24.8) | 983 (24.6) | 58 (29.7) |
| ≥56 | 931 (22.2) | 887 (22.2) | 44 (22.6) |
| Index cavity | |||
| No | 3,067 (73.1) | 2,924 (73.0) | 143 (73.7) |
| Yes | 1,130 (26.9) | 1,079 (27.0) | 51 (26.3) |
| Child contact age, years | |||
| 0–5 | 1,858 (43.6) | 1,786 (44.0) | 72 (36.2) |
| 6–10 | 1,214 (28.5) | 1,151 (28.3) | 63 (31.7) |
| 11–15 | 1,187 (27.9) | 1,123 (27.7) | 64 (32.2) |
| Child contact sex | |||
| Female | 2,117 (49.7) | 2,008 (49.5) | 109 (54.8) |
| Male | 2,142 (50.3) | 2,052 (50.5) | 90 (45.2) |
| Child contact BCG vaccination | |||
| No | 817 (19.2) | 778 (19.2) | 39 (19.6) |
| Yes | 3,442 (80.8) | 3,282 (80.8) | 160 (80.4) |
| Child contact nutritional status | |||
| Normal | 3,386 (80.3) | 3,254 (81) | 132 (66.7) |
| Underweight | 114 (2.7) | 111 (2.8) | 3 (1.5) |
| Overweight | 716 (17) | 653 (16.3) | 63 (31.8) |
| Household SES | |||
| High | 1,570 (37.9) | 1,514 (38.3) | 56 (28.9) |
| Medium | 1,875 (45.2) | 1,774 (44.9) | 101 (52.1) |
| Low | 702 (16.9) | 665 (16.8) | 37 (19.1) |
Data are represented as N (%)
Abbreviations: TB, tuberculosis; DM, diabetes mellitus; BCG, Bacillus Calmette-Guérin; SES, socioeconomic status
Table 2.
Univariable and multivariable models for risk of TB infection at enrollment among child contacts.
| Univariable (N=4,259) | Multivariable (N=3,914) | |||
|---|---|---|---|---|
| Variable | Prevalence Risk Ratio (95% CI) | P-value | Prevalence Risk Ratio (95% CI) | P-value |
| Index DM status | ||||
| No | Ref | Ref | ||
| Yes | 1.01 (0.77–1.32) | 0.957 | 1.05 (0.78–1.42) | 0.756 |
| Index age, years | ||||
| 15–30 | Ref | Ref | ||
| 31–45 | 1.29 (1.14–1.46) | <0.001 | 1.19 (1.05–1.36) | 0.007 |
| 46–60 | 1.10 (0.91–1.32) | 0.341 | 1.04 (0.85–1.28) | 0.675 |
| >60 | 0.60 (0.45–0.80) | <0.001 | 0.59 (0.44–0.79) | <0.001 |
| Index sex | ||||
| Female | Ref | Ref | ||
| Male | 0.82 (0.74–0.92) | 0.001 | 0.81 (0.72–0.91) | <0.001 |
| Index smoking status | ||||
| Nonsmoker | Ref | Ref | ||
| Light smoker | 1.06 (0.61–1.82) | 0.841 | 1.05 (0.61–1.80) | 0.856 |
| Heavy smoker | 1.46 (1.01–2.11) | 0.044 | 1.48 (1.06–2.09) | 0.023 |
| Index drinking status | ||||
| Non-drinker | Ref | – | – | |
| Drinker | 1.01 (0.91–1.14) | 0.799 | – | – |
| Index smear status | ||||
| Negative | Ref | Ref | ||
| + | 1.34 (1.14–1.57) | <0.001 | 1.19 (1.01–1.40) | 0.033 |
| ++ | 1.38 (1.16–1.64) | <0.001 | 1.32 (1.11–1.56) | 0.002 |
| +++ | 1.25 (1.07–1.47) | 0.006 | 1.14 (0.96–1.34) | 0.130 |
| Index coughing duration, days | ||||
| 0–13 | Ref | Ref | ||
| 14–28 | 1.4 (1.17–1.68) | <0.001 | 1.26 (1.05–1.51) | 0.012 |
| 29–55 | 1.51 (1.26–1.82) | <0.001 | 1.34 (1.11–1.61) | 0.002 |
| ≥56 | 1.63 (1.35–1.96) | <0.001 | 1.37 (1.13–1.67) | 0.001 |
| Index cavity | ||||
| No | Ref | Ref | ||
| Yes | 1.38 (1.23–1.55) | <0.001 | 1.33 (1.18–1.50) | <0.001 |
| Child contact age, years | ||||
| 0–5 | Ref | Ref | ||
| 6–10 | 1.44 (1.27–1.62) | <0.001 | 1.42 (1.26–1.61) | <0.001 |
| 11–15 | 1.76 (1.56–1.97) | <0.001 | 1.76 (1.56–1.99) | <0.001 |
| Child contact sex | ||||
| Female | Ref | – | – | |
| Male | 0.97 (0.88–1.07) | 0.544 | – | – |
| Child contact BCG vaccination | ||||
| No | Ref | – | – | |
| Yes | 0.94 (0.84–1.06) | 0.312 | – | – |
| Child contact nutritional status | ||||
| Normal | Ref | Ref | ||
| Underweight | 0.73 (0.51–1.06) | 0.099 | 0.81 (0.57–1.15) | 0.240 |
| Overweight | 1.01 (0.89–1.14) | 0.863 | 0.99 (0.87–1.13) | 0.926 |
| Household SES | ||||
| High | Ref | Ref | ||
| Medium | 0.80 (0.70–0.90) | <0.001 | 0.83 (0.73–0.94) | 0.003 |
| Low | 0.83 (0.70–0.97) | 0.021 | 0.80 (0.68–0.94) | 0.007 |
Abbreviations: TB, tuberculosis; DM, diabetes mellitus; BCG, Bacillus Calmette-Guérin; SES, socioeconomic status
Incident TB infection at 6 months among all HHCs
Among 4,812 HHCs uninfected at enrollment, 235 (4.9%) were exposed to DM index patients (Table 3 and S4). We found that HHCs exposed to DM index patients had a lower, though not statistically significant, risk of incident TB infection (adjusted cumulative rate ratio [aCRR]=0.85, 95% CI: 0.66–1.09) (Table 4). This estimate showed modest variation across all sensitivity analyses, including after excluding non-DM index patients with elevated serum fructosamine (aCRR = 0.85, 95% CI = 0.64–1.12), considering index patients without DM but with a serum fructosamine level ≥ 340 μmol/L as having DM (aCRR = 0.97, 95% CI = 0.78–1.21), and adjusting for Mtb lineage of index patients (aCRR = 0.83, 95% CI = 0.63–1.10), restricting to index patients aged ≥40 years (aCRR = 0.92, 95% CI = 0.70–1.20), and stratifying DM index patients by metformin use (not on metformin: aCRR = 0.93, 95% CI = 0.62–1.39; on metformin: aCRR = 0.82, 95% CI = 0.60–1.11), with none of the comparisons reaching statistical significance (Table S3).
Table 3.
Household contacts stratified by DM status of their index patients among those who were not TB infected at enrollment.
| Total (N = 4,812) | Exposed to index patients without DM (N = 4,577) | Exposed to index patients with DM (N = 235) | |
|---|---|---|---|
| Incident Infection at 6 months | |||
| No | 3,516 (73.1) | 3,338 (72.9) | 178 (75.7) |
| Yes | 1,296 (26.9) | 1,239 (27.1) | 57 (24.3) |
| Index age, years | |||
| 15–30 | 2,908 (60.4) | 2,905 (63.5) | 3 (1.3) |
| 31–45 | 869 (18.1) | 825 (18.0) | 44 (18.7) |
| 46–60 | 466 (9.7) | 376 (8.2) | 90 (38.3) |
| >60 | 569 (11.8) | 471 (10.3) | 98 (41.7) |
| Index sex | |||
| Female | 2,009 (41.7) | 1,905 (41.6) | 104 (44.3) |
| Male | 2,803 (58.3) | 2,672 (58.4) | 131 (55.7) |
| Index smoking status | |||
| Non-smoker | 4,622 (97.9) | 4,388 (97.8) | 234 (100) |
| Light smoker | 46 (1.0) | 46 (1.0) | 0 (0.0) |
| Heavy smoker | 51 (1.1) | 51 (1.1) | 0 (0.0) |
| Index drinking status | |||
| Non-drinker | 2,764 (59.2) | 2,617 (59.0) | 147 (63.4) |
| Drinker | 1,904 (40.8) | 1,819 (41.0) | 85 (36.6) |
| Index smear status | |||
| Negative | 1,272 (26.5) | 1,236 (27.1) | 36 (15.3) |
| + | 1,333 (27.8) | 1,280 (28.1) | 53 (22.6) |
| ++ | 843 (17.6) | 768 (16.8) | 75 (31.9) |
| +++ | 1,345 (28.1) | 1,274 (28.0) | 71 (30.2) |
| Index coughing duration, days | |||
| 0–13 | 1,071 (22.6) | 1,031 (22.9) | 40 (17.3) |
| 14–28 | 1,496 (31.6) | 1,423 (31.6) | 73 (31.6) |
| 29–55 | 1,173 (24.8) | 1,097 (24.4) | 76 (32.9) |
| ≥56 | 995 (21.0) | 953 (21.2) | 42 (18.2) |
| Index cavity | |||
| No | 3,515 (73.8) | 3,350 (73.9) | 165 (71.4) |
| Yes | 1,249 (26.2) | 1,183 (26.1) | 66 (28.6) |
| Household contact age, years | |||
| 0–15 | 2,266 (47.1) | 2,155 (47.1) | 111 (47.2) |
| 16–30 | 1,191 (24.8) | 1,142 (25) | 49 (20.9) |
| 31–45 | 667 (13.9) | 637 (13.9) | 30 (12.8) |
| 46–60 | 407 (8.5) | 383 (8.4) | 24 (10.2) |
| 51–60 | 281 (5.8) | 260 (5.7) | 21 (8.9) |
| Household contact sex | |||
| Female | 2,661 (55.3) | 2,526 (55.2) | 135 (57.4) |
| Male | 2,151 (44.7) | 2,051 (44.8) | 100 (42.6) |
| Household contact BCG vaccination | |||
| No | 738 (15.3) | 706 (15.4) | 32 (13.6) |
| Yes | 4,074 (84.7) | 3,871 (84.6) | 203 (86.4) |
| Household contact nutritional status | |||
| Normal | 3,039 (63.8) | 2,921 (64.4) | 118 (50.6) |
| Underweight | 88 (1.8) | 84 (1.9) | 4 (1.7) |
| Overweight | 1,639 (34.4) | 1,528 (33.7) | 111 (47.6) |
| Household SES | |||
| High | 1,629 (34.5) | 1,559 (34.8) | 70 (30.3) |
| Medium | 2,159 (45.8) | 2,046 (45.6) | 113 (48.9) |
| Low | 929 (19.7) | 881 (19.6) | 48 (20.8) |
Data are represented as N (%)
Abbreviations: TB, tuberculosis; DM, diabetes mellitus; BCG, Bacillus Calmette-Guérin; SES, socioeconomic status.
Table 4.
Univariable and multivariable models for incident TB infection at the 6-month follow-up among household contacts who were not infected at enrollment.
| Univariable (N = 4,812) | Multivariable (N = 4,608) | |||
|---|---|---|---|---|
| Variable | Cumulative Rate Ratio (95% CI) | P-value | Cumulative Rate Ratio (95% CI) | P-value |
| Index DM status | ||||
| No | Ref | Ref | ||
| Yes | 0.86 (0.67–1.10) | 0.239 | 0.85 (0.66–1.09) | 0.203 |
| Index age, years | ||||
| 15–30 | Ref | Ref | ||
| 31–45 | 0.88 (0.76–1.03) | 0.107 | 0.94 (0.81–1.10) | 0.465 |
| 46–60 | 1.08 (0.92–1.28) | 0.350 | 1.07 (0.90–1.27) | 0.448 |
| >60 | 0.90 (0.74–1.08) | 0.247 | 0.85 (0.70–1.03) | 0.105 |
| Index sex | ||||
| Female | Ref | Ref | ||
| Male | 0.98 (0.88–1.09) | 0.672 | 0.91 (0.82–1.01) | 0.090 |
| Index smoking status | ||||
| Nonsmoker | Ref | |||
| Light smoker | 0.86 (0.49–1.50) | 0.593 | – | – |
| Heavy smoker | 0.83 (0.47–1.44) | 0.500 | – | – |
| Index drinking status | ||||
| Non-drinker | Ref | – | – | |
| Drinker | 1.06 (0.95–1.18) | 0.293 | – | – |
| Index smear status | ||||
| Negative | Ref | Ref | ||
| + | 1.21 (1.05–1.41) | 0.011 | 1.20 (1.03–1.40) | 0.016 |
| ++ | 1.31 (1.11–1.54) | 0.001 | 1.36 (1.15–1.60) | <0.001 |
| +++ | 1.21 (1.04–1.41) | 0.012 | 1.24 (1.06–1.44) | 0.006 |
| Index coughing duration, days | ||||
| 0–13 | Ref | – | – | |
| 14–28 | 1.07 (0.92–1.24) | 0.373 | – | – |
| 29–55 | 1.14 (0.97–1.33) | 0.103 | – | – |
| ≥56 | 1.07 (0.91–1.27) | 0.407 | – | – |
| Index cavity | ||||
| No | Ref | Ref | ||
| Yes | 1.29 (1.16–1.45) | <0.001 | 1.20 (1.07–1.34) | 0.002 |
| Household contact age, years | ||||
| 0–15 | Ref | Ref | ||
| 16–30 | 2.13 (1.88–2.41) | <0.001 | 2.08 (1.83–2.36) | <0.001 |
| 31–45 | 2.67 (2.35–3.04) | <0.001 | 2.62 (2.28–3.01) | <0.001 |
| 46–60 | 2.91 (2.52–3.36) | <0.001 | 2.75 (2.35–3.23) | <0.001 |
| 51–60 | 2.54 (2.11–3.05) | <0.001 | 2.47 (2.01–3.02) | <0.001 |
| Household contact sex | ||||
| Female | Ref | – | – | |
| Male | 0.97 (0.88–1.06) | 0.481 | – | – |
| Household contact BCG vaccination | ||||
| No | Ref | – | – | |
| Yes | 1.64 (1.4–1.92) | <0.001 | – | – |
| Household contact nutritional status | ||||
| Normal | Ref | Ref | ||
| Underweight | 0.67 (0.41–1.08) | 0.101 | 0.81 (0.51–1.31) | 0.395 |
| Overweight | 1.50 (1.37–1.65) | <0.001 | 1.07 (0.97–1.19) | 0.166 |
| Household SES | ||||
| High | Ref | Ref | ||
| Medium | 1.03 (0.91–1.17) | 0.592 | 1.01 (0.89–1.13) | 0.929 |
| Low | 1.15 (1.00–1.33) | 0.054 | 1.00 (0.86–1.15) | 0.972 |
Abbreviations: TB, tuberculosis; DM, diabetes mellitus; BCG, Bacillus Calmette-Guérin; SES, socioeconomic status
Risk of TB disease among HHCs
Primary outcome—incident TB disease among all HHCs during 12-month follow-up
Among 12,442 patient who were free from TB disease at enrollment, 368 (3.0%) developed incident TB during the 12 months of follow-up (Tables 5 and S5). The Kaplan–Meier curves comparing the incidence of TB disease among HHCs exposed to index patients with and without DM are shown in Figure S1. Those exposed to index TB patients with DM were two-thirds less likely to develop incident TB compared to those exposed to TB patients without DM (aCRR=0.33, 95% CI=0.13–0.85) (Table 6). This strong effect remained consistent across sensitivity analyses, including after excluding non-DM index patients with elevated fructosamine (aCRR = 0.32, 95% CI: 0.11–0.90), classifying all index patients with a serum fructosamine level ≥ 340 μmol/L as having DM (aCRR = 0.38, 95% CI: 0.18–0.81), adjusting for index TB patients’ Mtb lineage (aCRR = 0.38, 95% CI: 0.13–1.08) or IPT use among HHCs (aCRR = 0.31, 95% CI: 0.12–0.8), and restricting to index patients aged ≥40 years (aCRR = 0.24, 95% CI: 0.07–0.77). Risk was similarly reduced among HHCs exposed to DM index patients not taking metformin (aCRR = 0.43, 95% CI: 0.10–1.76) and those on metformin (aCRR = 0.29, 95% CI: 0.09–0.98) (Table S6).
Table 5.
Characteristics of all household contacts stratified by the DM status of their index patients.
| Variable | Total (N = 12,422) | Exposed to index patients without DM (N=11,756) | Exposed to index patients with DM (N=666) |
|---|---|---|---|
| Developed Incident TB | |||
| No | 12,054 (97.0) | 11,394(96.7) | 660 (99.1) |
| Yes | 368 (3.0) | 362 (3.3) | 6 (0.9) |
| Index age, years | |||
| 15–30 | 7,339 (59.1) | 7,320 (62.3) | 19 (2.9) |
| 31–45 | 2,458 (19.8) | 2,325 (19.8) | 133 (20.0) |
| 46–60 | 1,345 (10.8) | 1,077 (9.2) | 268 (40.2) |
| >60 | 1,280 (10.3) | 1,034 (8.8) | 246 (36.9) |
| Index sex | |||
| Female | 5,083 (40.9) | 4,804 (40.9) | 279 (41.9) |
| Male | 7,339 (59.1) | 6,952 (59.1) | 387 (58.1) |
| Index smoking status | |||
| Non-smoker | 11,867 (97.3) | 11,213 (97.2) | 654 (99.1) |
| Light smoker | 141 (1.2) | 141 (1.2) | 0 (0.0) |
| Heavy smoker | 192 (1.6) | 186 (1.6) | 6 (0.9) |
| Index drinking status | |||
| Non-drinker | 6,920 (57.9) | 6,531 (57.8) | 389 (60.0) |
| Drinker | 5,022 (42.1) | 4,763 (42.2) | 259 (40.0) |
| Index smear status | |||
| Negative | 3,272 (26.5) | 3,151 (26.9) | 121 (18.2) |
| + | 3,379 (27.3) | 3,238 (27.7) | 141 (21.2) |
| ++ | 2,254 (18.2) | 2,090 (17.9) | 164 (24.6) |
| +++ | 3,459 (28.0) | 3,219 (27.5) | 240 (36.0) |
| Index coughing duration, days | |||
| 0–13 | 2,621 (21.5) | 2,497 (21.6) | 124 (19.0) |
| 14–28 | 3,996 (32.8) | 3,762 (32.6) | 234 (35.9) |
| 29–55 | 2,943 (24.1) | 2,774 (24.0) | 169 (26.0) |
| ≥56 | 2,639 (21.6) | 2,515 (21.8) | 124 (19.0) |
| Index cavity | |||
| No | 8,951 (73.1) | 8,487 (73.2) | 464 (71.8) |
| Yes | 3,294 (26.9) | 3,112 (26.8) | 182 (28.2) |
| Household contact age, years | |||
| 0–15 | 4,370 (35.2) | 4,162 (35.4) | 208 (31.2) |
| 16–30 | 3,348 (27.0) | 3,168 (26.9) | 180 (27.0) |
| 31–45 | 2,213 (17.8) | 2,080 (17.7) | 133 (20.0) |
| 46–60 | 1,629 (13.1) | 1,543 (13.1) | 86 (12.9) |
| 51–60 | 862 (6.9) | 803 (6.8) | 59 (8.9) |
| Household contact sex | |||
| Female | 6,878 (55.4) | 6,487 (55.2) | 391 (58.7) |
| Male | 5,544 (44.6) | 5,269 (44.8) | 275 (41.3) |
| Household contact BCG vaccination | |||
| No | 1,741 (14.0) | 1,650 (14.0) | 91 (13.7) |
| Yes | 10,679 (86.0) | 10,104 (86.0) | 575 (86.3) |
| Household contact nutritional status | |||
| Normal | 7,086 (57.6) | 6,770 (58.1) | 316 (47.6) |
| Underweight | 214 (1.7) | 207 (1.8) | 7 (1.1) |
| Overweight | 5,007 (40.7) | 4,666 (40.1) | 341 (51.4) |
| Household SES | |||
| High | 4,172 (34.4) | 3,982 (34.7) | 190 (29.1) |
| Medium | 5,409 (44.6) | 5,103 (44.5) | 306 (46.8) |
| Low | 2,534 (20.9) | 2,376 (20.7) | 158 (24.2) |
Data are represented as N (%)
Abbreviations: DM, diabetes mellitus; BCG, Bacillus Calmette-Guérin; SES, socioeconomic status
Table 6.
Univariate and multivariable analyses for incident TB disease at 12-month follow-up among household contacts who were not diagnosed with TB at enrollment.
| Univariable (N=12,422) | Multivariable (N=11,566) | |||
|---|---|---|---|---|
| Variable | Cumulative Rate Ratio (95% CI) | P-value | Cumulative Rate Ratio (95% CI) | P-value |
| Index DM status | ||||
| No | Ref | Ref | ||
| Yes | 0.31 (0.14–0.70) | 0.005 | 0.33 (0.13–0.85) | 0.021 |
| Index age, years | ||||
| 15–30 | Ref | Ref | ||
| 31–45 | 1.23 (0.92–1.65) | 0.154 | 1.23 (0.91–1.66) | 0.186 |
| 46–60 | 0.90 (0.61–1.35) | 0.622 | 1.05 (0.69–1.60) | 0.826 |
| >60 | 0.65 (0.42–1.01) | 0.055 | 0.65 (0.39–1.07) | 0.089 |
| Index sex | ||||
| Female | Ref | Ref | ||
| Male | 0.75 (0.59–0.94) | 0.014 | 0.77 (0.61–0.98) | 0.033 |
| Index smoking status | ||||
| Nonsmoker | Ref | |||
| Light smoker | 1.22 (0.55–2.70) | 0.620 | – | – |
| Heavy smoker | 1.42 (0.71–2.85) | 0.324 | – | – |
| Index drinking status | ||||
| Non-drinker | Ref | – | – | |
| Drinker | 0.89 (0.70–1.13) | 0.325 | – | – |
| Index smear status | ||||
| Negative | Ref | Ref | ||
| + | 1.48 (1.06–2.07) | 0.022 | 1.47 (1.03–2.09) | 0.032 |
| ++ | 1.58 (1.12–2.24) | 0.010 | 1.62 (1.13–2.34) | 0.009 |
| +++ | 1.37 (0.98–1.92) | 0.064 | 1.43 (1.00–2.04) | 0.051 |
| Index coughing duration, days | ||||
| 0–13 | Ref | Ref | ||
| 14–28 | 1.53 (1.06–2.21) | 0.024 | 1.33 (0.92–1.93) | 0.133 |
| 29–55 | 1.85 (1.28–2.68) | 0.001 | 1.57 (1.07–2.29) | 0.020 |
| ≥56 | 1.59 (1.05–2.42) | 0.029 | 1.42 (0.94–2.16) | 0.099 |
| Index cavity | ||||
| No | Ref | Ref | ||
| Yes | 1.47 (1.15–1.89) | 0.002 | 1.39 (1.08–1.79) | 0.009 |
| Household contact age, years | ||||
| 0–15 | Ref | Ref | ||
| 16–30 | 1.15 (0.89–1.49) | 0.281 | 1.37 (1.04–1.79) | 0.024 |
| 31–45 | 0.57 (0.40–0.82) | 0.002 | 1.03 (0.69–1.55) | 0.871 |
| 46–60 | 0.69 (0.47–1.00) | 0.052 | 1.25 (0.80–1.95) | 0.326 |
| 51–60 | 0.74 (0.47–1.18) | 0.206 | 1.19 (0.72–1.96) | 0.510 |
| Household contact sex | ||||
| Female | Ref | Ref | ||
| Male | 1.18 (0.98–1.43) | 0.087 | 1.12 (0.91–1.37) | 0.287 |
| Household contact BCG vaccination | ||||
| No | Ref | Ref | ||
| Yes | 0.63 (0.49–0.80) | <0.001 | 0.65 (0.50–0.84) | 0.001 |
| Household contact nutritional status | ||||
| Normal | Ref | Ref | ||
| Underweight | 1.43 (0.84–2.44) | 0.190 | 1.52 (0.87–2.64) | 0.142 |
| Overweight | 0.37 (0.29–0.48) | <0.001 | 0.39 (0.29–0.53) | <0.001 |
| Household SES | ||||
| High | Ref | Ref | ||
| Medium | 0.80 (0.61–1.04) | 0.089 | 0.85 (0.65–1.11) | 0.226 |
| Low | 0.73 (0.53–1.02) | 0.063 | 0.79 (0.56–1.10) | 0.162 |
Abbreviations: TB, tuberculosis; DM, diabetes mellitus; BCG, Bacillus Calmette-Guérin; SES, socioeconomic status
Secondary outcome— composite of co-prevalent and incident TB
Of the 210 HHCs with TB at enrollment, 195 (93%) were classified as co-prevalent. Combined with 374 incident cases, 569 HHCs met the composite outcome. HHCs exposed to index patients with DM were less likely to develop this outcome (aCRR = 0.47; 95% CI: 0.23–0.95).
Transmission dynamics among participants with DM
Among 306 HHCs exposed to index patients with DM, incident TB infection at 6 months was higher in contacts with DM (6/90; 6.7%) than without (4/179; 2.2%) (p = 0.09). Similarly, incident TB disease occurred in 3.3% of HHCs with DM vs. 0.8% without DM (p = 0.25), though neither difference was statistically significant.
Discussion
We evaluated the impact of DM in index TB patients on the risk of TB infection and disease among their HHCs. While index patients with DM were more likely to be smear-positive than those without DM, the risk of TB infection in their contacts did not differ by the DM status of the index case. In contrast, HHCs exposed to TB-DM index patients were substantially less likely to develop TB disease during follow-up. This unexpected finding raises the possibility that the metabolic state associated with DM in the index patient may influence disease progression in exposed HHCs, even after Mtb transmission has occurred. Notably, among HHCs exposed to a person with DM, transmission may be more easily facilitated when the HHCs also have DM.
Diagnosing DM in TB patients is challenging, particularly in resource-limited settings. Relying on self-report may miss undiagnosed DM, while only using glycemic markers at TB diagnosis may overestimate DM due to the stress hyperglycemia commonly caused by TB. Two prior studies have shown that among TB patients without known DM but with an HbA1c ≥6.5% at diagnosis, HbA1c levels fell below this threshold in 70–90% following TB treatment[21,22], suggesting that a single elevated HbA1c may reflect TB severity rather than true DM. To address possible DM misclassification, we conducted several sensitivity analyses that combined self-reported DM and baseline glycemic measurements. To further evaluate potential age-related confounding, we adjusted for the age of index patients and conducted a sensitivity analysis restricted to HHCs exposed to index patients aged ≥40 years. Our findings remained consistent across these approaches.
Given previous reports of metformin’s protective effects against TB infection, disease progression, and treatment outcomes through immunomodulation[23–26], we also stratified index patients with DM by metformin use to assess whether it could protect their contacts from developing TB. However, we found no evidence that metformin use among DM index patients modified TB outcomes in their HHCs.
Direct comparisons with prior studies are limited by differences in how DM and transmission were measured. Only one previous study, conducted in Brazil, assessed the impact of index DM on TB infection among HHCs. While baseline IGRA positivity was similar between those exposed to TB-DM versus non-DM index patients (40% vs 43%), conversion among initially IGRA-negative contacts was higher in the TB-DM group (17% vs 9%)[16]. This discrepancy may reflect the higher sensitivity of IGRA relative to TST[27], and their use of baseline HbA1c to define diabetes, which may capture TB-related hyperglycemia rather than chronic DM.
Our results align with another Peruvian cohort study in which Grandjean et al. followed 1,055 disease-free HHCs for two years. Their study found that contacts exposed to an index patient with DM had a fivefold reduced risk of developing TB disease compared to those exposed to an index patient without DM[28].
Other studies have assessed the relationship between index patient DM and transmission using molecular methods. In our prior work, whole-genome sequencing showed that TB patients with DM were less likely to be transmitters than those without DM. A systematic review and meta-analysis of six molecular epidemiologic studies also found that TB patients with DM were less likely to be clustered with other TB cases than patients without DM (global odds ratio = 0.84)[30]. An exception was a small Spanish study reporting higher DM prevalence among transmitters, though unstable estimates (0.12 to 23.7) suggest uncontrolled confounding[31].
These findings are unexpected given the well-established association of DM with high bacillary load and cavitary disease—features typically linked to increased transmissibility. One possible explanation is that Mtb adapts to the metabolic environment of diabetic hosts in ways that influence its behavior in subsequent hosts. During infection, Mtb responds to fluctuations in nutrient availability by inducing metabolic pathways that enable the use of multiple carbon sources, including glucose, glycerol, cholesterol, and triglycerides[32–34]. The dominant carbon source may influence its growth rate, virulence, and transcriptional profile[32,35–37]. For instance, studies have shown that using different carbon sources alters the abundance of propionyl-CoA and acetyl-CoA, which are precursors for virulence-related cell wall lipids of Mtb[38–40]. Likewise, Mtb extracted from host lung tissue shows increased synthesis of phthiocerol dimycocerosate (PDIM), a pattern consistent with growth on odd-chain fatty acids[41]. Alterations in central carbon metabolism may also confer resistance to host-derived stresses such as acidity, redox disruptions, or hypoxia—conditions Mtb routinely encounters during infection[42–45].
Our finding that transmission may be facilitated among HHCs with DM exposed to index patients with DM suggests that DM-induced metabolic changes could enhance TB transmission between such hosts. DM alters glucose, lipid, and fatty acid levels—potential carbon sources for Mtb—in complex ways[14,46], likely producing a distinct transcriptional profile in Mtb from hosts with DM. Although direct studies of how DM-adapted Mtb affects disease in subsequent hosts are lacking, other evidence indicates that the pre-transmission metabolic state of Mtb can influence outcomes; for example, zinc limitation primes Mtb to resist oxidative stress, leading to more severe granulomas in a mouse model[47]. If subsequent hosts do not have DM—which was likely in our study, given the <15% prevalence of DM in Peru[19]—Mtb adapted to the intracellular environment of a host with DM may need to re-adapt to a non-diabetic host. The timing of this process is unclear, but early immune interactions between Mtb and alveolar macrophages strongly influence disease progression [48,49].
Our study has several limitations. First, the cohort study was not specifically designed to measure the effect of DM on TB transmissibility, even though the TB infection and disease outcomes are expected to be accurately measured. Furthermore, the study was conducted before TB preventive therapy (TPT) was widely implemented; only about half of child contacts received 6 months of TPT, and adults were offered isoniazid preventive therapy only if they had serious comorbidities. While the current WHO recommendation for universal TPT among HHCs may limit the present-day generalizability of our findings, it also makes it ethically challenging to study transmission risk factors in HHC-based cohorts. Thus, despite being historical, these data offer a unique opportunity to examine the impact of DM on TB transmission under conditions that are no longer reproducible today. Secondly, in response to the unexpected findings, we explored alternative outcome definitions and conducted multiple sensitivity analyses to assess the robustness of our results. Due to the small sample sizes of some groups (e.g. TB-DM patients not on metformin, HHCs with DM), we had limited statistical power to explore all relevant questions. Finally, despite our efforts to address potential exposure misclassification and bias, we cannot fully exclude the possibility of residual confounding by age or other factors.
Conclusion
In this large cohort study, HHCs exposed to index patients with DM did not experience a higher risk of TB infection and instead had a markedly lower risk of progressing to active TB disease. These findings suggest that the influence of DM on TB transmission dynamics is more nuanced than previously assumed.
Supplementary Material
Acknowledgements
MBM and MCB designed the original parent study; CCH conceived this secondary data analysis. MBM, MCB, RC, CC, LL, JJ, RY, and JG implemented the original study and collected the data. ZZ served as the data manager. CCH and QT performed the analysis. NCH contributed to interpreting the findings from a microbiological perspective. MBM, CCH, NCH and QT wrote the original draft of this paper and all authors contributed to the final draft.
The authors thank the patients, their families, and the healthcare personnel at the 106 participating health centers in Lima, Peru. Their contribution and support were crucial to the study’s progress and outcomes. The content is solely the responsibility of the authors and does not necessarily represent the official views of the funders.
Funding
This work was supported by grants from the National Institute of Allergy and Infectious Diseases: grants U01AI057786 (MCB), U19AI076217 (MCB), U19AI109755 (JTG) (Center for Excellence in Translational Research), U19AI111224 (MBM) (Tuberculosis Research Unit), U19AI142793 (MBM) (Beat-TB), and by the William F. Milton Fund (C-CH).
Footnotes
Conflicts of Interest
The authors declare no conflicts of interest.
References
- [1].Jeon CY, Murray MB. Diabetes mellitus increases the risk of active tuberculosis: a systematic review of 13 observational studies. PLoS Med 2008;5:1091–101. 10.1371/JOURNAL.PMED.0050152. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [2].Foe-Essomba JR, Kenmoe S, Tchatchouang S, et al. Diabetes mellitus and tuberculosis, a systematic review and meta-analysis with sensitivity analysis for studies comparable for confounders. PLoS One 2021;16. 10.1371/JOURNAL.PONE.0261246. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [3].Baker MA, Lin HH, Chang HY, et al. The risk of tuberculosis disease among persons with diabetes mellitus: a prospective cohort study. Clin Infect Dis 2012;54:818–25. 10.1093/CID/CIR939. [DOI] [PubMed] [Google Scholar]
- [4].Chang JT, Dou HY, Yen CL, et al. Effect of type 2 diabetes mellitus on the clinical severity and treatment outcome in patients with pulmonary tuberculosis: a potential role in the emergence of multidrug-resistance. J Formos Med Assoc 2011;110:372–81. 10.1016/S0929-6646(11)60055-7. [DOI] [PubMed] [Google Scholar]
- [5].Baker MA, Harries AD, Jeon CY, et al. The impact of diabetes on tuberculosis treatment outcomes: a systematic review. BMC Med 2011;9. 10.1186/1741-7015-9-81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [6].Van Crevel R, Koesoemadinata R, Hill PC, et al. Clinical management of combined tuberculosis and diabetes. International Journal of Tuberculosis and Lung Disease 2018;22:1404–10. 10.5588/ijtld.18.0340. [DOI] [PubMed] [Google Scholar]
- [7].Mi F, Tan S, Liang L, et al. Diabetes mellitus and tuberculosis: pattern of tuberculosis, two-month smear conversion and treatment outcomes in Guangzhou, China. Trop Med Int Health 2013;18:1379–85. 10.1111/TMI.12198. [DOI] [PubMed] [Google Scholar]
- [8].Armstrong LR, Steve Kammerer J, Haddad MB. Diabetes mellitus among adults with tuberculosis in the USA, 2010–2017. BMJ Open Diabetes Res Care 2020;8. 10.1136/BMJDRC-2020-001275. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [9].Abdelbary BE, Garcia-Viveros M, Ramirez-Oropesa H, et al. Tuberculosis-diabetes epidemiology in the border and non-border regions of Tamaulipas, Mexico. Tuberculosis (Edinb) 2016;101S:S124–34. 10.1016/J.TUBE.2016.09.024. [DOI] [PubMed] [Google Scholar]
- [10].Zafar MI, Chen L-L, Xiaofeng Y, et al. Impact of Diabetes Mellitus on Radiological Presentation of Pulmonary Tuberculosis in Otherwise Non-Immunocompromised Patients: A Systematic Review. Curr Med Imaging Rev 2019;15:543–54. 10.2174/1573405614666180806124416. [DOI] [PubMed] [Google Scholar]
- [11].Gil-Santana L, Almeida JL, Oliveira CAM, et al. Diabetes Is Associated with Worse Clinical Presentation in Tuberculosis Patients from Brazil: A Retrospective Cohort Study. PLoS One 2016;11. 10.1371/JOURNAL.PONE.0146876. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [12].Melsew YA, Doan TN, Gambhir M, et al. Risk factors for infectiousness of patients with tuberculosis: a systematic review and meta-analysis. Epidemiol Infect 2018;146:345–53. 10.1017/S0950268817003041. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [13].Awad SF, Critchley JA, Abu-Raddad LJ. Impact of diabetes mellitus on tuberculosis epidemiology in Indonesia: A mathematical modeling analysis. Tuberculosis (Edinb) 2022;134. 10.1016/J.TUBE.2022.102164. [DOI] [PubMed] [Google Scholar]
- [14].Dhar N, McKinney J, Manina G. Phenotypic Heterogeneity in Mycobacterium tuberculosis. Microbiol Spectr 2016;4. 10.1128/MICROBIOLSPEC.TBTB2-0021-2016. [DOI] [PubMed] [Google Scholar]
- [15].Baker JJ, Johnson BK, Abramovitch RB. Slow growth of Mycobacterium tuberculosis at acidic pH is regulated by phoPR and host-associated carbon sources. Mol Microbiol 2014;94:56–69. 10.1111/MMI.12688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [16].Arriaga MB, Rocha MS, Nogueira BMF, et al. The Effect of Diabetes and Prediabetes on Mycobacterium tuberculosis Transmission to Close Contacts. J Infect Dis 2021;224:2064–72. 10.1093/INFDIS/JIAB264. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [17].Becerra MC, Huang C-C, Lecca L, et al. Transmissibility and potential for disease progression of drug resistant Mycobacterium tuberculosis: prospective cohort study. Bmj 2019;367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [18].Odone A, Calderon R, Becerra MC, et al. Acquired and Transmitted Multidrug Resistant Tuberculosis: The Role of Social Determinants. PLoS One 2016;11. 10.1371/JOURNAL.PONE.0146642. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [19].Calderon RI, Arriaga MB, Lopez K, et al. High prevalence and heterogeneity of Dysglycemia in patients with tuberculosis from Peru: a prospective cohort study. BMC Infect Dis 2019;19. 10.1186/S12879-019-4416-2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [20].Vinodchandra Patel V, Pallavbhai Desai T, Purushottam Patel H. A Comparative Analysis of HbA1c and Fructosamine for Short-Term Glycemic Monitoring in Diabetic Patients Undergoing Surgical Procedures. International Journal of Life Sciences 2025;14. 10.69605/ijlbpr_14.7.2025.29. [DOI] [Google Scholar]
- [21].Mtabho CM, Semvua HH, Van Den Boogaard J, et al. Effect of diabetes mellitus on TB drug concentrations in Tanzanian patients. J Antimicrob Chemother 2019;74:3537–45. 10.1093/JAC/DKZ368. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [22].Alkabab YMA, Biswas S, Ahmed S, et al. Differentiating transient from persistent diabetic range hyperglycemia in a cohort of people completing tuberculosis treatment in Dhaka, Bangladesh. PLoS One 2021;16. 10.1371/JOURNAL.PONE.0260389. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [23].Singhal A, Jie L, Kumar P, et al. Metformin as adjunct antituberculosis therapy. Sci Transl Med 2014;6. 10.1126/SCITRANSLMED.3009885. [DOI] [PubMed] [Google Scholar]
- [24].Salindri AD, Gujabidze M, Kipiani M, et al. Metformin reduces the risk of poor treatment outcomes among individuals with rifampicin-resistant tuberculosis and type-2 diabetes mellitus. MedRxiv 2024. 10.1101/2024.07.12.24310348. [DOI] [Google Scholar]
- [25].Lee YJ, Han SK, Park JH, et al. The effect of metformin on culture conversion in tuberculosis patients with diabetes mellitus. Korean J Intern Med 2018;33:933–40. 10.3904/KJIM.2017.249. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [26].Yu X, Li L, Xia L, et al. Impact of metformin on the risk and treatment outcomes of tuberculosis in diabetics: a systematic review. BMC Infect Dis 2019;19. 10.1186/S12879-019-4548-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [27].Pai M, Elwood K. Interferon-gamma release assays for screening of health care workers in low tuberculosis incidence settings: Dynamic patterns and interpretational challenges. Canadian Respiratory Journal : Journal of the Canadian Thoracic Society 2012;19:81. 10.1155/2012/420392. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [28].Grandjean L, Gilman RH, Martin L, et al. Transmission of Multidrug-Resistant and Drug-Susceptible Tuberculosis within Households: A Prospective Cohort Study. PLoS Med 2015;12. 10.1371/JOURNAL.PMED.1001843. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [29].Trevisi L, Brooks MB, Becerra MC, et al. Who Transmits Tuberculosis to Whom: A Cross-Sectional Analysis of a Cohort Study in Lima, Peru. Am J Respir Crit Care Med 2024;210:222–33. 10.1164/RCCM.202307-1217OC. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [30].Blanco-Guillot F, Delgado-Sánchez G, Mongua-Rodríguez N, et al. Molecular clustering of patients with diabetes and pulmonary tuberculosis: A systematic review and meta-analysis. PLoS One 2017;12. 10.1371/JOURNAL.PONE.0184675. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [31].Xu Y, Cancino-Munoz I, Torres-Puente M, et al. High-resolution mapping of tuberculosis transmission: Whole genome sequencing and phylogenetic modelling of a cohort from Valencia Region, Spain. PLoS Med 2019;16. 10.1371/JOURNAL.PMED.1002961. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [32].De Carvalho LPS, Fischer SM, Marrero J, et al. Metabolomics of Mycobacterium tuberculosis reveals compartmentalized co-catabolism of carbon substrates. Chem Biol 2010;17:1122–31. 10.1016/J.CHEMBIOL.2010.08.009. [DOI] [PubMed] [Google Scholar]
- [33].Xu Y, Pooja, Borah K. Mycobacterium tuberculosis carbon and nitrogen metabolic fluxes. Biosci Rep 2022;42:BSR20211215. 10.1042/BSR20211215. [DOI] [Google Scholar]
- [34].Griffin JE, Pandey AK, Gilmore SA, et al. Cholesterol catabolism by Mycobacterium tuberculosis requires transcriptional and metabolic adaptations. Chem Biol 2012;19:218–27. 10.1016/j.chembiol.2011.12.016. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [35].Pawełczyk J, Brzostek A, Minias A, et al. Cholesterol-dependent transcriptome remodeling reveals new insight into the contribution of cholesterol to Mycobacterium tuberculosis pathogenesis. Sci Rep 2021;11. 10.1038/S41598-021-91812-0. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [36].Chen Y, MacGilvary NJ, Tan S. Mycobacterium tuberculosis response to cholesterol is integrated with environmental pH and potassium levels via a lipid metabolism regulator. PLoS Genet 2024;20. 10.1371/JOURNAL.PGEN.1011143. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [37].Borah K, Mendum TA, Hawkins ND, et al. Metabolic fluxes for nutritional flexibility of Mycobacterium tuberculosis. Mol Syst Biol 2021;17. 10.15252/MSB.202110280. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [38].Lee W, VanderVen BC, Fahey RJ, et al. Intracellular Mycobacterium tuberculosis Exploits Host-derived Fatty Acids to Limit Metabolic Stress. Journal of Biological Chemistry 2013;288:6788–800. 10.1074/JBC.M112.445056. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [39].Koh EI, Oluoch PO, Ruecker N, et al. Chemical-genetic interaction mapping links carbon metabolism and cell wall structure to tuberculosis drug efficacy. Proc Natl Acad Sci U S A 2022;119. 10.1073/PNAS.2201632119. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [40].Wilburn KM, Fieweger RA, VanderVen BC. Cholesterol and fatty acids grease the wheels of Mycobacterium tuberculosis pathogenesis. Pathog Dis 2018;76:fty021. 10.1093/FEMSPD/FTY021. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [41].Jain M, Petzold CJ, Schelle MW, et al. Lipidomics reveals control of Mycobacterium tuberculosis virulence lipids via metabolic coupling. Proc Natl Acad Sci U S A 2007;104:5133–8. 10.1073/PNAS.0610634104. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [42].Singh A, Crossman DK, Mai D, et al. Mycobacterium tuberculosis WhiB3 maintains redox homeostasis by regulating virulence lipid anabolism to modulate macrophage response. PLoS Pathog 2009;5. 10.1371/JOURNAL.PPAT.1000545. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [43].Baker JJ, Johnson BK, Abramovitch RB. Slow growth of Mycobacterium tuberculosis at acidic pH is regulated by phoPR and host-associated carbon sources. Mol Microbiol 2014;94:56–69. 10.1111/MMI.12688. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [44].Healy C, Ehrt S, Gouzy A. An exacerbated phosphate starvation response triggers Mycobacterium tuberculosis glycerol utilization at acidic pH. MBio 2025;16. 10.1128/MBIO.02825-24. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [45].Nandy A, Mondal AK, Pandey R, et al. Adipocyte Model of Mycobacterium tuberculosis Infection Reveals Differential Availability of Iron to Bacilli in the Lipid-Rich Caseous Environment. Infect Immun 2018;86. 10.1128/IAI.00041-18. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [46].Sherry J, Rego EH. Phenotypic Heterogeneity in Pathogens. Annu Rev Genet 2024. 10.1146/ANNUREV-GENET-111523-102459. [DOI] [PubMed] [Google Scholar]
- [47].Dow A, Sule P, Odonnell TJ, et al. Zinc limitation triggers anticipatory adaptations in Mycobacterium tuberculosis. PLoS Pathog 2021;17. 10.1371/JOURNAL.PPAT.1009570. [DOI] [PMC free article] [PubMed] [Google Scholar]
- [48].Cohen RM, Holmes YR, Chenier TC, et al. Discordance between HbA1c and fructosamine: evidence for a glycosylation gap and its relation to diabetic nephropathy. Diabetes Care 2003;26:163–7. 10.2337/DIACARE.26.1.163. [DOI] [PubMed] [Google Scholar]
- [49].Cadena AM, Flynn JL, Fortune SM. The Importance of First Impressions: Early Events in Mycobacterium tuberculosis Infection Influence Outcome. MBio 2016;7. 10.1128/MBIO.00342-16. [DOI] [PMC free article] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.