To the Editors:
Diabetes mellitus increases the risk of TB disease and adverse TB outcomes [1]. Emerging evidence suggests diabetes is also associated with latent TB infection (LTBI), and population-based studies reported the prevalence of LTBI among US adults with diabetes to be more than twice that of adults without diabetes (11.6% vs 4.6%) [2, 3]. Given the rapid increase of global diabetes prevalence in regions with high TB burdens, clinical and public health interventions targeting this co-epidemic would avert substantial morbidity and mortality [4].
Metformin and statins are widely used inexpensive therapies to prevent metabolic and cardiovascular complications among patients with diabetes. Studies in euglycemic mice reported that metformin and statins reduced lung bacillary load in early and late phases of TB infection when administered either alone or in combination with anti-TB drugs [5, 6]. Retrospective data from patients with diabetes and TB from our and other studies provide evidence of metformin efficacy in human TB [5, 7–10]. These studies variously reported that use of metformin vs any other diabetic treatment was associated with lower risk of progressing to pulmonary TB disease, lower risk of cavitary TB, lower risk of death during anti-TB therapy, improved sputum conversion rates and lower risk of recurrent TB. Similarly, a population-based cohort study using Taiwanese insurance data reported nearly 50% lower incidence of TB disease in adults using statins compared to matched controls without statin use [11]. Only one small (n=220) study from Singapore examined metformin use in the context of LTBI and did not assess statin use [12].
Whether the relationship between diabetes and LTBI is modified by metformin or statins has not been thoroughly evaluated. If metformin or statin use reduces the risk of LTBI in patients with diabetes, there may be additional rationale for evaluating these therapies as TB prevention tools. This study aimed to determine if the association between diabetes and prevalence of LTBI was different by metformin or statin use.
We conducted a cross-sectional study using data from the National Health and Nutrition Examination Survey (NHANES) 2011–2012, a three-stage probability sample designed to be representative of non-institutionalized US adults [13]. Data collected from NHANES includes an in-person interview, a health examination, and laboratory measurements.
Diabetes and pre-diabetes status were defined by self-report and glycated hemoglobin. Participants who self-reported previous diabetes diagnosis by a healthcare professional were classified as having diabetes regardless of HbA1c. Participants without self-reported history of diabetes were classified by HbA1c as euglycemic (≤5.6%), prediabetes (5.7–6.4%) or diabetes (≥6.5%) following American Diabetes Association guidelines [14]. LTBI prevalence was measured by QuantiFERON-TB Gold In-tube (QFT) according to manufacturer instructions and by 0.1ml purified protein derivative tuberculin skin test (TST) which were read 46–76 hours after placement and indurations ≥10mm were defined as TST positive.
Metformin, statin, and non-metformin diabetes medication use (insulin, sulfonylureas, dipeptidyl peptidase 4 inhibitors) were defined by self-report. During NHANES interviews, all participants were asked to report use of prescription medications during a one-month period prior to the survey date. Those who answered “yes” were asked to present medication containers of all products used. For each medication presented, interviewers entered the product’s complete name into a Computer-Assisted Personal Interviewing system.
We estimated LTBI prevalence (with QFT and TST) stratified by diabetes and pre-diabetes status and by metformin, statin, and non-metformin diabetes drug use. We calculated prevalence differences (PD), odds ratios (OR), and 95% confidence intervals (95%CI) to estimate associations between diabetes and LTBI. We used two-sided Rao-Scott or Wald Chi-square p-values <0.05 to define significance. All analyses accounted for weighted probability designs of NHANES [15]. All data were publically available and de-identified and therefore determined exempt from institutional ethical review board review.
Overall weighted prevalence of LTBI among participants with diabetes was 11.6% (95%CI 7.9–15.3%) by QFT (n=4958) and 7.1% (95%CI 4.8–9.3%) by TST (n=4261), significantly higher than euglycemic participants (4.6% and 4.1%, p-value <0.05) (Table 1). Among participants with diabetes, 53.8% reported no metformin use, and LTBI prevalence was non-significantly higher in those without metformin use (by QFT [PD, 1.4% 95%CI −3.7–6.4%] and by TST [PD, 2.7%, 95%CI: −0.3–5.7]) compared to those self-reporting any metformin use. Among participants with diabetes, lower prevalence of LTBI was observed among participants with metformin plus two or more other diabetes medications (6.2% by QFT and 1.8% by TST) compared to those not using diabetes medications. After adjusting for age, sex, HbA1c, type of diabetes, income level, and duration of diabetes, the odds of TST positivity among participants with diabetes but without any diabetes medication (aOR 3.9, 95%CI 1.1–13.8) were significantly greater than participants using metformin plus two or more other diabetes medications (data not shown).
Table 1:
Diabetes and prevalence of latent tuberculosis by metformin and statin use, NHANES adult participants 2011–2012
| Diabetes status |
QuantiFERON-TB Gold In Tube | Tuberculin Skin Test | |||||
|---|---|---|---|---|---|---|---|
| QFT positivea % (95% CI) |
Prevalence differenceb PD% (95% CI) |
Odds ratio OR (95% CI) |
TST positivea % (95% CI) |
Prevalence differenceb PD% (95% CI) |
Odds ratio OR (95% CI) |
||
| ALL NHANES | Diabetes | 11.6 (7.9–15.3) | 7.0 (3.1–10.8)c | 2.7 (1.8–4.1)d | 7.1 (4.8–9.3) | 3.0 (0.6–5.4)c | 1.8 (1.2–2.8)d |
| Pre-diabetes | 7.0 (5.2–8.7) | 2.3 (1.0–3.7)c | 1.5 (1.2–1.9)d | 6.5 (2.6–10.4) | 2.4 (0.8–5.6)c | 1.6 (1.0–2.7) | |
| Euglycemic | 4.6 (3.7–5.6) | REF | REF | 4.1 (2.6–5.6) | REF | REF | |
| SUBGROUP | |||||||
| No metformin | Diabetese | 12.3 (8.0–16.6) | 1.4 (−3.7–6.4) | 1.1 (0.7–1.9) | 8.4 (5.6–11.2) | 2.7 (−0.3–5.7) | 1.5 (1.0–2.5) |
| Any metformin use | 10.9 (6.0–15.8) | REF | REF | 5.7 (3.1–8.3) | REF | REF | |
| No agent | Diabetese | 12.5 (7.5–17.5) | 6.3 (−4.4–17.0) | 2.2 (0.4–10.5) | 10.7 (6.2–15.3) | 8.9 (4.4–13.4)c | 6.5 (2.3–18.1)d |
| Non metformin | 11.7 (7.3–16.0) | 5.4 (−4.6–15.5) | 2.0 (0.4–9.2) | 4.7 (1.9–7.6) | 2.9 (−0.4–6.2) | 2.7 (0.9–8.2) | |
| Metformin only | 13.1 (4.8–21.3) | 6.8 (−6.0–19.6) | 2.3 (0.4–12.3) | 8.4 (4.7–12.2) | 6.6 (2.5–10.7)c | 4.9 (1.6–14.9)d | |
| Metformin+1f | 11.4 (4.6–18.2) | 5.2 (−4.8–15.2) | 1.9 (0.4–9.6) | 5.4 (1.3–9.4) | 3.6 (−0.5–7.6) | 3.1 (1.0–9.3)d | |
| Metformin+ ≥2f | 6.2 (0.0–15.8) | REF | REF | 1.8 (0.0–3.8) | REF | REF | |
| No metformin/statin | Diabetese | 12.6 (6.4–18.8) | 2.1 (−4.4–8.6) | 1.2 (0.7–2.3) | 9.6 (5.1–14.2) | 5.6 (0.4–10.9)c | 2.6 (1.1–5.9)d |
| Statin | 11.8 (7.2–16.4) | 1.4 (−6.5–9.2) | 1.1 (0.5–2.6) | 6.1 (2.5–9.6) | 2.1 (−2.9–7.0) | 1.6 (0.6–4.4) | |
| Metformin | 11.6 (3.8–19.5) | 1.2 (−8.3–10.7) | 1.1 (0.4–3.0) | 8.0 (3.0–13.1) | 4.0 (−1.5–9.5) | 2.1 (0.8–5.2) | |
| Statin + Metformin | 10.4 (4.6–16.3) | REF | REF | 4.0 (1.6–6.4) | REF | REF | |
| No statin use | Diabetese | 12.1 (6.9–17.3) | 9.1 (2.8–15.3)c | 4.4 (1.3–14.9)d | 8.9 (4.9–13.0) | 6.1 (−0.3–12.4) | 3.3 (0.6–18.3) |
| Simvastatin | 12.7 (6.4–19.1) | 9.7 (1.3–18.1)c | 4.7 (1.1–19.7)d | 4.2 (0.9–7.5) | 1.3 (−5.6–8.2) | 1.5 (0.2–11.5) | |
| Atorvastatin | 12.8 (5.6–20.1) | 9.8 (2.1–17.5)c | 4.7 (1.4–16.0)d | 4.4 (0.5–8.3) | 1.5 (−4.9–8.0) | 1.6 (0.2–10.2) | |
| Other statin | 9.7 (2.8–16.5) | 6.7 (−1.3–14.7) | 3.4 (0.8–14.0) | 7.9 (3.1–12.8) | 5.1 (−2.6–12.8) | 2.9 (0.4–19.4) | |
| Pravastatin | 3.0 (0.0–6.7) | REF | REF | 2.9 (0.0–7.9) | REF | REF | |
| Any statin useg | Diabetes | 11.0 (7.3–14.8) | 6.5 (2.0–11.1)c | 2.6 (1.4–5.1)d | 4.8 (3.1–6.5) | 1.2 (−2.3–4.7) | 1.3 (0.5–3.4) |
| Pre-diabetes | 5.4 (2.3–8.5) | 0.9 (−1.5–3.3) | 1.2 (0.8–2.0) | 5.0 (0.0–10.1) | 1.3 (−4.0–6.7) | 1.4 (0.4–4.7) | |
| Euglycemic | 4.5 (2.4–6.6) | REF | REF | 3.7 (0.8–6.5) | REF | REF | |
| No statin use | Diabetes | 12.2 (6.8–17.6) | 7.6 (2.2–13.0)c | 2.9 (1.7–4.8)d | 9.0 (4.8–13.3) | 4.9 (0.8–9.0)c | 2.3 (1.3–3.9)d |
| Pre-diabetes | 7.5 (5.5–9.5) | 2.8 (0.9–4.8)c | 1.7 (1.2–2.2)d | 7.1 (2.6–11.5) | 2.9 (−0.9–6.7) | 1.7 (1.0–3.0)d | |
| Euglycemic | 4.7 (3.7–5.6) | REF | REF | 4.2 (2.8–5.6) | REF | REF | |
abbreviations: QFT: QuantiFERON-TB Gold In Tube test; CI: confidence interval; PD: prevalence difference; OR: odds ratio; TST: tuberculin skin test
Among NHANES 2011–2012 adult participants, N=4958 had valid diabetes and QFT results; TST positive defined by induration ≥ 10 mm, N=4261 had valid diabetes and TST results
Taylor series variance estimation for 95% CI of prevalence difference
Rao-Scott Chi-square p-value <0.05
Wald Chi-square p-value <0.05
Among participants with diabetes and with QFT results (N=791) or TST results (N=685) available
Metformin in combination with any 1 or 2 other diabetes medications
Wald Chi-square test for interaction p-value < 0.03 between diabetes status and statin use with LTBI measured by TST
Any statin use among participants with diabetes was common (46.2%), and the lowest prevalence of LTBI was among those using pravastatin (3.0% by QFT and 2.9% by TST). Among those with diabetes, QFT positivity was significantly higher in participants without any statin use (OR 4.4, 95%CI 1.3–14.9) compared to those with pravastatin use. The association between no statin use and LTBI remained after adjusting for age, sex, income level, metformin use, and Hba1c (aOR 4.8 95%CI 1.4–16.5). The prevalence of TST positivity was also significantly greater among participants without combined metformin-statin use (9.6%) compared to those with combined metformin-statin therapy (4.0%) (p=0.02).
Among adult NHANES participants, the odds of QFT positivity among those with diabetes were significantly greater compared to euglycemic participants in those with (OR 2.6, 95%CI 1.4–5.1) and without statin use (OR 2.9, 95%CI 1.7–4.8). We observed multiplicative interaction between statin use and diabetes with prevalence of TST positivity, which was significantly greater among participants with diabetes and no statin use (9.0%) compared to those with diabetes and any statin use (4.8%) (p=0.03). Interaction with statin use remained significant in multivariable models adjusted for age, sex, BMI, and smoking status (p=0.03); the odds of TST positivity among participants with diabetes was greater in those without statin use (aOR 2.7, 95%CI 1.6–4.8) but not among those with diabetes that used statins (aOR 1.2, 95%CI 0.5–3.0).
Our results enhance recent findings that LTBI is more common among US adults with diabetes.[2] We report that combined metformin and statin use in patients with diabetes was associated with less than half the prevalence of LTBI (TST prevalence 4% among combined statin/metformin use vs. 10% with no statin/metformin use). Whether defined by QFT or TST, the highest prevalence of LTBI among participants with diabetes was observed among those who did not use either metformin or statins and the lowest prevalence was among those who used metformin in combination with two or more other diabetes medications. Among statin use, we report that pravastatin was associated with the lowest prevalence of LTBI by both QFT and TST. Our results also indicate the effect of diabetes on LTBI is different by statin use. Despite limitations of cross-sectional data and the potential for unmeasured confounding, when taken in the context of other studies that reported benefits of metformin and statins with TB disease, our results suggest that patients with diabetes at risk of LTBI may benefit from combination therapy with both metformin and statins. Preventing LTBI is an essential step in preventing TB disease, and both LTBI and TB disease are complications of diabetes that contribute to substantial morbidity and mortality.
ACKNOWLEDGMENTS
All authors made substantial contributions to the conceptual design, critically revised the manuscript, and approved the final version. MJM performed the analyses, drafted the initial version, and agrees to be accountable for all aspects of the work and ensures that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
FUNDING
This work was supported by the National Institute Of Allergy And Infectious Diseases of the National Institutes of Health [grant number R03AI133172 to M.J.M] and by Singapore Immunology Network A*STAR and A*STAR JCO-CDA [grant number 15302FG151 to A.S]. The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Institutes of Health.
Footnotes
CONFLICT OF INTERESTS
A.S. have filed a patent with respect to the use of metformin for controlling Mycobacterial infections (WO2017123161A1). All other authors have no conflict of interests to declare.
REFERENCES
- 1.Baker MA, Harries AD, Jeon CY, Hart JE, Kapur A, Lonnroth K, Ottmani SE, Goonesekera SD, Murray MB. The impact of diabetes on tuberculosis treatment outcomes: A systematic review. BMC Med 2011: 9(1): 81. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 2.Barron MM, Shaw KM, Bullard KM, Ali MK, Magee MJ. Diabetes is associated with increased prevalence of latent tuberculosis infection: Findings from the National Health and Nutrition Examination Survey, 2011–2012. Diabetes research and clinical practice 2018: 139: 366–379. [DOI] [PubMed] [Google Scholar]
- 3.Martinez L, Zhu L, Castellanos ME, Liu Q, Chen C, Hallowell BD, Whalen CC. Glycemic Control and the Prevalence of Tuberculosis Infection: A Population-based Observational Study. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2017: 65(12): 2060–2068. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Harries AD, Kumar AM, Satyanarayana S, Lin Y, Zachariah R, Lonnroth K, Kapur A. Diabetes mellitus and tuberculosis: programmatic management issues. The international journal of tuberculosis and lung disease : the official journal of the International Union against Tuberculosis and Lung Disease 2015: 19(8): 879–886. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Singhal A, Jie L, Kumar P, Hong GS, Leow MK, Paleja B, Tsenova L, Kurepina N, Chen J, Zolezzi F, Kreiswirth B, Poidinger M, Chee C, Kaplan G, Wang YT, De Libero G. Metformin as adjunct antituberculosis therapy. Sci Transl Med 2014: 6(263): 263ra159. [DOI] [PubMed] [Google Scholar]
- 6.Skerry C, Pinn ML, Bruiners N, Pine R, Gennaro ML, Karakousis PC. Simvastatin increases the in vivo activity of the first-line tuberculosis regimen. J Antimicrob Chemother 2014: 69(9): 2453–2457. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Degner NR, Wang JY, Golub JE, Karakousis PC. Metformin Use Reverses the Increased Mortality Associated With Diabetes Mellitus During Tuberculosis Treatment. Clinical infectious diseases : an official publication of the Infectious Diseases Society of America 2018: 66(2): 198–205. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 8.Pan SW, Yen YF, Kou YR, Chuang PH, Su VY, Feng JY, Chan YJ, Su WJ. The Risk of TB in Patients With Type 2 Diabetes Initiating Metformin vs Sulfonylurea Treatment. Chest 2018: 153(6): 1347–1357. [DOI] [PubMed] [Google Scholar]
- 9.Marupuru S, Senapati P, Pathadka S, Miraj SS, Unnikrishnan MK, Manu MK. Protective effect of metformin against tuberculosis infections in diabetic patients: an observational study of south Indian tertiary healthcare facility. Braz J Infect Dis 2017: 21(3): 312–316. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Lee YJ, Han SK, Park JH, Lee JK, Kim DK, Chung HS, Heo EY. The effect of metformin on culture conversion in tuberculosis patients with diabetes mellitus. Korean J Intern Med 2018: [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Su VY, Su WJ, Yen YF, Pan SW, Chuang PH, Feng JY, Chou KT, Yang KY, Lee YC, Chen TJ. Statin Use Is Associated With a Lower Risk of TB. Chest 2017: 152(3): 598–606. [DOI] [PubMed] [Google Scholar]
- 12.Leow MK, Dalan R, Chee CB, Earnest A, Chew DE, Tan AW, Kon WY, Jong M, Barkham T, Wang YT. Latent tuberculosis in patients with diabetes mellitus: prevalence, progression and public health implications. Experimental and clinical endocrinology & diabetes : official journal, German Society of Endocrinology [and] German Diabetes Association 2014: 122(9): 528–532. [DOI] [PubMed] [Google Scholar]
- 13.Johnson CL, Dohrmann SM, Burt VL, Mohadjer LK. National health and nutrition examination survey: sample design, 2011–2014. Vital and health statistics Series 2, Data evaluation and methods research 2014(162): 1–33. [PubMed] [Google Scholar]
- 14.Standards of medical care in diabetes−−2013. Diabetes care 2013: 36 Suppl 1: S11–66. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.CDC. National Health and Nutrition Examination Survey: Analytic Guidelines, 2011–2012. 2013. [cited; Available from: https://www.cdc.gov/nchs/data/nhanes/analytic_guidelines_11_12.pdf [PubMed]