Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

  • Loading metrics

Association between diabetes mellitus and active tuberculosis: A systematic review and meta-analysis

  • Rami H. Al-Rifai ,

    Roles Data curation, Formal analysis, Investigation, Methodology, Project administration, Resources, Software, Validation, Visualization, Writing – original draft, Writing – review & editing

    rrifai@uaeu.ac.ae, rha2006@qatar-med.cornell.edu

    Affiliations Infectious Disease Epidemiology Group, Weill Cornell Medical College–Qatar, Cornell University, Qatar Foundation–Education City, Doha, Qatar, Department of Healthcare Policy and Research, Weill Cornell Medical College, Cornell University, New York, New York, United States of America, Institute of Public Health, College of Medicine and Health Sciences, United Arab Emirates University, Al-Ain, United Arab Emirates

  • Fiona Pearson,

    Roles Investigation, Methodology, Validation, Writing – review & editing

    Affiliation Population Health Research Institute, St George’s, University of London, London, United Kingdom

  • Julia A. Critchley,

    Roles Conceptualization, Funding acquisition, Investigation, Methodology, Supervision, Validation, Writing – review & editing

    Affiliation Population Health Research Institute, St George’s, University of London, London, United Kingdom

  • Laith J. Abu-Raddad

    Roles Conceptualization, Data curation, Formal analysis, Funding acquisition, Investigation, Methodology, Project administration, Resources, Software, Supervision, Validation, Visualization, Writing – review & editing

    Affiliations Infectious Disease Epidemiology Group, Weill Cornell Medical College–Qatar, Cornell University, Qatar Foundation–Education City, Doha, Qatar, Department of Healthcare Policy and Research, Weill Cornell Medical College, Cornell University, New York, New York, United States of America, College of Public Health, Hamad bin Khalifa University, Qatar Foundation, Education City, Doha, Qatar

Abstract

The burgeoning epidemic of diabetes mellitus (DM) is one of the major global health challenges. We systematically reviewed the published literature to provide a summary estimate of the association between DM and active tuberculosis (TB). We searched Medline and EMBASE databases for studies reporting adjusted estimates on the TB–DM association published before December 22, 2015, with no restrictions on region and language. In the meta-analysis, adjusted estimates were pooled using a DerSimonian-Laird random-effects model, according to study design. Risk of bias assessment and sensitivity analyses were conducted. 44 eligible studies were included, which consisted of 58,468,404 subjects from 16 countries. Compared with non-DM patients, DM patients had 3.59–fold (95% confidence interval (CI) 2.25–5.73), 1.55–fold (95% CI 1.39–1.72), and 2.09–fold (95% CI 1.71–2.55) increased risk of active TB in four prospective, 16 retrospective, and 17 case-control studies, respectively. Country income level (3.16–fold in low/middle–vs. 1.73–fold in high–income countries), background TB incidence (2.05–fold in countries with >50 vs. 1.89–fold in countries with ≤50 TB cases per 100,000 person-year), and geographical region (2.44–fold in Asia vs. 1.71–fold in Europe and 1.73–fold in USA/Canada) affected appreciably the estimated association, but potential risk of bias, type of population (general versus clinical), and potential for duplicate data, did not. Microbiological ascertainment for TB (3.03–fold) and/or blood testing for DM (3.10–fold), as well as uncontrolled DM (3.30–fold), resulted in stronger estimated association. DM is associated with a two- to four-fold increased risk of active TB. The association was stronger when ascertainment was based on biological testing rather than medical records or self-report. The burgeoning DM epidemic could impact upon the achievements of the WHO “End TB Strategy” for reducing TB incidence.

Introduction

Despite the decline in the mortality rate of active tuberculosis (TB) since 1990, TB is ranked as one of the leading causes of death [1]. In 2015, there were an estimated 10.4 million incident TB cases worldwide [1]. The “End TB Strategy” launched by the World Health Organization (WHO) in 2016, aims to end the global TB epidemic by 2035 [1]. Targets set in this strategy include 90% reduction in TB deaths and an 80% reduction in TB incidence by 2030, compared with 2015 [1].

The growing epidemic of diabetes mellitus (DM) is set to become one of the major global health challenges [2]. The number of individuals with DM is projected to rise from 415 million in 2015 to 642 million by 2040 [3]. It is estimated that over a million TB cases among adults were affected by DM in 2012 [4]. The rising DM epidemic could contribute to an increase in TB burden. Although a few studies have failed to confirm a positive association between TB and DM [57], other studies reported a strong association between DM and active TB [812]. Based on earlier published summary effect estimates, DM increases the risk of active TB by 3.11–fold [13] and latent TB by 1.18–fold [14]. DM also has a major effect on TB treatment outcomes [15, 16]; in particular, it is associated with delayed sputum culture conversion, increased risk of treatment failure, and increased risk of TB relapse and mortality [17]. With the accumulation of recent evidence supporting the TB–DM association, there is a need for an updated understanding of the magnitude of the TB–DM association. This understanding is critical for implementation of comprehensive TB and DM control programs.

In this study, we aimed to systematically review the published literature on the association between active TB and DM, and to statistically summarize the evidence on the strength of the association.

Materials and methods

Search strategy and selection criteria

This review follows Cochrane Collaboration guidelines [18] and reports findings using the Preferred Reporting Items for Systematic reviews and Meta-analyses (PRISMA) guidelines [19].

We searched Medline (from 1945 to December 22, 2015) and EMBASE (from 1980 to December 22, 2015) databases, for studies on the TB–DM association. For inclusiveness, our search strategy covered studies that examined any risk factor for TB. The literature search protocol is summarized in the S1 Box.

Inclusion of studies was restricted to human studies that provided an estimate or allowed us to compute an estimate of the association, adjusted at least for one variable. No restrictions were made on study language, population, publication year, or region.

We excluded studies: amongst animals or children, if qualitative in design, case reports, case series, reviews, anonymous reports, editorials or author commentaries, with no appropriate control arm or comparator group, where TB patients with DM were not separated from those with other co-morbidities, of TB outcomes rather than the association, that did not report adjusted estimates of the TB–DM association, and duplicate reports.

We contacted authors of potentially eligible studies to provide the adjusted estimate for the association, if the adjusted estimate was not included in the publication. Studies whose authors did not respond were excluded. The flow diagram of study selection is shown in Fig 1. First author (RHA) screened all titles and abstracts, reviewed full-text articles, and assessed their eligibility for inclusion.

thumbnail
Fig 1. Flow diagram of study selection.

Published studies were retrieved from the MEDLINE-PubMed and EMBASE databases. TB: tuberculosis; DM: diabetes mellitus.

https://doi.org/10.1371/journal.pone.0187967.g001

Data extraction and quality assessment

Three authors (RHA, JAC, and LJA) designed the literature search terms and strategy. All potentially relevant full-text articles retrieved and reviewed to confirm eligibility. If eligible, data were extracted. All authors (RHA, FP, JAC, and LJA) contributed in assessing the eligibility of the included studies. The first author (RHA) extracted the data, which were re-extracted independently by at least one co-author (FP, JAC, or LJA). Discrepancies in data extraction were resolved by consensus or consultation with a third co-author. In addition to extracting baseline characteristics of included studies, we assessed methodological aspects, such as sampling strategy, characteristics of the study population, and TB and DM ascertainment. If a study reported more than one adjusted estimate or stratified adjusted estimates, we chose the most representative and relevant estimate with the most confirmatory DM (i.e. prioritizing HbA1c over fasting blood glucose (FBG)) and/or TB (i.e. prioritizing bacterial culture over X-ray) ascertainment strategy, or the estimate adjusted for the largest number of appropriate variables when the study reported multiple adjustment models. Five of the contacted authors have provided us with adjusted estimates [7, 2023]. Two adjusted estimates [24, 25] were extracted from a previous review [13].

We evaluated each study’s risk of bias (ROB) using nine domains for cohort and cross-sectional studies, and eleven domains for case-control studies. The ROB domains were adapted from Cochrane guidelines for systematic reviews [26, 27] and other validated quality assessment tools [28, 29]. The utilized ROB domains are presented in the S2 Box and were related to different quality criteria such as rigor of sampling strategy, TB and DM case definition and ascertainment, and DM timing in relation to TB. In case-control studies, convenience sampling of TB cases was considered as a probability-based sampling method as long as all cases in the sampling frame were selected. Based on the composite scores of the ROB domains, each cohort or cross-sectional study was classified as “potentially of low ROB” (score ≥7) or “potentially of high ROB” (score <7). Each case-control study was classified as “potentially of low ROB” (score ≥9) or “potentially of high ROB” (score <9) (S2 Box).

Statistical analysis

For studies reporting stratified crude estimates, we calculated an overall adjusted estimate by only one stratification if there was an overlap with other strata. In such studies, we prioritized pooling crude estimates that stratified by co-morbidity, location, age, or sex, consecutively. When there were two or more levels of stratification without overlap, we pooled estimates for this sub-levels stratification. If a study stratified estimates according to DM type, we pooled the overall adjusted estimate regardless of the DM type.

In sensitivity analyses, we pooled estimates of studies of “potentially low ROB”, studies unlikely to contain duplicate individual–level data, and studies in the general population rather than specific clinical populations. We further pooled estimates stratified by potential for duplicate data on same patients, country-income level, background TB incidence (≤50 or >50 cases per 100,000 person-year), region, and TB and DM ascertainment methodology. We obtained relevant data on TB incidence from the included studies or from the closest matching year made available by public databases.

We pooled adjusted estimates stratified by study design and regardless of study design using random-effects model [30]. Cochran’s Q statistic was used to test for evidence of heterogeneity [31, 32]. We estimated the I-squared (I2) as a measure of heterogeneity. We computed Tau-squared (τ2) to estimate the between-study variance of the true association between TB and DM [31, 32].

We assessed the presence of publication bias by examining the funnel plots using Egger’s t statistic to examine asymmetry (S2 Fig) [33]. We used the pooled effect estimate in four prospective studies to estimate the attributable risk fraction of DM in developing active TB among people with DM and to estimate the population attributable risk fraction of DM in developing active TB among the entire population in six high-TB-burden countries (India, Indonesia, China, Nigeria, Pakistan, and South Africa), that accounted for 60% of new TB cases in 2015 [1]. Calculations are presented in the S1 Text.

All Statistical analyses were performed using STATA SE 14 (Stata Corporation, College Station, TX) [34].

Results

We identified 19,963 publications, 44 of which were found relevant and included in this systematic review and meta-analysis (Fig 1) [512, 2025, 3564]. The included studies consisted of 58,468,404 subjects and 89,592 TB cases and they were set in 16 countries. Most studies were conducted in Taiwan (11 studies) and USA (11 studies), while only one study was in Africa [59]. Two studies of different designs stratified patients according to DM type (1 or 2) [9, 51], one study was in patients with type 1 DM [48], while the rest of the studies were either among type 2 DM patients or the type of DM was not specified (presumably, type 2 DM as it is more prevalent). There were four prospective [8, 23, 35, 36], 19 retrospective [57, 20, 22, 24, 3749], 17 case-control [911, 21, 25, 5061], and three cross-sectional [12, 62, 63] studies. One study was classified as “other” as the exact study design could not fit into the other categories [64]. One of the prospective studies was among people aged ≥65 years [35] and one was among renal allograft recipients [36]. Four of the retrospective studies were among renal patients [7, 24, 37, 41]. Several studies were national in scope, thereby including the national population as the study sample size, such as for a study from Australia [47]. Seven retrospective studies in Taiwan [22, 37, 4043, 48] and one retrospective and one case-control-study in the United kingdom, and three cross-sectional [12, 62, 63] studies. One study was classified as “other” as the exact study design could not fit into the other categories [64]. One of the prospective studies was among people aged ≥65 years [35] and one was among renal allograft recipients [36]. Four of the retrospective studies were among renal patients [7, 24, 37, 41]. Several studies were national in scope, thereby including the national population as the study sample size, such as for a study from Australia [47]. Seven retrospective studies in Taiwan [22, 37, 4043, 48] were potentially duplicate studies using the same database with overlapping years. One cross-sectional study was set in 46 countries and one retrospective and one case-control-study in the United kingdom [38, 53] were potentially duplicate studies using the same database with overlapping years. One cross-sectional study was set in 46 countries [62].

In prospective studies, estimates were adjusted at least for age except for one study that reported sex-specific crude estimates [8]. For the latter we pooled these for the present review. In one prospective study the effect estimate was not adjusted for sex [36]. In retrospective studies, all estimates were adjusted at least for age or sex. In case-control studies, estimates were adjusted at least for age and sex except in two studies where the estimate was adjusted for age but not for sex [54, 55]. All cross-sectional studies were adjusted at least for age and sex. Baseline characteristics of all included studies are shown in Tables 1 and 2.

thumbnail
Table 1. Baseline characteristics of 23 cohort studies, prospective and retrospective, that reported on the association between TB and DM and that were included in the meta-analyses.

https://doi.org/10.1371/journal.pone.0187967.t001

thumbnail
Table 2. Baseline characteristics of 17 case–control and 3 cross–sectional studies that reported on the association between TB and DM and that were included in the meta-analyses.

https://doi.org/10.1371/journal.pone.0187967.t002

The strongest estimate of the TB–DM association was 7.83 (95% CI 2.37–25.09) in a case-control study from Russia [58] followed by 7.58 (95% CI 2.94–19.49) in a prospective study from USA [23]. The lowest effect estimate was 1.00 (95% CI 0.69–1.44) in a retrospective study from Canada [49] followed by 1.16 (95% CI 0.58–2.32) in a case-control study from USA [61]. All four prospective studies demonstrated a positive association (p<0.05). Fourteen of the 19 retrospective [20, 22, 24, 3848] and 12 of the 17 case-control studies [10, 11, 25, 50, 5254, 5660] demonstrated a positive association (p<0.05) between DM and TB (Tables 1 and 2).

In the four prospective studies, DM was associated with 3.59–fold (95% CI 2.25–5.73) increased risk of TB. The I2 was estimated at 77.9% indicating that most variation across studies was due to heterogeneity in effect size rather than chance. In 16 retrospective studies, DM was associated with 1.55–fold (95% CI 1.39–1.72) increased risk of TB (I2 = 77.1%). In the 17 case-control studies, DM was associated with 2.09–fold (95% CI 1.71–2.55) increased risk of TB (I2 = 79.5%). In the three cross-sectional studies, DM was associated with 1.70–fold (95% CI 1.28–2.24) increased risk of TB (I2 = 28.9%) (Table 3). In all studies regardless of study design, DM was associated with 2.00-fold (95% CI 1.78–2.24) increased risk of TB (I2 = 90.5%). Forest plots of meta-analysis according to study design are shown in Fig 2 with summary findings presented in Table 3.

thumbnail
Fig 2. Forest plot of the meta-analyses.

Pooled findings of 44 studies reporting adjusted estimates of the association between TB and DM, stratified according to study design. Size of the square is proportional to the precision (weight) of the study-specific effect estimates. Circle is the study–specific effect point estimate. Arrows indicate that the bars are truncated to fit the plot. The diamond is centered on the summary effect estimate, and the width indicates the corresponding 95% CI. RRs: relative risk; RR: rate ratio; OR: odds ratio; HR: hazard ratio.

https://doi.org/10.1371/journal.pone.0187967.g002

thumbnail
Table 3. Summary findings of the meta-analyses for the association between DM and active TB, according to study design.

https://doi.org/10.1371/journal.pone.0187967.t003

All of the four prospective studies were judged of “potentially low ROB”. Except one study [6], the 16 retrospective studies reporting relative risk (RRs), hazard ratio (HR), or rate ratio (RR), were judged of “potentially low ROB”. Of the 17 case-control studies, 13 were judged of “potentially low ROB” [911, 25, 5054, 56, 57, 59, 61] (S1 Table).

In the sensitivity analyses presented in S1 Table, restricting the meta-analyses to studies judged of “potentially low ROB”, among only the general population, and with no potential for duplicate data on the same patients, DM patients were overall, that is by including all studies regardless of study design, at 2.00–fold (95% CI 1.77–2.27), 2.12–fold (95% CI 1.82–2.48), 1.63–fold (95% CI 1.45–1.82), respectively, increased risk of TB compared to the original overall estimate including all studies of 2.00–fold (95% CI 1.78–2.24) increased risk of TB (Table 3).

Moreover, overall, DM patients in low- or middle-income countries (3.16, 95% CI 2.20–4.53), in settings with TB incidence >50 cases per 100,000 person-year (2.05, 95% CI 1.80–2.33), or in Asian continent (2.46, 95% CI, 2.04–3.02) were at higher risk of TB than DM patients in high-income countries, in settings with TB incidence ≤50, or in Europe or USA and Canada, respectively (S1 Table).

In two prospective studies [8, 35], with microbiologically defined TB, and with blood testing for DM, patients were at 3.67–fold (95% CI 2.01–6.70) increased risk of TB. Overall, DM patients with microbiologically defined TB were at 3.03-fold (95% CI 2.31–3.98) increased risk of TB compared to 1.58–fold (95% CI 1.46–1.71) and 1.73–fold (95% CI 1.36–2.20) based on medical records or self-reported TB, respectively. Overall, blood tested DM patients were at 3.10–fold (95% CI 2.02–4.74) increased risk of TB compared to 1.60-fold (95% CI 1.18–2.17) and 1.95-fold (95% CI 0.90–4.25) based on medical records or self-reported DM, respectively (S1 Table).

Overall, DM patients with HbA1c ≥ 6.5%, FBG ≥ 120 mg/dl, or on insulin treatment, were at 1.87–fold (95% CI 1.19–2.93), 3.30-fold (95% CI 2.12–5.14), or 2.51–fold (95% CI 1.62–3.87) increased risk of TB, respectively (S2 Table).

In the four prospective (p = 0.495), 16 retrospective that reported RR, RRs, or HR (p = 0.439), three retrospective that reported an OR (p = 0.864), and three cross-sectional (p = 0.696) studies, reporting on the association between DM and TB, the Egger’s t statistic for asymmetry in the funnel plot indicated no evidence for the presence of a small-study effect (S2a, S2b, S2c and S2d Fig, respectively). However, in the 17 case-control studies, the Egger’s t statistic for asymmetry in the funnel plot indicated evidence for the presence of a small-study effect (p = 0.005) (S2e Fig).

With a summary RRs of 3.59 among DM patients in the four prospective studies, DM accounts for 72% of active TB cases among strictly DM patients (attributable risk fraction). In the six high-TB-burden countries (India, Indonesia, China, Nigeria, Pakistan, and South Africa), that accounted for 60% of new TB cases in 2015, 18%, 14%, 22%, 11%, 15%, and 15% of active TB cases in the entire population in these countries is attributed to DM, respectively. For the population attributable risk fraction, calculations are presented in S1 Text.

Discussion

In this systematic review and meta-analysis of studies on the TB–DM association, we identified a strong positive association, but with substantial heterogeneity in effect size between studies. Stronger associations were noted among DM patients where TB was confirmed microbiologically, where DM ascertainment was based on blood testing (rather than self-report or medical records only), and among DM patients with uncontrolled blood glucose. This most comprehensive review and meta-analysis included 44 observational studies, compared to nine and 13 [13] studies in the previously published two reviews, one of which included a meta-analysis of prospective studies [13]. With this increase in published evidence, our meta-analysis confirmed the result of the earlier meta-analysis [13] and strengthened the evidence base for a strong association between DM and active TB. Our results demonstrated consistent evidence of a two- to four-fold increased risk of developing TB disease for DM patients compared to non-DM patients.

As a result of aging and increasing prevalence of major DM risk factors; particularly obesity and tobacco use [1, 6769], it is projected that the number of individuals with DM will rise from 415 million in 2015 to 642 million by 2040 [3]. With the observed TB relative risk of 3.59 among DM patients in prospective studies, DM will therefore account for an increasing fraction of active TB cases in the entire population in the future. A frequent misperception is that chronic diseases such as DM are diseases of affluence [20, 70], in fact they are also common amongst poorer populations, where infectious diseases such as TB remain prevalent. Estimates suggest that the DM burden is increasing fastest in those regions where TB remains endemic [46]. From a public health perspective, it is of note that there are now more TB patients with concomitant DM than with HIV [71]. Given these findings, it may be challenging to control TB, particularly in settings that experience the double burden of the twin–epidemic of TB and DM. Robust public health intervention programs should consider tackling the underlying factors of DM such as lack of physical activity and obesity. As well as, programs to screen DM patients for TB alongside controlling blood glucose in TB patients to improve TB/DM treatment outcomes and to control this “twin epidemic”. Our findings strengthen the evidence base for how DM can impact upon the achievements of the WHO “End TB Strategy” [1].

The observed slight discrepancy in the summary estimate in the present and the previous meta-analysis [13] is partially due to the fact that one of the prospective studies [24] included in the previous meta-analysis was classified as retrospective in our review. Also, we used estimates from more confirmatory TB (bacterial culture confirmed rather than X-ray ascertained TB) and DM (HbA1c rather than FBG ascertained DM) ascertainment criteria that adjusted for the largest number of variables to pool strata-specific estimates, which in turn would produce more reliable association between TB and DM. In the previous meta-analysis [13], the estimate obtained from Kim et al., 1995 [8] was age-adjusted for all types of TB using a fixed-effect model, whereas we adjusted for sex for bacteriologically proven TB only, using a random-effects model. Moreover, we identified two more prospective studies [23, 35] that contributed 44% of the weight in our summary estimate. However, excluding one of these two studies, as DM and one-third of TB cases were ascertained by “self-report”, the summary estimate was 3.18 (95% CI 1.95–5.18), which is more comparable to that reported in the previous meta-analysis [13]. In the prospective studies, DM was mostly ascertained prior to the development of TB, suggesting that DM increases the risk of developing active TB, though some studies screened for DM at the time of TB diagnosis, and thus infection-related hyperglycemia could also explain some of the association.

We conducted sub-group sensitivity analyses to assess the heterogeneity in effect size. Several factors appeared to have contributed to this heterogeneity, including sampling methodology, study subjects, year of study, geographical location, exposure and outcome ascertainment methodology, variability within the specific subpopulation studied, sex and age-group representation in the sample, and publication bias. However, with the relatively small number of outcome measures according to study design, it was not possible to quantify the contribution of these sources of variation to the heterogeneity in the association through a meta-regression analysis.

All included studies were adjusted for at least age or sex, and estimates from majority of studies were also adjusted for different demographic and other potential confounders. This ensured that overall summary estimates were adjusted for at least the major confounding effects of age or sex. The strongest TB–DM association was observed from the four prospective studies [8, 23, 35, 36]. Data from almost three-quarters of included studies were representative of the general population. In studies reporting more than one adjusted estimate or strata-specific estimates, we included the estimates with more confirmatory ascertainment criteria for TB and/or DM, and that adjusted for the largest number of variables. This in turn produced summary estimates with lower potential of including false positive or false negative DM or TB cases. Overall, limiting the meta-analysis to studies judged as “potentially of low ROB” and excluding potentially duplicate studies did not change the direction nor magnitude of the association.

In the present review, the overall summary estimate in settings with TB incidence >50 cases per 100,000 person-year showed stronger association compared to that in settings with TB incidence ≤50 (S1 Table). This is in line with the findings of the previous meta-analysis [13]. Dobler et al., 2012 [47], hypothesized that the reason for the stronger association in settings with higher TB incidence could relate to the quality of diabetes management, assuming healthcare services may be poorer or harder to access in higher TB incidence settings.

We noticed a stronger association in blood-tested DM patients. DM patients with well-controlled glucose levels are less likely to be included when the definition of DM is based on blood glucose levels, which implies that hyperglycaemia rather than a DM diagnosis per se, increases the risk of TB [47]. DM patients suffer from immune system impairments, resulting in a lack of energy supply to immune cells, that subsequently increases virulence of infectious microorganisms [7275]. These impairments weaken the immune system response to Mycobacterium tuberculosis [7678]. This is supported by the observed stronger association in patients with uncontrolled blood glucose level (FBG ≥120 mg/dl or HbA1c ≥6.5%) (S2 Table).

There are several biological mechanisms that appear to alter the immune system and by which DM patients may develop TB [7284]. High levels of insulin were associated with a decrease in T helper 1 (Th1) immunity through a reduction in the Th1 cell to T helper 2 (Th2) cell ratio and interferon-c (IFN-c) to interleukin-4 (IL-4) ratio [80]. Other studies showed that nonspecific IFN-c levels were significantly reduced in people with diabetes compared to people with no diabetes [81], and that levels of IFN-c were negatively correlated with levels of HbA1c [82]. Neutrophils in people with diabetes were found with a lack in chemotaxis and oxidative killing potential compared to non-diabetic controls [83]. Leukocyte bactericidal activity was found to be reduced in people with diabetes, especially those with poor glucose control [84]. These observed immunologic alterations seen in people with diabetes have also been supported in experimental animal studies. Diabetic mice experimentally infected with Mycobacterium tuberculosis have higher bacterial loads compared to euglycemic mice [85, 86] with significantly lower production of IFN-c and interleukin-12 and fewer T cells [86].

Several of the included studies had methodological weaknesses. Eight studies [5, 12, 23, 52, 54, 55, 61, 62] relied on “self-reported” DM and four studies [12, 23, 54, 62] relied on “self-reported” TB. Studies that utilized blood tests to define DM may also have reported stronger associations between DM and TB, since they can identify undiagnosed DM [13], which is common in many low–and middle–income countries. In studies that relied on “self-reported” DM, subjects with controlled blood glucose (euglycemic) would be “misclassified” as DM patients. This assumption is supported in our sensitivity analyses (S1 Table). There is an additional potential misclassification of TB and DM cases as studies often used routinely collected data without validation using laboratory tests [38]. For instance, a single HbA1c measurement might misclassify individuals as either DM or non-DM patients. It is recommended that DM diagnosis should be confirmed with a repeat HbA1c test, unless clinical symptoms and plasma glucose levels >11.1mmol/l (200 mg/dl) are present [87]. Missing adjustment for potential confounders is also a noteworthy limitation. In six studies [16, 4548, 76], estimates were adjusted only for age and/or sex. Individual studies that controlled for the influence of age, sex, and smoking [35, 53, 58] produced stronger estimates than those controlled for age and sex [20, 43, 47]. Biased estimates on the TB–DM association may have occurred in studies among patients receiving dialysis [24, 41] or among subjects from specialty clinics or hospitals rather than the general population [88]. In almost all case-control studies, sampling of cases and/or controls was based on non-probability sampling. Studies using hospital-based controls reported weaker estimates for the association [13].

DM can affect different aspects of TB natural history and treatment outcomes, and therefore can impact TB transmission dynamics. An ongoing study has identified seven epidemiologically-relevant plausible effects for DM on TB natural history, and three for DM on TB treatment outcomes [89]. Our study, however, was focused on one major aspect of the TB–DM synergy, that of the association between DM status and active TB disease diagnosis—we did not assess other aspects of this synergy such as effects of DM on TB infection acquisition, TB reactivation among those latently infected, TB infectiousness, or TB treatment outcomes. A recent review, for example, reported that DM increases risk of latent TB by 1.18–fold [14], though with substantial heterogeneity across studies, and other studies have demonstrated major effects for DM on TB treatment outcomes [1517]. Comprehensive and granular characterization and quantification of the diverse effects of DM on TB is essential for a proper understanding and estimation of the impact of DM on TB epidemiology.

In most of included studies, type of DM was unclear, thereby limiting our ability to assess the association by DM type (1, 2, or both). Having said so, in the three studies that assessed the association of TB with type 1 DM [9, 48, 51], the effect size was comparable to that seen for type 2 DM studies. HIV/AIDS is a strong risk factor for TB [90], but only a fraction of included studies controlled for its effect in their assessment of the association. This may not affect appreciably our results, as HIV prevalence is very low in nearly all countries where the association was assessed. Age is another confounding factor for the TB–DM association, and nearly all studies controlled for this factor. Our study was focused on the overall effect of DM on TB disease, and we did not provide a pooled effect size stratified by age. Despite these limitations, our review and meta-analyses compiled and summarized important data and critically provided narrative information from a large number of studies that reported on the TB–DM association.

Conclusions

Our systematic review and meta-analysis demonstrated consistent evidence of a substantially increased risk of TB disease among people with DM. This evidence was based on data from studies using different designs and reported from six continents. DM patients with uncontrolled blood glucose (measured by higher FBG or HbA1c) appeared to be at higher risk of active TB than patients with controlled DM. Efforts to halt the burgeoning DM epidemic would have an accompanied benefit of alleviating the global burdens of DM and TB. The burgeoning epidemic of DM is likely to impact upon the achievements of the WHO “End TB Strategy”. Our findings inform strategy planning of health service provision and implementation of effective prevention programs to control the “twin epidemic” of DM and TB.

Supporting information

S1 Table. Summary findings of the meta-analyses for the sensitivity analyses of the association between DM and active TB in 44 studies, according to study design and overall.

https://doi.org/10.1371/journal.pone.0187967.s001

(DOCX)

S2 Table. Estimates and summary estimates of the association between DM and active TB, according to DM ascertainment in blood-tested patients and study design.

https://doi.org/10.1371/journal.pone.0187967.s002

(DOCX)

S2 Fig. Funnel plots assessing the risk of publication bias according to study design.

https://doi.org/10.1371/journal.pone.0187967.s004

(DOCX)

S1 Box. Data sources and search criteria for systematically reviewing literature reporting on active tuberculosis (TB) and diabetes mellitus (DM) association.

https://doi.org/10.1371/journal.pone.0187967.s005

(DOCX)

S2 Box. Criteria used to assess quality of included studies.

https://doi.org/10.1371/journal.pone.0187967.s006

(DOCX)

S1 Text. Calculation of attributable risk fraction of TB among DM patients and population attributable risk fraction of TB due to DM.

https://doi.org/10.1371/journal.pone.0187967.s007

(DOCX)

Acknowledgments

The authors are very grateful for infrastructure support provided by the Biostatistics, Epidemiology, and Biomathematics Research Core at Weill Cornell Medicine-Qatar. This publication was made possible by NPRP grant number 7-627-3-167 from the Qatar National Research Fund (a member of Qatar Foundation). The findings achieved herein are solely the responsibility of the authors. The authors are also grateful for infrastructure support provided by the Biostatistics, Epidemiology, and Biomathematics Research Core at Weill Cornell Medicine-Qatar. JAC is also supported by the Higher Education Funding Council for England.

References

  1. 1. World Health Organization. Global tuberculosis report 2016. http://wwwwhoint/tb/publications/global_report/gtbr2016_main_textpdf?ua=1 (accessed September 15, 2016).
  2. 2. Kassebaum NJ, Arora M, Barber RM, Bhutta ZA, Carter A, Casey DC, et al. Global, regional, and national disability-adjusted life-years (DALYs) for 315 diseases and injuries and healthy life expectancy (HALE), 1990–2015: a systematic analysis for the Global Burden of Disease Study 2015. Lancet. 2016;388(10053):1603–58. pmid:27733283
  3. 3. International Diabetese Fedration. International Diabetese Fedration Atals 2015. http://www.diabetesatlas.org/(accessed September 15, 2016).
  4. 4. Lonnroth K, Roglic G, Harries AD. Improving tuberculosis prevention and care through addressing the global diabetes epidemic: from evidence to policy and practice (vol 2, pg 730, 2014). Lancet Diabetes & Endocrinology. 2014;2(10):E16–E.
  5. 5. Chen W, Shu W, Wang M, Hou YC, Xia YY, Xu WG, et al. Pulmonary Tuberculosis Incidence and Risk Factors in Rural Areas of China: A Cohort Study. Plos One. 2013;8(3). pmid:23554875
  6. 6. Moran-Mendoza O, Marion SA, Elwood K, Patrick D, FitzGerald JM. Risk factors for developing tuberculosis: a 12-year follow-up of contacts of tuberculosis cases. International Journal of Tuberculosis and Lung Disease. 2010;14(9):1112–9. pmid:20819255
  7. 7. Rungruanghiranya S, Ekpanyaskul C, Jirasiritum S, Nilthong C, Pipatpanawong K, Mavichak V. Tuberculosis in Thai renal transplant recipients: A 15-year experience. Transplantation Proceedings. 2008;40(7):2376–9. pmid:18790239
  8. 8. Kim SJ, Hong YP, Lew WJ, Yang SC, Lee EG. Incidence of pulmonary tuberculosis among diabetics. Tubercle and Lung Disease. 1995;76(6):529–33. pmid:8593374
  9. 9. PablosMendez A, Blustein J, Knirsch CA. The role of diabetes mellitus in the higher prevalence of tuberculosis among Hispanics. American Journal of Public Health. 1997;87(4):574–9. pmid:9146434
  10. 10. Alisjahbana B, van Crevel R, Sahiratmadja E, den Heijer M, Maya A, Istriana E, et al. Diabetes mellitus is strongly associated with tuberculosis in Indonesia. International Journal of Tuberculosis and Lung Disease. 2006;10(6):696–700. pmid:16776459
  11. 11. Perez A, Brown HS, Restrepo BI. Association between tuberculosis and diabetes in the Mexican border and non-border regions of Texas. American Journal of Tropical Medicine and Hygiene. 2006;74(4):604–11. pmid:16606993
  12. 12. Marks SM. Diabetes and tuberculosis, US National Health Interview Survey, 2000–2005. International Journal of Tuberculosis and Lung Disease. 2011;15(7):982–4. pmid:21682976
  13. 13. Jeon CY, Murray MB. Diabetes mellitus increases the risk of active tuberculosis: A systematic review of 13 observational studies (vol 5, pg e152, 2008). Plos Medicine. 2008;5(8):1298-.
  14. 14. Lee MR, Huang YP, Kuo YT, Luo CH, Shih YJ, Shu CC, et al. Diabetes mellitus and latent tuberculosis infection: a systemic review and meta-analysis. Clin Infect Dis. 2016. pmid:27986673.
  15. 15. Baker MA, Harries AD, Jeon CY, Hart JE, Kapur A, Lonnroth K, et al. The impact of diabetes on tuberculosis treatment outcomes: A systematic review. Bmc Medicine. 2011;9. pmid:21722362
  16. 16. Faurholt-Jepsen D, Range N, Praygod G, Kidola J, Faurholt-Jepsen M, Aabye MG, et al. The role of diabetes co-morbidity for tuberculosis treatment outcomes: a prospective cohort study from Mwanza, Tanzania. Bmc Infect Dis. 2012;12. pmid:22839693
  17. 17. Faurholt-Jepsen D, Range N, PrayGod G, Jeremiah K, Faurholt-Jepsen M, Aabye MG, et al. Diabetes is a strong predictor of mortality during tuberculosis treatment: a prospective cohort study among tuberculosis patients from Mwanza, Tanzania. Trop Med Int Health. 2013;18(7):822–9. pmid:23648145
  18. 18. Higgins JP, Green S. Cochrane handbook for systematic reviews of interventions: Wiley Online Library; 2008.
  19. 19. Moher D, Liberati A, Tetzlaff J, Altman DG. Preferred reporting items for systematic reviews and meta-analyses: the PRISMA statement. Annals of internal medicine. 2009;151(4):264–9. pmid:19622511
  20. 20. Kamper-Jorgensen Z, Carstensen B, Norredam M, Bygbjerg IC, Andersen PH, Jorgensen ME. Diabetes-related tuberculosis in Denmark: effect of ethnicity, diabetes duration and year of diagnosis. International Journal of Tuberculosis and Lung Disease. 2015;19(10):1169–75. pmid:26459528
  21. 21. Ku NS, Choi YH, Kim YK, Choi JP, Kim JM, Choi JY. Incidence of and risk factors for active tuberculosis in human immunodeficiency virus-infected patients in South Korea. International Journal of Tuberculosis and Lung Disease. 2013;17(6):777–81. pmid:23676161
  22. 22. Chung WS, Lin CL, Hung CT, Chu YH, Sung FC, Kao CH, et al. Tuberculosis increases the subsequent risk of acute coronary syndrome: a nationwide population-based cohort study. International Journal of Tuberculosis and Lung Disease. 2014;18(1):79–83. pmid:24365557
  23. 23. Cegielski JP, Arab L, Cornoni-Huntley J. Nutritional Risk Factors for Tuberculosis Among Adults in the United States, 19711992. American Journal of Epidemiology. 2012;176(5):409–22. pmid:22791739
  24. 24. Chen CH, Lian JD, Cheng CH, Wu MJ, Lee WC, Shu KH. Mycobacterium tuberculosis infection following renal transplantation in Taiwan. Transplant Infectious Disease. 2006;8(3):148–56. pmid:16913973
  25. 25. Brassard P, Kezouh A, Suissa S. Antirheumatic drugs and the risk of tuberculosis. Clinical Infectious Diseases. 2006;43(6):717–22. pmid:16912945
  26. 26. Higgins JPT, Green S., Cochrane Collaboration. Cochrane Handbook For Systematic Reviews Of Interventions. England, Hoboken, NJ.: Wiley-Blackwell Chichester, 2008.
  27. 27. Higgins JP, Altman D.G., Gotzsche P.C., Juni P., Moher D., Oxman A.D., Savovic J., Schulz K.F., Weeks L., Sterne J.A.C., Cochrane Statistical Methods Group. The Cochrane Collaboration’s tool for assessing risk of bias in randomised trials. BMJ. 2011;343:d5928. pmid:22008217
  28. 28. Hoy D, Brooks P, Woolf A, Blyth F, March L, Bain C, et al. Assessing risk of bias in prevalence studies: modification of an existing tool and evidence of interrater agreement. J Clin Epidemiol. 2012;65(9):934–9. pmid:22742910.
  29. 29. Viswanathan M, Berkman ND, Dryden DM, Hartling L. Assessing Risk of Bias and Confounding in Observational Studies of Interventions or Exposures: Further Development of the RTI Item Bank. AHRQ Methods for Effective Health Care. Rockville (MD) 2013.
  30. 30. Harris RJ, Bradburn MJ, Deeks JJ, Harbord RM, Altman DG, Sterne JAC. metan: fixed- and random-effects meta-analysis. Stata Journal. 2008;8(1):3–28.
  31. 31. Borenstein M. Introduction to meta-analysis. Chichester, UK: John Wiley & Sons. 2009;xxviii,: 421 p.p.
  32. 32. Higgins JPT, Thompson SG, Deeks JJ, Altman DG. Measuring inconsistency in meta-analyses. British Medical Journal. 2003;327(7414):557–60. pmid:12958120
  33. 33. Egger M, Smith GD. Meta-Analysis. Potentials and promise. BMJ. 1997;315(7119):1371–4. pmid:9432250.
  34. 34. StataCorp. 2015. Stata Statistical Software: Release 14. College Station, TX: StataCorp LP.
  35. 35. Leung CC, Lam TH, Chan WM, Yew WW, Ho KS, Leung GM, et al. Diabetic control and risk of tuberculosis: A cohort study. American Journal of Epidemiology. 2008;167(12):1486–94. pmid:18400769
  36. 36. John GT, Shankar V, Abraham AM, Mukundan U, Thomas PP, Jacob CK. Risk factors for post-transplant tuberculosis. Kidney International. 2001;60(3):1148–53. pmid:11532111
  37. 37. Ou SM, Liu CJ, Teng CJ, Lin YT, Chang YS, Chiang SC, et al. Impact of pulmonary and extrapulmonary tuberculosis infection in kidney transplantation: a nationwide population-based study in Taiwan. Transplant Infectious Disease. 2012;14(5):502–9. pmid:22571347
  38. 38. Pealing L, Wing K, Mathur R, Prieto-Merino D, Smeeth L, Moore DAJ. Risk of tuberculosis in patients with diabetes: population based cohort study using the UK Clinical Practice Research Datalink. Bmc Medicine. 2015;13. pmid:26048371
  39. 39. Baker MA, Lin HH, Chang HY, Murray MB. The Risk of Tuberculosis Disease Among Persons With Diabetes Mellitus: A Prospective Cohort Study. Clinical Infectious Diseases. 2012;54(6):818–25. pmid:22238171
  40. 40. Kuo MC, Lin SH, Lin CH, Mao IC, Chang SJ, Hsieh MC. Type 2 Diabetes: An Independent Risk Factor for Tuberculosis: A Nationwide Population-Based Study. Plos One. 2013;8(11). pmid:24236069
  41. 41. Hu HY, Wu CY, Huang N, Chou YJ, Chang YC, Chu D. Increased risk of tuberculosis in patients with end-stage renal disease: a population-based cohort study in Taiwan, a country of high incidence of end-stage renal disease. Epidemiology and Infection. 2014;142(1):191–9. pmid:23510593
  42. 42. Lee CH, Lee MC, Shu CC, Lim CS, Wang JY, Lee LN, et al. Risk factors for pulmonary tuberculosis in patients with chronic obstructive airway disease in Taiwan: a nationwide cohort study. Bmc Infect Dis. 2013;13. pmid:23631563
  43. 43. Wu CY, Hu HY, Pu CY, Huang N, Shen HC, Li CP, et al. Aerodigestive tract, lung and haematological cancers are risk factors for tuberculosis: an 8-year population-based study. International Journal of Tuberculosis and Lung Disease. 2011;15(1):125–30. pmid:21276308
  44. 44. Demlow SE, Oh P, Barry PM. Increased risk of tuberculosis among foreign-born persons with diabetes in California, 2010–2012. Bmc Public Health. 2015;15. pmid:25884596
  45. 45. Suwanpimolkul G, Grinsdale JA, Jarlsberg LG, Higashi J, Osmond DH, Hopewell PC, et al. Association between Diabetes Mellitus and Tuberculosis in United States-Born and Foreign-Born Populations in San Francisco. Plos One. 2014;9(12). pmid:25478954
  46. 46. Young F, Wotton CJ, Critchley JA, Unwin NC, Goldacre MJ. Increased risk of tuberculosis disease in people with diabetes mellitus: record-linkage study in a UK population. Journal of Epidemiology and Community Health. 2012;66(6):519–23. pmid:21109542
  47. 47. Dobler CC, Flack JR, Marks GB. Risk of tuberculosis among people with diabetes mellitus: an Australian nationwide cohort study. Bmj Open. 2012;2(1). pmid:22331390
  48. 48. Shen TC, Lin CL, Wei CC, Liao WC, Chen WC, Chen CH, et al. Increased Risk of Tuberculosis in Patients With Type 1 Diabetes Mellitus: Results From a Population-Based Cohort Study in Taiwan. Medicine. 2014;93(16). pmid:25310745
  49. 49. Dyck RF, Klomp H, Marciniuk DD, Tan L, Stang MR, Ward HA, et al. The relationship between diabetes and tuberculosis in Saskatchewan—Comparison of registered Indians and other Saskatchewan people. Canadian Journal of Public Health-Revue Canadienne De Sante Publique. 2007;98(1):55–9. pmid:17278679
  50. 50. Lai SW, Lin CL, Liao KF, Tsai SM. Increased risk of pulmonary tuberculosis among patients with appendectomy in Taiwan. European Journal of Clinical Microbiology & Infectious Diseases. 2014;33(9):1573–7. pmid:24781006
  51. 51. Leegaard A, Riis A, Kornum JB, Prahl JB, Thomsen VO, Sorensen HT, et al. Diabetes, Glycemic Control, and Risk of Tuberculosis A population-based case-control study. Diabetes Care. 2011;34(12):2530–5. pmid:21972407
  52. 52. Jurcev-Savicevic A, Mulic R, Ban B, Kozul K, Bacun-Ivcek L, Valic J, et al. Risk factors for pulmonary tuberculosis in Croatia: a matched case-control study. Bmc Public Health. 2013;13. pmid:24144113
  53. 53. Jick SS, Lieberman ES, Rahman MU, Choi HK. Glucocorticoid use, other associated factors, and the risk of tuberculosis. Arthritis & Rheumatism-Arthritis Care & Research. 2006;55(1):19–26. pmid:16463407
  54. 54. Valerie Corris NUaJC. Quantifying the association between tuberculosis and diabetes in the US: a case-control analysis. Chronic Illness. 2012;8(2):121–34. pmid:22387690
  55. 55. Buskin SE, Gale JL, Weiss NS, Nolan CM. Tuberculosis Risk-Factors in Adults in King County, Washington, 1988 through 1990. American Journal of Public Health. 1994;84(11):1750–6. pmid:7977912
  56. 56. Mori MA, Leonardson G, Welty TK. The Benefits of Isoniazid Chemoprophylaxis and Risk-Factors for Tuberculosis among Oglala Sioux Indians. Archives of Internal Medicine. 1992;152(3):547–50. pmid:1546917
  57. 57. Viney K, Cavanaugh J, Kienene T, Harley D, Kelly PM, Sleigh A, et al. Tuberculosis and diabetes mellitus in the Republic of Kiribati: a case-control study. Trop Med Int Health. 2015;20(5):650–7. pmid:25598275
  58. 58. Coker R, McKee M, Atun R, Dimitrova B, Dodonova E, Kuznetsov S, et al. Risk factors for pulmonary tuberculosis in Russia: case-control study. British Medical Journal. 2006;332(7533):85–7. pmid:16339219
  59. 59. Faurholt-Jepsen D, Range N, PrayGod G, Jeremiah K, Faurholt-Jepsen M, Aabye MG, et al. Diabetes Is a Risk Factor for Pulmonary Tuberculosis: A Case-Control Study from Mwanza, Tanzania. Plos One. 2011;6(8). pmid:21912626
  60. 60. Wu HP P Y, Hua CC, Shieh WB, Jiang BY, Yu TJ. Pneumoconiosis and liver cirrhosis are not risk factors for tuberculosis in patients with pulmonary infection. Respirology. 2007;12(3):416–9. pmid:17539848
  61. 61. Rosenman KD, Hall N. Occupational risk factors for developing tuberculosis. American Journal of Industrial Medicine. 1996;30(2):148–54. pmid:8844044
  62. 62. Goldhaber-Fiebert JD, Jeon CY, Cohen T, Murray MB. Diabetes mellitus and tuberculosis in countries with high tuberculosis burdens: individual risks and social determinants. International Journal of Epidemiology. 2011;40(2):417–28. pmid:21252210
  63. 63. Wang QZ, Ma AG, Han XX, Zhao SL, Cai J, Ma YB, et al. Prevalence of Type 2 Diabetes among Newly Detected Pulmonary Tuberculosis Patients in China: A Community Based Cohort Study. Plos One. 2013;8(12). pmid:24367535
  64. 64. Ponce-De-Leon A, Garcia-Garcia MD, Garcia-Sancho C, Gomez-Perez FJ, Valdespino-Gomez JL, Olaiz-Fernandez G, et al. Tuberculosis and diabetes in southern Mexico. Diabetes Care. 2004;27(7):1584–90. pmid:15220232
  65. 65. Lo HY, Chou P, Yang SL, Lee CY, Kuo HS. Trends in Tuberculosis in Taiwan, 2002–2008. Journal of the Formosan Medical Association. 2011;110(8):501–10. pmid:21783019
  66. 66. U.S. Department of Health and Human Services, Public Health Service Centers for Disease Control. Case definitions for public health surveillance. MMWR. 1990;39:39–40.
  67. 67. Gaziano TA, Bitton A, Anand S, Abrahams-Gessel S, Murphy A. Growing epidemic of coronary heart disease in low- and middle-income countries. Curr Probl Cardiol. 2010;35(2):72–115. pmid:20109979.
  68. 68. Hu FB. Globalization of diabetes: the role of diet, lifestyle, and genes. Diabetes Care. 2011;34(6):1249–57. pmid:21617109.
  69. 69. Malik VS, Willett WC, Hu FB. Global obesity: trends, risk factors and policy implications. Nat Rev Endocrinol. 2013;9(1):13–27. pmid:23165161.
  70. 70. Ezzati M, Vander Hoorn S, Lawes CM, Leach R, James WP, Lopez AD, et al. Rethinking the "diseases of affluence" paradigm: global patterns of nutritional risks in relation to economic development. PLoS Med. 2005;2(5):e133. pmid:15916467.
  71. 71. Ruslami R, Aarnoutse RE, Alisjahbana B, van der Ven AJ, van Crevel R. Implications of the global increase of diabetes for tuberculosis control and patient care. Trop Med Int Health. 2010;15(11):1289–99. pmid:20955495.
  72. 72. Casqueiro J, Casqueiro J, Alves C. Infections in patients with diabetes mellitus: A review of pathogenesis. Indian J Endocrinol Metab. 2012;16 Suppl 1:S27–36. pmid:22701840.
  73. 73. Peleg AY, Weerarathna T, McCarthy JS, Davis TME. Common infections in diabetes: pathogenesis, management and relationship to glycaemic control. Diabetes-Metabolism Research and Reviews. 2007;23(1):3–13. pmid:16960917
  74. 74. Joshi N, Caputo GM, Weitekamp MR, Karchmer AW. Infections in patients with diabetes mellitus. N Engl J Med. 1999;341(25):1906–12. pmid:10601511.
  75. 75. Vardakas KZ, Siempos II, Falagas ME. Diabetes mellitus as a risk factor for nosocomial pneumonia and associated mortality. Diabetic Medicine. 2007;24(10):1168–71. pmid:17888136
  76. 76. Restrepo BI, Camerlin AJ, Rahbar MH, Wang WW, Restrepo MA, Zarate I, et al. Cross-sectional assessment reveals high diabetes prevalence among newly-diagnosed tuberculosis cases. Bulletin of the World Health Organization. 2011;89(5):352–9. pmid:21556303
  77. 77. Harries AD, Lin Y, Satyanarayana S, Lonnroth K, Li L, Wilson N, et al. The looming epidemic of diabetes-associated tuberculosis: learning lessons from HIV-associated tuberculosis. International Journal of Tuberculosis and Lung Disease. 2011;15(11):1436–44. pmid:21902876
  78. 78. Dooley KE, Chaisson RE. Tuberculosis and diabetes mellitus: convergence of two epidemics. Lancet Infectious Diseases. 2009;9(12):737–46. pmid:19926034
  79. 79. Geerlings SE, Hoepelman AIM. Immune dysfunction in patients with diabetes mellitus (DM). Fems Immunology and Medical Microbiology. 1999;26(3–4):259–65. pmid:10575137
  80. 80. Viardot A, Grey ST, Mackay F, Chisholm D. Potential antiinflammatory role of insulin via the preferential polarization of effector T cells toward a T helper 2 phenotype. Endocrinology. 2007;148(1):346–53. pmid:17008395.
  81. 81. Stalenhoef JE, Alisjahbana B, Nelwan EJ, van der Ven-Jongekrijg J, Ottenhoff TH, van der Meer JW, et al. The role of interferon-gamma in the increased tuberculosis risk in type 2 diabetes mellitus. Eur J Clin Microbiol Infect Dis. 2008;27(2):97–103. pmid:17962984.
  82. 82. Tsukaguchi K, Okamura H, Ikuno M, Kobayashi A, Fukuoka A, Takenaka H, et al. [The relation between diabetes mellitus and IFN-gamma, IL-12 and IL-10 productions by CD4+ alpha beta T cells and monocytes in patients with pulmonary tuberculosis]. Kekkaku. 1997;72(11):617–22. pmid:9423299.
  83. 83. Delamaire M, Maugendre D, Moreno M, Le Goff MC, Allannic H, Genetet B. Impaired leucocyte functions in diabetic patients. Diabet Med. 1997;14(1):29–34. pmid:9017350.
  84. 84. Rayfield EJ, Ault MJ, Keusch GT, Brothers MJ, Nechemias C, Smith H. Infection and diabetes: the case for glucose control. Am J Med. 1982;72(3):439–50. pmid:7036735.
  85. 85. Yamashiro S, Kawakami K, Uezu K, Kinjo T, Miyagi K, Nakamura K, et al. Lower expression of Th1-related cytokines and inducible nitric oxide synthase in mice with streptozotocin-induced diabetes mellitus infected with Mycobacterium tuberculosis. Clin Exp Immunol. 2005;139(1):57–64. pmid:15606614
  86. 86. Martens GW, Arikan MC, Lee JH, Ren FC, Greiner D, Kornfeld H. Tuberculosis susceptibility of diabetic mice. Am J Resp Cell Mol. 2007;37(5):518–24.
  87. 87. World Health Organization. Use of Glycated Haemoglobin (HbA1c) in the Diagnosis of Diabetes Mellitus. http://www.who.int/cardiovascular_diseases/report-hba1c_2011_edited.pdf?ua=1 (accessed October 15, 2017). 2011.
  88. 88. Feinstein AR, Walter SD, Horwitz RI. An Analysis of Berksons Bias in Case Control Studies. Journal of Chronic Diseases. 1986;39(7):495–504. pmid:3722313
  89. 89. Awad SF, Dargham SR, Omori R, Pearson F, Critchley J, Abu-Raddad LJ. Characterizing the Impact of Diabetes Mellitus on Tuberculosis Epidemiology: Analytical Insights. under preperation. 2017.
  90. 90. Lienhardt C, Rodrigues LC. Estimation of the impact of the human immunodeficiency virus infection on tuberculosis: tuberculosis risks re-visited? Int J Tuberc Lung Dis. 1997;1(3):196–204. pmid:9432364.