Incidence and prevalence of pulmonary tuberculosis among patients with type 2 diabetes mellitus: a systematic review and meta-analysis

Abstract Background The epidemic of type 2 diabetes mellitus (T2DM) poses a great challenge to pulmonary tuberculosis (PTB) control. However, the incidence and prevalence of PTB among T2DM patients has not been fully determined. This meta-analysis aimed to provide the estimation on the global incidence and prevalence of PTB among T2DM patients (T2DM-PTB). Methods Online databases including Web of Science, PubMed, China National Knowledge Infrastructure and Cochrane Library were searched for all relevant studies that reported the incidence or prevalence of T2DM-PTB through 31 January 2022. Pooled incidence and prevalence of T2DM-PTB with 95% confidence interval (CI) was estimated by the random-effect model. All statistical analyses were performed using R software. Results A total of 24 studies (14 cohort studies, 10 cross-sectional studies) were included. The pooled incidence and prevalence of T2DM-PTB were 129.89 per 100,000 person-years (95% confidence interval (CI): 97.55–172.95) and 511.19 per 100,000 (95% CI: 375.94–695.09), respectively. Subgroup analyses identified that the incidence of T2DM-PTB was significantly higher in Asia (187.20 per 100,000 person-years, 95% CI: 147.76–237.17), in countries with a high TB burden (172.04 per 100,000 person-years, 95% CI: 122.98–240.68) and in studies whose data collection ended before 2011 (219.81 per 100,000 person-years, 95% CI: 176.15–274.28), but lower in studies using International Classification of Diseases-10 codes (73.75 per 100,000 person-years, 95% CI: 40.92–132.91). The prevalence of T2DM-PTB was significantly higher in countries with a high TB burden (692.15 per 100,000, 95% CI: 468.75–1022.04), but lower in Europe (105.01 per 100,000, 95% CI: 72.55–151.98). Conclusions This systematic review and meta-analysis suggests high global incidence and prevalence of PTB among T2DM patients, underlining the necessity of more preventive interventions among T2DM patients especially in countries with a high TB-burden. Key messages A total of 24 studies (14 cohort studies, 10 cross-sectional studies) containing 2,569,451 T2DM patients were included in this meta-analysis. The pooled incidence and prevalence of T2DM-PTB are 129.89 per 100,000 person-years (95% CI: 97.55–172.95) and 511.19 per 100,000 (95% CI: 375.94–695.09) respectively. The incidence of T2DM-PTB was significantly higher in Asia, in countries with a high TB burden and in studies whose data collection ended before 2011, but lower in studies using International Classification of Diseases-10 codes. The prevalence of T2DM-PTB was significantly higher in countries with a high TB-burden, but lower in Europe.


Introduction
As a major global health problem, tuberculosis (TB) is a serious chronic infectious disease caused by Mycobacterium tuberculosis (MTB). TB occurs mostly in the lungs, leading to pulmonary TB (PTB). The 2021 Global TB reported that there were approximately 9.87 million new TB patients worldwide with an incidence rate of 127 per 100,000 in 2020 [1]. Although substantial efforts have contributed to the decline in the global TB epidemic, the pace of progress needs to be speed up to curb TB burden and achieve goals of WHO's "End TB strategy".
Diabetes mellitus (DM) is a group of metabolic diseases characterized by hyperglycaemia, which is known to be an important predisposing factor in the development of PTB. The incidence of DM is increasing rapidly. In 2019, the number of DM patients has exceeded 460 million, which is expected to reach 700 million by 2045 [2]. Type 2 diabetes mellitus (T2DM) accounts for almost 90% of DM that caused by a combination of resistance to insulin action and an inadequate compensatory insulin secretory response [3]. T2DM frequently causes adverse microvascular and macrovascular comorbidities [4], such as chronic kidney disease, hypertension, myocardial infarction. A previous meta-analysis demonstrated that DM patients have two-to four-fold increased risk of active TB in comparison to the non-DM population [5], which may be related to the impaired innate and adaptive immune system caused by the long-term poor glycemic control. Additionally, the rates of treatment failure and death of PTB patients with concomitant DM are higher [6]. Therefore, it highlights the importance to understand the incidence and prevalence of T2DM-PTB co-morbidity, especially in high TB-burden countries.
To the best of our knowledge, a growing body of research have explored the mutual relationship of these two diseases and paid more attention to the bidirectional screenings of TB and DM [7]. Risk factors have been determined for PTB in T2DM patients previously [8][9][10]. However, current evidence from epidemiologic studies has been limited. Besides, the global incidence and prevalence of PTB among T2DM patients has not been fully determined. In order to implement comprehensive prevention strategy, it is imperative to conduct a systematic review and metaanalysis to review the current knowledge about the incidence and prevalence, and associated risk factors of PTB among patients with T2DM.

Search strategy
This meta-analysis followed the guidelines of Preferred Reporting Items for Systematic Reviews and Meta-Analyses statement. A literature search was conducted through electronic and manual searches on Web of Science, PubMed, China National Knowledge Infrastructure (CNKI) and Cochrane Library to identify all relevant studies published from their inception to 31 January 2022. The following keywords included: ("tuberculosis" or "TB" or "pulmonary tuberculosis" or "PTB") and ("diabetes mellitus" or "DM" or "type 2 diabetes mellitus" or "T2DM") without restriction to regions and languages. The bibliographic database searches were supplemented by screening the reference lists of all relevant publications. All search records were inserted into EndNote X8 software.

Eligibility criteria
We included eligible observational studies (cohort and cross-sectional studies) that reported the incidence and/or prevalence of PTB in patients with T2DM or provided enough data (the number of T2DM-PTB cases and person-years of follow up) to compute the estimates. For studies with overlap population, only the study with the latest information or larger sample size was enrolled.

Exclusion criteria
The exclusion criteria for the studies included in this meta-analysis were as follows: (1) the full text could not be retrieved, (2) studies were performed in animals, (3) data were insufficient for analysis, (4) the diagnostic criteria of T2DM were not clearly reported, (5) studies were published as reviews, meta-analysis, case series with small sample size (<50), meeting abstracts or editorials.

Data extraction and quality assessment of included studies
Two investigators (Qian Wu and Yang Liu) independently screened the titles and abstracts of all articles for eligibility, and extracted the following data of each eligible study: first author's name, publication year, the prevalence of T2DM-PTB (or number of T2DM and T2DM-PTB cases), the incidence of T2DM-PTB (or number of T2DM and T2DM-PTB cases, person-years of follow up), study period, study design, region, TB burden of study countries, definitions or diagnostic criteria for PTB. Quality of cohort studies was assessed using Newcastle-Ottawa quality assessment scale (NOS), with a score of !5 out of 10 considered as high-quality score [11]. Risk of bias in prevalence studies were assessed by using Hoy's risk of bias tool [12]. Any disagreement between the researchers was resolved through discussion. The corresponding authors were contacted by e-mail for additional data.

Statistical analysis
The incidence and prevalence of PTB among T2DM patients was reported with 95% confidence intervals (CIs) and presented in the forest plot. For cohort studies, T2DM-PTB incidence should be corrected for the total follow-up time and expressed as cases per 100,000 person-years. T2DM-PTB prevalence should be expressed as cases per 100,000. I 2 statistic was used to quantify the effect of heterogeneity. In the case of significantly between-study heterogeneity (I 2 value > 50% and the p value for Cochrane Q test <0 .10), the random-effects model was selected to compute the pooled estimate of prevalence [13]. Otherwise, the fixed-effect model was used. To identify the possible source of heterogeneity, we conducted subgroup analysis stratified by region, the TB burden of study countries, study design, diagnostic criteria for PTB and ending year of data collection. In addition, the visual inspection funnel plot with Egger's linear regression test [14] and Begg's rank correlation test [15] were used to assess potential publication biases. In order to assess the stability and reliability of the estimates, sensitivity analysis was conducted by iteratively removing one study from the meta-analysis. For all tests, p < 0.05 was deemed to be statistically significant. All statistical analyses were conducted using R software (version 3.6.1; http://www.R-project.org) with the "meta" package.

Search results
After the comprehensive literature search, we identified 16,552 citations from Web of Science (6200), PubMed (5419), CNKI database (4600), Cochrane library (333). A total of 11,752 records remained after duplicates were excluded. After screening titles and abstracts, 67 eligible studies were retrieved. Finally, 43 studies were excluded accordingly and 24 studies  were included in this systematic review with meta-analysis ( Figure 1). A total of 2,569,451 patients with T2DM were included, out of which 12,642 patients were T2DM-PTB patients. There were 14 cohort studies and 10 cross-sectional studies. All of these included studies were conducted from 1998 to 2018. The sample size of these studies ranged from 630 to 840,899. In terms of study region, 19 were from Asia (79.17%), 3 from Africa (12.50%) and 2 from Europe (8.33%). The majority of the included cohort studies were of moderate or high quality. The overall risk of bias of the prevalence studies was low in 7 (29.17%) studies, moderate in 12 (50%) studies and high in 5 (20.83%) studies.

T2DM-PTB incidence
Fourteen cohort studies containing 2,533,555 T2DM patients reported the incidence of T2DM-PTB. Since the three studies among them were conducted in different groups, we analysed these studies separately [18,20,25]. The main characteristics of the eligible studies were extracted and listed in Table 1. For significant heterogeneity among the included studies (I 2 ¼99.6%, tau 2 ¼0.37, p < 0.001), random effects model was used for incidence pooling. As shown in the forest plot ( Figure 2), the overall pooled incidence of PTB among T2DM patients was 129.89 (95% CI: 97.55-172.95) per 100,000 person-years. The T2DM-PTB incidence of these studies ranged from 16 Figure S1). Sensitivity analysis showed that the estimates after omitting any study did not change significantly from the pooled estimates of 67 studies retrieved for further detailed assessment Reviews or meta-analysis (1759) Case reports, trials, meeting abstracts or editorials (1867) Excluded after titles/abstracts screen (8059) 14 studies included for incidence meta-analysis 16552 records identified from Web of Science (6200), PubMed (5419), the Chinese National Knowledge Infrastructure (CNKI) database (4600), Cochrane library (333) 11752 potentially relevant reports after duplicates removed Excluded by review of full-text (12) No sufficient data for analysis (15) Full text were not available (7) Not meet the diagnostic criteria (7) Duplicated publication (2) Identification Screening Eligibility Included 10 studies included for prevalence meta-analysis  T2DM-PTB incidence were robust (Supplementary Figure S2).

T2DM-PTB prevalence
Fourteen cohort studies and ten cross-sectional studies containing 2,569,451 T2DM patients reported the prevalence of T2DM-PTB cases. The main characteristics of the eligible studies were extracted and listed in  Figure S3). In addition, sensitivity analysis indicated that the pooled estimate of T2DM-PTB prevalence was stable (Supplementary Figure S4).

Subgroup analysis
The results of subgroup analysis are shown in Table 3.    However, no significant difference in the prevalence of T2DM-PTB was found when stratified according to study design (p ¼ .89).

Discussion
This review quantitatively synthesized the existing evidence based on 24 studies involving a total of 2,569,451 T2DM patients, and revealed that the estimated overall incidence and prevalence of PTB among T2DM patients were 129.89 (95% CI: 97.55-172.95) per 100,000 person-years and 511.19 (95% CI: 375.94-695.09) per 100,000 respectively. The mechanisms that explain the association between T2DM and increased susceptibility to PTB are barely understood. One of the main causes for the aggravation and even death of T2DM patients is their high susceptibility to many infectious diseases including PTB, which may be related to immune dysfunction [40]. In terms of the high incidence and prevalence of T2DM-PTB, preventive interventions among T2DM patients are imperative. It is worth noting that the incidence of PTB in T2DM patients varies significantly in different geographical regions, which was significantly higher in Asia countries than that in Europe countries. This is in accordance with the finding reported in another systematic review that DM patients in Asian continent (2.46, 95% CI: 2.04-3.02) were at higher risk of TB than DM patients in Europe (1.71, 95% CI: 1. 33-2.20). The incidence of T2DM-PTB was significant higher in countries with a high TB burden than in countries with a low TB burden, which may be related to the socioeconomic conditions and access to medical healthcare. It highlights the need to implement prevention interventions, such as T2DM-PTB bidirectional screening, to reduce the dual burden especially in low-income and middle-income countries. The literatures included in this meta-analysis have different criteria for PTB diagnosis. According to the subgroup analysis, the incidence of T2DM-PTB differed between subgroups. Our results showed that the T2DM-PTB incidence was higher in studies by reviewing medical or public health records and using ICD-9 codes. It indicated that different methods may affect the results of PTB diagnosis in T2DM patients. The first study to complete data collection ended in 2005 [39], but the latest study to complete data collection ended in 2018 [16]. In terms of the ending year of data collection, we found the estimated T2DM-PTB incidence was significantly lower for studies whose data collection ended after 2011 than before 2011. It can be speculated that the decline in T2DM-PTB incidence may be is associated with substantial progress made in the areas of TB prevention, diagnosis and treatment. The pooled prevalence of T2DM-PTB was much lower than the prevalence of DM-TB (4.72%, 95% CI: 3.62-5.83%) reported in a previous meta-analysis [41], which might be attributed to different study population and inclusion criteria. Our study found significant higher prevalence of T2DM-PTB in Africa and Asia, which consistent with findings reported by previous meta-analyses [41,42]. It might partly be attributable to the fact that most of the studies included in this meta-analysis were conducted in endemic regions. When evaluated by TB burden of study countries, the prevalence of T2DM-PTB appeared to be higher in countries with a high TB burden compared to countries with a low TB burden. This finding is supported by the fact that the number of new TB cases worldwide is mostly from countries with a high TB burden, such as India, China and Nigeria. In terms of study design, there was no significant difference in the prevalence of T2DM-PTB between cohort studies and cross-sectional studies.
There are several limitations should be acknowledged in this meta-analysis. First of all, there has been epidemiological evidence that age [43], the status of glycemic control [44], cigarette smoking [41], gender and body mass index [37] are factors related to the prevalence of TB among DM patients. However, in most of included studies lack of information about these variables, restricting our ability to identify specific traits in subgroup analyses. Second, the heterogeneity among the included studies was significant, which may be caused by unmeasured characteristics. Thus, the pooled incidence and prevalence need to be interpreted with caution. Third, most of these studies in this meta-analysis were cross-sectional and did not set up a control group, so there was a lack of reasonable comparison. Fourth, our comprehensive search strategy has identified many eligible studies, but the majority of them conducted in Asia. As we known, most of the 30 countries with high-TB burden were from Asia and Africa. In countries with low-TB burden, there is little research in the field of T2DM-PTB comorbidity. To some extent, it might have implications for the precision and representativeness of the estimates.
Despite these limitations, to the best of our knowledge, this is the first meta-analysis that quantitatively summarizes existing evidence about the global incidence and prevalence of PTB in patients with T2DM from a large number of studies. The findings of our study shed light on the impact of T2DM in the development of PTB, and can be further used for performing cost-effectiveness analysis and cost-utility analysis of screening programmes for PTB among T2DM patients.

Conclusions
Taken together, our findings provide a theoretical basis for the necessity of PTB screening in T2DM patients and implementation of effective prevention programs to mitigate the dual burden of PTB-T2DM co-morbidity, particularly in countries with a high TB burden. The mechanism by which T2DM increases the risk of PTB at the individual and population levels are complex and still needs to be further clarified.

Ethical approval
This article does not contain any studies with human participants performed by any of the authors.

Author contributions
The authors' responsibilities were as follows: QW, KL and SHC conceptualized the meta-analysis, QW, YL and YBM conducted the research and performed the statistical analysis; QW drafted the initial manuscript; QW, YL and YBM substantively revised the manuscript and have approved the submitted version.

Disclosure statement
The authors declare that they have no relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript.

Data availability statement
Data sharing is not applicable to this article, as this study is a meta-analysis.