Type 2 diabetes mellitus and recurrent Tuberculosis: A retrospective cohort in Peruvian military workers

Background

The role played by type 2 diabetes mellitus (DM2) in the occurrence of recurrent tuberculosis is still uncertain. Military personnel are an occupational group with an increased risk for tuberculosis exposure due to their activities. Methods. We conducted a retrospective cohort to study the association between DM2 and recurrent TB in military workers who have been previously treated for tuberculosis at the Central Military Hospital in Lima, Peru, between 2016 and 2017. We evaluated the risk between DM2 and recurrent TB using Nelson-Aalen graphical analysis and Cox regression stratified by TB cured with hazard ratio (HR) calculation adjusted for confounders. Results. We evaluated 220 workers with a mean age of 23.2 ± 7.8 years (96.8 % male). DM2 and recurrent TB frequency were 11.8 % and 5.0 %, respectively. Those with DM (36.5 %) presented a greater proportion of recurrent TB than those without DM2 (10.5 %). The cumulative risk for recurrent TB increases faster among workers with DM2 (p = 0.025, LR chi-squared test). Cox regression stratified by type of cured TB did not show an association between DM2 and recurrent TB (HR: 3.67; 95 %CI: 1.00–13.46). Conclusion. The cumulative risk for recurrent TB increases faster in patients with DM than in those without DM2. DM2 is not associated with the time of apparition of recurrent TB in military workers.

1. Introduction

Tuberculosis (TB) infection is known to be influenced by factors such as proximity to an individual with active TB, immunosuppression, malnutrition, and medical conditions such as HIV/SIDA, silicosis, and severe kidney disease, among others [1]. However, recent consistent evidence has shown that poorly controlled type 2 diabetes (DM2) increases the risk of TB disease threefold (RR: 3.11; 95 % CI: 2.27 to 4.26) and leads to unfavorable outcomes in its treatment [2]. Although this is the best-characterized aspect of the association between TB and DM2, these findings vary widely among studies, with hazard ratios ranging from 1.00 to 7.83 [3]. This situation illustrates the complexity of studying DM2 as a risk factor for TB, given the heterogeneity in populations worldwide, including age differences, access to medical care, level of glucose control, types and numbers of complications, and available medications. This aspect is worrisome, considering recent estimates that indicate the global prevalence of DM2 in people aged 20 to 79 years was 10.5 % (536.6 million people) in 2021, with a projected increase to 12.2 % (783.2 million) by 2045 [4].

Despite the established relationship between DM2 and the risk of TB, little research has been conducted on the impact of DM2 on the recurrence of TB (relapse or exogenous reinfection) [5]. Previous studies have reported a significant association between DM2 and TB recurrence due to relapse (OR: 1.96, 95 %CI: 1.22–3.15), especially in men (RR: 1.23, 95 %CI: 1.15–1.32) [6]. On the other hand, it has also been described that the increase in glycemia is related to the increase in the incidence of recurrent TB [7], so the critical factor in evaluating is glycemic control. According to a systematic review, stratified pooled recurrence rates increased from 1.47 (95 %CI 0.87–2.46) to 4.10 (95 %CI 2.67–6.28) per 100 person-years for studies conducted in low vs. high TB incidence settings [8]. The factors associated with recurrent TB are the recurrence time for relapse, HIV coinfection, the strain's genotype, incarceration, and immigration for exogenous reinfection [9]. According to complete genome sequencing analysis, studies carried out in Peru show a TB recurrence of 5.2 % [10], with 34 % attributed to relapses [11].

In this sense, this study aimed to evaluate the association between DM2 and recurrent TB in a cohort of military workers who received TB treatment in 2016 and 2017 at the Military Central Hospital in Lima, Peru. In this population, occupational exposure is a relevant factor associated with the development of TB, contributing to the disease burden partly due to training activities or fieldwork [12]. The military population can be exposed to environments where they can contact the agent that causes the disease or where these environments are conducive to disease development [13], [14]. This issue is noteworthy considering that in Peru, TB ranks among the top 20 causes of mortality [15], reaching peaks of 3.7 deaths per 100,000 inhabitants [15]. Additionally, DM2 has a prevalence of 2.9 % has been reported in the Peruvian population over 15 years of age, rising to 7.0 % in people over 40 [16]. Evidence on recurrent TB in this population is scarce and limited to describing the prevalence of TB [17], [18], [19], [20]. The results could improve the understanding of DM2 as a risk factor for recurrent TB and the importance of strengthening secondary prevention programs.

2. Methods

2.1. Study Area and Participants

A retrospective cohort of military workers treated at the Military Hospital of Lima, Peru (a military medical hospital) during 2016 and 2017 were studied. These patients were enrolled in the TB control and treatment program. This hospital is classified as a Category III-1 medical establishment (high complexity hospital), and it is the largest hospital in Peru that provides care to personnel linked to the Peruvian army.

The study population consisted solely of workers with military activity who had a history of a confirmed diagnosis of pulmonary and extrapulmonary TB and were cured. The initial diagnosis used acid-fast staining or mycobacterial culture of fluid and tissue samples, chest x-ray, or, when available, nucleic acid amplification testing (NAAT). Workers with activities other than military, those who died during the follow-up process, and those without glycemia and glycosylated hemoglobin results were excluded from the analysis.

The follow-up period lasted approximately between three and nine months. All military workers had sputum smear-negative results at baseline and a negative GeneXpert® test at the end of treatment within the past three months. Epidemiological surveillance records of military workers were evaluated, identifying the presence of type 2 Diabetes Mellitus (DM2) (exposure variable) at the cohort's admission. TB recurrence (outcome) was identified as the bacteriologically positive result both with smear and GeneXpert® during the follow-up period (see supplementary table 1). We show the flowchart for enrollment in Fig. 1.

Fig. 1.

Flowchart summarizing the process of enrollment and follow-up.

2.2. Techniques and procedures

Medical records of military workers from 2016 and 2017 who belonged to the tuberculosis control and treatment program at the Military Hospital of Lima were reviewed. A data collection form was used to obtain information such as age, sex, follow-up time (expressed in days), location of the cured TB (pulmonary or extrapulmonary), bacilloscopic results, GeneXpert® rifampicin susceptibility test, and presence of DM2 and recurrent TB. A double typing of the records to measure the quality control of the information was performed.

The diagnosis of DM2 was according to the criteria established by the American Diabetes Association, which defines the presence of hyperglycemia (glycemia ≥ 126 mg/dL in fasting conditions for 12 h prior to taking the blood sample) and elevated glycosylated hemoglobin (greater than 6.5 %) [21]. Likewise, a normal value of glycemia up to 110 mg/dL was used.

The diagnosis of recurrent TB was defined according to the recommendations of the Ministry of Health of Peru [22], which establishes the presence of smear microscopy of sputum and other biological fluids as positive (presence of acid-fast bacilli to Ziehl Nielsen staining and under microscopy at 100X), and scored in crosses (negative, paucibacillary: 1–9 BARR in 100 observed fields, 1+: 1 BAAR per field in 100 observed, 2+: 1–10 BARR per field in 100 observed or 10––99 AFB in 100 fields and 3+: >10 ABF per field in 20 observed) and positive result by molecular test. GeneXpert® (Cepheid Sunnyvale, CA, United States) was used to identify TB infection and resistance to rifampicin. Data on adenosine deaminase activity in pleural fluid from some military workers with a history of extrapulmonary TB was also obtained (see supplementary table 1). Visual light spectrophotometry (Wiener Lab, Argentina) measured ADA activity with a kinetic test (Spin React, Spain), and up to 40 U/L was considered a normal value.

2.3. Statistical analysis

A post-estimation of power in a Cox regression model of logarithmic risks [23] was performed with a binary response variable (Y: TB recurrence) in an independent variable (X: the presence of DM2) with a standard deviation of 1.12 and a sample size of 220 observations. This generated a power of 86.0 % at a significance level of 0.05, with the ability to detect a regression coefficient of 1.4 (equivalent to a Hazard ratio of 4.0). The calculation was fitted in a multiple Cox regression model, with an expected R2 of 0.8. Finally, the sample size was adjusted for an event proportion of 0.059 (approximate proportion of people with recurrence of TB and DM2 after a cured TB, according to data reported by Alarcón et al. [15]). The calculation was performed with the statistical module for estimation of sample size PASS (NCSS, LLC. Kaysville, Utah, USA) version 11 [24].

The general characteristics of the participants were presented descriptively. Missing values were excluded from the analysis. After evaluating the expected frequencies, the bivariate analysis used the chi-square test (Pearson, Yates, or Fisher). A p < 0.05 was considered to determine statistical significance. The accumulated risks were estimated using the Nelson-Aalen graph and representing the 95 % confidence interval (95 %CI). The Log Rank test evaluated the assumption of proportionality of risks in workers with and without DM2. We used the Cox regression stratified by cured TB in its bivariate (dependent and main independent variable) and multivariate (for adjustment through selected confounding variables according to epidemiological criteria: age and rifampicin resistance) to estimate the measure of association (Hazard Ratio: HR). A p-value less than 0.05 was considered for the statistical tests. All the calculations were conducted in the Stata Corp. LLC (TX, USA) version 17 program.

2.4. Ethical aspects

The Hospital Military Central Ethics Committee approved the research on October 2, 2018 (Official Letter No. 1034/AA-11/8/HMC/DADCI). Considering that the information from the records generated in the TB control and treatment program of the Military Hospital was used between 2016 and 2017, informed consent was not required. The data was used strictly to comply with the objectives set out in the investigation, the anonymity of the records was maintained, and the personal data of each military worker was not accessed.

3. Results

Table 1 shows the characteristics of the study population, where 96.8 % of the records came from male workers, and the mean age was 23.2 ± 7.8 years. Glycemia had a mean of 87.9 ± 17.1 mg/dL, and the proportion of DM2 was 11.8 % (95 %CI: 7.9–16.9 %). ADA activity in pleural fluid had a median of 59.5 U/L, and 96.4 % had levels above the normal range. The proportion of resistance to rifampicin was 8.7 % (95 %CI: 5.3–13.2 %), and recurrent TB was 5.0 % (95 %CI: 2.5–8.8 %). The follow-up period had a mean of 188.6 ± 45.5 days.

Table 1.

Descriptive characteristics of the study population.

Characteristic	N	%
Sex
Men	213	96.8
Women	7	3.2
Age (years) a	21.0 (19.0–26.0)
Type of cured TB
Pulmonary	125	56.8
Extrapulmonary	95	43.2
Adenosin deaminase activity (U/L) a,b	59.5 (53.6–84.9)
Rifampicin resistance c
No	200	91.3
Yes	19	8.7
DM2
No	194	88.2
Yes	26	11.8
Recurrent TB
No	209	95.0
Yes	11	5.0

^aMedian (interquartile range); ^b Only patients with extrapulmonary TB; ^c one missing reported (male, 18 years old).

Fig. 2 shows the survival of recurring TB. Within days 144 and 156, the first case of recurring TB was recorded (probability of survival: 0.995, 95 %CI: 0.966–0.999) and within days 408 and 420, a probability of survival of 0.744 (95 %CI: 0.456–0.895) was reached.

Fig. 2.

Kaplan-Meier survival estimate for recurrent TB.

Table 2 shows recurrent TB according to the study variables. There was a difference in the proportion of recurrent TB between patients with and without DM2, being that those with DM2 (36.5 %) presented a greater proportion of recurrent TB compared to those without DM2 (10.5 %).

Table 2.

Recurrent TB according to study variables.

Characteristic	Recurrent TB, n (%)		p-value
	No	Yes
Sex			0.792 a
Men	202 (96.7)	11 (100.0)
Women	7 (3.3)	0 (0.0)
Age, median (IQR)	21 (19–26)	21 (20–39)	0.302b
Type of cured TB			0.444c
Pulmonary	118 (56.5)	7 (63.6)
Extrapulmonary	91 (43.5)	4 (36.4)
Rifampicin resistance d			0.059c
No	192 (92.3)	8 (72.7)
Yes	16 (7.7)	3 (27.3)
DM2			0.029c
No	187 (89.5)	7 (63.6)
Yes	22 (10.5)	4 (36.4)

^aYates' adjusted chi-squared, ^b Mann-Whitney non-parametric test, ^c Fisher's exact test, ^d One missing reported (male, 18 years old).

Fig. 3 shows the accumulated risks (estimated using the Nelson-Aalen estimator) among military workers with and without DM2. It is observed that from the 144th day, the probability of occurrence of recurring TB increases in workers with and without DM2; However, the accumulated risk increases faster among workers with DM2 (p = 0.025, LR Chi-Squared Test).

Fig. 3.

Accumulated risks of recurring tuberculosis in military workers according to diabetes mellitus presence.

A Cox regression stratified by type of cured TB was carried out, since the proportionality of the hazards was not met. There was no significant association between DM2 and recurring TB (HR: 3.7, p = 0.05, 95 %CI: 1.00–13.46).

4. Discussion

This retrospective study evaluated the relationship between DM2 and the occurrence of recurrent TB in military workers who previously had TB and were successfully treated during 2016 and 2017. These populations were followed for a subsequent period of three to nine months to determine TB recurrence, considering collecting data related to DM2 diagnosis. We must consider that Peru is in second place with the highest burden of TB [25] and is one of the 30 countries with the highest burden of Multidrug-Resistant TB (MDR-TB) in the world [26]. In this context, the military population is exposed to training activities or fieldwork in endemic areas. Furthermore, TB and DM2 are known to have a high incidence in low- and middle-income countries such as Peru; in fact, between 95 % and 75 % of all global cases are found in these countries [1]. Additionally, although the relationship between DM2 and the risk of TB is already known up to three times in the general population [3], [27], no recurrent TB data has been reported in this type of occupational population.

Our main findings revealed a higher prevalence of recurrent TB in military personnel with DM2 compared to those without DM2. Additionally, we observed that the cumulative risk for recurrent TB increases more rapidly in patients with DM2 compared to those without DM2. This suggests that military personnel diagnosed with DM2 may be predisposed to developing recurrent TB. The HR found in this study was 3.7, a value close to those reported in studies evaluating the association between DM2 and TB risk in the general population rather than occupationally or with recurrent TB as an outcome. Given the high prevalence of TB and DM2 in Peru and other countries with limited to moderate resources, it is plausible that military personnel with DM2 face a greater susceptibility to developing recurrent TB due to their exposure to fieldwork. However, this relationship is not statistically significant.

Through a cross-sectional study of 2011–2012 National Health and Nutrition Examination Survey data, Barron et al. reported an association between DM2 and latent TB with an adjusted odds ratio (OR) of 1.90 (95 % CI: 1.15–3.14) [28]. In a study, Ko et al. found a high prevalence of DM2 among patients with newly diagnosed TB in a Taiwanese population [29]. In a case-control study, Pereira et al reported an association between DM2 and TB with an OR of 2.37 (95 % CI: 1.04–5.42) [30]. This study reflects the consistency in the relationship between DM2 and the risk of TB. Although the association between DM2 and the time to the appearance of recurrent TB in military workers was not statistically significant, the high prevalence of TB and DM2 in Peru and countries with low to medium resources [31], could predispose military personnel with DM2 to develop recurrent TB due to exposure in fieldwork.

One limitation was the duration of follow-up, three to nine months, which could underestimate the recurrence rate. The possibility of selection bias was minimized by confirming the results of the workers through bacilloscopic tests and GeneXpert® tests. However, the lack of information regarding DM2 status, including the presence of complications or the DM2 treatment received by the patients (such as the use of insulin, sulfonylurea, metformin, and others), diminishes the possibility of better characterizing the study population according to DM2 characteristics. Additionally, information on comorbidities was not available. Furthermore, there was a lack of information regarding TB resistance profiles, the current type of TB (pulmonary or extrapulmonary), the organ affected in extrapulmonary TB, the time of DM2 diagnosis, and whether it is the second or third episode of recurring TB. Also, no data were available to determine if the DM2 diagnosis preceded TB initiation. Despite these limitations, the data used are valuable for an initial study of the relationship between DM2 and recurrent TB. The study results must be interpreted with caution due to the described limitations.

It is relevant to evaluate the presence of DM2 and other comorbidities that cause an increased disease burden in this occupational population, as comorbidities could increase the probability of other health problems, including infections like TB. Despite the importance of TB diagnosis in Peru, the strategies used to evaluate TB impact on the population's health status do not emphasize the convergence between DM2 and TB in military personnel [32]. Hence, integrated actions should be considered to face the double burden of the disease and its consequences on the population's health, including military personnel and the healthcare system. Because militaries with a previous diagnosis of TB can be exposed to a new TB infection as part of military training or fieldwork [13], [14], the study findings could be used to promote programs focused on screening recurrent TB in military workers.

The increasing prevalence of DM2 poses a potential threat to TB control. This rise is attributed to population, aging, urbanization, dietary changes, and reduced physical activity patterns, all leading to increased obesity [33]. Given the evidence indicating that DM2 is a risk factor for the onset of TB and raises the possibility of reactivating latent cases [28], along with the observation that patients with DM2 and TB may experience delayed smear negativity after the initiation of treatment, it becomes essential to consider the heightened likelihood of complications and an increased risk of recurrent TB, especially in regions with a high prevalence of DM2. This association may become more evident, given the escalating trend of DM2 [34]. The concurrent occurrence of DM2 and TB poses a challenge for the control of both diseases and could represent a significant financial burden for public health.

In this study, the possibility of selection bias was minimized, considering that bacilloscopic tests and GeneXpert® tests confirmed the results of the workers. However, the lack of information regarding DM status, including complications or DM2 treatment received by the patients (use of insulin, sulfonylurea, metformin, and others), reduces the possibility of better characterizing the study population according to DM2 characteristics. Additionally, information on comorbidities and clinical manifestations of TB was not available. Furthermore, there was a lack of information regarding TB resistance profiles, the current type of TB (pulmonary or extrapulmonary), the organ affected in extrapulmonary TB, DM2 time of diagnosis, and whether it is the second or third episode of recurring TB. Also, no data was available to determine if the DM2 diagnosis was a previous TB initiation. Despite these limitations, the data used are helpful for an initial study of the relationship between DM2 and recurrent TB. The study results must be interpreted with caution due to the limitations described.

5. Conclusion

Recurrent TB cases are present in military workers with and without DM2. While the proportion of recurrent TB cases was greater in patients with DM2 compared to those without DM2, and the cumulative risk for recurrent TB increases faster in patients with DM2 compared to those without DM2, there was no association between DM2 and the time to the apparition of recurrent TB in military workers. This study is a first approximation of the association between DM2 and recurrent TB in military personnel in a country with a high burden of these diseases. Future studies can be developed using methodologies that consider data that allow a better characterization of the phenomenon studied.

ETHICAL STATEMENT

Institutional Review Board Statement: The research was approved by the Hospital Military Central Ethics Committee on October 2, 2018 (Official Letter No. 1034/AA-11/8/HMC/DADCI).

Informed Consent Statement: not required.

Declaration of competing interest: The authors declare no conflict of interest.

Author Contributions: Conceptualization, J.R.-R. and N.A.-V.; Formal analysis, J.R.-R., N.A.-V and G.B.-Q.; Supervision, G.B.-Q.; Writing—original draft, J.R.-R. and N.A.-V., J.A.-F. and J.I.; Writing—review & editing, J.R.-R. and N.A.-V., J.A.-F., F.S.-L. and J.I. All authors have read and agreed to the published version of the manuscript.

CRediT authorship contribution statement

Nataly T. Alvarado-Valdivia: Writing – original draft, Project administration, Investigation, Conceptualization. Juan A. Flores: Writing – review & editing, Visualization, Methodology, Investigation. Jorge L. Inolopú: Writing – original draft, Visualization, Methodology, Investigation. Jaime A. Rosales-Rimache: Writing – review & editing, Validation, Supervision, Formal analysis, Conceptualization.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Appendix A. Supplementary data

The following are the Supplementary data to this article:

Data Availability Statement:

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy considerations.