Abstract

Background:

To the best of our knowledge, imaging features of primary type 2 diabetes mellitus patients with rifampicin-resistant/multidrug-resistant tuberculosis (T2DM RR/MDR-TB) have not been reported in the literature to date. Hence this study investigated the imaging features of primary T2DM RR/MDR-TB patients.

Methods:

The clinical data and pulmonary CT findings of 87 primary T2DM RR/MDR-TB patients, 98 primary non-type 2 diabetes mellitus patients with rifampicin-resistant/multidrug-resistant tuberculosis (NT2DM RR/MDR-TB), 86 primary type 2 diabetes patients with drug-susceptible tuberculosis (T2DM DS-TB), and 93 primary pure drug-susceptible TB (DS-TB) patients without T2DM or RR/MDR-TB who were treated at Dalian Public Health Clinical Center from 2018 to 2023 were collected, and the clinical features and imaging differences among the four groups were compared using the chi-square test for categorical variables and the Kruskal-Wallis H test for ordinal or non-normally distributed continuous variables.

Results:

According to the results, among the four groups of patients, the NT2DM RR/MDR-TB patients were the youngest (p = 6.27e-10). Compared with non-diabetic patients, both T2DM RR/MDR-TB patients and T2DM DS-TB patients had a greater proportion of males (p = 1.28e-8) and a greater body mass index (BMI) (p = 6.01e-9) but lower albumin levels (p = 0.009). T2DM RR/MDR-TB patients presented the highest rates of smoking (p = 0.006) and alcohol abuse (p = 0.000281). In terms of imaging features, multinomial logistic regression analysis revealed that large nodules (p = 0.000239), patchy opacities (p = 0.004), non-calcified enlargement of mediastinal lymph nodes (p = 0.005), and especially multiple cavities (≥3 cavities) (p = 2.21e-8) were high-risk factors for T2DM in RR/MDR-TB patients (p < 0.05). As the glycated hemoglobin (HbA1c) levels increased, the incidence of large nodules (p = 0.029), multiple cavities (p = 0.000030), pleural effusion (p = 0.045), and non-calcified enlargement of mediastinal lymph nodes (p = 0.001) also increased in the T2DM RR/MDR-TB patients; conversely, the absence of cavities (“0” cavities) (p = 0.000009) decreased with increasing HbA1c levels.

Conclusion:

Imaging features, such as large solid nodules, patchy opacities, and particularly multiple cavities, are high-risk factors for T2DM in RR/MDR-TB patients. The presence of multiple cavities and extrapulmonary involvement increases with increasing blood glucose levels. Therefore, controlling blood glucose levels may reduce the risk and severity of T2DM RR/MDR-TB.

1 Introduction

According to the Global Tuberculosis Report 2025, in 2024, 10.7 million new tuberculosis cases and 1.23 million deaths were reported globally, positioning tuberculosis as the main cause of death due to a single infectious disease worldwide (1).

Given that RR-TB and MDR-TB (TB strains resistant to at least isoniazid and rifampin) patients have similar prognoses, the WHO has included a new drug classification to manage RR and RR/MDR (2, 3). According to the WHO, 400,000 cases (3.7%) of RR/MDR were reported worldwide in 2023 (1). Compared with patients with pure tuberculosis, patients with diabetes and RR/MDR-TB have a greater risk of treatment failure and death (4, 5). Computed Tomography (CT) is one of the diagnostic and therapeutic evaluation methods used for RR/MDR-TB and plays a crucial role in clinical practice. However, while many imaging studies have been conducted on RR/MDR-TB patients (6–12), no research on the imaging features of primary T2DM RR/MDR-TB patients has been reported. Previous studies, including the studies by our research group, have used either a mixture of NT2DM RR/MDR-TB and T2DM RR/MDR-TB patients as subjects, have not clearly specified the subjects, or have been conducted with a mix of primary and re-treated patients without further subgroup analysis, which might have influenced the conclusions of these studies.

This study retrospectively analyzed the clinical and imaging features of 87 primary T2DM RR/MDR-TB patients between 2018 and 2023. The resulting data are expected to serve as an important basis for imaging research on T2DM RR/MDR-TB patients.

2 Materials and methods2.1 Ethics statement

This study was reviewed and approved by the ethics committee of Dalian Public Health Clinical Center. All the data, including the records/information of the patients, were de-identified prior to data analysis and analyzed anonymously. Moreover, this study is a retrospective, non-interventional study for which additional informed consent was waived by the ethics committee.

2.2 Case definitions

T2DM RR/MDR-TB group: Patients with type 2 diabetes mellitus (T2DM) diagnosed according to the American Diabetes Association criteria reference, i.e., HbA1c ≥ 6.5% (48 mmol/mol), or fasting plasma glucose ≥7.0 mmol/L, or 2-h plasma glucose ≥11.1 mmol/L during an oral glucose tolerance test, or current use of antidiabetic medication. They also had rifampicin-resistant/multidrug-resistant tuberculosis (RR/MDR-TB), defined as Mycobacterium tuberculosis resistant to at least rifampicin (and isoniazid for MDR) by phenotypic or molecular drug susceptibility testing (DST).

NT2DM RR/MDR-TB group: Patients without type 2 diabetes mellitus, defined as HbA1c < 5.7% (39 mmol/mol) and fasting plasma glucose <5.6 mmol/L, with no prior diagnosis of diabetes and no antidiabetic treatment. They had RR/MDR-TB using the same DST criteria as group 1.

T2DM DS-TB group: Patients with T2DM (same definition as group 1) and drug-susceptible tuberculosis (DS-TB), i.e., M. tuberculosis susceptible to rifampicin and isoniazid (and to all first-line drugs) by DST.

DS-TB without T2DM or RR/MDR-TB group: Patients with DS-TB (same definition as group 3) and without T2DM (same definition as group 2), and without RR/MDR-TB.

2.3 Clinical information

The research was conducted at the Dalian Public Health Clinical Center.

The inclusion criteria for patients were as follows: complete clinical data, including demographic information, drug susceptibility testing (DST) results, and CT findings.

Patients with resistance patterns other than RR/MDR-TB and those with comorbid conditions such as chronic obstructive pulmonary disease (COPD), pneumonia, lung cancer, immune system disorders, AIDS, or other underlying systemic diseases were excluded from this study.

During the period between January 2018 and December 2023, a total of 4,229(100%) patients were diagnosed with pulmonary tuberculosis, which was bacteriologically confirmed through smear, culture, or GeneXpert positivity. Of these, 359 (8.49%) had RR/MDR-TB, and after excluding 156(43.45%) re-treatment cases, 203 (56.55%) remained. Another 3,870 (91.51%) were non-RR/MDR-TB; after excluding 909(23.49%) re-treatment cases, 525(13.57%) without DST results, and 303(7.83%) with other resistance, 2,133(55.12%) DS-TB patients remained.

Among the remaining 203 patients, 91(44.83%) patients with primary type 2 diabetes RR/MDR-TB and 112 (55.17%) patients were included in the study in chronological order. From the 2,133 DS-TB patients, 101(4.74%) with type 2 diabetes and 100 (4.69%) without were enrolled chronologically. Among these patients, after applying the inclusion and exclusion criteria, the final study cohort comprised 87(95.60%) T2DM RR/MDR-TB, 98(87.5%) NT2DM RR/MDR-TB, 86(85.15%) T2DM DS-TB, and 93(93.0%) pure DS-TB patients. All were HIV-negative. All enrolled patients were HIV-negative (Figure 1).

Screening of patients. RR, Rifampicin resistant; MDR-TB, Multidrug resistant tuberculosis; T2DM, Type 2 diabetes; DS, Drug susceptible; DST, Drug susceptibility testing.

2.4 Research methods2.4.1 General information

The clinical data of the included patients, including age, gender, smoking history, alcohol abuse history, glycated hemoglobin concentration (HbA1C) at the first diagnosis of TB or RR/MDR-TB, the time intervals between the first diagnoses of DM and TB, body mass index (BMI), and albumin levels, were collected.

2.4.2 CT image acquisition

All the subjects in this study underwent high-resolution computed tomography (HRCT) scans when they were diagnosed. The images were examined and recorded by two radiologists and two clinical experts, and decisions on the CT findings were reached by consensus.

Imaging assessments were performed for the following signs: (1) small nodules < 1 cm in diameter, including trees in buds, branch line nodes, nodules with poorly defined margins, cluster nodules, reticulonodular opacity (combined reticular and nodular opacity), and miliary nodules; (2) large nodules with diameters between 1 and 3 cm; (3) a mass, any circumscribed, mostly solid lesion that measured >30 mm in diameter; (4) patchy opacities, including patchy shadows, irregular flaky infiltration, and ground glass opacities (GGOs); (5) airway damage, including thickening of the airway wall, stenosis or expansion of the lumen, and thickening of the interstitium around the tracheal vascular bundle; (6) the number of cavities, divided into categories of zero cavities, one to two cavities, and three or more cavities; (7) consolidations without cavities, including focal consolidation, segment consolidation, and lobe consolidation; (8) pleural involvement, including pleural thickening or pleural effusion; (9) pericardial involvement, including pericardial thickening or pericardial effusion; (10) enlargement of the mediastinal lymph nodes, including calcified/non-calcified lymph nodes; (11) local calcified lesions; (12) bronchiectasis.

2.4.3 Glycemic control levels

The patients with T2DM RR/MDR-TB were divided into the following three groups based on their HbA1c levels at the initial diagnosis of tuberculosis: HbA1c < 7% (group H1), HbA1c 7–9% (group H2), and HbA1c > 9% (group H3). Pure DS patients were designated the baseline group H0.

2.5 Statistical methods

Statistical analyses were performed using SPSS version 26.0. Continuous variables with a normal distribution are expressed as the means ± standard deviations, whereas non-normally distributed continuous variables are presented as medians (P25, P75). Categorical variables are summarized as frequencies and percentages. Intergroup comparisons were conducted using independent t tests for normally distributed continuous variables and Mann–Whitney U tests for non-normally distributed variables. Categorical variables were compared via chi-square tests.

Multinomial logistic regression analysis was performed to identify the high-risk factors associated with pulmonary CT manifestations in T2DM RR/MDR-TB patients. Forest plots were generated using GraphPad Prism version 10.1.2 software to visualize the regression results. The correlation between the CT features and glycemic control levels was determined using Spearman’s rank correlation analysis, and correlation heatmaps were constructed using R Studio version 4.3.3.

3 Results3.1 Characteristics of the clinical data

Among the four patient groups, those with NT2DM RR/MDR-TB were the youngest, with a median age of 35.50 (26.00, 55.50) years, which was significantly lower than that of the diabetic groups (p < 0.05). Compared with non-diabetic patients, diabetic patients were more likely to be male and have a higher BMI but lower albumin levels. Additionally, T2DM RR/MDR-TB patients presented higher rates of smoking and alcohol consumption (Table 1).

Patient characteristicsT2DM RR/MDR-TB
(G1 = 87)NT2DM RR/MDR-TB
(G2 = 98)T2DM DS-TB
(G3 = 86)DS-TB
(G4 = 93)p-valueAge53.71 ± 11.3735(26.00, 57.25)56.98 ± 13.9749.31 ± 18.31<0.001BGender (Male)77(88.5%)54(55.1%)71(82.6%)52(55.9%)<0.001CTime intervalA5(0.17, 10.00)–5(0.58, 10.00)–0.537Smokers vs. nonsmokers33(37.9%)18(18.4%)24(27.9%)17(18.3%)0.006DAlcohol abuse26(29.9%)10(10.2%)14(16.3%)8(8.6%)<0.001EHbA1C (%)9.43 ± 1.89–9.80 ± 1.99–0.210BMI22.54 ± 3.6619.79 ± 2.8822.55 ± 3.3720.7(18.59, 23.01)<0.001FAlbumin (g/L)36.34 ± 6.0037.32 ± 5.4936.56 ± 5.7539.9(34.4, 43.05)0.009G

Comparisons of the clinical data of the four groups of patients.

G1: Group 1, T2DM RR/MDR-TB; G2: Group 2, NT2DM RR/MDR-TB; G3: Group 3, T2DM DS-TB; G4: Group 4, DS-TB. ATime interval between the first diagnoses of DM and TB (years). BG2 < G1, G3, G4, G1 = G3, G1 = G4, G3 > G4. CG1 = G3 > G2 = G4. DG 1 > G2 = G4, G1 = G3, G2 = G3 = G4. EG1 > G2 = G3 = G4. FG1 = G3 > G2 = G4. GG1, G2, G3 < G4, G1 = G2 = G3, G2 = G4.

3.2 Distribution of the affected lung lobes

In terms of the distribution of lung lesions, the distributions of lesions in the right middle lobe (p = 0.007), left upper lobe (p = 0.021), left lingular lobe (p = 0.001), and left lower lobe (p = 0.014) differed significantly among the four groups of patients. Compared with DS-TB patients, patients with T2DM RR/MDR-TB presented significantly greater involvement of the right middle lobe (p = 0.012), left upper lobe (p = 0.012), and left lingular lobe (p = 0.001104), along with a greater number of affected lobes, although no significant differences were observed compared with the other patient groups (p > 0.05) (Table 2).

Lung lobeT2DM RR/MDR-TB (G1 = 87)NT2DM RR/MDR-TB (G2 = 98)T2DM DS-TB (G3 = 86)DS-TB (G4 = 93)p-valueRUL71(81.6%)77(78.6%)67(77.9%)77(82.8%)0.813RML56(64.4%)42(42.9%)42(48.8%)38(40.9%)0.007ARLL60(69.0%)68(69.4%)55(64.0%)53(57.0%)0.253LUL70(80.5%)69(70.4%)60(69.8%)55(59.1%)0.021BLingula57(65.5%)55(56.1%)50(58.1%)35(37.6%)0.001CLLL56(64.4%)64(65.3%)57(66.3%)43(46.2%)0.014DTotal370(70.9%)375(63.8%)331(64.1%)301(53.9%)<0.001E

Distribution and comparison of the affected lung lobes among the four patient groups.

RUL, right upper lobe; RML, right middle lobe; RLL, right lower lobe; LUL, left upper lobe; LLL, left lower lobe. AG1 > G2 = G4, G1 = G3, G2 = G3 = G4. BG1 > G4, G1 = G2 = G 3, G2 = G3 = G4. CG1 = G3 > G4, G1 = G2, G2 = G3, G2 = G4. DG1 = G2 = G3, G1 = G4, G2 = G3 > G4. EG1 = G2 = G3 > G4.

3.3 Comparison of the imaging features of the T2DM RR/MDR-TB, NT2DM RR/MDR-TB, T2DM DS-TB, and pure DS-TB cases

Compared with the other three groups, T2DM RR/MDR-TB patients (p = 1.17e-10) presented an increased prevalence of cavities, with greater numbers, although no significant difference was observed compared with the T2DM DS-TB patients (p = 0.055). The incidences of large nodules (p = 0.001038) and patchy opacities (p = 0.024) were significantly greater in T2DM RR/MDR-TB patients than in pure DS-TB patients, although the difference was not significant compared with those in the other two groups (p > 0.05) (Table 3).

Pulmonary imaging featuresT2DM RR/MDR-TB (G1 = 87)NT2DM RR/MDR-TB
(G2 = 98)T2DM DS-TB (G3 = 86)DS –TB
(G4 = 93)p-valueNodulesSmall nodules87(100%)98(100%)83(96.5%)91(97.8%)0.084Large nodules69(79.3%)61(62.2%)60(69.8%)49(52.7%)0.002AMass7(8.0%)4(4.1%)3(3.5%)8(8.6%)0.342Patchy opacities56(64.4%)58(59.2%)50(58.1%)40(43.0%)0.024BAirway damage44(50.6%)55(56.1%)36(41.9%)40(43.0%)0.168Cavity0 cavities11(12.6%)40(40.8%)16(18.6%)48(51.6%)1–2 cavities22(25.3%)25(25.5%)32(37.2%)27(29.0%)≥3 cavities54(62.1%)33(33.7%)38(44.2%)18(19.4%)<0.001CConsolidationFocal consolidation35(40.2%)46(46.9%)39(45.3%)40(43.0%)0.814Pulmonary segment consolidation0 segments53(60.9%)66(67.3%)53(61.6%)63(67.7%)1 segments25(28.7%)23(23.5%)25(29.1%)20(21.5%)2 segments4(4.6%)7(7.1%)7(8.1%)8(8.6%)≥3 segments5(5.7%)2(2.0%)1(1.2%)2(2.2%)0.731Pulmonary lobe consolidation0 lobes79(90.8%)90(91.8%)83(96.5%)86(92.5%)1 lobes6(6.9%)6(6.1%)3(3.5%)7(7.5%)2 lobes2(2.3%)1(1.0%)0(0%)0(0%)≥3 lobes0(0%)1(1.0%)0(0%)0(0%)0.470Pleural involvementPleural effusion25(28.7%)16(16.3%)16(18.6%)17(18.3%)0.160Pleural thickening24(27.6%)20(20.4%)14(16.3%)14(15.1%)0.149Pericardial involvementPericardial effusion2(2.3%)1(1.0%)0(0%)2(2.2%)0.632Pericardial thickening1(1.1%)3(3.1%)1(1.2%)1(1.1%)0.733Enlargement of mediastinal lymph nodesEnlargement of mediastinal lymph nodes (non-calcified)30(34.5%)15(15.3%)15(17.4%)15(16.1%)0.004DEnlargement of mediastinal lymph nodes (calcified)7(8.0%)9(9.2%)16(18.6%)21(22.6%)0.011ELocal calcified lesions6(6.9%)19(19.4%)12(14.0%)30(32.3%)<0.001FBronchiectasis14(16.1%)32(32.7%)17(19.8%)18(19.4%)0.032

Comparison of the pulmonary imaging features among the four patient groups.

AG1 > G4, G1 = G2 = G3, G2 = G3 = G4; BG1 > G4, G1 = G2 = G3, G2 = G3 = G4; CG1 > G2, G1 = G3, G1 > G4, G2 = G3, G2 = G4, G3 > G4; DG1 > G2, G1 > G4, G1 = G3, G2 = G3 = G4; EG 1 < G4, G1 = G2 = G3, G2 = G3 = G4; FG1 = G3 < G4, G1 = G2, G2 = G3, G2 = G4.

With respect to calcified lesions, both calcified mediastinal lymph nodes (p = 0.042) and pulmonary calcifications (p = 0.000126) were less common in the T2DM RR/MDR-TB patients than in the pure DS-TB patients, although no significant differences were found compared with the other groups (p > 0.05). In contrast, non-calcified mediastinal lymphadenopathy was more common in T2DM RR/MDR-TB patients than in non-diabetic TB patients (p = 0.004) (Table 3).

Additionally, the statistical analysis conducted using the chi-square test revealed a significant difference in the distribution of bronchiectasis signs among the four groups of patients (p = 0.032). However, in the post hoc pairwise comparisons with Bonferroni correction, none of the comparisons reached the conventional level of statistical significance, which was attributed to the conservative nature of multiple comparison adjustment. No significant differences among the four groups were noted in the remaining imaging features (p > 0.05) (Table 3).

3.4 Multinomial logistic regression analysis of the imaging features in T2DM RR/MDR-TB patients

Before performing multinomial logistic regression analysis, a collinearity diagnosis was conducted on 25 factors: small nodules, large nodules, mass, patchy opacities, airway damage, 0 cavities, 1–2 cavities, ≥3 cavities, focal consolidation, 0 segmental consolidation, 1 segmental consolidation, 2 segmental consolidation, ≥3 segmental consolidation, 0 lobar consolidation, 1 lobar consolidation, 2 lobar consolidation, ≥3 lobar consolidation, pleural effusion, pleural thickening, pericardial effusion, pericardial thickening, Enlarged mediastinal lymph nodes (noncalcified), Enlarged of mediastinal lymph nodes (calcified), local calcified lesions, and bronchiectasis. Tolerance values ranged from 0.629 to 0.969, and all variances inflation factors were below 10, indicating the absence of multicollinearity among the variables. Compared with pure DS-TB patients, T2DM RR/MDR-TB patients presented significantly greater frequencies of large nodules, patchy opacities, non-calcified mediastinal lymphadenopathy, and especially multiple cavities (≥3 cavities) [p < 0.05; OR = 6.8 (1/0.147)]. In contrast, the absence of cavities (“0 cavities”) (p = 1.92e-7), calcified mediastinal lymphadenopathy (p = 0.010), and pulmonary calcifications (p = 0.000098) were less common in the T2DM RR/MDR-TB patients than in the pure DS-TB controls (Figure 2 and Table 4).

Multinomial logistic regression analysis of the imaging features in T2DM RR/MDR-TB patients.

VariableComparison groupβ(SE)OR(95% CI)Wald X2p-valueLarge nodulesThe model is significantT2DM RR/MDR-TB−1.2360.291(0.150–0.562)13.500<0.001NT2DM RR/MDR-TB−0.3920.675(0.379–1.202)1.7780.182T2DM DS-TB−0.7290.483(0.261–0.892)5.4030.020Patchy opacitiesThe model is significantT2DM RR/MDR-TB−0.8730.418(0.229–0.762)8.1050.004NT2DM RR/MDR-TB−0.6530.520(0.293–0.925)4.9520.026T2DM DS-TB−0.6100.543(0.300–0.984)4.0590.0440 cavitiesThe model is significantT2DM RR/MDR-TB1.9977.370(3.475–15.629)27.117<0.001NT2DM RR/MDR-TB0.4361.547(0.873–2.742)2.2300.135T2DM DS-TB1.5404.667(2.368–9.198)19.801<0.001≥3 cavitiesThe model is significantT2DM RR/MDR-TB−1.9200.147(0.075–0.287)31.304<0.001NT2DM RR/MDR-TB−0.7490.473(0.243–0.918)4.8990.027T2DM DS-TB−1.1940.303(0.155–0.591)12.276<0.001Enlarged of mediastinal lymph nodes (non-calcified)The model is significantT2DM RR/MDR-TB−1.0070.365(0.180–0.741)7.7760.005NT2DM RR/MDR-TB0.0621.064(0.488–2.320)0.0240.876T2DM DS-TB−0.0940.910(0.415–1.995)0.0550.814Enlarged of mediastinal lymph nodes (calcified)The model is significantT2DM RR/MDR-TB1.2043.333(1.338–8.304)6.6840.010NT2DM RR/MDR-TB1.0592.884(1.245–6.684)6.1030.013T2DM DS-TB0.2441.276(0.616–2.645)0.4300.512Local calcified lesionsThe model is significant