Demographic and behavioral characteristics of study participants

A total of 14,119 DM patients aged 15 years or older were screened for TB. More than half of them (54.3%) were females, and type II DM accounted for 89.0% of the cases. Of all TB-screened DM patients, 652 (4.62%) had signs and symptoms suggestive of TB and were enrolled in the study. Detailed patient-related socio-demographic, behavioral, clinical data and environmental data were collected from those 652 DM patients. Among them, the largest number were enrolled from Addis Ababa (225, 34.5%), followed by the Southern Ethiopia Region (130, 9.9%). The mean age was 55.47 years (± sd = 12.85), and the most frequent age group was above 65 years (201, 30.8%) followed by 55–64 years (180, 27.6%) and 45–54 (167, 25.6%) years. More than half of the participants (346, 53.1%) were females, and 88.5% (346) were married (577, 88.5%). Of all participants, 82.1% (535) were urban dwellers. Of all participants, 28.8% (188) had no formal education (Table 1).

Table 1 Demographic and behavioral characteristics, and environmental conditions of tuberculosis presumptive diabetic patients attending selected health facilities in Ethiopia from January to December 2023

Among all participants, 3.1% [20], 3.1% [20], 13.8% (90), and 5.4% (35) of them were cigarette smokers, alcohol consumers, raw milk consumers, and khat consumers users, respectively. Thirty-one (4.8%) of the study participants reported they were living or working in a high TB risk setting in general, and specifically, 12, 10, and four of them were working or living in health care settings, military barrack, and refugee camps, respectively (Table 1).

Clinical conditions of study participants

The most commonly observed symptoms were, cough lasting for two or more weeks (485, 74.4%), decreased appetite (426, 65.3%), weight loss (373, 57.2%), night sweet (366, 56.1%), and fever (357, 54.8%) (Fig. 1). Of all the study participants, 3.7% [24], 42.8% (279), 40.3% (263), and 13.2% (86) of them were underweight (BMI < 18.5 kg/m2), normal weight (BMI = 18.5–24.9 kg/m2), overweight (BMI = 25–29.9 kg/m2) and obese (BMI ≥ 30 kg/m2), respectively with a mean BMI of 25.7 kg/m2 ((± sd = 4.14). There was a BCG vaccination scar in 21.8% (142) of the participants. Thirty-eight (5.8%) participants had a previous TB treatment history. There was contact history with a known TB and MDR-TB cases in 18 (2.8%), and six (0.9%) participants, respectively. Among the participants, 3.4% [22], and 81.0% (528) were HIV seropositive and negative respectively and it was unknown for the remaining 15.6% (102) of them. Type II DM was the most common DM type (576, 88.3%), and 71.9% (469) had a FBS level above 126 mg/dL at the time of TB screening. The DM medication type included; oral medication (284, 43.6%), insulin injection (260, 39.9%), and a combination of both medications (108, 16.6%). The specific medications included; Metformin, Neutral Protamine Hagedorn (NPH) insulin, and Glibenclamide. Of all, 35.6% (232) of study participants had more than 10 years of DM follow-up (Table 2).

Fig. 1

Proportions of tuberculosis suggestive individual symptoms among tuberculosis presumptive diabetic patients attending selected health facilities in Ethiopia from January to December 2023

Table 2 Clinical characteristics of tuberculosis presumptive diabetic patients attending selected health facilities in Ethiopia from January to December 2023

There was DM-associated complication in 61.8% (403) of the study participants with heart disease (207, 31.7%), and neuropathy (30.2%) being the most frequent complication types (Fig. 2). Three-fourth (75.0%, 489) of study participants had at least one disease comorbidity, with hypertension (390, 59.8%) being the most common, followed by asthma (26.2%, 171), arthritis (24.2%, 158), and nephropathy (95, 14.6%) (Fig. 3).

Fig. 2

Frequency of complications observed among tuberculosis presumptive diabetic patients attending selected health facilities in Ethiopia from January to December 2023

Fig. 3

Frequency of comorbidities observed among tuberculosis presumptive diabetic patients attending selected health facilities in Ethiopia from January to December 2023

Prevalence of tuberculosis in diabetic patients

Overall, 73 TB cases (11.2%, 95%CI = 8.7–13.5%) were detected among the 652 TB presumptive DM patients, and 56 (8.6%, 95%CI = 6.4–10.6%) of them were bacteriologically confirmed TB cases. Among the culture-processed specimens, 47 (7.2%, 95%CI = 5.4–9.5) were culture-positive with MTBC, and 81.1% (529) were culture-negative. Non-tuberculous mycobacteria were identified in 45 cases (6.9%) and the remaining 31 specimens (4.8%) were contaminated. By the Xpert MTB/RIF Ultra assay, 55 MTBC cases (8.4%) were detected. Smear was made from the culture-processed sediment and became positive in 40 cases (6.1%). An additional 17 TB cases were clinically diagnosed. These clinically diagnosed cases were negative by bacteriological diagnostic methods. The clinical TB cases were those study participants who had X-ray findings suggestive of TB which were treated as a TB case based on the clinician’s decision. In total, 73 DM patients were found to have TB which yielded a point prevalence of 517/100,000 among all screened DM patients (Fig. 4). Among all 73 TB cases, 74.0% (54), 16 (21.9%), and 4.1% [3] were pulmonary, extrapulmonary, and disseminated TB cases, respectively. In a sub-group analysis, excluding 56 DM patients with known factors associated with TB; 18 HIV seropositive, 34 with previous TB treatment history, and four with both HIV seropositive and previous TB treatment history, the prevalence of TB was 9.6% (57/96), and the prevalence of bacteriological confirmed TB was 7.4% (44/596).

Fig. 4

Tuberculosis identified among tuberculosis presumptive diabetic patients

attending selected health facilities in Ethiopia from January to December 2023

Tuberculosis per participant characteristics

A comparably higher proportion of TB cases were identified in the younger age group (p = 0.05). The highest proportion of TB was detected among the underweight DM patients (29.2%, p = 0.07) compared to other groups. Based on behavioral associated factors, the proportions of identified TB cases among active cigarette smokers, alcohol consumers, and khat consumers were 45.0% (p < 0.001), 30.0% (p = 0.01), and 31.45% (p < 0.001), respectively. These proportions were higher compared to their counterparts being 10.7%, 10.6%, and 10.0% in their respective order. Based on the assessed TB suggestive symptoms, a comparably higher proportion of TB was detected among those participants who had a cough lasting for two or more weeks (12.8%, p = 0.03), malaise (19.7%, p = 0.03), and fatigue (14.9%, p = 0.09) compared to their counterparts (Table 3, Supplementary Table 1).

Table 3 Tuberculosis positivity per observed symptoms and behavioral, and clinical factors among tuberculosis presumptive diabetic patients attending selected health facilities in Ethiopia from January to December 2023

Compared to their counterparts, higher TB prevalence was detected among DM patients who had a previous TB treatment history (28.9%, p = 0.001), had contact with a known TB patient (55.6%, p < 0.001), had contact with an MDR-TB case (50.0%. p = 0.01), being HIV seropositive (27.3%, p = 0.04), had type I DM (19.7%, p = 0.01), taking insulin medication (14.2%, p = 0.08), had more than 10 years of DM follow-up (15.1%, p = 0.05), had FBS level > 126 mg/dL (13.3%, p = 0.01), and had irregular follow-up status (28.9%, p = 0.001) ( Table 3, Supplementary Table 1).

Factors associated with tuberculosis in diabetic patients

The bi-variable logistic regression model was computed for all the study variables that fulfilled the assumption, and 29 variables with a P-value < 0.25 in the bi-variable analysis were subjected to a multi-variable logistic regression model to identify the independently associated factors. In the adjusted model, younger age groups, current cigarette smoking, having contact with a known TB case, being HIV seropositive, having more than 10 years of DM follow-up, taking insulin medication, and having an FBS level > 126 mg/dL all had a statistically significant association with tuberculosis in DM patients (Table 4).

Table 4 Factors associated with developing tuberculosis among tuberculosis presumptive diabetic patients attending selected health facilities in Ethiopia from January to December 2023

Accordingly, those diabetic patients 15–24, 25–34, 35–44, and 45–54 years had 10.98 (95%CI = 1.90-63.56), 4.74 (95%CI = 1.12–20.13), 5.70 (95%CI = 2.09–15.55), and 2.68 (95%CI = 1.22–5.92) times the odds to have TB compared to those DM patients above 65 years old, respectively. When the age is categorized into two groups, DM patients < 65 years of age had 2.41 (95%CI = 1.22–4.80) times the odds of having TB compared to those DM patients above 65 years.

Likewise, DM patients who smoke cigarettes had a cough lasting two or more weeks and had a contact history with a known TB case had 7.50 (95%CI = 2.54–22.19), 2.73 (95%CI = 1.25–5.60), and 9.16 (95%CI = 2.83–29.70) times the odds to develop TB compared to their counterparts, respectively. Similarly, HIV-positive DM patients had 4.40 (95%CI = 1.36–14.46) times the odds of having TB compared to HIV negatives. In addition, the odds of developing TB were higher among DM patients with a FBS > 126 mg/dL (aOR = 2.72, 95%CI = 1.26–5.89). Besides, those who had more than 10 years of DM follow-up had 4.87 (95%CI = 2.06–11.52) times the odds of having TB compared to those who had < 5 years of follow-up in the diabetic clinic. Furthermore, those DM patients who took only insulin injections had three times (95%CI = 1.16–7.81) the odds of having TB compared to those who took both insulin injection and oral medication (Table 4).