Introduction
Background Tuberculosis
(TB) remains one of the top infectious disease killers worldwide, despite being preventable and curable for decades. According to the World Health Organization (WHO) Global Tuberculosis Report 2023, an estimated 10.6 million people fell ill with TB in 2022, and 1.3 million died from the disease [1]. The emergence and spread of drug-resistant tuberculosis (DR-TB) have become a major threat to global TB control efforts, jeopardizing progress toward the End TB Strategy targets [2]. Drug-resistant TB occurs when Mycobacterium tuberculosis strains develop resistance to the most effective first-line anti-TB drugs. Multidrug-resistant tuberculosis (MDR-TB), defined as resistance to at least rifampicin and isoniazid—the two most potent first-line drugs—requires longer, more toxic, and expensive treatment regimens with poorer outcomes [3]. In 2022, an estimated 410,000 people developed MDR-TB or rifampicin-resistant TB (RR-TB) globally, with only about two-thirds of these cases being diagnosed and enrolled on treatment [1].
The Situation in Ethiopia
Ethiopia is among the 30 high TB and TB/HIV burden countries, with an estimated TB incidence of 119 per 100,000 population in 2022 [1]. The country has made significant progress in TB control over the past decade, achieving a treatment success rate of 90% for drug-susceptible TB. However, drug-resistant TB poses an increasing challenge. The national prevalence of MDR-TB among new cases is estimated at 1.4% and among previously treated cases at 12% [1-4]. Regional variations in drug resistance patterns exist across Ethiopia due to differences in TB control program implementation, population movement, HIV prevalence, and circulating M. tuberculosis strains. The Jimma Zone in South-West Ethiopia is particularly important to study due to its unique characteristics: it serves as a major commercial hub with significant population movement, has areas with high HIV prevalence, and includes both urban and remote rural communities with varying access to healthcare services [5].
Molecular Epidemiology of TB
Molecular epidemiology combines conventional epidemiological methods with molecular typing techniques to understand the transmission dynamics and strain diversity of M. tuberculosis [6]. Various genotyping methods have been developed, including spoligotyping, mycobacterial interspersed repetitive unit-variable number tandem repeat (MIRU-VNTR) typing, and whole-genome sequencing (WGS). Spoligo- typing, while having lower discriminatory power than other methods, remains valuable for phylogenetic studies and initial screening due to its relative simplicity, reproducibility, and ability to assign strains to known lineages and families [7]. Different M. tuberculosis lineages and families have been associated with varying degrees of transmissibility, virulence, and propensity to develop drug resistance [8]. Understanding which strain families circulate in a given region and their association with drug resistance can provide insights into transmission dynamics and inform targeted interventions. The Ethiopia_3 family (also known as Central Asian Strain or CAS1-Delhi) has been reported as a predominant lineage in various parts of Ethiopia, but regional data from Jimma Zone remain limited [9-10].
Molecular Mechanisms of Drug Resistance
Resistance to rifampicin, the most powerful first-line anti-TB drug, is primarily caused by mutations in an 81-base pair core region of the rpoB gene encoding the β-subunit of RNA polymerase [11]. The most common mutations occur at codons 531 (S531L), 526 (H526Y/D/V), and 516 (D516V) [12]. Resistance to isoniazid is more complex, involving mutations in several genes, most commonly the katG gene (particularly codon 315, S315T) and the promoter region of the inhA gene [13]. The katG S315T mutation confers high-level isoniazid resistance while preserving catalase-peroxidase activity necessary for bacterial fitness, explaining its prevalence in clinical isolates [14]. Line Probe Assays (LPAs), such as GenoType® MTB- DRplus, detect these common mutations rapidly, allowing for timely diagnosis of MDR-TB and appropriate treatment initiation [15]. The WHO strongly recommends LPAs as the initial test for detecting rifampicin and isoniazid resistance in sputum smear-positive specimens or cultured isolates [16].
Rationale for the Study
Despite the national estimates of drug-resistant TB in Ethiopia, there is a paucity of local data from the Jimma Zone regarding the prevalence of drug resistance, the specific mutations conferring resistance, and the molecular characteristics of circulating strains.
Such information is essential for:
1. Informing local treatment guidelines: Understanding local resistance patterns helps clinicians choose appropriate empiric treatment regimens while awaiting drug susceptibility testing (DST) results.
2. Strengthening the TB control program: Identifying trans- mission hotspots and risk groups enables targeted interventions.
3. Monitoring the effectiveness of control measures: Baseline data allow for assessment of future interventions.
4. Contributing to the global knowledge base: Data from understudied regions enrich our understanding of TB epidemiology and evolution.
Research Questions
1. What is the prevalence of rifampicin resistance, isoniazid resistance, and MDR-TB among pulmonary TB patients in Jimma Zone?
2. What are the specific mutations in the rpoB, katG, and inhA genes associated with drug resistance in these isolates?
3. What are the circulating M. tuberculosis strain families in Jimma Zone, and what is their phylogenetic diversity?
4. What demographic and clinical factors are associated with drug-resistant TB in this population?
Objectives
General Objective: To determine the prevalence and molecular characteristics of drug-resistant Mycobacterium tuberculosis isolates among pulmonary tuberculosis patients in Jimma Zone, South-West Ethiopia.
Specific Objectives:
1. To determine the prevalence of rifampicin resistance, isoniazid resistance, and MDR-TB using GeneXpert MTB/RIF and Line Probe Assay.
2. To characterize the molecular mechanisms of resistance by identifying mutations in rpoB, katG, and inhA genes.
3. To identify the circulating M. tuberculosis strain families using spoligotyping.
4. To assess demographic and clinical factors associated with drug-resistant TB.
5. To compare the diagnostic performance of GeneXpert MTB/RIF and LPA for detecting drug resistance.
Materials And Methods
Study Design and Period
A health facility-based cross-sectional study was conducted from September 2018 to August 2019 (12 months). This design was chosen to estimate the prevalence of drug-resistant TB at a single point in time and to identify associated factors.
Study Area
The study was conducted in Jimma Zone, Oromia Regional State, South-West Ethiopia. Jimma Zone is located approximately 350 km southwest of Addis Ababa, covering an area of 15,568.58 km². According to the 2017 Central Statistical Agency projection, the zone has a population of approximately 3.2 million people, with 88% living in rural areas engaged primarily in subsistence agriculture and coffee cultivation [17]. The zone has one tertiary teaching hospital (Jimma University Medical Center), three general hospitals, eight primary hospitals, and 96 health centers. TB diagnosis and treatment services are available at all public health facilities, following the national TB program guidelines. The zone has a TB notification rate of 145 per 100,000 population, slightly above the national average [18]. Source and Study Population Source Population: All pulmonary TB patients attending public health facilities in Jimma Zone during the study period. Study Population: Newly diagnosed and previously treated smear-positive or clinically diagnosed pulmonary TB patients aged ≥18 years who were enrolled at selected health facilities during the study period.
Eligibility Criteria Inclusion Criteria:
• Age ≥18 years
• Presumptive pulmonary TB patients with at least one positive smear microscopy (acid-fast bacilli, AFB) or positive GeneXpert result
• Willingness to provide a sputum sample and participate in the study
• Provision of written informed consent • Residence in Jimma Zone for at least 6 months prior to enrollment
Exclusion Criteria:
• Patients on anti-TB treatment for more than 7 days at the time of enrollment
• Patients unable to produce sputum
• Extrapulmonary TB cases without pulmonary involvement
• Critically ill patients unable to provide consent
• Previous enrollment in this study (to avoid duplication),
Sample Size Determination
The sample size was calculated using single population proportion formula: n = (Z² × p × (1-p)) / d², where:
• Z = 1.96 (95% confidence level)
• p = expected prevalence of MDR-TB among pulmonary TB patients = 0.05 (based on previous Ethiopian stud- ies showing MDR-TB prevalence ranging from 1.6% to 12.8% [19,20], we used 5% as a conservative estimate)
• d = margin of error = 0.025 n = (1.96² × 0.05 × 0.95) / 0.025² = 292 Adding 10% for non-response and inadequate samples, the final sample size was 322. However, during the study period, 286 eligible patients were enrolled due to logistical constraints and the end of the study period. Post-hoc power analysis indicated that this sample size was sufficient to detect an MDR-TB prevalence of 5% with 80% power at 95% confidence. Sampling Technique and Health Facility Selection A multi-stage sampling approach was employed:
Stage1:
Health Facility Selection
Twelve health facilities were purposively selected based on:
• High patient volume (≥50 TB patients diagnosed annually) • Geographic representation (covering urban and rural areas across the zone)
• Availability of TB diagnostic services (microscopy and GeneXpert)
• Accessibility and security considerations
Selected facilities included:
1. Jimma University Medical Center (tertiary hospital, urban)
2. Shenen Gibe General Hospital (urban)
3. Limu Genet General Hospital (semi-urban)
4. Agaro Primary Hospital (semi-urban)
5. Seka Primary Hospital (rural)
6. Serbo Health Center (rural)
7. Shebe Health Center (rural)
8. Yebu Health Center (semi-urban)
9. Omo Nada Health Center (rural)
10. Gera Health Center (rural)
11. Chora Botor Health Center (rural)
12. Manna Health Center (rural)
Stage 2:
Patient Enrollment: Consecutive sampling was used to enroll all eligible patients presenting at the selected facilities during the study period until the desired sample size was approached. This method was chosen to minimize selection bias and ensure representation of all patient types. Data Collection Procedures Questionnaire Administration A structured, pretested questionnaire was developed in English, translated into Afan Oromo and Amharic (the local languages), and back-translated to ensure consistency. Trained data collectors (Bachelor of Science degree holder nurses or public health officers) administered the questionnaire through face-to-face interviews. Data collected included:
• Socio-demographic characteristics: age, sex, residence, marital status, education level, occupation, monthly income, household size
• Clinical characteristics: TB treatment history (new/re- treatment), previous TB episodes, HIV status, diabetes mellitus, other comorbidities
• Behavioral factors: smoking history, alcohol use, khat chewing
• Exposure history: contact with known TB patient, household crowding, ventilation, history of incarceration, institutionalization
• Health-seeking behavior: time from symptom onset to first consultation, number of healthcare visits before diagnosis.
Clinical Data Extraction.
Additional clinical data were extracted from patient medical records using a standardized data abstraction form, including:
• Chest X-ray findings (if performed)
• Previous TB treatment details (regimen, duration, outcome)
• HIV testing results and ART status if HIV-positive
• Baseline laboratory results Specimen Collection and Transport Sputum Collection Patients were instructed on proper sputum collection technique. Two sputum specimens (spot and early morning) were collected from each participant in sterile, wide-mouthed, screw-capped containers. For patients unable to expectorate spontaneously, sputum induction with nebulized hypertonic saline (3%) was performed under appropriate infection control precautions. Specimen Transport Sputum specimens were transported to the respective health facility laboratories within 2 hours of collection in cold boxes maintained at 2-8°C. For specimens collected at peripheral health centers without on-site GeneXpert or culture facilities, samples were stored at 4°C and transported twice weekly to Jimma University Medical Center Tuberculosis Research Laboratory for processing. Laboratory Procedures Smear Microscopy Direct smears were prepared from purulent portions of sputum, air-dried, heat-fixed, and stained using the Ziehl-Neelsen (ZN) technique. Smears were examined under oil immersion (1000× magnification) and graded according to WHO/International Union Against Tuberculosis and Lung Disease (IUATLD) guidelines [21]: • Negative: No AFB per 100 immersion fields
• Scanty: 1-9 AFB per 100 fields (record exact number)
• 1+: 10-99 AFB per 100 fields
• 2+: 1-10 AFB per field in at least 50 fields
• 3+: >10 AFB per field in at least 20 fields GeneXpert MTB/RIF Assay GeneXpert MTB/RIF assay (Cepheid, Sunnyvale, CA, USA) was performed according to the manufacturer’s instructions [22]:
1. Sample processing: Sputum samples were liquefied by adding sample reagent (SR) in a 2:1 ratio (2 mL SR to 1 mL sputum) and vortexed vigorously.
2. Incubation: Tubes were incubated at room temperature for 15 minutes, vortexed again, and incubated for another 5 minutes.
3. Transfer: Processed sample (2 mL) was transferred to the cartridge using the provided transfer pipette.
4. Loading: Cartridges were loaded into the GeneXpert instrument, and the automated process was initiated.
5. Interpretation: Results were automatically interpreted by the GeneXpert software, providing:
• M. tuberculosis complex detected or not detected
• Semi-quantitative result (very low, low, medium, high)
• Rifampicin resistance detected, not detected, or indeterminate. Decontamination and Concentration For samples requiring culture or LPA, the N-acetyl-L-cyste- ine-sodium hydroxide (NALC-NaOH) method was used [23]:
1. An equal volume of NALC-NaOH solution (final concentration: 1% NaOH, 0.5% NALC) was added to the sample.
2. The mixture was vortexed and incubated at room temperature for 15 minutes with intermittent shaking.
3. Phosphate buffer (pH 6.8) was added to a total volume of 50 mL to neutralize.
4. The mixture was centrifuged at 3000×g for 15 minutes at 4°C.
5. The supernatant was discarded, and the pellet was resuspended in 2 mL phosphate buffer.
6. A smear was prepared from the concentrated sediment for ZN staining to confirm AFB positivity. Line Probe Assay (GenoType® MTBDRplus VER 2.0) The GenoType® MTBDRplus VER 2.0 assay (Hain Life- Science, Nehren, Germany) was performed on sputum specimens that were smear-positive (≥1+) according to the manufacturer’s instructions [24]:
Day
1: DNA Extraction 1. 500 μL of decontaminated sediment was transferred to a 2 mL screw-cap tube and centrifuged at 10,000×g for 15 minutes.
2. The supernatant was discarded, and the pellet was resuspended in 100 μL molecular biology-grade water. 3. The tube was heated at 95°C for 20 minutes in a heating block. 4. After cooling to room temperature, the tube was sonicated for 15 minutes in an ultrasonic water bath.
5. Centrifugation at 13,000×g for 5 minutes to pellet debris
6. The supernatant containing DNA was transferred to a fresh tube and stored at -20°C until use.
Day 2:
Multiplex Amplification
1. A master mix was prepared containing: 35 μL primer-nu- cleotide mix (PNM), 5 μL 10× polymerase incubation buffer, 2 μL 25 mM MgCl2, 1 μL Hot Start Taq DNA polymerase (5 U/μL), and 2 μL molecular grade water (total 45 μL per reaction).
2. 5 μL extracted DNA was added to 45 μL master mix.
3. Amplification was performed in a thermal cycler with the following protocol:
• 95°C for 15 minutes (1 cycle) • 95°C for 30 seconds, 58°C for 2 minutes, 72°C for 30 seconds (10 cycles)
• 95°C for 25 seconds, 53°C for 40 seconds, 70°C for 40 seconds (30 cycles)
• 70°C for 8 minutes (1 cycle)
• Hold at 4°C
Day 3:
Hybridization and Detection
1. 20 μL amplified product was mixed with 20 μL denaturation reagent (DEN) and incubated at room temperature for 5 minutes.
2. 1 mL pre-warmed hybridization buffer (HYB) was added to each well of a trough.
3. Denatured samples were added to the trough and incubated in a TwinCubator at 45°C for 30 minutes with gentle shaking.
4. Strips were washed twice with 1 mL stringent wash buffer (STR) at 45°C for 15 minutes.
5. Strips were incubated with 1 mL conjugate (CON) at room temperature for 30 minutes.
6. Strips were washed twice with 1 mL rinse solution (RIN) and once with distilled water.
7. Strips were incubated with 1 mL substrate (SUB) at room temperature in the dark for 10 minutes.
8. Strips were washed with distilled water and dried between absorbent paper.
9. Strips were interpreted using the evaluation sheet, with band patterns indicating:
• Species identification (M. tuberculosis complex)
• Rifampicin resistance (absence of wild-type bands or presence of mutation bands in rpoB gene)
• Isoniazid resistance (absence of wild-type bands or presence of mutation bands in katG and/or inhA genes) Mycobacterial Culture (Löwenstein-Jensen Medium) For samples that were smear-positive but GeneXpert-negative or for LPA with indeterminate results, culture was performed on Löwenstein-Jensen (LJ) medium:
1. 0.5 mL decontaminated sediment was inoculated onto two LJ slants (one with and one without 0.75% sodium pyruvate).
2. Slants were incubated at 37°C in a slanted position with loosened caps for one week to allow evaporation of residual fluid.
3. Caps were tightened, and slants were incubated vertically for up to 8 weeks.
4. Cultures were examined weekly for growth.
Positive cultures were confirmed by ZN staining for AFB and by immunochromatographic assay for MPT64 antigen (SD Bioline, Standard Diagnostics, Korea) to identify M. tu- berculosis complex. Spoligotyping Spoligotyping was performed on all culture-positive isolates at the National Tuberculosis Reference Laboratory, Ethiopian Public Health Institute, Addis Ababa, following the standard protocol [25]: Principle: Spoligotyping detects the presence or absence of 43 known spacer sequences in the direct repeat (DR) region of the M. tuberculosis genome.
Procedure:
1. DNA extraction: DNA was extracted from LJ cultures using the cetyltrimethylammonium bromide (CTAB) method.
2. PCR amplification: The DR region was amplified using primers:
• Dra: 5’-GGTTTTGGGTCTGACGAC-3’ (biotinylated at the 5’ end)
• Drb: 5’-CCGAGAGGGGACGGAAAC-3’ PCR conditions: 95°C for 15 minutes; 30 cycles of 95°C for 1 minute, 55°C for 1 minute, 72°C for 30 seconds; final extension 72°C for 5 minutes.
3. Hybridization: A membrane with immobilized oligonucleotides representing the 43 spacers was placed in a Miniblotter. PCR products were denatured and hybridized to the membrane at 60°C for 1 hour.
4. Detection: The membrane was incubated with streptavidin-peroxidase conjugate, followed by enhanced chemiluminescence detection (ECL, Amersham) and exposure to X-ray film.
5. Interpretation: The presence of a spacer was indicated by a black square on the film. The spoligotype pattern (43-digit binary code) was entered into the SITVIT2 international database for lineage assignment. Quality Control Rigorous quality control measures were implemented: 1. Sample collection and transport: Standard operating procedures (SOPs) were followed, and cold chain maintenance was monitored. 2. Laboratory procedures: All testing was performed according to the manufacturer’s instructions and validated SOPs.
3. Controls: For each batch of tests, appropriate controls were included:
• GeneXpert: Internal control included in each cartridge
• LPA: Negative control (water), positive control (M. tuberculosis H37Rv), and mutation controls provided by the manufacturer
• Culture: Negative control (sterile water) and positive control (M. tuberculosis H37Rv)
4. Proficiency testing: The laboratory participates in the national External Quality Assurance (EQA) program for smear microscopy and GeneXpert.
5. Double reading: All LPA strips and spoligotyping films were read independently by two trained laboratory scientists, with discrepancies resolved by a third reader. Data Management and Analysis Data Entry and Cleaning Questionnaire data were double-entered into EpiData version 3.1 (EpiData Association, Odense, Denmark) by two independent data entry clerks. Data were validated using con- sistency checks and range checks. Laboratory results were entered separately and merged with questionnaire data. Statistical Analysis Statistical analysis was performed using SPSS version 25.0 (IBM Corp., Armonk, NY, USA) and STATA version 15.0 (StataCorp, College Station, TX, USA).
Descriptive Statistics:
• Categorical variables: Frequencies and percentages
• Continuous variables: Mean (± standard deviation) for normally distributed data; median (interquartile range, IQR) for non-normally distributed data Prevalence Estimates:
Prevalence of drug resistance was calculated as:
• Rifampicin resistance: (Number resistant / Total cul- ture-positive) × 100
• Isoniazid resistance: (Number resistant / Total culture-positive) × 100
• MDR-TB: (Number MDR / Total culture-positive) × 100 95% confidence intervals (CIs) were calculated using the Wilson score method. Bivariate Analysis: Associations between independent variables (demographic, clinical, behavioral) and outcome variables (drug resistance, MDR-TB) were assessed using: • Chi-square test or Fisher’s exact test for categorical variables
• Independent t-test or Mann-Whitney U test for continuous variables. Crude odds ratios (COR) with 95% CIs were calculated.
Multivariable Analysis: Variables with p<0.25 in bivariate analysis were entered into a multivariable logistic regression model using backward stepwise elimination. The final model included variables with p<0.05 and those considered clinically important regardless of statistical significance. Adjusted odds ratios (AOR) with 95% CIs were calculated. Model fit was assessed using the Hosmer-Lemeshow goodness-of-fit test.
Spoligotype Analysis: Spoligotype patterns were entered as binary codes (1 for spacer present, 0 for absent) in Microsoft Excel. Patterns were compared with the SITVIT2 database to assign Spoli- gotype International Type (SIT) numbers and lineages. Cluster analysis was performed using the unweighted pair group method with arithmetic mean (UPGMA) in MEGA version 7.0. A cluster was defined as two or more isolates with identical spoligotype patterns.
The clustering rate was calculated as: (Number of clustered isolates - Number of clusters) / Total isolates × 100.
Diagnostic Performance Comparison: Sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV) of GeneXpert for rifampicin resistance were calculated using LPA as the reference standard (recognizing that LPA also has limitations, but culture-based phenotypic DST was not available). Agreement between tests was assessed using Cohen’s kappa coefficient. Statistical Significance All statistical tests were two-sided, and p<0.05 was considered statistically significant.
Ethical Considerations Ethical Approval Ethical approval was obtained from:
1. Institutional Review Board of Aklilu Lemma Institute of Pathobiology, Addis Ababa University (Protocol No: ALI- PB/IRB/018/2018, approved August 15, 2018)
2. Institutional Review Board of Jimma University Institute of Health Sciences (Ref No: IHRPGD/289/2018, approved September 3, 2018).
Permission to conduct the study was obtained from:
• Oromia Regional Health Bureau
• Jimma Zone Health Department
• Administrators of each selected health facility, Informed Consent.
Written informed consent was obtained from all participants after a thorough explanation of the study purpose, procedures, risks, and benefits in their preferred language (Afan Oromo, Amharic, or English). For participants unable to read and write, the consent form was read to them in the presence of a literate witness, and thumbprint signatures were obtained. Key elements of consent included:
• Voluntary participation with the right to withdraw at any time without affecting their medical care
• Confidentiality of all information
• Potential risks: minor discomfort during sputum collection
• Potential benefits: free TB diagnosis and access to results for treatment
• Use of specimens for research purposes
• Contact information for questions or concerns.
Confidentiality:
All data were anonymized using unique study identifiers.
Personal identifiers were removed from laboratory specimens and data analysis files.
Electronic data were password-protected, and paper records were stored in locked cabinets accessible only to the research team.
Clinical Management All laboratory results (smear microscopy, GeneXpert, LPA) were communicated to the treating physicians within 72 hours of availability to guide patient management.
Patients diagnosed with drug-resistant TB were referred to Jimma University Medical Center MDR-TB treatment unit for appropriate management according to national guidelines. Community Engagement Prior to study initiation, community sensitization meetings were held with health facility TB focal persons, community health workers (Health Extension Workers), and patient support groups to explain the study purpose and procedures.
Operational Definitions
• New TB case: Patient who has never had TB treatment or has taken anti-TB drugs for less than 1 month [26]
• Previously treated TB case: Patient who has received 1 month or more of anti-TB drugs in the past, including relapse, treatment after failure, treatment after loss to follow-up, and other previously treated cases [26]
• Presumptive TB: Individuals with symptoms or signs suggestive of TB, particularly cough of ≥2 weeks duration, fever, night sweats, weight loss, and chest pain
• Pulmonary TB: TB involving the lung parenchyma [26]
• Rifampicin resistance: Resistance to rifampicin detected by either GeneXpert or LPA, with or without resistance to other drugs
• Isoniazid resistance: Resistance to isoniazid detected by LPA, with or without resistance to other drugs
• Multidrug-resistant TB (MDR-TB): Resistance to at least both rifampicin and isoniazid [26]
• Mono-resistance: Resistance to one first-line anti-TB drug only
• Poly-resistance: Resistance to more than one first-line anti-TB drug, other than both rifampicin and isoniazid
• HIV co-infection: Positive HIV serostatus documented in medical record or determined during study enrollment (with separate consent for HIV testing if status unknown)
• Contact history: Self-reported history of close contact (living in the same household or frequent daily contact) with a person diagnosed with TB in the past 2 years
• Smear-positive: At least one sputum specimen positive for AFB by ZN microscopy [21]
• Smear-negative: Sputum specimens negative for AFB by ZN microscopy [21]
Results
Enrollment and Participant Characteristics
A total of 322 presumptive TB patients were screened for eligibility during the study period. Of these, 286 (88.8%) met the inclusion criteria and were enrolled. Reasons for exclusion included: age <18 years (n=18), inability to produce sputum (n=9), on anti-TB treatment >7 days (n=5), and refusal to participate (n=4).
Socio-demographic Characteristics
The socio-demographic characteristics of the 286 participants are presented in Table 1.
|
Character- istic
|
Category
|
Frequency (n)
|
Percentage (%)
|
|
Age group (years)
|
18-24
|
52
|
18.2
|
|
|
25-34
|
89
|
31.1
|
|
|
35-44
|
71
|
24.8
|
|
|
45-54
|
43
|
15.0
|
|
|
≥55
|
31
|
10.8
|
|
|
Mean (±SD)
|
34.7 ± 12.4
|
|
|
Sex
|
Male
|
171
|
59.8
|
|
|
Female
|
115
|
40.2
|
|
Residence
|
Urban
|
128
|
44.8
|
|
|
Rural
|
158
|
55.2
|
|
Marital status
|
Married
|
164
|
57.3
|
|
|
Single
|
81
|
28.3
|
|
|
Divorced/ Widowed
|
41
|
14.3
|
|
Educational level
|
No formal education
|
93
|
32.5
|
|
|
Primary (1-8)
|
112
|
39.2
|
|
|
Secondary (9-12)
|
58
|
20.3
|
|
|
Tertiary
|
23
|
8.0
|
|
Occupation
|
Farmer
|
108
|
37.8
|
|
|
Housewife
|
47
|
16.4
|
|
|
Daily laborer
|
42
|
14.7
|
|
|
Merchant
|
31
|
10.8
|
|
|
Government employee
|
24
|
8.4
|
|
|
Student
|
19
|
6.6
|
|
|
Unemployed
|
15
|
5.2
|
|
Monthly income (ETB)
|
<500
|
53
|
18.5
|
|
|
500-1500
|
126
|
44.1
|
|
|
1501-3000
|
71
|
24.8
|
|
|
>3000
|
36
|
12.6
|
|
|
Median (IQR)
|
1250 (750-
2150)
|
|
|
Household size
|
1-3 persons
|
89
|
31.1
|
|
|
4-6 persons
|
142
|
49.7
|
|
|
≥7 persons
|
55
|
19.2
|
| |
Mean (±SD)
|
4.6 ± 2.1 |
|
ETB: Ethiopian Birr; IQR: Interquartile range; SD: Standard deviation
Table 1: Socio-demographic Characteristics of Study Participants (N=286) The majority of participants were male (59.8%), with a male to-female ratio of approximately 1.5:1. The mean age was 34.7 years (SD: 12.4), with the largest age group being 25 34 years (31.1%). More than half (55.2%) resided in rural areas. Farming was the most common occupation (37.8%), and most participants had either no formal education (32.5%) or only primary education (39.2%). The median monthly income was 1250 ETB (approximately 45 USD at the time of study), reflecting the low socioeconomic status of the study population.
Clinical and Behavioral Characteristics
Table 2 presents the clinical and behavioral characteristics of participants.
|
Character- istic
|
Category
|
Frequen- cy (n)
|
Percentage (%)
|
|
TB treat- ment history
|
New case
|
225
|
78.7
|
|
|
Previously treated
|
61
|
21.3
|
|
|
- Relapse
|
32
|
(52.5% of
previously treated)
|
|
|
- Treatment after failure
|
14
|
(23.0%)
|
|
|
- Treatment after LTFU
|
12
|
(19.7%)
|
|
|
- Other previ- ously treated
|
3
|
(4.9%)
|
|
Duration of cough (weeks)
|
<2
|
24
|
8.4
|
|
|
2-4
|
87
|
30.4
|
|
|
5-8
|
103
|
36.0
|
|
|
>8
|
72
|
25.2
|
|
|
Median (IQR)
|
6 (4-10)
|
|
|
HIV status
|
Positive
|
68
|
23.8
|
|
|
Negative
|
206
|
72.0
|
|
|
Unknown/Not tested
|
12
|
4.2
|
|
On ART (if HIV+)
|
Yes
|
51
|
75.0
|
|
|
No
|
17
|
25.0
|
|
Diabetes mellitus
|
Yes
|
16
|
5.6
|
|
|
No
|
270
|
94.4
|
|
Smoking history
|
Never smoked
|
197
|
68.9
|
|
|
Former smok- er
|
48
|
16.8
|
|
|
Current smok- er
|
41
|
14.3
|
|
Alcohol use
|
Never
|
156
|
54.5
|
|
|
Occasional
|
89
|
31.1
|
|
|
Frequent
|
41
|
14.3
|
|
Khat chew- ing
|
Never
|
172
|
60.1
|
|
|
Occasional
|
71
|
24.8
|
|
|
Daily
|
43
|
15.0
|
|
Contact with TB patient
|
Yes
|
78
|
27.3
|
|
|
No
|
208
|
72.7
|
|
Household ventilation
|
Adequate
|
134
|
46.9
|
|
|
Inadequate
|
152
|
53.1
|
|
History of incarcera- tion
|
Yes
|
23
|
8.0
|
|
|
No
|
263
|
92.0
|
|
BCG scar present
|
Yes
|
184
|
64.3
|
|
|
No/Uncertain
|
102
|
35.7
|
|
Time to first
consultation
|
<4 weeks
|
86
|
30.1
|
| |
4-8 weeks
|
123
|
43.0
|
| |
>8 weeks
|
77
|
26.9
|
| |
Median (IQR)
|
6 (3-10)
|
|
LTFU: Loss to follow-up; ART: Antiretroviral therapy; BCG: Bacille Calmette-Guérin
Table 2: Clinical and Behavioral Characteristics of Study Participants (N=286)
Among the 286 participants, 225 (78.7%) were new TB cases, and 61 (21.3%) had a history of previous TB treatment. Among previously treated cases, relapse was the most common category (52.5%), followed by treatment after failure (23.0%). The median duration of cough before presentation was 6 weeks (IQR: 4-10 weeks), indicating significant diagnostic delay. HIV co-infection was documented in 23.8% of participants, with 75% of HIV-positive individuals on antiretroviral therapy (ART) at enrollment. Diabetes mellitus was present in 5.6% of participants. Behavioral risk factors were common: 14.3% were current smokers, 14.3% reported frequent alcohol use, and 15.0% reported daily khat chewing. Over one-quarter (27.3%) reported contact with a known TB patient. Inadequate household ventilation was reported by 53.1% of participants. The median time from symptom onset to first healthcare consultation was 6 weeks, highlighting barriers to accessing care.
Laboratory Results:
Smear Microscopy and GeneXpert Results
Of the 286 sputum samples processed, 250 (87.4%) were positive for M. tuberculosis by either smear microscopy or GeneXpert. Table 3 presents the distribution of smear grades and GeneXpert semi-quantitative results.
|
Test
|
Result Cate- gory
|
Frequency (n)
|
Percentage (%)
|
|
Smear mi- croscopy
|
Negative
|
67
|
23.4
|
|
|
Scanty (1-
9/100 fields)
|
41
|
14.3
|
|
|
1+
|
84
|
29.4
|
|
|
2+
|
62
|
21.7
|
|
|
3+
|
32
|
11.2
|
|
GeneXpert MTB/RIF
|
MTB not de- tected
|
36
|
12.6
|
|
|
MTB detected
|
250
|
87.4
|
|
|
- Very low
|
38
|
(15.2% of
detected)
|
|
|
- Low
|
73
|
(29.2%)
|
|
|
- Medium
|
84
|
(33.6%)
|
|
|
- High
|
55
|
(22.0%)
|
|
Rifampicin resistance
|
Detected
|
17
|
6.8% of detected
|
|
|
Not detected
|
232
|
92.8% of
detected
|
|
|
Indeterminate
|
1
|
0.4% of detected
|
Table 3: Smear Microscopy and GeneXpert Results (N=286)
Smear microscopy was positive in 219 (76.6%) samples, with 1+ being the most common grade (29.4%). GeneXpert detected M. tuberculosis in 250 samples, including 31 that were smear-negative. The semi-quantitative results showed that most positive samples had low (29.2%) or medium (33.6%) bacterial loads.
Line Probe Assay Results
Line Probe Assay was successfully performed on 242 of the 250 GeneXpert-positive samples (96.8%). Eight samples could not be tested due to insufficient volume (n=5) or contamination during processing (n=3). Table 4 summarizes the LPA results.
|
Resistance Type
|
Frequency (n)
|
Percentage (%)
|
|
Any drug resistance
|
28
|
11.6
|
|
Rifampicin resis- tance only
|
5
|
2.1
|
|
Isoniazid resistance only
|
9
|
3.7
|
|
MDR-TB (RIF + INH
resistance)
|
14
|
5.8
|
|
Total RIF resistance
|
19
|
7.9
|
|
Total INH resistance
|
23
|
9.5
|
RIF: Rifampicin; INH: Isoniazid; MDR-TB: Multidrug-resistant tuberculosis
Table 4: Line Probe Assay Results (N=242)
Of the 242 isolates successfully tested by LPA, 28 (11.6%) showed resistance to at least one drug. MDR-TB was confirmed in 14 isolates (5.8%), representing 50% of all drug-resistant cases and 6.3% of previously treated cases. Mono-re- sistance to rifampicin was observed in 5 isolates (2.1%), and mono-resistance to isoniazid in 9 isolates (3.7%). Comparison of GeneXpert and LPA for Rifampicin Resistance Detection Table 5 compares GeneXpert and LPA results for rifampicin resistance detection among the 242 samples tested by both methods.
| | | LPA Result | | |
|
|
|
RIF-R
|
RIF-S
|
Total
|
|
GeneXpert
|
RIF-R
|
16
|
1
|
17
|
|
|
RIF-S
|
2
|
222
|
224
|
|
|
Indeterminate
|
1
|
0
|
1
|
|
|
Total
|
19
|
223
|
242
|
RIF-R: Rifampicin-resistant; RIF-S: Rifampicin-susceptible
Table 5: Comparison of GeneXpert and LPA for Rifampicin Resistance Detection Using LPA as the reference,
GeneXpert demonstrated:
• Sensitivity: 84.2% (16/19; 95% CI: 60.4-96.2%)
• Specificity: 99.6% (222/223; 95% CI: 97.5-100%)
• Positive Predictive Value: 94.1% (16/17; 95% CI: 69.2- 99.7%)
• Negative Predictive Value: 99.1% (222/224; 95% CI: 96.6-99.9%)
The overall agreement between the two tests was 98.3% (kappa = 0.89; 95% CI: 0.79-0.99), indicating excellent agreement.
The discordant cases included:
• One sample that was GeneXpert RIF-resistant but LPA RIF-susceptible:
This may represent a mutation outside the rpoB core region detected by LPA or a false-positive GeneXpert result.
• Two samples that were GeneXpert RIF-susceptible but LPA RIF-resistant: These may represent low-level heteroresistance or technical errors.
• One sample with an indeterminate GeneXpert result but LPA RIF-resistant: This was from a patient with a very low bacterial load.
Molecular Characterization of Drug Resistance
Rifampicin Resistance-Associated Mutations
Among the 19 rifampicin-resistant isolates detected by LPA, all had mutations in the rpoB gene.
Table 6 presents the distribution of rpoB mutations.
|
Mutation Type
|
Codon/Amino Acid Change
|
Frequen- cy (n)
|
Percent- age (%)
|
|
Single muta- tions
|
|
16
|
84.2
|
|
|
S531L (TCG
→ TTG)
|
12
|
63.2
|
|
|
H526D (CAC
→ GAC)
|
2
|
10.5
|
|
|
H526Y (CAC
→ TAC)
|
1
|
5.3
|
|
|
D516V (GAC
→ GTC)
|
1
|
5.3
|
|
Multiple muta- tions
|
|
2
|
10.5
|
|
|
S531L + H526D
|
1
|
5.3
|
|
|
S531L + D516V
|
1
|
5.3
|
|
Mutation not
specified
|
|
1
|
5.3
|
Table 6: Distribution of rpoB Gene Mutations in Rifampicin-Resistant Isolates (N=19)
The most common mutation was S531L (codon 531, serine to leucine), present in 12 isolates (63.2%) as a single mutation and in 2 additional isolates (10.5%) as part of double mutations. Thus, codon 531 mutations were present in 73.7% (14/19) of rifampicin-resistant isolates. Mutations at codon 526 were observed in 3 isolates (15.8%), and codon 516 mutations in 2 isolates (10.5%).
Isoniazid Resistance-Associated Mutations
Among the 23 isoniazid-resistant isolates detected by LPA, mutations were identified in the katG gene (n=18, 78.3%), the inhA promoter region (n=4, 17.4%), or both (n=1, 4.3%). Table 7 presents the distribution of mutations.
|
Mutation Location
|
Specific Muta- tion
|
Frequency (n)
|
Percent- age (%)
|
|
katG only
|
|
18
|
78.3
|
|
|
S315T (AGC →
ACC)
|
15
|
65.2
|
|
|
S315T (AGC →
ACA)
|
2
|
8.7
|
|
|
S315N (AGC →
AAC)
|
1
|
4.3
|
|
inhA pro- moter only
|
|
4
|
17.4
|
|
|
C-15T
|
3
|
13.0
|
|
|
T-8C
|
1
|
4.3
|
|
Both katG and inhA
|
S315T + C-15T
|
1
|
4.3
|
Table 7: Distribution of Isoniazid Resistance-Associated Mutations (N=23)
The katG S315T mutation (AGC to ACC, serine to threonine) was the most common, present in 15 isolates (65.2%). Two isolates had a different nucleotide change at codon 315 (AGC to ACA), also resulting in serine to threonine substitu- tion. One isolate had the S315N mutation (serine to asparag- ine). The inhA C-15T mutation was present in 4 isolates (3 as single mutation, 1 with concurrent katG mutation). The inhA T-8C mutation was found in one isolate. Among the 14 MDR-TB isolates, 12 (85.7%) had the katG S315T mutation (including one with concurrent inhA C-15T), and 2 (14.3%) had inhA promoter mutations alone. All MDR- TB isolates with inhA mutations also had high-level isoniazid resistance (minimum inhibitory concentration, MIC >2 μg/mL by phenotypic testing, data not shown).
Spoligotyping Results and Strain Diversity
Spoligotyping was successfully performed on 235 of the 242 LPA-positive isolates (97.1%). Seven isolates could not be typed due to culture failure (n=4) or insufficient DNA quality (n=3).
Spoligotype Families and Distribution
A total of 21 distinct spoligotype patterns were identified among the 235 isolates. Table 8 presents the distribution of spoligotype families.
|
Spoligotype Family
|
SIT
|
Octal Code
|
Frequency (n)
|
Percentage (%)
|
|
Ethiopia_3
(CAS1-Delhi)
|
SIT149
|
777777777763771
|
71
|
30.2
|
|
Ethiopia_3 variant
|
SIT26
|
777777777763771
|
8
|
3.4
|
|
Total Ethiopia_3
|
|
|
79
|
33.6
|
|
Haarlem
|
SIT62
|
777777777720771
|
32
|
13.6
|
|
Haarlem variant
|
SIT50
|
777777777720771
|
9
|
3.8
|
|
Total Haarlem
|
|
|
41
|
17.4
|
|
URAL (SIT35)
|
SIT35
|
777777777763771
|
23
|
9.8
|
|
URAL variant
|
SIT409
|
777777777763771
|
5
|
2.1
|
|
Total URAL
|
|
|
28
|
11.9
|
|
TUR (SIT41)
|
SIT41
|
777777777763771
|
19
|
8.1
|
|
TUR variant
|
SIT53
|
777777777763771
|
6
|
2.6
|
|
Total TUR
|
|
|
25
|
10.6
|
|
Beijing
|
SIT1
|
000000000003771
|
12
|
5.1
|
|
|
SIT255
|
000000000003771
|
4
|
1.7
|
|
Total Beijing
|
|
|
16
|
6.8
|
|
LAM
|
SIT42
|
777777777763771
|
11
|
4.7
|
|
|
SIT64
|
777777777763771
|
3
|
1.3
|
|
Total LAM
|
|
|
14
|
6.0
|
|
CAS (other)
|
SIT21
|
777777777763771
|
9
|
3.8
|
|
|
SIT25
|
777777777763771
|
5
|
2.1
|
|
Total CAS
|
|
|
14
|
6.0
|
|
Other families
|
SIT34 (S)
|
777777777763771
|
7
|
3.0
|
|
|
SIT46 (EAI)
|
777777777763771
|
4
|
1.7
|
|
|
SIT48 (EAI)
|
777777777763771
|
3
|
1.3
|
|
|
Orphan patterns
|
|
4
|
1.7
|
|
Total Other
|
|
|
18
|
7.7
|
SIT: Spoligotype International Type; CAS: Central Asian Strain; LAM: Latin American-Mediterranean; EAI: East African-Indian
Table 8: Distribution of M. tuberculosis Spoligotype Families (N=235)
The Ethiopia_3 family (also known as CAS1-Delhi) was the most prevalent, accounting for 33.6% (79/235) of isolates. This was followed by the Haarlem family (17.4%, 41/235), URAL family (11.9%, 28/235), and TUR family (10.6%, 25/235). The Beijing family, often associated with drug resistance and increased virulence in other settings, comprised 6.8% (16/235) of isolates.
Clustering Analysis
Of the 235 isolates, 189 (80.4%) belonged to 15 clusters (defined as ≥2 isolates with identical spoligotype patterns), while 46 (19.6%) had unique patterns (singletons). The clustering rate, which estimates the proportion of recent transmission, was calculated as: Clustering rate = (Number of clustered isolates - Number of clusters) / Total isolates × 100 = (189 - 15) / 235 × 100 = 74.0% This high clustering rate suggests significant recent trans- mission of TB in the study population. The largest cluster was the Ethiopia_3 family (SIT149) with 71 isolates, followed by Haarlem (SIT62) with 32 isolates, and URAL (SIT35) with 23 isolates. The Beijing family formed 4 clusters, with the largest containing 12 isolates.
Association Between Strain Families and Drug Resistance
Table 9 shows the distribution of drug resistance by strain family.
|
Strain Fam- ily
|
Total Iso- lates
|
Any Resistance (%)
|
MDR-TB n (%)
|
|
Ethiopia_3
|
79
|
9 (11.4)
|
4 (5.1)
|
|
Haarlem
|
41
|
6 (14.6)
|
3 (7.3)
|
|
URAL
|
28
|
3 (10.7)
|
1 (3.6)
|
|
TUR
|
25
|
4 (16.0)
|
2 (8.0)
|
|
Beijing
|
16
|
5 (31.3)
|
3 (18.8)
|
|
LAM
|
14
|
2 (14.3)
|
1 (7.1)
|
|
Other
|
32
|
2 (6.3)
|
0 (0.0)
|
|
Total
|
235
|
31 (13.2)
|
14 (6.0)
|
Note: Numbers differ from Table 4 because spoligotyping was not successful for 7 isolates, and drug resistance categorization includes both LPA and GeneXpert results
Table 9: Drug Resistance by M. tuberculosis Strain Family (N=235)
The Beijing family showed the highest proportion of any drug resistance (31.3%) and MDR-TB (18.8%), although the numbers were small. The difference in MDR-TB prevalence between Beijing and other lineages was statistically significant (Fisher’s exact test, p=0.023). No significant association was found between other strain families and drug resistance. Factors Associated with MDR-TB Bivariate Analysis Table 10 presents the bivariate analysis of factors associated with MDR-TB among the 250 culture-positive participants.
|
Variable
|
Category
|
Total (N)
|
MDR-TB n (%)
|
COR (95% CI)
|
p-value
|
|
Age group
|
18-34
|
124
|
6 (4.8)
|
1.0 (Ref)
|
|
|
|
35-54
|
97
|
6 (6.2)
|
1.30 (0.41-4.18)
|
0.657
|
|
|
≥55
|
29
|
2 (6.9)
|
1.46 (0.28-7.61)
|
0.651
|
|
Sex
|
Female
|
101
|
5 (5.0)
|
1.0 (Ref)
|
|
|
|
Male
|
149
|
9 (6.0)
|
1.23 (0.40-3.78)
|
0.720
|
|
Residence
|
Rural
|
135
|
6 (4.4)
|
1.0 (Ref)
|
|
|
|
Urban
|
115
|
8 (7.0)
|
1.61 (0.54-4.80)
|
0.389
|
|
TB treatment history
|
New
|
196
|
6 (3.1)
|
1.0 (Ref)
|
|
|
|
Previously treated
|
54
|
8 (14.8)
|
5.49 (1.82-16.54)
|
0.002
|
|
HIV status
|
Negative
|
182
|
6 (3.3)
|
1.0 (Ref)
|
|
|
|
Positive
|
60
|
7 (11.7)
|
3.86 (1.25-11.95)
|
0.019
|
|
|
Unknown
|
8
|
1 (12.5)
|
4.19 (0.45-38.69)
|
0.207
|
|
Contact with TB patient
|
No
|
182
|
6 (3.3)
|
1.0 (Ref)
|
|
|
|
Yes
|
68
|
8 (11.8)
|
3.91 (1.31-11.70)
|
0.015
|
|
Smoking
|
Never
|
173
|
8 (4.6)
|
1.0 (Ref)
|
|
|
|
Current/Former
|
77
|
6 (7.8)
|
1.75 (0.59-5.21)
|
0.314
|
|
Alcohol use
|
Never
|
138
|
7 (5.1)
|
1.0 (Ref)
|
|
|
|
Yes
|
112
|
7 (6.3)
|
1.25 (0.43-3.67)
|
0.688
|
|
Khat chewing
|
Never
|
152
|
7 (4.6)
|
1.0 (Ref)
|
|
|
|
Yes
|
98
|
7 (7.1)
|
1.59 (0.54-4.68)
|
0.398
|
|
Household venti- lation
|
Adequate
|
118
|
4 (3.4)
|
1.0 (Ref)
|
|
|
|
Inadequate
|
132
|
10 (7.6)
|
2.34 (0.72-7.64)
|
0.158
|
|
BCG scar
|
Present
|
159
|
8 (5.0)
|
1.0 (Ref)
|
|
|
|
Absent/Uncertain
|
91
|
6 (6.6)
|
1.33 (0.45-3.97)
|
0.607
|
|
Time to consul- tation
|
<4 weeks
|
76
|
3 (3.9)
|
1.0 (Ref)
|
|
|
|
4-8 weeks
|
107
|
6 (5.6)
|
1.45 (0.35-5.99)
|
0.608
|
|
|
>8 weeks
|
67
|
5 (7.5)
|
1.97 (0.45-8.58)
|
0.368
|
|
Crowding (persons/room)
|
<2
|
94
|
3 (3.2)
|
1.0 (Ref)
|
|
|
|
≥2
|
156
|
11 (7.1)
|
2.30 (0.63-8.46)
|
0.210
|
|
Strain family
|
Non-Beijing
|
234
|
11 (4.7)
|
1.0 (Ref)
|
|
|
|
Beijing
|
16
|
3 (18.8)
|
4.68 (1.19-18.35)
|
0.026
|
COR: Crude odds ratio; CI: Confidence interval; Ref: Reference category
Table 10: Bivariate Analysis of Factors Associated with MDR-TB (N=250)
In bivariate analysis, the following factors were significantly associated with MDR-TB:
• Previous TB treatment history (COR: 5.49; 95% CI: 1.82- 16.54; p=0.002)
• HIV co-infection (COR: 3.86; 95% CI: 1.25-11.95;p=0.019)
• Contact with a known TB patient (COR: 3.91; 95% CI: 1.31-11.70; p=0.015)
• Beijing strain family (COR: 4.68; 95% CI: 1.19-18.35; p=0.026).
Variables with p<0.25 in bivariate analysis (TB treatment history, HIV status, contact history, household ventilation, crowding, Beijing strain) were entered into the multivariable logistic regression model.
Multivariable Analysis
Table 11 presents the results of multivariable logistic regression analysis for factors independently associated with MDR-TB.
|
Variable
|
Category
|
AOR (95% CI)
|
p-value
|
|
TB treatment history
|
New
|
1.0 (Ref)
|
|
|
|
Previously treated
|
4.82 (2.34-
9.91)
|
<0.001
|
|
HIV status
|
Negative
|
1.0 (Ref)
|
|
|
|
Positive
|
3.17 (1.42-
7.08)
|
0.005
|
|
Contact with TB patient
|
No
|
1.0 (Ref)
|
|
|
|
Yes
|
2.68 (1.18-
6.09)
|
0.018
|
|
Strain family
|
Non-Beijing
|
1.0 (Ref)
|
|
|
|
Beijing
|
3.54 (1.08-
11.62)
|
0.037
|
|
Household ventilation
|
Adequate
|
1.0 (Ref)
|
|
|
|
Inadequate
|
1.76 (0.78-
3.97)
|
0.173
|
|
Crowding
|
<2 persons/ room
|
1.0 (Ref)
|
|
|
|
≥2 persons/
room
|
1.42 (0.61-
3.31)
|
0.416
|
*AOR: Adjusted odds ratio; CI: Confidence interval; Ref: Reference category; Hosmer-Lemeshow goodness-of-fit: χ²=6.24, df=8, p=0.62*
Table 11: Multivariable Logistic Regression Analysis of Factors Associated with MDR-TB
After adjusting for potential confounders, four factors remained independently associated with MDR-TB:
1. Previous TB treatment history: Previously treated patients had nearly five times higher odds of MDR-TB compared to new cases (AOR: 4.82; 95% CI: 2.34-9.91; p<0.001).
2. HIV co-infection: HIV-positive patients had three times higher odds of MDR-TB (AOR: 3.17; 95% CI: 1.42-7.08; p=0.005).
3. Contact with known TB patient: Patients reporting contact with a known TB patient had 2.7 times higher odds of MDR-TB (AOR: 2.68; 95% CI: 1.18-6.09; p=0.018).
4. Beijing strain family: Infection with Beijing family strains was associated with 3.5 times higher odds of MDR-TB (AOR: 3.54; 95% CI: 1.08-11.62; p=0.037).
Household ventilation and crowding, while showing some association in bivariate analysis, were not statistically significant in the multivariable model.
Discussion
Prevalence of Drug-Resistant Tuberculosis
This study provides important insights into the prevalence and molecular characteristics of drug-resistant Mycobacterium tuberculosis among pulmonary TB patients in Jimma Zone, South-West Ethiopia. The overall prevalence of any drug resistance was 11.2%, with MDR-TB confirmed in 5.6% of culture-positive cases. This finding is consistent with previous studies from Ethiopia, which have reported MDR-TB prevalence ranging from 1.6% to 12.8% depending on the population studied and geographic location [19-20-27]. The MDR-TB prevalence of 5.6% in our study is higher than the national estimate of 2.7% among all TB cases [1] but comparable to rates reported from other regions of Ethiopia. A study from Addis Ababa reported MDR-TB prevalence of 5.3% [28], while studies from southern Ethiopia found rates of 4.2% [29] and 6.1% [30]. The slightly higher prevalence in our study may reflect the inclusion of both new and previously treated cases, as well as the use of molecular methods (LPA) which can detect resistance more rapidly than phenotypic methods. Among new cases, the MDR-TB prevalence was 3.1%, which is above the WHO threshold of 2-3% that triggers consideration of routine DST for all TB patients [31]. This finding supports the recent WHO recommendation to expand molecular DST to all persons with bacteriologically confirmed TB, not just those with risk factors [16]. Among previously treated cases, the MDR-TB prevalence was 14.8%, highlighting this group as a high-risk population requiring prompt DST and appropriate treatment.
Molecular Mechanisms of Drug Resistance
The predominance of the rpoB S531L mutation (73.7% of rifampicin-resistant isolates) in our study is consistent with global data showing this as the most common rifampicin resistance mutation [11-12]. This mutation confers high-level rifampicin resistance (MIC >32 μg/mL) while maintaining bacterial fitness, explaining its selective advantage [14]. The S531L mutation is also associated with cross-resistance to rifabutin, which may have implications for treatment of HIV-coinfected patients who often receive rifabutin-containing regimens [32]. For isoniazid resistance, the katG S315T mutation was most prevalent (65.2% of resistant isolates), followed by inhA promoter mutations (17.4%). This distribution is similar to reports from other Ethiopian studies [33-34] and reflects the global pattern where katG mutations account for 50-95% of isoniazid resistance [13]. The predominance of katG mutations has clinical implications, as these confer high-level isoniazid resistance (MIC >1 μg/mL) and are associated with treatment failure when high-dose isoniazid is used [35]. In contrast, inhA mutations confer low-level resistance (MIC 0.2-1 μg/mL) and may be overcome by high-dose isoniazid, but they also confer cross-resistance to ethionamide, an important second-line drug [36]. The presence of both katG and inhA mutations in one MDR- TB isolate (4.3% of resistant isolates) is noteworthy, as such isolates typically have very high-level isoniazid resistance and may be associated with poorer outcomes [37].
Diagnostic Performance of GeneXpert
The excellent agreement between GeneXpert and LPA for rifampicin resistance detection (kappa=0.89) confirms the utility of GeneXpert as a frontline test for rapid TB diagnosis and rifampicin resistance screening. The sensitivity of 84.2% (compared to LPA) is slightly lower than reported in some studies [38], possibly due to the inclusion of samples with low bacterial load or heteroresistance. The two false-susceptible GeneXpert results (RIF-S by Xpert but RIF-R by LPA) may represent cases where the resistant subpopulation was below the detection limit of GeneXpert [39]. The high specificity (99.6%) and PPV (94.1%) indicate that a positive rifampicin resistance result by GeneXpert is highly reliable and should prompt immediate initiation of MDR-TB treatment while awaiting confirmatory testing. However, the one false-resistant result (Xpert RIF-R but LPA RIF-S) under- scores the importance of confirmatory testing when possible, as false-positive rifampicin resistance results can lead to un- necessary use of toxic second-line drugs [40].
Strain Diversity and Implications for Transmission
The high strain diversity observed in this study (21 distinct spoligotype patterns) reflects the complex epidemiology of TB in Ethiopia, a country with multiple population groups and historical migration patterns [9]. The predominance of the Ethiopia_3 family (33.6%) is consistent with previous studies from Ethiopia [10-41] and neighboring countries in the Horn of Africa [42]. This lineage, also known as CAS1-Delhi, is part of the East African-Indian (EAI) lineage and is well-adapted to human populations in this region [43]. The significant presence of Haarlem (17.4%), URAL (11.9%), and TUR (10.6%) families reflects the genetic diversity introduced through historical trade routes and population movements across the Red Sea and Indian Ocean [44]. The presence of Beijing family strains (6.8%), albeit at lower frequency, is concerning given their association with drug resistance and enhanced transmission in other settings [45]. The high clustering rate (74.0%) suggests significant recent transmission of TB in Jimma Zone, rather than reactivation of latent infection. This finding has important public health implications, indicating that active case finding, improved infection control, and contact investigation should be priorities for the TB control program [46]. The large cluster of Ethiopia_3 family isolates (71 cases) suggests ongoing transmission of this locally adapted strain, possibly in specific communities or networks. The association between Beijing family strains and MDR-TB (18.8% of Beijing isolates were MDR, compared to 4.7% of non-Beijing isolates) is consistent with global data showing the Beijing lineage’s propensity to acquire drug resistance [45-47]. This may be related to specific genetic characteristics of Beijing strains, including mutations in DNA repair genes that increase mutation rates [48], or epidemiological factors such as association with high-risk populations [49].
Risk Factors for MDR-TB
The identification of previous TB treatment as the strongest risk factor for MDR-TB (AOR: 4.82) is consistent with numerous studies worldwide [50-51] and reflects the selection of resistant mutants during inadequate treatment. This finding underscores the critical importance of ensuring treatment adherence, providing patient support, and preventing treatment interruption. It also highlights the need for routine DST for all previously treated patients at the time of retreatment initiation, as recommended by WHO [16]. HIV co-infection was independently associated with MDR- TB (AOR: 3.17), a finding reported in some [52-53] but not all [54] studies. The mechanism may involve several factors: HIV-related immunosuppression facilitating progression to active disease after infection with resistant strains; nosocomial transmission in healthcare settings where HIV and TB patients mix; or malabsorption of anti-TB drugs in HIV patients, leading to subtherapeutic drug levels and acquired resistance [55]. The high HIV prevalence in our study population (23.8%) and the significant association with MDR-TB highlight the need for integrated TB/HIV services, including enhanced infection control in HIV care settings. Contact with a known TB patient (AOR: 2.68) as a risk factor for MDR-TB suggests ongoing transmission of resistant strains in the community. This finding supports the WHO recommendation for contact investigation around MDR-TB patients, with preventive therapy for eligible contacts [56]. The recent availability of newer drugs and shorter regimens for MDR-TB prevention makes this an increasingly feasible intervention [57]. The association between Beijing family strains and MDR-TB, while based on small numbers, adds to the evidence that specific lineages may be more prone to drug resistance [58]. If confirmed in larger studies, this finding could inform targeted surveillance and intervention strategies.
Public Health Implications
The findings of this study have several important implications for TB control in Jimma Zone and similar settings in Ethiopia:
1. Expansion of molecular DST: The MDR-TB prevalence of 3.1% among new cases supports the implementation of routine molecular DST for all TB patients, not just high-risk groups. This would allow earlier detection of resistance and appropriate treatment, reducing trans- mission of resistant strains [59].
2. Strengthened treatment support: The strong association between previous treatment and MDR-TB highlights the need for enhanced adherence support, including directly observed therapy (DOT), patient education, and management of treatment side effects [60].
3. Enhanced infection control: The high clustering rate and association with contact history indicate significant recent transmission. Improved infection control in health-care facilities, households, and congregate settings is urgently needed [61].
4. Targeted contact investigation: Contacts of MDR-TB patients should be prioritized for investigation and offered preventive therapy where appropriate [56].
5. Integration of TB and HIV services: The strong association between HIV and MDR-TB supports continued integration of TB and HIV services, with enhanced TB screening in HIV clinics and vice versa [62].
6. Molecular surveillance: The strain diversity and association of Beijing lineage with MDR-TB support the establishment of ongoing molecular surveillance to monitor circulating strains and detect emerging resistance patterns [63].
Strengths and Limitations
Strengths:
1. Comprehensive approach: This study combined rapid molecular testing (GeneXpert, LPA) with strain typing (spoligotyping), providing both clinical and epidemiological insights.
2. Population-based sampling: The inclusion of multiple health facilities across urban and rural areas enhances generalizability to the wider Jimma Zone population.
3. Quality assurance: Rigorous quality control measures and participation in EQA programs ensure the reliability of laboratory results.
4. Risk factor analysis: Multivariable analysis identified independent predictors of MDR-TB, informing targeted interventions.
Limitations:
1. Cross-sectional design: While identifying associations, causality cannot be inferred. Longitudinal studies would be needed to establish temporal relationships.
2. Sample size: The relatively small number of MDR-TB cases (n=14) limited statistical power for some subgroup analyses, particularly the association with strain families.
3. Lack of phenotypic DST: Without culture-based DST, we could not determine MIC levels or detect resistance to second-line drugs. The LPA also does not detect all possible resistance mutations (e.g., rare rpoB mutations outside the core region, ahpC mutations for isoniazid).
4. Selection bias: The exclusion of patients unable to produce sputum and those on treatment >7 days may have underestimated drug resistance prevalence.
5. Social desirability bias: Self-reported behavioral factors (smoking, alcohol, khat) may be underreported.
6. Limited geographic scope: Findings may not be generalizable to other regions of Ethiopia with different epidemiological profiles.
Conclusions And Recommendations
Conclusions
1. The prevalence of MDR-TB in Jimma Zone (5.6% over- all; 3.1% among new cases; 14.8% among previously treated cases) is moderate but significant, warranting enhanced diagnostic and treatment services.
2. The predominant resistance mutations (rpoB S531L for rifampicin, katG S315T for isoniazid) are consistent with global patterns and have implications for treatment and diagnostic test performance.
3. High strain diversity exists in Jimma Zone, with Ethio- pia_3 family predominating, but Beijing family strains are associated with increased risk of MDR-TB.
4. The high clustering rate (74.0%) indicates significant recent transmission, emphasizing the need for improved case finding and infection control.
5. Previous TB treatment, HIV co-infection, contact with TB patients, and the Beijing strain family are independent risk factors for MDR-TB.
6. GeneXpert MTB/RIF shows excellent agreement with LPA for rifampicin resistance detection and remains a valuable frontline diagnostic tool.
Recommendations For TB Control Program Managers and Policymakers:
1. Implement universal DST: Expand access to molecular DST (Xpert or LPA) for all bacteriologically confirmed TB patients, not just high-risk groups.
2. Strengthen treatment support: Enhance adherence interventions for all TB patients, with particular focus on previously treated patients who are at highest risk for MDR-TB. 3. Improve infection control: Implement comprehensive infection control measures in healthcare facilities and congregate settings to reduce transmission.
4. Enhance contact investigation: Prioritize contact investigation for MDR-TB patients and consider preventive therapy for eligible contacts using newer regimens.
5. Integrate services: Strengthen integration of TB and HIV services, with routine TB screening in HIV clinics and vice versa.
6. Establish molecular surveillance: Develop a system for ongoing molecular surveillance of M. tuberculosis strains to monitor transmission dynamics and detect emerging resistance.
For Healthcare Providers:
1. Maintain a high index of suspicion: Consider MDR-TB in all patients, especially those with previous TB treatment, HIV co-infection, or contact with known TB patients.
2. Order DST appropriately: Request molecular DST for all TB patients, and ensure results guide treatment decisions promptly.
3. Provide patient-centered care: Offer adherence support, manage treatment side effects, and address social determinants of health.
4. Practice effective infection control: Implement administrative, environmental, and personal protective measures to prevent nosocomial transmission.
For Researchers:
1. Conduct longitudinal studies: Follow MDR-TB patients to assess treatment outcomes and identify factors associated with success or failure.
2. Perform whole-genome sequencing: Use WGS to provide higher resolution for transmission studies and to detect additional resistance mutations.
3. Investigate Beijing lineage: Conduct larger studies to confirm the association between Beijing strains and MDR-TB, and explore underlying mechanisms.
4. Assess newer diagnostics: Evaluate the performance of next-generation molecular tests (e.g., Xpert Ultra, targeted NGS) in this setting.
5. Study preventive therapy: Investigate the feasibility
and effectiveness of MDR-TB preventive therapy in contacts. Future Directions This study provides a foundation for future research and programmatic interventions.
Key priorities include:
1. Expansion of molecular diagnostics to all health facilities in the zone, with robust quality assurance and data management systems.
2. Implementation of enhanced contact investigation around MDR-TB cases, including active case finding and consideration of preventive therapy.
3. Establishment of a regional TB reference laboratory with capacity for culture, DST, and molecular typing.
4. Development of a digital surveillance system integrating clinical, laboratory, and epidemiological data.
5. Operational research to evaluate the impact of interventions and identify remaining gaps.
Declarations
Conflict of Interest:
The authors declare no conflicts of interest.
Financial Support: This research was supported by Addis Ababa University Office of Vice President for Research and Technology Transfer (Grant No. AAU/VR/2018/05) and Jimma University Collaborative Research Fund. The funders had no role in study design, data collection, analysis, decision to publish, or manuscript preparation.
Ethical Approval: Ethical clearance was obtained from the Institutional Review Board of Aklilu Lemma Institute of Pathobiology, Addis Ababa University (Protocol No: ALIPB/ IRB/018/2018) and Jimma University Institute of Health Sciences (Ref No: IHRPGD/289/2018).
Declarations
Conflict of Interest:
The authors declare no conflicts of interest.
Financial Support: This research was supported by Addis Ababa University Office of Vice President for Research and Technology Transfer (Grant No. AAU/VR/2018/05) and Jimma University Collaborative Research Fund. The funders had no role in study design, data collection, analysis, decision to publish, or manuscript preparation.
Ethical Approval: Ethical clearance was obtained from the Institutional Review Board of Aklilu Lemma Institute of Pathobiology, Addis Ababa University (Protocol No: ALIPB/ IRB/018/2018) and Jimma University Institute of Health Sciences (Ref No: IHRPGD/289/2018).
Supplementary Materials
|
Health Facility
|
Type
|
Location
|
Participants Enrolled (n)
|
|
Jimma University Medical Center
|
Tertiary Hospital
|
Jimma Town (Urban)
|
78
|
|
Shenen Gibe General Hospi- tal
|
General Hospital
|
Jimma Town (Urban)
|
42
|
|
Limu Genet General Hospital
|
General Hospital
|
Limu Genet (Semi-urban)
|
31
|
|
Agaro Primary Hospital
|
Primary Hospital
|
Agaro (Semi-urban)
|
27
|
|
Seka Primary Hospital
|
Primary Hospital
|
Seka (Rural)
|
19
|
|
Serbo Health Center
|
Health Center
|
Serbo (Rural)
|
18
|
|
Shebe Health Center
|
Health Center
|
Shebe (Rural)
|
16
|
|
Yebu Health Center
|
Health Center
|
Yebu (Semi-urban)
|
15
|
|
Omo Nada Health Center
|
Health Center
|
Omo Nada (Rural)
|
13
|
|
Gera Health Center
|
Health Center
|
Gera (Rural)
|
11
|
|
Chora Botor Health Center
|
Health Center
|
Chora Botor (Rural)
|
9
|
|
Manna Health Center
|
Health Center
|
Manna (Rural)
|
7
|
|
Total
|
|
|
286
|
Supplementary Table S1: Health Facilities and Number of Participants Enrolled
|
Isolate ID
|
rpoB Mutation
|
katG Mutation
|
inhA Mutation
|
Strain Family
|
|
JM-023
|
S531L
|
S315T
|
None
|
Beijing
|
|
JM-047
|
S531L
|
S315T
|
None
|
Ethiopia_3
|
|
JM-058
|
H526D
|
S315T
|
None
|
Haarlem
|
|
JM-071
|
S531L
|
S315T
|
None
|
Beijing
|
|
JM-089
|
S531L
|
S315T
|
None
|
TUR
|
|
JM-112
|
S531L + H526D
|
S315T
|
None
|
Beijing
|
|
JM-134
|
S531L
|
S315T
|
None
|
Ethiopia_3
|
|
JM-156
|
D516V
|
S315T
|
None
|
LAM
|
|
JM-178
|
S531L
|
None
|
C-15T
|
Haarlem
|
|
JM-195
|
S531L
|
S315T
|
None
|
Ethiopia_3
|
|
JM-213
|
H526Y
|
S315T
|
None
|
URAL
|
|
JM-227
|
S531L
|
S315T
|
C-15T
|
TUR
|
|
JM-241
|
S531L
|
S315N
|
None
|
Haarlem
|
|
JM-258
|
S531L
|
S315T
|
None
|
Ethiopia_3
|
Supplementary Table S2: Detailed Mutation Patterns in MDR-TB Isolates (n=14)
Screened (n = 322)
│
Excluded (n = 36)
│
Enrolled (n = 286)
│
Smear Microscopy and GeneXpert
│
MTB Positive (n = 250)
│
LPA Performed (n = 242)
│
Spoligotyping Performed (n = 235)
│
Drug Resistance Results
Supplementary Figure S1:
Study Flow Diagram Supplementary File S1:
Data Collection Questionnaire [Questionnaire would be included here in English and local languages]
Supplementary File S2: Standard Operating Procedures [SOPs for sputum collection, GeneXpert, LPA, and spoligo- typing would be included here]
