Clinical and Microbiological Characteristics Associated with Treatment Outcomes Among Rifampicin-Resistant Tuberculosis Patients in Limpopo Province

Introduction

Tuberculosis (TB) remains a major global public health challenge and the leading cause of death from a single infectious agent worldwide. Sub-Saharan Africa carries a disproportionate burden of disease, with South Africa ranking among the 30 high TB burden countries globally. In 2023, South Africa reported approximately 270,000 TB cases, including nearly 13,000 cases of rifampicin-resistant or multidrug-resistant tuberculosis (RR/MDR-TB). Despite advances in diagnostics and treatment, outcomes for DR-TB remain suboptimal because of delayed diagnosis, prolonged treatment regimens, drug toxicity, inconsistent adherence, and health system constraints.^1–3

Limpopo Province presents distinct epidemiological dynamics that justify focused investigation. The province is predominantly rural, with limited access to specialised diagnostic services and DR-TB treatment centres. It shares international borders with Botswana, Zimbabwe, and Mozambique, functioning as a major transit corridor for regional migration. Literature has shown that high population mobility contributes to delayed diagnosis, treatment interruption, and amplification of resistant strains. These structural and cross-border dynamics create a complex transmission environment that may differ substantially from national averages.^4,5

Sociodemographic and clinical factors, including age, sex, occupation, and HIV status, significantly shape TB risk, disease progression, and treatment outcomes. Literature has consistently demonstrated that TB disproportionately affects individuals within the economically active age group (15–44 years), with men more frequently affected than women. HIV co-infection remains a major driver of both incident TB and reactivation of latent infection, particularly in high-burden settings.⁶ In provinces such as Limpopo, where HIV prevalence and population mobility intersect, these factors may further influence DR-TB patterns and clinical outcomes.

Despite ongoing national surveillance efforts, province-specific data describing demographic, clinical, and microbiological factors associated with treatment outcomes among RR-TB patients remain limited in rural South African settings. Understanding local determinants of RR-TB treatment outcomes is essential for strengthening decentralized TB control strategies and optimizing integrated TB–HIV care in high-burden communities.^4,6 This study aimed to describe the demographic, clinical, and microbiological characteristics of rifampicin-resistant tuberculosis patients and evaluate factors associated with treatment outcomes in Limpopo Province, South Africa.

Addressing this evidence gap is critical for informing province-specific TB control strategies and strengthening implementation of South Africa’s National TB Recovery Plan. Generating locally relevant data on demographic distribution, clinical characteristics, and microbiological resistance patterns is essential for guiding targeted interventions and enhancing surveillance within a high-burden, resource-limited setting. Despite national surveillance, limited data exist on factors associated with treatment outcomes among DR-TB patients in rural South African settings. Understanding these determinants is critical for risk stratification and targeted interventions. Therefore, this study aimed to describe the demographic, clinical, and microbiological characteristics of rifampicin-resistant tuberculosis patients and evaluate factors associated with treatment outcomes in Limpopo Province, South Africa.

Methodology

Study Area

Limpopo Province, located in northern South Africa, is the country’s fifth most populous province, accounting for approximately 10% of the national population of over 60 million people.⁷ Predominantly rural, the province faces persistent disparities in access to specialized diagnostic services and DR-TB treatment centres, which may influence case detection and clinical outcomes.⁸ Recent analyses highlight that DR-TB continues to threaten TB control efforts in resource-limited provinces such as Limpopo, where weak laboratory infrastructure and limited genomic surveillance capacity hinder timely detection of resistance patterns.^4,6 In addition, high levels of cross-border migration with Botswana, Zimbabwe, and Mozambique contribute to population mobility and treatment interruption, further amplifying resistant strains.⁴

Ethical Approval

Ethical approval for this study was obtained from the Turfloop Research Ethics Committee (TREC) of the University of Limpopo (Reference: TREC/108/2024: PG), the National Health Laboratory Service (NHLS) (Reference: PR2456259), and the Limpopo Department of Health (Reference: LP_2025-03-011). Permission to access routine programmatic data was granted by the relevant provincial health authorities.

The requirement for informed consent was waived by all approving bodies, as the study utilized routinely collected laboratory and Department of Health data with no direct patient interaction. All data were fully anonymized prior to analysis to ensure confidentiality and privacy. The study was conducted in accordance with the principles of the Declaration of Helsinki, the South African National Health Act (No. 61 of 2003), and national Good Clinical Practice guidelines.

Study Site and Population

This study was conducted in Limpopo Province, South Africa, and included rifampicin-resistant tuberculosis (RR-TB) patients diagnosed between January 2023 and July 2025. Study samples were collected at National Health Laboratory Service (NHLS) Polokwane Microbiology laboratory.

Study Design

A retrospective cohort study was conducted to assess demographic, clinical, and microbiological characteristics associated with treatment outcomes among RR-TB patients in Limpopo Province, South Africa.

Data Collection

Data were compiled from complementary laboratory and routine programmatic sources. Microbiological data were obtained from the NHLS, while demographic, clinical, treatment, and treatment outcome data were sourced from Department of Health routine programmatic records. The variables collected included patient age, sex, HIV status, type of tuberculosis, drug resistance profile, treatment history, treatment regimen, and final treatment outcome. Treatment outcomes were defined according to national tuberculosis programme guidelines and categorized as cure, treatment completion, death, treatment failure, or loss to follow-up.

Inclusion and Exclusion Criteria

Eligible participants included all patients diagnosed with laboratory-confirmed DR-TB, between 1 January 2023 and 31 July 2025 in Limpopo Province. Cases were included if microbiological confirmation was available from NHLS Polokwane microbiology laboratory and corresponding demographic, clinical, treatment, and outcome data were available from Department of Health routine programmatic records. Patients with incomplete or missing key variables required for analysis (eg., demographic, clinical, microbiological, or treatment outcome data) or with unspecified TB classification were excluded.

Definitions

Treatment outcomes were classified according to national DR-TB treatment outcome definitions. Favourable outcomes included treatment completed, while unfavourable outcomes included death and loss to follow-up. Patients who were still on treatment at the time of data extraction were retained for descriptive analyses but excluded from regression analyses assessing predictors of treatment outcomes. Rifampicin-resistant tuberculosis was defined as tuberculosis caused by Mycobacterium tuberculosis resistant to rifampicin, with or without resistance to additional anti-tuberculosis drugs.

Data Analysis

Data were cleaned and coded using Microsoft Excel and subsequently exported to IBM SPSS Statistics version 30.0 for statistical analysis.

Descriptive statistics were used to summarize demographic and clinical characteristics. Categorical variables were expressed as frequencies and percentages. Associations between categorical variables were assessed using the chi-square test; where the assumption of expected cell counts ≥5 was violated, Fisher’s exact test was applied. A p-value <0.05 was considered statistically significant, and effect sizes (Cramer’s V) were reported to quantify the strength of associations.

Kaplan–Meier survival analysis was used to estimate time to death among DR-TB patients. Time-to-event was calculated as days from treatment initiation to death, with treatment completion, cure, or ongoing treatment treated as censoring events. Survival curves were compared using the Log rank test and stratified by HIV status. Kaplan–Meier survival analysis was performed as an exploratory descriptive approach to visualize differences in survival by HIV status.

To identify factors independently associated with unfavourable treatment outcomes, multivariable logistic regression analysis was performed. Patients who were still on treatment at the time of analysis were excluded from the regression model. Crude and adjusted odds ratios (ORs) with 95% confidence intervals (CIs) were reported. The adjusted model included key covariates such as age group, sex, HIV status, and smear status. Variables included in the adjusted model were selected based on clinical relevance and univariate associations. Complete-case analysis was performed for regression modeling, and no imputation procedures were undertaken because of the retrospective nature of the dataset and limited sample size. All analyses were two-tailed, and statistical significance was set at p <0.05. Reporting was conducted in accordance with the STROBE guidelines.

Results

Characteristics of DR-TB Patients

Among the 86 patients included in the study, most were aged 25–44 years (65.1%), and males accounted for 70.9% of cases (Table 1). Cases were distributed across all five districts of Limpopo Province, with the highest proportions observed in Capricorn (29.1%) and Mopani (26.7%). HIV co-infection was present in 55.8% of patients. All isolates were confirmed as Mycobacterium tuberculosis, with culture positivity observed in 98.8% of cases. Smear microscopy demonstrated varying bacillary loads, with 2+ smear positivity being the most common (48.8%). Drug susceptibility testing identified resistance to multiple first- and second-line anti-tuberculosis drugs, including isoniazid (46.5%), fluoroquinolones (19.8%), ethionamide (8.1%), amikacin (1.2%), bedaquiline (16.3%), and linezolid (1.2%).

Table 1 Demographic, Clinical, and Microbiological Characteristics of DR-TB Patients in Limpopo (2023–2025)

Treatment Outcomes of DR-TB Patients

Among the 86 RR-TB patients with available treatment outcome data, favourable outcomes (treatment completion) were observed in 62 (71.4%) patients. Unfavourable outcomes (death or loss to follow-up) occurred in 12 (14.3%) patients, while 12 (14.3%) patients were still on treatment at the time of analysis and were treated as censored observations.

Factors Associated with Treatment Outcomes

Demographic and Clinical Factors

In univariate analysis (Table 1), HIV co-infection was significantly associated with treatment outcomes (χ² = 9.7, p = 0.002), with a higher proportion of unfavourable outcomes observed among HIV-positive patients. Smear grade was also associated with treatment outcomes (χ² = 5.4, p = 0.04). Age group and sex were not significantly associated with treatment outcomes in univariate analysis. Effect size estimates based on Cramer’s V indicated predominantly small-to-moderate associations between categorical variables and treatment outcomes However, in multivariable logistic regression analysis (Table 2), only age 15–24 years (AOR = 9.97, 95% CI: 1.48–67.33, p = 0.018) and smear-negative status (AOR = 39.62, 95% CI: 3.33–471.83, p = 0.004) remained independently associated with unfavourable outcomes. HIV status and sex were not significantly associated after adjustment.

Table 2 Logistic Regression Analysis of Factors Associated with Unfavourable Treatment Outcomes Among DR-TB Patients

Microbiological and Resistance Factors

Among the included patients, MDR-TB, pre-XDR-TB, and XDR-TB phenotypes were identified based on additional resistance patterns. (Table 1). In univariate analysis, resistance profile was not significantly associated with treatment outcomes.

Isoniazid resistance was observed in 46.5% of isolates, while fluoroquinolone resistance was present in 19.8%. Resistance to bedaquiline was identified in 16.3% of cases. No statistically significant associations were observed between individual drug resistance patterns and treatment outcomes in univariate analysis.

In the multivariable logistic regression model (Table 2), microbiological resistance variables were not independently associated with unfavourable treatment outcomes.

Smear microscopy grading showed that 2+ smear positivity was the most frequent (48.8%), followed by 1+ (19.8%) and 3+ (15.1%). In univariate analysis, smear grade was significantly associated with treatment outcomes; however, in the multivariable model, smear-negative status (compared with other smear grades) remained independently associated with unfavourable outcomes.

Multivariable Analysis of Predictors of Unfavourable Outcomes

Multivariable logistic regression was performed to identify factors independently associated with unfavourable treatment outcomes among DR-TB patients, excluding those still on treatment at the time of analysis. In the adjusted model, patients aged 15–24 years had significantly higher odds of experiencing an unfavourable outcome compared with older age groups (AOR = 9.97, 95% CI: 1.48–67.33, p = 0.018). Similarly, smear-negative patients had markedly increased odds of unfavourable outcomes compared with non-smear-negative patients (AOR = 39.62, 95% CI: 3.33–471.83, p = 0.004). This should be interpreted more cautiously because estimates are imprecise. Female sex (AOR = 1.78, 95% CI: 0.36–8.84, p = 0.480) and HIV-positive status (AOR = 0.61, 95% CI: 0.12–3.00, p = 0.542) were not significantly associated with unfavourable outcomes after adjustment. The wide confidence intervals observed for some predictors indicate limited precision of the estimates, likely reflecting the relatively small number of unfavourable outcomes. These estimates should be interpreted cautiously because of the relatively small number of unfavourable outcomes and corresponding wide confidence intervals.

Survival Analysis by HIV Status

Kaplan–Meier survival analysis demonstrated differences in survival between HIV-positive and HIV-negative DR-TB patients (Figure 1), with HIV-positive patients exhibiting a lower probability of survival over time compared with HIV-negative patients.

Figure 1 Kaplan–Meier survival curves comparing HIV-positive and HIV-negative drug-resistant TB patients in Limpopo Province (2023–2025).

The difference in survival distributions between the two groups was statistically significant (Log rank test, p = 0.002). However, this association was not retained in the multivariable logistic regression model (Table 2). However, HIV status was not significantly associated with unfavourable outcomes in the adjusted logistic regression model.

Discussion

This province-wide analysis from Limpopo Province provides a comprehensive assessment of the demographic, clinical, and microbiological characteristics of drug-resistant tuberculosis (DR-TB). The study identified a predominance of RR-TB, a concentration of cases among working-age males, and a high prevalence of TB–HIV co-infection. These findings are consistent with national trends in South Africa and global estimates reported by the World Health Organization (WHO), while also reflecting context-specific transmission dynamics that warrant targeted public health interventions.^4,8,9 The predominance of male patients is consistent with previous studies reporting delayed healthcare-seeking behaviour, occupational mobility, treatment interruption and loss to follow-up among men; however, the present study was not sufficiently powered to robustly assess outcome-specific differences such as mortality versus loss to follow-up.

Rifampicin resistance remains a major programmatic concern because of its association with multidrug-resistant tuberculosis and its central role in frontline molecular diagnostics. The identification of additional MDR-TB, pre-XDR-TB, and XDR-TB phenotypes within this cohort highlights the ongoing evolution of resistance patterns in Limpopo Province and reinforces the importance of sustained resistance surveillance.^10,11 Rifampicin’s central role in first-line therapy generates sustained selective pressure, particularly in contexts of incomplete adherence, drug stock-outs, or delayed resistance detection.⁹ In addition, frontline molecular diagnostics such as Xpert MTB/RIF prioritize rifampicin resistance detection and operationally classify RR-TB as a proxy for MDR-TB, increasing its reported frequency.¹² Provincial genomic analyses similarly emphasize rifampicin resistance as the dominant phenotype in Limpopo, supporting the interpretation that primary transmission of RR strains substantially contributes to the local DR-TB burden.⁸

Demographically, DR-TB disproportionately affected males (70.9%) and individuals aged 25–44 years (65.1%), consistent with WHO global epidemiology indicating peak TB incidence among working-age adults.⁹ Male predominance has been associated with occupational exposure, labour migration, smoking, alcohol use, and delayed healthcare-seeking behaviour.^4,13 A similar male-skewed distribution of DR-TB has been reported in regional surveillance data, reinforcing the influence of gendered social determinants. The underrepresentation of children and older adults likely reflects diagnostic challenges in paediatric TB and healthcare access barriers among elderly populations.¹⁴

The resistance profile raises important programmatic considerations. High isoniazid resistance (46.5%) reflects global trends of increasing INH mono-resistance and its contribution to MDR evolution.^9,15 Fluoroquinolone resistance (19.8%) is concerning given the central role of this class in MDR-TB regimens. Encouragingly, most isolates remained susceptible to bedaquiline and linezolid, which are classified as Group A drugs in the current WHO DR-TB treatment guidelines.⁹ Low resistance to these agents supports the continued effectiveness of all-oral regimens in this setting. Nonetheless, emerging reports of bedaquiline resistance from South Africa and other high-burden countries highlight the need for continuous drug resistance surveillance to preserve the effectiveness of these critical agents.^8,16 The observed prevalence of bedaquiline resistance is clinically concerning given the central role of bedaquiline in contemporary DR-TB treatment regimens. However, the study design could not distinguish between primary transmission of resistant strains and acquired resistance resulting from prior drug exposure, treatment interruption, or regimen-related factors. These findings nevertheless underscore the importance of ongoing resistance surveillance and expanded access to drug susceptibility testing for newer anti-tuberculosis agents. Although resistance to amikacin and linezolid remained low, vigilance remains essential as expanded use of newer agents may increase selective pressure over time. Limited pretomanid susceptibility data reflect laboratory capacity constraints rather than confirmed susceptibility patterns. The high culture positivity rate (98.8%) strengthens confidence in bacteriological confirmation and laboratory performance within the province.

Smear microscopy grading showed a predominance of 2+ smear positivity, indicating moderate to high bacillary loads among patients at diagnosis. Smear grade is widely used as a proxy indicator of bacterial burden and potential infectiousness.^17,18 The predominance of 2+ smear grades is consistent with reports from other high-burden TB settings and suggests that many patients present with substantial bacillary loads at the time of diagnosis. The relatively lower frequency of smear-negative or scanty results may reflect the limited sensitivity of smear microscopy in detecting paucibacillary disease, particularly among HIV-positive individuals.¹⁹ The association was observed as imprecise estimates that may reflect delayed diagnosis, atypical disease presentation, or paucibacillary disease rather than a direct biological relationship.

The survival analysis highlighted differences in outcomes by HIV status, with Kaplan–Meier curves demonstrating significantly poorer survival among HIV-positive patients compared with HIV-negative individuals. However, this association was not retained in the multivariable logistic regression model, suggesting that the observed differences in survival may be influenced by confounding factors rather than HIV status alone. In contrast, younger age (15–24 years) and smear-negative status were independently associated with unfavourable treatment outcomes and should be interpreted more cautiously because estimates are imprecise. The increased risk observed among younger patients may reflect challenges related to treatment adherence, health-seeking behaviour, or delayed engagement with care. Similarly, smear-negative disease may be indicative of atypical or advanced disease presentation, particularly in immunocompromised individuals, which may contribute to delayed diagnosis and poorer clinical outcomes.^20–22 Although HIV-positive patients demonstrated poorer survival in the Kaplan–Meier analysis, HIV status was not independently associated with unfavourable outcomes in the adjusted logistic regression model. These approaches address related but distinct analytical questions and may therefore yield different findings. The relatively small number of unfavourable outcomes may also have limited the precision of adjusted estimates. These findings highlight the importance of considering multiple interacting determinants when evaluating treatment outcomes in drug-resistant tuberculosis.^6,23–25

Overall, this study provides important insights into the epidemiology and resistance patterns of DR-TB in Limpopo Province. The findings highlight the continued predominance of RR-TB, the high burden of TB–HIV co-infection, and the presence of resistance to both first-line and second-line anti-tuberculosis drugs. While microbiological resistance patterns were not independently associated with treatment outcomes, clinical factors such as younger age and smear-negative disease were associated with unfavourable outcomes. This association may reflect delayed diagnosis, atypical disease presentation, or paucibacillary disease rather than a direct biological relationship. Interpretation of some associations should be undertaken cautiously because the relatively small number of unfavourable outcomes contributed to wide confidence intervals and limited precision of certain regression estimates.

These findings underscore the need to strengthen integrated TB-HIV services, enhance early diagnosis, particularly in atypical presentations and expand access to comprehensive drug susceptibility testing. Sustained surveillance for emerging resistance, alongside targeted interventions for high-risk patient groups, will be critical for improving DR-TB control in this high-burden setting.

Strengths

This study has several strengths. It provides province-wide, laboratory-confirmed data, offering comprehensive regional coverage rather than isolated district-level analyses. The high culture confirmation rate enhances the reliability of microbiological findings. The inclusion of both first- and second-line drug susceptibility testing enables detailed characterization of resistance patterns. Integration of demographic and clinical variables allows for contextual epidemiological interpretation of DR-TB within the province. Furthermore, the consistency of these findings with prior provincial genomic studies⁸ supports the internal validity of the observed resistance patterns.

Limitations

This study has several limitations. The relatively small sample size reflects the number of DR-TB patients registered in the provincial programme during the study period and may limit the statistical power to detect smaller associations. As a retrospective, laboratory-based study, detailed clinical variables—including prior treatment history, adherence patterns, and socioeconomic status—were not available. Incomplete second-line drug susceptibility data, particularly for pretomanid, limit the comprehensive interpretation of emerging resistance patterns. The absence of routine whole-genome sequencing restricts the ability to differentiate between primary transmission and acquired resistance. Referral bias may have led to an overrepresentation of bacteriologically confirmed or more severe cases, while underestimating paucibacillary disease. Smear microscopy is subject to inter-observer variability, and the cross-sectional nature of the data limits causal inference regarding transmission dynamics and treatment outcomes. Additionally, the wide confidence intervals observed in the multivariable analysis indicate limited precision of some estimates, likely due to the relatively small sample size and low number of unfavourable outcomes.

Conclusion

The demographic and clinical profile of RR-TB patients in this study highlights a concentration of disease among economically active men, suggesting possible ongoing transmission within the community, although transmission dynamics could not be directly assessed. Strengthening decentralized molecular diagnostics, expanding genomic surveillance, and integrating TB–HIV services remain essential for improving early detection and treatment outcomes. Importantly, clinical factors particularly younger age and smear-negative disease - were associated with unfavourable outcomes, underscoring the need for risk-stratified patient management and improved recognition of atypical disease presentations. These findings provide actionable, context-specific evidence to support the strengthening of DR-TB diagnostics, surveillance systems, and integrated TB–HIV care in rural, high-burden settings in sub-Saharan Africa.