Paediatric tuberculosis treatment outcomes: a retrospective study at the Komfo Anokye teaching hospital in Ghana, 2014–2025

Abstract

Introduction

Tuberculosis (TB) remains a leading cause of morbidity and mortality among children, particularly those living with HIV in sub-Saharan Africa. Despite advances in integrating TB and HIV care in Ghana, long-term data on pediatric treatment outcomes across various age groups and nutritional statuses remain limited. This study aimed to determine TB treatment outcomes among children aged 0 to 14 years and assess the association of HIV co-infection and other factors on unsuccessful treatment outcomes.

Methods

A retrospective study was conducted using data from the Pediatric Tuberculosis Register at the Komfo Anokye Teaching Hospital between January 2014 and October 2025. All children included in this study were aged 0–14 years, confirmed with TB and on treatment were included. TB cases were clinically diagnosed, radiologically or bacteriologically confirmed. Treatment outcomes were classified according to the guidelines of the National TB Control Programme. Modified Poisson regression models with robust standard errors identified factors associated with unsuccessful outcomes, expressed as adjusted relative risk (ARR) with 95% confidence intervals (CI) using R statistical software version 4.4.1.

Results

Among 1,028 children analyzed, males were 56.8% [584/1028] and had a median age of 6.0 years (IQR 3–10). HIV prevalence stood at 53.2% [545/1,025]. Overall treatment success was 83.4% (857/1,028), with 16.6% [171/1,028] experiencing unsuccessful outcomes—comprising 8.9% [92/1,028] deaths and 7.7% [78/1,028] defaults. Treatment success rates ranged from 64.4% [56/87] in 2016 to 96.5% [110/114] in 2022. Multivariable analysis identified positive HIV status (ARR = 1.39, 95% CI 1.01–1.93, p = 0.044) and increasing age (ARR = 1.06 per year, 95% CI 1.02–1.10, p = 0.002) as significant predictors of unsuccessful treatment. Each additional year of age increased the risk of unsuccessful treatment by 8% (ARR = 1.08, 95% CI 1.04–1.12, p < 0.001). For underweight children with HIV, this interaction effect rose to 20% (ARR = 1.20, 95% CI 1.07–1.34, p = 0.002).

Conclusions

The pediatric TB treatment success rate at KATH still falls short of the WHO target of 90%. Older age and HIV co-infection significantly increase the risk of unsuccessful treatment, with underweight children facing a greater risk. Strengthening integrated pediatric TB/HIV services through improved age-related enhanced adherence and nutritional support for adequate weight gain is crucial for improving outcomes and reducing default and mortality rates among children with TB in Ghana.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12887-026-06971-8.

Introduction

Tuberculosis (TB) is a leading cause of death from infectious diseases worldwide [1], with children in sub-Saharan Africa being particularly affected [1, 2]. The World Health Organization (WHO) estimates that over one million children develop TB annually, resulting in more than 200,000 deaths [1]. Many of these deaths are preventable through prompt diagnosis and effective treatment [1]. The burden of childhood TB is especially high in sub-Saharan Africa, where factors such as HIV, poverty, overcrowding, and malnutrition converge to increase transmission and hinder management [3]. Childhood TB often manifests as a paucibacillary disease with non-specific symptoms that resemble those of other common childhood illnesses, leading to delays in diagnosis and missed cases [3, 4]. Nearly two-thirds of childhood TB cases in Africa might remain undiagnosed, creating a large reservoir for ongoing community transmission and substantial childhood morbidity [3, 4]. Children under 15 years constitute an estimated 5% of all cases in Ghana [5], though this figure is likely a significant underestimate due to diagnostic challenges. At the Komfo Anokye Teaching Hospital (KATH), a major urban tertiary hospital, the estimated prevalence is 28% [6] .

The Human Immunodeficiency Virus (HIV) epidemic has profoundly altered the TB landscape, increasing the risk of progression from latent TB infection to active disease and complicating clinical presentation, diagnosis, and TB treatment outcomes [7]. Recent studies have demonstrated that individuals with HIV and TB coinfection had approximately two [8] to five times [9] the risk of death compared to persons without HIV. For children, HIV-induced immunosuppression not only increases susceptibility to TB but also increases disease severity, reduces treatment adherence and worsens outcomes [3].

Ghana has made commendable progress in integrating TB and HIV care through differentiated services at treatment centres; however, a previous study at another teaching hospital in Accra, Ghana, reported an estimated 17.4% mortality rate among pediatric TB patients [8]. Nonetheless, evidence from long-term data on pediatric TB outcomes, particularly from major tertiary centres such as KATH, which manage complex pediatric cases in Ghana, remains scarce. Additionally, existing studies in Ghana have primarily focused on adult populations, leaving pediatric TB outcomes and the influence of HIV co-infection and nutrition on treatment results inadequately documented [3]. Consequently, assessing pediatric TB treatment outcomes over an extended period at KATH could provide valuable insights into the standards of care and inform quality improvement initiatives in clinical practice. Factors such as age, nutritional status, and HIV co-infection are believed to be key determinants [3, 6, 8]; however, their relative associations and possible interactions within the clinical setting are not well understood.

This study aimed to determine TB treatment outcomes among children aged 0 to 14 years and assess the associations of HIV co-infection and other factors on unsuccessful treatment outcomes. The goal was to generate evidence to inform effective, targeted clinical interventions that improve pediatric TB outcomes in Ghana.

Methods

Study design and setting

A retrospective study was carried out at KATH using data from January 2014 to October 2025. KATH is a 1,200-bed tertiary hospital situated in Kumasi, the capital of the Ashanti Region, Ghana. It functions as the main referral center for patients from 13 of the 16 regions of Ghana, with a catchment population of around 10 million people, covering both rural and urban areas across the country [10]. It hosts the largest TB treatment center in the Ashanti region, managing approximately 30% of all TB cases in the region [10]. The Pediatric Infectious Diseases clinic operates in accordance with the guidelines of the National Tuberculosis Control Program (NTP) and the National AIDS Control Program (NACP), offering comprehensive TB and HIV care services, including prevention, diagnosis and treatment. The clinic has over 500 children enrolled in HIV and TB care. Diagnostic services at KATH include sputum smear microscopy, chest radiography, HIV testing, Xpert MTB/RIF testing for rapid molecular detection of Mycobacterium tuberculosis and rifampicin resistance, culture and drug susceptibility testing (DST) through reference laboratory linkages and Computer Aided Detection for TB (CAD4TB).

Study participants and sample size

Inclusion criteria

The study population included all children aged 0 to 14 years who were diagnosed with confirmed drug-susceptible TB at KATH, initiated treatment between January 1, 2014 and April 30, 2025, and had a treatment outcome documented at the time of the study analysis (October 31, 2025). TB cases were clinically diagnosed, radiologically or bacteriologically confirmed with Xpert Mycobacterium Tuberculosis / Rifampicin Resistance (MTB/RIF) [11]. Among the TB cases, HIV status was confirmed in accordance with the Ghana National HIV testing algorithms and managed concurrently for both TB and HIV [5].

Exclusion criteria

Children with missing treatment outcome data or those transferred to other facilities were excluded from the analysis.

Sample size

A total of 1,089 TB cases (pulmonary and extrapulmonary) were initially identified, of which 1,028 met the inclusion criteria and were included in the final analysis (see Fig. 1).

Fig. 1.

Flowchart of inclusion of pediatric TB patients in the study

Study variables

The primary outcome variable was unsuccessful treatment outcomes, defined as TB cases that ‘died’, ‘failed treatment’ or ‘defaulted’ per the NTP guidelines [12]. Under NTP guidelines, ‘cured’ refers to a bacteriologically confirmed TB patient who is smear- or culture-negative in the last month of treatment and on at least one previous occasion; ‘treatment completed’ refers to a patient who completes treatment without evidence of failure but lacks bacteriological confirmation in the final month; and ‘default’ refers to a patient whose treatment was interrupted for two consecutive months or more [5]. Explanatory variables included sociodemographic factors (age, gender, residential status), clinical factors (site of TB infection, HIV status, year of diagnosis), and nutritional status (underweight, normal weight, overweight), calculated using weight-for-age Z-scores for children under 10 years of age.

Data sources and management

Data were extracted from the Pediatric Tuberculosis Register, which captures information on each TB case, including demographic details, clinical characteristics, and treatment outcomes. HIV status is routinely recorded for all patients with TB as part of integrated TB/HIV service delivery. To minimize selection bias, we included all eligible cases within the study period. Information bias was addressed by using standardized definitions for all variables as per the NTP guidelines [12].

Data analysis

Data were analyzed using R statistical software (version 4.4.1) [13]. Descriptive statistics were used to summarize the characteristics of the study population. The frequency of unsuccessful treatment outcomes was calculated with 95% confidence intervals. A bivariate modified Poisson regression model with robust standard errors was employed to assess the crude association between unsuccessful treatment outcomes and each independent variable, given that the frequency of unsuccessful treatment outcomes exceeds 10% [14, 15]. The statistical analysis involved fitting multiple-variable Poisson regression models to estimate relative risks and 95% confidence intervals (CI). Variable selection for the multivariable analysis was conducted a priori based on clinical and epidemiological relevance informed by existing literature. Model 1 included HIV status as the primary predictor, with subsequent models progressively adjusted for age and sex (Model 2), residence location (Model 3), and site of TB infection (Model 4). An interaction term between age and HIV status was incorporated into the final model to assess effect modification. All models were adjusted for the year of diagnosis. The variables were selected for inclusion in the multivariable modified Poisson regression model [14, 15] based on clinical relevance. Stratified analyses were performed for each model to evaluate the effect within each nutritional status category (overweight, underweight, and normal weight). The results were presented as relative risk and adjusted relative risk (ARR), with 95% confidence intervals, because an odds ratio would have overestimated the effect (prevalence > 10%) [15]. A p-value of < 0.05 was considered statistically significant in all analyses. Multicollinearity among covariates was assessed using variance inflation factors (VIFs), with a threshold > 7 indicating problematic collinearity. Model assumptions were evaluated by checking for overdispersion and verifying model convergence.

Ethics approval and consent to participate

This study was conducted in accordance with the ethical principles of medical research involving human subjects as outlined in the Declaration of Helsinki. The study protocol received full approval from the Committee on Human Research, Publications and Ethics (CHRPE) of Kwame Nkrumah University of Science and Technology (CHRPE/AP/185/23) and the Komfo Anokye Teaching Hospital Institutional Review Board (KATH IRB/097/24). Since this was a retrospective review of routinely collected data, the requirement for individual patient consent was waived by both CHRPE and Komfo Anokye Teaching Hospital Institutional Review Board. All data were de-identified prior to analysis to ensure patient confidentiality.

Results

Descriptive characteristics

The study involved 1,028 children with a median age of 6.0 years (IQR 3.0, 10.0). As shown in Table 1, most participants were male (56.8%) and resided in urban areas (80.2%). Pulmonary TB was the most common presentation (82.4%). A majority of children were living with HIV (53.2%). The overall treatment success rate was 83.4% (857/1028). Among the 171 children with unsuccessful treatment outcomes, death was the more frequent cause (53.8%), followed by defaulting (46.2%). The bivariate analysis revealed significant differences between the successful and unsuccessful treatment groups. Children with unsuccessful outcomes were older (median 8.0 years vs. 6.0 years, p = 0.004) and more likely to be in the 10–14 years age category (40.4% vs. 25.9%, p = 0.001). A significantly higher proportion of children in the unsuccessful treatment group were living with HIV (62.0% vs. 51.4%, p = 0.011).

Table 1.

Descriptive statistics of the study population and by treatment outcomes

Characteristic	N	Successful N = 8571	Unsuccessful N = 1711	Overall N = 10281	p-value2
Age in years	1,028	6.0 (3.0, 10.0)	8.0 (4.0, 11.0)	6.0 (3.0, 10.0)	0.004
Age Category	1,028				0.001
< 2years		121 (14.1)	23 (13.5)	144 (14.0)
2 - <5 years		216 (25.2)	29 (17.0)	245 (23.8)
5 - <10 years		298 (34.8)	50 (29.2)	348 (33.9)
10–14 years		222 (25.9)	69 (40.4)	291 (28.3)
Sex	1,028				0.300
Female		376 (43.9)	68 (39.8)	444 (43.2)
Male		481 (56.1)	103 (60.2)	584 (56.8)
Site of TB Infection	888				0.700
Ex-Pulmonary TB		126 (17.3)	30 (18.6)	156 (17.6)
Pulmonary TB		601 (82.7)	131 (81.4)	732 (82.4)
HIV status	1,025				0.011
Negative		415 (48.6)	65 (38.0)	480 (46.8)
Positive		439 (51.4)	106 (62.0)	545 (53.2)
Residence location	1,028				0.400
Urban		683 (79.7)	141 (82.5)	824 (80.2)
Rural		174 (20.3)	30 (17.5)	204 (19.8)
Nutritional Status	788				0.400
Overweight		53 (7.9)	12 (10.4)	65 (8.2)
Normal Weight		333 (49.5)	61 (53.0)	394 (50.0)
Underweight		287 (42.6)	42 (36.5)	329 (41.8)
Specific treatment outcome	1,028				< 0.001
Cured		43 (5.0)	0 (0.0)	43 (4.2)
Completed		814 (95.0)	0 (0.0)	814 (79.2)
Defaulted		0 (0.0)	79 (46.2)	78 (7.7)
Died		0 (0.0)	92 (53.8)	92 (8.9)

¹Median (Q1, Q3); n (%)

²Wilcoxon rank sum test; Pearson’s Chi-squared test; Fisher’s exact test

Trends in treatment outcomes

Figure 2 shows the annual trends in treatment outcomes. Success rates fluctuated notably, dropping to a low of 64.4% in 2016 before generally increasing in subsequent years, with rates surpassing 95% in 2020 and 2022. There was also a general rise in unsuccessful outcomes, reaching about 20% in 2025. The detailed breakdown of specific treatment outcomes is available in Supplementary Table 1. Differences in treatment outcomes before (≤ 2019) and after (> 2019) the COVID-19 pandemic were statistically significant (p-value < 0.001; see Supplementary Table 2).

Fig. 2.

Yearly trend of treatment outcomes (N = 1028)

When stratified by HIV status (Fig. 3), the outcomes for children co-infected with HIV were consistently less favourable than for their peers without HIV in most years. In 2019, the unsuccessful treatment outcome rate for children with HIV co-infection was 76.6%, compared to 23.4% for those living without HIV. After maintaining a low unsuccessful TB treatment outcome in children living with HIV 2021-2022, the proportion of unsuccessful outcomes in this population increased continuously from 25% in 2022 to 100% in 2025.

Fig. 3.

Yearly unsuccessful treatment outcomes by HIV status (N = 171)

Factors associated with unsuccessful treatment

The regression analysis of the entire population is presented in Table 2. In the unadjusted model (Model 1), a positive HIV status was associated with a 40% higher risk of unsuccessful treatment (ARR 1.40, 95% CI 1.03–1.91). This effect remained significant and largely unchanged, with a 39% higher risk (ARR 1.39, 95% CI 1.01–1.93), after further adjustments for age, sex, residence, and site of infection in the fully adjusted model (Model 4). Each additional year of age was independently associated with a 5–6% increase in the risk of an unsuccessful outcome across the adjusted models (Model 4, ARR 1.06, 95% CI 1.02–1.10). The inclusion of an HIV-age interaction term in Model 5 revealed that, for children living with HIV, the risk of unsuccessful treatment increased by 8% with each additional year of age (ARR 1.08, 95% CI 1.04–1.12). No significant trend was observed in children without HIV.

Table 2.

Regression analysis for the entire population

Characteristic	Model 1				Model 2				Model 3				Model 4				Model 5 (HIV × Age Interaction)
	RR	95% CI	p-value	VIF	ARR	95% CI	p-value	VIF	ARR	95% CI	p-value	VIF	ARR	95% CI	p-value	VIF	ARR	95% CI	p-value	GVIF	Adjusted GVIF
HIV status				1.0				1.0				1.0				1.0
Negative	—	—			—	—			—	—			—	—
Positive	1.40	1.03, 1.91	0.034		1.38	1.01, 1.89	0.042		1.39	1.02, 1.90	0.039		1.390	1.01, 1.93	0.044
Age in years					1.05	1.01, 1.09	0.007	1.0	1.05	1.01, 1.09	0.007	1.0	1.060	1.02, 1.10	0.002	1.0
Child’s sex								1.0				1.0				1.0
Female					—	—			—	—			—	—
Male					1.12	0.83, 1.53	0.500		1.12	0.83, 1.53	0.500		1.110	0.81, 1.53	0.500
Residence location												1.0				1.0
Urban									—	—			—	—
Rural									0.88	0.58, 1.29	0.500		0.970	0.62, 1.46	0.900
Site of TB Infection																1.0
Ex-Pulmonary TB													—	—
Pulmonary TB													0.930	0.63, 1.42	0.700
HIV status * Age in years																				1.0	1.0
Negative * Age in years																	1.03	0.99, 1.08	0.200
Positive * Age in years																	1.08	1.04, 1.12	< 0.001

Abbreviations: CI Confidence Interval, GVIF Generalized Variance Inflation Factor, IRR Incidence Rate Ratio, VIF Variance Inflation Factor, RR Relative Risk, ARR Adjusted Relative Risk

Model 1: Null deviance = 550; Null df = 887; Log-likelihood = -427; AIC = 868; BIC = 902; Deviance = 532; Residual df = 881

Model 2: Null deviance = 550; Null df = 887; Log-likelihood = -428; AIC = 869; BIC = 903; Deviance = 533; Residual df = 881

Model 3: Null deviance = 612; Null df = 1,024; Log-likelihood = -468; AIC = 948; BIC = 978; Deviance = 594; Residual df = 1,019

Model 4: Null deviance = 612; Null df = 1,024; Log-likelihood = -468; AIC = 947; BIC = 971; Deviance = 595; Residual df = 1,020

Model 5: Null deviance = 612; Null df = 1,024; Log-likelihood = -472; AIC = 951; BIC = 965; Deviance = 603; Residual df = 1,022

The role of nutritional status

The regression analysis was further stratified by nutritional status in children under 10 years old (Table 3 and Supplementary Tables 3–4). The most significant findings emerged in the underweight subgroup (Table 3). Among underweight children, the negative association of age was amplified by HIV co-infection; for an undernourished child living with HIV, each additional year of age increased the risk of an unsuccessful outcome by 20% (ARR 1.20, 95% CI 1.07–1.34). This interaction was not seen in the normal or overweight groups, where neither HIV nor age showed statistically significant associations with the outcome.

Table 3.

Stratified regression analysis of underweight population

Characteristic	Model 1				Model 2				Model 3				Model 4				Model 5 (HIV × Age Interaction)
	RR	95% CI	p-value	VIF	ARR	95% CI	p-value	VIF	ARR	95% CI	p-value	VIF	ARR	95% CI	p-value	VIF	ARR	95% CI	p-value	GVIF	Adjusted GVIF
HIV status				1.0				1.0				1.0				1.0
Negative	—	—			—	—			—	—			—	—
Positive	1.61	0.86, 3.21	0.200		1.530	0.81, 3.06	0.200		1.59	0.84, 3.17	0.200		1.78	0.92, 3.65	0.100
Year of Diagnosis	0.98	0.89, 1.07	0.600	1.0	0.990	0.90, 1.08	0.700	1.0	0.99	0.90, 1.09	0.800	1.0	1.01	0.91, 1.11	0.900	1.1
Age in years					1.160	1.04, 1.29	0.008	1.0	1.16	1.04, 1.29	0.007	1.0	1.17	1.05, 1.31	0.006	1.0
Child’s sex								1.0				1.0				1.0
Female					—	—			—	—			—	—
Male					1.080	0.59, 2.04	0.800		1.09	0.59, 2.06	0.800		1.00	0.53, 1.93	> 0.900
Residence location												1.0				1.0
Urban									—	—			—	—
Rural									0.72	0.29, 1.55	0.400		0.59	0.18, 1.48	0.300
Site of TB Infection																1.1
Ex-Pulmonary TB													—	—
Pulmonary TB													0.75	0.34, 1.90	0.500
HIV status * Age in years																				1.0	1.0
Negative * Age in years																	1.11	0.97, 1.27	0.120
Positive * Age in years																	1.20	1.07, 1.34	0.002

Abbreviations: CI  Confidence Interval, GVIF Generalized Variance Inflation Factor, IRR  Incidence Rate Ratio, VIF  Variance Inflation Factor, RR  Relative Risk, ARR Adjusted Relative Risk

Model 1: Null deviance = 158; Null df = 294; Log-likelihood = -112; AIC = 238; BIC = 264; Deviance = 146; Residual df = 288

Model 2: Null deviance = 158; Null df = 294; Log-likelihood = -112; AIC = 237; BIC = 263; Deviance = 145; Residual df = 288

Model 3: Null deviance = 173; Null df = 327; Log-likelihood = -123; AIC = 258; BIC = 281; Deviance = 162; Residual df = 322

Model 4: Null deviance = 173; Null df = 327; Log-likelihood = -123; AIC = 257; BIC = 276; Deviance = 163; Residual df = 323

Model 5: Null deviance = 173; Null df = 327; Log-likelihood = -127; AIC = 260; BIC = 272; Deviance = 170; Residual df = 325

Discussion

This retrospective review, spanning 11 years, provides important insights into tuberculosis treatment outcomes among children at a major tertiary hospital in Ghana. It shows how demographic and clinical factors, particularly age and HIV status, influence treatment success. The overall treatment success rate of 83.4% observed in this study closely matches recent national progress but remains below the WHO target of 90% for effective TB control [1, 12]. This moderate success rate indicates that, although there have been some improvements in pediatric TB detection and management in Ghana, significant challenges still exist, notably poor treatment outcomes (death and default), especially among older, underweight children and those living with HIV.

Over the period, this study revealed significant year-to-year variations in treatment outcomes, with the lowest success rate in 2016 (64. 4%) and peaks exceeding 95% in 2020 and 2022. Treatment outcomes before (≤ 2019) and after (> 2019) the COVID-19 pandemic were statistically different (p-value < 0.001). Such fluctuations likely reflect systemic challenges such as resource constraints, inconsistent supply of anti- TB drugs [16, 17], staff turnover, and disruptions in TB service delivery, including those linked to the COVID- 19 pandemic and its aftermath [12]. The high rates of “treatment completed” relative to “cured” cases (79.2% vs. 4.2%) also suggest diagnostic gaps in confirming bacteriological cure among children, consistent with findings in Accra [8] and other African nations [3]. The limited bacteriological confirmation reflects the difficulty in obtaining sputum samples from children, the paucibacillary nature of pediatric TB, and the frequent reliance on clinical and radiological diagnosis rather than microbiological confirmation [4]. Additionally, multiple pharmacokinetic and safety studies of HIV and TB drugs at the study site during the period may have contributed to the recorded TB cure rates through closer supervision of adherence [18–20].

The treatment success rate in this study (83. 4%), although below the desired global target of 90% among children, compares favourably with pediatric TB outcomes reported in other African settings, such as Nigeria (62. 9%) [21] and Ethiopia (78. 9%) [22]. Ghana’s relatively robust TB/HIV service integration may explain this performance [12, 23], with the Pediatric Infectious Diseases Unit coordinating such integrated pediatric TB and HIV services at KATH. However, the persistently high unsuccessful rate among co-infected children in this study highlights gaps in early diagnosis and management, including the fact that antiretroviral therapy (ART) uptake and ART regimen information were not available in the TB records at the time of the study, so the proportion on antiretroviral therapy could not be estimated. Similar to reports across sub- Saharan Africa, most unsuccessful TB treatment outcomes were observed in children co-infected with HIV, reaffirming the synergistic nature of these two infectious diseases [3]. Reports from the WHO indicate that children living with HIV are nearly twice as likely to die from TB compared to their peers without HIV, primarily due to late presentation, immune suppression, and severe disease [1]. The stratified analysis by nutritional status revealed that underweight children with HIV co-infection face particularly elevated risks, suggesting that malnutrition exacerbates the negative synergy between HIV and TB. This finding aligns with evidence from sub-Saharan Africa demonstrating that nutritional status is a critical predictor of mortality in children living with HIV [24, 25].

The mortality rate of 8.9% and the default rate of 7.6% among children in this study remain a significant concern, being higher than those observed in some comparable settings, such as Zambia, with 6.6% mortality [3], yet lower than the 17.4% mortality reported in a study at a teaching hospital in Accra, Ghana [8]. These outcomes highlight ongoing barriers to adherence monitoring and continuity of care, including socioeconomic constraints, long travel distances, and stigma linked to TB/HIV co-infection [6]. Children aged 10–14 years experienced particularly poor outcomes, highlighting the need for community-based follow-up strategies, adherence support as they gain more independence, and decentralization of pediatric TB care.

Age was a significant independent predictor of unsuccessful treatment outcomes. The risks increased notably with each additional year of age. The positive association between age and unsuccessful treatment outcomes (ARR = 1.06; 95% Cl: 1.02–1.10) indicates increasing vulnerability with age [26], potentially due to treatment fatigue from polypharmacy and long treatment duration, higher rates of non-adherence, and late diagnosis [27]. Furthermore, the interaction analysis showed that the effect of age was especially prominent among children with HIV, where each additional year raised the risk of unsuccessful treatment by 8%. This underscores the compounded association of immunosuppression [3, 8], drug interactions, and overlapping toxicities from anti-TB and long-term ART [26, 28]. These findings highlight the importance of enhanced adherence counselling, early ART initiation, and the integration of nutritional and psychosocial support into pediatric TB/HIV programmes.

Although males represented a slightly higher proportion of TB cases (56.8%), sex did not significantly affect treatment outcomes. This contrasts with some adult studies, in which males often exhibit poorer adherence and outcomes [29], but agrees with a pediatric study in Ethiopia [30], suggesting that gender disparities may be less evident in childhood.

Strengths and limitations of the study

The study setting, as a major tertiary center, had a relatively high proportion of participants with TB/HIV and offered valuable insights into TB/HIV treatment outcomes within a diverse population. The use of a large sample over 11 years ensured robust statistical power. Furthermore, employing advanced statistical methods was a strength of the study. However, there were some limitations. The single-center nature of the study restricts the generalizability of the findings to other settings with different healthcare systems and population profiles. Baseline anthropometric data (height) were missing for children aged 10–14 years, restricting the assessment of nutritional status (Body Mass Index) in this age group. The absence of records on HIV transmission routes, duration of HIV treatment (newly diagnosed versus failing ART), ART regimen details, and HIV viral load prevents a full assessment of the association of viremia on outcomes. Additionally, the lack of data on post-treatment results hampers understanding of long-term effects beyond survival, such as growth, development, and educational achievement. Despite adjusting for confounders, residual confounders may have influenced the regression analysis. Finally, associations in this study do not infer causality as this study was cross-sectional in design.

Conclusion

The successful pediatric TB treatment rate at KATH remains below the WHO target of 90%. Older age and HIV co-infection substantially increased the risk of unsuccessful treatment, with more risk among underweight children. Strengthening integrated pediatric TB/HIV services, improving enhancing age-related adherence and nutritional support are essential to improving outcomes and reducing TB-related default and mortality among children in Ghana.

Abbreviations

ARR

Adjusted Relative Risk

ART

Antiretroviral Therapy

CHRPE

Committee on Human Research, Publications and Ethics

CI

Confidence Interval

DOTS

Directly Observed Therapy –short course

HIV

Human Immunodeficiency Virus

KATH

Komfo Anokye Teaching Hospital

RR

Relative Risk

TB

Tuberculosis

WHO

World Health Organization

Authors' contributions

CMD: Conceptualization, Investigation, Writing - Original Draft JSI: Formal Analysis, Visualization, Writing - Original Draft MEM: Investigation, Writing - Original Draft AIC: Conceptualization, Investigation, Formal Analysis, Writing - Original Draft SAO: Validation, Writing - Review & Editing NKM: Data Curation, Validation, Writing - Review & Editing EOA: Data Curation, Validation, Writing - Review & Editing AM: Data Curation, Validation, Writing - Review & Editing AK: Methodology, Writing - Review & Editing, Supervision AE: Methodology, Writing - Review & Editing, Supervision.

Funding

This study received no specific funding.

Data availability

The data supporting the conclusions of this article are not publicly available due to patient confidentiality concerns. Access to the data may be considered upon reasonable request to the corresponding author.

Declarations

Ethics approval and consent to participate

This study was conducted in accordance with the ethical principles of medical research involving human subjects as outlined in the Declaration of Helsinki. The study protocol received full approval from the Committee on Human Research, Publications and Ethics (CHRPE) of Kwame Nkrumah University of Science and Technology (CHRPE/AP/185/23) and the Komfo Anokye Teaching Hospital Institutional Review Board (KATH IRB/097/24). Since this was a retrospective review of routinely collected data, the requirement for individual patient consent was waived by both CHRPE and Komfo Anokye Teaching Hospital Institutional Review Board. All data were de-identified prior to analysis to ensure patient confidentiality.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.