Clinical spectrum of extrapulmonary tuberculosis at a tertiary care centre in Ghana: a retrospective study Authors: Walter Appati, Divine Aseye Yao Amenuke, Adelaide Appati, Gifty Akua Baiden Published: 2026 ABSTRACT BackgroundExtrapulmonary tuberculosis (EPTB) presents with diverse clinical manifestations and remains a diagnostic and public health challenge, particularly in high tuberculosis (TB) burden settings such as Ghana. Understanding its clinical spectrum and associated factors is essential for improved case detection and management. This study aims to describe the epidemiology, clinical spectrum, and associated demographic and clinical factors of EPTB in Ghana.MethodsWe conducted a retrospective review of TB cases seen at the Chest Clinic of Komfo Anokye Teaching Hospital in Ghana between October 2023 and November 2025. Patients were categorised into pulmonary tuberculosis (PTB) and EPTB. The primary outcome was the clinical manifestation of EPTB, categorised by anatomical site of involvement. Multivariate logistic regression was used to identify factors associated with specific EPTB manifestations, and covariates included age categories, sex, and HIV status. Adjusted odds ratios (aORs) were reported, and statistical significance was set at p < 0.05.ResultsAmong 1,098 patients with TB, 179 (16.3%) had isolated EPTB. Patients with EPTB were less frequently male (52.0% vs. 68.8%; p < 0.001) and had lower HIV prevalence (9.0% vs. 14.9%; p = 0.035). Spinal TB was the predominant EPTB manifestation (73, 40.8%), followed by TB lymphadenitis (52, 29.1%), abdominal TB (34, 19.0%), TB pericarditis (13, 7.3%), and central nervous system (CNS) TB (7, 3.9%). Within the EPTB cohort, individuals aged 35–54 years had higher odds of spinal TB (aOR 3.2, 95% CI: 1.5–6.6, p = 0.002). Male sex (aOR 2.8, 95% CI: 1.4–5.7, p = 0.005) and HIV infection (aOR 4.1, 95% CI: 1.3–12.7, p = 0.014) were independently associated with TB lymphadenitis. No significant demographic or HIV-related associations were identified for abdominal TB or TB pericarditis.ConclusionIsolated EPTB accounted for approximately one-sixth of all TB cases. Spinal TB and TB lymphadenitis were the predominant manifestations. Distinct demographic and HIV-related associations across EPTB phenotypes underscore the clinical heterogeneity of EPTB and highlight the need for studies evaluating concurrent PTB and EPTB disease to better define the full spectrum of TB in Ghana. Background Tuberculosis (TB) is a highly infectious disease caused by the bacterium Mycobacterium tuberculosis. It generally targets the lungs (pulmonary tuberculosis), but can also spread beyond the lungs to affect other parts of the body (extrapulmonary TB) [1, 2]. While pulmonary tuberculosis (PTB) is the most common and transmissible form of the disease, extrapulmonary tuberculosis (EPTB) constitutes a substantial and increasingly recognised proportion of TB cases, contributing significantly to morbidity, disability, and health system costs [3, 4]. Globally, EPTB accounts for approximately 15–25% of all TB cases, with markedly higher proportions reported among people living with HIV, women, children, and individuals with immunosuppressive conditions [3]. There are three forms of TB, including PTB only, concurrent PTB and EPTB, and EPTB only; however, concurrent PTB and EPTB is captured as PTB in most registers worldwide, including Ghana [5]. EPTB is a consistently important but difficult-to-measure component of TB burden, with apparent increases in some settings largely reflecting improved detection and declining PTB rather than clear global growth [6]. In Ghana, TB remains a major public health concern and is managed by the Ghana Health Service through the National Tuberculosis Control Programme (NTP) [7]. The program operates within the framework of the global strategies of the World Health Organisation (WHO), particularly the End TB Strategy, and implements standardised policies for diagnosis, treatment, surveillance, and monitoring [8, 9]. Despite its clinical significance, data on the epidemiology, clinical spectrum, and outcomes of EPTB in Ghana remain limited, particularly in tertiary care settings. Most national surveillance systems prioritise PTB, resulting in insufficient characterisation of extrapulmonary disease patterns and outcomes. The national guideline regarding the diagnosis of EPTB is challenging due to nonspecific clinical signs and symptoms that may mimic other diseases, a lack of awareness among clinical providers of EPTB as a potential diagnosis, and difficulty in collecting samples from affected sites, often requiring surgical procedures [10, 11]. Low-Complexity automated Nucleic Acid Amplification Tests (LCaNAATs) can help identify patients with EPTB and rifampicin resistance. Although LCaNAATs are useful for detecting Mycobacterium tuberculosis in extrapulmonary specimens such as cerebrospinal fluid, peritoneal fluid, and pericardial fluid, sensitivity varies substantially depending on specimen type and bacillary burden [12]. The evidence base for EPTB-specific treatment is limited, and recommendations are often extrapolated from PTB [13]. Adults with EPTB are eligible for the 6-month 2HRZE/4 HR regimen, except for those with TB of the CNS, bone or joint, for which some expert groups suggest longer therapy (i.e., 9–12 months) [14]. EPTB is associated with considerable long-term morbidity, including neurological disability, spinal deformity, chronic pain, and cardiovascular complications, even after microbiological cure [15]. These sequelae are seldom captured in routine TB surveillance systems, which primarily focus on treatment completion and mortality, thereby underestimating the actual burden of disease [16, 17]. As global TB strategies increasingly emphasise patient-centred care and post-TB health, a clearer understanding of EPTB epidemiology and outcomes is urgently needed. This study aims to characterise the epidemiology and outcomes of EPTB by examining the demographic profile, risk factors, and clinical manifestations of extrapulmonary tuberculosis among patients attending a tertiary chest clinic in Ghana, thus contributing context-specific evidence to inform clinical decision-making, diagnostic strategies, and integrated TB care beyond pulmonary disease. Methods Study design and site This was a retrospective study of all adolescent and adult patients diagnosed with EPTB who underwent both microbiological and radiological investigations and were routinely followed up at the Chest Clinic of Komfo Anokye Teaching Hospital (KATH) from October 2023 to November 2025. Patients seen at the chest clinic include those with PTB, those with concurrent PTB and EPTB, and those with EPTB only. However, for the purposes of this study, we sought to analyse patients with only EPTB. The tests conducted at KATH for EPTB diagnosis included x-rays, excisional lymph node biopsies for histopathology, and pleural, pericardial, and CSF fluid analyses using GeneXpert, with assessment of LDH and ADA. This study involved reviewing patients’ records to obtain socio-demographic factors and sequelae Study population The study included a total of 1098 patients with TB, aged 15 years or older and who were routinely seen at the chest clinic between October 2023 and November 2025. These patients were diagnosed based on clinical symptoms, imaging, and microbiological testing. Patients with incomplete medical records for key study variables, including age, sex, HIV status or EPTB manifestation, were excluded from the final analysis. Because the proportion of missing data was small, complete-case analysis was performed to maintain data consistency and analytical reliability. Data collection Data was collected using an abstraction form by manually entering patient’s details. The data was subsequently extracted into an Excel spreadsheet for further analysis. Data included patient ID numbers, age, sex, comorbidities and risk factors, diagnosis, and other extrapulmonary manifestations. The collected data was checked for completeness and accuracy. Incomplete, inconsistent, or duplicate responses were reviewed and addressed. The analysed dataset contained no personal identifiers, such as names or phone numbers. Data analysis The data were analysed using Stata 17.0 (StataCorp, TX, USA). Characteristics of the study participants were summarised using descriptive statistics, including frequencies, percentages, and means. The primary outcome was the clinical spectrum of extrapulmonary tuberculosis, categorised by anatomical site of involvement as TB of the lymph nodes, abdomen, CNS, pericardium, and spine. All variables considered clinically relevant based on prior literature were entered into the multivariate logistic regression model. Covariates included age categories, sex, and HIV status. Multivariate logistic regression was performed to identify demographic and clinical factors associated with specific manifestations of EPTB. Adjusted odds ratios (aORs) with 95% confidence intervals (CIs) were reported. Statistical significance was set at p < 0.05. Confidence intervals were used to assess the precision and strength of associations. Ethical consideration Ethical clearance for the study was obtained from Komfo Anokye Teaching Hospital Institutional Review Board, with reference number KATH IRB/AP/019/26. Informed consent was waived as a result of the retrospective nature of the study. Gatekeeper consent, however, was sought from the Chest Clinic through the Internal Medicine Directorate of KATH. This study was conducted in accordance with the most recent version of the Declaration of Helsinki. Results Demographics and Risk factors for patients with extrapulmonary tuberculosis A total of 1098 patients with TB attended the Chest Clinic from October 2023 to November 2025. Of this number, 179 (16.3%) had isolated EPTB, and 919 (83.7%) had PTB or PTB with concomitant EPTB. Males constituted the majority of the overall study population (66.0%). However, the proportion of males was significantly lower among patients with EPTB compared with those with PTB (52.0% vs 68.8%, p < 0.001), indicating a relatively more balanced sex distribution in EPTB. The median age of patients with EPTB was slightly lower than that of patients with PTB. Most patients in both groups were aged 35–54 years, accounting for 44.7% of EPTB cases and 47.3% of PTB cases. Although a higher proportion of younger patients (15–34 years) was observed among EPTB cases compared with PTB cases (33.5% vs 25.6%), the overall age-group distribution did not differ significantly between the two groups (p = 0.069). HIV co-infection was present in 153 (13.9%) patients overall. The prevalence of HIV infection was significantly lower among patients with EPTB than among those with PTB (9.0% vs 14.9%, p = 0.035). This finding suggests that patients with isolated EPTB differ clinically or epidemiologically from those with PTB (Table 1).Table 1Socio-demographic characteristics of patientsCharacteristicsEPTBn = 179PTB*n = 919Totaln = 1098p-valueSexMale, n (%)93 (52.0)632 (68.8)725 (66.0)<0.001Median age (IQR), years42 (52–32)44 (56–34)44 (55–34)0.042Age group, n (%)0.06915–34 years60 (33.5)235 (25.6)295 (26.9)35–54 years80 (44.7)435 (47.3)515 (46.9)≥55 years39 (21.8)249 (27.1)288 (26.2)HIV StatusPositive, n (%)16 (9.0)137 (14.9)153 (13.9)0.035n = frequency, PTB = Pulmonary Tuberculosis, EPTB = Extrapulmonary Tuberculosis*PTB includes PTB with or without concurrent EPTB Table 1Socio-demographic characteristics of patientsCharacteristicsEPTBn = 179PTB*n = 919Totaln = 1098p-valueSexMale, n (%)93 (52.0)632 (68.8)725 (66.0)<0.001Median age (IQR), years42 (52–32)44 (56–34)44 (55–34)0.042Age group, n (%)0.06915–34 years60 (33.5)235 (25.6)295 (26.9)35–54 years80 (44.7)435 (47.3)515 (46.9)≥55 years39 (21.8)249 (27.1)288 (26.2)HIV StatusPositive, n (%)16 (9.0)137 (14.9)153 (13.9)0.035n = frequency, PTB = Pulmonary Tuberculosis, EPTB = Extrapulmonary Tuberculosis*PTB includes PTB with or without concurrent EPTB Socio-demographic characteristics of patients n = frequency, PTB = Pulmonary Tuberculosis, EPTB = Extrapulmonary Tuberculosis *PTB includes PTB with or without concurrent EPTB Among the 179 patients diagnosed with EPTB, spinal tuberculosis was the most common manifestation, accounting for 73 (40.8%) cases. This was followed by TB lymphadenitis, which constituted 52 (29.1%) cases, while abdominal TB accounted for 34 (19.0%) cases. Pericardial TB was less frequent, occurring in 13 (7.3%) patients, and CNS TB was the least common manifestation, representing 7 (3.9%) cases. Male predominance was observed in most EPTB manifestations, particularly among patients with TB lymphadenitis, where males accounted for 69.2% of cases. In contrast, abdominal TB showed a relatively higher proportion of females, with males constituting only 38.2% of cases. Pericardial and CNS TB demonstrated a more balanced sex distribution. The median age varied across manifestations. Patients with TB lymphadenitis and CNS TB tended to be younger, with median ages of 34 years (IQR: 28.5–49) and 31 years (IQR: 21–45), respectively. In contrast, patients with abdominal, spinal, and pericardial TB were generally older, with median ages above 40 years. Age-group distribution also differed by disease manifestation. More than half of patients with TB lymphadenitis (51.9%) and CNS TB (57.1%) were aged 15–34 years, whereas spinal TB was most common among individuals aged 35–54 years (57.5%). Pericardial TB had the highest proportion of older adults aged 55 years and above (38.5%). HIV co-infection varied across EPTB manifestations. CNS TB had the highest proportion of patients with HIV (28.6%), followed by TB lymphadenitis (15.4%). HIV prevalence was lower among patients with spinal TB (5.5%) and abdominal TB (2.9%) (Table 2).Table 2Characteristics of patients stratified by EPTB manifestationsCharacteristicsSpinen = 73Lymphadenitisn = 52Abdominaln = 34Pericardialn = 13CNSn = 7SexMale, n (%)34 (46.6)36 (69.2)13 (38.2)7 (53.9)3 (42.9)Median age (IQR), years43 (52–35)34 (49–28.5)44.5 (33–53)42 (60–39)31 (45–21)Age group, n (%)15–34 years16 (21.9)27 (51.9)10 (29.4)3 (23.0)4 (57.1)35–54 years42 (57.5)15 (28.9)16 (47.1)5 (38.5)2 (28.6)≥55 years15 (20.6)10 (19.2)8 (23.5)5 (38.5)1 (14.3)HIV StatusPositive, n (%)4 (5.5)8 (15.4)1 (2.9)1 (7.7)2 (28.6)n = frequency, HIV = Human Immunodeficiency Virus, CNS = Central Nervous System Table 2Characteristics of patients stratified by EPTB manifestationsCharacteristicsSpinen = 73Lymphadenitisn = 52Abdominaln = 34Pericardialn = 13CNSn = 7SexMale, n (%)34 (46.6)36 (69.2)13 (38.2)7 (53.9)3 (42.9)Median age (IQR), years43 (52–35)34 (49–28.5)44.5 (33–53)42 (60–39)31 (45–21)Age group, n (%)15–34 years16 (21.9)27 (51.9)10 (29.4)3 (23.0)4 (57.1)35–54 years42 (57.5)15 (28.9)16 (47.1)5 (38.5)2 (28.6)≥55 years15 (20.6)10 (19.2)8 (23.5)5 (38.5)1 (14.3)HIV StatusPositive, n (%)4 (5.5)8 (15.4)1 (2.9)1 (7.7)2 (28.6)n = frequency, HIV = Human Immunodeficiency Virus, CNS = Central Nervous System Characteristics of patients stratified by EPTB manifestations n = frequency, HIV = Human Immunodeficiency Virus, CNS = Central Nervous System Demographic/clinical variables and their associations with other EPTB manifestations Patients aged 35–54 years had significantly higher odds of spinal tuberculosis compared with those aged 15–34 years (aOR = 3.2, 95% CI: 1.5–6.6, p = 0.002). In contrast, patients aged 55 years and above did not show a statistically significant association with spinal TB (aOR = 1.8, 95% CI: 0.8–4.4, p = 0.177). Male sex was not significantly associated with spinal TB compared with female sex (aOR = 0.7, 95% CI: 0.4–1.3, p = 0.268) (Table 3).Table 3Demographic/clinical variables and their association with TB of the spineSpineAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–543.0 (1.5–6.3)0.0033.2 (1.5–6.6)0.002≥551.7 (0.7–4.1)0.2181.8 (0.8–4.4)0.177SexFemale (Reference)1–––Male0.7 (0.4–1.3)0.2330.7 (0.4–1.3)0.268Risk factorHIV0.5 (0.1–1.5)0.1870.4 (0.1–1.3)0.119Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Table 3Demographic/clinical variables and their association with TB of the spineSpineAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–543.0 (1.5–6.3)0.0033.2 (1.5–6.6)0.002≥551.7 (0.7–4.1)0.2181.8 (0.8–4.4)0.177SexFemale (Reference)1–––Male0.7 (0.4–1.3)0.2330.7 (0.4–1.3)0.268Risk factorHIV0.5 (0.1–1.5)0.1870.4 (0.1–1.3)0.119Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Demographic/clinical variables and their association with TB of the spine cOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Patients aged 35–54 years had significantly lower odds of TB lymphadenitis compared with those aged 15–34 years (aOR = 0.3, 95% CI: 0.1–0.6, p = 0.001). Similarly, patients aged 55 years and older were less likely to present with TB lymphadenitis than those aged 15–34 years (aOR = 0.4, 95% CI: 0.1–1.0, p = 0.041). These findings suggest that TB lymphadenitis was more common among younger individuals. Male patients had significantly higher odds of TB lymphadenitis than female patients (aOR = 2.8, 95% CI: 1.4–5.7, p = 0.005). In addition, patients with HIV were more likely to have TB lymphadenitis than patients without HIV (aOR = 4.1, 95% CI: 1.3–12.7, p = 0.014) (Table 4).Table 4Demographic/clinical variables and their association with TB lymphadenitisLymphadenitisAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–540.3 (0.1–0.6)0.0010.3 (0.1–0.6)0.001≥550.4 (0.2–1.0)0.0540.4 (0.1–1.0)0.041SexFemale (Reference)1–––Male2.8 (1.4–5.5)0.0042.8 (1.4–5.7)0.005Risk factorHIV2.7 (1.0–7.6)0.0614.1 (1.3–12.7)0.014Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Table 4Demographic/clinical variables and their association with TB lymphadenitisLymphadenitisAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–540.3 (0.1–0.6)0.0010.3 (0.1–0.6)0.001≥550.4 (0.2–1.0)0.0540.4 (0.1–1.0)0.041SexFemale (Reference)1–––Male2.8 (1.4–5.5)0.0042.8 (1.4–5.7)0.005Risk factorHIV2.7 (1.0–7.6)0.0614.1 (1.3–12.7)0.014Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Demographic/clinical variables and their association with TB lymphadenitis cOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval There were no statistically significant associations observed between age or sex and abdominal TB among patients with EPTB. Likewise, HIV infection was not significantly associated with abdominal TB (aOR = 0.2, 95% CI: 0.0–1.9, p = 0.168) (Table 5).Table 5Demographic/clinical variables and their association with TB abdomenAbdomenAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–541.3 (0.5–3.0)0.6161.2 (0.5–2.9)0.720≥551.3 (0.5–3.6)0.6281.3 (0.5–3.8)0.612SexFemale (Reference)1–––Male0.5 (0.2–1.1)0.0780.5 (0.2–1.1)0.075Risk factorHIV0.3 (0.0–2.1)0.2030.2 (0.0–1.9)0.168Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Table 5Demographic/clinical variables and their association with TB abdomenAbdomenAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–541.3 (0.5–3.0)0.6161.2 (0.5–2.9)0.720≥551.3 (0.5–3.6)0.6281.3 (0.5–3.8)0.612SexFemale (Reference)1–––Male0.5 (0.2–1.1)0.0780.5 (0.2–1.1)0.075Risk factorHIV0.3 (0.0–2.1)0.2030.2 (0.0–1.9)0.168Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Demographic/clinical variables and their association with TB abdomen cOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval No demographic or clinical variable was significantly associated with tuberculous pericarditis among patients with EPTB (Table 6).Table 6Demographic/clinical variables and their association with TB pericarditisPericarditisAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–541.3 (0.3–5.5)0.7531.3 (0.3–5.9)0.713≥552.8 (0.6–12.4)0.1772.9 (0.6–13.2)0.165SexFemale (Reference)1–––Male1.1 (0.3–3.4)0.8871.2 (0.4–3.6)0.802Risk factorHIV0.8 (0.1–6.9)0.8700.7 (0.1–6.0)0.760Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Table 6Demographic/clinical variables and their association with TB pericarditisPericarditisAgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–541.3 (0.3–5.5)0.7531.3 (0.3–5.9)0.713≥552.8 (0.6–12.4)0.1772.9 (0.6–13.2)0.165SexFemale (Reference)1–––Male1.1 (0.3–3.4)0.8871.2 (0.4–3.6)0.802Risk factorHIV0.8 (0.1–6.9)0.8700.7 (0.1–6.0)0.760Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Demographic/clinical variables and their association with TB pericarditis cOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Also, no statistically significant associations were observed between age or sex and central nervous system (CNS) tuberculosis among patients with EPTB. HIV infection showed a positive association with CNS TB, with patients with HIV having higher odds of CNS involvement compared with patients without HIV (aOR = 6.3, 95% CI: 1.0–40.6). However, this finding was borderline statistically significant (p = 0.051) and should be interpreted with caution, given the wide confidence interval and the small number of CNS TB cases (Table 7).Table 7Demographic/clinical variables and their association with CNS TBCentral Nervous System (CNS)AgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–540.4 (0.1–2.0)0.2460.2 (0.0–1.5)0.126≥550.4 (0.0–3.4)0.3800.2 (0.0–2.5)0.225SexFemale (Reference)1–––Male0.7 (0.1–3.1)0.6250.6 (0.1–3.0)0.559Risk factorHIV4.5 (0.8–25.4)0.0876.3 (1.0–40.6)0.051Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Table 7Demographic/clinical variables and their association with CNS TBCentral Nervous System (CNS)AgecOR (95% CI)p valueaOR (95% CI)p value15–34 (Reference)1–––35–540.4 (0.1–2.0)0.2460.2 (0.0–1.5)0.126≥550.4 (0.0–3.4)0.3800.2 (0.0–2.5)0.225SexFemale (Reference)1–––Male0.7 (0.1–3.1)0.6250.6 (0.1–3.0)0.559Risk factorHIV4.5 (0.8–25.4)0.0876.3 (1.0–40.6)0.051Adjusted for age, sex and HIV statuscOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Demographic/clinical variables and their association with CNS TB cOR = Crude Odds Ratio, aOR = Adjusted Odds Ratio, CI = Confidence Interval Discussion This retrospective study provides a comprehensive description of the clinical spectrum and associated demographic and clinical characteristics of EPTB among patients attending the chest clinic in a high TB burden setting in Ghana. Among all TB cases seen between October 2023 and November 2025, extrapulmonary disease accounted for 16.3% of all tuberculosis cases, which is consistent with global estimates indicating that EPTB constitutes 15–20% of TB in the general population and a substantially higher proportion in people with immunosuppression [3, 18]. The near-equal sex distribution observed among patients with EPTB contrasts with the male predominance typically reported for pulmonary TB and aligns with evidence suggesting that sex differences in TB presentation narrow in extrapulmonary disease [19, 20]. Spinal tuberculosis was the most common clinical manifestation of EPTB in this population, accounting for over 40% of cases, followed by lymph node tuberculosis. This distribution differs from that of many global series, in which lymphadenitis predominates, particularly in high-income settings [21, 22]. However, it is consistent with reports from parts of Africa and Asia, where delayed presentation and diagnostic constraints may favour detection of more advanced or symptomatic disease, such as spinal involvement [19]. The relatively low frequency of CNS TB observed may reflect diagnostic limitations rather than true rarity, given the known challenges in confirming TB of the CNS [23]. A key finding is the variation in demographic profiles across EPTB manifestations. TB lymphadenitis and CNS TB were more common among younger patients, whereas spinal and abdominal TB tended to occur in older age groups. This pattern is consistent with prior studies suggesting that lymphatic TB is more common in younger individuals due to stronger immune-mediated lymph node involvement, while skeletal and abdominal TB may reflect chronic reactivation or delayed dissemination [11, 20]. Sex distribution also varied by site, with a male predominance in lymphadenitis and spinal TB, whereas abdominal TB showed a relatively higher proportion of females. These differences may reflect a combination of biological susceptibility, behavioural factors, and healthcare-seeking patterns reported in TB epidemiology studies [24]. Overall, HIV prevalence was significantly lower among patients with isolated EPTB compared with those with pulmonary tuberculosis (PTB) involvement (9.0% vs 14.9%). This finding contrasts with several studies reporting higher HIV prevalence among EPTB patients and may reflect differences in patient populations, referral patterns, ART coverage, or the predominance of isolated rather than disseminated EPTB in this cohort [25, 26]. Among the different EPTB forms, CNS TB demonstrated the highest HIV prevalence, consistent with evidence that advanced immunosuppression predisposes individuals to severe disseminated TB and meningeal involvement [27]. Similarly, HIV infection was independently associated with TB lymphadenitis within the EPTB population. However, the overall observed association may reflect differences within EPTB subgroups rather than a substantially elevated HIV burden relative to PTB overall [25, 28]. These findings highlight the heterogeneity of EPTB and suggest that HIV-related susceptibility may vary according to disease site. Multivariate analyses further highlight important site-specific associations. Spinal TB was significantly associated with middle-aged adults (35–54 years), suggesting peak vulnerability in economically active populations. TB lymphadenitis was independently associated with younger age, male sex, and HIV infection, consistent with previous studies demonstrating strong immunological and demographic influences on lymphatic TB presentation [22, 24]. Male sex was independently associated with a significantly increased risk of lymphadenitis in this study, a finding that contrasts with some reports suggesting female predominance, highlighting the context-specific nature of EPTB manifestations [20, 29]. In contrast, abdominal, pericardial, and CNS TB showed no consistent statistically significant associations, likely reflecting small subgroup sizes and limited statistical power rather than the absence of biological relationships. Limitations The study was conducted at a single tertiary-level chest clinic, which may limit generalisability to peripheral health facilities or community-based settings with different referral patterns and diagnostic capacity. Important clinical variables, such as duration of symptoms before presentation, socioeconomic determinants, alcohol use, smoking, and body mass index (BMI), were not consistently available and could not be included in the analysis, raising the possibility of residual confounding. Given the small subgroup counts, the multivariate models may be underpowered, and results should be interpreted with caution. Conclusion Early recognition of EPTB should be prioritised in clinical algorithms, particularly in outpatient settings where diagnosis is often delayed. The strong association of CNS and lymph node TB with HIV infection reinforces the importance of routine HIV testing among all patients with TB, in line with global TB/HIV collaborative recommendations. Furthermore, the burden of EPTB among younger and middle-aged adults highlights its potential socioeconomic impact in endemic settings. Future studies should further characterise patients with concurrent PTB and EPTB, as this subgroup may represent a distinct clinical and immunological phenotype.