Treatment outcomes of Nigerian patients with tuberculosis: A retrospective 25-year review in a regional medical center.

Despite the general improvement in global health standards in the second decade of the 21 st Century, the prevalence and mortality of tuberculosis (TB) in low-and middle-income countries (LMICs) remain high [ 1 ]. Since 1993, when the World Health Organization (WHO) declared tuberculosis (TB), a disease of increasing global concern, millions of people are still being infected and die annually as a result of TB [ 1 , 2 ]. Multi-drug resistant Mycobacterium tuberculosis (MTB) infection is also on the rise, making TB infection of more concern [ 1 , 2 ]. The factors contributing to the persistence of high TB with poor outcomes in LMICs include a high level of extreme poverty, low-quality health care, inadequate nutrition, and overcrowding [ 1 , 3 ]. The third sustainable development goal (SDG3) target is to end the TB epidemic by 2030. In order to achieve this goal, LMIC TB programs must have comprehensive and ongoing reviews of successes and failures in order to guide systematic improvements that will aid in achieving the WHO set target of 85% treatment success. In the past three decades, studies reported varying TB treatment outcomes ranging from 34.0% to 85.45% in the LMICs [ 1 , 4 ]. However, comprehensive data of TB treatment in LMICs are scarce, often of poor quality, or of a limited time span, making these data sub-optimal for informed decision-making. The present study evaluated data collected over 25 years of a continuing TB treatment program and focuses on the outcomes and lessons that can be learned as the world strives towards achieving the SDG3 targets.

Of the 1,146,560 patients seen in the hospitals’ clinics during the study period, 24,330 presumptive cases were investigated for TB and 3,443 (14.2%) cases were confirmed following WHO TB diagnosis guidelines ( Fig 1 ). Fifty-nine (1.7%) patients were excluded because of insufficient data. There were no significant differences in demographic and clinical characteristics between study subjects and those excluded. The flow diagram of program enrolment and treatment outcome of TB cases. The final outcome data excludes 26 missing data points (TB:tuberculosis; EPTB: Extrapulmonary tuberculosis). The median age of patients was 35.7 (range 4 months to 101 years) with 52.5% in the 20–39 years ( Fig 2 ) range and 1,904(56.2%) giving a male:female ratio of 1.3:1. Children accounted for one-tenth 332 (9.8%) of TB cases. The female gender predominate the TB cases in childhood. Fewer children 141 (69.5%) were sputum positive compared with the adults 2020 (75.5%). Additionally, the most common extrapulmonary sites for TB differed between children and adults. New cases constituted 90.3% (3,059). The second largest category were patients who had previously failed TB treatment (132; 3.9%). Other clinical characteristics are shown in Table 1 and Fig 3 . The figure shows the annual incidence of tuberculosis for children and adult over the 25-year study period. The variables of gender 25 (0.7%), HIV status 26(0.8%), and pre-treatment status 21 (0.6%) have the respective missing data. Fig 3 shows the trend in incidence of TB over the 25 years under review. The annual number of newly diagnosed tuberculosis cases peaked in 1993 and 1998. The time trend incidences were 401.3 and 607.7 per 100,000 population respectively as estimated from hospital records ( S1 Appendix ). The incidence however declined after 2005 with a record of 175.6 cases per 100,000 population in 2014. The incidence in TB among the pediatric population declined by 0.003% (95% CI 0.001% to 0.005%) per year, p = 0.012. A decline of 0.010% (95% CI 0.004% to 0.016%) per year, p = 0.002 was also noted in adults. The mean decline in the entire population with TB was 0.007% (95% CI 0.002% to 0.011%) per year, p = 0.005. Of the 3,467/24,330 (14.2%) screened for HIV, 1,243 were positive giving a seroprevalence rate of 36.70%. The proportion of patients screened for HIV infection over the study period is shown in Table 2 . Of the 3,384 diagnosed with TB, only 1,004 were tested for HIV infection out of which 152 tested positive giving a seroprevalence rate of 15.0%. The prevalence shows unstable pattern (sine wave pattern) with peaked in 2009. Multidrug resistance TB identified via drug sensitivity test was found in only 6 patients (0.2%). The mean annual adherence rates over the 25-year study period was 91.4 (±5.8) %. This outcome was fairly constant except for declines in 1994 (78.5%) and 2013 (74.2%) ( Fig 4 ). The mean successful annual treatment outcome was 75.3(±8.8) %, varied between 91.8% in 1992 and 77.8% in 2017. The nadir of successful outcomes occurred in 1994 (58.9%). The mean annual death was 10.0(±3.6) %. Annual mean unsatisfactory treatment outcome was 14.8 (±7.4) % being at all-time low at the outset of the program (1992) but peaked in 1994 (41.1%) ( Fig 4 ). Successful treatment outcome = cure+ treatment completed; Potentially Unsatisfactory outcome = default + transferred out + treatment failure. Factors with statistically higher adherence rates were females (90.9% vs 87.4%, aOR = 1.4, 95% CI 1.1–1.7, p = 0.005), EPTB (93.7% vs 88.1%, aOR = 1.8, 95% CI 1.3–2.7, p = 0.002), newly diagnosed TB (94.2%, aOR = 3.1, 95% CI 1.4–6.7, p = 0.005) and relapsed TB (95.1%, aOR = 2.8, 95% CI 1.1–7.5, p = 0.037) subjects. Children and those with positive HIV also displayed higher adherence rates ( Table 3 ). About 301 (10.0%) of the patients with treatment adherent could not submit sputum sample at the end of their treatment course. * statistically significant at p < 0 . 05 , ref : Reference level . Significant variables on bivariate analysis for treatment adherence were gender, age, forms of TB, pre-treatment and HIV status were subjected to binary logistic regression with reference indicator as the male gender, adult, pulmonary tuberculosis, pre- treatment default and HIV negative respectively. Factors associated with successful treatment outcome were females (80.4% vs 76.1%, aOR = 1.3, 95% CI 1.1–1.5, p = 0.006), EPTB (83.0% vs 77.1%, aOR = 1.3, 95% CI 1.01–1.7, p = 0.036), newly diagnosed TB (79.1%, aOR = 2.7, 95% CI 1.3–5.4, p = 0.005) and relapsed TB (79.4%, aOR = 2.9, 95% CI 1.3–6.8, p = 0.012) compared with defaulters. Trend was close to significance for children and negative HIV patients ( Table 4 ). *statistically significant at p < 0.05, ref: Reference level. Significant variables on bivariate analysis for successful treatment were gender, age, forms of TB, pre-treatment and HIV status were subjected to binary logistic regression with reference indicator as the male gender, adult, pulmonary tuberculosis, pre- treatment default and HIV negative respectively. Adulthood (9.9% vs 5.4%, aOR = 1.9, 95% CI 1.2–3.1, p = 0.011) and HIV+ (16.4% vs 8.7%, aOR = 2.1, 95% CI 1.3–3.5, p = 0.003) were risk predictors for death ( S2 Appendix ). The male gender, pulmonary tuberculosis, relapse and smear positivity at two and five months of treatment were significantly associated with potentially unsatisfactory outcome ( S3 Appendix ).

Meeting SDG goals and having a world without TB requires tackling socio-economic factors including improving education, community sanitation/hygiene and reducing extreme poverty. It also requires increasing financial support to ensure adherence, equipping motivated trained personnel, providing subsidized/free TB diagnostics, and finally annual evaluation and monitoring of treatment programs throughout each region/country worldwide.

This is a retrospective review of the TB treatment program at Bowen University Teaching Hospital (BUTH) formally called Baptist Medical Centre Ogbomoso (BMCO), Nigeria from 1992 to 2017. In the early 1990s, using the National Tuberculosis Leprosy Control Program (NTBLCP) protocol, the Damien Foundation (Netherlands) in collaboration with Oyo State chose BMCO, a regional referral center, as one of the program’s initial healthcare facilities. Together BMCO/BUTH, the Damien Foundation and the Oyo State government provided free medications, education, diagnostics, counseling, contact tracing and ongoing updates and reviews. All patients were fully evaluated and had all or some of the following laboratory investigations performed depending on their presenting features: white blood cell count, erythrocyte sedimentation rate, tuberculin skin testing (TST), chest and spinal X-rays, sputum microscopy for acid-fast bacilli (AFB) stain and culture, fine needle aspirate for cytology, histology, and GeneXpert from 2015. Since 2006 all TB patients were screened for HIV1 and HIV2 by parallel testing using Determine® Kits (Alere Medical, Chiba-ken Japan), and Uni-GoldTM HIV (Trinity Biotech, Wicklow, Ireland), and tie-break with STAT-PAK® (Chembio Diagnostic System, Medford NY). A patient was confirmed to have active TB if any of the following were found. AFB positive sputum microscopy. TST induration >15 mm and/or ulceration or TST of 10mm plus background history suggestive of TB. AFB positive spinal cerebrospinal fluid (CSF) on microscopy. Spinal gibbus with characteristic x-ray. Fine needle aspiration cytology showing granulomatous giant Langhans cells. Positive histopathological tissue findings of chronic granulomatous inflammation with giant cell and caseous necrosis. Positive GeneXpert test on tissue/body fluids. A definitive diagnosis of TB in children was difficult as most children cannot produce sputum. The NTBLCP workers manual 5 and most recently the Childhood Tuberculosis Desk Guide for diagnosis and management of childhood tuberculosis was used, an adaptation from the NTBLCP workers manual. This entails a composite clinical and laboratory reference standard5 of history and exam plus one or more of the following. Chest X-ray, tuberculin skin test, gastric aspirate, non-response to routine antibiotics in a child with history and exam consistent with TB with subsequent response to TB drugs. All patients diagnosed with TB received directly observed therapy (DOT) daily in the TB Unit where they received individual education about their disease, treatment course and were observed for complications from the disease or treatment. Home visits were carried out prior to discharge with contact tracing and locating patients who defaulted. Patients were followed up by physicians at two weeks; one, two and five months and as needed unless “lost to follow-up” or transferred to another center. Modifications of the treatment protocol did occur over the 25 years in order to follow WHO protocol changes adopted by the NTBLCP. Treatment was divided into categories I and II. Category I was for newly-diagnosed pulmonary TB while category II was for extra-pulmonary TB, treatment failure, or relapse. Between 1992 and 2000, children (<18years) received intramuscular streptomycin instead of oral ethambutol during the intensive/initial 2-month treatment phase. Drug combinations evolved over the years to the current combination of rifampicin, isoniazid, pyrazinamide, and ethambutol (RIPE) for drug-sensitive MTB. Adherence was defined as patients taking ≥80% of the prescribed anti-TB via the DOT intervention or treatment completed. The study used a daily dosing treatment regimen and calculated adherence for the entire treatment course (no separation was made between the intensive and the continuation phase). In addition to DOT, patients were encouraged to eat a balanced diet, sleep alone in well-ventilated rooms during the intensive treatment phase and ensure that their sputa were appropriately disposed of to prevent spread of TB to others. Treatment outcomes as defined by WHO guidelines and the International Union Against Tuberculosis and Lung Diseases (IUATLD) follow [ 3 , 5 ]. Cure: Sputum smears negative on two occasions, one of which must be at treatment end. Treatment completed: Patient completed treatment but last smear unavailable or extrapulmonary TB. Successful treatment is a or b. Relapse: When a patient who had been declared cured later has a positive smear. Treatment failure: Patient remains smear positive ≥5-months after beginning treatment. Default: Patient had ≥1-month of treatment with >2month interruption. Death: Mortality from any cause during the course of receiving treatment. Transfer out: Patient referred out of catchment area and outcome of treatment unknown. Potentially unsatisfactory treatment outcome is a combination of treatment default, transferred out and treatment failure [ 3 ]. Statistical analyses were performed using IBM’s Statistical Package for Social Sciences (SPSS) TM , Version 23.0 for Windows with statistical significance set at p < 0.05. Descriptive statistics for categorical variables were presented as number (%). The significant variables of sociodemographic characteristics of age (adult or children), gender (male or female) and clinical variables of categories of diseases (pulmonary, extrapulmonary disease, pre-treatment status, and HIV status) on bivariate analysis were subjected to binary logistic regression to adjust for co-variance to determine the predictors of successful treatment outcomes, treatment adherence potentially unsatisfactory outcome and death. Interaction effects were built from the statistical perspective (only significant independent variables from the main effects multivariate analyses were interacted). The trend in incidence was assessed for significance using a linear regression model. This retrospective review was approved by the Bowen University Teaching Hospital Human Research Ethics Committee (Approval No. BUTH/REC-029). As a result of the retrospective nature of the study design, informed consent was waived by the ethics committee.

(DOCX) Click here for additional data file. (PDF) Click here for additional data file. (PDF) Click here for additional data file.