Childhood Tuberculosis Epidemiology in Urban Central Ethiopia: Death Predictors, Disease Determinants, and Unfavorable Treatment Outcomes
No Thumbnail Available
Date
2024-06
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Background: Limited evidence exists on the epidemiology of childhood tuberculosis (TB) in
Ethiopia. Even though the TB treatment success rate is one of the most important global metrics
for assessing the implementation of the End TB Strategy, little is known about the magnitude and
determinants of unfavorable TB treatment outcomes that affect TB treatment success rates
among children in Ethiopia. Additionally, despite Ethiopia's high national BCG coverage and
studies showing the vaccine's effectiveness against TB meningitis (TBM), the disease remains a
persistent problem.
Objectives: To describe the epidemiology of drug-susceptible TB and identify predictors of TB
death, identify determinants of TB disease, assess the magnitude and predictors of unfavorable
TB treatment outcomes, and evaluate the effectiveness of BCG vaccine against TBM among
children aged 16 and under in Urban Central Ethiopia.
Methods and materials: This study was conducted across healthcare facilities in Addis Ababa,
Adama, and Bishoftu from September 25 to June 24, 2022. A retrospective cohort design was
used to address the following specific objectives: describing TB epidemiology, identifying
predictors of death, and assessing the magnitude and determinants of unfavorable treatment
outcomes. Additionally, 1:1 matched case-control and 1:4 unmatched case-control designs were
used to identify determinants of TB disease and assess BCG vaccine effectiveness against TBM,
respectively. Sample sizes were statistically determined. Thirty-two healthcare facilities—26 in
Addis Ababa, three in Bishoftu, and three in Adama—were randomly selected. All children
treated for TB at these facilities were included in the descriptive epidemiology and predictors of
death analyses, as well as in the study on unfavorable treatment outcomes. In the matched case-
control study, cases (children with TB) were randomly selected from treated children identified
through TB registries. Controls (children never diagnosed with TB) were sequentially chosen
from the same facilities. Data on TB cases were collected through TB record reviews and phone
interviews with parents or caregivers. Data on controls were obtained through face-to-face
interviews with parents or caregivers. Descriptive epidemiology was presented using relative
frequency tables and graphs. Statistical analyses included Cox proportional hazards regression
(challenged with extended Cox regression) to compute adjusted hazard ratios (aHR) for death
predictors, conditional logistic regression for matched adjusted odds ratios (mORadj) of TB
xi
disease determinants, log-binomial analysis for risk ratios (aRR) of unfavorable outcomes, and
unconditional logistic regression for adjusted odds ratios (aOR) to assess BCG vaccine
effectiveness against TBM. This study was approved by the Institutional Review Board of the
College of Health Sciences, Addis Ababa University (protocol number: 057/19/SPH).
Results: Data from 640 children aged 16 years and under who underwent treatment for TB were
analyzed. Among them, 80 (12.5%) were under two years old, and a resurgence occurred starting
from around 12 years of age. Most of the enrolled children, 557 (87.0%), had no known
household TB contact. Out of 519 children with an identified place of stay, 396 (76.3%) were
attending school or daycare before being diagnosed with TB. Thirty-six (5.6%; 95% confidence
interval (CI) = 4.0–7.7%) of the 640 children died during the course of TB treatment. Among
those who died, nine (25%) were under two years old. Factors such as HIV infection (aHR = 4.2;
95% CI = 1.9–9.3), undernourishment (aHR = 4.2; 95% CI = 2.2–10.48), age below 10 years
(aHR = 4.1; 95% CI = 1.7–9.7), and relapsed TB (aHR = 3.7; 95% CI = 1.1–13.1) increased the
likelihood of death during TB treatment.
Children not vaccinated with BCG at birth or within two weeks of birth (mORadj = 2.11; 95%
CI = 1.28–3.48), those who lived with a TB-sick family member (mORadj = 4.28; 95% CI =
1.95–9.39), those who lived with a smoking family member (mORadj = 3.15; 95% CI = 1.07–
9.27), and HIV-infected children (mORadj = 8.71; 95% CI = 1.96–38.66) were more likely to
develop TB than their counterparts. A post-estimation analysis indicated that children who were
BCG-vaccinated at birth or within two weeks of birth had a lower risk of TB than their
unvaccinated counterparts until the age of 15, but there was no difference found between the two
groups when they turned 16.
Out of 640 children, 42 (6.6%; 95% CI = 4.8–8.8%) had unfavorable TB treatment outcomes,
with 31 (73.8%; 95% CI = 58.0–86.1%) occurring after the first two months of treatment
initiation. Children under ten years old (aRR = 2.69; 95% CI = 1.56–4.61), those with relapsed
TB (aRR = 3.19; 95% CI = 1.79–5.70), undernourished children at TB diagnosis (aRR = 2.68;
95% CI = 1.53–4.71), and HIV-infected children (aRR = 2.62; 95% CI = 1.50–4.59) had a higher
risk of unfavorable TB treatment outcomes than their counterparts. However, among children
who completed the first two months of TB treatment, relapsed TB was not significantly
associated (aRR = 2.81; 95% CI = 0.96–8.22) with unfavorable outcomes, while the remaining
xii
factors retained significance. No significant association was detected between BCG vaccination
and the risk of TBM (aOR = 0.46; 95% CI = 0.11–1.84).
Conclusions: Most children contract TB from the community, with morbidity resurging around
age 12, possibly due to waning BCG vaccine effectiveness. Despite a pooled global death rate of
0.9% in treated childhood TB patients, the death rate in Urban Central Ethiopia is high. Risk
factors for death and unfavorable outcomes include being under 10, suggesting redefining the
high-risk age category from 5 to 10 years. BCG vaccination is protective of childhood TB but
wanes with age. Second-hand smoking increases childhood TB risk, underscoring the need for
public health policies to reduce children's exposure to tobacco smoke. HIV remains a significant
morbidity and mortality risk factor, raising concerns about the coverage and efficacy of TB
preventive treatments. The rate of unfavorable TB treatment outcomes aligns with the WHO's
milestone, staying below 10%, but nearly three-quarters of these outcomes occur during the
continuation phase, highlighting the need for extended risk-focused follow-up. Undernutrition
persisting into the continuation phase of TB treatment predicts unfavorable outcomes, indicating
the need for nutritional interventions throughout both the intensive and continuation phases of
childhood TB treatment. BCG was not found to significantly protect against TBM compared to
other types of TB combined.
Recommendations: Expand TB contact tracing in children to include school communities and
maintain universal newborn BCG immunization. Consider legislation against smoking in
households with children. Include children under 10 as a high-risk group for TB-related death in
TB guidelines. Evaluate the coverage and effectiveness of TB preventive therapy intervention in
HIV-infected children. Children under 10, those who are HIV positive, and undernourished
children should be carefully assessed, treated, monitored, and provided with necessary support,
regardless of the phase of TB treatment they are in. This study recommends conducting larger
studies to assess the effectiveness of BCG vaccination against TBM compared to other types of
TB
Description
Keywords
Childhood TB, epidemiology, survival, disease determinants, treatment outcomes