|The impact of antibiotics on growth in children in low and middle income countries: systematic review and meta-analysis of randomised controlled trials
|Gough EK, Moodie EE, Prendergast AJ, Johnson SM, Humphrey JH, Stoltzfus RJ, Walker AS, Trehan I, Gibb DM, Goto R, Tahan S, de Morais MB, Manges AR
This review concluded that antibiotics had a growth promoting effect in prepubertal children from undernourished populations in low and middle income countries, particularly increases in weight (ponderal growth). This was a well-conducted review. The results are likely to be reliable but the authors' conclusions may be overstated, given the apparently small effect size of antibiotics.
To determine the impact of oral antibiotic treatment on growth in prepubertal children in low or middle income countries and to identify moderators of any treatment effect.
MEDLINE, EMBASE, Scopus and Cochrane CENTRAL were searched to December 2013 with no language restrictions. Search terms were reported in an appendix. Web of Science was searched for publications that cited the included studies. Reference lists of included studies and relevant reviews were hand searched.
Randomised controlled trials (RCTs) that examined the impact of oral antibiotic treatment on growth in children (aged one month to 12 years) in low or middle income countries were included. Indicators of growth were mean height (cm) or weight (g) at the final follow-up or mean change in height or weight per unit of follow-up time. Eligible comparators included placebo, no treatment or an intervention with no known antimicrobial effect. Studies of anthelmintic treatments were excluded, as were studies of neonates (aged <1 month) and those in which the condition being treated did not depend on the antimicrobial effect of antibiotic treatment.
The included studies were published between 1953 and 2013. Most of the administered antibiotics were classed as being bactericidal with a narrow or broad spectrum (some were classed as bacteriostatic narrow spectrum). Dose regimens varied across the studies. All studies except one were placebo controlled; one study gave controls as no treatment. Conditions treated included malnutrition, infection with Giardia lamblia, diarrhoea (with or without vomiting), environmental enteropathy and prophylaxis in children with HIV. Almost one third additionally provided nutritional supplements to participants in both arms. Mean age of participants ranged from four to 115 months; most were below the age standardised reference population mean for height or weight.
Two reviewers independently assessed studies for inclusion; any discrepancies were resolved by a third reviewer.
Assessment of study quality
Risk of bias was assessed using criteria from the Cochrane Handbook for sequence generation, allocation concealment, blinding, informative censoring and selective outcome reporting.
Two reviewers independently performed the assessment; any discrepancies were resolved by consensus.
Data on growth outcomes (increases in height or weight) were obtained to calculate mean differences and 95% confidence intervals. Mean differences and their 95% confidence intervals were scaled as the average effect for each month of follow-up. Study authors were contacted up to three times for individual patient data. When these data were not available, two reviewers independently extracted data from the papers of included studies; any discrepancies were resolved by consensus. All data were extracted/obtained according to the intention-to-treat principle.
Methods of synthesis
Where individual patient data were obtainable these were analysed using linear mixed models to calculate mean differences and 95% confidence intervals for each study. An individual patient data model was fitted to each study with baseline growth, age, sex, duration of follow-up and duration by treatment interaction as covariates. Mean differences and 95% confidence intervals from the individual patient data analyses were then pooled with those calculated from the aggregate data (from the studies without individual patient data) using random-effects models. Statistical heterogeneity was assessed using the Ι² statistic. Significant heterogeneity was explored using weighted meta-regression analyses and subgroup analyses (further reported in the review). Publication bias was assessed using Egger's test. Sensitivity analyses were conducted via removal of studies one by one from the meta-analysis, removal of studies with children admitted to hospital and by fitting linear mixed models to the studies that provided individual patient data.
Results of the review
Ten RCTs (4,316 participants) were included in the review: five provided individual patient data (1,795 participants) and five provided aggregate data (2,521 participants). Overall risks of bias for the studies were high (one RCT), unclear (four RCTs) and low (five RCTs).
Compared with placebo/no treatment, use of antibiotics led to statistically significantly greater increases in mean weight (MD 23.8 g/month, 95% CI 4.3 to 43.3; 10 RCTs; Ι²=84.4%) and height (MD 0.04 cm/month, 95% CI 0.00 to 0.07; eight RCTs; Ι²=84.8%). Meta-analyses only including data from children aged under two years demonstrated no significant difference between study groups in the height treatment effect; significantly greater increases in weight were observed with use of antibiotics.
Bivariate meta-regression analysis showed that studies conducted in Africa had a significantly larger weight treatment effect than studies conducted in other regions. No other covariates significantly explained the heterogeneity observed in the weight or height meta-analyses.
Removal of one particular study impacted on results for height and this reduced the treatment effect by 50%. No substantial differences were observed in the sensitivity and subgroup analyses relating to weight. No evidence of publication bias was found. Results were reported in greater detail in the review paper.
Antibiotics have a growth promoting effect in prepubertal children from undernourished populations in low and middle income countries; this relates particularly to increases in weight (ponderal growth).
The review question and inclusion criteria were clearly defined. A wide range of relevant data sources (including some unpublished literature) were accessed and no language restrictions were imposed, reducing the risk that relevant studies were missed. All review processes were performed in duplicate, minimising the risk of reviewer error and/or bias. Suitable quality assessment criteria were employed; results showed that study quality was variable.
The statistical methods of synthesis were appropriate and an extensive array of exploratory analyses were conducted. The authors acknowledged that heterogeneity between the studies means that the overall findings have limited generalisability. They also pointed out that they had limited power to identify moderators of treatment effects and emphasised that possible cross levels bias means that care must be taken in extending the treatment modifying effects to the individual level (particularly participant age).
This was a well-conducted review. The results are likely to be reliable but the authors' conclusions may be overstated, given the apparently small effect size of antibiotics.
Implications of the review for practice and research
Practice: The authors did not state any implications for clinical practice.
Research: The authors stated that more research was needed to improve understanding of the biological mechanisms behind antibiotic associated effects on growth, particularly in children aged under two years, acutely malnourished children and children exposed to or infected with HIV.
The leading author was supported by the Vanier Canada Graduate Scholarship. No other funding reported.
Gough EK, Moodie EE, Prendergast AJ, Johnson SM, Humphrey JH, Stoltzfus RJ, Walker AS, Trehan I, Gibb DM, Goto R, Tahan S, de Morais MB, Manges AR. The impact of antibiotics on growth in children in low and middle income countries: systematic review and meta-analysis of randomised controlled trials. BMJ 2014; 348: g2267
Subject indexing assigned by NLM
Anti-Bacterial Agents /therapeutic use; Child; Child Development /drug effects; Child, Preschool; Developing Countries /statistics & Growth Substances /therapeutic use; Humans; Infant; Infant, Newborn; Randomized Controlled Trials as Topic; numerical data
Date bibliographic record published
Date abstract record published
This is a critical abstract of a systematic review that meets the criteria for inclusion on DARE. Each critical abstract contains a brief summary of the review methods, results and conclusions followed by a detailed critical assessment on the reliability of the review and the conclusions drawn.