Research and analysis
Annex 1: Table 1
Updated 28 June 2021
Download CSV 6.07 KB
| | Associations between use of macrolide antibiotics during pregnancy and adverse child outcomes: A systematic review and meta-analysis (Fan et al. 2019) Focused on studies with comparison groups exposed to similar antibiotics (mostly penicillins and cephasporins) | Prenatal Exposure to Macrolides and Risk of Congenital Malformations: A Meta‑Analysis. (Mallah et al. 2019) Three comparison groups: Group 1: infants unexposed to any medicine in utero Group 2: infants exposed to non-macrolide antibiotics or non-teratogens in utero Group 3: mixed population of Groups 1 and 2 | |||
|---|---|---|---|---|---|
| Findings | • Fan and colleagues conducted a review of 10 observational studies and 11 RCTs assessing fetal and child outcomes in 228,556 study participants. • The analysis suggested that prescription of macrolides in pregnancy was associated with an increased risk of miscarriage (pooled OR 1.82, 95% CI 1.57–2.11, three studies, I2 = 0%), cerebral palsy and/or epilepsy (OR 1.78, 1.18–2.69; one study), epilepsy alone (OR 2.02, 1.30–3.14, one study; OR 1.03, 0.79–1.35, two studies), and gastrointestinal malformations (OR 1.56, 1.05–2.32, two studies) compared with alternative antibiotics (1st trimester). • no difference in risk was identified for major malformations (pooled OR 1.13, 0.99-1.29, three studies) or cardiovascular malformations (pooled OR 1.14, 0.8-1.62, four studies), all with 1st trimester exposure. • No association found for musculoskeletal malformations (pooled OR 1.18, 0.95-1.47, 2 studies) | • Mallah and colleagues conducted a review of 21 observational studies conducted between 1990-2019 assessing effects on macrolides antibiotics on congenital malformations in live births. • The analysis suggested a weak association between exposure to macrolides and any type of congenital malformation (OR 1.06, 95% CI 1.01–1.10) when compared to a mixed population (group 3). • This association for any congenital malformation was also observed for fetal exposure limited to the first trimester (ORgroup3 1.06, 95% CI 1.01–1.11, 21 studies) and when restricted to evidence from 17 cohort studies (OR 1.07, 95% CI 1.02–1.13). • Digestive system malformations were found to be weakly associated with prenatal exposure to macrolides (ORgroup3 1.14 [95% CI 1.02–1.26] based on 9 studies). • The musculoskeletal system was also found to be potentially affected (ORgroup2 1.21 [95% CI 1.08–1.35], 4 studies and (ORgroup3 1.15 [95% CI 1.05–1.26] 6 studies). • No association was found for cardiovascular malformations: ORgroup2 0.87, 0.81-0.95, 7 studies) | Not set | Not set | Not set |
| Research question | To determine the effects of macrolide treatment during pregnancy on fetal and child outcomes. | To assess the relation between prenatal exposure to macrolides and occurrence of congenital malformations. | Not set | Not set | Not set |
| Literature search, study selection, and quality analysis | The search strategy involved the use of PubMed, Embase, Cochrane Library, Conference Proceeding Citation Index-Science and ClinicalTrials.govto find included studies. Various search terms specific to the research question and the requirements of the database were applied to identify primary research articles. Bias was assessed using Cochrane Collaboration’s tool for assessing the risk of bias in randomised trials and ROBINS-I for observational studies. The selection criteria for all included and excluded studies are outlined and described. Furthermore, the quality of 10% of the studies was methodologically and independently assessed by two reviewers, achieving 81% of an inter-observer agreement. | The studies were searched in MEDLINE, Embase, five regional bibliographic databases of the World Health Organization, the Open Access Thesis and Dissertations, and Conference Proceeding Citation Index. Various search terms specific to the research question and the requirements of the database were applied to identify primary research articles. The quality of studies was assessed using Newcastle–Ottawa scale. No additional framework was reported for assessment of bias for observational studies. The selection criteria for all included and excluded studies are outlined and described. Furthermore, data extraction and analysis were independently performed by two epidemiologists. | Not set | Not set | Not set |
| Outcome measures and combination of studies: | All fetal and/or childhood outcomes were considered. Randomised controlled trials (n=9) and observational studies(cohort=12) were included in the review according to different eligibility criteria indicated and analysed separately. Data were pooled to estimate pooled ORs for each adverse outcome using a random-effects meta-analysis, considering the heterogeneity among studies. | Congenital malformations in live births only were considered. Case-control studies (n=4) and observational studies(cohort=17) were included in the review and their data were analyzed together using pooled ORs by weighting the log RRs and log ORs for cohort and case-control studies, respectively | Not set | Not set | Not set |
| Main results and tests of significance | Maternal exposure to macrolide antibiotics was associated with an increased risk of miscarriage, while evidence of its association with cerebral palsy and epilepsy was inconsistent. | Prenatal use of macrolides in early pregnancy was weakly associated with congenital malformations and was primarily limited to musculoskeletal and digestive systems. | Not set | Not set | Not set |
| Comments | • Focused on studies with comparison groups exposed to other antibiotics (mostly penicillins and cephalosporins) • Did indicate which study contributed to each pooled estimate • Finding for miscarriage based on three studies but dominated by Muanda and others (2017a) | • Included studies with mixed/unexposed comparison groups. • Did not restrict to first trimester exposure • Included studies solely investigating risk of pyloric stenosis which likely explains the association seen for digestive system malformations, • Included the most recent study using same population (such as Källén 2014, not Källén 2005). • Did not indicate which studies were included in pooled estimates (just numbers of studies). | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |
| Not set | Not set | Not set | Not set | Not set | Not set |