Metronidazole (all routes except local only)

Study	Type of data	Exposure measurement	Outcome assessment	Adjustment
Abdel-Salam (All routes), 2000	case control	Exposure data collected from 3 sources: a post-paid structured questionnaire sent to the parents requesting drugs taken during pregnancy, according to gestational months; maternal prenatal care logbook (in which obstetricians must record all prescribed drugs); nurses visited non-responding families.	The Hungarian Congenital Abnormality Registry (HCAR), in which notification by physicians of cases with Congenital anomalies is mandatory (including infant deaths and usual stillborn fetuses). Controls were selected from the National Birth Registry of the Central Statistical Office.	Controls matched according to sex, birth week, and district of parents' residence. POR adjusted for maternal age, birth order, and maternal diseases.
Alexander, 2001	case control	Maternal exposure data were obtained by interviews using a structured questionnaire. A pharmacist (blind to numbers of mothers and status) reviewed the list of medications reported by the mothers, but none were recognized as known topo-II inhibitors.	Case ascertainment was based on referrals to treating hospitals but was population based.	Controls were selected with the same center, same gender and similar dates of birth as soon as each case mother had been interviewed (normally shortly after diagnosis).
Andrews, 2003	randomized controlled trial	Subjects were instructed to take medications or placebo. Medication compliance was estimated at this visit by examining the pill containers for remaining capsules.	Pregnancy outcomes were assessed at the respective institutions.	No adjustment. Randomisation.
Caro-Paton, 1997	case control	Maternal drug use was assessed by the attendant doctor during an interview with the mothers carried out within the first 3 days of delivery.	Not specified.	A control is defined as the next new-born to a malformed child, with the same sex and born in the same hospital.
Czeizel, 1998	case control	Exposure data collected from 3 sources: a post-paid structured questionnaire sent to the parents requesting drugs taken during pregnancy, according to gestational months; maternal prenatal care logbook (in which obstetricians must record all prescribed drugs); nurses visited non-responding families.	The Hungarian Congenital Abnormality Registry (HCAR), in which notification by physicians of cases with Congenital anomalies is mandatory.	Controls were matched with every index case according to sex, birth week, and district of parents’ residence. Odds ratio adjusted for potential confounders, including maternal age, birth order, use of other drugs and maternal disorders.
Diav-Citrin, 2001	prospective cohort	Details of exposure were collected during pregnancy, using a structured questionnaire.	A telephone interview with the woman using a structured questionnaire was conducted to obtain details on pregnancy outcomes. In case of malformation, verifications were performed: medical documents, certified pediatrician contacted the mother again or the child’s physician was contacted.	None.
Heinonen (Controls exposed to penicillins), 1977	prospective cohort	Extensive information on drugs taken during pregnancy was collected before the birth of the child. Data on drug were recorded at each antenatal visit and confirmed, with few exceptions, by the attending physician or by the hospital or clinic record.	The baby was examined by a pediatrician every day during the first week after delivery or weekly if the child remained in the hospital longer. The mothers were interviewed for what was termed an interval history when the child was 4, 8, 12 and 24 months old, thereafter, annually.	None.
Heinonen (Controls unexposed, NOS), 1977	prospective cohort	Extensive information on drugs taken during pregnancy was collected before the birth of the child. Data on drug were recorded at each antenatal visit and confirmed, with few exceptions, by the attending physician or by the hospital or clinic record.	The baby was examined by a pediatrician every day during the first week after delivery or weekly if the child remained in the hospital longer. The mothers were interviewed for what was termed an interval history when the child was 4, 8, 12 and 24 months old, thereafter, annually.	None.
Jeffcoat, 2003	prospective cohort	Patients were randomly assigned to one of three treatment groups. An University of Alabama at Birmingham (UAB) research pharmacist, who provided a double packet with coding information for each patient, generated the randomization code.	Trained research obstetric nurses abstracted maternal records to determine the predefined age at delivery. These abstractors were completely blinded as to the periodontal status or the patients’ periodontal treatment.	None
Kamal, 2018	prospective cohort	All positive cases were treated with metronidazole administered orally 500mg twice a day for 7 days.	Delivery data, neonatal outcome, and medical records were obtained from hospital discharge records to collect information about labor complications and neonatal outcomes.	None. Exclusion criteria: Women with systemic diseases (e.g., diabetes mellitus, hypertension, ...), uterine anomalies, incompetent cervix, excess uterine contraction, oversized uterus (polyhydramnios, multiple pregnancies), abruption placenta, cervical bacterial infection, previous history of preterm labor, premature rupture of membrane, ...
Koss, 2012	retrospective cohort	Prenatal care and hospital records of mothers were reviewed to determine whether metronidazole treatment occurred. Chart reviewers were blind to the purpose of the study.	Congenital anomalies were ascertained from electronic birth certificates, an electronic Neonatal Intensive Care Unit database, and the local perinatal data system. Gestational age and birth weight were identified from mothers’ hospital records.	Multivariable analysis using logistic regression to adjust for variables with P values of < 0.10 in bivariate analysis: maternal age, race, alcohol consumption, and smoking; for preterm/LBW: addition of educational status, marital status, prior preterm birth, and prepregnancy BMI; for malfo: addition of preexisting diabetes mellitus, gestational diabetes mellitus, and alcohol consumption.
Mann, 2009	retrospective cohort (claims database)	Treatment status was determined by using Medicaid outpatient pharmacy billing records, which provided drug names and dates dispensed.	Children with intellectual disability (ID) were identified based on 3 sources: the Medicaid file, a data file from the South Carolina Department of Education (DOE) and the enrollment in services from the South Carolina Department of Disabilities and Special Needs (DDSN).	Singleton only. Exclusion of children who were diagnosed with a known or likely cause of Intellectual disabilities. Inverse propensity-score–weighted models including maternal age, race, alcohol use and tobacco use.
Medveczky, 2004	case control	Exposure data collected from 3 sources: a post-paid structured questionnaire sent to the parents requesting drugs taken during pregnancy, according to gestational months; maternal prenatal care logbook (in which obstetricians must record all prescribed drugs); nurses visited non-responding families.	The Hungarian Congenital Abnormality Registry (HCAR), in which notification by physicians of cases with Congenital anomalies is mandatory (including infant deaths and usual stillborn fetuses). Controls were selected from the National Birth Registry of the Central Statistical Office.	Controls matched according to sex, birth week, and district of parents' residence. Multivariable unconditional logistic regression model for maternal age, birth order, and and employment status.
Morgan, 1978	retrospective cohort	Not specified.	Not specified.	None.
Muanda a (Controls exposed to penicillins), 2017	retrospective cohort (claims database)	The Quebec Public Prescription Drug Insurance database (drug name, start date, dosage and duration). Filled prescription of antibiotics within the QPC has been validated against maternal reports with high positive and negative predictive value (PPV 86.7% and NPV 92.3%).	MCMs were identified in the Régie de l’assurance maladie du Québec (RAMQ) and the provincial hospital discharge database (MedEcho) databases. MCMs included in the QPC have also been validated against medical charts with high PPV (78.1%) and NPV (94.2%).	Potential confounders selected a priori. The following covariates were included: maternal age, maternal marital status, receipt of social assistance during pregnancy, education level in years, and area of residence); maternal chronic co-morbidities; endometriosis and maternal infections; healthcare utilization in the year before pregnancy; calendar year of delivery and infant gender.
Muanda a (Controls unexposed, NOS), 2017	retrospective cohort (claims database)	Exposure identified from the Quebec Public Prescription Drug Insurance database (drug name, start date, dosage and duration). Filled prescription of antibiotics within the QPC has been validated against maternal reports with high positive and negative predictive value (PPV 86.7% and NPV 92.3%).	Major congenital malformations were identified in the Régie de l’Assurance Maladie du Québec (RAMQ) medical file and the Quebec hospitalization archives (MedEcho) databases and defined according to International Classification of Diseases ICD-9 and ICD-10 codes.	Adjusted for socio-demographic variables (such as maternal age, education level, ...), maternal chronic co-morbidities (such as chronic hypertension, diabetes mellitus, epilepsy, ...), endometriosis and maternal infections, healthcare utilization in the year before pregnancy, year of delivery and infant gender. Singleton only. Pregnancies with exposure to known teratogens were excluded.
Muanda b (Controls exposed to penicillins), 2017	nested case control	The province’s Public Prescription Drug Insurance Plan database (drug name, start date, dosage and duration).	Régie de l’assurance maladie du Québec (RAMQ (diagnoses, medical procedures...), the provincial hospital discharge database (MEDÉCHO) (in-hospital diagnoses and procedures, and gestational age) and the Institut de la statistique du Quebec (birth weight).	Controls matched by gestational age (within 3 d) and year of pregnancy. ORs adjusted for maternal age, area of residence, social assistance during pregnancy, education level, marital status, maternal chronic comorbidities, maternal infections, use of other anti-infective agents, use of health services, and history of planned and spontaneous abortions.
Muanda b (Controls unexposed, NOS), 2017	nested case control	The province’s Public Prescription Drug Insurance Plan database (drug name, start date, dosage and duration).	Régie de l’assurance maladie du Québec (RAMQ (diagnoses, medical procedures...), the provincial hospital discharge database (MEDÉCHO) (in-hospital diagnoses and procedures, and gestational age) and the Institut de la statistique du Quebec (birth weight).	Controls matched by gestational age (within 3 d) and year of pregnancy. ORs adjusted for maternal age, area of residence, social assistance during pregnancy, education level, marital status, maternal chronic comorbidities, maternal infections, use of other anti-infective agents, use of health services, and history of planned and spontaneous abortions.
Piper, 1993	retrospective cohort (claims database)	Prescription drug use during pregnancy was ascertained from Medicaid pharmacy files, which included the date the prescription was dispensed and the days of supply.	Tennessee birth, death, and fetal death certificates. A trained nurse who had no knowledge of the exposure status of the pregnancy reviewed the infant's delivery record using a structured questionnaire.	Matched by year of delivery, hospital of delivery, mother's age and maternal race. Adjusted in marital status, time when prenatal care began, and number of prenatal visits.
Pombo-de-Oliveira, 2006	case control	Mothers were interviewed in person in the hospital with the aid of a well-structured questionnaire divided in two major sections, from which one related to exposures during pregnancy.	Mothers were interviewed in person in the hospital with the aid of a well-structured questionnaire divided in two major sections. The first part of the questionnaire was devoted to childbirth and nursing.	All controls were age frequency matched. All analyses were adjusted for region of residence, sex, income, maternal age, and birth weight.
Rosa, 1987	case control	Medicaid invoices for prescriptions.	Medicaid invoices for diagnoses.	None.
Ross, 2003	case control	Exposure information was collected from mothers using a structured telephone questionnaire. All prescription drugs recorded in the medical record were abstracted, including data for the trimester of pregnancy the drug was prescribed based upon gestational ages recorded in medical records.	Signed medical record release forms were obtained and complete copies of medical records were requested. Data were abstracted from medical records by two registered nurses using a structured protocol.v	Controls matched to cases by birth date (within one year) and telephone area and exchange. Adjusted for maternal age, education, and income.
Santos, 2011	nested case control	The Régie de l’assurance maladie du Québec (RAMQ) database, that provides information on prescriptions filled for residents insured by Québec’s Public Drug Insurance Plan. Data are available for physician-based diagnoses, therapeutic procedures, characteristics of patient, and healthcare providers.	The Med-Echo database that records hospitalisation data such as gestational age for deliveries for all Quebec residents. It also records the gestational age for planned abortions, miscarriages, and deliveries.	Adjusted for maternal age, maternal place of birth, gestational age, maternal RAMQ drug plan status; maternal health before and during pregnancy: comorbidities (infections, diabetes, hypertension, ...), number of medication used, number of different prescribers, emergency and doctor visits... Singletons only.
Santos, 2012	nested case control	The Régie de l’assurance maladie du Québec (RAMQ) database, that provides information on prescriptions filled for residents insured by Québec’s Public Drug Insurance Plan. Data are available for physician-based diagnoses, therapeutic procedures, characteristics of patient, and healthcare providers.	The Med-Echo database that records hospitalisation data such as gestational age for deliveries for all Quebec residents. It also records the gestational age for planned abortions, miscarriages, and deliveries.	Adjusted for maternal age, maternal place of birth, gestational age, maternal RAMQ drug plan status; maternal health before and during pregnancy: comorbidities (infections, diabetes, hypertension, ...), number of medication used, number of different prescribers, emergency and doctor visits... Singletons only.
Sorensen, 1999	retrospective cohort (claims database)	All prescription data have been stored in the Pharmaco-Epidemiological Prescription Database of North Jutland.	Pregnancy outcome data were collected from the Danish Medical Birth Registry; and the Danish Hospital Discharge Registry was used to identify congenital abnormalities which accordingly were all diagnosed after birth.	Adjustment for maternal age, birth order, gestational age and smoking habit.
Thapa, 1998	retrospective cohort	Medicaid pharmacy files were used to identify prescriptions of metronidazole filled by mothers of the cohort children.	Review of medical records. For each potential case, a trained nurse-abstractor, unaware of drug exposure status, reviewed medical records to confirm diagnosis and eligibility and completed a structured study abstract form.	Adjusted for maternal age, rural county of residence, white race, unwed status, maternal education, and being first born.
Werler, 2014	case control	Mothers were interviewed by telephone within 12 months after delivery about medication use, including indication, product, timing, and frequency.	Study subjects were ascertained from birth defect registries in Massachusetts, New York, and North Carolina. Mothers were then interviewed and an orthopedist reviewed the children’s pediatric and orthopedic records (77% agreed). Controls identified from birth certificates or hospital records.	Odds ratios were adjusted for study site, first born, sex, body mass index, and maternal smoking through Lunar Months 4. Confounding by the use of multiple medications was assessed. Analysis with exclusion of first-degree history of clubfoot.
Wright, 2012	case control	Medical information including antibiotic use were abstracted from hospital charts. Mothers were telephoned to confirm antibiotic exposure and to obtain consent to contact obstetrical care providers to validate antibiotic use data.	Medical information were abstracted from hospital charts.	None.
Zagorodnikova, 2017	cohort	Not specified.	Not specified.	None (disease-controlled population).

master protocol