López 2005.

Methods	Study design: RCT Location: Chile Setting: Public Health Clinic, Santiago, Chile Recruitment period: not stated
Participants	Inclusion criteria: healthy pregnant women with gingivitis aged 18 to 42; single gestation; ≤ 22 weeks of gestation; gingival inflammation with ≥ 25% of sites with bleeding on proving, and no sites with clinical attachment loss > 2 mm Exclusion criteria: < 18 natural teeth; indication of prophylactic antibiotics for invasive procedures; diabetes previous to pregnancy and the intention to deliver at a hospital other than that of the study Mean age (± standard deviation (years)): Group A = 25.54 ± 5.41, Group B = 24.98 ± 4.55 (P = 0.31) Gestational age: ≤ 22 weeks Previous preterm low birth weight (%): Group A = 3.44, Group B = 7.47 (P = 0.009) Number randomised: n = 870 Number evaluated: n = 834 (attrition n = 36: loss to follow‐up n = 5, withdrawal from treatment and study n = 9, spontaneous abortion n = 10, preterm delivery n = 11, stillbirth n = 1)
Interventions	A) Antenatal periodontal treatment (n = 580): plaque control instructions (toothbrushes and mouthrinse daily), supra and subgingival scaling, and crown polishing before 28 weeks of gestation + maintenance therapy (oral hygiene instruction and supragingival plaque removal by instrumentation) every 2 to 3 weeks until delivery B) Postnatal periodontal treatment (n = 290): monitoring 2 to 3 times during pregnancy All women: repeated periodontal examinations after 30 weeks of gestation No information on the expertise of the dental professional who administered intervention
Outcomes	Preterm birth (< 37 weeks gestational age with birth weight < 2500 g following spontaneous labour and/or rupture of the membranes, regardless of route of delivery); low birth weight; gestational age; infant birth weight; plaque; bleeding on probing; probing depth; clinical attachment loss
Funding	Not stated
Notes	290 women were required to detect a significant difference of preterm/low birth weight between groups with 80%. To increase statistical power a 2:1 allocation of participants to the treatment and control groups was adopted. Intention‐to‐treat principle was applied
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Quote: "Randomisation was done equalizing gingivitis as the relevant variable, and the percentage of bleeding on probing sites was selected as the variable describing gingivitis... One woman of each group of the three was selected by rolling a dice"
Allocation concealment (selection bias)	Unclear risk	Not stated
Blinding of participants and personnel (performance bias) All outcomes	High risk	Not feasible
Blinding of obstetric outcome assessment (detection bias)	Low risk	Obstetrician researcher who obtained pregnancy outcome data from hospital records was masked to the periodontal characteristics of the patients. Staff involved in labour and delivery management decisions had no knowledge that the patients were participating in a research study. However, it is not clear whether periodontal outcome assessment was blinded
Incomplete outcome data (attrition bias) All outcomes	Low risk	Attrition rates were similarly low and balanced across groups (1.7% versus 1.3%)
Selective reporting (reporting bias)	Low risk	Expected outcome reported
Other bias	High risk	More participants in the control group had a history of previous preterm/low birth weight compared to the treatment group (P = 0.009)