Kumar 2009.

Methods	Randomised, double‐blind, placebo‐controlled trial
Participants	Women were recruited between January 2005 and December 2007 at a hospital in New Delhi, India. Healthy normotensive primigravid women with uncomplicated single pregnancy; pregnancy 12 to 25 weeks' gestation, known date of the last menstrual period, and intention to deliver at Lok Nayak Hospital, New Delhi. Study population had a low dietary calcium. Exclusions: multiple pregnancy, polyhydramnios, fetal malformations, diabetes, chronic hypertension, renal disease, cardiovascular disease, urolithiasis, or BP of 140/90 mmHg or higher at first visit or at enrolment.
Interventions	4 tablets (2 g calcium or placebo) were taken daily.
Outcomes	Pre‐eclampsia (SBP > 140 mmHg and DBP > 90 mmHg on 2 occasions 4 hours apart after 20 weeks' pregnancy in women normotensive previously, together with proteinuria > 300 mg/24 h or 1+ on a clean‐catch dipstick in the absence of urinary infection); eclampsia; preterm delivery; caesarean section. Baseline characteristics comparable.
Notes	Imbalanced groups: 290 allocated to calcium, 262 to placebo group. 17 and 11 lost to follow‐up so 273 and 251 analysed respectively. See below. Source of funding: University Grant Commission, New Delhi, India. CoI: not stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Unclear risk	Simple randomisation sequence developed manually
Allocation concealment (selection bias)	Low risk	Coded numbers assigned to treatment packets and distributed to participants using the random number sequence.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind. Calcium and placebo tablets were identical. Randomisation code broken after completion of the trial.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Double‐blind. Calcium and placebo tablets were identical. Randomisation code broken after completion of the trial.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Attrition < 10%
Selective reporting (reporting bias)	Low risk	Expected outcomes reported.
Other bias	High risk	Imbalance in size of groups. The authors were contacted regarding the imbalance and they explained that a random sequence was generated for 600 participants (unblocked) but recruiting was stopped at 552 participants and so 48 numbers remained unallocated.