S‐Ramos 1994.

Methods	Double‐blind placebo‐controlled trial. 4/33 allocated calcium lost to follow‐up.
Participants	Women attending a hospital in Jacksonville, Florida, USA between 1989 and 1993. University hospital serving low‐income population Normotensive nulliparas; positive roll‐over test (281/1065) and positive angiotensin II infusion test at 20‐24 weeks' gestation (67/281). 67 were allocated to calcium (33) or placebo (34). Exclusion criteria: factors increasing the risk of gestational hypertension, including renal disease, collagen vascular disease, diabetes mellitus, chronic hypertension, multifetal pregnancy.
Interventions	Calcium supplementation with 2 g per day elemental calcium as 500 mg calcium carbonate tablets, vs identical placebo tablets. Compliance checked with electronic pillboxes. Compliance was 79% vs 81%.
Outcomes	Gestational hypertension (BP at least 140/90 mmHg on 2 occasions 4‐6 hours apart, on bedrest in hospital); pre‐eclampsia (gestational hypertension + proteinuria: 1+ or 300 mg/24 hours); severe pre‐eclampsia (pre‐eclampsia plus 1 of BP at least 160 mmHg systolic or 110 mmHg diastolic; proteinuria at least 5 g/24 hours; oliguria < 400 mL per day; elevated liver enzymes; thrombocytopenia < 100,000/microlitre; pulmonary oedema; severe epigastric pain). Birthweight (calcium 3245 (SD 414) vs placebo 3035 (542) g); mean gestational ages (35.6 vs 34.4 weeks); 5 minute Apgar < 7 (1/29 vs 1/34); cord arterial pH (7.25 (0.07) vs 7.20 (0.07)); fetal growth impairment (2/29 vs 4/34).
Notes	Jacksonville, Florida, USA. University hospital serving low‐income population Source of funding: not stated. CoI: not stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomisation by computer‐generated list
Allocation concealment (selection bias)	Low risk	Adequate
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Double‐blind. Tablets were prepared by a pharmaceutical company and were identical with respect to weight, size, flavour and appearance.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Double‐blind. Tablets were prepared by a pharmaceutical company and were identical with respect to weight, size, flavour and appearance.
Incomplete outcome data (attrition bias) All outcomes	Low risk	< 10% attrition
Selective reporting (reporting bias)	Low risk	Expected outcomes reported
Other bias	Low risk	Data entered before breaking the code. ITT analysis. 4/33 in the calcium group lost to follow‐up so 29 in calcium and 34 in placebo, however even if the 4 lost to follow‐up had PIH, results would still have significantly favoured the calcium group.