South Africa 1997b.

Methods	Sealed, numbered, opaque envelopes. Nursing sister not involved in clinical care then made up the allocated solution (4 mL). 8 women excluded (9%) as delivered without receiving antihypertensive therapy. CFU ‐ B, blinding ‐ B.
Participants	88 women at least 28 weeks' gestation, DBP > 110 mmHg or DBP > 100 mmHg for 30 mins.
Interventions	Ketanserin: 500 mL crystalloid IV over 15 min, then bolus 10 mg ketanserin in 4 mL IV. Bolus repeated every 20 min, until DBP 90 mmHg, to a maximum of 4 doses. Hydralazine: 500 mL crystalloid IV over 15 min, then bolus 5 mg hydralazine in 4 mL IV. Bolus repeated every 20 min, until DBP 90 mmHg, to a maximum of 4 doses.
Outcomes	Woman: death, persistent high BP (DBP > 90 mmHg after 4 bolus injections), delivery for fetal distress, caesarean section. Baby: death.
Notes	Trial stopped by 'monitoring committee', reason not stated.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	“Patients were then assigned to receive either 5 mg dihydralazine or 10 mg of ketanserin according to random numbers which had been previously generated by computer.”
Allocation concealment (selection bias)	Low risk	“Successively numbered sealed, opaque envelopes contained the instructions for the preparation of each new patient’s medication. A nursing sister not involved in the management of the particular prepared patient.”
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Participant/clinician appeared to be blinded, “In either case, the managing physician was given a syringe with four millilitre of clear fluid.”
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Not reported.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	The study was stopped early on the advice of the monitoring committee “the study stopped after the analysis of 88 consecutive patients who qualified for the study.”  8 of these were not included in the analysis – 6 patients who qualified for the study were not randomised because their BP was lower than 90 mm Hg after the fluid overload and 2 patients did not receive the medication after randomisation – in both the fetal heart rate pattern deteriorated to such a degree that emergency caesarean sections were performed.
Selective reporting (reporting bias)	Low risk	All expected outcomes reported.
Other bias	Unclear risk	“The groups were comparable regarding maternal age, gravidity and gestation age”  However, study stopped early ‐ reasons not described and no baseline characteristics for 8 patients who were not included in the analysis.