Pre‐eclampsia is diagnosed on the basis of hypertension (raised blood pressure) with proteinuria (protein in the urine) after 20 weeks' gestation (Davey 1988). This multi‐system disorder complicates between 2% and 8% of all pregnancies (WHO 1988). It is an important cause of maternal and perinatal death worldwide; 10 to 15% of maternal deaths in low‐ and middle‐income countries are associated with the hypertensive disorders of pregnancy (Duley 1992). In the United Kingdom, hypertensive disease of pregnancy remains the second cause of direct maternal deaths (NICE 2001). Even when death is avoided there is the risk of serious morbidity associated with the disease. Mothers can suffer from eclampsia (seizures), renal failure (kidney damage), disseminated intravascular coagulation (blood clotting disorders), myocardial infarction (heart attacks), cerebro‐vascular accident (stroke), retinal detachment (blindness) and the HELLP syndrome (haemolysis, elevated liver enzymes, low platelets). The baby can have intrauterine growth retardation (fail to grow properly within the uterus) or be born prematurely resulting in complications such as hyaline membrane disease (immaturity of the lungs), intraventricular haemorrhages (brain haemorrhages), and necrotising enterocolitis (bleeding in the bowel). Although many women with pre‐eclampsia have no long‐term problems and give birth to healthy babies, any treatment that can prevent or ameliorate pre‐eclampsia will be of great benefit to pregnant women and their babies. Currently, the only cure for the disease is delivery of the baby and the placenta but this leads to many babies being born prematurely and vulnerable to the problems described above.

Several interventions have been suggested for prevention of pre eclampsia. Other Cochrane reviews cover calcium supplementation (Hofmeyr 2003), magnesium supplementation (Makrides 2003), protein intake (Kramer 2003, Kramer 2003a), nutritional advice (Kramer 2003b), salt intake (Duley 2003) and antiplatelet agents (Knight 2003). Antiplatelet agents were found to have small to moderate benefits, and calcium supplementation appears to be beneficial for women at high risk of gestational hypertension and in communities with low dietary calcium intake.

Diuretics are commonly used to treat hypertension in non‐pregnant individuals. They were also formerly used in pregnancy to treat high blood pressure, and delay or prevent the onset of pre‐eclampsia. Historically, the rationale for their use in preventing pre‐eclampsia was based on the incorrect supposition that excess sodium (as salt) intake and retention caused the disease. As diuretics promote excretion of sodium and decrease oedema and blood pressure in non‐pregnant people, it was assumed that they would be beneficial in preventing pre‐eclampsia. In the 1960s, the prophylactic use of thiazide diuretics in pregnancy was advocated to protect the mother from pre‐eclampsia and reduce perinatal mortality and prematurity (Finnerty 1966). This practice became widespread, with up to 40% of pregnant women being given continuous diuretic therapy in pregnancy (Gray 1968). Women with essential hypertension are at increased risk of superimposed pre‐eclampsia, presenting as de novo proteinuria after 20 weeks' gestation. These women in particular were thought likely to benefit from diuretic therapy.

The use of diuretics in pregnancy became controversial with increasing evidence of the reduction of plasma volume (the non‐cellular part of the blood) in pre‐eclampsia (Hays 1985). Women who use diuretics from early pregnancy do not increase their plasma volume to the degree usual in normal pregnancy (Sibai 1984). There were theoretical concerns that diuretics might worsen the existing hypovolaemia (low blood plasma volume) in women with pre‐eclampsia, with adverse effects on the mother and fetus, particularly fetal growth (Gifford 2000). However, there is no firm clinical evidence that diuretics cause harm in pregnancy. Indeed, although the acute blood pressure lowering effect of thiazide diuretics is due to plasma volume contraction, over several weeks renal (kidney) compensatory mechanisms return plasma volume towards normal levels. Cardiac output returns to pretreatment levels, but total peripheral vascular (blood vessel) resistance remains low. The sustained blood pressure lowering effect of thiazide diuretics is thought to involve mobilisation of excess sodium from the arteriolar (blood vessel) wall, with widening of the vessel lumen (Mabie 1986). This might theoretically be of benefit in the pathological vasoconstriction (narrowing of blood vessels) which is an important characteristic of pre‐eclampsia.

There have also been case reports of maternal side‐effects of diuretics in pregnancy including hypokalaemia (low potassium levels), hyponatraemia (low sodium levels), decreased carbohydrate tolerance (increased tendency to diabetes), and pancreatitis. Jaundice and thrombocytopaenia (low platelet levels) have been reported in newborn babies exposed to diuretics in the uterus. It has been suggested that risks of serious side‐effects may have been overstated due to selected case reporting (Collins 1985).

A non‐systematic review of randomised trials of diuretics in pregnancy in nearly 7000 women appeared to show evidence of prevention of pre‐eclampsia (Collins 1985). This significant effect persisted when trials using oedema (which would be decreased by diuretics) as a diagnostic criteria for pre‐eclampsia were excluded. This might suggest a true decrease in the incidence of pre‐eclampsia with diuretics. However, the authors were concerned that the definitions of pre‐eclampsia used depended heavily on an increase in blood pressure. The apparent prevention of pre‐eclampsia might thus be due solely to diuretics lowering the blood pressure, and not be a true effect. The overall perinatal mortality was not reduced. This early example of a review with pooled analysis stated that many thousands of women would need to be included to assess effects of treatment on rare outcomes such as perinatal mortality (Collins 1985). Our review is primarily interested in the prevention of pre‐eclampsia using diuretics as some confusion remains over this question. A systematic review of the data is needed to address this.