Dietary calcium supplementation for prevention of pre-eclampsia and related problems: a systematic review and commentary

Authors


Dr GJ Hofmeyr, Effective Care Research Unit, East London Hospital Complex, University of the Witwatersrand/University of Fort Hare/Eastern Cape Department of Health, Private bag X9047, East London 5200, South Africa. Email gjh@global.co.za

Abstract

Background  Calcium supplementation during pregnancy may reduce the risk of hypertensive disorders of pregnancy.

Search strategy  We searched the Cochrane Pregnancy and Childbirth Group trials register and the Cochrane Central Register of Controlled Trials (March 2006).

Selection criteria  Randomised trials comparing at least 1 g of calcium daily during pregnancy with placebo. Eligibility and trial quality were assessed.

Data collection and analysis  Data were extracted and analysed using Review Manager software.

Main results  Twelve studies (15 528 women) were included, all of good quality. Most women were at low risk and had low dietary calcium. High blood pressure was reduced with calcium supplementation rather than placebo (11 trials, 14 946 women: relative risk [RR] random effects model 0.70; 95% CI 0.57–0.86), as was pre-eclampsia (12 trials, 15 206 women: RR 0.48; 95% CI 0.33–0.69). The effect was greatest for women at high risk (five trials, 587 women: RR 0.22; 95% CI 0.12–0.42) and for those with low baseline calcium intake (seven trials, 10 154 women: RR 0.36; 95% CI 0.18–0.70). There was heterogeneity, with less effect in the larger trials. The composite outcome maternal death or serious morbidity was reduced (four trials, 9732 women: RR 0.80; 95% CI 0.65–0.97). The syndrome of haemolysis, elevated liver enzymes and low platelets was increased (two trials, 12 901 women: RR 2.67; 95% CI 1.05–6.82). There was no overall effect on the risk of preterm birth or stillbirth or death before discharge from hospital.

Conclusions  Calcium supplementation appears to reduce the risk of pre-eclampsia and to reduce the rare occurrence of the composite outcome ‘maternal death or serious morbidity’. There were no other clear benefits or harms.

Commentary  We present the hypothesis that adequate dietary calcium before and in early pregnancy may be needed to prevent the underlying pathology responsible for pre-eclampsia. We suggest that the research agenda be redirected towards calcium supplementation at a community level.

Introduction

High blood pressure with or without proteinuria is a major cause of maternal death and morbidity worldwide,1,2 as well as perinatal morbidity and mortality. Hypertension has been estimated to complicate 5% of all pregnancies and 11% of first pregnancies. Half the women with hypertension have pre-eclampsia. Hypertensive disorders account for up to 40 000 maternal deaths annually.3 For this reason, strategies to reduce the risk of hypertensive disorders of pregnancy have received considerable attention.4–7

A widely accepted definition of gestational hypertension is a diastolic blood pressure greater than or equal to 90 mmHg appearing after 20 weeks of gestation. Gestational hypertension and significant proteinuria (2+ by dipstick testing, greater than or equal to 300 mg/24 hours, or greater than or equal to 500 mg/l) usually indicate the presence of pre-eclampsia. Urine protein/creatinine ratio is used increasingly as a measure of proteinuria.8 Predictors of poor outcome include low gestational age and high levels of proteinuria.9

An inverse relationship between calcium intake and hypertensive disorders of pregnancy was first described in 1980.10 This was based on the observation that Mayan Indians in Guatemala, who traditionally soak their corn in lime before cooking, had a high calcium intake and a low incidence of pre-eclampsia and eclampsia. A very low prevalence of pre-eclampsia had been reported from Ethiopia where the diet, among other features, contained high levels of calcium.11 These observations were supported by other epidemiological and clinical studies12–16 and led to the hypothesis that an increase in calcium intake during pregnancy might reduce the incidence of high blood pressure and pre-eclampsia among women with low dietary calcium. An association has been found between pre-eclampsia and hypocalciuria,17 lower urine calcium to creatinine ratio,18 hypocalcaemia,19 lower plasma and higher membranous calcium,20 lower dietary milk intake,21 and between eclampsia and hypocalcaemia.22

Low calcium intake may cause high blood pressure by stimulating either parathyroid hormone or renin release, thereby increasing intracellular calcium in vascular smooth muscle14 and leading to vasoconstriction. A possible mode of action for calcium supplementation is that it reduces parathyroid release and intracellular calcium and so reduces smooth muscle contractility. By a similar mechanism, calcium supplementation could also reduce uterine smooth muscle contractility and prevents preterm labour and delivery.23 Calcium might also have an indirect effect on smooth muscle function by increasing magnesium levels.24

This hypothesis was tested in several randomised trials commencing in the late 1980s, which suggested a promising beneficial effect for calcium supplementation. Early systematic reviews highlighted the need for larger trials to assess the effects on important clinical outcomes in addition to pre-eclampsia and preterm delivery, such as perinatal mortality.4,25 A subsequent systematic review5 came to more enthusiastic conclusions, but this optimism was not confirmed by a large trial in the USA among women with adequate dietary calcium intake.26 For women at low risk, a modest reduction of around 20% in the relative risk (RR) of pre-eclampsia remained possible but required further confirmation.27 New evidence is now available from a large trial conducted in communities with low dietary calcium intake.28

This paper summarises all the available evidence. It is based on a Cochrane review,29 which is a regularly updated electronic publication. The objective of this review was to assess the effect of calcium supplementation during pregnancy on the risk of high blood pressure and related maternal and fetal/neonatal adverse outcomes.

Methods

Criteria for considering studies for this review

We included published, unpublished, and continuing trials with random allocation to calcium supplementation during pregnancy versus placebo. We excluded quasi-random designs. The participants were pregnant women less than 35 weeks of gestation, regardless of their risk of hypertensive disorders of pregnancy or their previous calcium intake. Women with diagnosed hypertensive disorders of pregnancy were excluded. The interventions were supplementation with at least 1 g of calcium compared with placebo. Studies with no placebo were excluded.

Outcome measures for the women were as follows:

  • 1High blood pressure as defined by trial authors, with or without proteinuria. Ideally, high blood pressure was defined as diastolic blood pressure greater than or equal to 90 mmHg, an increase in systolic blood pressure of 30 mmHg or more, or in diastolic blood pressure of 15 mmHg or more.
  • 2High blood pressure with significant proteinuria, as defined by trial authors. Ideally, proteinuria was defined as 2+ by dipstick testing, greater than or equal to 300 mg/24 hours, or greater than or equal to 500 mg/l. Although the strict definition of pre-eclampsia includes confirmation of no hypertension or proteinuria outside pregnancy, for pragmatic reasons, we used the above criteria to define pre-eclampsia.
  • 3Maternal death or serious morbidity: A composite outcome of death or at least one measure of serious morbidity. Serious morbidity included Eclampsia, renal failure, syndrome of haemolysis, elevated liver enzymes and low platelets (HELLP syndrome), and admission to intensive care. In addition, each individual outcome is presented.
  • 4Other maternal outcomes included placental abruption, caesarean section, severe pre-eclampsia as defined by trial authors, proteinuria, and woman’s hospital stay for 7 days or more.

Outcome measures for the child were as follows:

  • 1Stillbirth or death before discharge from hospital.
  • 2Death or severe neonatal morbidity.
  • 3Preterm birth (birth before 37 weeks of estimated gestation).
  • 4Neonate small for gestational age, as defined by trial authors.
  • 5Other neonatal outcomes: birthweight less than 2500 g, admission to neonatal intensive care unit, and neonate in intensive care unit for 7 days or more.
  • 6Long-term outcome for the child: childhood disability, systolic blood pressure greater than 95th percentile during childhood, and diastolic blood pressure greater than 95th percentile during childhood.

The prespecified primary outcomes were high blood pressure, pre-eclampsia, preterm birth, admission to neonatal intensive care unit, and stillbirth or neonatal death.

Search strategy for identification of studies

We searched the Cochrane Pregnancy and Childbirth Group trials register (March 2006).

The Cochrane Pregnancy and Childbirth Group’s trials register is maintained by the Trials Search Coordinator and contains trials identified from: (1) quarterly searches of the Cochrane Central Register of Controlled Trials (CENTRAL); (2) monthly searches of MEDLINE; (3) handsearches of 30 journals and the proceedings of major conferences; and (4) weekly current awareness search of a further 37 journals.

Trials identified through the searching activities described above are given a code (or codes) depending on the topic. The codes are linked to review topics. The Trials Search Coordinator searches the register for each review using these codes rather than keywords.30 We searched the Cochrane Central Register of Controlled Trials (The Cochrane Library, Issue 4, 2005) using the terms calcium AND pregnan* AND (hypertens* or blood press*).

No language restrictions were applied. Abstracts were considered for inclusion if the methods were adequately described.

Assessment of studies and data extraction

Two reviewers independently assessed the methodological quality and other inclusion criteria of the identified trials. Disagreements were resolved by consensus. Whenever possible, additional information was obtained from the trialists. Data were extracted onto preprepared sheets, cross-checked by two reviewers independently, and then entered into the Review Manager software.31

Statistical analyses

We compared categorical data using RRs and their 95% CI. We tested for statistical heterogeneity between trials using the I2 statistic, with values greater than 50% indicating significant heterogeneity. In the absence of significant heterogeneity, data were pooled using a fixed effects model. If there was significant heterogeneity, a random effects model was used, and an attempt was made to identify potential sources of heterogeneity32,33 based on subgroup analyses by risk of hypertensive disorders, baseline dietary calcium intake, trial quality, and trial size. For continuous data, we calculated pooled estimates of effect size from a weighted average, with weight based on the inverse of the variance.34

Prespecified subgroup analyses were based on whether during trial entry women (1) had low dietary calcium intake (assessed by 24 hour dietary recall in most studies, as defined by trial authors, or if not defined, mean intake less than 900 mg/day) or adequate dietary calcium intake (as defined by trial authors, or if not defined, mean intake greater than and equal to 900 mg/day) or (2) were at low/average risk of hypertensive disorders, at high risk, or if risk status was unclear. Women at high risk were defined by the trial authors definitions and included as being teenagers, women with previous pre-eclampsia, women with increased sensitivity to angiotensin II, or women with pre-existing hypertension. Primiparity alone was not regarded as a high risk factor. Subgroup analyses are limited to the primary outcomes.

Results

Twelve trials were included (CPEP 1997,26 WHO 2006,28 Belizan et al.1991,35 Villar et al. 1987,16 Lopez-Jaramillo et al.1989,36 Villar, Repke 1990,37 Lopez-Jaramillo et al. 1990,38 Sanchez-Ramos et al. 1994,39 Purwar et al. 1996,40 Lopez-Jaramillo et al. 1997,41 Crowther et al. 1999,42 and Niromanesh et al. 200143), with a total of 15 528 women (Table 1). Twenty-four studies were excluded (see Cochrane review29 for references). The reasons for exclusion were: allocation to treatment group was not randomised (seven studies), there were no relevant outcome data (six studies), the interventions were not at least 1 g/day calcium compared with placebo (four studies), information about the study methods was unclear or inadequate (two studies), women recruited for treatment rather than prevention of hypertensive disorders of pregnancy (two studies), more than 20% of participants were excluded from the analysis (one study), the trial appeared to have been cancelled (one study), and the paper was a comment rather than a trial report (one study).

Table 1.  Characteristics of included trials
Study IDMethodsParticipantsInterventionsAllocation concealment
Belizan et al. 199135Multicentre trial. Numbered, sealed opaque envelopes, containing randomisation codes. Of 593 women (calcium) vs 601 women (placebo) enrolled, 577 women vs 588 women had at least partial follow upNulliparous women, less than 20 weeks pregnant; blood pressure < 140/90 mmHg (mean of five measurements); no present or past disease; not taking medication; normal oral glucose tolerance tests2 g calcium as 500 mg calcium carbonate tablets vs identical looking placebo tablets. Compliance was 84% (calcium) and 86% (placebo)Adequate
CPEP 199726Numbered treatment packs in computer-generated simple randomisation sequence. Loss to follow up: calcium 132/2295 women vs placebo 121/2294 womenPregnant nulliparas (45% black, 35% non-Hispanic white, and 17% Hispanic white). Passed compliance test (took 75% of placebo over 6–14 days); blood pressure 134/84 mmHg or less; urine protein dipstick negative or trace; 13–21 weeks pregnant2 g/day elemental calcium as calcium carbonate or placebo. All women took 50 mg calcium per day as normal supplementationAdequate
Crowther et al.199942Central telephone randomisation, stratified by centre using variable blocks. Double-blindNulliparous women; singleton pregnancy; less than 24 weeks of gestation; blood pressure < 140/90 mmHg; expected to give birth at a collaborating centreCalcium carbonate 1.8 g daily or lactose placebo tablets from 20–24 weeks till birthAdequate
Lopez-Jaramillo et al.198936Assigned independently in sequence using a table of random numbers. Fourteen women who delivered at 36–38 weeks were excluded from the trial report and were included in this reviewNulliparous women; age 25 years or less; certain menstrual dates; clinic attendance before 24 weeks of gestation; residence in Quito; normotensive; no medical disorders; not taking medication or vitamin/mineral preparationsCalcium supplementation with four calcium gluconate tablets daily, each containing 500 mg elemental calcium from after 23 weeks of gestation till delivery vs identical placebo tabletsAdequate
Lopez-Jaramillo et al. 199038Double-blind placebo-controlled trial. Stated ‘Each patient was assigned independently in sequence’ and ‘All women completed the study’Healthy nulliparous women with positive roll-over test at 28–30 weeks of gestational age – judged at high risk for gestational hypertension2000 mg elemental calcium daily from 28–32 weeks till delivery vs placebo starch tabletsUnclear
Lopez-Jaramillo et al. 199741Double-blind placebo-controlled trial. Fourteen women withdrew after randomisation: 12 women by change to another hospital or private medical doctor, two women by noncomplianceNulliparous women; age < 17.5 years; first prenatal visit before 20 weeks of gestation; certain menstrual dates; residency in Quito for at least 1 year; blood pressure ≤120/80 mmHg; average daily calcium intake in this population is 51% of the recommended dietary allowanceElemental calcium 2 g daily as calcium carbonate from 20 weeks vs placebo tablets 
Niromanesh et al. 200143Double-blind placebo-controlled trialWomen at high risk for pre-eclampsia: positive ‘roll-over’ test and at least one risk factor for pre-eclampsia; 28–32 weeks pregnant; blood pressure < 140/90 mmHg. Not defined as low or adequate calcium intake (from the table dairy intake appears to be about 200 ml + 400 g/day)Elemental calcium 2 g daily (500 mg 6-hourly) or placebo, coded by the pharmacyAdequate
Purwar et al. 199640Double-blind placebo-controlled trial. Allocated by means of a computer-generated randomisation list. After randomisation, 11/201 (5.5%) women lost to follow up (calcium six, placebo five)Nulliparous women; normal single viable pregnancy; known dates; antenatal clinic before 20 weeks; intending to deliver in the same institute; normal glucose tolerance test; no hypertension; no underlying medical disorders. Calcium intake mean 336 mg (calcium) and 352 mg (placebo group) per dayOral calcium containing 2 g elemental calcium daily, compared with identical placebo tablets, taken from 20 weeksAdequate
Sanchez-Ramos et al. 199439Double-blind placebo-controlled trial. Randomisation by computer-generated list. Outcome data entered before breaking the code. 4/33 women allocated calcium lost to follow upNormotensive nulliparas; positive roll-over test (281/1065) and positive angiotensin II infusion test at 20–24 weeks of gestation (67/281)Calcium supplementation with 2 g/day elemental calcium as 500 mg calcium carbonate tablets vs identical placebo tabletsAdequate
Villar et al. 198716Double-blind placebo-controlled trial. Random numbers in closed envelopesInclusion criteria: nulliparous or primiparous; known menstrual dates; age 18–30 years; singleton pregnancy; negative roll-over test. Mean calcium intake at 26 weeks was as follows: calcium group 1129 (SD 736) and placebo group 914 (478)2 g elemental calcium as 500 mg calcium carbonate tablets vs placebo tablets. All women were prescribed prenatal vitamin tablets containing 200 mg calcium and 100 mg magnesium per dayAdequate
Villar, Repke 199037Double-blind placebo-controlled trial. Allocation by opaque envelopes; ordered by a computer-generated list of random numbersPregnant women of 17 years or younger; no underlying medical disorders; most women were nulliparous with known last menstrual period and singleton pregnancy2 g elemental calcium as 500 mg calcium carbonate tablets vs placebo tablets. All women were prescribed prenatal vitamin tablets containing 200 mg calcium and 100 mg magnesium per dayUnclear
WHO 200628Double-blind placebo-controlled trial. Randomisation stratified by centre, with computer-generated blocks of six to eight. Allocation by consecutively numbered treatment packsPopulations with median daily calcium intake less than 600 mg; primiparous women less than 20 weeks pregnant. Exclusion criteria: renal disease or urolithiasis; parathyroid disease; blood pressure >140 mmHg systolic or >90 mmHg diastolicChewable calcium carbonate tablets with 500 mg elemental calcium, three daily, or identical placebo tablets from enrolment till deliveryAdequate

For the included trials, about one-third of the women had adequate calcium intake (5275 adequate calcium intake; 10 253 low calcium). Most women were at low risk during trial entry (14 923 women). The reasons for classification as high risk were age (two trials, 765 women) and positive ‘roll-over’ and/or angiotensin II infusion tests (three trials, 149 women). There was no heterogeneity in results between the trials classified as high risk. The dose of calcium evaluated was primarily 1.5–2 g daily. One study conducted long-term follow up of the children,35,44 although only children born to the subset of women recruited in private clinics were eligible to be contacted. Data from follow up in another trial were excluded from the review due to large losses to follow up.26,45

Overall, these trials were of high quality with 83% (12 914/15 528 women) of the data from the two large well-conducted trials. For one small study, there is a large discrepancy in the size of the two allocated groups,38 the reason for which is unclear. In some trials, individual denominators were not given for specific outcomes. Where it was clear that the outcomes were not measured in the entire group, we have adjusted the denominators accordingly.

There was significant heterogeneity for four outcomes: pre-eclampsia, high blood pressure, preterm birth, and birthweight less than 2500 g. In view of this, we used a random effects model for these four outcomes.

High blood pressure with or without proteinuria

Overall, there was less high blood pressure associated with calcium supplementation rather than placebo (11 trials, 14 946 women: RR random effects model 0.70; 95% CI 0.57–0.86). The reduction in RR appeared greatest for women at high risk of developing pre-eclampsia (four trials, 327 women: RR 0.47; 95% CI 0.22–0.97) and for those with low baseline dietary calcium (six trials, 9894 women: RR 0.47; 95% CI 0.29–0.76), although the CI overlap.

Pre-eclampsia

There was a halving in the RR of pre-eclampsia associated with calcium supplementation rather than placebo (12 trials, 15 206 women: RR 0.48; 95% CI 0.33–0.69) (Figure 1). Again, this reduction appeared greatest for women at high risk of pre-eclampsia (five trials, 587 women: RR 0.22; 95% CI 0.12–0.42) and for those with low baseline calcium intake (seven trials, 10 154 women: RR 0.36; 95% CI 0.18–0.70), although the CI overlap. For women with adequate dietary calcium, the reduction in pre-eclampsia does not achieve statistical significance (four trials, 5022 women: RR 0.62; 95% CI 0.32–1.20).

Figure 1.

Effect of calcium supplementation during pregnancy on pre-eclampsia, subgrouped by dietary calcium intake.

When the trials are grouped by both dietary calcium intake and study size, the effect size appears to be associated most strongly with study size (in the small studies with <400 women, RRs 0.21 for the low calcium trials, and 0.26 for the adequate calcium trials and in the large studies 0.87 and 0.70, respectively) (Figure 2).

Figure 2.

Effect of calcium supplementation during pregnancy on pre-eclampsia, ranked by sample size.

Maternal death or serious morbidity

The RR of having the composite outcome maternal death or serious morbidity was reduced for women allocated calcium supplementation compared with placebo (four trials, 9732 women: RR 0.80; 95% CI 0.65–0.97) (Figure 3).

Figure 3.

Effect of calcium supplementation during pregnancy on ‘maternal mortality and severe morbidity’, subgrouped by dietary calcium intake.

HELLP syndrome

HELLP syndrome was reported only by the two largest studies.26,28 The RR was higher for women allocated calcium supplementation rather than placebo (two trials, 12 901 women: RR 2.67; 95% CI 1.05–6.82).

Maternal death

Maternal deaths were reported only in one trial.28 One death occurred in the calcium group and six in the placebo group, a difference which was not statistically significant (RR 0.17; 95% CI 0.02–1.39).

Other maternal outcomes

Other maternal outcomes are summarised in Table 2. There were no clear differences between the groups for any of these outcomes.

Table 2.  Maternal and child outcomes
Outcome measureTrialsParticipantsRelative risk (95% CI) size
  • *

    Random effects model.

Outcome for the women
High blood pressure (with or without proteinuria)1114 9460.70 (0.57, 0.86)*
Pre-eclampsia1215 2060.48 (0.33, 0.69)*
Maternal death/serious morbidity497320.80 (0.65, 0.97)
Placental abruption514 3090.86 (0.55, 1.34)
Caesarean section714 7100.95 (0.88, 1.01)
Proteinuria (with or without hypertension)183121.01 (0.88, 1.15)
Severe pre-eclampsia183020.74 (0.48, 1.15)
Eclampsia212 9010.73 (0.41, 1.27)
HELLP syndrome212 9012.67 (1.05, 6.82)
Intensive care unit admission183120.84 (0.66, 1.07)
Maternal death183120.17 (0.02, 1.39)
Outcome for the children
Preterm birth1014 7510.81 (0.64, 1.03)*
Birthweight <2500 g814 3590.84 (0.68, 1.03)*
Neonate small for gestational age313 0911.10 (0.88, 1.37)
Admission to neonatal intensive care unit413 4061.05 (0.94, 1.18)
Stillbirth or death before discharge1015 1410.89 (0.73, 1.09)
Systolic blood pressure greater than 95th percentile at age 7 years15140.59 (0.39, 0.91)
Diastolic blood pressure greater than 95th percentile at age 7 years15140.81 (0.50, 1.31)

Outcome for the child

Outcomes for the child are summarised in Table 2. There was no overall effect on the RR of a stillbirth or the child dying before discharge from hospital (ten trials, 15 103 women: RR 0.89; 95% CI 0.73–1.09). There was no overall effect on preterm birth (ten trials, 14 751 women: RR 0.81; 95% CI 0.64–1.03). However, the RR of preterm birth was reduced among women at high risk of developing pre-eclampsia recruited to four small trials (568 women: RR 0.45; 95% CI 0.24–0.83).

At about 7 years of age, diastolic blood pressure greater than 95th percentile was reduced (one trial,35,44 514 women: RR 0.59; 95% CI 0.39–0.91). There were no clear differences between the groups for any other outcomes.

Discussion

Calcium supplementation with at least 1 g of calcium is associated with a halving in the RR of pre-eclampsia, with the CIs putting the true effect anywhere between a 31% reduction and a 67% reduction. Women with an adequate dietary intake of calcium were the only subgroup for which this was not statistically significant; nevertheless, the point estimate for this subgroup of women was a 38% reduction. The greatest reduction in risk appeared to be for women at high risk and for those with low baseline dietary calcium intake. There was also a 30% reduction in the risk of gestational hypertension, with again the greatest effect being among women at high risk and those with a low calcium intake during trial entry.

There was no overall effect on the RR of preterm birth, although a moderate reduction associated with calcium supplementation remains possible. For women at high risk, there was a halving in the RR of preterm birth. As the number of women in this subgroup is small, these data should be interpreted with caution and they may merely reflect the play of chance.

Although the RR of pre-eclampsia was reduced, this was not clearly reflected in any reduction of severe pre-eclampsia, eclampsia, or admission to intensive care. Nevertheless, the point estimates for these outcomes favour calcium supplementation and so moderate reductions in these outcomes remain possible. Also, the RR of the composite outcome ‘maternal death or severe morbidity’ was reduced by 20% (95% CI 35–3%) for women allocated calcium supplementation. In the two trials reporting HELLP syndrome, which is included in the composite outcome of death and severe morbidity, the RR of this outcome was increased with calcium supplementation.

No adverse effects of calcium supplementation have been recorded in the trials reviewed. There is little information about the long-term follow up of children within these trials, with the exception of a reduction in childhood systolic hypertension in the one study to measure this outcome. There is no information about any possible changes in the use of healthcare resources associated with calcium supplementation. It would seem plausible that a reduction in gestational hypertension and pre-eclampsia might lead to fewer antenatal visits, less admission for antenatal care, and fewer inductions of labour. However, these trials do not provide data on these outcomes.

As illustrated in the analysis of trials ranked by sample size, heterogeneity in the results seems to be largely associated with study size, with the small studies having the most strongly positive results. As the small studies tended to recruit high-risk women, at least some of the heterogeneity may be explained by calcium having a greater effect for high-risk women. An alternative explanation may be that there is publication bias, with small studies that failed to report an effect for calcium supplementation not being published. This suggestion should be interpreted with caution, as the analysis by study sample size was based on a post hoc hypothesis.

There are no clear differences in any other outcomes, although for several outcomes the CIs approach statistical significance. So for caesarean section, a small (5%) reduction in RR associated with calcium supplementation is possible, and for stillbirth and death before discharge from hospital, an 11% reduction is possible, although for both these outcomes no effect or a small increase in risk have not been excluded.

The relatively modest effects found in the large trials contrast with the large differences reported in observational studies between populations with adequate and those with low dietary calcium intake.10–12 Possible explanations for these differing results include dietary calcium being a marker for other aetiological factors, that starting supplementation in the middle trimester of pregnancy may be too late to be fully effective, and that the earlier observational studies may have exaggerated the potential benefits of calcium, perhaps due to unresolved confounding factors.

The finding of reduced childhood hypertension needs replication but, if true, has far-reaching implications for public health. Although based on only a partial follow up in one study, this finding is supported by a very limited follow up in two studies,26,45 as well as observational46 and animal studies.47

Conclusion

Taken together, these trials show a halving in the RR of pre-eclampsia. This is reflected in more modest reductions in the RR of gestational hypertension and of maternal death or serious morbidity. There are no clear effects on other substantive outcomes for either mother or child during discharge from hospital.

Implications for practice

The reduction in pre-eclampsia and in ‘maternal death or severe morbidity’ supports the use of calcium supplementation during pregnancy for women with low dietary intake. For women with adequate dietary intake, the evidence of benefit from calcium supplementation is not conclusive.

Implications for research

Any future trials should collect information about the use of health service resources, as well as other clinical outcomes. The minimum dose in this review was 1 g of calcium daily. It would be relevant to assess whether supplementation via dietary modification, for women with low calcium intake, has the same benefits as the tablets administered in these trials.

Further research is also needed to provide reassurance that calcium supplementation during pregnancy does not have any adverse effects on the children exposed when in utero and to verify whether it reduces blood pressure in childhood.

Commentary

Possible mechanisms by which calcium might reduce pre-eclampsia include inhibition of the pathological processes, such as endothelial damage, underlying the development of pre-eclampsia or alternatively just prevention of its manifestation by reducing blood pressure. This latter explanation is supported by a study of the biochemical measures commonly elevated in pre-eclampsia (G. J. Hofmeyr et al., unpubl obs.), which showed no effect of calcium supplementation on serum urate, platelet count, and the urine protein/creatinine ratio. It is also supported by the large WHO trial28, which reported no effect of calcium supplementation on the RR of gestational proteinuria (RR 1.01; 95% CI 0.88–1.15).

How can the reduction in pre-eclampsia associated with calcium supplementation be reconciled with the lack of evidence supporting any effect on either proteinuria or other markers for pre-eclampsia? Our hypothesis is that calcium supplementation in the second half of pregnancy reduces blood pressure and thus the diagnosis and clinical manifestations of pre-eclampsia, without having any significant effect on the underlying pathologies. This hypothesis might also explain our finding in this review of an increase in the risk of HELLP syndrome in association with calcium supplementation. If calcium supplementation reduces only blood pressure, this would reduce the frequency of diagnosis of pre-eclampsia and some of its hypertension-related complications, while the effects on other organ systems, particularly the endothelium and platelets might continue, the more so in the calcium supplementation group in which fewer early deliveries for hypertension would take place.

Calcium supplementation during pregnancy for women with deficient dietary calcium intake offers modest benefit for individual women. However, a public health policy of calcium supplementation during pregnancy is unlikely to have a major impact on the incidence of pre-eclampsia. The main reason is that communities where dietary calcium is inadequate are often poor, and the women at highest risk of pre-eclampsia often attend antenatal clinics late in pregnancy or not at all and so would not benefit from strategies directed at those attending for antenatal care. We therefore suggest that future research be directed towards evaluation of improving calcium intake at a population level, for example by food fortification, rather than at an individual level.

Acknowledgements

We thank the trial authors who have contributed additional data for this review and Jose Villar for constructive criticism of the protocol.

Potential conflict of interest

G.J.H. is a collaborator in the WHO Calcium Trial28, which is included in this review. He did not participate in decisions regarding that trial.

Ancillary