A comparison of maternal calcium and magnesium levels in pre-eclamptic and normotensive pregnancies: an observational case–control study

Authors


Abstract

Objective

Supplementing pregnant women at high risk of developing pre-eclampsia with calcium may reduce the incidence of the disease. This study examines differences in serum and hair concentrations of calcium and magnesium between women with pre-eclamptic and normotensive pregnancies.

Design

Observational case–control study.

Setting

Two teaching hospitals in Cape Town, South Africa.

Population

Women with pre-eclamptic (N = 96) or normotensive (N = 96) pregnancies, who delivered a single, live infant.

Methods

Demographic and current pregnancy details were retrieved from clinical notes. Each participant completed a dietary questionnaire. Venous blood samples were taken from each participant to assess serum calcium and magnesium concentrations. Hair samples were obtained from all participants and calcium and magnesium levels were measured by inductively coupled plasma optical emission spectrometry (ICPOES).

Main outcome measure

Hair and serum calcium and magnesium concentrations were compared between women with pre-eclamptic and normotensive pregnancies.

Results

Diet and socio-economic status in the two groups were similar. There was no significant difference in the hair calcium level between women with pre-eclamptic [1241 parts per million (ppm); range, 331–4654 ppm] and normotensive (1146 ppm; range, 480–4136 ppm) pregnancies (P = 0.5). Hair calcium levels in both groups were not affected by HIV infection.

Conclusion

Woman with pre-eclampsia showed no difference in chronic calcium status relative to normotensive women. This finding does not support the current belief that the mechanism by which calcium supplementation reduces the risk of developing pre-eclampsia is by correcting a nutritional deficiency.

Introduction

Pre-eclampsia and eclampsia are leading causes of maternal mortality and morbidity in South Africa.[1] Although much research has been performed to investigate interventions to reduce the incidence of pre-eclampsia, at present it seems that this condition cannot be prevented. Studies investigating calcium supplementation of pregnant women at risk of developing pre-eclampsia have shown that this intervention reduces the incidence and severity of the disease.[2] Although the evidence surrounding calcium supplementation is conflicting, the most beneficial effect has been demonstrated in women living in low dietary calcium settings.[2]

Low serum calcium may predispose women to pre-eclampsia for the following reasons: (1) by increasing parathyroid hormone release and thus increasing vascular smooth muscle intracellular calcium, which promotes vasodilatation; (2) by stimulating renin release, which increases angiotensin II levels; (3) by decreasing serum magnesium levels, inducing vasoconstriction in vascular smooth muscle; (4) by diminishing the effect of endothelial nitric oxide synthase, a calcium-dependent enzyme with vasodilatory action; and, finally (5) by reducing circulating prostacyclin, a calcium-dependent enzyme and a potent vasodilator.

In the current literature, the chronic calcium status of woman recruited to studies investigating the relationship between calcium and pre-eclampsia was not investigated prior to the commencement of supplementation.[3-6] Previous studies that reviewed the relationship between calcium and pre-eclampsia limited their assessment of calcium status to daily dietary calcium content, serum calcium measurements or a dietary survey.[3, 7] Serum calcium concentration is affected by numerous factors and is an unreliable indicator of chronic calcium status. Hair analysis is an accurate and well-recognised method of determination of long-term nutritional status.[8-10] By comparing hair calcium levels between women with pre-eclamptic and normal pregnancies, more information about the relationship between calcium and pre-eclampsia may be obtained. There is little information about the status of elements in the hair of pregnant women. Although the effect of calcium supplementation on hair calcium levels in pregnant women has been described, the effect of pregnancy on hair calcium uptake has not been studied.

Calcium metabolism may be affected by serum magnesium concentrations and nutritional factors. Serum calcium and magnesium are related as their concentrations usually increase or decrease simultaneously.[11] The influence of HIV infection on nutritional status is recognised in an extensive published literature.[12, 13] Insufficient caloric and nutrient intake associated with HIV infection may alter calcium intake and long-term calcium status. For this study, it was important to assess both serum and hair magnesium concentrations and to document HIV status.

Methods

A case–control study was conducted in the Maternity Centre at Groote Schuur Hospital and Mowbray Maternity Hospital in Cape Town, South Africa, between January and November 2010. All women of 18 years and older with pre-eclampsia [defined as hypertension (any single systolic blood pressure reading ≥160 mmHg or diastolic blood pressure reading ≥110 mmHg, or two systolic readings ≥140 mmHg or two diastolic readings ≥90 mmHg taken 4 hours apart) plus significant proteinuria (2 + proteinuria on reagent strip or sulphosalicylic acid test 4 hours apart, or a 24-hour urine protein collection ≥300 mg)] after 20 weeks of gestation in a previously normotensive and aproteinuric patient, who had delivered a live infant in a period not exceeding 96 hours from birth to enrolment, were eligible for the study.[14] All healthy normotensive woman who had delivered a live infant in a period not exceeding 96 hours from birth to enrolment were eligible to be recruited as controls. Control participants were selected by matching them to recruited ‘cases’ using ethnicity, gravidity, age ±2 years and gestational age at delivery ±3 weeks as criteria. Strict exclusion criteria were applied (Table 1). Gestational age was determined with the most accurate obstetric estimate, using menstrual history, symphysis–fundal height at the booking visit or ultrasound scan, where available. Women who had undergone chemical straightening and other hair treatments containing hydrogen peroxide were excluded from the study.

Table 1. Inclusion and exclusion criteria
 CasesControls
Inclusion criteria

≥18 years of age

Pre-eclamptic[10]

Singleton pregnancy

≥18 years of age

Normotensive and aproteinuric

Singleton pregnancy

Exclusion criteria

Unbooked pregnancy

Eclampsia

Multiple pregnancy

Intrauterine death

Antenatal vitamin or mineral supplementation

Chronic hypertension

Diabetes mellitus

Renal disease

Known disorders of metabolism or digestion

Known thyroid or adrenal disease

Bleaching of hair during pregnancy

Eating disorders

The questionnaire used in this study was piloted and adjusted before being utilised. Informed consent was provided by each participant. With the assistance of a trained interviewer, each participant provided answers to a piloted questionnaire.[10] The questionnaire comprised six sections: demographic details, booking details of the index pregnancy, delivery details, general medical and health details, a dietary questionnaire and details relating to hair care. HIV status was obtained from the voluntary counselling and testing information noted on the antenatal card.

Blood pressure (using the fifth Korotkoff sound) and dipstick urine protein measurements were recorded from a participant's observation chart to confirm the diagnosis of pre-eclampsia. Four blood pressure readings taken 4 hours apart over 12 hours, and four measurements of urinary protein taken 4 hours apart, were documented by the interviewer. In those participants in whom urine protein had been quantified, a measurement of more than 0.3 g per 24 hours was taken as being significant. In cases in which women presented to hospital with normal blood pressure and no detectable protein on their initial urine dipstick, further urine protein testing was discontinued. When magnesium sulphate was prescribed as part of the management of pre-eclampsia, this was carefully recorded.

Hair collection and analysis

Up to 0.5000 g of hair was cut from the sub-occipital region of the head of each participant in an identical manner using sterile surgical-grade steel scissors. The hair was cut as close to the scalp as possible to ensure that the most recent growth was included in the sample. In samples in which the hair was longer than 4 cm, the cut end was identified. Each hair sample was placed into a numbered, clean polythene bag, excess air was expelled and the bag was sealed. Samples were stored in a dry, dark area at room temperature until collection was complete. All samples were couriered together to the Department of Chemistry, University of Hull, UK. The entire laboratory process was supervised by the Department of Chemistry, University of Hull using the methodology described to measure calcium and magnesium in hair.[15]

Hair samples were added with 2 ml of nitric acid (trace metal SpA grade, Romil, Cambridge, UK) to Teflon microwave vessels and slow room-temperature digestion was allowed to begin. The vessel was then sealed and heated in a microwave (CEM Microwave Technology Ltd, Milton Keynes, UK) to 200°C for 10 minutes. When cooled to room temperature, the digest was diluted by weight with pure water (UHQPS, Elga Veolia Labwater, Marlow, UK). Sample and digest weights were used to calculate the dilution factor of each sample. The digests were analysed using an Optima 5300DV inductively coupled plasma optical emission spectrometry (ICPOES) instrument (PerkinElmer Life and Analytical Sciences, Shelton, CT, USA), and all digests were analysed in a single batch to avoid interassay variation. Three interference-free emission wavelengths for calcium and magnesium (Ca 317.933 nm, Ca 315.887 nm, Ca 393.366 nm, Mg 285.213 nm, Mg 279.077 nm, Mg 280.271 nm) were included in the analysis method. The calibration was performed up to 10 parts per million (ppm) in radial view mode, based on three replicates per measurement. As good quantitative agreement was obtained between the wavelengths indicated, it was decided to proceed with the recommended wavelengths of Ca 393.366 nm and Mg 280.271 nm for the analysis of the samples. The final concentrations of calcium and magnesium in the hair samples were calculated by multiplying the measured concentration of each element in solution (ppb) by the dilution factor (ratio of the digest weight and sample weights), and altering the units to per cent by dry weight of the calcium and magnesium in the original solid samples provided.

Blood collection and analysis

Five millilitres of venous blood were taken in a sterile manner from the antecubital fossa of each subject. The area was prepared with isopropyl alcohol and, in most cases, a tourniquet was not used. In those cases in which a tourniquet was necessary, a needle was inserted into the vein, the tourniquet was removed and blood was withdrawn after waiting for a period of 5–10 seconds. This was performed to avoid any disturbance in albumin concentration that may occur when a tourniquet is applied for a period exceeding 1 minute. Blood was placed into BD Vacutainer SST II Advance tubes (BD, Franklin Lakes, NJ, USA) and processed by the National Health Laboratory Service at Groote Schuur Hospital within 12 hours of taking the specimen. Where patients had been treated with magnesium sulphate to prevent eclamptic seizures, blood was taken after a minimum of 12 hours after the treatment had been stopped to ensure complete excretion of exogenous magnesium.[16]

The serum calcium, magnesium and albumin concentrations were measured and the corrected calcium level was calculated by the laboratory. All tests were performed using a Roche Modular P analyser (Roche Holdings Limited, Basel, Switzerland). Corrected calcium was calculated in millimoles per litre (mmol/l) using the formula: corrected calcium = measured total calcium (mmol/l) + 0.02[40 – serum albumin (g/l)] (http://en.wikipedia.org/wiki/Calcium_metabolism#Corrected_calcium_level ).

Data analysis

Data were entered into Microsoft® Excel version 2007. Statistical analyses were performed using Stata® 11.1 (StataCorp LP, College Station, TX, USA). Data were analysed in three groups: the total group (which included all participants), HIV-positive women and HIV-negative women. The Shapiro–Wilk test was used to assess whether the distribution of the data was normal. Student's t and Wilcoxon rank-sum tests were used to compare means and medians between the two groups, respectively. Chi-squared tests were used to compare proportions and, where cell frequencies were less than five subjects, Fisher's exact test was employed.

Ethical consideration

This study was approved by the Research Ethics Committee of the Faculty of Health Sciences of the University of Cape Town, reference 016/2010. Permission to recruit participants was obtained from the Heads of the Obstetric Units. This study was performed in accordance with the World Medical Association Declaration of Helsinki of 2008.

Results

Recruitment

During the study period, 264 recently delivered women were asked to participate in the study. Thirty-six (14.6%) elected not to participate for a variety of reasons. Of the 228 (86.4%) women who agreed to participate, 14 (5.3%) were excluded: one did not fulfil the diagnostic criteria for pre-eclampsia and 13 had taken calcium-containing supplements during pregnancy.

In total, 214 women were recruited: 106 had pre-eclampsia (‘cases’) and 108 were controls. Of those recruited, 166 were HIV negative and 48 were HIV positive. Ninety-six of the cases were matched to controls. Ten cases and 12 controls that were recruited could not be matched and were not included in the final analysis. One hundred and ninety-two participants contributed to the final analysis, 152 of whom were HIV negative and 40 HIV positive (Figure 1). Data were analysed in three groups: the total group, a HIV-negative group and a HIV-positive group.

Figure 1.

Flow chart of participants.

Demographics

In the total group, the mean age of women with pre-eclampsia was 24 ± 4.3 years (range, 18–36 years) and of normotensive controls was 24 ± 4.4 years (range, 18–37 years; P = 0.7). There was no statistically significant difference in age between the HIV-negative and HIV-positive groups. As cases and controls were matched for gravidity and ethnicity, there was no difference between the groups in these parameters. The majority of participants were primigravid (79.2%). Eighty-five (88.5%) matched pairs were black African and 11 (11.5%) pairs were of mixed ethnicity. There were no Indian or white women recruited to the study. All 20 pairs in the HIV-positive group were black African.

Socio-economic status was assessed using four parameters: highest level of education attained, monthly household income, source of domestic water and whether participants lived in formal or informal dwellings (Table 2). Overall, there were no differences in any of these parameters in the total group. When analysed by HIV status, there were two significant differences. In the HIV-negative group, fewer women completed grade 12 or achieved tertiary qualifications in those affected by pre-eclampsia. In the HIV-positive group, fewer women had access to domestic water supply in those affected by pre-eclampsia.

Table 2. Socio-economic parameters*Thumbnail image of

There was no difference in gestational age at booking in the total group (P = 0.4) and this measure was not affected by HIV status. The gestational age at delivery of women with pre-eclampsia was significantly lower in the total (P < 0.05) and HIV-negative (P < 0.05) groups, but there was no statistically significant difference in women with HIV infection (Table 3).

Table 3. Gestational age (in weeks) at booking and delivery
 TotalHIV negativeHIV positive
Case (N = 96)Control (N = 96)Diff P Case (N = 76)Control (N = 76)Diff P Case (N = 20)Control (N = 20)Diff P
Booking mean ± SD20.88 ± 6.4921.77 ± 6.800.900.421.32 ± 6.8122.43 ± 6.061.110.319.15 ± 4.8819.25 ± 8.900.101.0
Delivery mean ± SD37.35 ± 2.3838.36 ± 2.011.010.00137.61 ± 2.3238.59 ± 2.000.990.00636.40 ± 2.4437.50 ± 1.821.100.1

HIV infection

The degree of immune system compromise in the HIV-positive group was assessed using the CD4 count. Four cases and five controls had CD4 counts of <200 × 106 l (P = 0.1). Four cases and nine controls were receiving highly active antiretroviral therapy (HAART) and 15 cases and 10 controls received antenatal monotherapy with zidovudine and nevirapine in labour. Two HIV-positive women, one case and one control, received no antiretroviral therapy, and the reason for this deviation from normal protocol remains unexplained.

Dietary parameters

To compare the nutritional status and dietary calcium intake of the three groups analysed, a number of variables were investigated (Table 4). Body mass index (BMI) was used as an indicator of nutritional status and to identify those participants who were underweight. Dietary intake of calcium-rich foods, such as fish, dairy products, green vegetables, fruit, nuts, beans and rooibos tea, and calcium-fortified foods, such as bread and cereals, were analysed. Dietary factors that might affect calcium absorption, such as red meat, coffee and Ceylon tea, as well as alcohol use and cigarette smoking, were also included in the analysis.

Table 4. Dietary parameters [where results are not given in the form n (%), the alternative unit of measurement is indicated in parentheses]Thumbnail image of

There was no difference in BMI at booking between women with pre-eclampsia and matched controls in the groups. One participant in the control arm of the HIV-positive group was underweight with a BMI of <18.5 kg/m2. There was no difference in the consumption of calcium-rich foods, including fish, milk, yoghurt, green vegetables, nuts, beans and rooibos tea, in the three groups. In the total and HIV-negative groups, normotensive women consumed more cheese (P = 0.01) and more fruit (P = 0.04). No women in the HIV-positive group with pre-eclampsia consumed cheese.

There were no statistically significant differences in the consumption of bread, cereal, Ceylon tea and red meat between women with pre-eclampsia and matched controls in the groups. In the HIV-positive group, women with pre-eclampsia consumed less coffee each day than those with normotensive pregnancies (P = 0.02).

Fourteen (7.3%) of 192 women who participated in this study were smokers and six (3.1%) consumed alcohol during their pregnancy. No statistically significant difference was seen between cases and controls in participants who smoked or consumed alcohol.

Blood and hair analysis

Blood and hair levels of calcium and magnesium are given in Tables 5 and 6. Twelve serum assays were not performed as a result of laboratory oversight and are indicated in the relevant tables. In the total group, the mean serum calcium concentration of women with pre-eclampsia was statistically significantly lower than that of matched controls, although both mean values fell within the normal range (2.05–2.56 mmol/l) accepted by the laboratory. One matched control, in the HIV-negative group, had a serum calcium concentration slightly below the lower limit of normal.

Table 5. Calcium and magnesium levels in serum and hair of the total group
 Cases (N = 96)Controls (N = 96) P
  1. a

    Magnesium data were not normally distributed and were transformed in order to perform parametric statistics using 1/([Mg]2). The untransformed means are reported.

Calcium
Serum (mmol/l), corrected (mean ± SD) (normal range, 2.05–2.56 mmol/l)2.23 ± 0.09 (n = 95)2.26 ± 0.08 (n = 93)0.04
Hair (ppm) (median, range)1241, 331–4654 (n = 96) 1146, 480–4136 (n = 96)0.5
Magnesium
Serum (mmol/l)a (mean ± SD) (normal range, 0.65–1.10 mmol/l)0.87 ± 0.21 (n = 94)0.72 ± 0.07 (n = 90) <0.01
Hair (ppm) (median, range)117, 33–896 (n = 96)106, 18–819 (n = 96)0.4
Table 6. Calcium and magnesium levels in serum and hair by HIV status
 HIV negativeHIV positive
Cases (N = 76)Controls (N = 76) P Cases (N = 20)Controls (N = 20) P
  1. a

    Magnesium data were not normally distributed and were transformed in order to perform parametric statistics using 1/([Mg]2). The untransformed means are reported.

Calcium
Serum (mmol/l) (corrected) (mean ± SD) (normal range, 2.05–2.56 mmol/l)2.24 ± 0.09 (n = 75)2.27 ± 0.08 (n = 73)0.032.22 ± 0.08 (n = 20) 2.23 ± 0.10 (n = 20) 0.7
Hair (ppm) (median, range)1162, 479–4136 (n = 76)1168, 331–4654 (n = 76) 0.11098, 498–3231 (n = 20)1443, 532–2378 (n =20)0.1
Magnesium
Serum (mmol/l)a (mean ± SD) (normal range, 0.65–1.10 mmol/l)0.87 ± 0.22 (n = 74)0.71 ± 0.07 (n = 70)<0.010.87 ± 0.18 (n = 20) 0.72 ± 0.07 (= 20) <0.05
Hair (ppm) (median, range)117, 33–896 (n = 76)104, 17–667 (n = 76)0.2122, 37–76 (n = 20)153,25–819 (n = 20)0.6

There was no significant difference in the concentration of hair calcium between women with pre-eclampsia (median, 1241 ppm; range, 331–4654 ppm) and their matched controls (median, 1146 ppm; range, 480–4136 ppm; P = 0.5).

Eighty-six of 92 women with pre-eclampsia were treated with intravenous magnesium sulphate to offer primary protection against eclamptic seizures. The mean serum magnesium concentration of women with pre-eclampsia was significantly greater than that of matched controls (< 0.01) in the total and HIV-negative groups. There was no statistically significant difference in the level of magnesium in hair between women with pre-eclampsia and matched controls in the total group.

In the HIV-negative group, differences observed in calcium and magnesium levels in serum and hair between women with pre-eclampsia and their matched controls reflected those found in the total group. In the HIV-positive group, there were no statistically significant differences in serum corrected calcium, hair calcium and hair magnesium levels. The only difference that achieved statistical significance in this group was the serum magnesium concentration (< 0.05), where women with pre-eclampsia had higher levels than their matched controls. Seventeen of 20 HIV-positive women with pre-eclampsia received magnesium sulphate therapy.

Intergroup comparisons were performed which showed that serum and hair calcium and magnesium levels in women with pre-eclampsia were not influenced by HIV infection. Similarly, in women with normotensive pregnancies, HIV infection had no influence on these levels.

No relationship between serum and hair calcium concentrations could be detected using Spearman's rank correlation coefficient: −0.0076 (P = 0.92). This also held true for serum and hair magnesium concentrations: −0.0274 (P = 0.71).

Discussion

Main findings

The literature surrounding calcium supplementation in pre-eclampsia is conflicting and the mechanism by which calcium prevents pre-eclampsia remains unclear. Pre-eclampsia has a high prevalence in developing countries, which is where the greatest effect of calcium supplementation has been seen. A possible explanation may be that body calcium status and the need for supplementation in pregnancy vary by geographical location and diet within an area.[17] Poverty in developing countries contributes to poor calcium intake, and therefore an association between developing countries, insufficient dietary calcium and a high prevalence of pre-eclampsia has been made.

Serum calcium is a poor indicator of chronic calcium status and cannot reflect calcium level over time. Hair is a stable and reliable indicator of long-term calcium status. Although there is no recognised normal range of hair calcium, the values obtained in this study have been used as a means to compare long-term calcium status between groups.

By matching women with pre-eclampsia to normotensive women with regard to age, gravidity, race and gestational age at delivery, no difference in demographic measurements was found between the groups. This was important as total body calcium levels may differ in women of varying ages and gestational age, and calcium intake may differ between ethnic groups as a result of dietary customs.[17] The mean gestational age at booking for all participants in the study was 21 ± 6.64 weeks with few women having early pregnancy scans. Late booking and limited access to ultrasound scanning are common problems in our antenatal service. Differences in gestational age at delivery reached statistical significance in the total and HIV-negative groups, where woman suffering from pre-eclampsia delivered approximately 1 week earlier than matched controls. This result was anticipated, as women who are diagnosed with pre-eclampsia at Mowbray Maternity Hospital or Groote Schuur Hospital have their delivery expedited after 34 weeks of gestation.

The similarities in socio-economic status between the two groups make it possible to assume that women with pre-eclampsia and matched controls had equal opportunity to access similar diets and thus calcium intake.

The dietary questionnaire, previously used in a similar study, provided a broad dietary history of each participant.[10] Although the questionnaire does not focus on exact quantities of food, it indicates definite deficiencies if they exist. During statistical analysis, it was found that the median value of the quantity of certain foodstuffs that were consumed was often zero, and data were converted from a continuous to a dichotomous variable (consumers or non-consumers) to facilitate analysis.

The dietary parameters measured in our study reviewed three groups of food: those naturally rich in calcium, foodstuffs that have been fortified with calcium and foods that impair calcium absorption in the gut. Generally, there was little difference in diet composition between women in the groups. In the HIV-positive group, some dietary differences reached statistical significance; however, this group contained few participants and these results were viewed with some reserve.

Insufficient caloric and nutrient intake associated with malnutrition in HIV infection may alter dietary calcium and long-term calcium status. In our study, calcium and magnesium levels in serum and hair did not differ between women with pre-eclampsia or matched controls when compared by HIV status. The similar values in calcium and magnesium levels in hair between the HIV-negative and HIV-positive groups might be because any real difference was obscured by the small numbers in the HIV-positive group or because immune suppression is not necessarily accompanied by weight loss.[13]

In this study, serum calcium concentrations of non-supplemented women at delivery were significantly lower in those with pre-eclampsia than in matched controls in the total and HIV-negative groups. Although this difference was statistically significant, its clinical relevance is difficult to assess, given that all participants, with one exception, had serum calcium levels within the normal range.

Hair calcium levels showed no difference between women with pre-eclampsia and matched controls in any of the three groups, supporting the null hypothesis that there is no difference in the long-term calcium status between women with pre-eclamptic and normotensive pregnancies. Our analysis confirms that neither socio-economic status nor dietary intake had any influence on the outcome of the hair results.

Serum magnesium was higher in women with pre-eclampsia than in matched controls in all three groups. This finding was unexpected and was not attributable to impaired renal function. All women with pre-eclampsia in this study had normal or marginally raised creatinine levels (80–100 μmol/l) at recruitment. Women with pre-eclampsia who received magnesium sulphate therapy had significantly higher mean serum magnesium levels (0.88 ± 0.22 mmol/l) than women with pre-eclampsia who did not receive magnesium sulphate (0.74 ± 0.08 mmol/l). There was no statistically significant difference in serum magnesium level between those women with pre-eclampsia who did not receive magnesium sulphate therapy (0.74 ± 0.08 mmol/l) and the total matched control group (0.72 ± 0.07 mmol/l). Blood samples of participants who had received magnesium sulphate were drawn after a period of not less (and often longer) than 12 hours following completion of magnesium sulphate therapy to ensure complete excretion of exogenous magnesium. There was no significant difference in hair magnesium levels between women with pre-eclampsia and matched controls in the three groups. Residual exogenous magnesium probably contributed to the higher serum levels in women with pre-eclampsia. In future studies, the waiting period before sampling blood will need to be extended.

Strengths

To date, this is the first study to investigate the long-term calcium status of women with pre-eclampsia through hair analysis. Using hair as an indicator of chronic calcium status in pre-eclamptic women is an accurate and reproducible method which future studies may use.

Limitations

Women with early-onset pre-eclampsia were not recruited to this study as it would not be possible to match them to suitable controls, as women who deliver prior to 34 weeks of gestation cannot be considered to have had an uncomplicated pregnancy. Gestational age could be better determined with improved resources and access to antenatal care. Approximately 10% of recruited women were not included in the final analysis as they were unable to be matched, although statistical calculations showed that the findings of this study would not have been altered significantly had they been included. We acknowledge that the HIV group is too small to perform definitive analysis, and that the results presented in this study may not be applicable to ethnic groups other than black African women.

Interpretation

As this is the only study to investigate the relationship between hair calcium and pre-eclampsia, there is no literature against which we can compare our findings. It is reassuring to note that the hair calcium values in our study (range, 331–4654 ppm) were similar to those of studies performed in America (range, 188–4900 ppm),[18] China (mean, 946 ppm)[8] and Sweden (median, 1290 ppm; range, 231–5360 ppm).[19] A study to compare the long-term calcium status of pregnant women in Cape Town and in developed countries will be necessary to investigate and confirm these results.

Conclusions

Chronic calcium status was found to be no different in women with pre-eclamptic and normotensive pregnancies. This finding refutes the current belief that the mechanism by which calcium supplementation reduces the risk of developing pre-eclampsia is by correcting a pre-existing nutritional deficiency. Other mechanisms need to be investigated, such as the pharmacological effects of calcium.

Acknowledgements

Sisters Anne Hoffman and Shane Moore of the Reproductive Medicine Unit at Groote Schuur Hospital are acknowledged for their involvement in recruitment and data collection.

Disclosure of interests

None.

Contributions to authorship

DGDR performed research design and development of the protocol, data collection and preparation of the article. SWL carried out research design, funding and review of the article. HC performed statistical analysis. RK carried out scientific design of the hair analysis methodology and processing of all hair specimens. SJH performed scientific design of the hair analysis methodology and supervision of laboratory processes. ZMvdS carried out research design and development of the protocol, funding, supervision of the study, analysis of the data and preparation of the article.

Details of ethics approval

This study was approved by the Research Committee of the Department of Obstetrics and Gynaecology and then submitted to the Research Ethics Committee of the Faculty of Health Sciences of the University of Cape Town, reference 016/2010. Permission to recruit participants was obtained from the heads of the Obstetric Units. This study was performed in accordance with the World Medical Association Declaration of Helsinki of 2008.

Funding

Funding was provided from the research funds of Z. M. van der Spuy, University of Cape Town, South Africa and S. W. Lindow, University of Hull, UK.

Ancillary