Objective
We postulated that calcium supplementation of calcium-deficient pregnant women would lower vascular resistance in uteroplacental and fetoplacental circulations.
Study Design
Pulsatility index (PI) and resistance index (RI) (uterine and umbilical arteries) and presence of bilateral uterine artery diastolic notching were assessed by Doppler ultrasound between 20-36 weeks’ gestation in 510 healthy, nulliparous Argentinean women with deficient calcium intake in a randomized, placebo-controlled, double-blinded trial.
Results
Average umbilical and uterine artery RI and PI tended to be lower in the supplemented group at each study week. Differences became statistically significant for umbilical artery RI and PI from 32 and 36 weeks, respectively. Estimated probabilities of bilateral uterine artery diastolic notching trended toward lower values in calcium-supplemented women.
Conclusion
Calcium supplementation of pregnant women with deficient calcium intake may affect uteroplacental and fetoplacental blood flow by preserving the vasodilation of normal gestation.
Hypertensive disorders complicate approximately 9% of all pregnancies and are responsible for substantial morbidity and mortality in both mother and fetus with the majority of these bad outcomes due to pure or superimposed preeclampsia/eclampsia. Despite recent and existing progress related to causality, the etiology and pathophysiology of preeclampsia remain unclear, and we are yet to develop targeted prevention strategies. Thus, faced with overt disease, delivery remains the only definitive treatment.
There are reports that intracellular free calcium concentration is increased in women with preeclampsia. The observations have been reported in platelets and lymphocytes and interpolation of these findings to smooth muscle cells have led to the hypothesis that such changes underlie the widespread maternal vasoconstriction and hypertension. It has further been suggested that increased intracellular free calcium levels are a consequence of the increase in parathyroid hormone levels secondary to the low calcium intake, a problem correctable by calcium supplementation and proposed as a preventive strategy for preeclampsia, specifically among low-calcium-intake women.
A plausible approach toward understanding the mechanisms linking dietary calcium intake to vasoconstriction during normal and abnormal pregnancy is to prospectively assess the effects of different levels of calcium intake on the uterine and umbilical circulations. Doppler velocimetry allows for noninvasive estimation of blood flow characteristics in these uteroplacental vessels. These are ideal vascular systems to study, as abnormal placentation has been implicated in the etiology of preeclampsia and poor placental perfusion may be responsible for the aberrant placental production of circulating maternal proteins that play a role in the preeclamptic process.
In 2006, the World Health Organization (WHO) published the results of a large multicenter, randomized, double-blinded and placebo-controlled trial indicating that daily calcium supplementation of 1500 mg in pregnant populations whose mean daily calcium intake was <600 mg reduced the severity of the maternal preeclamptic syndrome and decreased fetal loss. An ancillary study was performed at the Argentine center in Rosario (Centro Rosarino de Estudios Perinatales [CREP]), with investigators serially assessing the uterine and placental circulations and hypothesizing that calcium supplementation would reduce vascular resistance. Below we report the results of the ancillary study.
Materials and Methods
The WHO calcium supplementation trial in low-calcium-intake women for the prevention of preeclampsia randomized women to receive either elemental calcium supplements of 1500 mg/day, as calcium carbonate, or placebo starting at enrollment until delivery, initiation of magnesium sulfate therapy, or clinical suspicion of urolithiasis. Appearance of preeclampsia or hypertension was not reason to discontinue treatment. Chewable tablets (Nikomed Pharma Ltd, Elverum, Norway), containing 500 mg of calcium or placebo, similar in size and taste, were ingested 3 times daily with meals.
Randomization for all sites was performed in Geneva, Switzerland, at the WHO Statistics and Informatics Unit of the Special Programme of Research, Development and Research Training in Human Reproduction and provided to the labeling and packaging company, Magistra SA Laboratories (Geneva, Switzerland). Randomization sequences were produced using computer-generated random number blocking (SAS Institute Inc, Cary, NC). Randomization codes remained at WHO until the time of analysis and were not available to any person until the analyses were completed.
Boxes containing tablet bottles (7 bottles/box, 100 tablets/bottle), prepared by Magistra SA according to the randomization sequence, were distributed by WHO to each of the centers and kept in a designated area within a locker. Boxes were provided consecutively to the randomized subjects who returned their bottles at each 4-weekly antenatal care visit, regardless of whether she had taken all the tablets or not, and a new bottle was issued. Compliance was evaluated by calculating the number of tablets returned to the total number of tablets provided.
Subjects
In all, 1565 women participating in the main WHO multicenter calcium supplementation trial were enrolled at CREP. Healthy, nulliparous pregnant women able to give consent were recruited <20 weeks’ gestation. Gestational age at trial entry was determined with the “best obstetric estimate,” which included ultrasound examination if required by the attending obstetrician.
Eligibility also demanded that their average calcium intake be <600 mg/day (recommended intake of 1000 mg/day). Before initiation of the calcium trial, a calcium intake assessment was conducted on all subjects attending antenatal care clinics at CREP. Assessment was conducted by nutritionists through a 24-hour dietary recall survey. After fulfilling selection criteria, each woman was then randomly assigned to receive calcium-containing supplement tablets or placebo from enrollment until day of delivery. Methods further detailing the selection and exclusion criteria for subjects in the trial have been published elsewhere.
All women enrolling in the main trial between November 2002 and July 2003 were approached to participate in the ultrasound study. In all, 511 of the total 1565 women enrolled to participate in the calcium supplementation trial were also separately consented to undergo ultrasound examination to assess uteroplacental blood flow and fetal growth at 20, 24, 28, 32, and 36 weeks’ gestation.
The institutional review board of the CREP approved this addition to the main trial’s protocol.
Outcomes
Doppler velocimetry was performed using standard techniques. The specific measurements and observations sought were the pulsatility index (PI) and resistance index (RI) of the uterine and umbilical arteries and the presence of uterine artery bilateral diastolic notching. PI and RI are measures of blood flow that relate to the systolic and diastolic components of the flow velocity waveform describing the characteristics of the blood flow in a specific vessel during 1 cardiac cycle. These indices represent the downstream flow impedance, which is an indicator of vascular resistance. The higher the values of PI and RI, the greater the resistance to flow. The presence of an early diastolic notch in the uterine artery recording is a qualitative assessment of the blood flow and in normal pregnancy disappears during early second trimester. The persistence of the early diastolic notch after that period is indicative of increased uterine vascular resistance. Bilateral diastolic notching specifically refers to presence of this finding in both main uterine arteries on Doppler examination.
Data acquisition
Before the initiation of this study, the participating ultrasonographers were trained and certified through a course conducted by 2 of the coauthors (R.D., R.G.) to standardize the testing and assure high-quality data collection. Interobserver and intraobserver variability were not formally measured. Visualization of at least 3 similar consecutive waveforms was required before recording a measurement and all images were stored digitally. The ultrasound equipment used to perform the scans and Doppler velocimetry was the Sonosite 180 Plus (Sonosite Inc, Bothell, WA).
Measurements from the main branch of the uterine artery were taken at the crossover point with the iliac vessels. The average value from combined left and right uterine artery measurements was used in the analysis. Umbilical artery measurements were taken from a free-floating central part of the umbilical cord during a state of fetal quiescence.
Statistical analysis
To detect a biologically meaningful reduction in uterine artery RI between supplement and placebo groups (30% difference with 80% power and .05 significance level), a power analysis determined a minimum sample size of 109 women in each arm. Women were included in the analysis if they contributed at least 1 ultrasound examination. Ultrasounds performed 1 week before or after the scheduled examination were included in both cross-sectional and longitudinal analyses; however, those separated by greater periods were excluded from cross-sectional analysis.
Cross-sectional analyses to assess the effects of calcium supplementation at each gestational age were conducted by comparison of Doppler velocimetry indices for the 2 groups. Measures considered for the analyses were PI and RI in uterine and umbilical circulations. For each measurement and gestational age, the mean difference between groups and 95% confidence intervals (CI) were calculated.
Longitudinal analyses were performed to evaluate the treatment effects of calcium supplementation on the rate of change of PI and RI. Model-based estimates of the indices for both treatment and placebo groups were derived using the parameter estimates. The general linear model for repeated measurements was used to account for correlation in measurements of the same subject. A quadratic effect of gestational age on the measurement provided a better fit than a linear effect for all the indices except for RI of the umbilical artery where a linear effect was used. Estimated differences and their 95% CI were based on the appropriate model-based SE.
Differences in rate of change due to calcium supplementation were assessed by testing the interaction between treatment group and the gestational age effect. Compound symmetry matrices were developed for all measurements after evaluating structure selection using the Schwarz Bayesian information criteria. Regression coefficients were estimated using maximum likelihood as per SAS system (version 9; SAS Institute, Cary, NC). The model was used to derive adjusted measurements at weeks 20, 24, 28, 32, and 36 for each treatment group and their differences.
In evaluating for the presence of bilateral diastolic notches in the uterine arteries, a generalized linear model with a logit link was adjusted with correlation structures following a compound symmetry pattern. Generalized estimation equations were used to obtain the parameter estimates.
Results
The WHO calcium supplementation trial enrolled a total of 8325 women, of whom 1565 mothers were recruited at the CREP center in Argentina. Of the latter, 511 subjects were enrolled during the specified time window (November 2002 through July 2003) and were eligible for participation in the ancillary ultrasound trial. There were no cases of women consenting to participate in the main trial who refused to participate in the substudy. After exclusion of 1 multiple pregnancy, 510 women were deemed eligible for ongoing ultrasound examinations ( Figure 1 ).
Of the 510 mothers eligible for the trial, 31 women (18 from the placebo group; 13 from the calcium group) were withdrawn due to early termination of pregnancy (n = 26) or having been discovered not to be pregnant (n = 5). Thirteen women (7 in the placebo group; 6 in the calcium group) were lost to follow-up due to moving from the area, their withdrawal, or inability to keep appointments. Five more were excluded when congenital malformations were identified during the ultrasound examination. Therefore, 461 women with at least 1 ultrasound examination (230 in the placebo and 231 in the calcium group) contributed to the final analyses ( Figure 1 ).
Maternal baseline characteristics at entry
There were no differences in age, education, weight, height, body mass index, and gravidity. Gestation, presence of proteinuria, and systolic and diastolic blood pressure on enrollment was also similar between groups ( Table 1 ). Calcium intake assessed by dietary recall survey demonstrated a median calcium intake of 481 mg/day (25th centile: 246 mg/day; 75th centile: 767 mg/day) with 82% of surveyed women (95% CI, 73-91%) found to have inadequate dietary intake. Compliance with treatment was assessed after the withdrawal of 31 women who had experienced abortion or who were not pregnant ( Table 2 ). Compliance, measured by the ratio of tablets taken to tablets expected to be taken, was 72% for the calcium group and 76% for the placebo group.
Maternal characteristic | Calcium (n = 251) | Placebo (n = 259) | ||||
---|---|---|---|---|---|---|
n | Mean | SD | n | Mean | SD | |
Age, y | 251 | 20.8 | 4.7 | 259 | 20.2 | 4.5 |
Education, y | 250 | 10.1 | 2.9 | 259 | 9.7 | 2.7 |
Weight, kg | 249 | 57.9 | 12.1 | 259 | 58.3 | 11.9 |
Height, cm | 245 | 158.8 | 6.2 | 256 | 158.9 | 6.0 |
Body mass index, kg/m 2 | 245 | 22.9 | 4.4 | 256 | 23.1 | 4.1 |
Smoking during pregnancy | 41/251 (16.3%) | 55/258 (21.3%) | ||||
Gestation on enrollment, wk | 251 | 12.6 | 3.8 | 259 | 12.9 | 3.7 |
Gravidity ≥1 | 23/251 (9.2%) | 20/259 (7.7%) | ||||
SBP on enrollment, mm Hg | 251 | 104 | 11.4 | 259 | 103 | 11.4 |
DBP on enrollment, mm Hg | 251 | 59 | 8.8 | 259 | 58 | 9.4 |
Proteinuria ≥++ on enrollment | 6/245 (2.4%) | 14/246 (5.7%) |
Compliance with treatment | |||||
---|---|---|---|---|---|
Variable | Calcium (n = 238) | Placebo (n = 241) | P value | ||
Mean | SD | Mean | SD | ||
Tablets taken | 389.6 | 143.8 | 409.7 | 143.4 | .137 |
Tablets expected | 545.4 | 90.8 | 537.3 | 91.5 | .332 |
Taken/expected | 0.719 | 0.245 | 0.764 | 0.239 | .046 |
Average daily intake | 2.16 | 0.745 | 2.29 | 0.717 | .046 |
Compliance with ultrasound scans | |||||||
---|---|---|---|---|---|---|---|
Gestational age, wk | Calcium | Placebo | P value | ||||
Performed | Expected | % | Performed | Expected | % | ||
20 | 195 | 238 | 81.9 | 208 | 241 | 86.3 | .212 |
24 | 197 | 238 | 82.8 | 206 | 241 | 85.5 | .454 |
28 | 193 | 238 | 81.1 | 207 | 240 | 86.3 | .138 |
32 | 187 | 235 | 79.6 | 194 | 236 | 82.2 | .484 |
36 | 185 | 222 | 83.3 | 178 | 214 | 83.2 | 1.000 |
Total | 957 | 1171 | 81.7 | 993 | 1172 | 84.7 | .059 |