*Dr J. M. Beliz´n, Latin American Centre for Perinatology, Pan American Health Organization, World Health Organization, Hospital de Clínicas s/n, 11000 Montevideo, Uruguay.
Objective To assess in an animal model the effect of maternal dietary calcium intake during pregnancy on the blood pressure of offspring.
Design Randomised controlled trial.
Sample Ninety-eight 20-week-old female Wistar–Kyoto rats, giving birth to a total of 119 pups that were included in the follow up study.
Methods Rats were randomised to a calcium deficient diet, a diet with the recommended calcium levels, or a diet with calcium content much higher than the recommended levels. After one month on the experimental diet they were bred. After birth, systolic blood pressure in the offspring was measured monthly till 52 weeks of age.
Main outcome measures Blood pressure of the offspring.
Results The difference in blood pressure of the offspring between the normal and low maternal calcium group increased 0.49 mmHg per month (95% CI 0.18 to 0.84), 0.38 (0.07 to 0.68) between the low and high calcium group, and 0.12 (−0.20 to 0.43) between the normal and high calcium group. At the end of the follow up (52 weeks of age) blood pressure of the offspring in the deficient calcium group was 12.1 mmHg (95% CI 8.8 to 15.4, P < 0.0001) higher than in the normal calcium group and 7.5 mmHg (95% CI 4.4 to 10.5, P < 0.001) higher than the high calcium group. Blood pressure of the offspring in the high calcium group was 4.3 mmHg (95% CI 1.0 to 7.5, P= 0.0l) higher than in the normal calcium group. In a multiple regression model maternal calcium intake during pregnancy was the strongest predictor of blood pressure of the offspring during adulthood.
Conclusions This experiment supports previous studies in humans suggesting a link between calcium intake during pregnancy and blood pressure in the offspring, and provides an animal model to explore the mechanisms involved in such association.
Our interest in the relationship between calcium intake and blood pressure originated in observation of indigenous women in Guatemala. Those women have a high calcium intake due to the Mayan tradition of treating corn with lime and have a very low incidence of hypertensive disorders of pregnancy1. After this epidemiological observation we performed a series of studies in which we showed an association between dietary calcium intake and blood pressure in rats2, young individuals3 and pregnant women4. These observations lead us to state the hypothesis that the incidence of one of the most severe forms of hypertension during pregnancy, pre-eclampsia, can be reduced in populations of low calcium intake by calcium supplementation5. In agreement with this hypothesis, a recent meta-analyses including 10 randomised controlled trials has shown that in communities with low dietary calcium intake calcium supplementation during pregnancy is associated with a 68% (from 51% to 79%) reduction in the incidence of pre-eclampsia6.
In 1987 our group performed a large, randomised, placebo controlled trial of calcium supplementation during pregnancy7 in a population with low calcium intake, and seven years later we conducted the follow up of children born from these mothers8. This study showed that the offspring of calcium supplemented mothers had a lower incidence of high blood pressure than offspring of mothers receiving placebo, suggesting a role of fetal calcium restriction on the genesis of hypertension in later life. This observation is supported by a large number of epidemiological studies linking an adverse fetal environment with an increased risk of adult hypertension9–11, and by recent animal studies showing that offspring life-long hypertension can be also determined during fetal life by maternal protein restriction12.
In the present study an animal model is used to assess the effect of maternal dietary calcium intake during pregnancy on blood pressure of the offspring.
Virgin Wistar–Kyoto rats (Taconic, Germantown, New York, USA) were housed individually in wire mesh cages, and maintained at 24°C on natural light cycles. All rats had free access to water. Animals were housed in the centre's animal facility in strict compliance with institutional regulations.
At 20 weeks of age, the rats were randomly assigned to one of three experimental synthetic diets: a calcium deficient diet, a diet with the recommended calcium levels for rat chow, or a diet with a calcium content much higher than the recommended levels. Synthetic diets (Dyets Inc, Bethlehem, Pennsylvania, USA) were prepared by supplementing a calcium deficient rat chow with 0 ppm added calcium (low calcium diet), 5000 ppm added calcium (normal calcium diet) and 10,000 ppm added calcium (high calcium diet).
A sample of each synthetic diet was sent to an independent laboratory (Biochemistry Faculty, National University of Rosario, Rosario, Argentina) to determine calcium and protein content. The measurement in the samples showed a value of 10.5mg% calcium and 16.8g% protein in the low calcium diet, 473mg% calcium and 16.1g% protein in the normal calcium diet, and 785mg% calcium and 17.7g% protein in the high calcium diet. Synthetic diets were identical in colour, odour and taste.
To conceal the randomisation, a unique number was assigned to each dam and imprinted in the tail. An external statistician used this number to allocate the dams to three groups using a random-numbers table from a statistical textbook. Rats were then housed in groups of four dams per cage, according to the allocated group, and cages were labelled as group I, II, and III. An external statistician labelled rat chow containers as group I, II and III, and kept treatment codes concealed till the end of the study. All the investigators involved in implementing the study remained blinded with respect to the calcium content of the experimental diets, including the statistician that analysed the results. A period of 30 days was allowed to habituate the female rats to the experimental diets before mating. The experimental feeding regimen was maintained throughout the mating period and pregnancy.
Systolic blood pressure and weight was measured before randomisation and at weekly intervals till birth. After birth, we recorded the birthweight for all surviving pups and their mothers were transferred to a standard laboratory rat chow. The pups were removed from their mothers when they were four weeks old. In litters with two or more pups, two pups were selected at random from each litter, and when possible, one male and one female pup were chosen. The excess was discarded. Selected pups were identified by unique numbers and housed in groups of four regardless of maternal diet. After weaning, the pups were fed on a diet with the recommended calcium levels for standard rat chow, and maintained under standard laboratory conditions. The same person recorded the systolic blood pressure and weight of the offspring monthly until they were 52 weeks old.
Blood pressure was measured in conscious rats using an automatic rat tail blood pressure monitor (Kent Scientific, Litchfield, USA). The rats were placed in a darkened restraint tube, maintained at 28°C. To reduce the effects of stress on blood pressure, the animals were conditioned to handling and measurements were taken routinely 10 minutes after initial handling. An occlusion cuff was placed over the tail and inflated to 300 mmHg. As the rats were different sizes, care was taken to use cuffs of appropriate size for each animal. A separate piezo-electric pulse sensor was also attached to the tail. The blood pressure monitor was linked to a personal computer that recorded deflation rate and pulses and computed systolic blood pressure. The computer produced a graph of cuff pressure and pulses that was used to assess whether the recordings were valid. At least five valid measurements were recorded for each rat at each time and the average was used for the analysis.
The original hypothesis was that the minimal difference worth detecting between a low or high calcium diets, compared with a normal calcium diet, was 5 mmHg. The sample size required to detect such difference, with a type I error of 5% and type II error of 10%, was 30 mothers per group. To evaluate the statistical significance of different blood pressure patterns of the offspring over time, a multilevel modelling approach was used13,14. Multilevel modelling is an extension of ordinary multiple regression where data have a hierarchical or clustered structure. A hierarchy consists of units grouped at different levels. Repeated measures are one example of hierarchical structured data. Here, monthly blood pressure measurements are clustered within rats that represent the level 2 unit, with the monthly measurements being the level 1 units. Thus, because rats were measured on more than one occasion, two levels of variability accounted for a single rat's departure from the fitted curve (level 1) and the differences between blood pressure curves of different rats (level 2). The model fitted accounted for complex level 2 variation that allows each rat to have their own intercept and slope14. Because two rats per litter were selected, and blood pressure variability within offspring belonging to the same litter might be smaller than variability in offspring from different litters, litter was included in the multilevel model as level 3 in hierarchy. The fitted model included maternal calcium diet as a fixed effect, the interaction between maternal calcium diet and rat's age, and a cuadratic and cubic term for the rat's age. Multilevel residuals were calculated for model checking and diagnosis14.
Multiple linear regression analysis was used to assess the statistical significance of the differences in systolic blood pressure between treatment groups at the end of follow up (52 weeks), and to compare the magnitude of the effect of treatment with other predictors of blood pressure. The models included maternal calcium diet during pregnancy, offspring's weight at 52 weeks of age, sex, birthweight, and maternal blood pressure during pregnancy and maternal blood pressure at 28 weeks of age. Regression residuals were used to assess the validity of the regression models. Regression coefficients were standardised to allow the comparison of independent variables with different scale of measurement, and the coefficient represents the change in rat's systolic pressure (in mmHg) for one standard deviation shift in the value of the independent variable. Statistical analyses were performed using MLWIN (Multilevel Models Project, Institute of Education, London, UK) and SAS (Cary, North Carolina, United States) packages for IBM-PC.
Ninety-eight female rats were included in the experiment. Pregnancy was not achieved in 18 rats (six in each group), and in another 18 rats (six in the high calcium group, five in the low calcium group and seven in the normal calcium group) all pups died soon after birth due to cannibalism and neglect (Table 1). The number of pups per litter and the average birthweight was similar among experimental groups (Table 1). In five litters only one pup was available for follow up. A total of 119 pups were finally included in the study: 40 in the high calcium group, 42 in the low calcium group and 37 in the normal calcium group. Offspring groups were similar with respect to proportion of male pups and birthweight (Table 1).
Table 1. Number and characteristics of rats included in the study.
No. of randomised mothers
Pregnancy not achieved
Pregnancy achieved but all pups dead
No. of mothers with at least one live pup
Total no. of pups
Average no. of pups per litter
Mean [SD] birthweight (g)
Pups included in the follow up study
No. of pups
No. of male pups
Mean [SD] birthweight (g)
Figure 1 shows offspring systolic blood pressure patterns over time in the three groups of maternal dietary calcium during pregnancy. At one month of age systolic blood pressure was similar between offspring of rats with different dietary calcium intake during pregnancy, but differences between groups became evident with increasing age.
A multilevel model was used to analyse the patterns of blood pressure of the offspring over time, according to maternal calcium during pregnancy (Fig. 1). These analyses showed a significant interaction between maternal calcium diet during pregnancy and offspring age (P < 0.001). The interaction coefficients were used to evaluate if the difference between groups changes with increasing age. The difference in blood pressure of the offspring between the normal and low maternal calcium group increased 0.49 mmHg per month (95% CI 0.18 to 0.84). The difference between the high and low calcium group also significantly increases over time (difference between groups increased 0.38 mmHg per month, 95% CI 0.07 to 0.68). The difference between the normal and high maternal calcium group did not significantly change over time (difference between groups increased 0.12 mmHg per month, 95% CI −0.20 to 0.43).
At the end of follow up (i.e. 52 weeks of age) blood pressure of the offspring in the low calcium groups was significantly higher than in the normal calcium group. On average, the blood pressure of offspring in the low calcium group was 12.1 mmHg (95% CI 8.8 to 15.4, P < 0.0001) higher than in the normal calcium group (Fig. 1). Blood pressure in the high calcium group was 4.3 mmHg (95% CI 1.0 to 7.5, P= 0.01), higher than in the normal calcium group (Fig. 1).
In order to compare the contribution of maternal dietary calcium intake to the development of high blood pressure in offspring with other predictors of high blood pressure, a multiple regression model was fitted with blood pressure of offspring at 52 weeks of age as the dependent variable and maternal calcium diet during pregnancy, maternal blood pressure during pregnancy, maternal blood pressure at 28 weeks of age, offspring weight at 52 weeks of age, offspring sex and offspring birthweight as explanatory variables. In this model maternal dietary calcium during pregnancy was the strongest predictor of blood pressure of offspring at 52 weeks of age, followed by rat's weight (Fig. 2). Birthweight showed an inverse association with blood pressure that was borderline significant (Fig. 2). No association was found between blood pressure of the offspring and maternal blood pressure (Fig. 2). An analysis fitting a multilevel model, with mother as level 2 and offspring as level 1, produced similar results.
The present study shows in an animal model that maternal dietary calcium intake during pregnancy has a modelling effect on the offspring's blood pressure. Calcium deficit during pregnancy involves offspring with higher blood pressure values, an effect that amplifies in adult life. Furthermore, maternal dietary calcium during pregnancy was the main predictor of blood pressure in the adult rat.
These findings are in agreement with findings of our previous study on children and provide support to our statement that the effect of maternal calcium intake during pregnancy on blood pressure of the offspring could be amplified throughout later life and could contribute to hypertension on adulthood8. On the other hand a dietary calcium intake much higher than the recommended levels showed no further benefits.
There are differences between this experimental model and human populations. In human populations with a calcium deficit, there is usually a chronic restriction in calcium consumption, from birth to adult life, and for both mother and offspring. In contrast, in this animal model the restriction was only for a short period of time before and during pregnancy, and offspring received an adequate amount of calcium in the diet. On the other hand, calcium restriction during pregnancy was more severe in our model than in human populations, because the low calcium diet was almost calcium-free. Further research is needed to address the impact of chronic calcium restriction on mother's and their offspring, in a model with conditions more similar to human populations. This model should also explore the effect of increasing dietary calcium beyond the recommended levels.
To our knowledge, there is only one randomised study in rats that assessed the effect of maternal calcium diet on blood pressure of the offspring15. This study is in agreement with our results, showing no effect of maternal calcium restriction during pregnancy on blood pressure of the offspring at one month of age15.
The association between birthweight and elevated blood pressure in later life found in our study is in agreement with a weight of evidence in humans and in animals16. Rat models with low protein feeding during pregnancy have been used extensively to study mechanisms that may be involved in such an association17. Such studies have provided a considerable body of evidence implicating glucocorticoids in the programming of blood pressure by the maternal diet18. Maternal under-nutrition reduces activity of 11 beta-hydroxysteriod dehydrogenase in the placenta and may hence lead to overexposure of the fetus to glucocorticoids19. Offspring of rats which receive low amounts of protein during pregnancy remain hypersensitive to glucocorticoids into adult life, and have increased glucocorticoids receptor numbers at several sites, including the vasculature20. Intrauterine steroid exposure thus may establish an increased sensitivity to angiotensin II in early postnatal life, which in turn establishes lifelong raised blood pressure17. In a recent study it has been shown that losartan, a specific angiostensin II receptor antagonist, prevents raised blood pressure in offspring of rats that received low amounts of protein during pregnancy, while nifedipine, a calcium-channel blocker with negligible long term effects upon the renin–angiostensin system, had no effect on the blood pressure elevation in these animals21. The data are consistent with the hypothesis that angiostensin II plays a major role in the prenatal programming of hypertension by low levels of protein, and that a different mechanism might be involved in low calcium models. It has been shown that treatment of hypertension simultaneously with calcium antagonists and a high calcium diet is synergistic, with a mechanism of action that might be mediated by calciotropic hormones regulating calcium-channel activity in vascular smooth muscle cells22. In line with these arguments, it has been suggested that all forms of hypertension are associated with and dependent on cytosolic-free calcium excess that is either extra cellular or intracellular in origin23. In intracellular calcium-dependent hypertension (identified clinically with low renin forms of hypertension) the operative mechanism seems to involve excess net cellular calcium accumulation from the extracellular space, mediated by the action of calcium regulating hormones such as 1,25(OH)2D, and parathyroid hypertensive factor (PHF)23,24. The mechanism of action of PHF involves an increase in calcium-channel activity in vascular smooth muscle cells. PHF level explains why a high calcium diet may be effective in lowering blood pressure in patients who respond to calcium-channels blockers: dietary calcium might inhibit the production of PHF (and parathyroid hormone), whereas calcium-channels blockers would inhibit PHF at its target site24,25. In support for this hypothesis, studies in SHR rats have shown that the effects of dietary calcium on blood pressure may be mediated by PHF, such that a high calcium diet inhibits, and a low calcium diet stimulates, the expression of this factor25. In summary, an animal model of protein restriction during pregnancy have been used to explore the mechanisms linking a deficient fetal environment and hypertension in adult life and evidence suggest the effect of maternal calcium restriction on blood pressure of the offspring operates through a different mechanism of action. In our model calcium restriction during pregnancy might generate alterations in cellular ion transport systems inducing a metabolic set-point of calcium regulating hormones that can result in a predisposition to high blood pressure. These are intriguing hypotheses and the detailed mechanism of the involvement of calcium-regulating hormones in the genesis of hypertension and in the programming of blood pressure of the offspring await further studies. The present study provides an animal model for such studies.
The confirmation of these findings could have strong public health implications. The deleterious effects of hypertension on survival, disabilities, quality of life, and costs of health care are widely known. On the other hand calcium intake in the world is well below the requirements during pregnancy.
It is estimated that the mean calcium intake in the world is 472 mg per day, with an average intake of 860 mg per day in the developed world and 346 mg per day in developing countries26. These figures contrast with the recommended calcium intake during pregnancy of 1200 mg per day27. In view of the present findings, the achievement of such requirements could imply relevant effects on the survival and quality of life of future generations.
The authors would like to thank Dr R. Perez for her excellent contributions to the experiment; Dr J. Jost for his valuable methodological suggestions; and N. Dorf and A. Decker for their laboratory support. They would also like to thank Dr H. Piriz for providing facilities to conduct the experiment. The study was partially supported by a grant from the Argentinean Research Council (CONICET).