Cardiovascular Disease
We're Out of Milk: Dietary Calcium and CVD
Abstract & Commentary
By Howell Sasser, PhD, Adjunct Lecturer, Department of Epidemiology & Community Health, School of Health Sciences & Practice, New York Medical College, Valhalla, NY. Dr. Sasser reports no financial relationships relevant to this field of study.
Synopsis: A large observational study conducted in Germany has found little evidence that higher levels of dietary calcium are associated with a reduced risk of cardiovascular disease events. The additional finding of an elevation in risk with the use of calcium supplements suggests that caution is warranted when recommending them.
Source: Li K, et al. Associations of dietary calcium intake and calcium supplementation with myocardial infarction and stroke risk and overall cardiovascular mortality in the Heidelberg cohort of the European Prospective Investigation into Cancer and Nutrition study (EPIC-Heidelberg). Heart 2012;98:920-925.
Between 1994 and 1998, a German research group assembled a cohort of 23,980 people between the ages of 35 and 64 who were free of existing cardiovascular disease and very low- or high-calorie dietary patterns. Study participants completed a food frequency questionnaire (FFQ) to assess dietary sources of calcium, and an interview and follow-up questionnaires to measure use of calcium supplements. Total calcium intake was divided into quartiles 0-603, 604-748, 749-924, and > 924 mg/day for men, and 0-610, 611-738, 739-898, and > 898 mg/day for women. Supplement use was categorized as no use, calcium only, calcium plus other supplements, and other supplements only. Cardiovascular events (myocardial infarction [MI], stroke, and cardiovascular disease-related death) were recorded over an average follow-up time of 11 years. Relative hazards (relative risks taking into account survival time) were calculated using Cox Proportional Hazards regression. The regression models adjusted for a number of potential confounding factors, including age, sex, education level, physical activity level, vitamin D intake, total caloric intake, self-reported diabetes at baseline, and calcium supplement use (when modeling total calcium intake). Separate models considered all cardiovascular events, and only those occurring more than 2 years after the study period began.
There were few statistically significant associations, positive or negative, between total calcium intake in any amount and cardiovascular events. Those reported were as follows. As compared with those in the lowest quartile of calcium intake (Quartile 1), those in Quartile 3 had a lower relative risk of MI (hazard ratio [HR] = 0.69, 95% confidence interval [CI] 0.50 to 0.94). This association persisted when events in the first 2 years were excluded (HR = 0.67, 95% CI 0.48 to 0.94). When compared with those in Quartile 1, those in Quartile 2 had a higher relative risk of stroke (HR = 1.50, 95% CI 1.06 to 2.11), but this association did not remain significant when the first 2 years of events were excluded. Finally, in a model that excluded the first 2 years of events, those in Quartile 2 had a higher relative risk of death from cardiovascular causes as compared with those in Quartile 1 (HR = 1.51, 95% CI 1.05 to 2.17). There was no statistical evidence of a dose-response relationship for any outcome. In models assessing the effect of supplement use, those who reported taking only calcium supplements had a higher risk of MI as compared with those reporting no use of any supplements (HR = 2.39, 95% CI 1.12 to 5.12 when considering all events, and HR = 2.70, 95% CI 1.26 to 5.79 when excluding the first 2 years). There was no similar association with the other outcomes or for those taking other supplements as well as calcium.
Commentary
Simply because of its size and design, this study carries considerable weight. Findings from small, tightly controlled experiments (such as clinical feeding studies) must be interpreted with caution until they are validated in larger samples and under more natural conditions. This study's size builds confidence in the precision of its estimates. Its prospective design helps to minimize numerous kinds of potential bias in its results. At the same time, its observational nature (dietary elements were not "prescribed," simply recorded) adds greatly to its generalizability. The dietary patterns of the participants are likely to reflect those of a much larger population. Finally, the use of hard endpoints, events that are unlikely to go unnoticed, rather than markers that require active efforts to identify, increases the probability of complete ascertainment. However, there are also a few issues in the design of this study that should affect how we evaluate its conclusions.
First, diet data were collected only once, at baseline. This ignores any changes in diet over the course of the study. It could be argued that any such changes would have been too proximal to have much effect on the outcomes, but there is evidence that even short-term changes can have measurable effects on some relevant parameters, such as markers of inflammation. Such changes may also have been motivated by health states related directly or indirectly to the outcomes of interest. Information on supplements was collected at each follow-up interval as well as at baseline. This permitted a richer analysis, including a separate set of models for each participant's most recent reported supplement use. Balancing this is the fact that these separate analyses did not produce results that were materially different from those obtained with baseline data.
Second, the FFQ method of collecting diet information is good, but not perfect. In this study, the FFQ results were validated using twelve 24-hour food recall exercises. The correlation between the FFQ and what was reported for the 24-hour recall periods for the two most important categories for calcium intake dairy foods and non-alcoholic beverages were 0.58 and 0.70, respectively. This suggests that information about general eating habits and information about specific consumption in multiple intervals did not match up ideally. Balancing this to some degree is the large sample size, which would tend to minimize the impact on precision of individual mismeasurements. However, it is difficult to assess the likely impact on the results of any systematic bias in the assessment of dietary exposures.
Third, the authors note that there may have been underreporting of calcium supplement use. They point out that their reported rate (3.6%) is lower than that of another study of older Germans (8% for men and 27% for women) and lower than the rate reported in U.S. National Health Interview Survey data (11% overall). This may imply that supplement use was actually higher in the EPIC study than it appeared. If so, a critical finding increased risk with supplement use might have been understated.
These issues notwithstanding, this paper should dampen enthusiasm for the use of calcium in the prevention of cardiovascular disease. There was limited earlier support for such a relationship, but this paper is generally in line with the finding of no association reported in most studies.1-4 Although adequate calcium intake remains important for bone health, calcium from dietary sources appears to have little or no impact on cardiovascular events, and calcium supplementation may add some risk without any concomitant benefit.
References
1. Umesawa M, et al. Dietary calcium intake and risks of stroke, its subtypes, and coronary heart disease in Japanese: The JPHC Study Cohort I. Stroke 2008;39:2449-2456.
2. Bostick RM, et al. Relation of calcium, vitamin D and dairy food intake to ischemic heart disease mortality among postmenopausal women. Am J Epidemiol 1999;149:151-161.
3. Umesawa M, et al. Dietary intake of calcium in relation to mortality from cardiovascular disease: The JACC Study. Stroke 2006;37: 20-26.
4. Larsson SC, et al. Potassium, magnesium, and calcium intakes and risk of stroke in women. Am J Epidemiol 2011;174:35-43.