At the end of July (2010) there was another nasty swipe at dietary supplements (Natural Health Products, for our Canadian friends). Again, the negative report came not from a single study, but from a particularly troubling type of report called a meta-analysis. Rising like a spider on a lethal web, this type of report keeps popping out of nowhere, attempting to make connections between studies where none may have been found before. In fact, we rarely see a competently designed meta-analysis on nutrients because there are many variables that come into play that simply don’t exist for isolated pharmaceutical drugs that aren’t (or at least aren’t supposed to be) present in our normal food supply. The interplay of nutrients affects each others’ metabolism, serum levels, and activities in a live human body. So when I see yet another meta-analysis getting major press coverage based on fairly flimsy evidence, I cringe and wonder why researchers and journalists fail to see the obvious flaws in their big story. Perhaps drug researchers are trying to branch out into nutrient research and just get in over their heads because they fail to see the complexity of nutrient research design.
The alleged danger this time: supplemental calcium was associated with a 30% increased risk of myocardial infarction, so the risks now outweigh the benefits. But was this proven? NO! Absolutely not. Nada. Zilch. Even the authors of this analysis know better than to claim that their report was definitive proof of these alleged dangers.
First of all, there was no increase in deaths in the groups given supplemental calcium. There was an increase in non-fatal heart attacks; but only in people who had already high calcium intake from their diets and also took high doses of supplemental calcium that they apparently didn’t need. And there was no problem demonstrated when calcium was given along with supplemental vitamin D, so maybe the problem was really that some folks had inadequate levels of that essential vitamin to deal with a high calcium intake.
But since vitamin D is both in the diet (as is calcium) and made from sunshine under the right conditions, these meta-analysis authors don’t really know how much vitamin D these people actually had, confounding their data. They don’t know because it would have required a rigorous study design that looked at all three intakes: food, supplements, and sunlight exposure at the times of day and year where vitamin D could be internally produced. Or at least measuring the serum before treatment to detect vitamin D levels and eliminate that as a variable. But that wasn’t done, perhaps because these researchers only looked at other people’s studies rather than running a human clinical trial themselves.
This points out perhaps the major problem with this type of study: since a meta-analysis is only a statistical model that cannibalizes previously published work to “mine” data in ways that were not planned by the original study designers, it has many built-in limitations. For example, cardiovascular outcomes were admittedly not intended as the primary endpoints in any of the 15 studies ‘Mixmastered’ together to make up this meta-analysis, so data on cardiovascular events were not gathered in the usual standardized manner. In plain English: the few cherry-picked studies included in this report had not done the type of standardization and control of variables needed to properly design a robust cardiovascular study because the original studies actually were looking at calcium’s effects on bone health, not heart health. This was, in fact, an admitted limitation of the current meta-analysis.
A good meta-analysis tries to pick the largest possible number of studies with similar designs to pool their results and in effect try to get a larger, hopefully more significant number of virtual test subjects. The farther from that model the meta-analysis gets, the more variables get introduced to confound the researchers. And since control of variables is the essential competence of the scientific method, we unfortunately find that an inexpertly executed meta-analysis is a misleading and erroneous scientific tool; like a ruler that has not been calibrated correctly and implies that our measurements are accurate when they may be way off the mark.
Researchers utilizing the imprecise tool of meta-analysis should not delude themselves that their work is definitive, because it almost never is. Nor should they become media darlings because they have espoused new theories - based on an unproven brew of mathematical models - that usually are already contradicted by a lot of better designed primary science. Nor should their attempts to re-examine previously published studies for results which were never intended to be measured scare us away from taking essential nutrients, which are commonly more healthful than harmful. As the current meta-analysis reported, the problem was that people who ate a lot of calcium who also took a lot of supplemental calcium without the benefits of taking vitamin D or other bone-forming cofactors had more non-fatal heart attacks. No one died. This implies that, at the most, one should not take a single nutrient in excess as if it were a magic drug, especially when they probably already get plenty in the diet. Don’t most of us know that already?
Nutrients are synergistic; therefore an imbalanced diet, including unneeded supplements that may tilt one even more into imbalance, may not properly support good health. Heck, I could have told you that before all this media fuss. But, with enough vitamin D the problem magically disappeared!. Supplementation of vitamin D apparently re-established good calcium metabolism, even at high levels of intake. The true issue then isn’t so much related to variations in calcium and vitamin D supplies, not to mention the unknown availability of other essential bone nutrients. It’s the misapplication of a drug model to a nutrient while ignoring known variables that affect the body’s proper use of that nutrient. One shouldn’t give a lot of calcium to those with adequate dietary intake, nor to those with insufficient vitamin D (and K, magnesium, et al) intake. This type of thought process is elementary to a nutritionist, of course, but apparently not to the well credentialed authors of many flawed drug model meta-analyses and their often overreaching conclusions.
You may be interested to know that a world-class researcher who has been working on calcium metabolism and osteoporosis for more than 50 years, the respected scientist who drafted the World Health Organization’s dietary calcium recommendations, strongly opposed the conclusions of the current calcium meta-analysis. Professor Chris Nordin from the Royal Adelaide Hospital in Australia was interviewed by ABC News. This news report cited Professor Nordin as saying that the meta-analysis was misleading because it improperly included studies involving a mixture of men and women, and the findings were not statistically significant. "Men are much more liable to heart attacks than women but women need calcium far more than men, so it is absurd to publish a study of the effect of calcium on the heart without separating men from women," he was quoted. Professor Nordin noted the fact that calcium supplements are predominantly recommended for and used by postmenopausal women because their bone loss is due to an increase in bone breakdown, which responds well to calcium supplementation (and vitamin D, if necessary). But he reports that calcium is seldom recommended for elderly men because their bone loss has a different cause, which is seldom caused by a need for more of that mineral. "Concluding that calcium supplements can lead to a 30 per cent increase in heart attack risk is quite premature and alarmist and can only set back the cause of osteoporosis prevention which should be our primary objective," he said.
According to ABC News, Osteoporosis Australia has also questioned the findings of the meta-analysis because many long-term studies have shown calcium supplements are safe and effective. In a statement posted on its website, Osteoporosis Australia says the weight of evidence to date indicates no increased risk of heart attacks from taking calcium supplements, which it says are an effective way of reducing fracture risk and bone loss in older men and women who have diets low in calcium.
This should point out the folly of well-meaning researchers who design studies to investigate topics that they don’t properly understand. In those cases, they simply don’t realize that there are other essential factors that can change everything and require a completely different and more comprehensive study design. That’s par for the course when meta-analyses are used to sort through previously published nutrient studies. Basic design flaws often make meta-analyses fatally flawed, despite the apparently sophisticated mathematical models and methods that researchers incorrectly try to apply to the data. Is it any wonder that we’re so confused about nutrition when researchers gain worldwide prominence for issuing highly questionable sensational reports that contradict the scientific consensus - but never have to say they’re sorry?
Calcium does not cause more heart attacks in well-designed primary human clinical studies. Healthy people should not be concerned about taking essential nutrients just because certain studies were flawed or performed on sick or at-risk populations and may have had negative results. Many of those negative reports have been challenged and may not have been verified and replicated in well designed clinical trails. The conclusions of such reports are sensational and well publicized precisely because they seem to negate our previous scientific consensus; which should make them more, not less, suspect. But the modern news cycle seems to thrive on such controversy, without caring how confused we get about what’s healthy and what’s not. But don’t worry; you have Honest Nutrition to help you sort it all out!
REFERENCES:
Bolland MJ, Avenell A, Baron JA, Grey A, Maclennan GS, Gamble GD, Reid IR. Effect of calcium supplements on risk of myocardial infarction and cardiovascular events: meta-analysis. BMJ. 2010 Jul 29;341:c3691. doi: 10.1136/bmj.c3691. PubMed PMID: 20671013.
http://www.bmj.com/cgi/content/full/341/jul29_1/c3691?view=long&pmid=20671013
http://www.abc.net.au/news/stories/2010/08/03/2972399.htm?section=justin
No comments:
Post a Comment