Search Results for: folic acid

Are Low Doses of Niacin Dangerous?

Posted by on April 25, 2012 109 comments

In Food Fortification: A Risky Experiment?, Mar 23, 2012, we began looking at the possibility that fortification of food, especially the enriched flours used in commercial baked goods, with niacin, iron, and folic acid may have contributed to the obesity and diabetes epidemics.

As this plot shows, fortification caused intake of per capita niacin intake in the United States to rise from about 20 mg/day to about 32 mg/day:

Multivitamins typically contain about 20 mg niacin, so (a) a typical American taking a multivitamin is getting 52 mg/day niacin, and (b) if the increase of 12 mg/day due to fortification is dangerous, then taking a multivitamin would be problematic too.

There wasn’t evidence of niacin deficiency at 20 mg/day. The RDA was set at 16 mg/day for men and 14 mg/day, levels that equalize intake with urinary excretion of niacin metabolites [source: Dietary Reference Intakes]. Fortification of grains with niacin was designed to make refined white wheat have the same niacin content as whole wheat, not to rectify any demonstrated deficiency of niacin.

B-vitamins are normally considered to have low risk for toxicity, since they are water soluble and easily excreted. But recently, scientists from Dalian University in China proposed that niacin fortification may have contributed to the obesity and diabetes epidemics. [1] [2]

Niacin, Oxidative Stress, and Glucose Regulation

The Chinese researchers note that niacin affects both appetite and glucose metabolism:

[N]iacin is a potent stimulator of appetite and niacin deficiency may lead to appetite loss [10]. Moreover, large doses of niacin have long been known to impair glucose tolerance [23,24], induce insulin resistance and enhance insulin release [25,26].

They propose that niacin’s putative negative effects may be mediated by oxidative stress, perhaps compounded by poor niacin metabolism:

Our recent study found that oxidative stress may mediate excess nicotinamide-induced insulin resistance, and that type 2 diabetic subjects have a slow detoxification of nicotinamide. These observations suggested that type 2 diabetes may be the outcome of the association of high niacin intake and the relative low detoxification of niacin of the body [27].

The effect of niacin on glucose metabolism is visible in this experiment. Subjects were given an oral glucose tolerance test of 75 g glucose with or without 300 mg nicotinamide. [1, figure source]

Dark circles are from the OGTT with niacinamide, open circles without. Plasma hydrogen peroxide levels, a marker of oxidative stress, and insulin levels were higher in the niacinamide group. Serum glucose was initially slightly higher in the niacinamide group, but by 3 hr had dropped significantly, to the point of hypoglycemia in two subjects:

Two of the five subjects in NM-OGTT had reactive hypoglycemia symptoms (i.e. sweating, dizziness, faintness, palpitation and intense hunger) with blood glucose levels below 3.6 mmol/L [64 mg/dl]. In contrast, no subjects had reactive hypoglycemic symptoms during C-OGTT. [1]

Of course 300 mg is a ten-fold higher niacinamide dose than most people obtain from food, but perhaps chronic intake of 32 mg/day (52 mg/day with a multivitamin) daily over a period of years have similar cumulative effects on glucose tolerance as a one-time dose of 300 mg.

Is There a Correlation with Obesity?

OK. Is there an observable relationship between niacin intake and obesity or diabetes?

There may be, but only with a substantial lag. Here is a figure that illustrates the possible connection [2, figure source]:

Niacin intake maps onto obesity rates with a 10-year lag. After niacin intake rose, obesity rates rose 10 years later. Note the scaling: a 60% increase in niacin intake was associated with a doubling of obesity rates 10 years later.

Obesity leads diabetes by about 15 years, so we could also get a strong correlation between niacin intake and diabetes incidence 25 years later. The scaling in this case would be a 35% increase in niacin associated with a 140% increase in diabetes prevalence after a lag of 25 years.

How seriously should we take this? As evidence, it’s extremely weak. There was a one-time increase in niacin intake at the time of fortification. A long time later, there was an increase in obesity, and long after that, an increase in diabetes. So we really have only 3 events, and given the long lag times between them, the association between the events is highly likely to be attributable to chance.

It was to emphasize the potential for false correlations that I put the stork post up on April 1 (Theory of the Stork: New Evidence, April 1, 2012). Just because two data series can be made to line up, with appropriate scaling of the vertical axis and lagging of the horizontal axis, doesn’t mean there is causation involved.

Is There Counter-Evidence?

Yes.

If niacin from wheat fortification is sufficient to cause obesity or diabetes, with an average intake of 12 mg/day, then presumably the 20 mg of niacin in multivitamins would also cause obesity or diabetes.

So we should expect obesity and diabetes incidence to be higher in long-time users of multivitamins or B-complex vitamins.

But in fact, people who take multivitamins or B-complex vitamins have a lower subsequent incidence of obesity and diabetes.

One place we can see this is in the Iowa Women’s Health Study, discussed in a previous post (Around the Web; The Case of the Killer Vitamins, Oct 15, 2011). In that post I looked at a study analysis which was highly biased against vitamin supplements; the authors chose to do 11-factor and 16-factor adjustments designed to make supplements look bad. The worst part of the analysis, from my point of view, was using obesity and diabetes as adjustment factors in the regression analysis. As you can see in the table below, multivariable adjustment including obesity and diabetes significantly raises the mortality associated with consumption of multivitamins or B-complex supplements:

This increase in hazard ratios (“HR”) with adjustment for obesity and diabetes almost certainly indicates that the supplements reduce the incidence of these diseases.

Multivitamins are protective in other studies too. The relation between multivitamin use and subsequent incidence of obesity was specifically analyzed in the Quebec Family Study, which found that “nonconsumption of multivitamin and dietary supplements … [was] significantly associated with overweight and obesity in the cross-sectional sample.” [3]

Does this exculpate niacin supplementation? I don’t think so. In general, improved nutrition should reduce appetite, since the point of eating is to obtain nutrients. So it’s no surprise that multivitamin use reduces obesity incidence. But multivitamins contain many nutrients, and it could be that benefits from the other nutrients are concealing long-term harms from the niacin.

Conclusion

At this point I think the evidence against niacin is too weak to convict in a court of law.

Nevertheless, we do have:

  • Clear evidence that high-dose (300 mg) niacinamide causes oxidative stress and impaired glucose tolerance. If niacinamide can raise levels of peroxide in the blood, what is it doing at mitochondria?
  • No clear evidence for benefits from niacin fortification or supplementation.

Personally I see no clear evidence that niacin supplementation, even at the doses in a multivitamin, is likely to be beneficial. Along with other and stronger considerations, this is pushing me away from multivitamin use and toward supplementation of specific individual micronutrients whose healthfulness is better attested.

I also think that food fortification was a risky experiment with the American people, and stands as yet another reason to avoid eating grains and grain products. (And to rinse white rice before cooking, to remove the enrichment mixture.)

References

[1] Li D et al. Chronic niacin overload may be involved in the increased prevalence of obesity in US children. World J Gastroenterol. 2010 May 21;16(19):2378-87. http://pmid.us/20480523.

[2] Zhou SS et al. B-vitamin consumption and the prevalence of diabetes and obesity among the US adults: population based ecological study. BMC Public Health. 2010 Dec 2;10:746. http://pmid.us/21126339.

[3] Chaput JP et al. Risk factors for adult overweight and obesity in the Quebec Family Study: have we been barking up the wrong tree? Obesity (Silver Spring). 2009 Oct;17(10):1964-70. http://pmid.us/19360005.

Food Fortification: A Risky Experiment?

Posted by on March 23, 2012 81 comments

We’ve learned enough in the last two years to revisit the supplementation advice from our book, and toward that end I am starting a series on micronutrients.

I’ve recently been looking at some papers studying the effects of food fortification with micronutrients. These changes provide a sort of “natural experiment” which may provide insight into the benefits and risks of supplementation.

Fortification of Food

Grain products are the most important category of fortified foods. Industrially produced baked goods must generally use enriched flour, and Wikipedia (“Enriched Flour”) tells us what they’re enriched with:

According to the FDA, a pound of enriched flour must have the following quantities of nutrients to qualify: 2.9 milligrams of thiamin, 1.8 milligrams of riboflavin, 24 milligrams of niacin, 0.7 milligrams of folic acid, and 20 milligrams of iron.

This is an ironic choice of nutrients. While thiamin and riboflavin are harmless, niacin, folic acid, and iron are three micronutrients we recommend NOT supplementing in the book. Another nutrient we recommend NOT supplementing, vitamin A, is also a fortified nutrient, although not in flour.

Sales Cartoon #6021 by Andertoons

Perhaps not even for that!

A history of nutrient fortification over time can be found at this USDA site. Enrichment has a long history, but the amount of fortification has increased substantially since the 1960s. Enrichment mixtures were added to rice, cornmeal/grits, and margarine beginning in 1969, and to ready-to-eat cereals, flour, and semolina beginning in 1973. Inclusion of high levels of folic acid in all enriched foods became mandatory in 1998.

You may have noticed that when putting raw rice in water, a white powder comes off the rice. This is the enrichment mixture which contains folic acid. According to the American Rice Company (hat tip: Matthew Dalby),

The enrichment mixture is applied to rice as a coating. Therefore, it is recommended that rice not be rinsed before or after cooking and not be cooked in excessive amounts of water and then drained. The enrichment … would be lost.

This is useful information: We can remove the enrichment coating by rinsing rice before cooking. That may turn out to be a good idea!

The Contribution of Fortification to Nutrient Intake

Using USDA data for the four nutrients most likely to be harmful in excess, I made up a chart of the contribution of fortified nutrients to total nutrient intake among Americans. It looks like this:

You can see sharp rises in fortified niacin and folic acid in 1973, in iron in 1983, and again in folic acid in 1998. By 1998, folic acid in fortified foods constituted 44% of all dietary folate, and enrichment mixtures provided one-third of all iron and niacin. Fortified vitamin A provided about 10% of all dietary vitamin A from 1964 through 2000.

Folic Acid

Here is a chart of per capita daily intake of fortified folic acid plus natural food folate in the United States since 1950:

Folate intake from foods has always been around 300 mcg per day, and jumped sharply when folic acid intake became mandatory in 1998. The USDA estimates that intake of folate, including folic acid, jumped from 372 mcg per person per day in 1997 to 678 mcg in 1998, and has remained above 665 mcg ever since (source).

For those who eat a lot of wheat products, intake may be even higher. A pound of enriched white flour has 770 mcg folic acid along with its 1660 calories. If Americans were getting 372 mcg folate from food prior to folic acid fortification, then someone eating a pound of enriched wheat products per day would be getting about 1,142 mcg folate from all food sources.

It’s not uncommon to eat substantial amounts of enriched wheat. The typical American eats 474 g (1800 calories) carbohydrate per day. Most of that is from enriched grains. Those eating industrially produced breads, cookies, crackers, and breakfast cereals may have a very high folic acid intake.

Add in a multivitamin – most multivitamins have 400 mcg and prenatal vitamins have 800 mcg – and a sizable fraction of the population has a folate intake of 1,500 to 1,900 mcg per day, 1200 to 1600 of it as synthetic folic acid. This is well above the tolerable upper limit (UL) for folic acid of 1000 mcg (Wikipedia, “Folate”).

Averaged over all Americans, folic acid from fortified foods comprises 44% of all food-sourced folate, but for Americans taking a multivitamin folic acid becomes 65% of all folate and, for those taking a prenatal vitamin, 75%.

There are several potential health problems that could arise from excessive intake of folic acid, and I’ll explore a few in this series.

Iron and Niacin

Iron intake has risen by about 50% due to fortification:

Niacin intake has also risen about 50%:

These two nutrients have similar concerns:

  • An excess of each promotes infections. Niacin (in the NAD+ form) is the rate-limiting factor in bacterial metabolism. Iron is a critical mineral for oxygen handling and is needed by most infectious pathogens; in fact the immune response tries to lock up iron in ferritin during infections.
  • Both niacin and iron are involved in oxygen handling during metabolism and an excess of each can aggravate oxidative stress.

Vitamin A

Although fortification never increased vitamin A intake by more than 10%, it may serve as a marker for consumption of artificial sources of vitamin A from supplements. Moreover, total food intake of vitamin A was apparently affected by fortification; food intake of vitamin A rises in the 1960s when fortification was growing, and falls after 2000 when intake of fortified vitamin A decreased:

In the book we noted studies showing that people whose intake of vitamin A was above 10,000 IU/day tended to have higher mortality. This was most commonly observed in people taking multivitamins.

There was a period of enthusiasm for vitamin A supplementation between the 1960s and 2000. Multivitamins had more vitamin A in that period. After studies showed negative results, the vitamin A content of multivitamins was reduced.

It is possible that the source of problems may not be vitamin A per se, but degradation products of vitamin A. I’ve previously blogged about how vitamin A plus DHA (a fatty acid in fish oil) plus oxidative stress can produce highly toxic degradation products (see DHA and Angiogenesis: The Bottom Line, May 4, 2011; Omega-3s, Angiogenesis and Cancer: Part II, Apr 29, 2011; Omega-3 Fats, Angiogenesis, and Cancer: Part I, Apr 26, 2011).

Naturally occurring vitamin A in foods is located in lipid fractions and protected from oxidation by accompanying antioxidants (eg vitamin E) and oxidation-resistant lipids. Vitamin A from fortification is not so carefully protected. The Food and Agriculture Organization of the United Nations comments:

Foods which have been successfully fortified with vitamin A include margarine, fats and oils, milk, sugar, cereals, and instant noodles with spice mix. Moisture contents in excess of about 7-8% in a food are known to adversely affect the stability of vitamin A. Beyond the critical moisture content there is a rapid increase in water activity which permits various deteriorative reactions to occur. Repeated heating, as may be experienced with vegetable oils used for frying, is known to significantly degrade vitamin A. The hygroscopic nature of salt has prevented its use as a vehicle for vitamin A fortification in countries of high humidity. In trying overcome this problem, a new vitamin A fortificant, encapsulated to provide an additional moisture barrier, was evaluated with limited success. The cost of using highly protected fortificants can be prohibitive in many cases.

There aren’t many foods that don’t contain 7% water, or acquire it after fortification, so degradation is a real concern.

Vitamin A in multivitamins may also be exposed to degradation. The possibility of vitamin A degradation, especially in combination with DHA from fish oil and oxidative stress, is why I’m skeptical of the health merits of fermented cod liver oil.

Conclusion

I think exploring the effects of fortification will be an interesting topic.

We will consider whether fortification may play a role in various diseases that have become more common since 1970 or 1998, such as obesity, diabetes, and autism.

And we will consider what the health effects of food fortification may tell us about how to optimize micronutrient supplementation.

 

Around the Web; PaleoFX Edition

Posted by on March 10, 2012 40 comments

I’ll be speaking and sitting on two panels at the Paleo(fx) conference this week, Wednesday-Friday in Austin, Texas. The meeting has a great lineup of speakers and I’m looking forward to meeting a lot of interesting people there. If you’re there, be sure to introduce yourself!

[1] Music to Read By: Glen Campbell has Alzheimer’s, and is doing a farewell tour. His final concert will be June 30 in Bayfield, Wisconsin. Here is “Wichita Lineman”:

And “Rhinestone Cowboy”:

[2] Interesting posts this week:

My favorite post of the week: Emily Deans reports on the Pathogen Host Defense theory of depression. Depression is part of the immune response to infection: it is how the immune system gets us to rest and avoid other people so as to keep the infection from spreading. The supporting evidence: every genetic allele known to increase the incidence of depression, also increases immunity to infectious disease. Moral of the story: if you’re depressed, find and treat your infections.

Bix discusses the role of histamine in depression. Antihistamines relieve depression. Of course, they may also inhibit the immune system from fighting an infection.

Kamal Patel has a new website on pain management, paindatabase.com; he opens with an excellent discussion of nightshades. Peter Frost notes that redheads are more sensitive to pain.

Dr. Steve Parker reviews the classic study in which diabetic Australian aborigines were returned to their ancestral lifestyle.

Angelique Corthals presents an offbeat theory of multiple sclerosis. Another offbeat theory: Via Melissa McEwen, a paper argues that Crohn’s may result from bacteria which survive in refrigerators.

Chris Kresser discusses folic acid dangers, a topic I’ve recently done some thinking about.

Allan Balliet has a podcast interview with Dr Thomas Cowan, author of The Fourfold Path to Healing: Working with the Laws of Nutrition, Therapeutics, Movement and Meditation in the Art of Medicine.

Tony Federico was able to fix his cholesterol by adding carbs. CarbSane finds she sleeps better on 100g per day starches (PHD proportions) than she did on VLC. Cameron also had trouble on VLC.

Serum IL-6 levels predict mortality and disability. IL-6 is part of the immune response to many infections.

An 87-year-old billionaire aims to live to 125. He eats a lot of fruit and vegetable smoothies.

Parvovirus B19 is present in the thyroid glands of 90% of Hashimoto’s patients.

Stephan Guyenet got some nice exposure at Boing Boing, talking about how the cafeteria diet seduces rats. Matt Metzgar’s personal experiments are supportive of Stephan’s view.

Sean at PragueStepChild has a terrific post pointing out the flawed logic underlying a new “food reward in a pill” drug.

Seth Roberts warns that supplement quality can vary widely.

Apparently Dr Steve Phinney thinks the best fat is a mix of butter, olive oil, and canola oil. I think it’s possible to do better.

The Daily Mail reports that passing weak electric currents through the brain can lift patients out of depression for up to six weeks.

Via Dave Asprey, an “M2A” (mouth to anus) capsule enables photography of the intestine. He’s got video. Dave also has a rapid fat loss protocol: a diet of coffee, butter, and supplements. I prefer our Food for a Fast.

Warning: Dr Clark says that coffee is bad for people with gluten sensitivity, due to cross-reactive antibodies.

Dr Mercola defends free speech, and objects to an FDA assertion that walnuts are drugs. Yahoo reports that the United States Department of Agriculture (USDA) is buying 7 million pounds of “pink slime” beef for the school lunch program.

I didn’t know this: “In healthy individuals, blood sugar rises to ~135 mg/dL after 15 minutes of vigorous exercise…. In Type 1 diabetics, the same intensity and duration of exercise produced blood sugars that exceeded 150 mg/dL in spite of continuous insulin infusion.” (from Sam Knox)

Via PaleoHacks, how a pig gets divided up into meats.

Matthew Green satirizes Paleo.

Finally, Apple came out with a new product last week, and Yoni Freedhoff has the marketing video:

[3] Cute animal:

Via Yves Smith.

[4] Meet Ray Audette: The author of NeanderThin: Eat Like a Caveman to Achieve a Lean, Strong, Healthy Body and one of the pioneers of Paleo:

[5] Not the weekly video: Dr. Mark Cucuzzella offers “The Principles of Natural Running”:

[6] Shou-Ching’s Photo Art:

[7] Weekly Video: Ever wondered what auroras look like from space?

Around the Web; The Case of the Killer Vitamins

Posted by on October 15, 2011 56 comments

I’d like to thank Patrick Timpone for a very enjoyable interview on The Morning Show at One Radio Network. Here is the MP3; I’m on for the second half of the show. You can find a zip file at the archive for October 13. Patrick’s producer Sharon tells me that she’s already benefited from our book:

I was following The Primal Diet and since I read the book, I’ve been allowing myself potatoes and rice and doing very very well on them among doing some other things you recommend.

Also, I’d like to thank Jimmy Moore once more for hosting his highly entertaining “safe starch” symposium (Jimmy’s original post; my response, here and at Jimmy’s). It was great to get the opportunity to explain ourselves to so many people in the low-carb and Paleo movements.

Jimmy is planning to try our diet for a week in November, which will be a good occasion for us to publish a 7-day meal plan. We’ll invite anyone who’s curious to try the diet along with Jimmy, and compare notes.

[1] Interesting posts this week:

Angelo Coppola on Latest in Paleo wonders if Denmark’s saturated fat tax will apply to mother’s milk. If so, it’s bad news for unemployed infants! (He also discusses the “safe starch” debate.)

I once knew a French astronomer who died from snorting cocaine while observing at 14,500 feet. Emily Deans makes me wonder:  Did he have Crisco for dinner?

Stan the Heretic offers his mitochondrial dysfunction theory of diabetes. Peter Dobromylskyj and JS Stanton are also developing ideas along this line. Speaking of JS, his post this week has some great photos of Sierra wildflowers and reflections on the state of the Paleo community.

CarbSane partially confirms Dr. Ron Rosedale: eating carbs does raise leptin levels compared to eating fat, but it is a mild rise over an extended period of time, not a “spike.”

Beth Mazur explains why her bathroom door is always closed.

Chris Kresser discusses why chronic illness often generates a form of hypothyroidism, low T3 syndrome.

Joshua Newman knows how to flatter.

How solid is the case against Andrew Wakefield? Autism is certainly characterized by intestinal dysfunction, and Age of Autism notes that distinguished scientists are citing Wakefield’s work.

Richard Nikoley claims he doesn’t know the words to “Kumbayah.”

Seth Roberts points out that the Specific Carbohydrate Diet has been curing Crohn’s for 80 years, but still no clinical trial.

Jamie Scott, That Paleo Horse Doctor, asks: Why do horses get laminitis?

We’ve quoted vegetarian Dr. Michael Greger’s concerns about arsenic in eggs. I’m more concerned about soy protein in eggs.

Following Steve Jobs’s death, Tim asked for an opinion about the unconventional cancer therapies of Dr Mercola’s friend Nicholas Gonzalez. David Gorski, toward the end of a detailed examination of Jobs’s medical condition and treatment, links to his own claim that the Gonzalez protocol is “worse than useless.”

[2] Music to read by:

[3] Cute animal photo:

[4] Notable comments this week:

PeterC’s dad, who has diabetes, is doing well on our diet. Daniel’s stepdad had a similar experience.

Helen informs us that sweet potato intolerance may be due to raffinose.

Mario Iwakura gives us his infectious theory of diabetes. I think a lot of the cases of disrupted glucose regulation, where people get frequent hyperglycemic and hypoglycemic episodes, may be due to occult infections.

Dr Jacquie Kidd (who blogs at drjacs.com) has gotten some great advice from Jamie Scott.

Ellen tells us of cases of iodine supplementation controlling diabetes.

Ned is looking for grass-fed cowbells.

[5] Do Vitamins Kill?: An analysis of the Iowa Women’s Health Study came out this week, and it purported to show that nearly all supplements except calcium and vitamin D increased mortality, with iron being the worst. Oskar asked us to look into it, so we did.

The study followed a large number of women in Iowa, and queried them several times about supplement use. In 1986, the baseline, the women had an average age of 62 (range of 55 to 69) and 66% were taking supplements. By 2004, the surviving women had an average age of 82 and 85% were taking supplements.

Here is the data on overall mortality vs supplement use:

“Cases” are instances of someone dying. “HR” or hazard ratio is the likelihood of dying if you supplement divided by the likelihood of dying if you don’t. Note that all the hazard ratio estimates are “adjusted.”

Unadjusted Hazard Ratios

The left columns of the table give us death statistics and allow us to calculate raw hazard ratios, with no adjustment whatsoever. Seven of the supplements have unadjusted HRs below 1.00, eight have unadjusted HRs above 1.00. The 15 HRs average to 1.01. Without copper, which has an unadjusted HR of 1.17, they average to 0.998. In short, death rates among supplementers were almost identical to death rates among non-supplementers.

This is interesting because supplement usage rose rapidly with age. It was 66% at age 62 and 85% at age 82. Supplement users were, on average, older than non-supplement users. But mortality rises rapidly with age. So there should have been a lot more deaths among the supplement users, just because of their more advanced age.

The paper should have, but didn’t, report age-adjusted hazard ratios. Adjusting for age is very important, since mortality depends strongly on age, and so does supplement use. However, it’s obvious what the result of age-only adjustment would have been. Supplement usage would have shown a substantial reduction in the risk of dying.

Hazard Ratios Adjusted for Age and Energy Intake

The least-adjusted hazard ratios reported in the paper are adjusted for age and energy intake.

The energy intake adjustment is disappointing, because energy intake is affected by health: healthier people are more active and eat more, and obese people also eat more. Including indices of health as independent variables in a regression analysis will tend to mask the impact of the supplements on health, creating misleading results.

However, let’s go with what we have. Based on “Age and Energy Adjusted” hazard ratios, supplements generally decrease mortality. Nine of the fifteen supplements decreased mortality, five increased mortality. At the 95% confidence interval, five supplements decreased mortality, only one increased mortality.

Looking at the specific supplements, results are mostly consistent with our book analysis. Let’s start with the five that showed harm:

  • Folic acid and iron – two nutrients we regard as dangerous and recommend not supplementing – both elevate mortality, as we would expect. Iron is particularly harmful, and should generally be avoided by women once they have stopped menstruating.
  • Multivitamins slightly increase mortality, a result that has been found before and that we acknowledge in the book. This is probably due to (a) an excess of folic acid, (b) an excess of iron (if the women are taking iron-containing multis after menopause), (c) an excess of vitamin A (this is no longer the case – multi manufacturers have reduced the A content of vitamins in response to data – but in 1986-2004 most multis contained substantial amounts of A) which is harmful in women with vitamin D and/or K2 deficiencies (both extremely common, and D deficiency in this cohort is supported by the benefits of D and calcium in the study and the northerly latitude of Iowa) or (d) imbalances in other nutrients; for reasons of bulk multis tend to lack certain minerals, notably magnesium and calcium.
  • Vitamin B6 is an anomaly, as we wouldn’t expect B6 to be harmful in moderation. I’m guessing B6 would have been taken to reduce high homocysteine and for this purpose would often have been taken along with folic acid, a harmful supplement. Also, B6 should be balanced by vitamin B12 and biotin, and may not have been. Perhaps people with cancer were unaware that B6 promotes tumor growth; (UPDATE: See comments; I was misremembering studies, B12 and folic acid can promote tumor growth, but in other studies B6 looks protective against cancer) indeed, in the breakdown by cause of death in Table 3, B6 increases cancer mortality by 6%, but CVD mortality by only 1%. (Folic acid and vitamin A were other cancer-promoting supplements.) The harm from B6 was not statistically significant and I wouldn’t read much into it.
  • Copper is another anomalous result, but this was the least popular supplement, taken by only 229 women or 0.59%. Copper’s hazard ratios were dramatically affected by adjustment: in the raw data, mortality is only 17% higher among copper supplementers, but after age and energy adjustment it is 31% higher, and multivariable adjustment increases it substantially again. Clearly the effect of copper is highly sensitive to adjustment factors, indicating that copper was being taken by an unusual population. I think the hazard ratio for copper is impossible to interpret without knowing why these women were supplementing copper. If we knew their situation, there would probably be an appropriate adjustment that would make a huge difference in mortality. I would say the numbers are too small, the population too skewed, and the information too limited to draw any conclusion here.

Overall, I would interpret the nine that showed benefits as being highly supportive of micronutrient supplementation. The fact that vitamin A, vitamin B complex, vitamin C, vitamin D, vitamin E, calcium, magnesium, selenium, and zinc all reduced mortality suggests that a well-formulated multivitamin would likely have reduced mortality.

Hazard Ratios After Multivariable Adjustment

Now, what about the “Multivariable Adjusted” results, which were responsible for the headlines?

We have to keep in mind a famous aphorism from the mathematician John von Neumann:

With four parameters I can fit an elephant, and with five I can make him wiggle his trunk.

The multivariable adjustments use 11 parameters and 16 parameters respectively. Using so many parameters lets the investigators generate whatever results they want.

I don’t think it’s a coincidence that both multivariable adjustments substantially increased the hazard ratio of every single one of the 15 supplements. The 11-variable adjustment increased hazard ratios by an average of 7%, the 16-variable adjustment by an average of 8.2%.

Rest assured, it would have been easy enough to find multivariable adjustments that would have decreased hazard ratios for every single one of the 15 supplements.

I believe it verges on the unethical that the variables chosen include dangerous health conditions: diabetes, high blood pressure, and obesity. These three health conditions just happen to be conditions that are often improved by supplementation.

Anyone familiar with how regression analyses work will immediately recognize the problem. The adjustment variables serve as competing explanations for changes in mortality. If supplementation decreases diabetes, high blood pressure, and obesity, and through these changes decreases mortality, the supplements will not get credit for the mortality reduction; rather the decreased diabetes, blood pressure, and obesity will get the credit.

Imagine we had a magic pill that completely eliminated diabetes, obesity, and high blood pressure, and reduced mortality by 20%, with no negative health effects under any circumstances. But if regression analysis showed that non-diabetic, non-obese, and non-hypertensive people had 25% less mortality, then a multivariable adjusted analysis would show that the magic pill increased mortality. Why? Because the elimination of diabetes, obesity, and hypertension should have decreased mortality by 25% (the regression analysis predicts), but mortality was only decreased 20%, so adjusted for diabetes, obesity, and hypertension the magic pill must be credited with the additional 5% dead. The multivariable adjusted HR for the magic pill becomes 0.8/0.75 = 1.067.

Of course, what ordinary people want to know is: Will this magic pill improve my health? The answer to that would be yes.

What (too many) scientists want to know is: Which methodology for analyzing this magic pill data will get me grant money? That depends on whether the funding authorities are positively or negatively disposed toward the magic pill industry. Once you know that, you search for the 16-variable multivariable regression that generates the hazard ratios the authorities would like to see.

My take? Judging by the data in Table 2 plus corroborating evidence from clinical trials reviewed in our book, I would say that a well-formulated supplement program, begun at age 62, may increase the odds of survival to age 82 by something on the order of 5% to 10%. Perhaps not a magic pill; but worthwhile.

[6] Not the weekly video: An exceptional magic show:

[7] Shou-Ching’s Photo Art:

[8] Weekly video: A new tool for stroke recovery: