You don't need to be signed in to read BMJ Blogs, but you can register here to receive updates about other BMJ products and services via our site.


Primary Care Corner with Geoffrey Modest MD: OTC supplements — not what they say they are…

19 Feb, 15 | by EBM

By: Dr. Geoffrey Modest 

The New York Times had a rather remarkable article on Feb 3rd noting that in a NY attorney general’s investigation, approx 80% of herbal supplements from 4 retailers “contained none of the herbs listed on their labels”, but included “cheap fillers like rice and house plants, or substances that could be hazardous to people with food allergies”. Specifically, they analyzed products from GNC, Target, Walmart, and Walgreens (see here). This raises several issues.

–there is a total lack of oversight/regulation for these “dietary supplements”: ie, no FDA or other agency assuring the quantity or quality. I have previously sent out an article on vitamin D supplementation of milk (see N Engl J Med 1992; 326: 1178), finding that only 29% of 42 samples of 13 brands of milk had the amount of vitamin D advertised on their label (within 20%), 62% had < 80% of what was advertised, and 3 of 14 samples of skim milk had NO vitamin D!! So, people relying on any of the “dietary supplements” are at risk of getting none of what they want.

–there is clearly a potential health risk if one has an allergy to an added (but not listed) ingredient to the supplements

–and, unfortunately, I do find that even in the relatively poor urban community where I work, many patients are buying these overpriced supplements (which it turns out are often fraudulently labeled).


Will append the list from the article below (O’Connor, 2015):

gncFrom GNC, Herbal Plus brand- 

Gingko Biloba:

  • No gingko biloba found
  • Did detect allium (garlic), rice, spruce and asparagus

St. John’s Wort

  • No St. John’s Wort found
  • Did detect allium (garlic), rice and dracaena (a tropical houseplant)


  • No ginseng found
  • Did detect rice, dracaena, pine, wheat/grass and citrus


  • Contained garlic


  • No echinacea found
  • Did detect rice in some samples

Saw Palmetto

  • One sample contained the clear presence of palmetto
  • Other samples contained a variety of ingredients, including asparagus, rice and primrose

From Target, Up & Up brand-

Gingko Biloba

  • No gingko biloba found
  • Found garlic, rice and mung/French bean

St. John’s Wort

  • No St. John’s Wort found
  • Found garlic, rice and dracaena (houseplant)


  • Contained garlic
  • One test identified no DNA


  • Most but not all tests detected Echinacea
  • One test identified rice

Saw Palmetto

  • Most tests detected saw palmetto
  • Some tests found no plant DNA

Valerian Root

  • No valerian root found
  • Found allium, bean, asparagus, pea family, rice, wild carrot and saw palmetto

From Walgreens, Finest Nutrition brand-

Gingko Biloba

  • No gingko biloba found
  • Did detect rice

St. John’s Wort

  • No St. John’s Wort found
  • Detected garlic, rice and dracaena


  • No ginseng found
  • Detected garlic and rice


  • No garlic found
  • Detected palm, dracaena, wheat and rice


  • No echinacea found
  • Identified garlic, rice and daisy

Saw Palmetto

  • Contained saw palmetto

From Walmart, Spring Valley brand-

Gingko Biloba

  • No gingko biloba found
  • Found rice, dracaena, mustard, wheat and radish

St. John’s Wort

  • No St. John’s Wort found
  • Detected garlic, rice and cassava


  • No ginseng found
  • Found rice, dracaena, pine, wheat/grass and citrus


  • One sample showed small amounts of garlic
  • Found rice, pine, palm, dracaena and wheat


  • No echinacea or plant material found

Saw Palmetto

  • Some samples contained small amounts of saw palmetto
  • Also found garlic and rice​

Primary Care Corner with Geoffrey Modest MD: Non-cow’s milk and low vitamin D levels

26 Jan, 15 | by EBM

By: Dr. Geoffrey Modest 

No shocker here, but it turns out that non-vitamin D fortified non-cow’s milk (eg goat’s milk or plant-based milk such as soy, rice, almonds…) leads to lower 25-OH vit D levels in the blood. In this Canadian study, they looked at serum 25-OH in kids who were drinking non-cow’s milk (which is only voluntarily fortified with vit D and, if so, with no regulation of content) vs. those on fortified cow’s milk, by law required to have 40 IU vit D/100 ml (see DOI:10.1503/cmaj.140555​). Note that this was an observational study, not RCT — so there might be other confounders (eg, those on non-cow’s milk are on it for specific reasons which could also affect their vitamin D levels, or have other dietary changes in addition to non-cow’s milk).



–2268 children aged 1-6 yo, coming in for routine well-child visits, had dietary history and blood tests. of these, 1950 drank only cow’s milk, 146 only non-cow’s milk, 88 both, 109 neither. Also, 50-60% of each group were on vit D supplementation
–Non-cow’s milk consumption was associated with a 4.2 nmol/L (1.7 ng/ml) decrease in 25-OH vit D per 250 ml milk consumed, vs. those on cow’s milk (p=0.008), and there was a linear gradient in those consuming both, reflecting the amount of non-cow’s milk drunk
–Those drinking exclusively non-cow’s milk were at higher risk of 25-OH vit D levels below 50 nmol/L (20 ng/ml), 11% vs. 4.7% with odds ratio of 2.7 (1.6-4.7)
–And (also not a shocker), there was variation in 25-OH vit D levels depending on whether there was either vit d supplementation (which, unfortunately, they did not quantify) and having darker skin pigmentation.

So, no big surprises. I bring this up because of the increasing understanding of the role of vitamin D in health in general and especially in kids. For example, the American Association of Pediatrics just released a clinical report for clinicians entitled “Optimizing Bone Health in Children and Adolescents” (see here), which suggests that “Adequate vitamin D intake for infants younger than 1 year is 400 IU/d. The RDA is 600 IU for children 1 year and older”.​ And, since fortified cow’s milk has been shown in several studies to be the main dietary source of vitamin D in early childhood, it may be very important for us clinicians to be particularly attentive to kids drinking non-cow’s milk (which anecdotally I have found occasionally but increasingly in our inner-city community) as well as the quantity of fortified dairy products consumed. Of course, one concern in a Canadian study is the remarkable lack of good sunlight in those northern climes, making the issue there (and a lot of northern US) even more important. Another perhaps significant issue is that an old study found a remarkable discordance between the advertised supplementation on fortified milk and the actual content (see DOI: 10.1056/NEJM199204303261802​), with only 29% of 42 samples having between 80-120% of the advertise content.

Primary Care Corner with Geoffrey Modest MD: Vitamin D Screening Recommendations USPSTF

10 Dec, 14 | by EBM

By: Dr. Geoffrey Modest

The US Preventive Services Task Force just published their recommendation for vitamin D screening: “the current evidence is insufficient to assess the balance of benefits and harms of screening for vitamin D deficiency in asymptomatic adults”, which applies to community-dwelling, nonpregnant adults aged >18 yo seen in a primary care setting and without either signs/symptoms of vitamin D deficiency or conditions where vitamin D treatment is recommended (for review of recommendations, see here; for full systematic review, see either here or here). This recommendation is largely based on the fact that no study has directly evaluated clinical outcomes or harms comparing those screened for vitamin D deficiency or not. They note:

–There is no consensus about what the cutoff is for vitamin D deficiency. Some use <50 nmol/L (<20 ng/mL), others <75 nmol/L (<30 ng/mL), though they do note that some studies do show low levels are associated with increased risk for fractures, functional limitations, cancer (colorectal and others), diabetes, cardiovascular disease, cognitive decline, depression and death.

–There is no standard test used and there are some variations in results depending on the test done. One issue, for example, is that although African-Americans have lower 25(OH)D levels, they also have lower levels of vitamin D binding protein, therefore similar levels of bioavailable vitamin D)

–Data are lacking that there is clear benefit for screening of the general community-dwelling population. Current studies have found that there is no benefit in finding/treating vitamin D deficiency on cancer, diabetes, risk of death, risk of fractures in those not at high risk of fractures. Data are inadequate about other outcomes, such as psychosocial and physical functioning. Also, they did not find any studies on cardiovascular or immune diseases that met their inclusion criteria. They do note that there are some data that treatment of asymptomatic patients is adequate for a few limited outcomes. “Vitamin D treatment, with or without calcium, may be associated with decreased risk for mortality and falls in older or institutionalized adults”.

–Harms: evidence is pretty consistent that the harms of vitamin D supplementation are “small to none”, with no studies overall finding an increase in adverse events, hypercalcemia, kidney stones, or GI symptoms

–These recommendations only address screening for the general population, not those specific populations at high risk of falls, fractures, cardiovascular disease, or cancer without first determining 25(OH)-vitamin D levels.

A few points:

–This recommendation only addresses screening for vitamin D deficiency, not the need to have adequate vitamin D levels.

–This recommendation addresses only the general asymptomatic population, not those who are symptomatic or at high risk of vitamin D deficiency, and calls for “more research to determine vitamin D treatment effects in younger noninstitutionalized adults and to clarify the subpopulations that are most likely to benefit from treatment”.

–For example, these results may not apply to other populations: vitamin D treatment is associated with decreased risk of death (pooled RR 0.83, with CI 0.70-0.00) in older, institutionalized people, and is associated with lower rate of falls in institutionalized people

–There is pretty clear consensus that vitamin D is an important component of health, with the Institute of Medicine suggesting that adults aged 18-70 yo should take 600 IU of vitamin D and those >70yo should have 800 IU, and that this amount should be sufficient for 97.5% of the population. The US Endocrine Society clinical practice guideline on vitamin D deficiency (see doi: 10.1210/jc.2011-0385​) also suggested that vitamin D deficiency is remarkably common, and that the primary approach should be generalized supplementation, with testing for vitamin D deficiency in those at risk for deficiency and treating those deficient (<20 ng/mL). They recommend the same intake as the IOM, but comment that these levels may not be enough to raise 25(OH)D levels to 30 ng/mL and may not provide all of the nonskeletal health benefits (which are unproven).

–And I do think there is concern about the correct test to do — eg, the African-Americans who have low 25(OH)D​ but normal vitamin D binding protein levels, or that obese patients may have higher adipose tissue stores which may be bioavailable.

So, what does this all mean? Mostly that there is not enough evidence to screen the general population. It is pretty clear that vitamin D is important, and unlike other vitamins/minerals, is not naturally occurring in many foods, but is generated by UV light exposure. So, unlike other vitamins, where I tend to rant about eating healthfully instead of trying to supplement an individual ingredient here or there, vitamin D levels are an accident of one’s latitude. And there are vitamin D receptors throughout the body, including the immune system, with at least one article finding that those patients randomized to vitamin D supplementation who had active pulmonary tuberculosis had accelerated sputum conversion and improvement of several markers of TB-associated inflammatory responses — see here​ . So, perhaps it is correct that we should not be spending the money to check 25(OH)D levels in everyone and that we should just strongly recommend adherence to the supplementation guidelines of the Institute of Medicine —  with the only caveat being that there may be better adherence with vitamin D supplementation if the patient knows that they themselves are vitamin D deficient instead of just being told to drink more milk/orange juice/take supplements as one of many health different provider suggestions.

Primary Care Corner with Geoffrey Modest MD: Selenium and vitamin E may increase risk of prostate cancer

28 Feb, 14 | by EBM

New study in JNCI attempted to sort out diverse findings on prostate cancer development in men supplemented with selenium or vitamin D or both. (see DOI:10.1093/jnci/djt456). Previous data has shown:

–From the selenium and vitamin E cancer prevention trial (SELECT) in 2001, a prospective study of 35,533 men randomized to one of 4 groups (selenium 200 mcg/d plus vitamin E 400 IU/d, vitamin E plus placebo, selenium plus placebo, or just placebo), there was no benefit from these interventions, and a significantly increased risk of prostate cancer by 17% with vitamin E supplementation alone
–The nutritional prevention of cancer trial (NPC), a study of 928 men who lived in an area of the US with low soil selenium levels, found that men with moderate or low selenium levels and given supplementation with selenium had a decrease in prostate cancer risk by 75%.  This was not found in men who had high plasma selenium levels at baseline.

this current study was a cohort study comparing 1739 men with prostate cancer (489 with Gleason 7-10) from the SELECT trial with a randomly selected subcohort of 3117 men, also a from the SELECT trial.  They assessed whether selenium and vitamin E supplementation effect on prostate cancer was conditional on baseline selenium status.  Findings:

–Toenail selenium levels at baseline was not associated with prostate cancer risk
–Supplementation with selenium only or selenium plus vitamin E had no effect on prostate cancer in men with low baseline selenium levels in their toenails, but found a 91%  increased risk of high grade prostate cancer in men with a higher baseline selenium status.
–Vitamin E supplementation alone had no effect among men with high selenium status at baseline, the but had an increased risk of prostate cancer by 63% (total, low-grade, and high-grade) in men with lower selenium baseline status.
–When these investigators assessed the data using the cutpoints of the NPC trial above for baseline selenium concentrations, they could not reproduce any positive effect of supplementation. (though note that the NPC study was much smaller than the SELECT one).
–a couple of caveats to the current study: Toenail selenium concentrations are  not a perfect measure of actual selenium absorption and metabolism, and does not measure the functional activity of selenium-dependent proteins in prostate and other tissues.  Also there were very few black patients involved in the SELECT trial.

So, bottom-line, supplementation with selenium and vitamin E did not help, and in certain groups (e.g. selenium supplementation in those with high baseline selenium levels, and vitamin E supplementation alone in those with low selenium levels) was associated with increased prostate cancer risk.  We do know from many different studies that good nutrition is associated with lower levels of heart disease and cancer.  However, we repeatedly find in the studies that micronutrient supplementation often does not reproduce these results, and in some cases (e.g. beta-carotene and lung cancer in smokers) is harmful.  The issue here, as I pointed out in many prior blogs about micro-nutritional supplementation, is that there is likely an important interplay between the different nutritional components to foods as well as a balance of the amounts of each one, and there is been a long evolutionary history of humans consuming an array of appropriate vegetables, for example, which provide the appropriate mix.  The reductionist approach of trying to isolate specific micronutrients as a magic bullet to prevent disease and then supplementing with a supraphysiologic dose is therefore both conceptually problematic and typically clinically unsuccessful (with the possible exception to vitamin D, which is not largely nutritionally-derived, and deficiency may well create problems in those in untoward northern climates).



Primary Care Corner with Dr. Geoffrey Modest: Multivitamins. They still don’t work.

2 Jan, 14 | by EBM

3 articles on vitamin supplementation in dec 17th issue of annals of internal medicine. (all with same conclusions).

1. high dose vitamins and minerals post-MI (see doi:10.7326/0003-4819-159-12-201312170-00004). 1700 pts >50yo with MI at least 6 weeks earlier and serum creat <2.  Randomized to high dose vitamins/minerals, with composition designed by alternative med practitioners (vitamins significantly above “daily value” include: vitamins A, C, E, B6, niacin, pantothenic acid, manganese, selenium). results:

–mostly white males (82% male, 94% white) from Brigham and Duke hospitals

–unusually terrible adherence (which involved 3 pills bid), with 46% discontinuation rates in both the placebo and treatment arms

–nonsignificant difference in outcomes: 5-year event rates (total death, recurrent MI, stroke, coronary revasc, hosp for angina) in 34.2% on vitamins and 37% on placebo

2. vitamins and cognitive function (see  doi:10.7326/0003-4819-159-12-201312170-00006). 12 year study of 6000 men >65 yo in the Physicians’ Health Study II, randomized to multivitamins (Centrum Silver or placebo), with 4 repeated cognitive assessments. concept is that several vitamins may prevent oxidative damage to the brain (eg, vits C, E, b-carotene), are involved in development of neurotransmitters, DNA, and neuronal membranes (B vitamins), and are involved in neuronal survival and plasticity (vitamin A). overall  this study was of a pretty healthy group (96% nonsmokers, 60% with vigorous exercise at least 1/wk, 8% with diabetes, 53% with htn and 40% with hyperlipidemia). results:

–no diff in telephone assessment of cognitive function or in secondary outcome of verbal memory

but, this is a pretty select group of highly educated and baseline healthy people (and cognitive testing is less sensitive in those with more education), who likely ate well during their lives, and, if they didn’t, are receiving vitamin supplementation only late in their lives (negating a potential benefit since there is already a significant age-related decline by their age of >65yo at the start of the study).

3. a systematic review of the literature on vitamin and mineral supplementation in primary prevention of cardiovasc dz and cancer, reviewed for the US Preventive Service Task Force (see doi:10.7326/0003-4819-159-12-201312170-00729).  findings:

–1/2 of Americans use vitamin/mineral supplements, spending $11.8B/yr

–26 studies reviewed (5 assessed use of multivitamins, rest looked at individual vitamins or small combinations)

–for multivitamins: 2 very large studies predominated the analysis. no effect found for fatal or nonfatal cardiovasc events. one study (PHS-II, same study group as in the second article above) did find a protective effect in men (31% decreased risk), but not women. some studies with adverse events in vitamin group (melanomas in women, hip fracture rate in women). the investigators are hesitant to overgeneralize the cancer benefit in men, since it was only marginally significant and applied only to men (and there was no significant benefit on any specific subgroup of cancers, such as prostate cancer).

–single and  paired vits/minerals: no consistent protective effect for cardiovasc dz or cancer risk, inconsistent or contradictory results for calcium/vitamin D. other than b-carotene (assoc with higher incidence of lung cancer in smokers or those with asbestos exposure), there were differing adverse events in the different trials, without a consistent pattern.

major issue here is that the preponderance of evidence suggests that in people eating a “healthy” diet without evident micronutrient deficiency, there really is no indication at this time for vitamin supplementation (the data are still equivocal and awaiting large RCTs with regard to vitamin D, especially beyond bone and muscle development, decreasing falls in the elderly). the case of homocysteine provides an interesting case-in-point. hyperhomocysteinemia is associated with a variety of vitamin deficiencies (B12, folate, pyridoxine), hyperhomocysteinemia was pretty consistently found to be associated with adverse vascular events, those vitamins are really important ones for life, there are no known significant toxicities to them, and surrogate markers improved (not just decreasing the homocysteine levels, but studies also showing improved endothelial function, for example), BUT many different studies looking at real clinical outcomes surprisingly (at least to me) found no efficacy with vitamin supplementation.  the other issue was elaborated by walt willett at harv school of pub  health after the b-carotene supplementation study found the surprising increase in lung cancer in smokers (surprising in that b-carotene is also an important anti-oxidant which one would think might at least somewhat counter the toxicity of cigarettes, felt in part to be associated with their oxidant action). His position (with me taking significant liberties) was that there are dozens of naturally occurring carotenoids, and that it is fundamentally reductionist to try to purify a single ingredient of foods and then market it as beneficial: the reality is that there is a complex interplay of different vitamins, minerals, etc in naturally occurring healthy foods, and that we should eat these foods (eg fruits and vegetables) instead of looking for the magic pill…  although i must admit i do test and treat for vitamin b12 deficiency, persisting with the flawed logic i mentioned above:  B12 deficiency occurs relatively frequently in those over 50 years old (which is typically above the age that evolution acts…), is easily treated (in most cases orally), seems to be nontoxic in high doses, and is associated with significant clinical problems (esp anemia and neuropathy)…..   unfortunately, there are no long-term studies, starting with early finding of B12 deficiency through routine screening, and following cognitive decline (or neuropathy) over many years.

so, what is the bottom line? hard to be certain. at this point it makes sense not to suggest multivitamins for someone without a compelling medical or social reason which would predispose them to vitamin deficiency. but…. would they be useful if started at age 30 (before most people have significant atherosclerotic disease or cognitive decline)? are there subgroups of people who might benefit (should i give vitamin supplements to my kid who assiduously avoids eating anything green, like most vegetables)? these studies have not been done.


Primary Care Corner with Dr. Geoffrey Modest: vitamin B12 deficiency with acid suppressive therapy

12 Dec, 13 | by EBM

Although gastric acid is required to cleave vitamin B12 from ingested dietary proteins, and the same parietal cells that produce acid also produce intrinsic factor, studies have been mixed as to whether acid suppressive therapy leads to low vitamin B-12 levels. a large community-based case-controlled study was done by Kaiser Permanente North California comparing 26,000 patients with an incident diagnosis of vitamin B12 deficiency with 185,000 without (see doi:10.1001/jama.2013.280490).  Results:


–Vitamin B12 deficiency was more commonly diagnosed in patients with at least a 2 year supply of PPIs compared to nonusers (OR, 1.65).

–among people taking PPIs for at least 2 years, those on a higher dose (greater than 1.5 pills per day) had a higher OR of 1.95, as compared to those who were on lower doses of less than 0.75 pills per day, with OR of 1.63, and a statistically significant trend.

–a similar trend was found with people on H2 blockers, though there was no increased risk in those taking less than 0.75 pills per day.  The odds ratios were also significantly less than with PPIs.

–In addition, there was an increasing trend of vitamin B12 deficiency with longer duration of use of PPIs, but no trend with H2 blockers
–The strength of the association between PPI use and B12 deficiency decreased with discontinuation of the PPI

–The association between PPIs and B12 deficiency was strongest among those who were younger, especially those younger than 30 years of age.  The association was also stronger for women than men.  No differences by race/ethnicity.  No difference if diagnosis of GERD or not.


so, although the study was not a randomized controlled trial, it has the benefit of being very large, did find that the likelihood of B12 deficiency increased with higher potency of acid suppression (PPI>H2 blocker) and higher doses of the PPI, more likely with longer duration of taking the PPI, decreased on stopping PPI,  and was more profound than with other conditions known to be associated with B12 deficiency such as thyroid disease.  There there also is a reasonably plausible mechanism. This again reinforces that PPI’s are very strong drugs and should be used only when necessary — prior studies have shown that it is pretty uncommon for patients initially put on PPIs to be “stepped-down” to less aggressive therapies (H2 blockers or calcium tablets), reinforcing the “step-up” approach of starting with less potent and increasing potency as clinically needed (though i also do find that even in this case, over time some patients can be down-titrated)



EBM blog homepage

Evidence-Based Medicine blog

Analysis and discussion of developments in Evidence-Based Medicine Visit site

Creative Comms logo

Latest from Evidence-Based Medicine

Latest from EBM