On 2014 May 14, David Keller commented:

Still waiting for anyone to answer my criticisms of this USPSTF report

When I first read the 2014 USPSTF update on vitamins for disease prevention, I expected, based on the headlines, to find evidence that there is no reason to take a multivitamin. Instead, I became convinced by the data presented by the USPSTF that the evidence of benefits versus harms favors men over 50 taking a multivitamin to prevent cancer and possibly reduce overall mortality. At the very least, the USPSTF would be fully justified in recommending that men over 50 who do not consume a diet rich in vegetables and fruits should consider adding a multivitamin. For some reason, most editorials and comments have completely ignored the significant reductions in cancer for men randomized to multivitamins in the 2 studies cited by USPSTF, and the significant reduction in overall mortality for men randomized to the high dose multivitamin tested in the French study. Instead, we saw headlines and editorials stating or implying that we now have proof that multivitamins are useless. I request that an expert in this area reply to my comments, giving good reasons why you are not convinced by the data we have, and what it would take to convince you. If you are knowledgeable in this area, and especially if you are a member of the USPSTF, I would greatly appreciate your pointing out where my thinking on this issue is wrong or even debatable. For details, data and references, please see my Open Letter to the USPSTF on the following PubMed Commons web page:

http://www.ncbi.nlm.nih.gov/pubmed/24566474#cm24566474_4093

Lastly, it is not helpful to tag a comment as "not helpful" without specifying why. PubMed Commons should foster meaningful debate, not merely anonymous unexplained contradiction of each other.

On 2014 Apr 23, David Keller commented:

Open letter to the USPSTF: the evidence shows multivitamins reduce cancers in men

The United States Preventative Services Task Force (USPSTF) recently updated their report on multivitamin supplements, and again found "insufficient evidence" to recommend their general use. The USPSTF report dismissed the significant reduction in cancers seen in men in both the SU.VI.MAX and Physicians' Health Study (PHS II), stating: "The lack of effect in women and the use of different supplement formulations in the 2 trials make extrapolating these findings to the general population difficult."(1) However, since men constitute approximately half of the general population, there is no need to extrapolate any benefit which men obtain from supplements to women. A benefit proven to accrue to men would justify a recommendation that men should take supplements, and the USPSTF could state that the issue was still unresolved for women. After all, the USPSTF makes other sex-specific recommendations, such as aortic ultrasound screening for male smokers but not females. Alternatively, the USPSTF may have doubted the biological plausibility of supplements preventing cancers in men but not in women. The SU.VI.MAX authors explained that difference by pointing out that women have a higher baseline nutritional status than men (2), and thus have less to gain by adding a multivitamin supplement. The USPSTF failed to mention or rebut that seemingly reasonable explanation.

The second reason USPSTF gave for not being able to interpret the results of the PHS II and SU.VI.MAX studies was that these two clinical trials tested 2 different multivitamin supplement formulations. However, the 5 antioxidants used in the SU.VI.MAX supplement are a subset of the micronutrients in the Centrum Silver administered in PHS II. I have pointed out in PubMed Commons (3) that there is evidence of a dose-response effect when one compares the effects of the Centrum Silver supplement versus the higher-dose SU.VI.MAX supplement, with regard to the significant reduction in cancer seen in men in both studies. Each of the 5 ingredients of the SU.VI.MAX supplement (vitamin C, vitamin E, beta carotene, selenium and zinc) is present in a higher dose than in Centrum Silver, yielding a correspondingly higher reduction in the risk of cancer in men (see Table 1). A dose-response effect tends to corroborate the findings of the individual studies, and it also suggests the need for a dose-ranging study of the SU.VI.MAX supplement. Would increasing the doses of these 5 antioxidant nutrients reduce cancer rates and mortality even further in men? Would a significant effect in women become evident?

Lastly, the updated USPSTF report completely omitted any mention of the significant reduction in all-cause mortality which benefitted the men taking the SU.VI.MAX supplement (2).

At this time, there is consistent evidence from 2 large, prospective, randomized, placebo-controlled trials that the low-dose multivitamin supplement used in PHS II significantly reduces the incidence of cancer in men, and that the higher-dose SU.VI.MAX supplement (consisting of 5 antioxidant nutrients in higher doses than in Centrum Silver) reduces cancer rates even more in men, and adds a significant reduction in all-cause mortality. The USPSTF report did not indicate how many more studies must replicate these results before they find the data persuasive. Until such trials are completed, the evidence we have indicates that men will continue to suffer cancers and deaths which are preventable by taking a multivitamin. If future studies fail to replicate these results (perhaps as a result of improved nutritional status among men) then the only known harm or cost would be the 6 cents per pill retail price of the Centrum multivitamin used in PHS II, or the inexpensive combination of supplements which duplicates the SU.VI.MAX formula.

The USPSTF should modify their assessment of multivitamin supplements to reflect the significant dose-related benefits multivitamins have demonstrated for men over the age of 50, who can benefit from reduced cancer rates and overall mortality, according to the best evidence we have. The USPSTF should call for dose-ranging studies to determine whether the benefits of the five antioxidants administered in SU.VI.MAX can be increased by increasing their doses. If the USPSTF has evidence or reasons other than those I have refuted above for not recommending multivitamins for men over age 50, they should make them public.

At the very least, the USPSTF report should be amended to include the fact that overall mortality was significantly decreased in men taking the SU.VI.MAX supplement. The absence of that data from their report was a disservice to men who are deciding whether or not to take a supplement.

Table 1: Dose-Response Effect? Vitamin doses versus relative risk of cancer:

Centrum Silver 50+............SU.VI.MAX multivitamin....................................................

Beta-Carotene 1000 IU.......Beta-Carotene 6mg = 9960 IU...............................................

Vitamin C 60 mg.................Vitamin C 120 mg..........................................................

Vitamin E 50 IU...................Vitamin E 30 mg = 67 IU ..................................................

Zinc 11 mg..........................Zinc 20 mg................................................................

Selenium 55 mcg..................Selenium 100 mcg..........................................................

RR of cancer = 0.93............RR of cancer = 0.69.......................................................

References

1: Moyer VA. Vitamin, Mineral, and Multivitamin Supplements for the Primary Prevention of Cardiovascular Disease and Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med. 2014 Feb 25. doi: 10.7326/M14-0198. [Epub ahead of print] PubMed PMID: 24566474.

2: Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briançon S. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med. 2004 Nov 22;164(21):2335-42. Erratum in: Arch Intern Med. 2005 Feb 14;165(3):286. PubMed PMID: 15557412.

3: Keller DL. Multivitamins: "Expensive Urine" or inexpensive cancer prevention? Comment posted on PubMed Commons, last edit dated 3/29/2014. http://www.ncbi.nlm.nih.gov/pubmed/24566474#cm24566474_3700

On 2014 Apr 19, David Keller commented:

Men taking multivitamins had significantly lower overall mortality and lower cancer risk

The USPSTF report did not disclose that men randomized to antioxidant supplements in the large prospective SU.VI.MAX trial had significantly lower risk of death, in addition to significantly fewer overall cancers than men assigned to placebo (1). This important information should be considered by any man over age 50 who is deciding whether or not to take multivitamins. In addition, please see my accompanying comments regarding the apparent dose-response effect of the SU.VI.MAX supplement with regard to cancer prevention in men.

(1) Hercberg S, Kesse-Guyot E, Druesne-Pecollo N, Touvier M, Favier A, Latino-Martel P, Briançon S, Galan P. Incidence of cancers, ischemic cardiovascular diseases and mortality during 5-year follow-up after stopping antioxidant vitamins and minerals supplements: a postintervention follow-up in the SU.VI.MAX Study. Int J Cancer. 2010 Oct 15;127(8):1875-81. doi: 10.1002/ijc.25201. PubMed PMID: 20104528.

On 2014 Mar 29, David Keller commented:

Multivitamins: "Expensive Urine" or inexpensive cancer prevention?

The USPSTF guideline statement on multivitamins and cancer risk (1) includes the following statements:

Statement 1: “Two large trials, the Physicians' Health Study II (PHS II) and the SU.VI.MAX (Supplementation in Vitamins and Mineral Antioxidants) study, showed a decrease in overall cancer incidence in men (pooled unadjusted relative risk, 0.93 [95% CI, 0.87 to 0.99])“

Statement 2: “Use of dietary supplements is common in the U.S. adult population. Forty-nine percent of adults used at least 1 dietary supplement between 2007 and 2010, and 32% reported using a multivitamin–multimineral supplement. Supplement use is more common among women and older adults than men and younger adults.”

Statement 3: “The lack of effect in women and the use of different supplement formulations in the 2 trials make extrapolating these findings to the general population difficult.”

The lack of benefit of multivitamins and mineral supplements (MVMS) in women might have been due to the higher background use of MVMS by women (Statement 2). Intention-to-treat analysis would count women in control groups who took MVMS in violation of experimental protocol as if they were not taking MVMS; this would tend to reduce the apparent benefit of MVMS in women, perhaps explaining Statement 3. A hypothesis-generating per-protocol analysis of these trials is warranted; if an anti-cancer effect of MVMS is thereby discerned in women, a more rigorous follow-up study would be justified.

As a male physician, I will continue to take a MVMS, based on Statement 1, unless evidence emerges which disproves the results of these 2 large trials. While awaiting further information, and considering the minimal potential harms and cost of multivitamins, and the possible benefits, I see no reason to dissuade women from taking a MVMS.

The USPSTF report also states that "the use of different supplement formulations in the 2 trials makes extrapolating these findings to the general population difficult", which refers to the fact that the Physician's Health Study tested "a commercially available multivitamin that contained 30 ingredients" (which was Centrum Silver), while the SU.VI.MAX Study (2) tested a supplement which "included nutritional doses of vitamins C and E plus β-carotene, selenium, and zinc".

In fact, the 5 nutrients included in the SU.VI.MAX supplement are a subset of the 30 nutrients included in Centrum Silver, and each of these 5 nutrients is present at a substantially higher dose in the SU.VI.MAX supplement than in Centrum Silver (3). Furthermore, use of the SU.VI.MAX supplement led to a larger reduction in the average relative risk of cancer than did the use of Centrum Silver, suggesting a possible dose-response effect for the ingredients of the SU.VI.MAX supplement with regard to lowering the relative risk of cancer. See Table 1:

Table 1: Dose-Response Effect? Vitamin doses versus relative risk of cancer:

Centrum Silver 50+............SU.VI.MAX multivitamin....................................................

Beta-Carotene 1000 IU.......Beta-Carotene 6mg = 9960 IU...............................................

Vitamin C 60 mg.................Vitamin C 120 mg..........................................................

Vitamin E 50 IU...................Vitamin E 30 mg = 67 IU ..................................................

Zinc 11 mg..........................Zinc 20 mg................................................................

Selenium 55 mcg..................Selenium 100 mcg..........................................................

RR of cancer = 0.93............RR of cancer = 0.69.......................................................

A dose-response effect, if present, would tend to support the hypothesis that the decrease in cancers observed in these 2 studies was real and not due to the play of chance. Proving a dose-response effect requires more than just two data points taken from 2 different studies on 2 different populations. However, this observed trend suggests the need for a follow-up study to determine whether further increases in the doses of the ingredients of the SU.VI.MAX supplement will lead to further declines in the average relative risk of cancer. Of course, all relevant cautions must be taken, such as not administering beta-carotene to persons with a smoking history.

References

1: Moyer VA. Vitamin, Mineral, and Multivitamin Supplements for the Primary Prevention of Cardiovascular Disease and Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med. 2014 Feb 25. doi: 10.7326/M14-0198. [Epub ahead of print] PubMed PMID: 24566474.

2: Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briançon S. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med. 2004 Nov 22;164(21):2335-42. Erratum in: Arch Intern Med. 2005 Feb 14;165(3):286. PubMed PMID: 15557412.

3: Centrum Silver 50+ website, accessed on 3/24/2014:<br> http://www.centrum.com/centrum-silver-adults-50-plus#tablets

4: The following websites were referenced for converting vitamin doses from mg to IU, accessed on 3/24/2014: http://ods.od.nih.gov/factsheets/VitaminE-HealthProfessional/

http://ods.od.nih.gov/factsheets/VitaminA-HealthProfessional/

http://dietarysupplementdatabase.usda.nih.gov/ingredient_calculator/help.php

On 2014 Mar 29, David Keller commented:

Multivitamins: "Expensive Urine" or inexpensive cancer prevention?

The USPSTF guideline statement on multivitamins and cancer risk (1) includes the following statements:

Statement 1: “Two large trials, the Physicians' Health Study II (PHS II) and the SU.VI.MAX (Supplementation in Vitamins and Mineral Antioxidants) study, showed a decrease in overall cancer incidence in men (pooled unadjusted relative risk, 0.93 [95% CI, 0.87 to 0.99])“

Statement 2: “Use of dietary supplements is common in the U.S. adult population. Forty-nine percent of adults used at least 1 dietary supplement between 2007 and 2010, and 32% reported using a multivitamin–multimineral supplement. Supplement use is more common among women and older adults than men and younger adults.”

Statement 3: “The lack of effect in women and the use of different supplement formulations in the 2 trials make extrapolating these findings to the general population difficult.”

The lack of benefit of multivitamins and mineral supplements (MVMS) in women might have been due to the higher background use of MVMS by women (Statement 2). Intention-to-treat analysis would count women in control groups who took MVMS in violation of experimental protocol as if they were not taking MVMS; this would tend to reduce the apparent benefit of MVMS in women, perhaps explaining Statement 3. A hypothesis-generating per-protocol analysis of these trials is warranted; if an anti-cancer effect of MVMS is thereby discerned in women, a more rigorous follow-up study would be justified.

As a male physician, I will continue to take a MVMS, based on Statement 1, unless evidence emerges which disproves the results of these 2 large trials. While awaiting further information, and considering the minimal potential harms and cost of multivitamins, and the possible benefits, I see no reason to dissuade women from taking a MVMS.

The USPSTF report also states that "the use of different supplement formulations in the 2 trials makes extrapolating these findings to the general population difficult", which refers to the fact that the Physician's Health Study tested "a commercially available multivitamin that contained 30 ingredients" (which was Centrum Silver), while the SU.VI.MAX Study (2) tested a supplement which "included nutritional doses of vitamins C and E plus β-carotene, selenium, and zinc".

In fact, the 5 nutrients included in the SU.VI.MAX supplement are a subset of the 30 nutrients included in Centrum Silver, and each of these 5 nutrients is present at a substantially higher dose in the SU.VI.MAX supplement than in Centrum Silver (3). Furthermore, use of the SU.VI.MAX supplement led to a larger reduction in the average relative risk of cancer than did the use of Centrum Silver, suggesting a possible dose-response effect for the ingredients of the SU.VI.MAX supplement with regard to lowering the relative risk of cancer. See Table 1:

Table 1: Dose-Response Effect? Vitamin doses versus relative risk of cancer:

Centrum Silver 50+............SU.VI.MAX multivitamin....................................................

Beta-Carotene 1000 IU.......Beta-Carotene 6mg = 9960 IU...............................................

Vitamin C 60 mg.................Vitamin C 120 mg..........................................................

Vitamin E 50 IU...................Vitamin E 30 mg = 67 IU ..................................................

Zinc 11 mg..........................Zinc 20 mg................................................................

Selenium 55 mcg..................Selenium 100 mcg..........................................................

RR of cancer = 0.93............RR of cancer = 0.69.......................................................

A dose-response effect, if present, would tend to support the hypothesis that the decrease in cancers observed in these 2 studies was real and not due to the play of chance. Proving a dose-response effect requires more than just two data points taken from 2 different studies on 2 different populations. However, this observed trend suggests the need for a follow-up study to determine whether further increases in the doses of the ingredients of the SU.VI.MAX supplement will lead to further declines in the average relative risk of cancer. Of course, all relevant cautions must be taken, such as not administering beta-carotene to persons with a smoking history.

References

1: Moyer VA. Vitamin, Mineral, and Multivitamin Supplements for the Primary Prevention of Cardiovascular Disease and Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med. 2014 Feb 25. doi: 10.7326/M14-0198. [Epub ahead of print] PubMed PMID: 24566474.

2: Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briançon S. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med. 2004 Nov 22;164(21):2335-42. Erratum in: Arch Intern Med. 2005 Feb 14;165(3):286. PubMed PMID: 15557412.

3: Centrum Silver 50+ website, accessed on 3/24/2014:<br> http://www.centrum.com/centrum-silver-adults-50-plus#tablets

4: The following websites were referenced for converting vitamin doses from mg to IU, accessed on 3/24/2014: http://ods.od.nih.gov/factsheets/VitaminE-HealthProfessional/

http://ods.od.nih.gov/factsheets/VitaminA-HealthProfessional/

http://dietarysupplementdatabase.usda.nih.gov/ingredient_calculator/help.php

On 2014 Apr 19, David Keller commented:

Men taking multivitamins had significantly lower overall mortality and lower cancer risk

The USPSTF report did not disclose that men randomized to antioxidant supplements in the large prospective SU.VI.MAX trial had significantly lower risk of death, in addition to significantly fewer overall cancers than men assigned to placebo (1). This important information should be considered by any man over age 50 who is deciding whether or not to take multivitamins. In addition, please see my accompanying comments regarding the apparent dose-response effect of the SU.VI.MAX supplement with regard to cancer prevention in men.

(1) Hercberg S, Kesse-Guyot E, Druesne-Pecollo N, Touvier M, Favier A, Latino-Martel P, Briançon S, Galan P. Incidence of cancers, ischemic cardiovascular diseases and mortality during 5-year follow-up after stopping antioxidant vitamins and minerals supplements: a postintervention follow-up in the SU.VI.MAX Study. Int J Cancer. 2010 Oct 15;127(8):1875-81. doi: 10.1002/ijc.25201. PubMed PMID: 20104528.

On 2014 Apr 23, David Keller commented:

Open letter to the USPSTF: the evidence shows multivitamins reduce cancers in men

The United States Preventative Services Task Force (USPSTF) recently updated their report on multivitamin supplements, and again found "insufficient evidence" to recommend their general use. The USPSTF report dismissed the significant reduction in cancers seen in men in both the SU.VI.MAX and Physicians' Health Study (PHS II), stating: "The lack of effect in women and the use of different supplement formulations in the 2 trials make extrapolating these findings to the general population difficult."(1) However, since men constitute approximately half of the general population, there is no need to extrapolate any benefit which men obtain from supplements to women. A benefit proven to accrue to men would justify a recommendation that men should take supplements, and the USPSTF could state that the issue was still unresolved for women. After all, the USPSTF makes other sex-specific recommendations, such as aortic ultrasound screening for male smokers but not females. Alternatively, the USPSTF may have doubted the biological plausibility of supplements preventing cancers in men but not in women. The SU.VI.MAX authors explained that difference by pointing out that women have a higher baseline nutritional status than men (2), and thus have less to gain by adding a multivitamin supplement. The USPSTF failed to mention or rebut that seemingly reasonable explanation.

The second reason USPSTF gave for not being able to interpret the results of the PHS II and SU.VI.MAX studies was that these two clinical trials tested 2 different multivitamin supplement formulations. However, the 5 antioxidants used in the SU.VI.MAX supplement are a subset of the micronutrients in the Centrum Silver administered in PHS II. I have pointed out in PubMed Commons (3) that there is evidence of a dose-response effect when one compares the effects of the Centrum Silver supplement versus the higher-dose SU.VI.MAX supplement, with regard to the significant reduction in cancer seen in men in both studies. Each of the 5 ingredients of the SU.VI.MAX supplement (vitamin C, vitamin E, beta carotene, selenium and zinc) is present in a higher dose than in Centrum Silver, yielding a correspondingly higher reduction in the risk of cancer in men (see Table 1). A dose-response effect tends to corroborate the findings of the individual studies, and it also suggests the need for a dose-ranging study of the SU.VI.MAX supplement. Would increasing the doses of these 5 antioxidant nutrients reduce cancer rates and mortality even further in men? Would a significant effect in women become evident?

Lastly, the updated USPSTF report completely omitted any mention of the significant reduction in all-cause mortality which benefitted the men taking the SU.VI.MAX supplement (2).

At this time, there is consistent evidence from 2 large, prospective, randomized, placebo-controlled trials that the low-dose multivitamin supplement used in PHS II significantly reduces the incidence of cancer in men, and that the higher-dose SU.VI.MAX supplement (consisting of 5 antioxidant nutrients in higher doses than in Centrum Silver) reduces cancer rates even more in men, and adds a significant reduction in all-cause mortality. The USPSTF report did not indicate how many more studies must replicate these results before they find the data persuasive. Until such trials are completed, the evidence we have indicates that men will continue to suffer cancers and deaths which are preventable by taking a multivitamin. If future studies fail to replicate these results (perhaps as a result of improved nutritional status among men) then the only known harm or cost would be the 6 cents per pill retail price of the Centrum multivitamin used in PHS II, or the inexpensive combination of supplements which duplicates the SU.VI.MAX formula.

The USPSTF should modify their assessment of multivitamin supplements to reflect the significant dose-related benefits multivitamins have demonstrated for men over the age of 50, who can benefit from reduced cancer rates and overall mortality, according to the best evidence we have. The USPSTF should call for dose-ranging studies to determine whether the benefits of the five antioxidants administered in SU.VI.MAX can be increased by increasing their doses. If the USPSTF has evidence or reasons other than those I have refuted above for not recommending multivitamins for men over age 50, they should make them public.

At the very least, the USPSTF report should be amended to include the fact that overall mortality was significantly decreased in men taking the SU.VI.MAX supplement. The absence of that data from their report was a disservice to men who are deciding whether or not to take a supplement.

Table 1: Dose-Response Effect? Vitamin doses versus relative risk of cancer:

Centrum Silver 50+............SU.VI.MAX multivitamin....................................................

Beta-Carotene 1000 IU.......Beta-Carotene 6mg = 9960 IU...............................................

Vitamin C 60 mg.................Vitamin C 120 mg..........................................................

Vitamin E 50 IU...................Vitamin E 30 mg = 67 IU ..................................................

Zinc 11 mg..........................Zinc 20 mg................................................................

Selenium 55 mcg..................Selenium 100 mcg..........................................................

RR of cancer = 0.93............RR of cancer = 0.69.......................................................

References

1: Moyer VA. Vitamin, Mineral, and Multivitamin Supplements for the Primary Prevention of Cardiovascular Disease and Cancer: U.S. Preventive Services Task Force Recommendation Statement. Ann Intern Med. 2014 Feb 25. doi: 10.7326/M14-0198. [Epub ahead of print] PubMed PMID: 24566474.

2: Hercberg S, Galan P, Preziosi P, Bertrais S, Mennen L, Malvy D, Roussel AM, Favier A, Briançon S. The SU.VI.MAX Study: a randomized, placebo-controlled trial of the health effects of antioxidant vitamins and minerals. Arch Intern Med. 2004 Nov 22;164(21):2335-42. Erratum in: Arch Intern Med. 2005 Feb 14;165(3):286. PubMed PMID: 15557412.

3: Keller DL. Multivitamins: "Expensive Urine" or inexpensive cancer prevention? Comment posted on PubMed Commons, last edit dated 3/29/2014. http://www.ncbi.nlm.nih.gov/pubmed/24566474#cm24566474_3700