Two identical unmarked pill bottles sit between a thick stack of clinical folders and a thinner stack on a dark study desk.
Supplements 8 min read

Centrum Silver’s Cancer Signal Weakened in a Larger, Shorter Trial

One trial of Centrum Silver found fewer cancer diagnoses in male doctors. A larger, shorter trial left the size and reliability of any benefit uncertain.

SUPPLEMENT NOTICE

The supplements discussed in this article are not intended to diagnose, treat, cure, or prevent any disease. Dosages mentioned reflect those used in specific research studies and should not be interpreted as recommendations. Always consult a healthcare professional before beginning any supplement regimen, especially if you have existing health conditions or take medications.

In 2012, researchers counting new cancers among nearly 15,000 male physicians reported 89 fewer diagnoses in the group assigned Centrum Silver than in the group assigned a look-alike placebo. A later trial with more than 21,000 participants found a raw difference of just 17 cases over a shorter period, leaving anyone eyeing the bottle for cancer protection with two estimates and considerable uncertainty. Physicians’ Health Study II · COSMOS randomized trial

The bigger trial may sound like the natural tiebreaker. Yet it enrolled a different population and observed participants for roughly a third as long. Both trials estimated slightly fewer cancer diagnoses with the multivitamin, and their confidence intervals overlapped. The apparent disagreement is chiefly about how much confidence to place in a modest effect and how far the results may travel beyond the people who generated them. Physicians’ Health Study II · COSMOS randomized trial

The doctors’ trial found a modest difference over 11 years

The Physicians’ Health Study II enrolled 14,641 U.S. male physicians aged 50 or older. Their average age at enrollment was 64, and 1,312 had a history of cancer. Participants received either Centrum Silver or a matching placebo as part of a blinded randomized trial. The investigators counted new cancers other than nonmelanoma skin cancer and analyzed participants according to their original assignment. That approach preserves the randomized comparison even when pill-taking habits change during follow-up. Physicians’ Health Study II

Over a median of 11.2 years, the investigators recorded 1,290 cases in the multivitamin group and 1,379 in the placebo group. The rates were 17.0 and 18.3 cases per 1,000 person-years, respectively. The denominator matters. One thousand person-years could represent 1,000 people observed for a year or 100 people observed for ten years. The rate difference describes the accumulated observation time in this study and gives a sense of scale alongside the headline figure of an 8% lower estimated hazard. Physicians’ Health Study II

The estimated hazard ratio was 0.92, with a 95% confidence interval of 0.86 to 0.998. A ratio of 1 would indicate equal estimated hazards. The upper bound was almost exactly there, so the data were compatible with a reduction near zero as well as a more meaningful one. The investigators also found no clear reduction in deaths from cancer. A study may count fewer diagnoses while still leaving open what happened to the cancers most likely to shorten life. Physicians’ Health Study II

COSMOS recruited more people and followed them for less time

COSMOS randomized 21,442 U.S. adults. Women could enroll from age 65 and men from age 60. Women made up about 59% of the study, and the eligibility criteria excluded people with recently diagnosed cancer or major cardiovascular disease. The study also independently assigned a cocoa extract or its placebo. That design let investigators compare everyone assigned the multivitamin with everyone assigned its placebo while distributing the cocoa assignment across both groups. COSMOS randomized trial

After a median of 3.6 years, 518 of 10,720 participants assigned the multivitamin and 535 of 10,722 assigned placebo had a confirmed invasive cancer. Those counts translate to about 48 versus 50 diagnoses per 1,000 people over the trial’s follow-up. The hazard ratio was 0.97, with a 95% confidence interval of 0.86 to 1.09. The data allowed for a modest reduction, essentially equal hazards, or a modest increase. The 17-case difference gives a sense of the observed gap, with the interval showing how imprecisely the study estimated any underlying benefit. COSMOS randomized trial

A study with more people can still record fewer events. COSMOS counted 1,053 invasive cancers, whereas the physicians’ trial counted 2,669 cancers under its primary definition. Follow-up matters for a second reason. Some cancers diagnosed in the first few years may already have been developing at enrollment, so a trial of that length may offer limited opportunity to see an effect on the earliest stages of disease. The studies did not randomly assign different follow-up periods, and their results cannot establish that taking the supplement for longer would yield a larger benefit. Physicians’ Health Study II · COSMOS randomized trial

The lung cancer result was more striking than the overall result

COSMOS reported an estimated lung cancer hazard ratio of 0.62, with a 95% confidence interval of 0.42 to 0.92. That represents a 38% lower estimated hazard for one cancer site, compared with a 3% lower estimate for all invasive cancers. The absolute difference in lung cancer was about two cases per 1,000 people. A large relative estimate can coexist with a modest absolute difference when the event is uncommon in the study population. COSMOS randomized trial · USPSTF evidence review

Lung cancer was one of several secondary outcomes. When researchers examine multiple cancer sites, chance has more opportunities to produce a result that stands out. The physicians’ trial also estimated fewer lung cancers in the multivitamin group, with uncertainty that included no effect. Pooling the two estimates in the evidence review yielded lower odds of lung cancer, but the result still rested on only those two trials. Both assignments were randomized. The open question is how reliably a finding for one cancer site would recur if it were the main outcome of another trial. Physicians’ Health Study II · USPSTF evidence review

Smoking history makes the lung result especially tempting to personalize. COSMOS estimated fewer total invasive cancers among participants who had ever smoked than among those who had never smoked when each group was compared with its own placebo group, but the evidence that the supplement effect differed by smoking history was inconclusive. The distinction is easy to miss. A subgroup can produce an interesting estimate even when the trial has too little information to show that the subgroup responds differently. COSMOS randomized trial

A bottle contains several hypotheses at once

One plausible explanation for any benefit is that a multivitamin could fill nutrient gaps that matter to processes such as DNA synthesis and repair. Testing that explanation would require knowing which participants entered the study with low levels of particular nutrients, whether the tablet corrected those levels, and whether the correction tracked with later cancer incidence. A trial that assigns the whole tablet can test its net effect. It has limited ability to tell whether one ingredient, several ingredients together, or differences in starting nutrient status account for an observed result. COSMOS randomized trial · USPSTF recommendation

The physicians’ trial offers a similar lesson about who may benefit. The estimate was more favorable in men who entered with a prior cancer diagnosis than in men who entered without one. The comparison between those subgroup effects remained uncertain, and the result concerned new cancers as defined by the study. A trial designed to evaluate recurrence after treatment would need to specify the original cancer, the treatment received and how recurrence would be confirmed. Those details could materially change what a supplement trial can tell a survivor. Physicians’ Health Study II

The product name also has limits as a scientific description. These trials evaluated the formulations supplied to participants during specific study periods. Evidence for those mixtures cannot establish what would happen with a different ingredient list, a single nutrient taken separately, or several overlapping products. The task force treated beta carotene and vitamin E supplements separately in its review and recommended against using either for cancer or cardiovascular prevention. The beta carotene evidence included increased lung cancer risk in trials involving people at elevated risk. The formulation and the population both matter when carrying a finding from one study to another. USPSTF evidence review · USPSTF recommendation

The pooled result leaves a clinical judgment open

The evidence review prepared for the U.S. Preventive Services Task Force pooled four randomized trials with 48,859 participants and estimated an odds ratio of 0.93 for any cancer, with a 95% confidence interval of 0.87 to 0.99. The absolute differences in the adequately powered trials ranged from about 0.2 to 1.2 percentage points in favor of supplementation. Pooling can provide a more precise average, while differences in the products, groups of volunteers and follow-up periods limit what that average may predict for one person or one formulation. USPSTF evidence review

The review also pooled cancer deaths. Its estimate was 0.94, with a confidence interval of 0.81 to 1.09. That wider interval leaves room for both fewer and more deaths. Counting cancer diagnoses and counting deaths answer different questions because the cancers included in the first count vary in severity and treatment outcomes. Longer follow-up and more events may help clarify whether any difference in diagnoses also shows up in outcomes such as advanced cancer or cancer mortality. USPSTF evidence review

The task force judged the evidence insufficient to determine the balance of benefits and harms of multivitamins for preventing cancer or cardiovascular disease in adults without known deficiencies. Its recommendation addresses routine use for those purposes. A documented deficiency raises a separate question with a different outcome, such as whether an intervention can restore an inadequate nutrient level. For cancer, the next informative trial would prespecify the outcome, characterize nutrient status at enrollment and follow participants long enough to measure diagnoses and deaths. That design could help show whether the modest average difference masks a benefit in a definable group or reflects a signal too uncertain to guide routine use. USPSTF recommendation

Frequently asked questions

Did both trials find fewer cancers with Centrum Silver?

Both counted fewer cases in the multivitamin group. Only the physicians’ trial found a statistically significant difference in its main cancer outcome, and that result narrowly cleared the conventional threshold.

Did Centrum Silver reduce lung cancer in COSMOS?

COSMOS estimated a lower lung cancer hazard with the multivitamin. Because lung cancer was one of several secondary outcomes, the result warrants confirmation in a trial designed around that outcome.

Did either trial show fewer deaths from cancer?

Neither trial established a reduction in cancer deaths. The pooled estimate also remained compatible with little or no difference.


This article summarises published research and is not medical advice. Consult your healthcare provider before changing your routine.

Frequently Asked Questions

Did both trials find fewer cancers with Centrum Silver?
Both counted fewer cases in the multivitamin group. Only the physicians’ trial found a statistically significant difference in its main cancer outcome, and that result narrowly cleared the conventional threshold.
Did Centrum Silver reduce lung cancer in COSMOS?
COSMOS estimated a lower lung cancer hazard with the multivitamin. Because lung cancer was one of several secondary outcomes, the result warrants confirmation in a trial designed around that outcome.
Did either trial show fewer deaths from cancer?
Neither trial established a reduction in cancer deaths. The pooled estimate also remained compatible with little or no difference.

Sources

  1. Physicians’ Health Study II randomized trial, JAMA(2026)
  2. COSMOS randomized trial, American Journal of Clinical Nutrition(2026)
  3. U.S. Preventive Services Task Force evidence review(2026)
  4. U.S. Preventive Services Task Force recommendation on vitamin supplementation(2026)
multivitamins cancer incidence randomized trials Centrum Silver COSMOS

Stay Updated on Longevity Science

Weekly research digests. No spam, unsubscribe anytime.

Subscribe

Related Articles