Read Original Release

News release hypes weak association between magnesium and pancreatic cancer

IU researchers find magnesium intake may be beneficial in preventing pancreatic cancer

Our Review Summary

iStock_000044962300_SmallThis news release from Indiana University summarizes findings from a large database of health data gathered through a public survey of vitamin and supplement use — the VITamins and Lifestyle study. The aim of the study is to help ascertain whether supplements have any impact on reducing cancer risk. This news release focuses on the main findings which are overstated in the British Journal of Cancer article (shame on the editors of that journal for not asking the authors to state the limitations of the study more clearly). There are many limitations of this cohort study, including that there were relatively few cancer cases overall (151), that there were substantial baseline differences in smoking and alcohol use according to magnesium intake (two of the most clear risk factors for pancreatic cancer), and that when the statistical analysis excludes cases in the first 2 years of the study (those who may have already had a cancer growing), the findings became statistically non-significant. Moreover, the difference in absolute risk between groups was tiny. Somewhere along the line of this study’s journey from manuscript to news release, someone should have taken it upon themselves to scrutinize the findings more carefully and present a balanced report on the results.


Why This Matters

Pancreatic cancer represents just 3% of all new cancer cases in the United States (or about 49,000 new cases annually) but it has one of the worst outcomes. Only about 7.2% of those who develop the cancer survive longer than 5 years after diagnosis, according to the National Cancer Institute. Emerging evidence of any potential preventative measures people can take to reduce their risk is newsworthy and should be discussed, but with the understanding that the recommendations are speculative. It would be wonderful if a supplement prevented cancer, but it probably won’t.


Does the news release adequately discuss the costs of the intervention?

Not Satisfactory

No cost information is provided but there are cost implications to following the recommendations. The release concludes that everyone “should strive to get the daily recommendations of magnesium through diet” and “adding a magnesium supplement to their diet may prove beneficial in preventing this disease.”

Does the news release adequately quantify the benefits of the treatment/test/product/procedure?

Not Satisfactory

There’s very little quantification offered in the release. The release states that the 66,000 people who missed the recommended daily intake for magnesium by 100 mg were at a 24% increased risk for developing pancreatic cancer. Then it makes a strong recommendation to increase magnesium through diet and supplements.

The release would have been much stronger had it included measures of absolute risk. Unfortunately, the abstract and paper both make it difficult to calculate the absolute numbers. From Table 2, we can see that 44/14395 (0.30%) people in the cohort who were in the <75% of the RDA group developed pancreatic cancer, compared with 64/35348 (0.18%)people in the >100% of RDA group.  The absolute risk difference is 0.12%.  That is extremely small.  Giving only relative reductions is very, very misleading.  And, when the appropriate exclusions are done (the people who had cancers diagnosed in the first two years of the study), the results are not statistically significant anyway.

Does the news release adequately explain/quantify the harms of the intervention?

Not Satisfactory

The release encourages people to add magnesium through their diets and through supplementation but doesn’t tell them what the recommended daily allowance is. In the United States, the current magnesium recommended intakes are 400-420 mg/day for men and 310-320 mg/day for women.

Exceeding the recommended levels of magnesium through supplementation can cause side effects. Although these are not likely to be life-threatening, excessive intake of supplemental magnesium is harmful to  individuals with impaired kidney functions. In addition, magnesium can interfere with certain medications so patients should always consult with their doctor before adding a supplement.

From WebMD: “Doses less than 350 mg per day are safe for most adults. When taken in very large amounts, magnesium is POSSIBLY UNSAFE. Large doses might cause too much magnesium to build up in the body, causing serious side effects including an irregular heartbeat, low blood pressure, confusion, slowed breathing, coma, and death.”

Does the news release seem to grasp the quality of the evidence?

Not Satisfactory

The study is longitudinal in that data is collected and then tracked over several years. It doesn’t show causation but a trend. One of the major drawbacks of studies based on questionnaires, including this one, is that it relies on individuals’ ability to accurately recall and record their own data, which can lead to lots of inaccuracy. In this case, participants were asked to describe their supplement use for the previous 10 years. Moreover, this observational study shows a very weak association only.  One of the main shortcomings is likely to be that the baseline differences observed in the groups (low, medium, high magnesium use) , although accounted for in the regression analyses, are probably still driving the findings.  This problem of residual confounding (not being able to fully account statistically for differences in the people according to what their nutrient intake was) is common in this type of study.  The authors of the paper should have spent some time in their discussion on this, but they did not. The news release authors also would have done well to put more emphasis on limitations.

Does the news release commit disease-mongering?

Not Satisfactory

The release calls pancreatic cancer “the fourth leading cause of cancer-related death in both men and women in the United States.”  But this statement should have been accompanied by some statistics that demonstrate the relative rarity of the disease.  According to the National Cancer Institute, the lifetime risk of developing pancreatic cancer is approximately 1.5 percent.

Does the news release identify funding sources & disclose conflicts of interest?

Not Satisfactory

Funders of the study aren’t named in the release. According to the published findings, the research was supported by grants from the National Cancer Institute and the National Institute of Health Office of Dietary Supplements.

Does the news release compare the new approach with existing alternatives?

Not Satisfactory

The focus of this release is on a study that tries to compare the effect of meeting RDAs of magnesium and not meeting that daily target.

But the release doesn’t mention any other suspected risk factors for pancreatic cancer such as obesity, smoking, alcohol abuse, inactivity, genetics, and so on, which might have been helpful.

Does the news release establish the availability of the treatment/test/product/procedure?


The release briefly mentions dietary magnesium can be found in “dark, leafy greens or nuts” but a longer list of dietary sources might have been helpful. It’s generally understood that magnesium supplements are widely available on any grocery or drug store shelf.

Does the news release establish the true novelty of the approach?


The release acknowledges that previous work has examined the relationship between magnesium intake and risk for pancreatic cancer.

Previous studies have found that magnesium is inversely associated with the risk of diabetes, which is a risk factor of pancreatic cancer. But few studies have explored the direct association of magnesium with pancreatic cancer; of those that did, their findings were inconclusive, said Daniel Dibaba, a Ph.D. student at the School of Public Health-Bloomington, who led the IU study.

Of course, the implication seems to be that while other studies were “inconclusive,” this study has put the matter to rest. That’s certainly not the case, but we’ll give the benefit of the doubt since the release later states that more study is needed.

Does the news release include unjustifiable, sensational language, including in the quotes of researchers?


The release doesn’t come right out and say that taking magnesium will reduce pancreatic cancer risk but it comes close. Instead, it hedges on making that claim with phrases like “may be beneficial” and “may prove beneficial.” Based on limited available knowledge, it’s appropriate to share the information that getting the recommended daily allowance of magnesium may have specific health benefits.

Total Score: 3 of 10 Satisfactory

Comments (3)

Please note, comments are no longer published through this website. All previously made comments are still archived and available for viewing through select posts.


December 28, 2015 at 8:34 pm

Do you just think the authors were not asked some of the questions that you mention? I bet they were asked and could convincingly answer them. Most of the criteria you used to evaluate the paper are irrelevant to the study. Rather than using a hard and fixed criteria for all studies, you should have specific and relevant criteria to evaluate the study and the press release based on the study type and objective of the study. As someone who closely follows similar studies, your criteria are irrelevant to the study and shallow.


    Gary Schwitzer

    December 29, 2015 at 7:32 am


    Somehow you have missed the point that we review media messages on this site.

    Our criteria are not intended for the review of studies, as you assert. They are intended solely for the review of media messages – such as news releases. Such media messages are frequently about studies, but our purpose is to try to improve the accuracy, balance and completeness of news stories and news releases that include claims about health care interventions.

    Gary Schwitzer


    Kevin Lomangino

    December 29, 2015 at 7:40 am


    Beyond what Gary said about the purpose of our site, I’d say that the points raised in the review are very much relevant to the study itself. The weakness of the association, the small absolute difference in risk, and likelihood of residual confounding are all issues that should have been addressed in the paper.

    Kevin Lomangino
    Managing Editor