Junkfood Science: One more time: fatness not linked to overall cancer risks

November 18, 2007

One more time: fatness not linked to overall cancer risks

Readers may have overdosed on the fat and cancer stories over recent weeks, so we’ll make this brief. Despite attempts to make things sound confusing or complicated, and convince us that the science is contradictory concerning a link between fatness and cancers, all of the major, long-awaited papers that have been released this month have found the same thing: nothing.

There is no support for blaming fat people for getting cancer or for using cancer to try and scare them thin.

We’ve looked at the Second Expert Report by the World Cancer Research Fund and American Institute for Cancer Research, which reviewed 2,471 studies on 17 cancers and the data found no tenable associations between cancers and foods or body fatness, nor between cancer deaths and BMIs.

That was followed days later by the study from scientists at the National Center for Health Statistics at the Center for Disease Control and Prevention which reviewed 40 years of NHANES and vital statistics data. It also found no tenable or statisticaly significant associations between overall cancer mortality and any BMI category.

There was a third study we haven’t yet mentioned: the Million Women Study, published online at the British Medical Journal on November 6th. This epidemiological study also examined relative risks for 17 cancers and mortality associated with BMI. The researchers culled through data on 1.22 million women, aged 50-64, recruited between 1996-2001 to the Million Women Study. The women had self-reported their height and weight at enrollment and the researchers only had updated information on about one-third of the women, unlike the CDC study. It was also a shorter study. Cancer incidences and deaths among these women were followed for only 5.4 and 7 years, respectively. The researchers adjusted for known confounding factors, including age, geographical region, socioeconomic status, reproductive history, smoking status, alcohol intake, physical activity and, as appropriate, time since menopause and use of hormone replacement therapy. They also did a secondary analysis to rule out reverse causation (low weight preceeding clinical diagnosis of cancer) by elminating the first two years of follow-up and their findings were unaffected.

Their statistical findings? They found no tenable associations between any BMI category and incidences of all cancers — relative risks were 1.12 overall among those with BMIs >30, compared to relative risks of 1.00 among BMIs of 22.5-24.9. Some individual cancers were on one side of the line and some on the other, but they were all hugging null and none exceeded chance, random coincidence or statistical error.

Looking at cancer deaths, they reported only a negligible 0.06 difference in relative risks between the thin and fattest women, with 7,812 deaths among those with BMIs<25 and 3,436 deaths among those with BMIs >30.

Do you want the splitting hairs breakdown? While the press may try to scare us into believing these were meaningful correlations, we know better. But, thin women with BMIs <22.5 (within ‘normal’) were associated with higher risks for squamous cell carcinoma of the esophagus, stomach, colorectal, lung, melanoma, premenopausal breast and brain cancers. The fattest women with BMIs >30 were associated with higher risks for adenocarcinoma of the esophagus, pancreas, postmenopausal breast, cervix, endometrium, ovary, kidney, bladder, non-Hodgkins lymphoma, and leukemia. The only cancers they reported relative risks just over 2 associated with obesity were endometrial and adenocarcinoma of the esophagus — which are rarer cancers than more common cancers where obesity was associated significantly lower risks: lung and squamous cell carcinoma of the esophagus.

It’s undertandable that consumers may have gotten a different impression of this study, however, as it was reported as finding: “Increasing body mass index is associated with a significant increase in the risk of cancer for 10 out of 17 specific types examined.” Technically true, but as with everything, the devil is in those details.

© 2007 Sandy Szwarc

Bookmark and Share