October 3, 2018

expert reaction to study looking at processed meat and breast cancer

Research published in the International Journal of Cancer provides evidence that higher consumption of processed meat, but not red meat, is associated with higher risk of breast cancer.

Dr Gunter Kuhnle, Associate Professor in Nutrition and Health, University of Reading, said:

“This study seems quite robust, but of course relies on the data of the original studies.  The study reviews the evidence currently available for a possible link between processed meat intake and breast cancer risk.  Processed meat – for example sausages or bacon – has already been classified as ‘carcinogenic to humans’ by the WHO based on overwhelming evidence, but has been mainly linked to an increased risk in bowel cancer.  There are different theories about the underlying mechanism, and it is likely that they all have a role to play: the curing salt in many processed meats, e.g. bacon, can react with the protein in meat and form carcinogenic nitrosamines, or haem in the meat can cause direct damage to colon cells.  It is however important to note that while the evidence for classifying processed meat as carcinogen is strong, the actual risk to the individual is very small and it is more relevant on a population level (David Spiegelhalter explained this well at the time as follows: 6 out of 100 people in the UK will get bowel cancer in their lives – it will be 7 out of 100 if they eat 50 g of processed meat daily).

“This new study shows that there is a risk increase for breast cancer, but it is even smaller than for bowel cancer – a 10% difference between those consuming the highest amount compared to those consuming the lowest amount.  Whether this justifies a change in the current recommendation of 70 g of red and processed meat is questionable, especially as it is not clear how much processed meat the high consumer group consumed.

“However, it is important to note that breast and bowel cancer are the most common types of cancer in the UK with almost 100,000 cases per year, and even small changes to the disease risk will have a huge impact.  It is therefore important to follow up on these findings and investigate whether the risk of cancer associated with processed meat consumption could be reduced, e.g. by developing new food production methods.  Changes to the diet might also be helpful – for example if curing salt has an important role in the development of cancer, most British sausages or Parma ham would be fine to eat.”

Prof Kevin McConway, Emeritus Professor of Applied Statistics, The Open University, said:

“This is generally a competent piece of research, but it’s really not easy to interpret its findings.  Given the inconsistent results from previous studies, it makes sense to see what happens when they are looked at together.  But there are several difficulties.  We can’t really know from this work whether eating particular types of meat actually causes an increased risk of breast cancer.  Also, whether the observed relationship is one of cause and effect or not, it’s pretty well impossible to tell from these results how much smaller the risk of breast cancer would be if consumption of processed meat were lower.  The new study has moved knowledge forward to some extent, but inevitably it leaves many questions unanswered.

“The studies that were included in this new analysis were almost all observational – that is, the researchers in those studies observed what people chose to eat, and followed them up and recorded which of them developed breast cancer.  In studies like that, a major problem is that people who choose to eat different amounts of meat are very likely to differ in other ways too, and an apparent relationship between meat eating and breast cancer might be because of these other differences rather than because eating some types of meat actually causes breast cancer to be more likely.  Researchers can adjust their comparisons statistically to take account of these other differences – ‘confounders’ in the jargon – but they can do this only for factors that they have data on, and these adjustments can’t completely establish what causes what.  So since this new study is based on studies that couldn’t say whether meat eating actually causes breast cancer to be more likely, the new study can’t establish that either.

“The included studies were pretty diverse – different countries, different populations, and different statistical approaches.  They adjusted for often quite different sets of confounders, and they used different amounts to categorize the meat consumption.  The new research was bound to find it awkward to deal with this diversity.  In fact the data they put together was the comparison, for each study, between the highest and lowest categories of meat consumption in that study, whatever those categories were, and based on the statistical adjustments made in that study, whatever they were.  If one of the studies had used different consumption categories, its results would have been different, and so the result from pooling all these studies could also have been different.  So I just can’t tell you how many more cases of breast cancer there would be if everyone ate an extra bacon sandwich a day – this research just can’t give that information.  The researchers found statistically significant evidence of heterogeneity in the studies that they analysed, indicating that the relationships between meat consumption and breast cancer were different in different studies.  That’s hardly surprising given the differences between the individual studies, but it means that the 9% figure for the difference in breast cancer risk between the highest and lowest categories of processed meat consumption is an average over studies that differ in potentially important ways.  The 9% figure could be considerably different (in either direction) if the individual studies had used different meat consumption categories, had adjusted for different confounders, or had involved different populations in different places.  The researchers did carry out investigations to try to establish some specific reasons for this heterogeneity, but they appear not to have found such reasons.  So it’s hard to say what the 9% increase in risk for higher amounts of processed meat actually means in detail.

“It’s also awkward to interpret the fact that there was no statistically significant relationship between unprocessed red meat consumption and breast cancer risk.  That relationship is also affected by the heterogeneity.  To say it’s not statistically significant means that the data are consistent with there being no real relationship.  But they are also consistent with there being a relationship of the same size as for processed meat, and indeed with many other possibilities.  The difference between the sizes of the relationships for unprocessed red meat and processed meat is not, itself, statistically significant.  All that really demonstrates, perhaps, is the difficulty of getting any very definite information out of such a diverse set of individual studies.”

‘Consumption of red and processed meat and breast cancer incidence: A systematic review and meta-analysis of prospective studies’ by Maryam S. Farvid et al. was published in the International Journal of Cancer on Tuesday 2 October 2018. 

Declared interests

Dr Gunter Kuhnle: “I have conducted research into the association between red/processed meat and cancer for more than a decade and was involved in an EU funded research project to investigate new food production methods to reduce cancer risk.”

Prof Kevin McConway: “I am a member of the SMC advisory committee.”