Role of diet in type 2 diabetes incidence: umbrella review of meta-analyses of prospective observational studiesBMJ 2019; 366 doi: https://doi.org/10.1136/bmj.l2368 (Published 03 July 2019) Cite this as: BMJ 2019;366:l2368
All rapid responses
In 2015, 30.3 million Americans, or 9.4% of the population, had diabetes. This study suggests that if everyone in the USA ate two extra slices of bacon a day, 18.8% of the population would be diagnosed with diabetes.
The evidence for this rests in the meta-analysis Misha 2010. This combines 5 studies. 4 of these are from the USA, one is from China. In the Chinese study, "we found that salted/preserved meat consumption (yes/no) was associated with a higher risk of type 2 DM (RR 1.16; 95%CI: 1.04-1.31, P<0.01), while consumption of bacon/smoked meat or Chinese sausage was unrelated to the risk of type 2 DM". This is not a measurement by slices.
The same authors, Hu and Willett, appear in 3 of the 4 US studies. These studies (NHS, HPFS, NHS 2) are entirely populated by health professionals, so that bacon eating would be expected to be a sign of low conscientiousness and associated with other behaviours more reliably linked to risk, such as inactivity or alcohol consumption.
The one study not in health professionals, WHS, does not define bacon consumption in terms of slices. (1.21, 1.06–1.39 for ≥2/week vs. <1/week, P = 0.004 for trend). In fact none of the studies used a "slices" measure or calculation of portion size, all seem to have used crude measures of frequency.
Bryan criticised the Misha 2010 paper's CVD findings, stating in part "Furthermore, the Micha analysis extrapolates beyond the dietary intake ranges described in the Sinha study." Much the same thing seems to have happened with regard to bacon and the risk of type 2 diabetes in Misha 2010. The "2 slices of bacon per day" figure can only be described as an imaginative leap.
The current paper takes the Misha calculations on trust and rates that evidence as "high", using the AMSTAR system for grading meta-analyses, rather than the GRADE system (which rates it as "low- to very low-certainty evidence" for the combined processed meat category).
But really, the only proper way to grade a meta-analysis involves take the time to read the papers it contains. Epidemiological studies into diet and disease, and meta-analyses of them, are often designed by the enthusiasts for a particular point of view, and carried out in the very populations whose beliefs they have influenced. This is certainly the case with Misha 2010. It seems to be less the case with the recent GRADE analysis of the same question.
 Micha R, Wallace SK, Mozaffarian D. Red and processed meat consumption and risk of incident coronary heart disease, stroke, and diabetes mellitus: a systematic review and meta-analysis. Circulation. 2010; 121:2271–2283
 Villegas R, Shu XO, Gao YT, et al. The association of meat intake and the risk of type 2 diabetes may be modified by body weight. Int J Med Sci. 2006;3(4):152–159. Published 2006 Oct 27. doi:10.7150/ijms.3.152
 Bryan, N. Letter by Bryan Regarding Article, "Red and Processed Meat Consumption and Risk of Incident Coronary Heart Disease, Stroke, and Diabetes Mellitus: A Systematic Review and Meta-Analysis" Circulation 2011. 10.1161/CIRCULATIONAHA.110.967604
 Johnston BC, Zeraatkar D, Han MA, et al. Unprocessed Red Meat and Processed Meat Consumption: Dietary Guideline Recommendations From the Nutritional Recommendations (NutriRECS) Consortium. Ann Intern Med. 2019;171:756–764. [Epub ahead of print 1 October 2019]. doi: https://doi.org/10.7326/M19-1621
Competing interests: No competing interests
Re: Role of diet in type 2 diabetes incidence: umbrella review of meta-analyses of prospective observational studies
Neuenschwander et al.1, in their umbrella review of systematic reviews and meta-analyses (SRMAs), have added to the substantial body of moderate and high quality evidence implicating red meat and especially processed red meats in the aetiology of type 2 diabetes. More generally, they have confirmed significant evidence of diets richer in plant protein v. animal protein being protective.
However, when diets are characterized by their content of animal protein v plant protein, they gloss over the fact that animal protein almost always refers to muscle meats, since this is the part of the animal protein mass that is generally consumed. Therefore it adds little to consider the difference between “unprocessed” and “processed meats”, in terms of the addition of salts and preservatives, when the most consequential bit of processing is cutting the muscle off the bones and discarding the bones. This is because muscle meats proteins are comprised of very different amino acids than are bones and connective tissues. Specifically, muscle meats are relatively methionine-rich and glycine-poor, whereas the bones’ protein content (mostly collagen) is glycine-rich and methionine-poor. This has generally been overlooked owing to glycine’s being viewed as a non-essential amino acid, and attention has always been focused on the 8 essential amino acids (e.g., methionine), as the determinants of protein’s nutritional quality. But when (muscle) meat consumption is high, the intake of methionine far surpasses biochemical needs, and the liver gets rid of most of it, via the enzyme glycine-N-methyltransferase pathway2. Importantly, the clearance of each molecule of excess methionine requires at least two molecules of glycine. Hence it makes sense to hypothesize that diets highest in animal (muscle) protein, even though they represent the highest intake of all the protein amino acids, may actually result in decreasing the blood levels of glycine, owing to the depletion of glycine for methionine clearance.
This is important because glycine is an important endogenous regulator of macrophages, the immune cells that effect the primary inflammatory response3. Hence, lower glycine levels should make for a greater vulnerability to inflammation, and thus for a host of chronic conditions rooted in inflammation, including type 2 diabetes.
Meanwhile, the importance of glycine in Type 2 diabetes has been established by both observational and experimental evidence. For example, a 2016 Harvard SRMA4 reviewed and compiled data from 8 prospective studies on 1940 diabetic subjects compared to over 6,000 controls. Among significant findings was an inverse association between blood glycine concentration and type 2 diabetes.
While this finding is of limited value in the absence of dietary intake data, recently, the hypothesis linking low glycine levels with high meat consumption has been supported by the UK arm of the EPIC study in a 2016 study of 392 middle-aged men5. Equal numbers of men representing four dietary groups (meat-eaters, fish-eaters, vegetarians and vegans) were recruited into the study which recorded their dietary consumption of all the protein amino acids, as well as their plasma levels of the free amino acids. The finding that glycine was highest in the diets of meat-eaters (20% higher consumption than that of vegans), yet lowest in the plasma of meat-eaters (14% lower than that of vegans), can be explained by the depletion of glycine by the methionine clearance pathway.
Experimentally, a randomized, placebo controlled trial of oral glycine supplementation in 20086 found a significant reduction in markers of inflammation and diabetes, including a reduction in hemoglobin A1C in the glycine treatment group from 8.3 to 6.9%.
I would therefore suggest that future studies on diet and the development of type 2 diabetes and other inflammation-related disorders focus on the amino acid content of proteins consumed, perhaps including the collection of data on the intake of gelatin and bone broths.
1. Neuenschwander M, Ballon A, Weber KS, et al. Role of diet in type 2 diabetes incidence: umbrella review of meta-analyses of prospective observational studies. BMJ 2019;366:l2368
2. Martinov MV, Vitvitsky VM, Banerjee R, Ataullakhanov FIl. Review: The logic of the hepatic methionine metabolic cycle. Biochim Biophys Acta 2010; 1804: 89–96. doi:10.1016/j.bbapap.2009.10.004
3. Wheeler MD, Ikejema K, Enomoto N, et al. Glycine: a new anti-inflammatory immunonutrient (Review). Cell Mol Life Sci 1999;56:843–856.
4. Guasch‐Ferré M, Hruby A, Toledo E et al. Clish (2016). Metabolomics in prediabetes and diabetes: A systematic review and meta‐analysis. Diabetes Care 2016;39:833–46. https://doi.org/10.2337/dc15-2251
5. Schmidt JA, Rinaldi S, Scalbert A, et al. Plasma concentrations and intakes of amino acids in male meat-eaters, fish-eaters, vegetarians and vegans: a cross-sectional analysis in the EPIC-Oxford cohort. Eur J Clin Nutrition 2016;70:306–12. doi:10.1038/ejcn.2015.144
6. Cruz M, Maldonado-Bernal C, R. Mondragón-Gonzalez R, et al. Glycine treatment decreases proinflammatory cytokines andincreases interferon- in patients with Type 2 diabetes. J Endocrinol Invest 2008;31:694-99
Competing interests: Author is also President and CEO of Natural Food Science, LLC, which manufactures and markets a glycine supplement product.