Bonaccio M, Di Castelnuovo A, Costanzo S, Persichillo M, De Curtis A, Donati MB, de Gaetano G, Iacoviello L, on behalf of the MOLI-SANI StudyInvestigators.
Adherence to the traditional Mediterranean diet and mortality in subjects with diabetes. Prospective results from the MOLI-SANI study. European Journal of Preventive Cardiology 2015. Pre-publication. DOI: 10.1177/2047487315569409.
Background: Adherence to the Mediterranean diet is associated with lower mortality in a general population but limited evidence exists on the effect of a Mediterranean diet on mortality in subjects with diabetes. We aim to examine the association between the Mediterranean diet and mortality in diabetic individuals.
Design: Prospective cohort study on 1995 type 2 diabetic subjects recruited within the MOLI-SANI study.
Methods: Food intake was recorded by the European Project Investigation into Cancer and Nutrition food frequency questionnaire. Adherence to the Mediterranean diet was appraised by the Greek Mediterranean diet score. Hazard ratios were calculated using multivariable Cox-proportional hazard models.
Results: During follow-up (median 4.0 years), 109 all-cause including 51 cardiovascular deaths occurred. A 2-unit increase in Mediterranean diet score was associated with 37% (19%–51%) lower overall mortality. Data remained unchanged when restricted to those being on a hypoglycaemic diet or on antidiabetic drug treatment. A similar reduction was observed when cardiovascular mortality only was considered (hazard ratio=0.66; 0.46–0.95). A Mediterranean diet-like pattern, originated from principal factor analysis, indicated a reduced risk of overall death (hazard ratio=0.81; 0.62–1.07). The effect of Mediterranean diet score was mainly contributed by moderate alcohol drinking (14.7% in the reduction of the effect), high intake of cereals (12.2%), vegetables (5.8%) and reduced consumption of dairy and meat products (13.4% and 3.4% respectively).
Conclusions: The traditional Mediterranean diet was associated with reduced risk of both total and cardiovascular mortality in diabetic subjects, independently of the severity of the disease. Major contributions were offered by moderate alcohol intake, high consumption of cereals, fruits and nuts and reduced intake of dairy and meat products.
Many scientists have reported on health benefits associated with the traditional diet of much of the Mediterranean area: the primary basis of the diet is on plant-based foods (whole grains, vegetables, fruits), olive oil, moderate wine consumption, and limited intake of meat or dairy products. In 2003, Antoniao Trichopoulou and her associates (Trichopoulou et al, 2003) created a scoring system based on the number of components of the Mediterranean Diet (Med-Diet) consumed by subjects, using a total score to judge the degree that individuals were following a Med-Diet pattern. They found that greater adherence to such a diet (a higher score) had a favorable effect on the risk of mortality. Most studies since then have shown the same, but generally have related the full dietary pattern to outcomes; there are limited data on the relative merits of each component of the diet. The authors report here on their attempt to judge the proportion of the protection against mortality that can be attributed to each component of the score.
In the present study, the investigators used this scoring system that includes each component of the Med-Diet assessed from dietary records, with the intake of 7 items (cereal intake, ratio of monounsaturated to saturated fats, vegetables, fruits, nuts, fish, and moderate alcohol), considered to have positive effects and 2 items (dairy products, meat and meat products) considered to have negative effects. Subjects whose intake of each positive component was above the median level of intake were given 1 point on the Med-Diet score, with those consuming less than the median values given 0 points (for negative items, intake below the median point resulted in 1 point). For alcohol, men reporting 10-50 g/day and women reporting 5-25 g/day received 1 point, with subjects reporting less or more than these levels being given 0 points. Possible scores ranged from 0 to 9 points, with higher scores indicating greater adherence to a Med-Diet pattern.
The authors’ calculations estimated that, in descending order of importance, moderate alcohol consumption (associated with a reduction of 14.7% in the protection against mortality when it was removed from the total score), cereal intake (12.2% reduction), ratio of monounsaturated to saturated fats (5.8%), and consumption of vegetables (5.8%), fruits and nuts (5.2%) and fish (5.0%) lowered mortality risk. Low intakes of dairy products (reduction of 13.4%) and meat and meat products (3.4%) were also associated with lower mortality
Specific Comments on Study by Forum Members: Most Forum members considered this to be a well-done attempt to tease apart the relative importance of each component of the Med-Diet as it relates to health outcomes. Said reviewer Finkel: “Diabetes provides a severe test for any regimen aspiring to cardiovascular health benefits: in this study, most components of the Med-Diet provided protection among such subjects.” Reviewer Goldfinger agreed: “In contemporary medical practice, diabetics are already presumed to have established vascular disease and thus are subject to more stringent targets for cholesterol and LDL cholesterol lowering and more aggressive pharmaceutical and lifestyle intervention, as their health is expected to be more easily compromised. This is indeed a high-risk group.”
In regards to the present study, Forum member Skovenborg stated: “Kudos to these researchers from the MOLI-SANI study with this convincing analysis extending the known benefits of the Mediterranean diet to a cohort of type 2 diabetics. Serge Renaud was among the pioneers in the study of the Cretan Mediterranean diet for prevention of coronary heart disease (Renaud et al). The beneficial effects of the Mediterranean dietary pattern have been studied in US populations within both smoking and BMI strata (Mitrou et al). The results of the MOLI-SANI Study are in accordance with the Greek EPIC prospective cohort study analysis of the contributions of the individual components of the Mediterranean diet regarding moderate alcohol consumption, the major contributor with 23.5%, while minimal contributions were found for cereals and dairy products (Trichopoulou et al, 2009).”
Forum member Goldfinger agreed that this study provides important data regarding the management of diabetes mellitus: “The Med Diet has been proposed especially as a tool for primary prevention. With respect to diabetics, this study provides useful data to suggest an important role of the Med-Diet in secondary prevention, that is, reduced cardiovascular and other clinical end points among those with established disease. In addition to reinforcing benefits of lifestyle modification in diabetics, it specially reinforces the benefit of moderate alcohol in these subjects, who are often ill advised not to drink.”
Alcohol and diabetes mellitus: An inverse association between alcohol consumption and the development of diabetes mellitus has been consistently demonstrated for decades. For example, Stampfer and colleagues reported in 1988 that moderate drinkers in the Nurses’ Health Study had a lower risk for development of diabetes than did abstainers. Numerous studies since then have confirmed such an association (Rimm et al, Perry et al); good summaries of the effect have been provided by Howard et al and Koppes et al. Forum member Ellison noted that cumulative evidence shows a J-shaped or U-shaped curve, with usually an approximately 30% reduction in the risk of developing diabetes among light to moderate drinkers in comparison with abstainers (Koppes et al).
Reviewer Puddey commented: “The authors in discussing their results mention the potential beneficial effect of moderate alcohol intake on the incidence of type II diabetes mellitus. Any commentary needs to consider evidence that there is a U-shaped relationship between alcohol use and incidence of type II diabetes, with a meta-regression of 20 cohort studies by Baliunas et al concluding that when compared to lifetime abstainers, the relative risk (RR) for Type 2 Diabetes among men is lowest in those consuming 22 g/day of alcohol (RR 0.87 [95% CI 0.76 –1.00]) but deleterious at just over 60 g/day of alcohol. Among women, RR was at its lowest when consuming 24 g/day of alcohol (0.60 [95% CI 0.52– 0.69]) and became deleterious at about 50 g/day alcohol.”
Reviewer Van Velden noted: “The Med Diet plays an important role in primary and secondary prevention of CVD. Pharmacological treatment to increase insulin secretion may increase inflammation, and more attention should be placed on dietary interventions. The Med-Diet is important in this regard, as is wine.”
How precise an estimate of effect can come from the Med-Diet score? Forum member Lanzmann-Petithory insists that “There is no such thing as a single Med-Diet, but many variations throughout the Mediterranean region. In this heterogeneity, Serge Renaud identified alpha-linolenic acid (ALA) and wine as key factors (Renaud et al). In the present analysis, not only is the alcohol not identified as mainly wine, but the fat score does not take into account alpha-linolenic acid (ALA), nor the ratio of omega-6 to omega-3 fatty acids. ALA effect is diluted among the other groups: vegetables (primary source of ALA), meat and dairy products (following the food chain of the country), fruits and nuts. ALA was the principal factor in the cardiovascular protection of the diet in the Lyon Diet-Heart Study, a transposition of the Cretan traditional diet, rich in wild greens, and with an ALA enriched food made with canola oil (de Lorgeril et al). Many studies have focused on the benefits of ALA, the most recent being that of Cespedes et al, in which ALA, but not EPA and DHA, had a protective effect against development of the metabolic syndrome.”
Forum member Puddey had some criticisms of the scoring system used in this study to judge the effects of each component of the Med-Diet. “The observation in this paper that moderate alcohol consumption is a major contributor to an observed protective effect of a Mediterranean diet against mortality in Type II diabetic subjects is consistent with an already strong literature indicative of a protective effect of moderate alcohol consumption on overall and cardiovascular mortality in subjects with Type II diabetes mellitus. However, the article would be more informative from a preventative viewpoint if there was a comparison of mortality outcomes for those who drank either above or below the defined moderate drinking limits. As it stands, low or no alcohol consumers are lumped together with high alcohol consumers, with both groups scoring a zero on the Mediterranean diet score for alcohol intake. We therefore do not know if the 14.7% contribution of moderate alcohol intake to the reduction in overall mortality with a Mediterranean diet is due to not drinking heavily or whether it is possibly due to a specific protective effect of moderate alcohol use relative to low or no alcohol intake.”
Reviewer Ellison agreed that, “Theoretically, the broad categories of the Med-Diet score of Trichopoulou et al could make it difficult to interpret the implications of alcohol consumption on mortality risk. However, in the present study, the subjects given a 0 value for an alcohol intake that was outside the moderate limits were predominantly abstainers or light drinkers. Of the 676 women in this study, the large majority were abstainers (n=363) or very light drinkers (n=100), and there were only 64 women (9%) whose consumption exceeded the ‘moderate’ category of ≥ 25 g/d. Of the 1,319 men in the study, 397 (30%) were abstainers or light drinkers, while 245 (18.5%) exceeded the moderate limits. Given that, overall, the large majority of subjects not in the moderate category in this study were non-drinkers or light drinkers, I would interpret the findings to indicate that the demonstrated reduction in risk is more likely to be due to a specific protective effect of moderate alcohol use relative to low or no alcohol intake.”
Reviewer Zhang had some uncertainties about the approach used to interpret the individual components of the Med-Diet score, noting that nutritional analysis is especially complicated, as it may be difficult to interpret effects of not consuming a single item without considering the effect of its omission from the diet affecting the input of other foods. He also had a comment on another aspect of the analyses: “If the Med-Diet reduces the risk of diabetes, and we assume a person’s diet, in general, is relatively stable, examining the association of Med-Diet and risk of mortality conditioning on diabetes (akin to adjusting for an intermediate variable) would not only be susceptible to collider stratification bias, but also change the interpretation of the effect estimate. Such an effect estimate is reflecting the association of Med-Diet on mortality not through diabetes. This could mean that the total effect of the diet (and alcohol) on mortality might be even greater than estimated in this paper.”
Differential effects according to type of alcoholic beverage consumed: Professor Puddey commented further: “In the present analysis, there has been no account taken of the pattern of alcohol use or type of alcohol beverage. In terms of the latter, wine drinkers make different dietary choices compared to beer or spirits drinkers (Johansen et al) and higher wine intake associates with a higher intake of healthy food items (Tjonneland et al). Beverage preference may therefore be an unmeasured confounder in studies that associate alcohol in the Mediterranean diet with its protective effect.” Forum member Ellison noted: “While specific proportions are not given in the present study, most of the alcohol consumption in the Moli-Sani Study has been shown to be from the consumption of wine.”
Reviewer Skovenborg had some comments on Professor Puddey’s views on beverage preference and diet. “In a cross sectional study of customers in Danish supermarkets wine buyers made more purchases of healthy food items than people who buy beer; however, in Denmark the change from beer to wine was inspired by the increasingly popular holiday travels to Mediterranean countries. The same trend was observed in Finland where wine drinkers have more antioxidants in their diet. (Männistö et al).”
Skovenborg continued: “The traits of persons who choose wine might well be different in other nations – especially in the Mediterranean countries. For example, a cross-sectional analysis of the relation between wine drinking and intake of selected foods in Italy did not find an association of wine preference with indicators of healthy diet (Chatenoud et al). The conclusion of the SUN cohort study of alcoholic beverage preference and dietary pattern in Spanish university graduates found no relevant differences in adherence to the Mediterranean food pattern according to alcoholic beverage preference (Alcácera et al). In a sample of middle-aged French males, there was a linear trend between increasing alcohol intakes and worsening of quality of diet, while no difference was observed according to beverage preference (Herbeth et al).”
Differences between the response of men and women to the Med-Diet: Reviewer Puddey also wondered about differences in the relation between alcohol and mortality between men and women. He stated: “There were substantial differences in the definition of moderate drinking for men and women. In fact, men could have been drinking up to 10 times as much as women (50g vs 5g per day) and still be categorised within the same moderate drinking category. Given previous evidence of gender differences in alcohol – mortality relationships it would have been useful to know if the Mediterranean diet – mortality relationships in this study were of identical magnitude in men and women.”
References from Forum comments
Alcácera MA, Marques-Lopes I, Fajó-Pascual1 M, et al. Alcoholic beverage preference and dietary pattern in Spanish university graduates: the SUN cohort study. European Journal of Clinical Nutrition 2008;62:1178-1186
Baliunas DO, Taylor BJ, Irving H, Roerecke M, Patra J, Mohapatra S, Rehm J. Alcohol as a Risk Factor for Type 2 Diabetes A systematic review and meta-analysis. Diab Care 2009;2:2123-2132.
Cespedes E, Baylin A, Campos H. Adipose tissue n-3 fatty acids and metabolic syndrome. Eur J Clin Nutr 2015:69:114-120. doi:10.1038/ejcn.2014.150.
Chatenoud L, Negri E, La Vecchia C, Volpato O, Franceschi S. Wine drinking and diet in Italy. Eur J Clin Nutr 2000;54:177-179.
de Lorgeril M, Renaud S, Mamelle N. Mediterranean alpha-linolenic acid-rich diet in secondary prevention of coronary heart disease. Lancet 1994;343:1454–1459.
Herbeth B, Samara A, Stathopoulou M, Siest G, Visvikis-Siest S. Alcohol consumption, beverage preference, and diet in middle-aged men from the STANISLAS Study. Journal of Nutrition and Metabolism 2012: Article ID 987243. doi:10.1155/2012/987243.
Howard AA, Arnsten JH, Gourevitch MN: Effect of alcohol consumption on diabetes mellitus: a systematic review. Ann Intern Med 2004;140:211-219.
Johansen D, Friis K, Skovenborg E, Gronbaek M. Food buying habits of people who buy wine or beer: cross sectional study. Br Med J 2006;332:519-522.
Koppes LL, Dekker JM, Hendriks HF, et al: Moderate alcohol consumption lowers the risk of type 2 diabetes: a meta-analysis of prospective observational studies. Diabetes Care 2005;28:719-725.
Männistö S, Uusitalo K, Roos E, Fogelholm M, Pietinen P. Alcohol beverage drinking, diet and body mass index in a cross-sectional survey. European Journal of Clinical Nutrition 1997;51:326-332.
Mitrou PN, Kipnis V, Thiébaut ACM, et al. Mediterranean Dietary Pattern and Prediction of All-Cause Mortality in a US Population. Results From the NIH-AARP Diet and Health Study. Arch Intern Med 2007;167:2461-2468. doi:10.1001/archinte.167.22.2461.
Perry IJ, Wannamethee SG, Walker MK, et al: Prospective study of risk factors for development of non–insulin dependent diabetes in middle aged British men. BMJ 1995;310:560-564.
Renaud S. de Lorgeril M, Delaye J, et al. Cretan Mediterranean diet for prevention of coronary heart disease. Am J Clin Nutr 1995;61(suppl):1360-1367S.
Rimm EB, Chan J, Stampfer MJ, et al: Prospective study of cigarette smoking, alcohol use, and the risk of diabetes in men. BMJ 1995; 310:555-559.
Stampfer MJ, Colditz GA, Willett WC, et al: A prospective study of moderate alcohol drinking and risk of diabetes in women, Am J Epidemiol 1988;128:549-558.
Tjonneland A, Gronbaek M, Stripp C, Overvad K. Wine intake and diet in a random sample of 48763 Danish men and women. Am J Clin Nutr 1999; 69:49-54.
Trichopoulou A, Costacou T, Bamia C, Trichopoulos D. Adherence to a Mediterranean diet and survival in a Greek population. N Engl J Med 2003;348:2599-2608.
Trichopoulou A, Bamia C, Trichopoulos D. Anatomy of health effects of Mediterranean diet: Greek EPIC prospective cohort study. BMJ 2009;338:b2337.
Many scientists have reported on health benefits associated with the traditional diet of much of the Mediterranean area: the primary basis of the diet is on plant-based foods (whole grains, vegetables, fruits), olive oil, moderate wine consumption, and limited intake of meat or dairy products. In 2003, Antonia Trichopoulou and her associates created a scoring system based on the number of components of the Mediterranean Diet (Med-Diet) consumed by subjects, giving a total score to judge the degree that individuals were following a Med-Diet pattern. The score is based on the intake of 7 items (cereal intake, ratio of monounsaturated to saturated fats, vegetables, fruits, nuts, fish, and moderate alcohol), considered to have positive effects, and 2 items (dairy products, meat and meat products) considered to have negative effects. Many groups have found that greater adherence to such a diet (a higher score) is associated with a lower risk of many diseases, and lower mortality. There are limited data on the relative merits of each component of the Med-Diet. In the present paper, the authors attempt to judge the proportion of the protection against mortality that can be attributed to each component of the score.
The authors’ calculations estimated that, in descending order of importance, moderate alcohol consumption (associated with a reduction of 14.7% in the protection against mortality when it was removed from the total score), cereal intake (12.2% reduction), ratio of monounsaturated to saturated fats (5.8%), and consumption of vegetables (5.8%), fruits and nuts (5.2%) and fish (5.0%) lowered mortality risk. Lower intakes of dairy products (reduction of 13.4%) and meat and meat products (3.4%) were also associated with lower mortality. Alcohol has long been known to relate to a lower risk of developing diabetes, and the present study indicates that it is an important factor in reducing the risk of mortality among subjects who have already developed diabetes, as has been shown in previous studies. This study indicates further that the full Med-Diet has very favorable effects on mortality among diabetics.
Forum members considered this to be a well-done attempt to tease apart the relative importance of each component of the Med-Diet as it relates to health outcomes among diabetics. They felt that these researchers from the MOLI-SANI study have published a convincing paper extending the known benefits of the Mediterranean diet to a cohort of type 2 diabetics, a group of subjects at high risk for cardiovascular and all-cause mortality.
There were some questions raised about the precision of the Med-Diet Score, as published and used in this study, in judging the relative contributions of the different dietary constituents, especially because of known complexities in judging the effects of avoiding one particular food when its absence may affect the intake of other foods. And for alcohol, the score used did not permit an evaluation of the role of the pattern of drinking or even the type of beverage consumed, which are known to affect the net effects of drinking on health.
Nevertheless, the results of this study add to an accumulating base of knowledge of the importance of the Mediterranean-type diet in reducing the risk of many health outcomes, including mortality among diabetics.
* * * * * * *
Comments on this critique by the International Scientific Forum on Alcohol Research were provided by the following members:
Ian Puddey, MD, Dean, Faculty of Medicine, Dentistry & Health Sciences, University of Western Australia, Nedlands, Australia
Yuqing Zhang, MD, DSc, Clinical Epidemiology, Boston University School of Medicine, Boston, MA, USA
Erik Skovenborg, MD, specialized in family medicine, member of the Scandinavian Medical Alcohol Board, Aarhus, Denmark
Harvey Finkel, MD, Hematology/Oncology, Boston University Medical Center, Boston, MA, USA
Tedd Goldfinger, DO, FACC, Desert Cardiology of Tucson Heart Center, University of Arizona School of Medicine, Tucson, AZ, USA
Pierre-Louis Teissedre, PhD, Faculty of Oenology–ISVV, University Victor Segalen Bordeaux 2, Bordeaux, France
Arne Svilaas, MD, PhD, general practice and lipidology, Oslo University Hospital, Oslo, Norway
Gordon Troup, MSc, DSc, School of Physics, Monash University, Victoria, Australia
David Van Velden, MD, Dept. of Pathology, Stellenbosch University, Stellenbosch, South Africa
Dominique Lanzmann-Petithory,MD, PhD, Nutrition/Cardiology, Praticien Hospitalier Hôpital Emile Roux, Paris, France
R. Curtis Ellison, MD. Section of Preventive Medicine & Epidemiology, Department of Medicine, Boston University School of Medicine, Boston, MA, USA
Note on Potential Conflict of Interest: It should be pointed out that one of the authors of this paper (Giovanni de Gaetano) is a member of the International Scientific Forum on Alcohol Research. Dr. de Gaetano has not provided comments for the critique of this paper.