No short-term effects of calorie-controlled Mediterranean or fast food dietary interventions on established biomarkers of vascular or metabolic risk in healthy individuals

Fast food is commonly accepted to be a health threatening, "atherogenic" diet, regardless of the amount of caloric intake or the mode of preparation. Fast food is characterized by high amounts of saturated fat and a high caloric density. Several studies demonstrated the association with weight gain, poor dietary indicators, insulin resistance and obesity [1]. In contrast, epidemiological studies, but not intervention studies, pointed to a lower prevalence of various diseases in populations on the Mediterranean diet. Lower incidence rates were demonstrated for cardiovascular diseases [2, 3, 4, 5], chronic neurodegenerative diseases [6], malignomas [5, 7], obesity and Diabetes mellitus type 2 [8, 9, 10, 11]. There is no doubt that diet in general is an important part of the therapeutic regime in metabolic diseases and that certain eating habits are associated with development and progression of metabolic disease [12, 13, 14].

In a recent intervention study, Estruch et al. [15] reported that a Mediterranean diet supplemented with either extra virgin olive oil or nuts has beneficial effects in the primary prevention of cardiovascular disease in a high risk population. It is of importance to realize that there was no significant effect on death from any cause or myocardial infarction, the only statistically significant protective effect reported was for stroke. The margins of risk were very small and the authors' interpretation of the data provoked substantial critique: The benefit reported for stroke was apparent only in the first study year and vanished during further follow-up, while apparent differences in the baseline characteristics between the study groups might have contributed to the minor, short-term differences in absolute risk [16, 17]. Furthermore, the results presented by Estruch et al. are only statistically significant as long as calculated as a projection for 1,000 person years. Therefore we calculated the data for the entire follow-up period without applying a projection for 1,000 patient years and found the primary end point to be no longer statistically significant; neither for the comparison between the nut-based Mediterranean diet (83 events in 2,454 individuals) and control condition (109 events in 2,450 individuals), P = 0.054 (2-tailed), nor for the comparison between extra virgin olive oil Mediterranean diet (96 events in 2,543 individuals) and control condition (109 events in 2,450 individuals), P = 0.23. The discrepancy between the interpretation as presented by the authors and the more objective event based calculation used above, indicates that an effect of diet in the primary prevention of cardiovascular disease in high risk populations might be minor. Thus we need to learn more about determinants of atherosclerosis potentially applicable to medical or lifestyle intervention.

Numerous studies addressed the issue of protective effects of "healthy food" for healthy people, but prospective randomized trials proving efficiency of e.g. low fat, high fiber intake in healthy people with respect to hard endpoints are not yet available. Thus the question arises whether healthy people can become even healthier by a specific diet or whether fast food per se has a negative impact on a healthy person's individual risk profile independent of caloric intake [18]. It was recently shown that there is no beneficial effect of a low fat diet compared to either increased protein or carbohydrate intake, with respect to weight loss in type 2 diabetic patients [19]. Another study demonstrated the lack of favorable effects of lifestyle intervention in young patients with type 2 diabetes [20]. While guidelines for healthy dietary regimes recommend restriction of sucrose intake, a weight-maintaining diet with high sucrose intake had no effect on insulin sensitivity, glycemic profiles or measures of vascular compliance in healthy, non diabetic subjects [21]. In a recent systemic review, only the consumption of whole grain products showed moderate-grade evidence for protection against type 2 diabetes and cardiovascular disease [22]. In addition, the Palaeolithic ("old stone age") diet, which is rich in meats, fish, fruits, vegetables, rood vegetables, eggs and nuts, improved glucose tolerance in patients with heart disease and either glucose intolerance or type 2 diabetes, more than the Mediterranean diet [23, 24]. These data indicate that the concept of healthy vs. unhealthy diet for healthy individuals should be viewed more critically.

In order to address the question of short-term effects of dietary regimes in healthy individuals, we conducted a two-week intervention study in healthy male volunteers (Table 1), comparing three dietary regimes (Table 2) after a standardized run-in period of two weeks (Fig. 1). The Mediterranean diet as an example of a supposedly health improving diet was compared to fast food, and the German cooking style - reflecting the traditional local eating regime - served as a reference group. Importantly, all groups received equal amounts of calories per day.

Fig. 1
Schematic timeline of the study and randomization of n = 39 participants in three groups with different dietary portfolio.

• Click for larger image

Table 1
Subject characteristics: body weight and body mass index (BMI) measured before and after two weeks of dietary intervention

• Click for larger image

Table 2
Main features of the three diet groups

• Click for larger image

Main parameters to detect changes in the individual risk profile were classical biomarkers for vascular risk such as LDL- and HDL-cholesterol and homocysteine, previously shown to be affected by the Mediterranean diet [10, 25, 26, 27]. Since oxidative stress has been shown to contribute to vascular disease and aging in general, we included biomarkers of oxidative stress, namely 8-hydroxy-2-deoxyguanosine and malondialdehyde. 8-hydroxy -2-deoxyguanosine derives from deoxyguanosine oxidation and reflects the level of DNA damage by oxidation, while Malondialdehyde is a reactive aldehyde that results from the degradation of polyunsaturated lipids by reactive oxygen species [28, 29, 30]. Furthermore methylglyoxal, a highly cytotoxic α-oxoaldehyde derived from triose phosphate intermediates of glycolysis, was studied. Methylglyoxal is a potent glycating agent and is believed to be of main importance in aging and degenerative neuronal disease. In experimental settings, methylglyoxal has been shown to mediate vascular disease and pain by modification of amino acids [31]. Thus methylglyoxal is believed to be a major toxic metabolite derived from energy flux and a biomarker for overloaded metabolic detoxification mechanisms, predicting development of pain and endothelial cell dysfunction. The effects of dietary intervention on methylglyoxal as a biomarker for changes in metabolic flux and detoxification processes have not been studied jet. The relatively short intervention phase of two weeks was chosen because numerous studies in healthy subjects demonstrated that e.g. high-fat vs. low-fat interventions or changing the fatty acid composition result in diet-induced changes in the vascular and metabolic risk profile as early as two weeks after intervention, suggesting a significant immediate effect of substantial dietary changes on the individual risk profile [32, 33, 34, 35, 36, 37, 38, 39]. Importantly, two intervention studies comparing low-fat/high-protein and low-fat/high-carbohydrate diets with regard to markers of vascular risk report a more pronounced short-term response after 2 or 4 weeks when compared to the effects in the long run [33, 36], pointing towards the ability of healthy individuals to metabolically compensate for changes in the dietary composition.

The first objective of this study was to assess whether changes in the food composition and the mode of preparation - reflected by different dietary regimes - have an immediate effect on the individual risk profile, independent of overall energy consumption. The second objective was to study the effect of different diets on the metabolic marker methylglyoxal, not studied in the context of dietary intervention in healthy subjects before. Thus this study was planned to understand whether calorie-adjusted diets affect biomarkers reflecting the risk of vascular and metabolic disease in healthy individuals, or whether healthy individuals possess the metabolic capacity to balance changes in the dietary composition.

With this pilot study the authors addressed one main question: do changes in the dietary regime affect parameters reflecting aging, vascular risk or metabolic disease in healthy young men independent of differences in caloric intake? This study did not address the already extensively investigated question of diet-induced impacts on health by excessive caloric intake, the effect of diet on metabolically ill patients or the effect of unbalanced nutrition. The short duration of the study is a limitation. However, we chose a two-week intervention period in order to detect potential immediate diet-induced changes in biomarkers of vascular and metabolic risk, previously shown to be affected by a two-week intervention [32, 33, 34, 35, 36, 37, 38, 39]. Another limitation is that potential effects of temperature difference in the process of preparation of the dishes were not investigated in this study, also potential effects of cooling and rewarming [47] that occur in fast food restaurants, have been excluded by the study design. Thus this study focused exclusively on the short-term effect of three different diets on markers reflecting potential risk of disease in healthy volunteers. An important feature and strength of this study is the investigation of dietary intervention under stringent control for equal energy consumption in all study groups, assuring the investigation of effects highly specific to the dietary composition itself and not due to differences in calorie consumption coming along with changes in the dietary regime (fast food: high energy density, short-lasting in satisfying peoples' appetite; Mediterranean diet: low energy density, long-lasting satisfaction of appetite). Furthermore, the inclusion of fast reacting biomarkers of oxidative and metabolic stress such as methylglyoxal, which has not been studied in the setting of dietary intervention in healthy individuals before, allowed the detection of potential immediate metabolic effects of the dietary composition.

In this study, biomarkers of vascular and metabolic risk showed no difference between study groups or compared to baseline in the setting of freshly prepared dishes with equal amounts of calories during a two-week intervention period. The minor, but statistically significant increase in uric acid in the Mediterranean group (+ 0.59 mg/dl when compared to baseline) could, at least in part, be explained by the relatively high amount of purines (Table 3) and is unlikely to be of significant clinical relevance in healthy individuals. The same is true for the increases over baseline for urea observed in the fast food and in the Mediterranean group, since the levels remained well within the reference value of < 45 mg/dl. But we cannot completely rule out the possibility that changing from the traditional German baseline diet to Mediterranean or fast food might have caused slight changes in protein metabolism. It is noteworthy, but not the focus of the study, that the participants in the fast food group complained about feeling constantly hungry when consuming 2,500 kcal per day and demanded increased amounts of food, while participants on the German or Mediterranean diet did not. The observation of feeling malnourished by fast food is in accordance with other studies proving that fast food is less long-lasting in satisfying peoples' appetite, thus leading towards excessive caloric intake, obesity and metabolic disorders in the long run [48].

In our approach with study groups controlled for equal energy consumption, significant effects were observed neither for classical biomarkers of cardiovascular risk nor for the sensitive and fast reacting biomarkers of oxidative and metabolic stress 8-hydroxy-2-deoxyguanosine, malondialdehyde and especially the toxic reactive carbonyl methylglyoxal [49, 50], which has not been studied in such a setting before. Even in studies with longer intervention periods [51, 52] or in studies focused on very early acute effects of different diets [53], no significant effects were observed previously. Additionally, studies using multicompound diet interventions comparing Mediterranean diet with other diets [54] or dietary interventions with red wine [55] did not show significant benefits on total cholesterol, LDL-cholesterol or triglyceride levels in healthy subjects. Certainly there were some differences with respect to the content of individual nutrients in the three cooking styles tested, but for most components the contents of the three diets were not as different as generally believed (Table 3). One should keep in mind that the differences between individuals, even healthy individuals, with respect to carbohydrate metabolism [56], insulin secretion and sensitivity [57, 58], lipid [59, 60] and vitamin metabolism [61, 62, 63, 64, 65, 66, 67] are by far greater than the differences between the nutrient contents in the three groups tested. Thus healthy individuals have the capacity to balance the differences in the nutrient composition of common diets as long as changes in the caloric uptake are avoided. Therefore, the question arises whether recommended daily doses for certain nutrients are applicable and valid for each and every healthy person, since these recommendations neglect the huge metabolic variability between individuals and most importantly are not based on intervention studies in healthy people. Concerning nutrition counseling, our data indicate that short-term noncompliance to a recommended "healthy" dietary regime is unlikely to substantially increase a healthy person's individual risk profile. Therefore, preventive nutrition counseling should take into account that extrinsically motivated changes in a person's eating habits might induce psychic and social stress, while an occasional return to the previous dietary regime, excluding increased energy consumption, could be a feasible way to avoid those stresses. Others and we have shown recently that psychosocial stress significantly affects signaling events mediated by reactive oxygen species (ROS), culminating in activation of the ROS-sensitive transcription factor NF-κB and finally in atherosclerosis [68]. These data in human and mice are consistent with the data of the "Interheart" Study [69], demonstrating an almost nine fold stronger effect of psychic stress compared to nutrition. The role of psychosocial factors has been highlighted recently in social intervention studies, demonstrating that social intervention can reduce the incidence of obesity and diabetes [70, 71]. The psychosocial effect was stronger and resulted in a lower number needed to treat compared to published nutritional intervention studies in obesity or diabetes [71]. Therefore it is not surprising that the recently published PREDIMED study [15], while assuming a potential effect of Mediterranean cooking on stroke, had no effect on death or even myocardial infarction in the high risk population studied. While on the one hand the issue of healthy food turning healthy individuals even healthier is questioned by the authors of this study, there is on the other hand no doubt that wrong nutrition does contribute to disease. Thus further studies are needed to investigate the differences in energy and nutrient metabolism in healthy people, with the perspective of identifying new surrogate markers for clinically relevant end points. Having such studies at hand, it would become easier to give valid nutrition counseling to healthy individuals [72].

In summary, there were neither diet-specific negative effects of fast food on the parameters studied compatible with a threat to health, nor a beneficial effect of Mediterranean food in healthy young men when consuming the same amount of calories. We conclude that, while there is no doubt about the health threatening effects of overnutrition, undernutrition or malnutrition, there is no solid evidence that the cooking style itself impacts on the health of young healthy men studied here and in other cohorts [73, 74, 75]. Thus healthy individuals can, at least for two weeks, compensate for the relatively small differences in nutrient composition of supposedly 'unhealthy' dietary regimes. It remains to be studied how social factors and enzymatically controlled pathways of energy metabolism cooperate and might have the capability to affect the results of this study.