2. Materials and Methods
2.1. Literature Search
The PubMed database was searched at the end of March 2017, by combining various search terms for dairy foods (“milk” or ”dairy” or “dairy products” or “yoghurt” or “yogurt” or “yoghourt” or “cheese” or “cultured milk products”) with search terms for non-communicable diseases (“cardiovascular diseases” or “cardiovascular disease” or “heart disease” or “heart diseases” or “stroke” or “myocardial infarction” or “hypertension” or “blood pressure” or “diabetes” or “mortality” or “death” or “obesity” or “metabolic syndrome” or “diabetes mellitus, type 2” or “type 2 diabetes” or “cancer”). The search was limited to the last five years (January 2012–March 2017) and to human studies. Only meta-analyses and systematic reviews were taken into consideration. Search results were evaluated, and studies in children, or those that focused on single nutrients or individual microorganisms only, were excluded.
The identified references were analyzed by specifically focusing on fermented dairy products, including yoghurt, cheese, and fermented milk and comparing the outcomes of these analyses with milk (as a reference for non-fermented dairy) and total dairy (for the total contribution of fermented and non-fermented dairy). Consequently, physiological effects of fermented dairy products which are mediated by the fermentation process, as assessed by one of the clinical indications investigated in this report, should be observed with fermented milk, yoghurt and cheese but not with milk. Such effects could also be observed with total dairy intake, provided fermented products contribute significantly to the total intake of dairy products. Finally, one should note that effects observed with the consumption of cheese, but not milk, might not be uniquely attributed to the impact of milk fermentation since cheese making also involves the removal of whey.
For probiotics, the same search was conducted as on dairy products, except that the key words “devoted to dairy products” were replaced by the key word “probiotic*”(the asterisk being the PubMed truncation symbol).
For fermented foods of plant origin, the same search strategy was used that is described above, except the key words “devoted to dairy products” were replaced by the key words “plant or soy*” or “sauerkraut or olive*” or “coffee” or “wine” or “fermented plant food*”. As the outcome of this search strategy was not satisfactory, a manual search was subsequently performed.
2.2. Rating of the Results from the Meta-Analyses
To rate the results of meta-analyses summarizing observational studies, the characterization of an effect as ‘beneficial’ or ‘detrimental’ was based on the direction of the effect and on the statistical significance of the dataset provided by each meta-analysis. Non-significant effects were characterized as ‘neutral’. The size of the effect was qualified as ‘weak’ for a risk reduction of ≤10% and ‘moderate’ for a risk reduction >10%. For meta-analyses summarizing three or more prospective cohort studies (grade IIa), the quality of the evidence was rated as ‘moderate’ without consideration of other important parameters of relevance to an estimation of the evidence, such as the heterogeneity of the studies, the relevance of the population studied, or the qualities of the investigated foods or probiotics tested. For meta-analyses summarizing less than three prospective cohort studies or summarizing case-control studies (grade IIIa) or retrospective studies (IIIb) the evidence was rated as ‘weak’ (see also Table S1
To rate the results of meta-analyses summarizing randomized controlled intervention studies, the characterization of an effect as ‘beneficial’ or ‘detrimental’ was based on the direction of the effect and on the statistical significance of the dataset provided by each meta-analysis. Non-significant effects were characterized as ‘neutral’. The size of the effect was qualified as ‘weak’ for a risk reduction of ≤10% and ‘moderate’ for a risk reduction >10% compared to normal range values. All meta-analyses were summarizing three or more controlled randomized trials (grade Ia), and the quality of the evidence was therefore rated as ‘moderate’ without consideration of other important parameters of relevance to an estimation of the evidence, such as the heterogeneity of the studies, the relevance of the population studied, or the qualities of the investigated foods or probiotics tested.
This report does not focus on particular brands of products (e.g., Swiss cheese), particular bacterial strains, probiotic products, or nutraceuticals. As such, the products described in this report (e.g., cheese or yoghurt) are to be considered from a generic point of view.
The first clear outcomes of this report are that (i) the consumption of fermented foods is not associated with negative risks to humans with regard to the discussed clinical indications, i.e., cardio-metabolic health and cancer; and (ii) the majority of the results are associated with neutral effects.
Most, if not all, of the clinical data summarized in meta-analyses have been obtained with fermented products as part of total dairy intake (mostly prospective cohort studies), with other sources of fermented foods being much less investigated so that studies of sufficient quality are not available to perform meta-analyses.
Generally, the associations of fermented dairy with cardio-metabolic diseases and cancer comply with those of dairy products—this association being neutral for cancer and weakly beneficial, though inconsistent, for specific aspects of cardio-metabolic health. The strongest evidence for a beneficial effect of fermented dairy products was found for the effects of yoghurt on risk factors associated with type 2 diabetes. However, for all combinations of indications and dairy products reviewed in this report, the heterogeneity in the design of the studies as well as in the products being investigated, impedes a definitive assessment of the association between product consumption and health. In that context, it is unclear, and rather unlikely that additional studies will provide more robust answers to these questions. The above statement is not only restricted to the evaluation of fermented dairy products but is also relevant to other fermented foods.
The literature provides some information on the bioactivity and mechanisms mediating the impact of fermented foods in health and disease prevention, in particular, in regard to dairy products [87
]. The health benefits associated with the fermentation process may be the result of direct interactions between the ingested live microorganisms and the host (probiotic effect), or indirectly, through ingestion of microbial metabolites and products of fermentation (biogenic effect) [88
]. During the fermentation process, a wide range of peptides are released by proteolysis from caseins and whey proteins, some of them with bioactive effects, like blood pressure lowering or thrombin inhibition [89
]. Furthermore, a cholesterol lowering effect has been described in relation to exopolysaccharide-producing bacteria during fermentation, which results in increased synthesis of bile acids from cholesterol, decreasing the circulating level of cholesterol [90
]. Other proposed mechanisms are the assimilation of cholesterol by bacterial cells, also resulting in reduced absorption of exogenous cholesterol in the small intestine, or the conversion of cholesterol to coprostanol [91
]. Finally, bacterial cultures used for fermentation are known to synthesize vitamins, like folate, riboflavin, vitamin B12, or vitamin K2 (menaquinones), which are all involved in pathways important for cardiovascular health [21
More specifically, several mechanisms associated with the fermentation process may contribute to the inverse association between yoghurt intake and risk of T2D. On one hand, probiotic bacteria have been shown to improve lipid profiles and antioxidant status in T2D patients [92
] and have beneficial effects on cholesterol levels [93
]. On the other hand, the accumulation of insulinotrophic peptides and amino acids [94
] and the microbial synthesis of vitamins like menaquinones [8
] may improve insulin sensitivity. For the negative association between cheese intake and stroke, the evidence points more towards a consequence of technology, rather than fermentation, as cheese making significantly enriches calcium and magnesium, which have been associated with reduced risk of stroke [96
Thus, the literature is characterized by a wealth of data on the bioactivity of fermented foods, including metabolites and bacteria, which might beneficially contribute to a large range of physiological properties. However, the mechanistic data is not strongly linked to the results of the studies reviewed by meta-analyses. This gap can be explained by a range of factors, including the complexity of the investigated food–health interactions, the quality of the studies, and the real magnitude of the health benefits offered by these products. Further research should specifically integrate the impact of fermentation on risk factors in the study design, including the statistical approach. Such an approach may indeed provide a better indication of the impact of fermented foods on health.