Cognitive decline has become a major health problem with the aging of the population worldwide [1
]. By 2050, the total estimated prevalence of Alzheimer’s disease (AD) is expected to be 13.8 million, whereas the proportion of death resulting from AD is still increasing [2
]. Considering the undiscovered precise etiology of and the limited therapy of cognitive disorders, interventions on modifiable factors that may slow or prevent the development of cognitive decline have become an important public health issue.
A substantial amount of evidence has indicated that nutrition is associated with longevity and age-related diseases. Several healthy eating patterns and individual nutrients have shown preventive effects on cognitive decline [3
]. In recent years, accumulating epidemiological studies have investigated the potential role of milk, which is one of the most popular nourishments around the world. The associations of milk or dairy intake and chronic diseases have been under investigation, but the results are not conclusive. Previous studies reported that milk intake was inconsistently associated high blood pressure, type II diabetes, cardiovascular disease (CVD), and mortality [7
Milk and dairy foods have been inversely linked with the risk of stroke [13
]. Hemorrhagic stroke is a potential risk factor of dementia; as a result, milk is also likely to be associated with dementia. In fact, previous epidemiological studies have explored the association of milk intake and cognitive disorders [15
], but the conclusions remain contradictory. Some researchers have reported that the increased risk of cognitive disorders were significantly associated with lower intake of milk [15
]; however, other studies did not corroborate the inverse association [19
]. To the best of our knowledge, no quantitative assessment has been performed to evaluate the potential association between milk intake and cognitive disorders. Therefore, we conducted a systematic review and meta-analysis to summarize the evidence from observational studies on the relationship of milk consumption with the risk of age-related or vascular cognitive disorders (AD, dementia, cognitive decline, and cognitive impairment). Additionally, we performed subgroup and meta-aggression analyses to clarify whether the associations differed according to study characteristics.
2. Materials and Methods
2.1. Literature Search
We conducted the present systematic review and meta-analysis following the standard guidelines [22
]. From the inception to October 2016, the PubMed and the Embase databases were searched for observational studies reporting the association of milk consumption and cognitive disorders. Our search was limited to studies written in English. Search terms included “dairy”, “milk”, “yogurt”, “AD”, “dementia”, “Alzheimer*”, “aphronesia”, “cognitive*” and “cognition” (Table A1
in Appendix A
). We manually searched the reference lists of the relevant articles to identify more potential articles. When multiple publications from the same study were identified, the article with the longest follow-up duration was included.
2.2. Selection Criteria and Data Extraction
Initial screening was independently administered by two authors. After removing the duplicate articles, the title and abstract of each article was searched. Each article was independently assessed as exclusion or requiring further assessment. Disagreements were resolved by discussion within the two authors.
Inclusive criteria: (1) studies reporting relative risks (RRs) or hazard ratios (HRs) or odds ratios (ORs) and their corresponding 95% confidence intervals (CIs) of the cases of cognitive disorders involving intake of milk; (2) studies including a quantitative assessment of milk intake or milk combined other dairy products; (3) studies in which participants are aged ≥18 years old. Articles were excluded if (1) the data described a dietary pattern indicating a higher intake of milk, but did not describe the specific levels of milk intake; (2) the data described surrogate nutrients of milk, such as protein or fat; (3) the data described non-animal milk, such as breast milk or soy milk.
Data extraction was independently performed by two authors. The following data were extracted from each article: the first author and published year, study design, study location, number of participants and outcome cases, race, sex, the baseline age of participants, the duration of follow-up, the measurement method, the type of exposure and outcome, the category of exposure, adjustment variables, and the largest adjusted RRs, HRs, or ORs with corresponding 95% CIs of the risk of cognitive disorders.
2.3. Quality Assessment
Quality assessment was independently performed by two authors. We used a guideline of Methodological Evaluation of Observational Research [24
] and a quality assessment tool of observational cohort and cross-sectional studies [25
] to evaluate the study quality of the eligible articles. Five domains were assessed for each article: design bias, selection bias, information bias, confounding, and analysis bias. The highest score of the five domains were 3, 4, 5, 3, and 2 points, respectively. The total scale scored from 0 to a possible maximum of 17 points, and a higher point indicated a higher study quality. Disagreements were resolved through consensus.
2.4. Statistical Analysis
Stata, version 12.0 (StataCorp LP, College Station, TX, USA), and the Review Manager, version 5.2 (The Nordic Cochrane Centre, Copenhagen, Denmark), were used to perform the statistical analyses. ORs (95% CIs) for the highest level of milk intake compared with the lowest were used to measure the effect sizes for articles that reported the outcome of cognitive disorders. A generic inverse-variance random-effects method was used to pool the outcome data. A two-sided p-value of less than 0.05 was considered significant.
Q test and I2
statistic were used to evaluate the between-study heterogeneity, an I2
statistics of 50% or above was judged as statistically significant. Subgroup analysis was performed based on pre-specified characteristics: race (African, Asian, or Caucasian), gender (male, female, or both sexes), study design (cross-sectional or cohort), type of intake (milk or milk combined with other dairy products), the exposure assessment method (food-frequency questionnaire or others), and the type of participants (general population or patients). Meta-regression analysis was used to assess the heterogeneity between subgroups, and p-
values of less than 0.1 were considered significant results. By omitting one article at every turn, sensitivity analysis was used to evaluate the influence of a single article on the overall pooled results. Publication bias was estimated through Begg and Egger’s tests [26
This systematic review and meta-analysis identified seven studies involving a total of 10,941 subjects. The combined analysis showed that milk intake was inversely associated with the risk of cognitive disorders. Compared with the lowest level of milk consumption, the risk of cognitive disorders was decreased by 28% with the highest level of milk consumption, but the result was based on limited number of studies with considerable heterogeneity.
Although the stratified analysis by race did not show significant group-difference in the three subgroups, there was some evidence that the inverse association between milk intake and cognitive disorders was limited to Asian subjects, and the African populations showed an intermediate non-significant trend. A clear difference in the amount of milk and dairy consumption among Asian and Western countries was reported [28
]. Higher level of milk intake may significantly decrease the risk of cognitive disorders in populations (such as Asian) with relatively lower intake of overall milk and dairy products. In fact, excessive consumption of saturated fat in dairy products may adversely link with many chronic diseases [29
], but low-fat dairy foods were reported to be associated with beneficial health outcomes [30
]. Fat content was not described in all seven included studies, and only two studies reported more than three categories of milk intake; as a result, we could not separately evaluate the effect of whole-fat or low-fat milk. It is worth noting that the association was more pronounced in ischemic stroke patients based on a single study [17
]. Tu et al. indicated that a higher level of milk consumption might be a potential preventive measure to reduce the prevalence of vascular cognitive disorders among ischemic stroke patients. However, limited to the cross-sectional nature and small sample size [17
], false association should be considered. Further studies are still warranted to identify whether milk intake is more beneficial in some specific populations and diseases.
The preventive role of a diet rich in milk may be attributed to its protein, minerals, vitamins, and essential amino acids. Berg et al. reported that type 2 diabetes, hypertension, dyslipidemia, and obesity were associated with an increased risk of age-related cognitive dysfunction. Milk intake may reduce the risk of cognitive impairment via modifying neurovascular dysfunction, reducing weight and metabolic risks [32
]. Animal studies also reported the effect of dairy products in anti-obesity [34
]. Although no evidence from randomized controlled trials has investigated the preventive effect of milk consumption on cognitive disorders, several trials which showed the protective effect of milk intake against hypertension and obesity might partially support our finding [36
This is the first systematic review and meta-analysis specifically evaluated the association between milk intake and cognitive disorders. We extended the work by Crichton et al., who quantitatively showed an inverse association between milk intake and cognitive disorders [39
]. However, this study has several limitations. First, considerable heterogenicity was presented in our pooled analysis. We performed meta-regression analyses to investigate the possible explanation for the significant heterogeneity, and the results revealed that the heterogeneity might be associated with the characteristics of the subjects. Moreover, the diverse categories of milk intake (“times per week”, “gram per day”, “serving per day”, “high/low intake”, “tertiles”, etc.), the application of different dietary questionnaires, and different adjusted confounders may also lead to heterogeneous results. Second, we were unable to quantify the possible dose-response patterns of the association because of the different measurement units of milk consumption across studies. Furthermore, more than three categories of milk intake were not available in most of the included studies. High-quality prospective studies with standardized measurement unit should be administrated to further evaluate the potential dose-response effect of milk intake on the development of cognitive disorders. Third, because this is a meta-analysis based on observational studies, it is possible that the observed association is affected by unmeasured or residual confounding. For instance, two included studies only adjusted for a socio-demographic variable. A higher intake of milk may be linked with other healthy behaviors, such as regular physical exercise, a lower consumption of cigarettes and alcohol, and a lower intake of sugar and processed meat. All of the above behaviors were protective factors of cognitive disorders, but several included studies did not adjust for these possible confounding factors. Moreover, because of the nature of the cross-sectional design in most of the included studies, recall bias and selection bias cannot be avoided. Fourth, only two included studies used FFQ to measure the milk intake, and none of the studies corrected for measurement error. Fifth, whole-fat, low-fat, non-fermented, and fermented milk may have different associations with cognitive disorders; however, these details were poorly reported in the included studies. Finally, although we did not find publication bias in the statistical tests, having only seven included studies limits the interpretability of our results. Additional studies should assess the association between broader dietary patterns rather than sole ingredients (such as milk) and the risk of cognitive disorders with more confounding adjustments.