Adherence to Mediterranean Diet Reduces Incident Frailty Risk: Systematic Review and Meta-Analysis.

OBJECTIVES: To conduct a systematic review of the literature on prospective cohort studies examining associations between adherence to a Mediterranean diet and incident frailty and to perform a meta-analysis to synthesize the pooled risk estimates.

DESIGN: Systematic review and meta-analysis.

SETTING: Embase, MEDLINE, CINAHL, PsycINFO, and Cochrane Library were systematically searched on September 14, 2017. We reviewed references of included studies and relevant review papers and performed forward citation tracking for additional studies. Corresponding authors were contacted for additional data necessary for a meta-analysis.

PARTICIPANTS: Community-dwelling older adults (mean age ≥60).

MEASUREMENTS: Incident frailty risk according to adherence to a Mediterranean diet.

RESULTS: Two reviewers independently screened the title, abstract, and full text to ascertain the eligibility of 125 studies that the systematic search of the literature identified, and four studies were included (5,789 older people with mean follow-up of 3.9 years). Two reviewers extracted data from the studies independently. All four studies provided adjusted odds ratios (ORs) of incident frailty risk according to three Mediterranean diet score (MDS) groups (0-3, 4-5, and 6-9). Greater adherence to a Mediterranean diet was associated with significantly lower incident frailty risk (pooled OR = 0.62, 95% CI = 0.47-0.82, P = .001 for MDS 4-5; pooled OR = 0.44, 95% CI = 0.31-0.64, P < .001 for MDS 6-9) than poorer adherence (MDS 0-3). Neither significant heterogeneity (I2  = 0-16%, P = .30) nor evidence of publication bias was observed.

CONCLUSION: Greater adherence to a Mediterranean diet is associated with significantly lower risk of incident frailty in community-dwelling older people. Future studies should confirm these findings and evaluate whether adherence to a Mediterranean diet can reduce the risk of frailty, including in non-Mediterranean populations.