Adherence to a Mediterranean-style diet and incident fractures: pooled analysis of observational evidence

Purpose The Mediterranean diet is associated with decreased morbidity and mortality from various chronic diseases. Adherence to a Mediterranean-style diet has been suggested to have protective effects on bone health and decreases the incidence of bone fractures, but the evidence is not clear. We conducted a systematic review and meta-analysis of available observational studies to quantify the association between adherence to a Mediterranean-style diet, as assessed by the Mediterranean Diet Score (MDS), and the risk of fractures in the general population. Methods Relevant studies were identified in a literature search of MEDLINE, EMBASE, Web of Science, and reference lists of relevant studies to October 2016. Relative risks (RRS) with 95% confidence intervals (CIs) were aggregated using random-effects models. Results Five observational studies with data on 353,076 non-overlapping participants and 33,576 total fractures (including 6,881 hip fractures) were included. The pooled fully adjusted RR (95% CI) for hip fractures per 2-point increment in adherence to the MDS was 0.82 (0.71–0.96). Adherence to the MDS was not associated with the risk of any or total fractures based on pooled analysis of only two studies. Conclusion Limited observational evidence supports a beneficial effect of adherence to a Mediterranean-style diet on the incidence of hip fractures. Well-designed intervention studies are needed to elucidate the relationship between adherence to a Mediterranean-style diet and the risk of adverse bone health outcomes such as fractures.


Introduction
The traditional Mediterranean diet which is characterized by high consumption of olive oil, fruits, vegetables, nuts, legumes, and cereals; moderate consumption of fish, poultry, and alcohol; and low consumption of processed food, red meat, dairy, and sweets [1] has been suggested as the optimal diet for the primary prevention of various noncommunicable diseases. To assess the degree of adherence to a Mediterranean diet, the Mediterranean Diet Index was developed [2]; this index and its modification [the alternate or modified Mediterranean Diet Score (MDS)], which can be applied to non-Mediterranean populations, have been shown to have beneficial effects on health outcomes [2]. Adherence to a Mediterranean-style diet has been suggested to have beneficial effects on bone health [3]. To our knowledge, there is no published evidence of a clinical trial which shows a beneficial effect of a Mediterranean-style diet on adverse bone health outcomes such as fractures and osteoporosis. However, a limited number of epidemiological observational studies have suggested a protective effect of a high MDS on the risk of fractures, but the available evidence to date is inconsistent and inconclusive [4,5]. We aimed to clarify the existing evidence by pooling data from

Abstract
Purpose The Mediterranean diet is associated with decreased morbidity and mortality from various chronic diseases. Adherence to a Mediterranean-style diet has been suggested to have protective effects on bone health and decreases the incidence of bone fractures, but the evidence is not clear. We conducted a systematic review and metaanalysis of available observational studies to quantify the association between adherence to a Mediterranean-style diet, as assessed by the Mediterranean Diet Score (MDS), and the risk of fractures in the general population. Methods Relevant studies were identified in a literature search of MEDLINE, EMBASE, Web of Science, and reference lists of relevant studies to October 2016. Relative risks (RRS) with 95% confidence intervals (CIs) were aggregated using random-effects models. Results Five observational studies with data on 353,076 non-overlapping participants and 33,576 total fractures (including 6,881 hip fractures) were included. The pooled fully adjusted RR (95% CI) for hip fractures per 2-point increment in adherence to the MDS was 0.82 (0.71-0.96). Adherence to the MDS was not associated with the risk of any or total fractures based on pooled analysis of only two studies.

3
available published observational cohort studies which have examined the associations between adherence to a Mediterranean-style diet and the risk of fractures in general population settings.

Methods
This review was conducted in line with PRISMA and MOOSE guidelines (Appendices 1, 2). We searched MED-LINE, EMBASE, and Web of Science electronic databases up to October 17, 2016, using free and medical subject headings and combination of key words related to "Mediterranean diet" and "fracture." There were no restrictions on language. Bibliographies of all retrieved articles and other relevant publications, including reviews, were manually scanned for citations missed by the electronic search. Details on our search strategy are presented in Appendix 3. Summary measures were presented as relative risks (RRs) with 95% confidence intervals (CIs). To ensure consistency in the analysis, relevant risk estimates from each study were standardized to compare a two-point increment in the Mediterranean Diet Score (MDS), using methods previously described (Appendix 4). Where studies reported differing degrees of adjustment, the multivariable-adjusted estimate that included adjustment for fracture risk factors was used. Summary RRs were calculated by pooling study-specific estimates using a random effects model. Statistical heterogeneity across studies was quantified using the Cochrane χ 2 statistic and the I 2 statistic. All analyses were performed using STATA release 14 (StataCorp LP, College Station, TX, USA) software.

Results
The search strategy identified 174 potentially relevant articles. After the initial screening of titles and abstracts, 12 articles remained for further evaluation. Following detailed evaluation which included full text reviews, 7 articles were excluded. Five observational (four prospective cohort and one case-control) studies based in general populations were found to be eligible (Appendix 5). Eligible studies were published between 2013 and 2017. The studies involved 353,076 individuals aged 35-80 years at baseline, with 33,576 fractures (including 6881 hip fractures), collected over median or average follow-up periods that ranged from 8 to 15.9 years (Table 1) [4][5][6][7][8]. All five studies reported on hip fractures, with two of them additionally reporting on any or total fractures [4,5]. Only one study reported on other bone health outcomes such as bone mineral density (BMD) and muscle mass [4]. Three studies were based in Europe, one in North America (USA), and one in Asia (China  (Fig. 1). There was evidence of substantial heterogeneity (>70%) among the included studies. Egger's regression test showed no statistical evidence of publication bias (P = 0.603). When analysis was restricted to the two studies that reported on any or total fractures (comprising 91,496 individuals and 28,873 fractures), the corresponding pooled RR was 1.00 (95% CI 0.99-1.02). The absolute risk reduction (ARR) of hip fractures associated with a two-point increment in adherence to the MDS was 0.18%, which translates into a number needed to treat (NNT) of 556 (95% CI 345-2500) to prevent one hip fracture.

Discussion
Emerging evidence from observational cohort studies published only within the last 4 years and involving apparently healthy participants indicates that increasing adherence to a Mediterranean-style diet is associated with lower risk of hip fractures; however, the risk reduction is low. Our results add to the existing evidence that adherence to a Mediterranean diet is protective of adverse health outcomes such as cardiovascular disease, cancer, and neurodegenerative diseases [9], as well as all-cause mortality [9]. Although a limited number of studies have suggested a beneficial effect of the Mediterranean-style diet on the incidence of bone fractures, the results have mostly been inconsistent. By pooling the few published studies on the topic, we have shown that increased adherence to a Mediterranean-style diet is associated with reduced incidence of hip fractures among general population settings. However, pooled analysis of the only two published studies reporting on any or total fractures showed no statistically significant evidence of an association. Feart and colleagues in analysis of a cohort of French elderly people showed no evidence of associations of adherence to a Mediterranean-style diet with risk of any as well as hip fractures; however, their analysis was hampered by the small size [5]. In a recent post hoc analysis of over 90,000 participants recruited in the Women's Health Initiative (WHI) observational study, Haring and colleagues showed that higher adherence to a Mediterranean diet was associated with a reduced risk of hip fractures but not total fractures [4]. In the WHI study, the lack of an association between adherence to a Mediterranean diet and total fracture risk was potentially attributed to the wide variation of fracture types included in the analyses. Outcome events on any fractures from these two studies were self-reported, which increased the likelihood of misclassification bias. It has been suggested that the protective effects of the Mediterranean diet on fracture risk may be via its effect on BMD and muscle mass [4,10]. However, in the WHI study, the authors found no significant changes in BMD and lean body mass over time with adherence to a Mediterranean diet [4]. The Mediterranean diet has been suggested to have a beneficial effect on bone health, and this has been attributed to the antioxidant, anti-inflammatory, and alkalinising properties of the naturally occurring bioactive compounds within this diet [11]. Although the bone protective effects of the Mediterranean-style diet are attributed to the combination of the individual components of the diet, it has been suggested that key components of this diet may be responsible for its protective effect on bone mineral density (i.e., osteoporosis) and fracture occurrence [6]. Our findings have potential clinical implications, as hip fractures (particularly osteoporotic fractures) are one of the leading worldwide causes of disability and morbidity, especially in elderly patients, and increase the burden on health systems. The prevention of fractures is therefore of public health importance. Our ARR estimate of 0.18% as suggested by the pooled analysis translates to about 5,004 people having a two-point increment in adherence to the MDS to prevent one hip fracture in a year. However, this estimate assumes that the effect of adherence to the MDS is constant over time and with hip fracture events occurring at a constant rate over time [12]. The ARR estimate does not seem encouraging; however, it is well known that adherence to the Mediterranean-style diet has beneficial effects on several outcomes. Although bone mass and the risk of fractures are determined by a combination of aging, heritability, mechanical (such as physical activity), and hormonal factors, nutrition plays an important role in bone health. The evidence of a protective effect of nutrition on bone health has mostly been based on specific dietary factors such as calcium, vitamin D, or other isolated nutrients [13,14], though the role of proteins remains controversial [15,16]. The current findings suggest that the combined beneficial effects of the individual dietary components which make up the Mediterranean-style diet may represent an appropriate and feasible dietary intervention for the prevention of bone fractures, rather than the promotion of isolated nutrients. Although residual confounding may have explained part of the findings, at least adherence to a Mediterranean diet did not have a harmful effect on bone health. Given that the Mediterranean diet does not emphasize nutrients that have been suggested to have a beneficial effect on bone health such as calcium or protein intake, it is assuring to see that beyond other well established benefits of a Mediterranean diet; there are no detrimental effects of this diet on bone health.
To our knowledge, this is the first study to evaluate relevant studies that have assessed associations between adherence to a Mediterranean-style diet and the risk of fractures using a systematic meta-analytic approach. We were able to harmonize data from the limited studies conducted on the topic to perform a quantitative analysis, thereby obtaining reliable estimates of the nature and magnitude of the association between adherence to a Mediterranean-style diet and the risk of fractures. There were no relevant clinical trials published on this specific topic; therefore our review was based on only observational evidence. Substantial heterogeneity was observed between contributing studies and which could not be explored because of the limited number of studies. We acknowledge the country-specific characteristics of the Mediterranean dietary pattern, which may explain the different study-specific effect sizes as well as substantial heterogeneity among studies. Indeed, it has been shown that different dietary patterns even exist among Mediterranean countries [17]. Although each eligible study adjusted for a comprehensive panel of confounders including vitamin D, history of fracture, and physical activity (which are major risk factors for hip fracture), the study estimates are still prone to residual confounding because of the observational nature of the study designs. For example, studies did not take into account the mechanisms of fracture occurrence such as falls in their analysis; falls are known to influence hip fracture risk beyond BMD [18]. In addition, adherence to a Mediterranean diet may rather reflect a healthier lifestyle which was not completely captured by confounders that were included in the various analyses. Inadequate data on sex-specific estimates precluded assessment of the associations in males and females separately. However, limited data from the individual studies suggest that the protective effect of adherence to a Mediterraneanstyle diet on hip fractures is more evident in men compared with women. Even though we detected no evidence of publication bias, we were unable to adequately explore for this given that tests for publication bias are unlikely to be useful for analysis involving limited number of studies. Finally, our NNT estimate was calculated from an observational design; ideally, it should have been based on findings from a randomized controlled trial. The findings should therefore be interpreted with caution given these limitations.
In conclusion, available evidence suggests a beneficial effect of adherence to a Mediterranean-style diet on the incidence of hip fractures; however, the pooled risk reduction is low. This review also highlights the limited evidence on the topic in the existing literature and therefore the need for robust well-designed intervention studies to elucidate the relationship between adherence to a Mediterraneanstyle diet and the risk of adverse bone health outcomes such as fractures and osteoporosis.
Author contribution SKK, JAL, MRW, and AWB conducted and designed research; SKK analyzed and interpreted data. SKK and JAL contributed to data acquisition. SKK wrote the paper, and JAL, MRW, and AWB contributed to the interpretation of data. All authors read and approved the final manuscript.

Conflict of interest There is no conflict of interest in this study.
Open Access This article is distributed under the terms of the Creative Commons Attribution 4.0 International License (http:// creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution, and reproduction in any medium, provided you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license, and indicate if changes were made. State the process for selecting studies (that is, screening, eligibility, included in systematic review, and, if applicable, included in the meta-analysis)

Methods
Data collection process 10 Describe method of data extraction from reports (such as piloted forms, independently, in duplicate) and any processes for obtaining and confirming data from investigators

Data items 11
List and define all variables for which data were sought (such as PICOS and funding sources) and any assumptions and simplifications made

Risk of bias in individual studies
12 Describe methods used for assessing risk of bias of individual studies (including specification of whether this was done at the study or outcome level), and how this information is to be used in any data synthesis For each study, the present characteristics for which data were extracted (such as study size, PICOS, and follow-up period) and provide the citations    Assessment of study quality, including blinding of quality assessors; stratification or regression on possible predictors of study results Study quality was assessed based on the nine-star Newcastle-Ottawa Scale using pre-defined criteria namely: population representativeness, comparability (adjustment of confounders), ascertainment of outcome √ Assessment of heterogeneity Heterogeneity of the studies was quantified with I 2 statistic that provides the relative amount of variance of the summary effect due to the between-study heterogeneity and explored using meta-regression and stratified analyses √ Description of statistical methods in sufficient detail to be replicated Description of methods of meta-analyses, sensitivity analyses, meta-regression, and assessment of publication bias are detailed in the methods. We performed random effects meta-analysis with Stata 14 √ Provision of appropriate tables and graphics