Infertility is a disability in the reproductive system, which can be defined as one year or more of unprotected intercourse without pregnancy (Boivin
et al., 2007). Several factors can lead to men infertility, especially decrease of the semen quality which allocated 25 percent of infertility to itself (Evers, 2002). Furthermore, other factors such as environmental (Merzenich
et al., 2010), genetic (Wong
et al., 2000b), endocrine disorder (Wong
et al., 2000b), age (Wong
et al., 2000b), smoking (Wong
et al., 2000a), heavy alcohol use (Homan
et al., 2007, Wong
et al., 2000b), and nutrition (Vujkovic
et al., 2009a) may affect semen quality. In the past years, we observed a decline in the trend of semen quality as a result of nutrition transition (Pastuszak and Lipshultz, 2012). Studies established that the prevalence of unhealthy diets, characterized by low intakes of fruits and vegetables and high intakes of foods rich in saturated fats, has increased in men within the reproductive age (Jurewicz
et al., 2016b, Vujkovic
et al., 2007). High consumption of red meat, processed meat, full-fat dairy, sweets, trans fatty acid, saturated fatty acid, and soy foods were associated with lower sperm quality (Cutillas-Tolín
et al., 2015). In return, high intakes of fruits, vegetables, skim milk, seafood, and cereal were associated with healthy semen quality (Gaskins
et al., 2012). In addition, some studies investigated the effects of individual nutrients such as folate (Wong
et al., 2002), zinc(Wong
et al., 2002), vitamin C (Piomboni
et al., 2008), carotenoids (Eskenazi
et al., 2005), vitamin E (Suleiman
et al., 1996), and food groups (Attaman
et al., 2012, Mendiola
et al., 2009a) on semen quality.
Since nutrients and food groups are not consumed alone, considering the overall diet in which the combined effects of foods and nutrients are examined, can give us better and more comprehensive view of the relationship between diet and the disease (Hu
et al., 1999). However, we found few studies in relation to dietary patterns and semen quality (Salas-Huetos
et al., 2017). The results of these studies are contradictory, for example some studies indicated that the western dietary pattern which contains high fat dairy, processed meat, refined grains, pizza, and fried food could decrease the sperm concentration (Cutillas-Toĺn
et al., 2015, Liu
et al., 2015b), sperm motility (Eslamian
et al., 2016b), and sperm morphology (Liu
et al., 2015b); whereas, this result was not found in the other study (Oostingh
et al., 2017b). We also observed conflicting results regarding the healthy dietary patterns (Eslamian
et al., 2016a, Gaskins
et al., 2012) such as prudent (Gaskins
et al., 2012) and Mediterranean (Karayiannis
et al., 2016) dietary patterns, which are determined by high level of low fat dairy, poultry, fish, vegetables, whole grains, and fruits. Some studies established that healthy dietary patterns could improve the semen quality (Cutillas-Toĺn
et al., 2015, Jurewicz
et al., 2016c), but others could not show the same result (Karayiannis
et al., 2016, Vujkovic
et al., 2009b).
Several mechanisms are responsible for this result. For instance, healthy dietary patterns with high levels of antioxidants and carotenoids could overcome the stress oxidative and inflammatory marker, so they improved the male fertility (Eskenazi
et al., 2005, Mínguez-Alarcón
et al., 2012). However, the western dietary pattern may have adverse effects on male fertility due to the saturated and trans fatty acids (Afeiche
et al., 2014a, Mendiola
et al., 2009a).
The relationship of dietary patterns and nutrient with male infertility was investigated in only one review without meta-analysis. This study focused on individual nutrients, but not the dietary pattern (Salas-Huetos
et al., 2017). To the best of our knowledge, the present study is the first systematic review and meta-analysis on the relationship between dietary patterns and semen quality.
Materials and Methods
We conducted this systematic review and meta-analysis according to the MOOSE (Meta-Analysis of Observational Studies in Epidemiology). Moreover, this review was registered in the PROSPERO (http:// www. crd. york. ac. uk/ PROSPERO), an international prospective registration website of systematic reviews with the code number of CRD42017068581 (Hosseinzadeh and Hasanizadeh, 2017).
We conducted an electronic search throughout the PubMed, Web of Science, Scopus, and Google Scholar databases to obtain the articles published up to 22 May 2017. To identify the relevant articles, we used medical subject heading terms (MeSH) and non MeSH terms including: “dietary pattern”, “food pattern”, “eating pattern”, “dietary habit”, “eating habit”, “dietary behavior”, “fertility”, “infertility”, “fecundability”, “abnormal sperm”. Moreover, we did not limit our search by language.
Study selection: The process of study selection is summarized in
Figure 1. First, we selected the relevant studies on the basis of articles' titles and abstracts. In the next step, the full text of all related articles was considered by reviewers to extract the studies on dietary pattern and semen quality parameters. All the steps were performed by two authors independently (Hasanizadeh H and Hosseinzadeh M) and any disagreements were discussed and resolved by consensus with the third researcher (Salehi –Abarghuei A).
Finally, we entered the studies that met the following criteria into our meta-analysis: 1) examined dietary patterns using standard instruments such as a 24-h dietary recall, food record, and frequency questionnaire (FFQ); 2) calculated scores of dietary patterns or identified dietary patterns using a statistical method such as principal component analysis, and 3) calculated the mean value or median value for semen quality parameters in each quartile. Studies that considered only nutrients, food groups, or supplements rather than the dietary patterns were excluded.
Data extraction: The following information was obtained from each study: author’s family name, publication year, country where the study was conducted, age of participants, sample size, design, dietary assessment tools, race/ ethnicity, obtained dietary pattern, adjusted mean with standard deviation (SD), adjusted median with range, adjusted ods ratio/relative risk with confidence interval (CI), adjusted β coefficient with CI, and confounding factors adjusted in the analysis.
In order to enhance the comparability of results included in the meta-analysis, we selected the results based on the models with the most number of confounding variables. In addition, we only extracted the patterns of dietary intake identified in multiple studies such as the western and healthy dietary patterns. We also ensured that the selected dietary patterns were similarly based on the loading of food groups.
Quality assessment: In this study, the Newcastle–Ottawa scale was used to assess the quality of studies (Lo
et al., 2014). In this scale a “star system” is used to compare the studies. Studies are given star based on three sections: 1) selection of the study groups, 2) comparability of the groups, and 3) ascertainment of either the exposure or outcome of interest. Then, if the study acquired 3 or 4 stars in the selection part, 1 or 2 stars in the comparability part, and 2 or 3 stars in the outcome/exposure part, it was considered as a high quality study (Lo
et al., 2014).
Data synthesis and analysis: Statistical analyses were conducted by STATA software, version 11.2 (STATA Corp, College Station, TX). Mean value and SD were also calculated as the effect size of different dietary patterns and sperm concentration, motility, and, morphology. We selected the mean value that compared the highest and the lowest categories of adherence to the dietary patterns. In addition, to incorporate the between-study variation, a random effects model was applied to combine the effect sizes. This model takes the study heterogeneity into account.
I
2 and Q statistic was used to evaluate the statistical heterogeneity among studies (Higgins and Thompson, 2002). For the Q statistic, a p-value of < 0.1 was considered as statistically significant heterogeneity (Higgins and Thompson, 2002). Publication bias was evaluated by examination of the funnel plot (Egger
et al., 2008). Sensitivity analysis was also performed to identify whether a specific study or a particular group of studies affected the conclusions (Egger
et al., 2008). P-values less than 0.05 was considered significant.
Results
Literature search: A total of 1313 studies were identified by our electronic search throughout PubMed, Web of Science, Scopus, and Google Scholar prior to 22 may 2017. After removing the duplicates and screening the titles and abstracts of the studies, 53 full-text articles were reviewed. Of these, 44 studies were excluded. Therefore, 9 studies met the inclusion criteria and were included in our systematic review. For the meta-analysis, only 5 (n = 7679 participants) of the 9 articles were included. Of the four excluded studies, two studies had different categories of dietary patterns (Jurewicz
et al., 2016a, Jurewicz
et al., 2016b) and two calculated the odds ratio (Eslamian
et al., 2016a) and Beta coefficient (Oostingh
et al., 2017a)
. So, to obtain more reliable results, we pooled the results from 5 cross-sectional studies involving 7679 participants.
Figure 1 shows the flow chart of the study selection process.
Study Characteristics: Characteristics of all included studies are represented in
Table 1. Studies were conducted in different continents: one study was performed in USA (Cutillas-Tolín
et al., 2015), three in Europe (Gaskins
et al., 2012, Karayiannis
et al., 2016, Vujkovic
et al., 2009a), and one in Asia (Liu
et al., 2015a). All the selected studies had a cross-sectional design. Studies were published during 2009 - 2016. Confounding factors such as total energy intake, body mass index, age, ethnicity, and smoking were adjusted in most of the studies. Dietary consumptions were determined by validated FFQ. In most studies, the western and healthy dietary patterns were derived according to the similarity of foods and food groups highly loaded on each pattern.
According to the Newcastle–Ottawa quality assessment scale, the score quality of the included studies was 9 to 10 and all the studies had a high quality based on the: 1) selection of the study groups, 2) comparability of the groups, and 3) ascertainment of either the exposure or outcome of interest.
Healthy dietary patterns and semen quality parameters: The healthy dietary patterns are highly loaded with vegetables, fruits, whole grains, poultry, and low-fat dairy. Our pool analysis demonstrated no significant relationship between the healthy dietary patterns and the semen quality parameters (
Figure 2) such as sperm concentration (MD = 0.11,
P = 0.11), morphology (MD = -0.02,
P = 0.42) and motility (MD = 0.35,
P = 0.09). Pooled analysis on 5 studies (n = 7679 participants) found a substantial relationship between healthy dietary patterns and sperm motility. However, the sensitivity analysis indicated that results of the sperm concentration and motility could change in the case that we remove Liu et al. study (Liu
et al., 2015a) (MD = 0.23,
P = 0.01, MD = 0.48,
P = 0.03, respectively). In regard to sperm morphology, no heterogeneity was observed among the studies (
P = 0.7, I
2 = 0), whereas, moderate heterogeneity was recognized for sperm concentration (
P = 0.19, I
2 = 33.91). However, high heterogeneity was reported for sperm motility (
P < 0.001, I
2 = 90.1).
Figure 2 Forest plot demonstrates mean differences (represented by the black square) and 95% confidence interval (represented by a horizontal line) for sperm parameters, such as a) concentration, b) morphology, c) motility in participants with the highest and lowest adherence to “Healthy “dietary patterns. Weights are from random effects analysis. The area of the black square is proportional to the specific- study weight to the overall meta-analysis. The center of the diamond displays the pool mean differences and its width shows the pooled 95% CI.
Western dietary patterns and semen quality parameters: The western dietary patterns are determined by high levels of red and processed meat, refined grains, high-fat dairy products, and low intakes of fruits and vegetables. Our meta-analysis showed that higher adherence to the western dietary patterns could significantly reduce the sperm concentration (MD = -0.079,
P = 0.015,
Figure 3). However, we found no significant relationship between this dietary pattern and sperm morphology (MD = 0.2,
P = 0.396,
Figure 3) and motility (MD = 0.01,
P = 0.716,
Figure 3). In terms of sperm concentration and morphology, sensitivity analysis, we found that the results changed significantly by excluding the study conducted by Liu et al. (Liu
et al., 2015a), but exclusion of this study did not have any effect on sperm motility. Moreover, no heterogeneity was observed among the studies conducted on the relationship of this dietary pattern with sperm concentration and motility (
P = 0.68 I
2 = 0 and
P = 0.94 I
2 = 0, respectively). However, high heterogeneity was found among studies in relation to sperm morphology (
P < 0.001, I
2 = 86.99).