Asthma remains a global public health problem of epic proportions, especially in children and adolescents. Asthma is a complex respiratory disorder characterized by symptoms that include wheezing, chest tightness, coughing (especially at night) and breathlessness resulting from air flow obstruction triggered by the interaction of genetic and environmental factors( Reference Devereux 1 , Reference Hopkin 2 ).
Diet has been implicated as one of the environmental factors contributing to the pathogenesis of this disease( Reference Eder, Ege and von Mutius 3 , 4 ). In many Westernized countries today, dietary intakes are generally low in fruit, vegetables, wholegrain cereals and fish, and high in fast foods, sweets and salty snacks that are high in saturated fat, sugar and salt and low in fibre and antioxidants. Earlier studies suggested that the low prevalence of asthma symptoms in children from Mediterranean countries compared with English-speaking countries is possibly due to the different dietary patterns of these regions( Reference Asher, Anderson and Stewart 5 , Reference Lai, Beasley and Crane 6 ).
The traditional Mediterranean diet is a collection of eating habits traditionally followed by people in countries around the Mediterranean basin during the early 1960s( Reference Trichopoulou 7 , Reference Simopoulos 8 ). The main characteristics of traditional Mediterranean diets are that all foods consumed are minimally processed, seasonally fresh and locally grown. Furthermore, this pattern is characterized by high intakes of fruits, vegetables, wholegrain cereals, legumes and nuts, low to moderate consumption of dairy products (mainly as cheese and yoghurt), fish, poultry and red wine (usually with meals), fewer than four eggs weekly, small amounts of red meat and liberal use of olive oil as the primary source of added fat( Reference Willett, Sacks and Trichopoulou 9 ). Consequently, this pattern is low in SFA, high in MUFA and n-3 fatty acids, and rich in fibre, vitamins D and E, Mg and antioxidants( Reference Simopoulos 8 , Reference Willett, Sacks and Trichopoulou 9 ).
Most studies on diet and asthma have focused on individual nutrients or food groups and results have been inconsistent. A limitation of these studies is that they fail to take account of the interaction or synergistic effect between dietary components( Reference Hu 10 ). People consume a variety of foods and food groups in a meal, rather than individual nutrients or food groups, and for this reason it has been stated that dietary patterns represent a better picture of the true diet and nutrient intakes of a given population( Reference Sofi, Cesari and Abbate 11 , Reference Trichopoulou and Lagiou 12 ). Hence, it may be more appropriate to better understand the effect of a dietary pattern such as the Mediterranean diet, rather than specific foods or nutrients, on childhood asthma.
Few studies have investigated the association of dietary patterns with childhood asthma and most published reviews are narrative and reflect an opinion rather than an objective analysis of the evidence. To date, only one systematic review and meta-analysis of eight observational studies up to 2013 has been published which examined the influence of the Mediterranean diet on childhood asthma( Reference Garcia‐Marcos, Castro‐Rodriguez and Weinmayr 13 ). Results of the meta-analysis showed that adherence to a Mediterranean diet during childhood might protect against ‘asthma ever’ and ‘current wheeze’. Therefore, the purpose of the present systematic review was to conduct an up-to-date extensive database search to collate and analyse the literature on the effectiveness of a Mediterranean dietary pattern in childhood asthma. The review will add further to the evidence base that links the Mediterranean diet with lower asthma symptoms in children.
Methods
Literature search/data collection
The current systematic review was prepared and reported according to PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines( Reference Moher, Liberati and Tetzlaff 14 ). Relevant studies were identified by systematic search from PUBMED, MEDLINE, EMBASE, Cochrane Library, EBSCO (Ovid) and SCOPUS databases for articles published up to 29 October 2016 and extended back to 1946. Furthermore, in an effort to capture all published studies from a variety of international journals, no restrictions were applied on language, age or publication dates. Additional studies were sought from conference proceedings as well as clinical trial registries (international, European, Australia and New Zealand) in order to identify published/unpublished trials focusing on the Mediterranean dietary pattern and asthma in children. Other citations were identified by hand-searching the reference lists of potential articles. Finally, cross-checks were done to compare relevant articles in our review with previous reviews and systematic reviews.
Search terms were as follows: [‘Mediterranean dietary pattern’ OR ‘Mediterranean diet’ OR ‘Mediterranean-type diet’ OR ‘Mediterranean food pattern’ OR ‘dietary pattern’ AND ‘asthma’ AND ‘children’] OR [‘Mediterranean dietary pattern’ OR ‘Mediterranean diet’ OR ‘Mediterranean-type diet’ OR ‘Mediterranean food pattern’ OR ‘dietary pattern’ AND ‘childhood AND asthma’]. The full search strategy is provided in Appendices 1 to 4.
Study eligibility criteria
Type of studies
Since no randomized controlled trials were identified, the literature search included ‘all study types’, primarily observational studies (intervention, case–control, cohorts and cross-sectional) investigating Mediterranean diet as the primary exposure variable associated with asthma as the outcome in the child population.
Outcomes of interest
Primary outcomes were asthma symptoms, risk and prevalence.
Inclusion criteria
Inclusion criteria were based on PICO (population, intervention, comparator, outcomes)( Reference Liberati, Altman and Tetzlaff 15 ). A study was included when the population under investigation was ‘children’, ‘Mediterranean diet’ was the intervention and ‘asthma symptoms’ was the outcome. A publication was considered for the systematic review if it focused on human subjects, was published in English or any other language with an English translation available, and the dietary intervention was described in sufficient detail, highlighting the main components of the dietary pattern and being evaluated by a questionnaire or diet history. Regarding assessment of adherence to the Mediterranean dietary pattern, only studies using a score and results presented as odds ratio or relative risk and corresponding 95 % confidence interval were considered. Composite outcomes (e.g. allergy) were eligible when diet was part of a multifaceted intervention and results on the Mediterranean diet and asthma in children could be separated.
Exclusion criteria
Exclusion criteria were based on PICO characteristics and reasons for study exclusion were summarized. Types of studies not included were: reviews, systematic reviews, editorials, comments, letters, dissertations, newspapers, case studies and animal studies, as well as those with no abstract, full text or English translation available. Studies conducted in adults, infants and pregnant women were not the population under investigation. In addition, interventions or risk factors other than Mediterranean diet such as supplementation, specific nutrients, intake of single food groups, other dietary patterns (e.g. Western), probiotics, medication, pollution, urban environment, obesity and BMI, as well as studies focusing on maternal nutrition or the role of genetics in asthma, were excluded. Also deemed inappropriate were studies investigating the role of diet in relation to other outcomes, for example dermatitis, eczema, CVD, rheumatoid arthritis, obstructive lung disease, metabolic syndrome, cognitive function, or growth and development.
Selection of studies
The titles and abstracts of all articles identified were reviewed by two authors (M.M.P., C.I.) to assess eligibility and duplicates removed. Papers were screened for relevance based on the information contained in titles and abstracts. Potentially eligible abstracts were selected for full-text reading provided that all inclusion criteria were satisfied. Also, when there was insufficient information in the abstract to warrant exclusion of the article, the full text was retrieved. Finally, full-text articles satisfying inclusion criteria were reviewed by both reviewers and details were extracted. Discrepancies were resolved by discussion and consensus that led to agreement.
Quality assessment
Two authors (M.M.P., N.H.S.) assessed the quality of each study independently using a validated quality assessment tool modified from Zaza et al.( Reference Zaza, Wright-De Aguero and Briss 16 ). The quality assessment tool included a checklist of: description of the study design; description of the study population and how it was selected; how the exposure (Mediterranean diet score) and outcome (asthma in children) were measured; whether the measurements were valid and reliable; the appropriateness of the statistical analysis to answer the research question; and finally how each study controlled for potential bias (i.e. recall bias, measurement bias) or any potential bias that may have been missed. Results were evaluated as the percentage of how many points each study scored out of the maximum number of points it could have achieved, as not every study had the same maximum score due to factors such as different study design or statistical analysis. Finally, scores from both authors were averaged and became the final score. We used a score of 70 % or above to define high study quality and a score below 70 % as low study quality. Any disagreements that arose between the authors were resolved through discussion or with a third author (B.E.).
Data extraction
The data extracted from relevant studies included specific details about the author, year of publication, study design, study name, geographic area, sample size, age of target population, Mediterranean dietary assessment tool, outcome measure, exposure estimate, confounders and main findings. Where available, data analysis results were reported as odds ratios and 95 % confidence intervals, with P value at 5 % significance level.
Results
Electronic search
The database searches identified a total of 435 potential articles, of which 318 original articles remained after removal of duplicates (Fig. 1). One citation was found by cross-checking of reference lists. No trials (published or unpublished) were identified from international clinical trial registries. Of the remaining eligible articles, 319 were screened by two authors (M.M.P., C.I.) independently scanning titles and abstracts. A total of 270 articles were excluded as they did not examine the specific exposure (Mediterranean diet) in relation to the outcome (asthma symptoms) in the study population (children) or satisfy the publication type (original article) (not outcome, n 117; not intervention, n 20; not study population, n 9; no abstract, n 16; reviews, n 107; no English translation, n 1), leaving forty-nine full-text articles to be read for relevance. Of the remaining forty-nine potential studies, thirty-four full texts were considered inappropriate based on the exclusion criteria (not study population, n 6; not intervention, n 12; not study type, n 11; no abstract or full text available, n 5), leaving a total of fifteen studies relevant to the topic. Specifically, eleven cross-sectional, one intervention (with no control group), one case–control and two cohort studies investigating the association between children’s adherence to a Mediterranean dietary pattern and asthma( Reference Akcay, Tamay and Hocaoglu 17 – Reference Silveira, Zhang and Prietsch 31 ).
Flowchart detailing the study search for the present systematic review on the Mediterranean dietary pattern (Med diet) and childhood asthma
Study characteristics
The database search revealed fifteen original studies investigating children’s adherence to the Mediterranean dietary pattern and asthma (Tables 1 and 2). Ten of these studies were conducted in Mediterranean regions including Greece (n 4)( Reference Antonogeorgos, Panagiotakos and Grigoropoulou 18 , Reference Arvaniti, Priftis and Papadimitriou 19 , Reference Chatzi, Apostolaki and Bibakis 22 , Reference Grigoropoulou, Priftis and Yannakoulia 27 ), Spain (n 5)( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 , Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 , Reference Chatzi, Torrent and Romieu 23 , Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Gonzalez Barcala, Pertega and Bamonde 26 ) and Turkey (n 1)( Reference Akcay, Tamay and Hocaoglu 17 ); four were conducted in non-Mediterranean countries, namely Mexico (n 2)( Reference de Batlle, Garcia-Aymerich and Barraza-Villarreal 24 , Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 ), Peru (n 1)( Reference Rice, Romero and Galvez Davila 29 ) and Brazil (n 1)( Reference Silveira, Zhang and Prietsch 31 ); and the ISAAC (International Study on Allergies and Asthma in Childhood)( Reference Nagel, Weinmayr and Kleiner 28 ) involved twenty countries globally. Collectively, 103 248 children and adolescents aged 1 to 19 years participated, with sample sizes ranging from 104 to 50 004. In all these studies children’s dietary intake was evaluated using an FFQ that was either self-administered or completed by parents. Regarding assessment of asthma symptoms, thirteen studies( Reference Akcay, Tamay and Hocaoglu 17 – Reference Nagel, Weinmayr and Kleiner 28 , Reference Silveira, Zhang and Prietsch 31 ) used the ISAAC respiratory questionnaire( Reference Asher, Keil and Anderson 33 ) and two studies used spirometry( Reference Rice, Romero and Galvez Davila 29 , Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 ). A variety of outcomes were measured which included: prevalence of asthma, wheeze, exercise wheeze, any asthma symptoms, night-time cough, sleep disturbance due to wheeze, persistent asthma, atopic asthma, current severe asthma, spirometry (forced expiratory volume 1, forced vital capacity), bronchial inflammation, medication use, number of episodes/attacks, hospital admissions, severity of attacks, asthma control, asthma severity and bronchial hyper-responsiveness. All studies, except for one( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 ), adjusted for confounding factors although the confounding variables differed between studies. Children’s adherence to the Mediterranean dietary pattern was assessed using three indices. Three studies( Reference Chatzi, Apostolaki and Bibakis 22 , Reference de Batlle, Garcia-Aymerich and Barraza-Villarreal 24 , Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 ) used the original Mediterranean diet score developed by Trichopoulou et al.( Reference Trichopoulou, Costacou and Bamia 34 ), seven studies( Reference Akcay, Tamay and Hocaoglu 17 , Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 , Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Gonzalez Barcala, Pertega and Bamonde 26 , Reference Nagel, Weinmayr and Kleiner 28 , Reference Rice, Romero and Galvez Davila 29 , Reference Silveira, Zhang and Prietsch 31 ) used the scoring system by Psaltopoulou et al.( Reference Psaltopoulou, Naska and Orfanos 35 ) and six studies( Reference Antonogeorgos, Panagiotakos and Grigoropoulou 18 , Reference Arvaniti, Priftis and Papadimitriou 19 , Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 , Reference Chatzi, Apostolaki and Bibakis 22 , Reference Chatzi, Torrent and Romieu 23 , Reference Grigoropoulou, Priftis and Yannakoulia 27 ) used the KIDMED index developed by Serra-Majem et al.( Reference Serra-Majem, Ribas and Ngo 36 ). In the studies that used the scoring system based on Psaltopoulou, dietary intake was further categorized into two eating patterns: (i) a ‘Pro-Mediterranean’ pattern which correlated with intakes of fruit, vegetables, fish, cereals, pasta, rice and potatoes; and (ii) an ‘Anti-Mediterranean’ pattern as intakes of milk, meat, eggs, fast foods, soft drinks and butter( Reference Akcay, Tamay and Hocaoglu 17 , Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 , Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Gonzalez Barcala, Pertega and Bamonde 26 , Reference Nagel, Weinmayr and Kleiner 28 , Reference Rice, Romero and Galvez Davila 29 , Reference Silveira, Zhang and Prietsch 31 ).
Characteristics of cross-sectional studies included in the present systematic review on the Mediterranean dietary pattern and childhood asthma
Med diet, Mediterranean diet; Pro-Med, Pro-Mediterranean dietary pattern; Anti-Med, Anti-Mediterranean dietary pattern; Contra-Med, Contra-Mediterranean dietary pattern; BHR, hyper-responsiveness.
* KIDMED index, a Mediterranean diet quality index evaluating food habits in Spanish children developed by Serra-Majem et al.( Reference Serra-Majem, Ribas and Ngo 36 ).
† Mediterranean diet score developed by Castro-Rodriguez et al.( Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 ) also based on the scoring system of Psaltopoulou et al.( Reference Psaltopoulou, Naska and Orfanos 35 ).
‡ Mediterranean diet score based on the score developed by Trichopoulou et al.( Reference Trichopoulou, Costacou and Bamia 34 ).
§ Mediterranean diet score developed by Garcia-Marcos et al.( Reference Garcia-Marcos, Canflanca and Garrido 25 ) and based on the scoring system of Psaltopoulou et al.( Reference Psaltopoulou, Naska and Orfanos 35 ).
║ Mediterranean dietary pattern developed by Gonzalez Barcala et al.( Reference Gonzalez Barcala, Pertega and Bamonde 26 ).
Characteristics of cohort and case–control studies included in the present systematic review on the Mediterranean dietary pattern and childhood asthma
Med diet, Mediterranean diet; –, not indicated in study; Pro-Med, Pro-Mediterranean dietary pattern; Anti-Med, Anti-Mediterranean dietary pattern; BHR, bronchial hyper-responsiveness; FEV1, forced expiratory volume 1 (spirometry measure); FVC, forced vital capacity (spirometry measure); eNO, exhaled nitric oxide level; EBC, exhaled breathe condensate; AOR, adjusted odds ratio.
* KIDMED index, a Mediterranean diet quality index evaluating food habits in Spanish children developed by Serra-Majem et al.( Reference Serra-Majem, Ribas and Ngo 36 ).
† Mediterranean diet score based on the score developed by Trichopoulou et al.( Reference Trichopoulou, Costacou and Bamia 34 ).
‡ Mediterranean diet score developed by Castro-Rodriguez et al.( Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 ) also based on the scoring system of Psaltopoulou et al.( Reference Psaltopoulou, Naska and Orfanos 35 ).
In six of these studies( Reference Akcay, Tamay and Hocaoglu 17 , Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 , Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Gonzalez Barcala, Pertega and Bamonde 26 , Reference Nagel, Weinmayr and Kleiner 28 , Reference Rice, Romero and Galvez Davila 29 ) a score was allocated according to frequency of intake, whereas in the study undertaken by Silveira et al., adherence to the Mediterranean dietary pattern was measured qualitatively. Intake of at least five foods in each dietary pattern at a frequency of at least 3 times/week was classified as ‘yes’( Reference Silveira, Zhang and Prietsch 31 ).
Summarizing the findings of the literature search, the majority of studies reported a beneficial effect of adherence to a Mediterranean dietary pattern on asthma in children. Overall, twelve of these studies reported an inverse association between children’s adherence to the Mediterranean diet and asthma symptoms( Reference Antonogeorgos, Panagiotakos and Grigoropoulou 18 – Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Grigoropoulou, Priftis and Yannakoulia 27 – Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 ) and/or improvement in lung function( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 , Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 ), although the results of two studies were not statistically significant( Reference Chatzi, Apostolaki and Bibakis 22 , Reference Chatzi, Torrent and Romieu 23 ). On the other hand, two studies documented no association between Mediterranean diet and asthma symptoms( Reference Akcay, Tamay and Hocaoglu 17 , Reference Silveira, Zhang and Prietsch 31 ) and one an increase in asthma symptoms( Reference Gonzalez Barcala, Pertega and Bamonde 26 ).
Quality assessment
After assessment of the quality of each study included the averaged quality score was 79 %, the highest score was 88 % and lowest score was 64 % (Table 3). Almost all studies (14/15) were of high quality( Reference Akcay, Tamay and Hocaoglu 17 – Reference Arvaniti, Priftis and Papadimitriou 19 , Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 – Reference Silveira, Zhang and Prietsch 31 ) as they described the study population, measured the exposure and outcome clearly, used appropriate statistical analysis and considered bias in the analysis or discussed any potential bias. The only low-quality study (Calatayud-Saez et al.) suffered from inappropriate statistical analysis and lack of consideration of potential bias such as non-response and dropout bias that led to unreliable results and ambiguity in study population characteristics which made it difficult to compare the results against other studies( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 ). Due to heterogeneity among study methodologies, namely differences between age, exposure measurements and outcome measures, meta-analysis of the fifteen observational studies was deemed inappropriate.
Summary table of quality assessmentFootnote * of relevant studies included in the present systematic review on the Mediterranean dietary pattern and childhood asthma
✓, assessment criteria satisfied; ×, assessment criteria not satisfied.
* Using validated quality assessment tools modified from Zaza et al.( Reference Zaza, Wright-De Aguero and Briss 16 ).
Discussion
The current systematic review identified recent evidence suggesting that adherence to a Mediterranean dietary pattern is inversely related to, and provides a potentially protective effect against, asthma symptoms in children residing in Mediterranean and non-Mediterranean regions( Reference Antonogeorgos, Panagiotakos and Grigoropoulou 18 – Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Grigoropoulou, Priftis and Yannakoulia 27 – Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 ). The majority of studies documented a reduction in wheeze( Reference Arvaniti, Priftis and Papadimitriou 19 , Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 – Reference de Batlle, Garcia-Aymerich and Barraza-Villarreal 24 , Reference Grigoropoulou, Priftis and Yannakoulia 27 , Reference Nagel, Weinmayr and Kleiner 28 ), nocturnal cough( Reference Chatzi, Apostolaki and Bibakis 22 ), exercise wheeze( Reference Arvaniti, Priftis and Papadimitriou 19 , Reference Grigoropoulou, Priftis and Yannakoulia 27 ) asthma episodes( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 ), hospital admissions( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 ), medication use( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 ) and improvement in lung function( Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 , Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 ). A one-unit increase in the KIDMED score was associated with a 14–16 % lower likelihood of having asthma symptoms( Reference Arvaniti, Priftis and Papadimitriou 19 , Reference Grigoropoulou, Priftis and Yannakoulia 27 ) irrespective of potential confounders.
Our qualitative synthesis showed that for ten out of the twelve observational studies documenting a protective effect, statistically significant findings were observed for high adherence to the Mediterranean dietary pattern and asthma symptoms even though heterogeneity existed among study methodologies, namely Mediterranean diet scores, cut-off points, FFQ and asthma outcomes( Reference Antonogeorgos, Panagiotakos and Grigoropoulou 18 , Reference Calatayud-Saez, Calatayud Moscoso Del Prado and Gallego Fernandez-Pacheco 20 , Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 , Reference de Batlle, Garcia-Aymerich and Barraza-Villarreal 24 , Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Grigoropoulou, Priftis and Yannakoulia 27 – Reference Romieu, Barraza-Villarreal and Escamilla-Nunez 30 , Reference Arvaniti, Priftis and Papadimitriou 32 ). These findings coincide with two earlier systematic reviews( Reference Garcia‐Marcos, Castro‐Rodriguez and Weinmayr 13 , Reference Nurmatov, Devereux and Sheikh 37 ) and one meta-analysis( Reference Garcia‐Marcos, Castro‐Rodriguez and Weinmayr 13 ) investigating the role of the Mediterranean diet in relation to childhood asthma that was performed on seven observational studies (cross-sectional)( Reference Arvaniti, Priftis and Papadimitriou 19 , Reference Castro-Rodriguez, Garcia-Marcos and Alfonseda Rojas 21 – Reference Garcia-Marcos, Canflanca and Garrido 25 , Reference Nagel, Weinmayr and Kleiner 28 ) included in our manuscript. Both systematic reviews documented that despite the heterogeneity and inherent limitations of cross-sectional studies included, the evidence is suggestive of a protective effect between the Mediterranean diet and asthma in children, particularly in Mediterranean areas( Reference Garcia‐Marcos, Castro‐Rodriguez and Weinmayr 13 ).
A possible explanation for the beneficial/prophylactic effects observed is that the Mediterranean diet is characterized by low consumption of red meat and saturated fats, high intakes of fruit, vegetables, wholegrain cereals, legumes and fish, and abundance of olive oil, which are rich in antioxidants (vitamins A, C and E, β-carotene, polyphenols, glutathione, lycopene, flavonoids), micronutrients (Mg, Se, Zn) and vitamin D( Reference Willett, Sacks and Trichopoulou 9 , Reference Trichopoulou and Lagiou 12 ). These bioactive compounds may prevent or limit inflammatory responses in the airways by reducing reactive oxygen species and inhibiting lipid peroxidation, thus reducing asthma symptoms. In addition, the high content of long-chain n-3 PUFA found in fish triggers the production of EPA-derived eicosanoids which have anti-inflammatory effects and influence the differentiation of T-lymphocytes modulating immune responses, thereby also improving pulmonary function and decreasing asthma symptoms( Reference Patterson, Wall and Fitzgerald 38 ).
In contrast, two studies reported no association( Reference Akcay, Tamay and Hocaoglu 17 , Reference Silveira, Zhang and Prietsch 31 ) and one an adverse effect( Reference Gonzalez Barcala, Pertega and Bamonde 26 ) of adherence to a Mediterranean dietary pattern in asthmatic children. In the study undertaken by Gonzales Barcala et al.( Reference Gonzalez Barcala, Pertega and Bamonde 26 ), adherence to the Mediterranean diet was associated with an increase in asthma symptoms in children with ‘severe asthma’. This outcome may have been due to a reverse-causal effect. It is well established that the family environment, particularly parents, plays a major role in the dietary habits of young children( Reference Birch and Fisher 39 , Reference Patrick and Nicklas 40 ). More specifically, in those families with children suffering from severe asthma, parents may improve the quality of the child’s diet in an effort to improve overall health. No association was reported by Silveira et al.( Reference Silveira, Zhang and Prietsch 31 ) in the study examining the impact of adherence to a Mediterranean diet in children with persistent and intermittent asthma. A possible explanation may be the coexistence of other environmental factors not analysed in their study that could modify the prevalence or severity of asthma. Possible limitations of that study are the small sample size and no tool such as the Asthma Control Questionnaire was used to evaluate asthma severity. Also, mild, moderate and severe persistent asthma were categorized collectively as ‘persistent asthma’ rather than being assessed separately, which may have led to some effects being overlooked. Medication remains the cornerstone of asthma management. Perhaps medication use might have masked any beneficial effects of the Mediterranean diet. Regarding the study undertaken by Akcay et al.( Reference Akcay, Tamay and Hocaoglu 17 ) in adolescents, no association was reported between adherence to the Mediterranean diet and asthma symptoms. Adolescence is a period when youth defy authority, especially parental advice, gain autonomy and are influenced by peer-group pressure. Adolescents are known to practise poor dietary habits including high intakes of fast foods, sweets and soft drinks, which are rich sources of saturated fats, sugar and salt, and have low intakes of nutritious foods such as fruit, vegetables, cereals and fish that are high in fibre, antioxidants and n-3 fatty acids( Reference Jenkins and Horner 41 , Reference Story, Neumark-Sztainer and French 42 ). These poor dietary habits may have concealed the beneficial effects of adherence to a Mediterranean dietary pattern, hence explaining why no association was documented. In addition, Akcay et al. mentioned that adolescents had a high consumption of pickled and salted foods. Research has shown that a high intake of salt is associated with asthma symptoms( Reference Arvaniti, Priftis and Papadimitriou 32 , Reference Corbo, Forastiere and De Sario 43 ). In addition, adolescents have been known to misreport dietary intake( Reference Forrestal 44 , Reference Livingstone, Robson and Wallace 45 ).
Strengths/limitations of studies reviewed
Limitations must be considered in the present systematic review, some of which may be due to gaps in the literature. There was heterogeneity in study design, age groups considered, time periods of exposure, sample population (asthmatics v. non-asthmatics), outcome measures, outcome parameters and associations studied, thus limiting the possibility of drawing strong and consistent conclusions. The majority of studies reviewed were cross-sectional and hence a cause–effect relationship cannot be determined; however, hypotheses might be suggested for further exploration in more robust clinical trials. Diversity in results documented may be due to differences in the design of FFQ used to collect information on the dietary habits of children; for example, number of food items included and frequency of food consumption categories. Some FFQ were not validated( Reference Akcay, Tamay and Hocaoglu 17 , Reference Chatzi, Apostolaki and Bibakis 22 , Reference Rice, Romero and Galvez Davila 29 ), however research has shown that FFQ produce valid and reproducible estimates of dietary intake of children and adolescents( Reference Rockett and Colditz 46 ). Another weakness in these studies is that dietary questionnaires were self-administered or completed by parents, which may have led to recall or information bias. On the other hand, research often uses parents as proxy reporters for children’s dietary intake( Reference Livingstone and Robson 47 ) and parental report of children’s fruit and vegetable intake is an accurate estimate( Reference Byers, Trieber and Gunter 48 ). Regarding assessment of asthma outcome, respiratory function was evaluated using a questionnaire and by parent report of symptoms. This method is inferior to the use of pulmonary function tests such as spirometry( Reference Katsardis, Koumbourlis and Anthracopoulos 49 ). In studies undertaken in non-English speaking countries, the ISAAC questionnaire had been translated but not validated for that specific population( Reference Akcay, Tamay and Hocaoglu 17 , Reference Antonogeorgos, Panagiotakos and Grigoropoulou 18 , Reference Chatzi, Apostolaki and Bibakis 22 ). Cultural differences cannot be excluded in the assessment of asthma symptoms and report of wheezing, as was delineated by the differences in asthma outcome in the fifteen studies analysed. It is well recognized that, in children younger than 3 years, wheezing is often transient due to respiratory infection and diminishes as the child grows older( Reference Katsardis, Koumbourlis and Anthracopoulos 49 , Reference Heymann, Carper and Murphy 50 ). Regarding differences in Mediterranean dietary scores used to assess adherence, not all scores have been developed for use in children and adolescents( Reference Trichopoulou, Costacou and Bamia 34 , Reference Psaltopoulou, Naska and Orfanos 35 ).
A strength is the extensive literature search focusing exclusively on the Mediterranean dietary pattern as opposed to food groups or nutrients, although publication bias cannot be dismissed. Nevertheless, the studies reviewed were of high quality. In comparison to studies investigating the effect of nutrients or single food groups, the use of a Mediterranean diet score takes account of the synergistic effect or interactions between foods and nutrients( Reference Trichopoulou and Lagiou 12 ), as dietary scores are designed to reflect the whole dietary pattern. In addition, the use of scores improves the statistical power as compared with single nutrients or food groups that might account for small effects. The present systematic review highlights the need for future clinical trials investigating the effect of a Mediterranean dietary pattern on asthma symptoms in order to verify the promising findings documented in the literature. To enhance validity and reliability of the results published, there is a need for homogeneity between study methodologies and statistical tests in evaluation of the data compiled.
Conclusion
In conclusion, the present systematic review assessed evidence for the effectiveness of a Mediterranean dietary pattern in childhood asthma. Although the available evidence is limited, it is nevertheless supportive that adherence to a Mediterranean dietary pattern may reduce asthma symptoms in children. However, well-designed randomized controlled trials are warranted to confirm the prophylactic effects of this dietary pattern. The present findings have important public health implications because they suggest a non-pharmacological means for preventing childhood asthma.
Acknowledgements
Financial support: This research received no specific grant from any funding agency in the public, commercial or not-for-profit sectors. Conflict of interest: None. Authorship: All authors have contributed equally in the literature search, analysis of data published and manuscript writing, and each has seen and approved the final version of the manuscript submitted. Ethics of human subject participation: Not applicable.
Appendix 1
Websites accessed
PUBMED: http://www.ncbi.nlm.nih.gov/ (accessed 29 October 2016).
MEDLINE: http://www.ovidsp.tx.ovid.com (accessed 29 October 2016).
EMBASE: http://www.ovidsp.tx.ovid.com (accessed 29 October 2016).
Cochrane Central Registry of Controlled trials: http://onlinelibrary.wiley.com (accessed 29 October 2016).
EBSCO (CINAHL): http://www.web.b.ebxco.host.com (accessed 29 October 2016).
SCOPUS: https://www-scopus-com (accessed 29 October 2016).
International clinical trials registry: http://www.clinicaltrials.gov (accessed 29 October 2016).
Australian & New Zealand Clinical Trials Registry (ANZCTR): http://www.anzctr.org.au (accessed 29 October 2016).
European Clinical Trials Registry: https://www.clinicaltrialsregister.eu (accessed 29 October 2016).
Appendix 2
Details of PUBMED database search: Medical Subject Headings (MESH) terms
(Mediterranean [All Fields] AND (‘diet’ [MeSH Terms] OR ‘diet’ [All Fields] OR ‘dietary’ [All Fields]) AND pattern [All Fields]) OR (‘diet, mediterranean’ [MeSH Terms] OR (‘diet’ [All Fields] AND ‘mediterranean’ [All Fields]) OR ‘mediterranean diet’ [All Fields] OR (‘mediterranean’ [All Fields] AND ‘diet’ [All Fields])) OR (Mediterranean-type [All Fields] AND (‘diet’ [MeSH Terms] OR ‘diet’ [All Fields])) OR (Mediterranean [All Fields] AND (‘food’ [MeSH Terms] OR ‘food’ [All Fields]) AND pattern [All Fields]) OR ((‘diet’ [MeSH Terms] OR ‘diet ’[All Fields] OR ‘dietary’ [All Fields]) AND pattern [All Fields]) AND (‘asthma’ [MeSH Terms] OR ‘asthma’ [All Fields]) AND (‘child’ [MeSH Terms] OR ‘child’ [All Fields] OR ‘children’ [All Fields])) OR ((((Mediterranean [All Fields] AND (‘diet’ [MeSH Terms] OR ‘diet’ [All Fields] OR ‘dietary’ [All Fields]) AND pattern [All Fields]) OR (‘diet, Mediterranean’ [MeSH Terms] OR (‘diet’ [All Fields] AND ‘mediterranean’ [All Fields]) OR ‘mediterranean diet’ [All Fields] OR (‘mediterranean’ [All Fields] AND ‘diet’ [All Fields])) OR (Mediterranean-type [All Fields] AND (‘diet’ [MeSH Terms] OR ‘diet’ [All Fields])) OR (Mediterranean [All Fields] AND (‘food’ [MeSH Terms] OR ‘food’ [All Fields]) AND pattern [All Fields]) OR ((‘diet’ [MeSH Terms] OR ‘diet’ [All Fields] OR ‘dietary’ [All Fields]) AND pattern [All Fields])) AND (‘Childhood’ [Journal] OR ‘childhood’ [All Fields])) AND (‘asthma’ [MeSH Terms] OR ‘asthma’ [All Fields])) Sort by: Relevance
Appendix 3
PUBMED database search details
Appendix 4
Screenshot of EMBASE database search
