During the first one thousand days of life, from conception until the end of the second year, adequate nutritional status has been shown to be critical for both maternal health as well as optimal child’s growth, development and health later in life(Reference Huda, Brodie and Sattar1–Reference Gaillard, Rifas-Shiman and Perng3). Among indicators for nutritional status, accumulating evidence suggested that excessive gestational weight gain (GWG) is a risk factor for many adverse pregnancy outcomes such as gestational diabetes mellitus (GDM), pre-eclampsia, gestational hypertension, depression and cesarean birth(Reference Marchi, Berg and Dencker4,Reference Ma, Tutino and Lillycrop5) . The negative consequences of excessive GWG seem to reach beyond the period of pregnancy and contribute to transgenerational obesity. For instance, women who exceed the recommended weight gain during pregnancy are more likely to enter the next pregnancy with a higher BMI, to deliver heavier babies, who in turn may have higher odds to become overweight or obese adults later in life(Reference Goldstein, Abell and Ranasinha6,Reference Goldstein, Abell and Ranasinha7) . GWG was also found to be a strong determinant to postpartum weight retention (PPWR)(Reference Nehring, Schmoll and Beyerlein8–Reference Fadzil, Shamsuddin and Puteh10), the latter postulated to affect the long-term weight trajectory in women of childbearing age hence increasing the risk of overweight and obesity(Reference Nicklas, Rosner and Zera11,Reference Tahir, Haapala and Foster12) . Together excessive GWG and PPWR constitute potential risk factors for obesity-related disorders including insulin resistance, metabolic syndrome and cardiovascular diseases(Reference Catalano and Shankar13–Reference Wahabi, Fayed and Tharkar15). In view of such adverse effects of excessive GWG and PPWR on the health of the mother and child, prevention programmes and public health nutrition interventions were geared towards the promotion of adequate weight gain during gestation(Reference Hill, Skouteris and Boyle16). To that end, such interventions mainly targeted lifestyle behaviours and dietary patterns during pregnancy as these were shown to be important modifiable risk factors directly linked to GWG and PPWR(Reference Dalrymple, Flynn and Relph17,Reference Vincze, Rollo and Hutchesson18) .
Amongst the various dietary patterns, the Mediterranean diet (MD) has been heralded as one of the healthiest diets in the world(Reference Trichopoulou, Martínez-González and Tong19). This pattern reflects the eating habits of populations living in countries located around the Mediterranean basin(Reference Lăcătuşu, Grigorescu and Floria20). Despite variations in what foods/ food groups constituted a typical MD, such a pattern is generally characterised by high intakes of fruits, vegetables, whole grains, olive oil and legumes; moderate to high consumption of fish; a limited amount of red meat and processed meat and red wine(Reference Widmer, Flammer and Lerman21). Available evidence indicated a multitude of health benefits of the MD including lower risk of obesity(Reference Mancini, Filion and Atallah22), diabetes and associated metabolic abnormalities(Reference Schwingshackl, Missbach and König23), preventions of certain cancers(Reference Schwingshackl, Schwedhelm and Galbete24) and more recently a decrease in the risk of mental disorders, including cognitive decline and depression(Reference Ventriglio, Sancassiani and Contu25). Among the health benefits cited for the MD are those related to pregnancy outcomes and the health of mother and child. More specifically, a higher adherence to the MD was associated with the prevention of GDM(Reference Melero, García de la Torre and Assaf-Balut26,Reference Zhang, Xia and Weng27) , decreasing the risk of compromised fetal growth, prematurity, neural tube defects, asthma, allergy as well as a decrease in the cardio-metabolic risks for both mother and child(Reference Biagi, Nunzio and Bordoni28,Reference Amati, Hassounah and Swaka29) . Pregnancy has been considered as a physiological low-grade systemic inflammatory status with the placenta producing a range of inflammatory cytokines(Reference Parisi, Milazzo and Savasi30). The protective effect of the MD against metabolic abnormalities during pregnancy could be due to the fact that this pattern is rich in antioxidants nutrients (vegetables and fruits) or antioxidant phenols (olive oil and red wine). Further, these antioxidant compounds render adherence to MD to be inversely associated with inflammatory markers (interleukin-6, tumor necrosis factor-α and C-reactive protein)(Reference Bonaccio, Costanzo and Di Castelnuovo31). This aspect of the MD is speculated to protect against gestational metabolic disorders and improve maternal health and neonatal outcomes(Reference Billingsley and Carbone32). In addition, a few studies have addressed the association of adherence to the MD during pregnancy with GWG and PPWR, and their results suggested a protective effect of this dietary pattern(Reference H. Al Wattar, Dodds and Placzek33–Reference Cano-Ibáñez, Martínez-Galiano and Luque-Fernández35). It is important to note that these studies were carried out among women living in various countries in Europe; therefore, their findings may not be directly extrapolated to other contexts and settings such as countries of the Middle East.
Following the discovery of the oil, the United Arab Emirates (UAE), similar to many Gulf Cooperation Council countries in the Middle East, experienced rapid economic growth that precipitated drastic shifts in diet and lifestyle habits(Reference Mokdad, Jaber and Aziz36,Reference Salti37) . Such shifts were characterised by increased consumption of energy-dense, high fat and high sugar foods coupled to high rates of sedentary and physical inactivity. Concomitant to these lifestyle changes are surging rates of obesity and non-communicable diseases in the Gulf Cooperation Council countries(Reference Mokdad, Jaber and Aziz38–Reference Mahmoud and Sulaiman41). A cross-sectional study of 3064 adults, in the UAE, showed that the prevalence of overweight and obesity were 43·0 and 32·3 %, respectively, with women having significantly higher obesity rates than males (36·5 % v. 31·4 %)(Reference Sulaiman, Elbadawi and Hussein42). This alarming situation calls for evidence-based and culture-specific interventions aiming to prevent obesity, especially among women. In this context, GWG and PPWR seem to present plausible targets especially among women of childbearing age. In the UAE, limited evidence existed on GWG, PPWR and their associated factors.
Given the significant adverse effects of excessive GWG and PPWR on maternal and child health and the potential protective effect of the MD, this study aims to explore the associations of the MD with GWG and PPWR among women in the UAE. Using data from a prospective 2-year mother and child nutrition cohort conducted in the UAE (the Mother and Infant Study Cohort (MISC)(Reference Radwan, Hashim and Shaker Obaid43), the specific objectives of this study are to (1) examine adherence to the MD during pregnancy, (2) characterise GWG and PPWR and (3) investigate the associations between adherence to the MD during pregnancy with GWG and PPWR.
Materials and methods
Study design
This work was part of the Mother and Infant Study Cohort (MISC) in the UAE. Details about the protocol and data collection of this cohort were described elsewhere(Reference Radwan, Hashim and Shaker Obaid43,Reference Naja, Nasreddine and Yunis44) . The MISC cohort was a 2-year longitudinal prospective study, during which pregnant women were recruited in their third trimester, between December 2015 to December 2017. Over the duration of the study, data collection was conducted at six time points (visits): third trimester, at delivery and 2, 6, 12 and 18 months postpartum. The protocols used in the MISC cohort were reviewed and approved by the Research Ethics Committee, University of Sharjah (REC/14/01/1505) and by Al Qassimi Clinical Research Centre Ethical Research Committee (REC Reference Number: 215 12015–03), by the Ministry of Health Ethical Research Committee(R02) and by Dubai Health Authority (DSREC-0/2016). Moreover, the protocol of the research study was registered at ClinicalTrials.gov, identifier # NCT03578094. All MISC participants provided a written signed consent form. Subjects were reassured of anonymity and confidentiality, that they can withdraw at any time and that their participation in this study/lack of will not affect their medical care.
For the purpose of this study, data were used from the following time points: visit 1 (third trimester, after the 27th week of gestation, n 243), visit 2 (hospital medical records, n 243), visit 3 (2 months postpartum, n 150) and visit 4 (6 months postpartum, n 131).
Study population
Subjects were recruited using a convenient sampling approach by trained research assistants from three main public governmental hospitals and were followed up in seven Primary Health Care clinics and Mother and Child Centers in the Emirates of Sharjah, Dubai and Ajman. To be eligible to participate, women aged 19 to 40 years and pregnant during their third trimester, with a singleton, and not suffering from any chronic conditions.
Study protocol and data collection
During visit 1, information about the following characteristics was collected: age (in years), nationality (Emirati or Arab), occupation (employed v. housewife), education (intermediate or less, high school/technical diploma and university), parity (primiparous v. multiparous), income (< 10 000 Arab Emirates Dirham (AED) or ≥ 10 000 AED), presence of GDM, number of children (0, 1 or 2, > 2), maternity Leave (49 d (˜ 1·5 months), 70 d (˜ 2·5 months), > 3 months, Do not know/not applicable. From the medical charts of the participants (visit 2), information regarding pre-pregnancy weight (kg), gestational age (weeks), weight at delivery (kg) and GDM (yes, no) were obtained. For the diagnosis of GDM, the National Institute for Health and Care Excellence Diabetes in Pregnancy criteria were used(45). At visit 3 (2 months postpartum) and visit 4 (6 months postpartum), mothers were asked whether their infants were breast-feeding (yes, no). During these two visits (3 and 4), the height and weight of the mother were obtained following standard protocols, using Seca 220 Telescopic Measuring Rod for Column Scales. Measurements were recorded to the nearest 0·1 cm and 0·1 kg for height and weight, respectively.
GWG was calculated as the difference between the pre-pregnancy weight of the mother and the last weight recorded before delivery, both of which were extracted from the medical records. GWG was classified into three categories (insufficient, adequate or excessive) taking into account the participants’ prepregnancy BMI as specified by the Institute of Medicine in 2009(46). Gestational weight gain rate (GWGR, kg/week) is defined as the rate of weight gain per week following week 12 of gestation(46). Given that, in this study, there was no information available regarding weight gain at 12 weeks, it was assumed that women have gained 1 kg during their first trimester. This assumption was based on the Institute of Medicine weight gain recommendation of 0·5–2 kg over the first trimester(46). Accordingly, the GWGR was calculated using the pre-pregnancy weight, gestational age and weight before delivery, as per the following equation:
$${\rm{GWGR}} = {{\left( {{\rm{weight}}\;{\rm{before}}\;{\rm{delivery}}\;\left( {{\rm{kg}}} \right) - 1} \right) - {\rm{pre}}\;{\rm{pregnancy}}\;{\rm{weight}}\;\left( {{\rm{kg}}} \right)} \over {{\rm{gestational}}\;{\rm{age}}\;\left( {{\rm{weeks}}} \right) - 12}}$$
PPWR at 2 months (PPWR2, kg) was calculated as the difference between the mother’s weight measured at visit 3 (2 months postpartum) and pre-pregnancy weight (kg) (n 150). Similarly, PPWR at 6 months postpartum (PPWR6, kg) was calculated as the difference between the weight obtained at visit 4 (6 months postpartum) and pre-pregnancy weight (kg) (n 131).
Assessment of dietary intake
Maternal dietary intake was examined during the third trimester (at visit 1), using a culture-specific semi-quantitative FFQ. The FFQ assessed habitual intake over the three months prior to visit 1. Within the FFQ, the food list consisted of eighty-six items and included single food items as well as composite (mixed) traditional recipes. Subjects were asked to indicate the number of times each food item was consumed either per day, per week, per month or never. In order to estimate the portion size, the participants were asked to estimate the amount of the food eaten either in grams or in comparison with reference portion sizes listed for each item. A two-dimensional visual aid(Reference Posner, Smigelski and Duggal47) as well as pictures of common household measures, measuring cups and spoons were also used to assist the participants in estimating the portion size(48). The indicated frequency of each consumed food and beverage was converted to daily intake. The food composition database of the Nutritionist Pro. software (Axxya Systems LLC, Sta_ord) was then used to compute the total energy intakes of participants. A copy of the FFQ used in this study is presented in appendix A.
Mediterranean diet scores
In order to examine the adherence of study participants to the MD, scores for two MD indexes were calculated: the alternative Mediterranean Diet (aMED) and the Lebanese Mediterranean Diet (LMD) scores. In this study, no specific questionnaire was used for the assessment of adherence to the MD. Rather, data derived from the FFQ, obtained during visit 1, were used to examine various food groups intakes and to calculate the MD sores. More specifically, for each of the food groups included in the selected MD scores, corresponding items of the FFQ were identified and their dietary intakes were summed to obtain the food group’s intake. For example, for the food group ‘Whole Grains’, the intakes of the following items were added ‘Whole wheat bread, bran or whole grain breakfast cereals, burghul’. The detailed description of food groups and their corresponding FFQ items is presented in Supplementary Table 1. Following the estimation of the food groups intakes, the scores of the MD were calculated as per their corresponding references(Reference deKoning and Anand49,Reference Naja, Hwalla and Itani50) and as described in the next paragraph.
The aMED, a widely used score for the MD, was adapted from the traditional MD score developed by Trichopoulou et al. (Reference deKoning and Anand49). It consisted of eight components: whole grains, MUFA:SFA ratio, fruits, vegetables, legumes, nuts, fish and sea food and red and processed meats. The daily intake of each component was divided into two groups based on the median cut-off. For the first seven components, 0 or 1 points were given for participants consuming below or above the median, respectively. For red and processed meats, a presumed harmful food, a reverse scoring was assigned. Total aMED score was computed by adding the points obtained from the eight components. A higher score indicates higher adherence. In addition to the aMED, the LMD score was used to examine adherence to the MD. The LMD is an index for a variant of the MD which originated from Lebanon(Reference Naja, Hwalla and Itani50), a country belonging to the same region (Middle East) as the UAE, where the study is conducted. The LMD score consisted of nine components: whole grains/bulgur, olive oil, fruits, dried fruits, vegetables, starchy vegetables, legumes, eggs, as well as milk and dairy products. For each component, women whose consumption was in the first, second and third tertile were assigned 1, 2 and 3 points, respectively. Points are summed to obtain the total LMD score whereby higher scores simply better adherence to the MD Pattern.
Statistical analysis
Data analysis was conducted using the Statistical Package for Social Sciences (SPSS) software version 25 for Windows. Descriptive statistics were presented as frequency and percentage for categorical variables and mean ± standard error (se) for continuous variables. Adherence to the MD scores was classified as low or high based on the upper tertile cut-off. Similarly, high and low PPWR correspond to PPWR above or below the upper tertile cut-off, respectively. Simple and multiple multinomial logistic regression analyses were conducted to examine the associations between adherence to the MD (as per the aMED and LMD scores) with GWG and GWGR. As for the associations of adherence to the MD with PPWR2 and PPWR6, simple and multiple logistic regression analyses were used. In all regression analyses, adherence to the MD was considered as independent variable, while GWG, GWGR, PPWR2 and PPWR6 were the dependent variables. For the analysis, the sample of 243 participants was used to examine adherence to the MD during pregnancy, the sample of 150 women who attended visit 3 was used to examine PPWR at 2 months and its association with MD and the sample of 131 women who attended visit 4 was used to examine PPWR at 6 months and its association with MD. Statistical significance was considered at P-values lower than 0·05.
Results
Table 1 describes the baseline characteristics of the study population, including age, education level, employment status, income, nationality, maternity leave, presence of GDM and breast-feeding at 2 and 6 months. (Table 1).
Table 1. Baseline characteristics of the study population
(Numbers and percentages, n 243)
AED, United Arab Emirates Dirham.
The main outcomes of this study including GWG, GWGR, PPWR2 and PPWR6 are described in Table 2. In the total sample, mean GWG was 12·0 kg ± 0·5 with over half of the populations belonging to the excessive GWG category, 18·3 % had adequate GWG and 24·2 % had insufficient GWG. For GWGR, the mean was 0·5 kg/week ± 0·02. GWGR was classified as insufficient, adequate and excessive for 35 %, 25·4 % and 39·6 % of participants, respectively. Mean PPWR2 was 5·4 kg ± 0·7 and decreased at 6 months postpartum with mean PPWR6 at 3·8 kg ±0·7. The percentage of women retaining any weight at 2 months postpartum was 76·7 % and those retaining at least 5 kg was 50·7 %. These figures at 6 months postpartum were 73·3 % and 45 %, respectively. (Table 2).
Table 2. Gestational weight gain (GWG), gestational weight gain rate (GWGR), and postpartum weight retention (PPWR) at 2 and 6 months among study participants
(Numbers and percentages; mean values and standard deviations)
Figure 1 illustrates the association between GWG and PPWR at 2 and 6 months postpartum. Positive correlations were observed at both points in time where higher GWG was associated with higher PPWR (age-adjusted R2 = 0·21 and 0·22, respectively, at 2 and 6 months postpartum, P-values < 0·001). (Fig. 1).
Fig. 1. Association between gestational weight gain (GWG) and postpartum weight retention (PPWR) at 2 and 6 months postpartum among study participants. 
, Crude; 
, age adjusted
Table 3 describes adherence to the MD in the study population as per the aMED and LMD scores. Common components between aMED and LMD scores include whole grains/bulgur, vegetables and legumes. Considering both scores, the main contributors to higher adherence to the MD were vegetables and olive oil/(MUFA/SFA). On the other hand, legumes and whole grains were among the lowest contributors. The distribution of the total aMED and LMD scores in the study population followed a bell-shaped curve as presented in Fig. 2.
Table 3. Adherence to the Mediterranean diet (MD) in the study population, as per the alternate Mediterranean diet (aMED) and Lebanese Mediterranean diet (LMD) scores
(Mean values and standard errors, n 243)
* A higher score was assigned to a lower intake: the corresponding % contribution refers to the contribution of a lower intake of red and processed meat.
Fig. 2. Distribution of the alternate Mediterranean diet (aMED) and Lebanese Mediterranean diet (LMD) scores in the study population. (n 243).
Figure 3 shows the adherence to the MD (defined as having the corresponding total score above the upper tertile cut-off for aMED and LMD) by categories of GWG, GWGR and PPWR at 2 and 6 months among study participants. A similar trend is shown for both MD scores, whereby for each outcome the proportion of women who are adherent to the MD is highest for those belonging to adequate GWG/GWGR and low PPWR2 and PPWR6. (Fig. 3).
Fig. 3. Adherence to the Mediterranean diet (MD) (according to alternate Mediterranean diet (aMED) and Lebanese Mediterranean diet (LMD) scores) by categories of gestational weight gain (GWG), gestational weight gain rate (GWGR) and postpartum weight retention (PPWR) at 2 and 6 months among study participants.
The association between adherence to the MD (as examined by aMED and LMD scores) and GWG, GWGR among study participants was examined using simple and multiple multinomial regression analysis and results are summarised in Tables 4 and 5. The multiple regression models, in addition to age and energy intake variables, significantly associated with MD scores at the univariate level were included.
Table 4. Association between adherence to the Mediterranean diet (MD) (as examined by alternate Mediterranean diet (aMED) and Lebanese Mediterranean diet (LMD) scores) with Gestational weight gain (GWG) and gestational weight gain rate (GWGR) among study participants
(Odds ratio and 95 % confidence intervals)
Data are presented as Odds ratio (95 % CI).
* Adjusted for age, energy and number of children.
† Reference categories are adequate GWG and adequate GWGR.
‡ Low and high adherence to MD scores are defined as having the total score below or above the corresponding upper tertile cut-off, respectively.
Table 5. Association between adherence to the Mediterranean diet (MD) (as examined by alternate Mediterranean diet (aMED) and Mediterranean diet (LMD) scores) with high postpartum weight retention (PPWR)* at 2 and 6 months among study participants
(Odds ratio and 95 % confidence intervals)
Data are presented as odds ratio (95 % CI). PPWR2, 2 months postpartum weight rate; PPWR6, 6 months postpartum weight retention.
* High PPWR is defined as belonging to the third tertile of postpartum weight retention.
† Adjusted for age, energy and number of children.
‡ Low and high adherence to MD scores are defined as having the total score below or above the corresponding upper tertile cut-off, respectively.
After adjustment for potential confounders, high adherence to the MD, as examined by both scores, was associated with lower odds of excessive GWG. High adherence to the MD, according to the LMD, was associated with 66 % lower odds of excessive GWGR (Table 4).
Table 5 summarises the associations of adherence to the MD during pregnancy (as examined by aMED and LMD scores) with high PPWR2 and PPWR6 among study participants. The results of the adjusted models showed that while only a high aMED score was significantly associated with PPWR at 2 months, both scores showed significant associations with PPWR at 6 months. (Table 5).
Discussion
This study examined adherence to the MD during the third trimester of pregnancy and its association with GWG and weight retention at 2 and 6 months postpartum among women participating in the MISC cohort in the UAE. The results of this study showed that higher adherence to the MD during pregnancy was associated with lower odds of both insufficient and excessive GWG. Similarly, a negative association was observed between high adherence to the MD during pregnancy with weight retention at 2 and 6 months postpartum. In the study population, while an adequate intake of vegetables and olive oil/ (MUFA/SFA) was the main contributors to adherence to the MD, those of legumes and whole grains were among the lowest contributors. Among participants, high prevalence rates of excessive GWG and GWGR during pregnancy were observed. Furthermore, at 2 months postpartum, 3 out of 4 women retained weight (> 0 kg) and 1 in 2 women retained substantial weight (> 5 kg). Similar estimates for weight retention trends were observed at 6 months postpartum. The findings of this study also showed a strong association between GWG and PPWR at 2 and 6 months.
The findings of this study showed that over 80 % of women had inadequate GWG, and over 50 % of participants retained 5 kg or more at 2 and 6 months postpartum. Such alarming rates are higher than those reported by the Mother and Infant Nutrition Assessment (MINA) cohort conducted in Lebanon and Qatar (64 % had inadequate GWG and 36 % retained more than 5 kg at 6 months postpartum)(Reference Nasreddine, Ayoub and Abbas51). Compared with other countries around the world, the mean PPWR at 6 months in the study was also higher than those reported from Taiwan (2·1 kg)(Reference Shao, Hwang and Huang52), Malaysia (3·12 kg)(Reference Fadzil, Shamsuddin and Puteh10) and the USA of American (3·3 kg)(Reference Walker53). The excessive GWG and PPWR observed in this study are concerning, especially in light of the mounting evidence for their association with a higher risk of lifetime obesity, especially central obesity among women of childbearing age(Reference Smith, Lewis and Caveny54). Therefore, it is critical to research the context and culture-specific determinants and the factors associated with excessive GWG and PPWR in order to develop evidence-based interventions and public health nutrition programmes.
In this context, the results of this study revealed a negative association between higher adherence to the MD during pregnancy with insufficient as well as excessive gestational weight gain. Despite the considerable variability in the definition and assessment of MD across the literature, the protective effect of adherence to the MD against suboptimal GWG was reported in many previous studies. For instance, a study among pregnant women in Spain showed that one-point increase in adherence to the MD was associated with higher odds of appropriate weight gain among participants (OR: 1·39, CI: 1·06, 1·82)(Reference Silva-del Valle, Sánchez-Villegas and Serra-Majem34). Another more recent study from Spain also indicated a negative association between adherence to the MD and GWG in a sample of 503 women(35). Furthermore, the results of PRINCESA cohort showed that higher adherence to the MD was protective against an insufficient (OR: 0·63; CI:0·42, 0·95) and excessive GWG (OR: 0·62; CI, 0·41, 0·94) among 660 pregnant women in Mexico(Reference Ancira-Moreno, Vadillo-Ortega and Rivera-Dommarco55). Adherence to MD like patterns has shown similar results, whereby in a sample of 538 pregnant women in Urban South Africa, adherence to a dietary pattern rich in whole grains, legumes and vegetables was associated with a 30 % decrease in the odds of excessive weight gain(Reference Wrottesley, Pisa and Norris56). Such an observed association between the MD and GWG was not only described in observational studies but was also reported in intervention studies, whereby a multicentre trial in the UK showed that simple MD dietary advice was effective in reducing GWG among 593 pregnant women(Reference H. Al Wattar, Dodds and Placzek33). Furthermore, a meta-analysis including four clinical trials (2277 patients) and aiming to evaluate the effect of adherence to the MD during pregnancy on weight gain during pregnancy concluded that the MD was associated with a reduced GWG(Reference Zhang, Xia and Weng27). This effect of adherence to the MD on optimal GWG could be attributed to the anti-obesity effect of this pattern and its constituents. Using either of the two scores considered in this study, the MD included whole grains, olive oil and MUFA, fruits, vegetables and legumes. These constituents provide a considerable quantity of dietary fiber. The latter has been associated with increased satiety via several physiological mechanisms, including longer mastication time, increased gastric detention, slower gastric emptying and a greater release of cholecystokinin(Reference Buckland, Bach and Serra-Majem57). As for the fat content, the MD does not seem to fit the definition of low-fat diets as it contains 30–35 % of energy as fat; however, it is argued that the quality of the fat in the MD that seems to exert a protective effect against excessive weight gain. More specifically, the MD has a low content of saturated and trans fats and is high in monounsaturated fats (approximately 67 % of fat energy) as found in oleic acid in olive oil(Reference Buckland, Bach and Serra-Majem57). Furthermore, fruits and vegetables, integral components of the MD, are rich sources of polyphenols such as hydroxycinnamic derivatives, quercetin, resveratrol, oleuropein and hydroxytyrosol. These polyphenols are known for their antioxidant and anti-inflammatory activities, which are consequently implicated in preventing excessive weight gain and obesity(Reference Nani, Murtaza and Sayed Khan58). It remains important to note that the beneficial effect of the MD on GWG may be greater than the sum of the benefits derived from each of the components. For instance, in a systematic review of randomised trials where the effect of polyphenols supplementation on weight gain and obesity was considered in isolation of a whole dietary pattern, the results did not support polyphenol supplementation as a complementary strategy for enhancing weight loss(Reference Llaha and Zamora-Ros59). Therefore, it seems plausible to suggest that the beneficial effects of the MD could be attributed to the potential synergistic actions of its multiple ingredients(Reference Trichopoulou, Bamia and Trichopoulos60–Reference Tapsell62).
The findings of this study indicated that in addition to its association with GWG, a high adherence to the MD during pregnancy was associated with significantly lower odds of high PPWR at 2 and 6 months postpartum. This finding suggested that the protective effect of adherence to the MD during pregnancy against excessive weight gain extended well beyond delivery to reach 2 and 6 months postpartum. Few studies have examined the impact of dietary intake during pregnancy on PPWR. A prospective cohort in Singapore, The Growing Up in Singapore Towards healthy Outcomes (GUSTO) study, showed that a lower diet quality during pregnancy, as examined by the Healthy Eating Index was, in fact, associated with higher odds of substantial PPWR (5 kg or higher)(Reference Loy, Cheung and Marjorelee63). A possible explanation of the association of adherence to the MD with lower odds of high PPWR could be explained by the fact that women who were adherent to the MD during their pregnancy would most likely also be adherent to this diet after delivery and hence would benefit from its protective effect against PPWR then. In fact, available evidence indicates that dietary patterns tend to be stable across time(Reference Borland, Robinson and Crozier64). Another explanation of the protective effect of adherence to the MD during pregnancy against high PPWR could relate to the strong association between GWG and PPWR. More specifically, the adherence to the MD protect seems to protect against excessive GWG, the latter being closely associated with PPWR. In fact, in this study, a strong positive association was found between GWG and PPWR at both 2 and 6 months. This association has been repeatedly reported in the literature. In fact, GWG has been postulated as the single most important factor in predicting PPWR(Reference Shao, Hwang and Huang52,Reference Abebe, Von Soest and Von Holle65–Reference Koh, Ee and Malhotra67) . A closer examination of the various contributors to the MD in the study population indicated that, while an adequate intake of vegetables and olive oil/ (MUFA/SFA) were the main contributors, legumes and whole grains were among the lowest contributors. Aside from this study, little is known about the intake of legumes in the UAE. As for whole grains, according to Afshin et al., in the UAE, the average intake of whole grains falls below 40 g/d which is far lower than the recommended 125 g/d(Reference Afshin, Micha and Khatibzadeh68).
In light of the findings of this study and others underscoring the important role of MD in ensuring adequate GWG and PPWR rates, it seems sensible to recommend integrating the elements of this dietary pattern into national health programmes aiming to promote optimal weight gain among pregnant women. Such programmes may target clinical practice as well as public health interventions. More specifically, based on the findings of this study, health care professionals could be trained to include the MD recommendation within their routine antenatal care. For example, obstetricians, who have frequent encounters with pregnant women, are ideally situated to promote and support a higher adherence to the MD in this population group. At the public health level, although there exist no national dietary guidelines for pregnant women in the UAE, the available evidence of this research could justify integrating the MD within such guidelines once developed in the future. Among the factors that could facilitate the promotion of the MD among pregnant women in the UAE is the fact that the UAE hosts a multitude of nationalities, not the least of which are those coming from Mediterranean countries(69), making up around 10 % of the total population in the UAE. As such, residents of the UAE may be familiar with many culinary habits native to these Mediterranean countries and which are part of the MD.
Strengths and limitations
This is the first study to investigate the effect of adherence of the MD during pregnancy on both GWG as well as PPWR at two different time points after delivery. The cohort design of the study allows causation inference and minimises the effect of recall bias, whereby the exposure (adherence to the MD) was measured before the outcome (PPWR). The measurements of the maternal weight during pregnancy and after delivery were conducted by trained research assistants, rather than being reported or recalled by the mother. Such an approach enhances the accuracy of the data and improves the validity of the results. That said, the findings of this study ought to be considered in light of a few limitations. First, in this study, the number of participants dropped significantly between visit 1 (n 243), visit 2 (n 150) and visit 3 (n 131). High dropout rates are a common concern in cohort studies. In this study, every effort was exerted to retain subjects, whereby the timing of the visits was scheduled based on the participants’ preference. In case of no-show, the research assistant would conduct a minimum of multiple attempts before considering the participant as a drop-off. Second, in this study, adherence to the MD was not examined using a questionnaire specifically designed and validated for this purpose, rather it was estimated using an FFQ that assessed overall dietary intake. The use of short questionnaires as opposed to using long FFQ has been advocated for clinical and epidemiological research purposes as they are less burdensome and time consuming and consequently leading to a higher response rate(Reference Gnagnarella, Draga and Misotti70). In the context of this cohort study, the breadth of dietary exposures of interest favoured the use of FFQ. It is important to note that the FFQ method of dietary assessment relies considerably on the memory as well as on the motivation of the participants, and as such could be subject to recall error. However, the FFQ remains the most suitable dietary data assessment tool in epidemiological studies as it provides information on an individual’s habitual diet and dietary patterns over long periods of time and also allows ranking of individuals according to intake(Reference Willett71), which in the case of this study was applied to rank individuals according to their MD scores. Third, though not required for the calculations of the MD scores considered in this study, there are characteristics of certain food groups that could affect their nutritional value and ultimately their effect on weight gain, which were not accounted for in this study. An example of such foods is olive oil, whereby refined olive oil is a poor source of vitamins, polyphenols and phytosterols, while extra virgin olive oil contains the highest level of polyphenols(Reference Gorzynik-Debicka, Przychodzen and Cappello72). Taking such characteristics of olive oil into account in future studies of the health benefits of the MD is therefore warranted. Lastly, the interview-based approach in data collection could have led to a social desirability bias, as is the case of most questionnaire-based epidemiological research(Reference Okamoto, Ohsuka and Shiraishi73). In this study, the research assistants were thoroughly trained to maintain a neutral attitude and avoid judgemental comments(Reference Radwan, Hashim and Shaker Obaid43).
Conclusions
GWG and PPWR are critical factors that not only play a pivotal role in fostering optimal nutrition during the critical 1000 days’ period of life but are also implicated as causal factors for obesity among women. In a country with soaring rates of overweight and obesity, such as the UAE, it is critical to unearth the modifiable risk factors for excessive GWG and PPWR. In this context, the findings of this study confirmed the high rates of excessive GWG and high PPWR among women in the UAE and showed that a higher adherence to the MD during pregnancy protected against excessive weight gain during pregnancy. Furthermore, such an effect seemed to extend beyond the pregnancy period, whereby adherence to the MD was found to be associated with lower odds of PPWR. These findings lay the foundation for the development of evidence-based and culture-specific interventions and programmes for the promotion of the MD. According to the findings of this study, low intake of legumes and whole grains could be plausible targets for the interventions and programmes aiming to increase the adherence to the MD among women in the UAE. Future follow-up studies are warranted to examine the effect of maternal adherence to the MD on infants’ health and reveal the long-term health implication for both the mother and child.
Acknowledgements
We would like to acknowledge all the research assistants, the mothers, nurses and hospital and health clinic staff as well as the research assistants who participated in this study. We are grateful to Medala for their support and help and very special thanks to the Sharjah Baby-Friendly Office director and staff for their support of the MISC research project.
This research was funded by Al Jalila Foundation (AJF 201510) and the Vice-Chancellor Research and Graduate Studies Office/University of Sharjah grant no (1501057003-P). The study sponsors had no involvement in the research design, data collection, interpretation of the data or publication.
Study concept and design: F. N. and H. R.; acquisition of data: H. R., M. H., H. A. G. and R. S. O.; analysis and interpretation of data: F. N., N. A., R. O. and H. H. drafting of the manuscript: F. N., H. R. and M. H. critical revision of the manuscript for important intellectual content: all authors.
The authors declare no conflict of interest.
Supplementary material
For supplementary materials referred to in this article, please visit https://doi.org/10.1017/S0007114521002762
