Dietary patterns and odds of Type 2 diabetes in Beirut, Lebanon: a case–control study

Type 2 Diabetes (T2D) has reached epidemic proportions in both developed and developing countries and is now recognized as a major global health problem [1‐4]. Lebanon, similar to other counties of the Middle East North Africa Region (MENA) region, is witnessing a polarization in disease patterns with under nutrition and its related diseases coexisting with chronic diseases such as T2D and cardiovascular diseases typically associated with developed countries [5]. Even though the prevalence of T2D among adults in Lebanon (15.8%) [6] appears to be lower than that reported from most other countries in the MENA region including Bahrain (25.5%), United Arab Emirates (23.3%) and Saudi Arabia (23.7%) [5], available data in the country suggested that prevalence rates of the disease are following an alarming trend over time, increasing from 11.6% in 1999 to 15.8% in 2004 [6, 7]. According to the International Diabetes Federation, the projected prevalence of T2D among adults in Lebanon for year 2020 is 20.4% [8]. This is corroborated by a recent report suggesting that, worldwide, the greatest relative increase in the number of people with diabetes is expected to occur in countries of the Middle Eastern crescent [9]. This escalating trend in disease prevalence coupled to significant downstream pathophysiologic effects and enormous financial liabilities poses a major public health problem that necessitates the search for mitigating factors and strategies to address them.

Dietary intake is an important modifiable environmental risk factor in the development and prevention of T2D [10]. Although there is convincing evidence that excess consumption of calories and high glycemic index foods that are also low in phytochemical content are associated with an increased risk of T2D [11‐14], the evidence regarding the association of other dietary related risk factors with T2D is far less complete [15]. More recently, several nutritional epidemiologic studies have examined the association between dietary patterns rather than single nutrients in relation to the risk of T2D. Dietary patterns may exert greater effects on health than individual foods, nutrients, or food groups [16]. This approach has been used to study obesity, diabetes, cancer, cardiovascular diseases, and all-cause mortality [17‐24]. Most of the studies investigating the association of various dietary patterns with the risk of T2D –whether patterns were derived a posteriori (using cluster or factor analyses) or according to hypothesis-driven methods (such as index analysis) - showed that the adoption of a Western and Western-like dietary patterns, characterized by high intakes of processed foods, red meat, and refined grains, were related to a higher risk of glucose intolerance and T2D [17, 21‐27]. A recent study in Lebanon also showed that adherence to a Western dietary pattern was associated with higher odds of hyperglycemia and metabolic syndrome [28].

Furthermore, a Mediterranean dietary pattern, characterized by high consumption of fruits, vegetables, fish, and olive oil, has also been suggested to be protective against the risk of metabolic abnormalities and T2D [29‐32]. In the East Mediterranean region, studies evaluating the association between the different dietary patterns and the risk of T2D are scarce [33, 34].

The aim of this study was to examine the association of various dietary patterns with the odds of T2D among Lebanese adults. Findings from this study could inform prevention strategies among Lebanese adults with high risk of T2D.

This study was approved by the Institutional Review Board at the American University of Beirut. All participants signed a consent form prior to participation and had the right to withdraw from the study at any time.

During the study period, 65 newly diagnosed cases of T2D were identified, and 58 agreed to participate in the study (response rate: 89%). Out of 139 control participants approached, 116 were enrolled in the study (response rate: 82%). The main reasons for declining participation were “lack of time” and “no interest”.

Table 1 presents the means and proportions of socio-demographic, lifestyle, and anthropometric characteristics of the study population. Mean age of the total study population was 56.18 ± 10.93 years (range: 40–79 years). Cases and controls were comparable across age, sex, income, marital status, education, and smoking. There were significant differences between cases and controls with regard to family history of T2D, physical activity, BMI, waist circumference, and percent body fat. Compared to controls, a higher proportion of cases reported a family history of T2D (58.6% vs. 13.8%). Cases who reported engaging in moderate to high physical activity levels were fewer than the controls (19.0% vs. 51.7%). Compared to controls, cases had a significantly higher mean BMI (30.79 ± 5.51 vs. 26.81 ± 4.48 Kg/m2), higher percent body fat (33.43 ± 8.76 vs. 28.88 ± 7.12), and a larger waist circumference (106.53 ± 12.01 vs. 95.92 ± 12.00 cm).

Factor analysis revealed four dietary patterns: Refined Grains & Desserts, Traditional Lebanese, Fast Food and Meat & Alcohol which explained 13.68%, 10.8%, 9.13%, and 8.67% of the dietary intake variance respectively (Table 2). The factor loading matrix derived from PCFA is shown in Table 2. The Refined Grains & Desserts pattern featured high consumption of pasta, desserts, fried fish, pizzas and pies, in addition to breakfast cereals and white bread. The second pattern, Traditional Lebanese, was mostly associated with consumption of olives and olive oil, fruits and vegetables, low fat milk and milk products, traditional Lebanese dishes, whole wheat bread, Arabic sweets, and “Shawarma” & ”Falafel”. The Fast Food pattern was characterized by high intakes of French fries, added fat, fast food sandwiches, mixed nuts, and full fat milk and milk products. The fourth pattern Meat & Alcohol loaded mostly on eggs, alcohol, red meat, juices and carbonated beverages, chicken, and fish (not fried). Eggs is often cited with meats as one food group (in most of the food guide pyramids, MyPlate, MyCedar, etc.), hence this pattern’s name “Meat & Alcohol” encompasses eggs as well.

Table 3 shows the associations between the factor scores of the four dietary patterns with obesity markers, energy, and energy adjusted nutrient intakes as assessed by Pearson correlation coefficients. Nutrients which have been previously reported to be associated with T2D were included in this analysis. While the Refined Grains & Desserts, the Fast Food, and the Meat & Alcohol patterns were directly associated with BMI and waist circumference, inverse associations were noted between the Traditional Lebanese pattern and these obesity markers. Percent body fat was only directly associated with the Fast Food pattern (r = 0.249). The scores of the Refined Grains & Desserts, the Traditional Lebanese, the Meat & Alcohol patterns were directly correlated with energy, with those of the Refined Grains & Desserts pattern having the highest correlation (r = 0.741). In addition, the scores of the Refined Grains & Desserts pattern were directly correlated with total carbohydrates (r = 0.221) and sucrose (r = 0.161). The Traditional Lebanese pattern scores correlated directly with monounsaturated fat (r = 0.350) and dietary fiber (r = 0.411) intakes. The Fast Food pattern had the highest correlation with fat intake (r = 0.344). Compared to other patterns, the Meat & Alcohol pattern had the highest correlation with protein intake (r = 0.358).

Table 4 showed the association between the identified dietary patterns scores with the odds of T2D. In the final multivariate logistic regression model (adjusted for energy, sociodemographic and lifestyle characteristics, BMI, waist circumference, and percent body fat), a unit increase in the scores of the Refined Grains & Desserts pattern was associated with more than 3 folds increase in the odds of T2D (OR: 3.85, CI: 1.31-11.23). Similarly, scores of the Fast Food pattern were associated with higher odds of T2D (OR: 2.80, CI: 1.41-5.59). The odds of T2D were lower with increasing scores of the Traditional Lebanese pattern (OR: 0.46, CI: 0.22-0.97). No significant association was noted between the scores of the Meat & Alcohol dietary pattern and the odds of T2D.

This case control study aimed at evaluating the association between dietary patterns and the odds of T2D in a sample of 58 recently diagnosed cases with the disease and 116 matched controls residing in Beirut, the capital of Lebanon. Studying the association between dietary patterns and T2D overcomes the inconsistent findings of single nutrient analysis in evaluating diet-disease associations, takes into consideration the synergistic effects of nutrients, and produces culture specific recommendations. Furthermore, behavior change could be better achieved when discussed in terms of patterns, rather than nutrients or specific foods, because it gives food context, which in turn may facilitate the development of public health recommendations [16]. Our results revealed four major dietary patterns in this study population: Refined Grains & Desserts, Traditional Lebanese, Fast Food and Meat & Alcohol patterns which together explained 42.28% of the total variance in the dietary intake. The percentage of variance explained by the first four dietary patterns identified is similar in magnitude to what has been reported in others studies [47, 48]. It should be noted, however, the variance explained by all factors is a function of the number of factors researchers choose to retain [49]. Both the Refined Grains & Desserts and the Fast Food patterns were associated with higher odds of T2D, while the Traditional Lebanese was inversely associated with the odds of the diseases. No association was observed between the Meat & Alcohol pattern with the odds of T2D.

Previous studies have reported on the association between dietary patterns and the risk of T2D [21‐24, 50]. Overall, their results featured two main dietary patterns, the Western and Prudent. The Western pattern, characterized by a high consumption of red meat, eggs, refined grains, desserts, French fries, and high fat dairy products, was associated with higher plasma insulin [47, 51], C-peptide levels [52], and overall risk of T2D [19, 21, 47, 51]. A recent cohort study showed that adherence to a Western pattern characterized by high intakes of desserts, meats, and refined grains during adolescence was associated with 29% increase in the risk of T2D in middle aged women [53]. The Prudent pattern contained high amounts of fruits, vegetables, legumes, fish, poultry, and low fat dairy products and was not always reported as associated with reduced risk of disease [19, 21, 23, 25]. Hence, the association between the prudent pattern and the risk of T2D was less evident than that of the Western pattern with the risk of disease. Data describing the association between dietary patterns and T2D in the MENA region are scarce and limited to countries with particular patterns that are country and culture specific [33, 34]. In the present study, the Refined Grains & Desserts and the Fast Food patterns were the two patterns shown to have a direct association with the odds of T2D. Both of these patterns shared important features with the Western pattern previously described in the literature and shown to directly influence the risk of T2D [17, 21‐27]. The Refined Grains & Desserts pattern, found in this study, was characterized by high intake of pasta, desserts, pies, and pizzas, in addition to breakfast cereals and white bread and was directly correlated with energy and carbohydrates intake. The Fast Food pattern was characterized by high intake of chips and fries, fast food sandwiches, full fat milk and milk products, and added fats, and was correlated mainly with energy and fat intake. There are a few mechanisms by which adherence to these patterns may increase the risk of T2D. Many of the foods and food groups loading high on these patterns possess a high glycemic index as well as high glycemic load. Such foods have been repeatedly reported to have a direct association with T2D [54‐61]. Foods with a high glycemic load trigger higher postprandial insulin levels, which may lead to an accelerated age related decline in insulin secretion and ultimately the onset of T2D [13]. Furthermore, high consumption of sugar has been independently associated with the odds of T2D [62, 63]. In line with the findings of this study were the results of the National Nutrition Survey conducted in Lebanon in 2009, which indicated an association between a Western dietary pattern and increased odds of obesity [41], and the metabolic syndrome in Lebanon, conditions postulated to increase the risk of T2D [28].

The Traditional Lebanese pattern identified in this study was inversely associated with the odds of T2D. Although Mediterranean diets could be different from one country to another in the Mediterranean basin [64], the Traditional Lebanese pattern encompasses many features of the generally known Mediterranean diet. These features include consumption of fruits, vegetables, olives and olive oil, and whole bread [65]. Adherence to a Mediterranean diet has been repeatedly shown to exert a beneficial role regarding the development and progression of T2D [29‐32, 66]. Additionally, individual food groups and components of the diet, such as monounsaturated fatty acids, fruits, vegetables, and whole grain cereals may protect against the development of diabetes, possibly through the amelioration of insulin sensitivity and their anti-inflammatory actions [67]. A recent study among adults in rural Lebanon indicated that adherence to a Mediterranean diet, as assessed using the widely used Mediterranean diet score (MDS), is negatively associated with two of T2D known risk factors: obesity and visceral adiposity [68].

Our results indicated that the Meat & Alcohol pattern did not show an association with the odds of T2D. This pattern included food groups that were shown to be associated with varying odds of T2D. For instance, frequent consumption of meat, in particular red meat, has been reported to increase the risk of diabetes in prospective studies [21, 69, 70]. Furthermore, recent evidence, though inconclusive, suggested that increased consumption of eggs may raise blood cholesterol concentrations and thereby increase the risk of T2D by impairing islet cell function [71‐73]. On the other hand consumption of fish was shown to decrease the risk of the disease [50]. N-3 fatty acids (including α-Linolenic acid found to be positively correlated with the scores of this pattern) have been suggested to have beneficial effects on health outcomes such as reduction in concentrations of inflammatory markers and improvement of insulin sensitivity [74]. This inverse association between fish intake and the risk of T2D was challenged by a recent meta-analysis by Patel et al., which showed that lean fish, total fish, and shellfish intakes were not associated with incident diabetes but that fatty fish intake may be weakly inversely associated [75]. Moderate alcohol consumption (1–3 drinks per day) has been shown to be associated with lower risk of diabetes [76] and lower levels of pro inflammatory markers [77]. Hence the negative effects of meat and egg consumption on the odds of T2D could have been neutralized by the protective effect fish and moderate alcohol consumption might have had on the odds of the disease.

A few limitations should be considered when interpreting the findings of this study. The recall of past diet rises concerns regarding the possibility of recall bias among cases if their diet changed as a result of their diagnosis, or because they experienced symptoms as a result of the presence of subclinical disease before diagnosis. It is also possible that the dietary intake interview reflected changes due to medical advice. To minimize this recall bias, cases were constantly reminded to report on their dietary intake before diagnosis with T2D. Another limitation stems from the use of factor analysis in the derivation of the dietary patterns. This method involves many subjective decisions in selecting and grouping foods for analysis from the questionnaire, in determining the number of factors to retain, in choosing the method of rotation of the initial factors and in labeling dietary patterns according to their factor loadings [78]. To address the issue of subjectivity using factor analysis, the food groupings used in this study were similar to those reported by others in the region [28, 33, 34, 41] and the selection of the factor solution was done after evaluating the scree plots and eigenvalues. It remains important to note that the food frequency questionnaire used in this study was not validated in our study population. The fact that the food frequency questionnaire was administered by a trained dietitian and not self-completed offers many advantages in that it does not require a literate population and results in consistent interpretations and higher response and completion rates, each of which may enhance the validity of the data [78, 79]. Furthermore, any non-differential measurement error of the diet is likely to have attenuated the associations that we have observed. Finally, although significant associations were found between the Refined Grains & Desserts, the Traditional Lebanese and the Fast Food patterns with the odds of T2D, the small sample size of this study might have lead to insufficient power to detect an association between the Meat & Alcohol pattern and the disease.

This study provides insight to the association between dietary patterns and T2D, particularly among Lebanese adults, a currently understudied population. Our findings suggested that Refined Grains & Desserts and Fast Food dietary patterns were associated with increased odds of T2D while a Traditional Lebanese pattern was indirectly associated with the odds of the disease. Thus our results allow for evidence-based recommendations to limit the consumption of refined grains products, desserts, fast food sandwiches, French fries, and added fats as well as encouraging the consumption of a Traditional pattern based on fruits and vegetables. These recommendations lay grounds for planning intervention strategies that are culture based and in line with local dietary habits. The findings of this study are important not only in Lebanon but also in other countries of the Middle East region where such patterns are predominantly available.