Greater CUMULATIVE EXPOSURE TO A PRO-INFLAMMATORY DIET IS ASSOCIATED WITH HIGHER METABOLIC SYNDROME SCORE AND BLOOD PRESSURE IN YOUNG MEXICAN ADULTS

Abstract

Low-grade inflammation is a suggested mechanism in the development of Metabolic Syndrome (MetS), and diet could act as a regulator. Therefore, we hypothesized that the cumulative DII exposure from diet during infancy through adulthood would be positively associated with the MetS and its components in young Mexican adults. 100 participants from the ELEMENT cohort were included in this analysis. The dietary inflammatory potential of the diet (without supplements) was assessed using 27 nutrients obtained from repeated food frequency questionnaires (1–22 years) using the dietary inflammatory index (DII®), a validated score. The cumulative exposure of DII was constructed using the Area Under the Curve (AUC of DII). The MetS was defined using the International Diabetes Federation criteria and the Metabolic Syndrome Risk Z-score (MetRisk Z-score) was estimated. Linear regressions were conducted to assess the association between the AUC of DII with MetRisk Z-score and MetS components, adjusting for sex, socioeconomic status, smoking status, physical activity, birth weight and BMI. In adulthood, the mean age was 21.5 years, 54% were male and 17% had MetS. Positive associations were found between AUC of DII with MetRisk Z-score (β= 0.12; 95% CI: 0.03, 0.22; p=0.009), systolic (β=0.33; 95% CI: 0.05, 0.61; p=0.023) and diastolic blood pressure (β=0.24; 95% CI: 0.01, 0.47; p=0.040). A cumulative proinflammatory diet from infancy through young adulthood is associated with higher MetRisk Z-scores, as well as blood pressure. These findings may provide evidence for the implementation of anti-inflammatory diet interventions throughout early life for prevention of cardiometabolic risk.

1. Introduction

Metabolic Syndrome (MetS) is defined as a cluster of cardio-metabolic risk factors, including abdominal obesity, hyperglycemia, dyslipidemia and elevated blood pressure [1]. Subjects with MetS have an increased risk for the development of type 2 diabetes and cardiovascular disease [2].

MetS is considered a public health problem with increasing prevalence on a global scale [3]. According to the International Diabetes Federation (IDF), it is estimated that 25% of the world’s population has MetS [4]. In Mexico, the 2006 National Health and Nutrition Survey (ENSANUT) reported a national prevalence of MetS of 26% in the group of young adults (20–29 years) [5]. The prevalence of early-onset MetS has increased in parallel with the incidence of obesity in Mexico [6].

Evidence suggests that MetS pathology could in part be a consequence of a low-grade systemic inflammatory process, determined by an elevation of cytokines with pro-inflammatory activity [7]. Diet is a modifiable factor that has an important role as a regulator of the balance between pro and anti-inflammatory cytokines [8]. Unhealthy dietary patterns (high intake of fat, refined carbohydrates and proteins) have been associated with higher levels of inflammation, while healthier diets (high content of fruits, vegetables, and fish) have been linked to lower inflammation in adults [9].

The Dietary Inflammatory Index (DII®) estimates the inflammatory potential of the diet based on the association between dietary components and biomarkers involved in inflammation [10]. Several studies have evaluated the association of DII with MetS in adults. Two cross-sectional studies in Korean and Iranian populations found positive associations of DII with MetS [11, 12].In contrast, two other cross-sectional studies in Polish and American adults observed associations only with some of the individual MetS components, such as glucose, diastolic blood pressure and waist circumference [13–15]. From two longitudinal studies in Europe that evaluated the lifetime association between DII and MetS, only one found that more proinflammatory diets were associated with the diagnosis of MetS and with some of its components (triglycerides, HDL-C, systolic and diastolic blood pressure) [15, 16].

In Mexico, only one previous cross-sectional study has evaluated the DII with health outcomes in the adult population, finding that subjects in the highest quintile of the DII had higher odds of type 2 diabetes mellitus [17]. To our knowledge, there are no previous studies that have evaluated the role of DII exposure during childhood and adolescence with MetS risk in young Mexican adults. The first years of adult life are crucial for identifying and targeting cardiometabolic risk factors, and in the context of the high prevalence of MetS in Mexico, the objective of the present study was to evaluate the association between the cumulative exposure (from ages 1 to 21–22 years) to a pro-inflammatory diet with MetS and its components in young Mexican adults.

We hypothesized that a greater cumulative pro-inflammatory diet from infancy through adulthood would be positively associated with the MetS score and its components in this population. First, to evaluate the inflammatory potential of diet during childhood, we calculated the area under the curve of the DII at each childhood visit. Second, we explored the association of the cumulative DII and the metabolic risk syndrome, with multiple linear regressions.

2. Methods and materials

2.1. Study population

The present analysis includes a subsample from the first cohort of the Early Life Exposures in Mexico to Environmental Toxicants (ELEMENT) [18]. The cohort started in 1994 with the recruitment of pregnant women (n=631) who attended three prenatal clinics that served a low to a middle-income population in Mexico City. Exclusion criteria has been described elsewhere [18]. After birth, infants were followed every six months until 4 years of age (childhood follow-up). Then, between the period of 2008 and 2011, a second follow-up of the cohort was carried out, when children were between 14 to 16 years of age (adolescence follow-up, n=206). Finally, between 2016–2017, of those 206 participants, 100 subjects (21–22 years of age) who agreed to participate were selected for a third follow-up (adulthood follow-up) (Figure 1).

Figure 1:

Flowchart of the study population.

The Ethics, Biosafety and Research Committees of the National Institute of Public Health (INSP) approved the study (approval number 1377). Mothers and participants (age 18 and older) provided written informed consent at each wave of the study.

2.2. Dietary Assessment

The dietary intake during childhood and adolescence follow-ups was assessed by a validated semi-quantitative food frequency questionnaire (FFQ) with a recall of 3 months [19]. The questionnaire was administered up to 7 times to the mothers by trained and standardized personnel, at 1, 1.5, 2, 2.5, 3, 3.5 and 4 years of age. In the adolescence follow-up (between 14–16 years of age), the FFQ was administered directly to the study participants with the help of the mothers. The FFQ included a list of 116 foods (each item presented in standard portion size of one customary unit, cup, slice, piece, etc.) grouped into 10 categories (dairy products, fruits, vegetables, legumes, cereals, sweets, beverages, fats, snacks, and eggs and meats). Participants or mothers reported the frequencies of intake for each food item, which ranged from never to 6 or more times per day.

In the adulthood follow-up (21–22 years of age) dietary data was collected electronically using a validated, semi-quantitative FFQ based on the instrument used in the National Health and Nutrition Survey (ENSANUT) 2006, where 140 food and beverages classified into 14 food groups, with a recall of 7 days [20]. For each food item, the frequency of consumption (weekly and daily) was registered, specifying the portion size consumed using standard portion sizes or measures of weight (g) and volume (ml). The total weight or volume consumed was divided by seven days to obtain the average daily intake.

For each food item reported in the FFQ, energy, macro, and micronutrient intakes were calculated using a nutritional composition database of foods compiled by the National Institute of Public Health [21].

2.2.1. Dietary Inflammatory Index (DII®) and Children’s DII (C-DII™) calculation

FFQ-derived dietary data was used to calculate the scores for all participants at multiple time points. For the dietary information collected at childhood (1 to 4 years) and adolescence (14 to 16 years), we used the methodology of the validated Children’s Dietary Inflammatory Index (C-DII™) that included 25 components (called food parameters in the DII method) and used a food consumption database from children of 14 countries as reference [22]. For the adulthood dietary information, the original version of the DII, which comprises 45 components and uses a food consumption database from adults in 11 countries, was used [10].

For the C-DII and DII calculation, 24 and 27 components were consistently available, respectively. Z-scores of the dietary intake of each DII component were calculated, using a globally representative databases as a reference. The C-DII and DII calculations used separate databases. The Z-scores were calculated for standardized intakes per 1000 calories to account for inter-individual differences in energy intake [23]. Subsequently, centered proportions were calculated to minimize right-skewed bias. Each DII component proportion was then doubled and 1 was subtracted to achieve a distribution centered on 0 (null) and bounded between −1 (maximum anti-inflammatory) and +1 (maximum proinflammatory). Next, the resulting values were multiplied by their respective “inflammatory effect score” to obtain the specific DII of each component. Finally, for each subject, the component “specific DII scores” were added to create the overall DII score, where a greater positive DII score represents a higher pro-inflammatory potential of the diet, and a more negative DII value corresponds to a more anti-inflammatory diet [10].

2.3. Outcomes

At the young adulthood follow-up, anthropometric measurements and biomarker determinations were performed to define the MetS status.

2.3.1. Anthropometric assessment in adulthood

Standardized personnel performed the anthropometric measurements in duplicate. Weight and height were measured using a Tanita digital scale with a height rod (model WB-3000m). Weight and height were recorded to the nearest 0.1 kg and 0.5 cm, respectively. Waist circumference was measured twice to the nearest 0.1 cm with a SECA (model 201) measuring tape in the facilities of the Research Center. Blood pressure was measured twice with subjects seated, using an OMRON digital monitor (model HEM-7120); the average systolic and diastolic values were used.

2.3.2. Biomarker assessment in adulthood

Fasting venous blood samples of 10 ml were taken (fasting time ≥8 h). The samples were centrifuged and transported within the following 5 hours on ice to the Nutrition Laboratory of the National Institute of Perinatology (INPer). Glucose and lipids were determined using a bench clinical chemistry analyzer (DiaSys respons 910); high-density lipoprotein (HDL)-cholesterol and triglycerides were measured by enzymatic techniques (Roche Diagnostics, Mannheim, Germany). Additionally, insulin levels were determined by Elisa chemiluminescence method with INMULTIPLE-1000 equipment.

2.3.3. Metabolic Syndrome (MetS)

The MetS was diagnosed in the last follow-up using the IDF criteria for adults, that includes elevated waist circumference (men ≥94 cm, women ≥80 cm) plus two of four other risk factors: high triglycerides (≥150 mg/dL); low HDL-cholesterol (men <40, women <50 mg/dL); increased blood pressure (≥130 and/or ≥85 mmHg) and impaired fasting glucose (≥130 mg/dL) [4].

The definition of MetS is particularly controversial among young populations because there are no clear thresholds above which cardiometabolic risk factors start to worsen [24]. For that reason, recent studies have calculated a continuous MetS score instead of using a definition based on dichotomous diagnosis. We calculated the validated metabolic syndrome risk Z-score (MetRisk z-score), defined as the sum of Z-scores by sex of waist circumference (WC), fasting glucose, insulin, triglycerides/HDL-C ratio and the average systolic blood pressure (SBP) and diastolic blood pressure (DBP) [25].

2.4. Covariates

Birth weight and sex information were collected from the historical records of the cohort. At the young adulthood follow-up, the participants completed a questionnaire regarding socioeconomic status (SES) [26, 27] and lifestyle information, such as smoking status [28] and physical activity. The physical activity was evaluated using the short version of the International Physical Activity Questionnaire (IPAQ) that divides the population in three categories: inactive, minimally active and active) [29]. Based on measurements of weight and height made at the adult follow-up, we calculated the BMI (kg/m²). A dichotomous BMI variable was later created, where a value of 0 was assigned if the BMI was less than 25.0 kg/m² (normal); and a value of 1 if the BMI was equal or greater than 25 kg/m² (overweight or obesity) [30].

2.5. Statistical Analyses

We conducted a descriptive analysis of the main characteristics of the study sample. The categorical variables are expressed as frequencies and proportions, and the continuous variables are presented as means or medians, according to their distribution. We estimated statistical differences between the general characteristics of the subjects by MetS diagnosis using Fisher’s exact tests for the categorical variables, and Student’s Test or Wilcoxon for the continuous variables.

To determine the cumulative exposure of the DII during childhood, adolescence and young adulthood we calculated the Area Under the Curve for the DII scores (AUC of DII) over time using the trapezoidal rule [31]. This method estimates the value of a definite integral. The rule is based on approximating the value of the integral by the linear function, which passes through several points (DII scores). This integral is equal to the area of the trapezoid under the linear function graph. The individual scores of the DII measured over time include negative (more anti-inflammatory) and positive (more pro-inflammatory) values. In order to calculate the AUC, the original DII variable (range −2.64 to 4.71) was transformed into a positive value by adding the minimum DII value of |−2.64|. As this constant was subtracted from all DII values in the sample, the ranking from the original DII variable was preserved. In this way, the definite integral (the cumulative exposure to DII) is continuous and positive, where higher AUC values mean higher DII scores over time (more proinflammatory).

We computed successive linear regression models to examine the association of the AUC of DII with the MetRisk Z-score and its components at the young adulthood follow-up (WC, triglycerides, HDL-cholesterol, fasting glucose, SBP, DBP, and insulin). The first model was adjusted for sex (Model 1); the second model was further adjusted for SES (low, medium and high), smoking status (never, current and former), physical activity and birth weight (Model 2); and the third model was additionally adjusted for BMI (<25.0 vs. ≥25.0 kg/m²) (Model 3).

We additionally evaluated the association between the AUC of DII with the presence of MetS, using logistic regression models, adjusting by the same variables described above, in order to determine if this association remained using the dichotomous MetS diagnosis. Finally, we performed a sensitivity analysis where we replicated the linear and logistic regression models using an AUC of DII calculation only for the childhood period (1–4 years), to determine if the use of different FFQs at adolescence and adulthood could have affected our results (the adolescents were assisted by their mothers to answer the FFQ, and at the adulthood visit the patients self-reported their diet using a different FFQ). The sensitivity analyses were adjusted by the same covariates used in the main analysis.

All p-values presented are derived from two-tailed hypothesis tests. The analyses were performed using STATA software version 14.0.

3. Results

The final analytic sample was not statistically different from the 631 recruited women, in terms of age, BMI, marital status, delivery mode, the child’s gestational age and size at birth; however, they had attained more years of education (data not shown). The main characteristics of the participants according to their MetS diagnosis are shown in Table 1. A total of 100 subjects between 21–22 years (21.5 ± 0.5) were included in the present analysis. The study population had slightly more men (54%), most had a medium socioeconomic level (77%) and around half were current smokers (45%). The overall prevalence of MetS was 17%, and the prevalence of each component was: 51% with high WC (men ≥94 cm, women ≥80 cm), 16% with elevated glucose (≥130 mg/dL), 23% with increased triglycerides (≥150 mg/dL), 43% with low HDL-C (men <40, women <50 mg/dL), 11% with high DBP (≥130 mmHg) and 13% with high SBP (≥ 85 mmHg) (results not show). Additionally, participants with MetS presented higher BMI (30.1 kg/m² vs 23.9 kg/m²) and MetRisk Z-score values (2.7 vs −1.5), when compared to those without MetS.

Table 1.

General characteristics of a subsample of young Mexican adults from the ELEMENT cohort (n=100).

		Subgroups for diagnosis of MetS a
	Total (n=100)	Subjects without MetS (n= 83)	Subjects with MetS (n=17)	p-value b
Sociodemographic characteristics c
Sex				0.089
Male	54 (54%)	48 (58%)	6 (35%)
Female	46 (46%)	35 (42%)	11 (65%)
SES				0.696
Low	11 (11%)	9 (11%)	2 (12%)
Medium	77 (77%)	63 (76%)	14 (82%)
High	12 (12%)	11 (13%)	1 (6%)
Lifestyle habits c
Smoker, n (%)				0.764
Never	14 (14%)	11 (13%)	3 (18%)
Current	41 (41%)	35 (42%)	6 (35%)
Former	45 (45%)	37 (45%)	8 (47%)
Physical activity				0.102
Inactive	18 (18%)	12 (14%)	6 (35%)
Minimally active	38 (38%)	34 (41%)	4 (24%)
Active	44 (44%)	37 (45%)	7 (41%)
Anthropometry and biomarkers
Birth weight (kg)d	3.1 ± 0.4	3.1 ± 0.4	3.1 ± 0.4	0.762
Cholesterol (mg/dL)d	165.4 ± 4.5	163.4 ± 38.9	175.2 ± 36.9	0.200
PAS (mmHg)d	111.8 ± 10.0	111.0 ± 9.4	115.2 ± 12.3	0.115
PAD (mmHg)d	73.3 ± 6.6	73.1 ± 6.8	74.3 ± 5.7	0.507
BMI (kg/m2)e	24.3 (21.9, 27.3)	23.9 (21.6, 26.1)	30.1 (26.5, 31.3)	0.001
WC (cm)e	86.7 (80.2, 95.6)	84.5 (78.8, 93.1)	100.2 (93.6, 102.1)	0.008
Insuline	9.2 (13.2, 26.0)	8.1 (5.4, 12.5)	22.3 (13.2, 26.0)	0.002
Fasting glucose (mg/dL)e	90.1 (81.4, 96.4)	88.1 (80.7, 95.2)	100.2 (90.8, 102.1)	0.033
TG (mg/dL)e	91.5 (66.0, 145.5)	81.0 (60.0, 123.0)	168.0 (149.0, 195.0)	0.000
HDL-C (mg/dL)e	45.0 (39.1, 52.2)	46.7 (41.7, 53.5)	37.3 (33.2, 39.1)	0.001
Met-Risk Z-Scoree,f	−6.8 (−2.3, 2.0)	−1.5 (−2.5, 1.0)	2.7 (1.3, 6.5)	0.000

Abbreviations: MetS, metabolic syndrome; SES, socioeconomic status; BMI, body mass index; WC, waist circumference; TG, triglycerides; HDL-C, high-density lipoprotein cholesterol; SBP, systolic blood pressure; DBP, diastolic blood pressure.

^aMetS was defined by the International Federation of Diabetes (IDF) criteria.

^bStatistical significance was assessed using Fisher’s Exact tests for the categorical variables and Student’s t or Wilcoxon tests for continuous variables.

^cValues are presented as n (%)

^dValues are presented as means ± standard deviations (SD)

^eValues are presented as medians (Q1, Q3)

^fMet-Risk Z-Score was defined as the sum of the sex-specific Z-scores of waist circumference, fasting glucose, insulin, triglycerides/HDL-C ratio and the average of SBP and DBP.

Table 2 describes the DII score distribution by age and diagnosis of MetS. Our results suggest a positive association between age and the DII in the whole sample.

Table 2.

Dietary inflammatory Index (DII) distribution by age and presence of MetS

		All	Subjects without MetS	Subjects with MetS
	Age (years)	(n=100)	(n= 83)	(n=17)	p-value a
DII
	1b	−1.8 (−2.4, −0.9)	−1.9 (−2.3, −1.2)	−1.8 (−2.7, −0.5)	0.955
	1.5b	−1.7 (−2.3, −0.8)	−1.7 (−2.5, −0.9)	−1.3 (−2.1, −0.7)	0.219
	2c	−1.7 ± 0.9	−1.8 ± 0.9	−1.4 ± 1.0	0.126
	2.5c	−1.4 ± 1.0	−1.4 ± 1.0	−1.8 ± 1.2	0.244
	3c	−1.6 ± 0.9	−1.6 ± 0.9	−1.3 ± 1.1	0.331
	3.5c	−1.5 ± 1.0	−1.5 ± 1.0	−1.3 ± 0.9	0.408
	4c	−1.6 ± 1.0	−1.7 ± 1.0	−1.3 ± 0.9	0.163
	Adolescence (14−16) c	0.02 ± 1.5	0.03 ± 1.5	−0.02 ± 1.78	0.811
	Adulthood (21−22) b	0.7 (−1.1, 1.8)	0.8 (−1.2, 1.8)	0.4 (−1.0, 1.2)	0.790
AUC of DII	1−22d	25.0 ± 5.6	24.7 ± 5.7	26.4 ± 5.1	0.277

Abbreviations: MetS, metabolic syndrome; AUC of DII, area under the curve of the dietary inflammatory index.

^aThe statistical differences were evaluated using Student’s t or Wilcoxon tests for the normally and non-normally distributed variables, respectively.

^bDII is presented in its original scale as medians (Q1, Q3).

^cDII is presented in its original scale as means ± SD.

^dThe AUC of DII is presented as means ± SD.

When comparing DII scores by MetS diagnosis within specific age strata, none of these differences were statistically significant (Table 2). Table 3 shows the association between the AUC of DII with MetRisk Z-score and each component of the MetS, using multiple linear regressions. In Model 3, we found a significant positive association between AUC of DII and MetRisk Z-score (β= 0.12; 95% CI: 0.03, 0.22; p<0.01), SBP (β= 0.33; 95% CI: 0.05, 0.61; p<0.05) and DBP (β= 0.24; 95% CI: 0.01, 0.47; p<0.05). No statistical differences were found for Models 1 and 2.

Table 3.

Associations between the AUC of DII, MetRisk Z-score and components of the MetS during adulthood.

	Model 1 (n=100)			Model 2 (n=100)			Model 3 (n=100)
Outcomes	β	95% CI	p-value	β	95% CI	p-value	β	95% CI	p-value
MetRisk Z-score a	0.07	(−0.05, 0.19)	0.273	0.078	(−0.00, 0.156)	0.051	0.12	(0.03, 0.22)	0.009
Individual MetS components
WC (cm)	0.00	(−0.44, 0.44)	0.996	−0.03	(−0.17, 0.11)	0.694	0.15	(−0.11, 0.43)	0.247
Fasting glucose (mg/dL)	0.65	(−0.29, 1.59)	0.173	0.63	(−0.26, 1.53)	0.162	0.77	(−0.16, 1.70)	0.103
TG (mg/dL)	1.98	(−0.67, 4.64)	0.140	1.95	(−0.63, 4.54)	0.137	2.30	(−0.34, 4.92)	0.087
HDL-C (mg/dL)	−1.88	(−0.52, 0.15)	0.267	−0.18	(−0.50, 0.15)	0.288	−0.24	(−0.57, 0.07)	0.130
SBP (mmHg)	0.28	(−0.03, 0.59)	0.076	0.25	(−0.02, 0.52)	0.065	0.33	(0.05, 0.61)	0.023
DBP (mmHg)	0.21	(−0.02, 0.45)	0.069	0.21	(−0.01, 0.42)	0.062	0.24	(0.01, 0.47)	0.040
Insulin (μIU/mL)	0.24	(0.36, 0.41)	0.903	0.04	(0.30, 0.39)	0.819	0.14	(−0.22, 0.50)	0.448

Abbreviations: AUC of DII, area under the curve of the dietary inflammatory index; 95% CI, 95% confidence intervals; MetS, metabolic syndrome; WC, waist circumference; TG, triglycerides; HDL-C, high-density lipoproteins-cholesterol; SBP, systolic blood pressure; DBP, diastolic blood pressure.

The beta estimates and 95% confidence intervals were determined using multiple linear regressions.

^aMetRisk Z-score was defined as the sum of the sex-specific Z-scores of waist circumference, fasting glucose, insulin, triglycerides/HDL-C ratio and the average of SBP and DBP (n=92).

Model 1: adjusted for sex.

Model 2: Model 1 + SES (low, medium and high), tobacco use (never, current and former), physical activity (inactive, minimally active and active) and birth weight.

Model 3: Model 2 + BMI (<25.0 vs. ≥25.0 kg/m²).

Figure 2 shows the association between AUC of DII with MetRisk Z-score from Model 3, where each unit increase in the AUC of DII corresponded to a 0.12-unit higher MetRisk Z-score.

Figure 2.

Multiple linear regression for the association between the AUC of DII with MetRisk Z-score, adjusted by sex, SES (low, medium and high), tobacco use (never, past and active), physical activity (inactive, minimally active and active), birth weight and BMI (<25.0 vs. ≥25.0 kg/m²).

There were no statistically significant associations between the diagnosis of MetS and the AUC of DII (data not shown). In the sensitivity analysis, we observed consistent results in the linear regressions between the AUC of DII during childhood (1–4 years) with MetRisk Z-core and DBP (p<0.05). However, the association with SBP was attenuated, and we observed a negative association with HDL-C (p<0.05) (data not shown).

4. Discussion

Our results suggest that long-term exposure to a pro-inflammatory diet is positively associated with the MetRisk Z-score, which is in line with our initial hypothesis. We found that a unit increase in the AUC of DII (towards inflammation) corresponded to a 0.12 unit increase in the MetRisk Z-score (β= 0.12; 95% CI: 0.03, 0.22; p<0.01). This finding is consistent with previous studies that have evaluated the association of DII with MetS in adults [11, 12, 15]. In a study including 3,726 French adults from SU.VI.MAX cohort, a pro-inflammatory diet was significantly associated with a higher risk of MetS over 13 years (Q4 vs Q1: OR=1.39; 95% CI= 1.01,1.92, p<0.05) [15]. Furthermore, this association has also been evaluated in cross-sectional studies. A study in 606 Iranian adults showed that a higher DII score was significantly associated with MetS (Q4 vs Q1, OR= 2.26; 95% CI= 1.03, 4.92, p<0.05) [12]. Another study based on data from the sixth Korean National Health and Nutrition Examination Survey (KNHANES) (n=9291) found that the top DII quartile (Q4) was positively associated with MetS prevalence in men (OR= 1.40; 95% CI= 1.06, 1.85, p<0.01) and postmenopausal women (OR= 1.67; 95% CI= 1.15, 2.44; p<0.01) [11], when compared to the fist quartile (Q1). All of these previous studies adjusted for BMI. The DII score of our study is comparable to the scores from these studies (−4 to +4), however, it should be noted that the ages were different, since these studies included DII measured in adults.

The DII score is based on the inflammatory capacity of individual dietary components that may be found in other dietary scores such as the Mediterranean Diet Score; thus, our results could be considered in light of other studies that have evaluated diet and MetS [15]. A cross-sectional study conducted in a subset of the ELEMENT cohort (n=250) in children 8–14 years of age, showed that a prudent dietary pattern (vegetables, fruit, fish, legumes, and chicken) was associated with a lower MetRisk Z-score in boys (β:−0.14; p<0.05) [32]. Similarly, in a cohort study in Finland with a follow-up of 27 years including 2,128 individuals, the frequency of vegetable consumption during childhood was inversely associated with MetS odds (OR=0.86; 95% CI= 0.77–0.97, p<0.05), blood pressure and triglycerides concentrations in adulthood [33]. Similarly, a decreased frequency of childhood vegetable consumption predicted high blood pressure (OR=0.88; 95% CI= 0.80–0.98, p<0.01) and high triglycerides concentrations (OR=0.88; 95% CI=0.79–0.99; p<0.05). These findings suggest that food choices established early in childhood may act as risk or protective factors for the development of cardiovascular diseases in adulthood. Therefore, the first years of life are essential to identify possible risk indicators for the development of chronic diseases and to promote healthy dietary behaviors [34].

Our estimates only became statistically significant in models adjusting for overweight/obesity status. Other studies that have studied this association have included BMI as a confounder [11, 12, 14–16]. However, we cannot exclude the possibility that the stronger estimates could be due to collider-stratification bias, since BMI could also mediate the effect of diet-associated inflammation, as estimated by the DII effect on MetS. Other explanations exist, for example, obesity status may be an important source of reporting bias, because subjects with obesity are more likely to misreport their true dietary intake [35] and this could distort the measure of association. Despite the strong correlation between BMI and WC, the analysis of WC adjusting by BMI could help distinguish the effects of overall body size and muscularity from those of central obesity [36]. Because we found a positive association with SBP and DBP after adjusting for several potential confounders, including socio-demographic data and lifestyle habits, the connection between DII with the MetRisk Z-score in this study could be mainly explained by the association between these cardiovascular risk components of MetS with inflammation. Prior studies suggest a positive relationship between markers of chronic inflammation, including C-reactive protein (CRP), interleukin 6 (IL-6), and tumor necrosis factor α (TNF-α) with higher blood pressure [37]. Similarly, compared to normotensive subjects, higher concentrations of CRP have been found in hypertensive patients [14, 15].

The observed relationship between dietary inflammation and higher blood pressure is consistent with the SU.VI.MAX cohort study which found that higher DII scores were associated with higher systolic and diastolic blood pressure [15]. Furthermore, in the Polish-Norwegian Study (PONS), a more anti-inflammatory diet was associated with a decreased prevalence of SBP component among males [14].

Some limitations of our study should be mentioned. First, the small sample size could have affected the precision of our estimates and means that we may have been underpowered. Second, we focused on the inflammatory potential of the diet, but other factors such as stress and environmental toxicants could cause inflammation and thus potentially confound [8]. Third, we did not have data concerning previous family medical history, and other time-varying variables such as physical activity during childhood, which could be important confounders in this context. Finally, the measurement of diet is prone to a variety of biases in epidemiologic studies, and we were only able to obtain 27 of 45 original dietary inflammatory components to calculate the DII in adulthood. Components missing from our DII calculation included rosemary, saffron, pepper, garlic, thyme/oregano, turmeric, eugenol, flavan-3-ol, anthocyanidins, ginger, flavonols, onion and green/black tea (anti-inflammatory) and trans-fat (pro-inflammatory), as well as supplements.

Nonetheless, it is important to mention that some spices, such as saffron, turmeric and ginger, are not usually consumed in the Mexican population [32] and none of the children of this sample reported the intake of supplements during the first 5 years of age, and only 10 reported its consumption during the adulthood visit. Furthermore, the predictive capacity of the DII seemed unimpaired in this and other studies that used around 23–32 components obtained from FFQs, since they also found similar associations between the DII ad MetS [12, 38]. Also, the study population was limited to young adults enrolled in Mexico City; thus, the generalizability of the results might be limited to this group.

An important strength of the present study includes the evaluation of the cumulative exposure to the inflammatory potential of diet, using several DII measurements from infancy through young adulthood, which, to the best of our knowledge, had never been used with the DII and represents a unique and novel approach. Another strength of the present study is the prospective design entailing a long follow-up period of 21 years, which meant a very low potential of reverse causation bias. Furthermore, the use of the DII and C-DII, which have been validated with CRP concentrations in US adult [39] and children (25% Mexican American) population [22], and was designed to be representative of the dietary intake of populations around the world, facilitates quantitative comparisons across studies. Finally, the DII has been associated with inflammatory biomarkers in adolescents, such as TNF- α, IL-1, IL-2, IFN- γ, vascular cell adhesion molecule [40], IL-6 and the complement component 4 (or C4) [41].

The results of the present study suggest that a cumulative proinflammatory diet in childhood and adolescence is associated with increased MetRisk score, as well as SBP and DBP in young adulthood. These findings support the idea of the role of inflammation as an important underlying mechanism linking diet and the development of MetS, and may provide evidence for the design of interventions promoting anti-inflammatory diets among early life stages.

Highlights.

• The dietary inflammatory index (DII®) estimates the inflammatory potential of diet.

• The cumulative DII (1–22 years) was calculated using the area under the curve.

• Cumulative DII and the Metabolic Syndrome Risk Z-score are positively associated.

• Cumulative DII is associated with higher blood pressure in young adults.

• Inflammation might be an important mechanism linking diet and metabolic syndrome.

Abbreviations

MetS

Metabolic Syndrome

DII®

Dietary inflammatory index

FFQ

Food Frequency Questionnaire

AUC of DII

Area Under the Curve of the Dietary Inflammatory Index

IDF

International Diabetes Federation

MetRisk Z-score

Metabolic Syndrome Risk Z-score

ENSANUT

National Health and Nutrition Survey, Mexico

ELEMENT

Early Life Exposures in Mexico to Environmental Toxicants

INSP

National Institute of Public Health, Mexico

C-DII™

Children’s Dietary inflammatory index

INPer

National Institute of Perinatology

SES

Socioeconomic Status

BMI

Body Mass Index

WC

Waist Circumference

TG

Triglycerides

HDL-C

High-density Lipoproteins-cholesterol

SBP

Systolic Blood Pressure

DBP

Diastolic Blood Pressure

KNHANES

Korea National Health and Nutrition Examination Survey

PONS

Polish-Norwegian Study