Introduction

The Omega-3 Index has been suggested as a physiologically relevant, modifiable, and independent risk factor for cardiovascular disease [1]. The Index is equal to the content of long-chain omega-3 fatty acids—eicosapentaenoic acid (EPA, 20:5n-3) and docosahexaenoic acid (DHA, 22:6n-3)—in red blood cell (RBC) membranes, as a percentage of the total fatty acids. The omega-3 fatty acid content of RBC membranes reflects the fatty acid content of cardiac membranes [2]. Dietary intake of fats is thought to modify the Index [3], and supplementation with long-chain omega-3 fatty acids has been shown to increase the Omega-3 Index in a randomised trial [4]. Other factors such as age and smoking habits may also affect the Index [3].

Long-chain omega-3 fatty acids in RBC membranes exert beneficial metabolic effects, in part, by altering membrane characteristics and the activity of membrane-bound proteins [5]. They reduce the risk of cardiovascular disease through their strong anti-inflammatory effects, their capacity to reduce platelet adhesiveness, and by benefitting blood pressure and vascular reactivity, cardiac function, lipid metabolism, platelet and leukocyte function, cytokine production and oxidative stress [6–9]. Harris and von Schacky have shown, through analyses of epidemiological and randomised controlled trials, that an Omega-3 Index of <4% is associated with a high risk, 4–8% an intermediate risk and >8% a low risk of coronary heart disease mortality in adults [5].

Existing literature on investigations of the Omega-3 Index is predominantly focussed on adults [10–12]. To our knowledge, there have been no reports describing associations between the Omega-3 Index and cardiovascular and metabolic risk, dietary, lifestyle and socioeconomic factors in an adolescent population. To assess its potential value as an early predictor of adult cardiovascular and metabolic disease, we have investigated cross-sectional associations between cardiometabolic risk factors and the Omega-3 Index in a large population-based adolescent cohort in Western Australia.

Materials and Methods

Subjects

As a longitudinal observational study, the Raine Study recruited 2,900 pregnant women from May 1989 to November 1991 through the public antenatal clinic at the King Edward Memorial Hospital and private clinics in Perth, Western Australia. Further details on the Raine Study have been previously published [13]. Of the initial cohort of 2,868 live births, assessments occurred at birth and at ages one, two, three, five, eight, ten and 14 years. This study utilises data collected at the 14-year follow-up when assessments of RBC fatty acids and dietary intake were conducted. The ethics committees of King Edward Memorial Hospital and Princess Margaret Hospital approved the protocol for all aspects of the study. Each adolescent, as well as their parent or guardian, provided written consent for participation in the study.

Red Blood Cell Analysis

Fatty acids of interest for this study were EPA and DHA. RBC fatty acid analysis was performed as previously described [14]. Briefly, chloroform:methanol (2:1) was used to extract total lipids and fatty acid methyl esters were prepared by treatment of extracts with 4% H2SO4 in methanol at 90 °C for 20 min. Samples were analysed by gas chromatography using an Agilent 7890A gas chromatograph. The column was a BPX70 (25 m × 0.32 mm, 0.25 μm film thickness) (SGE, Ringwood, Victoria, Australia) with programmed temperatures of 150–210 °C at 4 °C/min. N2 was used as the carrier gas at a split ratio of 30:1. Peaks were identified by comparison with a known standard mixture. Reproducibility for duplicate analysis was ~10–20%. Coefficients of variation were 2% for DHA and 4% for EPA. The fatty acids were expressed as a percentage of the total fatty acids measured (C14–C22).

Assessment of Cardiovascular Factors

Blood pressure readings were taken using a Dinamap ProCare 100 automatic oscillometric blood pressure recorder (GE Healthcare Technologies, Rydalmere, Australia) while subjects were seated after a 5-min rest. Over a 10-min period, six blood pressure measurements were taken. The first was disregarded and blood pressure was calculated as the mean of the next five measurements. Trained phlebotomists visited the adolescents at their homes and obtained fasting blood samples prior to breakfast. The biochemistry assays for the study were conducted by PathWest Laboratories at Royal Perth Hospital. Serum triglycerides were measured using the Cobas MIRA analyser (Roche Diagnostics, Basel, Switzerland). Glucose was measured using an automated Technicon Axon Analyzer (Bayer Diagnostics, Sydney, Australia) and insulin was measured on an Immunlite 2000 Insulin Analyzer (Siemens Medical Solutions Diagnostic, LA, USA). Insulin resistance was determined using the homeostasis model assessment of insulin resistance (HOMA-IR), calculated as fasting plasma insulin (mU/L) × plasma glucose (mmol/L)/22.5 [15]. High density lipoprotein (HDL) cholesterol was determined on a heparin–manganese supernatant [16]. Cholesterol ratios were calculated for total/HDL, low density lipoprotein (LDL)/HDL and triglycerides/HDL [17, 18].

Dietary Intake

Dietary intake at this follow-up was assessed from three-day food records in household measures, as previously reported [19]. In brief, subjects were provided with a record booklet with instructions and a set of metric measuring cups and spoons. Food records were individually checked by a dietitian as they were returned in order to clarify any ambiguous or potential omissions [20]. The Australian Food and Nutrient database through the FoodWorks dietary analysis program (Professional Version 5, 2007, Xyris Software, Brisbane) was used to analyse the food record data. Individual nutrients as well as overall diet was considered from the food diaries, with dietary patterns used to investigate overall diet. Using exploratory factor analysis, two major dietary patterns were identified: a ‘healthy’ pattern high in fresh fruit, vegetables, whole grains and grilled or canned fish, and a ‘western’ pattern high in takeaway foods, confectionery, soft drinks, crisps and fried potato [21]. Each adolescent received a z-score for both western and healthy dietary patterns. A positive score indicated a greater intake of foods representative of that pattern. To assess differences in the Omega-3 Index due to early infant feeding, information on breastfeeding cessation was obtained from the year 1, 2 and 3 Raine follow-up questionnaires. At the time of these follow-ups it was not mandatory for infant formulas to be enriched with long-chain omega-3 fatty acids in Australia.

Additional Factors

Anthropometry and Puberty

Adolescents were dressed in running shorts and singlet tops for anthropometric measurements. Height was measured to the nearest 0.1 cm with a Holtain Stadiometer, and body weight was measured to the nearest 100 g using a Wedderburn Digital Chair Scale. The body mass index (BMI) was calculated as weight in kilograms divided by height in meters squared. The Tanner stages of pubic hair development [22, 23] was used to assess puberty; adolescents selected their corresponding developmental stage in a questionnaire completed privately. Adolescents were asked to choose from a set of standard drawings depicting the different Tanner stages from two (sparse) to five (adult) (stage one was omitted as this corresponds to pre-pubescent age younger than 10 years).

Sociodemographic and Family Characteristics

Family income, mother’s age at conception and mother’s education level were obtained by parent report. Maternal education was assessed by the highest school year completed. Current family income, defined as the annual income for the household before tax at the time of the follow-up, was determined as ($AUD): <$35,000 pa, $35,000–70,000 pa, or >$70,000 pa. Family history of cardiovascular disease was assessed as either yes or no, depending on whether a biological parent or sibling of the adolescent had been medically diagnosed with diabetes mellitus, hypertension, hypercholesterolemia, or other cardiac condition.

Fitness

The Physical Working Capacity 170 (PWC 170) test was applied to estimate aerobic fitness [24]. PWC170 was measured using a bicycle ergometer to determine the power output (watts) required at a heart rate of 170 beats per minute. This measure of fitness is highly correlated with self-reported physical activity level in the Raine cohort at the 14-year follow-up [25].

Statistical Analysis

Independent t tests or cross-tabs analysis were used to compare characteristics of Raine Study adolescents between Omega-3 Index risk categories, with low and intermediate categories combined due to the small number of subjects in the low category. Mann–Whitney tests were used to assess significance of dietary intakes of omega-3, the omega-6 to 3 ratio and EPA + DHA due to the skewed distribution of these variables. Associations between continuous variables and the Omega-3 Index were described with Pearson’s or Spearman’s correlations. Linear regression was used to examine associations between Omega-3 Index as a continuous independent variable and cardiovascular risk factors as dependent variables. General Linear Modelling was used to analyse associations with Omega-3 Index categories and cardiovascular risk factors. Log values were used for HOMA, triglycerides, and insulin due to their skewed distribution. To evaluate the relationship between the Omega-3 Index and cardiometabolic risk factors, parsimonious models were created which adjusted for age, sex, BMI, maternal education, total energy intake and family history of cardiovascular disease or diabetes, based on the relationship of the variables with the risk factors and the Omega-3 Index. Other variables considered for the model included puberty, aerobic fitness, physical activity, family income, single parent family, maternal age, but were excluded due to strong correlation between covariates or lack of association with risk factors and the Index. Subjects who were on cardiac or diabetes related medications (n = 4) or diagnosed with diabetes (n = 9) were excluded from analysis investigating cardiometabolic disease risk factors. The Statistical Package for Social Sciences for Windows, Rel.15.0.0. 2006 (Chicago: SPSS Inc) was used for analyses; statistical significance was set at P ≤ 0.05.

Results

From the initial Raine Study cohort of 2,868 live births, 1,861 subjects completed at least one aspect of the 14-year follow-up (mean age 14.0 ± 0.2 years, range 13.0–15.0 years), with the remainder lost to follow-up, withdrawn or temporarily deferred from the study (n = 975), or deceased (n = 32). RBC fatty acids were determined for 1,301 subjects. The mean ± SD Omega-3 Index was 4.90 ± 1.04% (range 1.41–8.42%). Relative to the published Omega-3 Index risk categories [5], 15.6% of adolescents were in the high risk category (<4%), 84.0% were in the intermediate risk category (4–8%) and 0.4% were in the low risk category (>8%). Omega-3 Index values in the population were normally distributed, while dietary intake of EPA and DHA was not (Fig. 1). Characteristics of this sample across high and low/intermediate Omega-3 Index risk categories are shown in Table 1.

Fig. 1
figure 1

a Histogram of Omega-3 Index in Raine Study adolescents (n = 1,301), showing Omega-3 Index risk categories (5), and b subjects who also reported dietary intake of EPA plus DHA from completed and followed-up 3-day food records (n = 689)

Full size image
Table 1 Characteristics and energy adjusted dietary intakes of Raine Study adolescents in high risk (n = 203) compared with low/medium risk (n = 1,098) Omega-3 Index risk categories [5]
Full size table

When associations with subject characteristics were examined with the Omega-3 Index as a continuous variable, age at assessment (r = 0.12, P < 0.01) and maternal education (r = 0.09, P < 0.01) showed significant correlations with the Index. BMI showed borderline significance (r = 0.05, P = 0.05). The Omega-3 Index as a continuous variable showed significant positive associations with healthy eating pattern scores (r = 0.14, P < 0.01), energy adjusted dietary intakes of EPA + DHA (r = 0.22, P < 0.01), protein (r = 0.12, P < 0.01) and total omega-3 fats (r = 0.08, P = 0.04), and a significant negative association with western eating pattern scores (r = −0.13, P < 0.01) (data not shown). The significant association with EPA + DHA was also observed in comparison of Omega-3 Index risk categories (P < 0.01), along with a positive association with glycemic index (P = 0.02) (Table 1).

Several food group intakes were correlated with the Omega-3 Index (Fig. 2). Fish and seafood (r = 0.18, P < 0.01), wholegrain foods such as multigrain bread, brown rice and pasta (r = 0.11, P < 0.01) and vegetables (r = 0.10, P < 0.01) showed the strongest positive associations with the Omega-3 Index. Intakes of soft drinks (r = −0.11, P < 0.01) and crisps (r = −0.10, P < 0.01) were negatively correlated with the Omega-3 Index.

Fig. 2
figure 2

Bivariate correlations between food group intakes (g/day) and the Omega-3 Index in Raine Study adolescents. **P < 0.01, *P < 0.05. Wholegrains includes wholegrain crackers and bread, oats, wholegrain breakfast cereal, brown rice and pasta. Refined grains includes white bread, crackers, rice and pasta, savory crackers. Confectionery includes ice and chocolate confection. Takeaway foods includes hamburgers, pizzas, fried chicken, savory pastries. Sweet baked goods includes cakes, biscuits, sweet pastries. Crisps includes popcorn, corn chips, extruded cheese snacks

Full size image

After adjustment for potential confounding factors in a linear regression model, the Omega-3 Index was positively associated with total cholesterol (β = 0.064, P = 0.01) and HDL-cholesterol (β = 0.029, P = 0.01) in boys and girls (Table 2). There were no significant associations with other cardiovascular or metabolic risk factors. When analysed according to gender, significant associations were observed with boys for total cholesterol (β = 0.074, P = 0.03), HDL-cholesterol (β = 0.036, P = 0.01) and diastolic blood pressure (β = −0.684, P = 0.04); insulin (β = −0.019, P = 0.05) and HOMA-IR (β = −0.019, P = 0.07) showed borderline significance. No significant associations were observed with girls. Associations with risk factors were also examined according to Omega-3 Index risk categories [5]. General linear modelling suggested boys and girls in the low/intermediate risk category were more likely to have lower HDL-cholesterol levels, with borderline significance (β = −0.057, P = 0.07); no other risk factors were significant when boys and girls were analysed together or separately (data not shown).

Table 2 Linear regression analysis results for the Omega-3 Index as the independent variable with cardiovascular factors as the dependent variables, in Raine Study adolescents
Full size table

Discussion

Due to variation in the methods used for red blood cell analysis in different laboratories, direct comparison of results is not possible. However, we found that the mean Omega-3 Index in the Raine Study adolescents (4.9 ± 1.0%) was consistent with the value of 4.7 ± 0.4% previously reported in a smaller control sample of healthy Australian children and adolescents aged 9–18 years [26]. Our value is also similar in value to that reported in US adults (4.9 ± 2.1% [27] and 3.5 ± 1.2% [28]), but lower than that in young overweight and obese adults from Iceland, Spain and Ireland (7.0 ± 1.9% [29] and Korean pre-school chidren (9.1 ± 0.8%) [30]. Although results are not directly comparable, dietary variations between cultures, particularly in regards to fish intake, may contribute to the variation in the Omega-3 Index observed in different populations.

Dietary fish intake was the strongest food group predictor of the Omega-3 Index in the Raine adolescents (P < 0.01), and has previously been shown to be a determinant of the Index in older US adults [3]. Fish is a good source of dietary EPA and DHA, and this finding is consistent with trials showing increased dietary EPA and DHA leads to higher concentrations of RBC omega-3 fatty acids [4, 31]. Further, our results showed daily intake of EPA + DHA positively correlated with the Omega-3 Index (P < 0.01). Fish is also a good source of protein, along with legumes, nuts and eggs which may have contributed to the positive association observed with protein and Omega-3 Index. The wholegrain food group showed the next strongest positive association with the Omega-3 Index (P < 0.01). Wholegrains contain omega-3 fats in the form of alpha-linolenic acid (18:3n-3) in the bran outer layer of the grain. Interestingly, refined grain products that are missing the bran layer were also observed to have a significant positive association with the Omega-3 Index (P < 0.05). This may be partially due to the consumption of DHA enriched white bread observed in the Raine Study cohort at this follow up [19]. However, given the additional positive associations with vegetables and fruit, it is more likely that these associations with the Omega-3 Index are related to an overall ‘healthy’ pattern of eating. Similarly, a high intake of soft drinks and crisps, along with a high dietary glycemic index, may reflect a less ‘healthy’ eating pattern. This is reflected in the associations observed with the dietary pattern scores identified in this cohort: the Index showed a positive association with the healthy pattern score and a negative association with the western pattern score. Food groups loading on the healthy pattern included fresh fruit, vegetables, wholegrains and grilled or canned fish, while the ‘western’ pattern score represented a higher intake of takeaway foods, crisps and soft drinks [21]. Although soft drinks do not contain fat, their consumption has been shown to result in a rapid increase in glucose and insulin concentrations in adolescents [32], and the fructose component of sucrose in soft drinks may lead to enhanced fatty acid synthesis, contributing to higher circulating triglycerides [33]. Higher GI diets, as observed in subjects in the high risk Omega-3 Index category, may also contribute to higher plasma triglycerides [34]. Together, these dietary factors may play a role in alteration of the fatty acid content of RBC membranes.

Results of our study also extend the findings of previous research in adults [3, 31] by confirming that higher dietary intake of EPA and DHA through consumption of fish is associated with a higher Omega-3 Index in adolescents. However, the difference in distributions of dietary EPA and DHA compared to Omega-3 Index values (as shown in Fig. 1) highlight the contribution that non-dietary factors may also make in determination of the Omega-3 Index.

The Omega-3 Index has been proposed to predict coronary heart disease mortality in adults [5]. Our data show it is also associated with total and HDL-cholesterol concentrations and diastolic blood pressure in boys in our adolescent cohort. The positive relationship between increasing Omega-3 Index and higher HDL-cholesterol supports Australia’s National Heart Foundation’s position statement that marine omega-3 PUFA supplementation increases HDL-cholesterol levels [35]. This result is also consistent with previous research in adult populations [36]. In contrast, the Omega-3 Index was not significantly associated with HDL-cholesterol or other cardiovascular risk factors in a study of overweight and obese European adults [29], although borderline significance was observed with lower LDL. The findings, however, may have been influenced by weight status, as our results suggest that the BMI is positively associated with the Omega-3 Index (P = 0.05). Consistent with previous research showing a positive association between age and the Index [3], the adults in the European study had a noticeably higher Omega-3 Index than we observed in our adolescent population (7.0 ± 1.9% vs. 4.9 ± 1.0%), which may have affected associations with disease risk. Although high total cholesterol is commonly considered a risk factor for cardiovascular disease, high HDL-cholesterol is beneficial [37], and the two measures are related as HDL-cholesterol contributes to total cholesterol. Ratios may therefore be considered more useful as indicators of cardiovascular disease risk [17], however we found no significant associations between the Index and ratios of total/HDL-cholesterol, LDL/HDL-cholesterol or triglycerides/HDL-cholesterol.

The significant negative association observed with diastolic blood pressure in boys in our cohort supports a role for omega-3 fatty acids in optimisation of blood pressure. Omega-3 fatty acids have been shown to have a variety of actions that can lead to improved vasodilation and arterial compliance, including increased membrane fluidity, suppression of vasoconstrictors, and changes in mobilisation of intracellular calcium [9, 38, 39].

In our adolescent cohort, we observed a gender difference in the relationship between Omega-3 Index and risk factors for cardiovascular and metabolic disease. Analysis with boys showed significant or borderline associations with measures of cholesterol, blood pressure, and insulin resistance, whereas no significant associations were observed for measures in the girls analysis. Boys were more likely to be in the high risk category of Omega-3 Index (<4%) compared to the low risk (≥4%, P = 0.10), and similar gender differences have previously been identified in the long-chain omega-3 fatty acid composition of RBC membranes in rats, with a link between ovarian hormones and DHA composition proposed [40]. Sex hormones may play a role in modification of the omega-3 content of tissues, possibly by altering expression of enzymes in the liver [41], while also contributing to gender specific pathophysiological differences in cardiovascular and metabolic disease [42].

To our knowledge, this is the first study to evaluate the Omega-3 Index with cardiovascular and metabolic risk factors in a large cohort of adolescents. Interpretation of our study results is limited by the cross-sectional study design, however an important strength of our study is that it represents a large, population based cohort and has data on a wide range of cardiometabolic, socioeconomic and dietary variables. Our results demonstrate a significant and independent association between the Omega-3 Index and total and HDL-cholesterol, as well as blood pressure, in Australian adolescent boys. Although the Omega-3 Index did not demonstrate a statistically significant association with the cholesterol ratios, it is considered to be a risk factor in its own right and has not been shown to mediate through effects on traditional risk factors. Therefore the Index may still be useful in the prediction of cardiovascular disease later in life, and our results support further long term investigation of this concept.