Validation of an equation predicting highly unsaturated fatty acid (HUFA) compositions of human blood fractions from dietary intakes of both HUFAs and their precursors

Abstract

Proportions of omega-3 (n-3) and omega-6 (n-6) in 20- and 22-carbon highly unsaturated fatty acids with 3 or more double bonds (HUFA) accumulated in tissue HUFA (e.g., the %n-6 in HUFA) are biomarkers reflecting intakes of n-6 and n-3 fatty acids. An empirical equation, referred to here as the Lands’ Equation, was developed previously to use dietary intakes of n-6 and n-3 HUFA and their 18-carbon precursors to estimate the %n-6 in HUFA of humans. From the PubMed database, we identified clinical trials reporting (a) dietary intake of at least linoleic acid (18:2n-6) and alpha-linolenic acid (18:3n-3), and (b) the amounts of at least arachidonic acid (20:4n-6), eicosapentaenoic acid (20:5n-3), and docosahexaenoic acid (22:6n-3) in lipids of plasma, serum, or red blood cell. Linear regression analyses comparing reported and predicted %n-6 in HUFA gave a correlation coefficient of 0.73 (p < 0.000000) for 34 studies with 92 subject groups. These results indicate that circulating HUFA compositions can be reliably estimated from dietary intake data that not only includes n-3 and n-6 HUFA consumption, but also includes consumption of 18 carbon n-3 and n-6 precursor fatty acids.

1. INTRODUCTION

Omega-6 (n-6) and omega-3 (n-3) polyunsaturated fatty acids (PUFA) are essential vitamin-like components of the human diet. They have important roles in both the prevention and the development of various diseases [1]. A potential biomarker for both monitoring dietary intake of PUFA and assessing disease risk is the percentage of n-6 highly unsaturated fatty acids (HUFA; 20- and 22-carbon fatty acids with 3 or more double bonds) in total tissue HUFA (i.e., the %n-6 in HUFA), defined as follows:

$$\%\mathrm{n}-6\text{in HUFA}=100\times \mathrm{n}-6\text{HUFA}\left(20:3+20:4+22:5\right)\mathrm{n}-3\text{HUFA}\left(20:5+22:5+22:6\right)+\mathrm{n}-6\text{HUFA}\left(20:3+20:4+22:4+22:5\right)+20:3\mathrm{n}-9$$ (Equation 1)

where the fatty acids are expressed as concentrations or percentages of total fatty acids in any tissue or blood fraction.

Tissue HUFA are precursors of signalling molecules called eicosanoids that regulate many physiological processes [2]. The wide impact of these signalling molecules is evident from the occurrence of receptors responding differently to n-6 and n-3 eicosanoids on cells in nearly every human tissue [3]. Numerous studies suggest that overproduction of n-6 eicosanoids contributes to cardiovascular disease [4, 5], cancer [6], and other chronic diseases [7, 8], as well as psychiatric and neurodegenerative disorders [9, 10]. Because the intensity of n-6 eicosanoid-mediated events depends on the intensity of n-6 HUFA release from tissue phospholipids [11], the %n-6 in HUFA serves as a biomarker that predicts n-6 eicosanoid overproduction and therefore can potentially be used to monitor health status [5, 12].

In addition to its clinical relevance in health risk assessment, the %n-6 in HUFA is useful in monitoring essential fatty acid intake [13]. Tissue HUFA composition is influenced by dietary intake of n-6 and n-3 HUFA and their 18-carbon PUFA precursors, linoleic acid (LA; 18:2n-6), gamma-linolenic acid (GLA; 18:3n-6), alpha-linolenic acid (ALA; 18:3n-3), and stearidonic acid (SDA; 18:4n-3) [1]. Accumulation of arachidonic acid (AA; 20:4n-6) in human tissues depends on dietary intake of AA as well as LA and GLA, which are elongated and desaturated to form AA (primarily in the liver). In addition, metabolic competition from intake of ALA, SDA and n-3 HUFA alters the proportion of AA accumulated in human tissue HUFA. Therefore, metabolic interactions among all four main types of essential fatty acids (n-6 and n-3 HUFA plus n-6 and n-3 18-carbon PUFA) should be included to estimate quantitatively the %n-6 in HUFA that results from dietary information. Such estimates allow researchers, clinicians, and patients to design new dietary interventions with a predicted impact on tissue composition and health status prior to actually doing the intervention and measuring the blood fatty acids.

Quantitative evidence from Mohrhauer and Holman (14, 15) showed a hyperbolic competitive relationship by which LA and ALA affect the accumulation of HUFA in liver and other tissues (reviewed in Figure 4 of [16]). An often overlooked fact from those 50-year old studies is that the increased proportions of n-3 or n-6 in HUFA in response to dietary increases in ALA or LA had a mid-point near 0.1 % of food energy (0.1 en%), and relatively little increase in HUFA was observed with dietary supplies above 2% of food energy (2 en%). Lands et al. [17] confirmed this competitive hyperbolic relationship for rats and developed an empirical equation to describe quantitatively the proportions of n-3 and n-6 in HUFA accumulated in rat liver, plasma and RBC phospholipids. In 1992, Lands et al. [13] extended the empirical equation to include dietary n-3 HUFA effects on the %n-6 in HUFA in human, rat, and mouse tissues. The equation and constants allowed successful use of blood fatty acid analyses to predict the number of n-3 HUFA capsules that some patients were taking in a small blinded clinical study [13]. They also accurately predicted the observed HUFA proportions in the blood of infants fed diets containing only 18-carbon precursors [18]. Later, two studies of Japanese consuming a wide range of n-3 HUFA [19] led to adjusting three of the eight constants [20] to give a better fit with the much broader set of data.

Here we test the ability of the empirical equation and revised constants [20] to fit data from additional, more recent, clinical studies that reported both dietary intake of fatty acids and the HUFA blood compositions of subjects consuming a variety of different diets. For each study group, we compare the %n-6 in HUFA predicted by the empirical equation, referred to here as the Lands’ Equation, to values reported for the HUFA observed in diverse blood fractions.

2. METHODS

2.1 Literature search

We conducted a literature search of the US National Library of Medicine’s PubMed database to identify clinical trials that report dietary intake of PUFA and blood fatty acid compositions of adult subjects. The search terms used were “polyunsaturated fatty acids AND (diet OR diets OR dietary) AND blood composition,” and the results were restricted to human studies published in English after 1990. We scanned titles and abstracts for potentially relevant studies and included studies if they (a) reported enough dietary data to calculate the intake of at least LA and ALA as a percentage of food energy (en%), and (b) reported enough data to calculate at least arachidonic acid (AA; 20:4n-6), eicosapentaenoic acid (EPA; 20:5n-3), and docosahexaenoic acid (DHA; 22:6n-3) as a percentage of total fatty acids in phospholipids or total lipids of either plasma, serum, or red blood cell (RBC). We excluded studies if subjects consumed more than one fish meal per week and the dietary intake of n-3 HUFA was not reported. However, for studies that did not report n-6 and/or n-3 HUFA dietary data, we used an average intake of n-6 HUFA (0.07 en%) and n-3 HUFA (0.11 en%) for typical Western diets [4].

2.2 Dietary and tissue composition data and calculations

The Lands’ Equation (Equation 2) [13] predicts a likely %n-6 in HUFA by combining dietary intakes (expressed as en%) of four categories of essential fatty acids: 18-carbon n-6 PUFA (P₆; 18:2n-6 and 18:3n-6), 18-carbon n-3 PUFA (P₃; 18:3n-3 and 18:4n-3), n-6 HUFA (H₆; 20:3+20:4 +22:4+22:5), n-3 HUFA (H₃; 20:5+22:5+22:6). The other fatty acids (O) were calculated by subtracting the en% of n-3 and n-6 nutrients from the percent of calories as fat (details are shown in an additional file, Table S1). The revised constants from [20] are

$$\text{Predicted}=100+100\%\mathrm{n}-6\text{in HUFA }1+\left({\mathrm{HC}}_6/\text{en}\%{\mathrm{H}}_6\right)\left(1+\text{en}\%{\mathrm{H}}_3/{\mathrm{HC}}_3\right)1+\left({\mathrm{PC}}_6/{\text{en}\%\mathrm{P}}_6\right)\left(1+\text{en}\%{\mathrm{P}}_3/{\mathrm{PC}}_3+\text{en}\%{\mathrm{H}}_3/{\mathrm{HI}}_3+\text{en}\%\mathrm{O}/{\mathrm{C}}_{\mathrm{o}}+\text{en}\%{\mathrm{P}}_6/{\mathrm{K}}_5\right)$$ (2)

HC₃ = 3.0, HC₆ = 0.70, PC₃ = 0.0555, PC₆ = 0.0441, HI₃ = 0.005, C_O = 5.0, and K_s = 0.175. For studies that did not report dietary intake of any of the minor nutrients (18:3n-6, 18:4n-3, 20:3n-6, 22:4n-6, 22:5n-6, 22:5n3), these fatty acids were regarded as zero in calculating P₆, P₃, H₆, and H₃ for Equation 2. To compare limited data from studies that did not report dietary intake of major HUFA, we used rough estimates for typical Western diets: 0.07 en% for H₆ and 0.11 en% for H₃ [4]. The reported %n-6 in HUFA was calculated using Equation 1, where the amount of each tissue HUFA was expressed as a percentage of total fatty acids reported in phospholipids or total lipids of plasma, serum, or RBC. Unreported tissue HUFA were regarded as zero in Equation 1.

2.3 Statistical analysis

Data were evaluated for normality of distribution and examined for outliers. Univariate and multivariate Pearson and Spearman linear regression analyses were performed using Statistica 8.0 (StatSoft, Tulsa, OK). P-values less than 0.05 were considered statistically significant.

3. RESULTS

3.1 Literature search

The literature search resulted in 92 study groups from 34 separate studies [13,21–53] conducted in 11 different countries. The complete data are given in an additional file, Table S1. Examples of fatty acid intake and blood composition results are provided in Table 1, illustrating the variety of studies included in this analysis. Studies considered were both interventional (23 studies) [13, 21–42] and observational (11 studies) [43–53], with an average of 69 subjects per group (range 7 to 2114). The interventional studies ranged from 3 to 26 weeks, with an average of 9 weeks.

Table 1.

Dietary fatty acid intake and blood compositions for example data sets.

Reference	Country	N	Duration (weeks)	LA	ALA	n-6 HUFA	n-3 HUFA	Blood Fraction	20:3n-6	20:4n-6	22:4n-6	20:5n-3	22:5n-3	22:6n-3	Reported %n-6 in HUFA1	Predicted %n-6 in HUFA2
				en%					% Total Fatty Acids
[46]	Japan	229	Obs	2.10	0.42	0.07	0.86	S-TL	NR	5.90	NR	3.80	1.34	8.40	30.35	27.21
	Japan	202	Obs	2.54	0.51	0.06	0.50	S-TL	NR	5.30	NR	3.20	1.20	7.70	30.46	37.90
[21]	Sweden	19	4	4.88	0.33	0.05	0.28	S-PL	3.15	8.14	NR	1.37	1.10	4.48	61.90	59.73
	Sweden	19	4	7.33	2.87	0.04	0.27	S-PL	2.82	8.12	NR	1.79	1.13	4.75	58.79	57.52
[18]	US	31	6	5.21	0.57	0.06	0.12	P-PL	3.45	11.62	0.44	0.94	0.97	3.8	73.09	70.85
	US	31	6	3.15	0.31	0.04	0.13	P-PL	3.81	12.10	0.52	0.99	1.01	5.15	69.68	62.70
	US	30	6	5.47	0.64	0.06	0.10	P-PL	3.45	11.19	0.45	0.99	0.98	4.14	71.18	72.48
	US	30	6	5.46	0.71	0.05	0.19	P-PL	3.12	10.83	0.38	1.11	0.84	5.63	65.40	64.38
[37]	France	276	Obs	4.70	0.37	0.08	0.20	P-TL	1.54	7.90	0.21	1.36	0.54	2.49	68.73	66.57
	France	257	Obs	4.96	0.40	0.08	0.21	P-TL	1.51	8.05	0.20	1.29	0.54	2.65	68.54	66.57
[20]	Spain	17	26	8.14	0.12	0.00	0.10	P-PL	NR	10.20	NR	0.81	0.68	3.96	65.18	71.19
	Spain	20	26	6.55	0.12	0.00	0.00	P-PL	NR	12.30	NR	0.84	0.71	3.15	72.35	76.59
[42]	Germany	2114	Obs	5.86	0.75	0.07	0.11	RBC-PL	1.50	13.10	2.70	0.75	2.30	4.70	69.06	73.05

ALA: alpha-linolenic acid; en%: percent of energy; HUFA: highly unsaturated fatty acids; LA: linoleic acid; n-3: omega-3; n-6: omega-6; NR: not reported; O: other fatty acids; Obs: observational study; P-PL: plasma phospholipids; P-TL: plasma total lipids; RBC-PL: red blood cell phospholipids; S-PL: serum phospholipids; S-TL: serum total lipids. % Total Fatty Acids was not reported for 22:5n-6 and 20:3n-9 in these studies.

¹Reported %n-6 in HUFA = (20:3n-6 + 20:4n-6 + 22:4n-6 + 22:5n-6)/(20:5n-3 + 22:5n-3 + 22:6n-3 + 20:3n-6 + 20:4n-6 + 22:4n-6 + 22:5n-6 + 20:3n-9) × 100

²Predicted %n-6 in HUFA = [100][1 + HC₆/en%n-6HUFA/(1 + en%n-3HUFA/HC₃]⁻¹ + [100][1 + PC₆/en%LA(1 + en%ALA/PC₃) + en%n-3HUFA/HI₃ + en%O/C_o + en%LA/K_s]⁻¹; HC₃ = 3.0, HC₆ = 0.70, PC₃ = 0.0555, PC₆ = 0.0441, HI₃ = 0.005, C_o = 5.0, K_s = 0.175

3.2 Dietary and tissue composition data and calculations

Of the 34 studies included in this analysis, 19 studies (48 groups) [21–29, 43–52] published enough data to calculate all four dietary variables in the Lands’ Equation. The correlation for those 19 studies was compared with the overall result from including the remaining 15 incomplete studies (44 groups) [13, 30–42, 53] which did not report dietary intake of n-6 HUFA (and 12 of the latter studies (36 groups) [13, 33–42, 53] also did not report dietary intake of n-3 HUFA). The blood fractions analyzed included plasma phospholipids (15 studies, 52 groups) [13, 21–23, 26, 28–30, 32, 35, 37–40], plasma total lipids (5 studies, 13 groups) [31,43,44,49,50], serum phospholipids (7 studies, 14 groups) [27, 31, 41, 42, 46, 47, 51], serum total lipids (2 studies, 5 groups) [36, 52], RBC phospholipids (2 studies, 3 groups) [48, 53], RBC ethanolamine phosphoglycerides (1 study, 1 group) [45], and RBC total lipids (2 studies, 4 groups) [25, 34]. Only one study (2 groups) [31] included blood composition data for all HUFA. Many studies did not provide values for 20:3n-9 (32 studies, 86 groups) [21–30, 32–53], 22:5n-6 (29 studies, 84 groups) [13, 21–30, 32–36, 38–44, 47–52], 22:4n-6 (23 studies, 68 groups) [22, 23, 25–30, 32–36, 38–42, 44, 47, 49, 50, 52], 20:3n-6 (17 studies, 51 groups) [22, 25, 26, 29, 30, 33–36, 38, 39, 41, 42, 47, 49, 50, 52], and/or 22:5n-3 (10 studies, 35 groups) [23, 25, 28, 29, 33, 36, 41, 42, 46, 49].

Figure 1 compares the observed %n-6 in HUFA derived from blood composition data and the predicted %n-6 in HUFA estimated from dietary data using the Lands’ Equation. The regression analysis including all study groups (n = 92) gave a correlation coefficient of r = 0.73 (p < 0.000000), with a slope of 0.83 and a y-intercept of 9.92. Restricting the analysis to groups with complete dietary data (n = 48) resulted in a higher correlation coefficient (r = 0.78; p < 0.000000) and a slope and y-intercept of 0.94 and 1.33, respectively. In contrast, for groups with incomplete dietary data (n = 44), the correlation coefficient was lower (r = 0.40; p < 0.007), with a slope and y-intercept of 0.35 and 44.97, respectively (data not shown), indicating that using generalized estimates for n-6 and n-3 HUFA intake reduces the accuracy of the Lands’ Equation.

Figure 1.

Correlation between the reported and predicted %n-6 in HUFA. The reported %n-6 in HUFA was calculated from blood fatty acid compositions expressed as percentages of total fatty acids in plasma, serum, or RBC phospholipids or total lipids. The predicted %n-6 in HUFA was calculated from dietary intakes using the Lands’ Equation [13] with revised constants [20]. The solid line represents the regression analysis for study groups with complete dietary data (n = 48, y = 0.94x + 1.33, r = 0.78, p < 0.000000). The dashed line represents the regression analysis for all study groups (n = 92, y = 0.83x + 9.92, r = 0.73, p < 0.000000).

For study groups with complete dietary data (n = 48), we also evaluated each dietary fatty acid component (n-6 HUFA, n-3 HUFA, 18-carbon n-6 PUFA, and 18-carbon n-3 PUFA) as a predictor of the %n-6 in HUFA. Using univariate regression models, only n-3 HUFA intake was significantly correlated with the reported %n-6 in HUFA (r = −0.85; p < 0.000000). We then entered all four dietary fatty acid components into a multivariate regression analysis to evaluate the apparent influence of each component on the observed %n-6 in HUFA. For the range of dietary intakes studied, both n-3 HUFA (β = −0.89) and 18-carbon n-3 PUFA (β = −0.18), but not AA or 18-carbon n-6 PUFA, were significant predictors in the overall model (r = 0.87; p < 0.000000).

4. DISCUSSION

The present analysis provides further evidence that the Lands’ Equation can reliably predict the %n-6 in HUFA in human plasma, serum, or RBC blood fractions using dietary intakes of n-6 and n-3 HUFA and their 18-carbon precursors. Our results confirm that the %n-6 in HUFA is a suitable biomarker for the relative intake of n-6 and n-3 fatty acids in the human diet. In the 1992 study that developed the quantitative, competitive, hyperbolic relationship for humans [13], the authors used an empirical equation with constants that fitted HUFA accumulation in liver, plasma, and RBC of rats [17]. The equation closely estimated the proportions of HUFA in plasma of human subjects who ate little n-6 or n-3 HUFA.

To describe the tissue HUFA of 8 additional patients taking n-3 HUFA supplements, a term was added to the equation to account for competitive interactions of n-6 and n-3 HUFA [13]. At the time, the report noted that: “Future controlled studies of humans with a wide range of en% fat and a widely varied intake of (n - 3) and (n - 6) fatty acids could provide data needed to establish the relationship more firmly.” The present report (like the 1992 report) adds to the growing evidence of extensive similarities between rats and humans in the selectivity of essential fatty acid metabolism.

In 2003, Lands [19] applied the equation to two sets of data from Japanese with a wider range of HUFA in their diets ([47], n = 87; [49], n = 79) and adjusted three of the eight original constants to fit the additional data. The regression line comparing the reported and predicted %n-6 in HUFA values for the combined original and Japanese data [19] had a slope and y-intercept of 0.90 and 1.43 (r = 0.90), respectively, which are similar to the slope and y-intercept found for studies with complete dietary data in the present analysis (slope = 0.94 and y-intercept = 1.33; r = 0.78; Figure 1) using the same constants.

In this study, a high inverse linear correlation (r = −0.85) was observed between dietary intake of n-3 HUFA and the %n-6 in HUFA for subject groups with complete dietary data. This finding is consistent with other reports of significant correlations between dietary intake of n-3 HUFA and the level of n-3 HUFA in plasma or serum [43, 47, 50], since any increase in tissue n-3 HUFA decreases the relative proportion of n-6 HUFA in total tissue HUFA. Multivariate regression analysis of the present data set indicates that dietary intake of 18-carbon n-3 PUFA also associates with the %n-6 in HUFA, although the effect was small (β = −0.18) compared to that for dietary n-3 HUFA (β = −0.89).

The small apparent influence of 18-carbon n-3 PUFA intake on the %n-6 in HUFA, as well as the insignificant apparent influence of 18-carbon n-6 PUFA and n-6 HUFA intakes, can be understood by comparing the diets included in this study. For each diet, LA intake far exceeded the intake of all other essential acids, with an average intake of 6 en% (range 1.77 to 15.16 en%). Our studies following those of Mohrhauer and Holman [14, 15] noted a mid-point of response to dietary abundance near 0.1 en% as reflected in the currently used values for PC₃ (0.0555) and PC₆ (0.044). At levels of LA intake above 1.77 en%, the enzymes involved in PUFA elongation and desaturation were likely saturated with LA and its metabolites. As a result, we do not expect the differences in LA intake among the diets in this study to appreciably affect the accumulated %n-6 in HUFA [16].

The intake of n-6 HUFA was the lowest of any of the four types of n-3 and n-6 nutrients (P₃, P₆, H₃, H₆). Only one subject group consumed more than 0.16 en%, whereas intake of n-3 HUFA ranged from 0 to 0.86 en%. Table 1 (and an additional file, Table S1) illustrate how different ethnic lifestyles give different average intakes of n-3 and n-6 nutrients and cause different outcomes of HUFA balance. To estimate the likely impact of changing any of the four types of nutrient intake, the equation and constants were embedded into a simple spreadsheet calculator ([54], Figure 10 in [16]). For example, lowering P₆ from 6.95 to 2 en% may lower HUFA balance from 80% n-6 in HUFA (near typical USA values) to 64% n-6 in HUFA (near traditional Mediterranean values). Alternatively, raising H₃ intake from 0.03 to 0.3 en% may lower the former outcome of 80 % to 67% n-6 in HUFA whereas it could be expected to lower the latter 64% to 43% n-6 in HUFA.

Importantly, voluntary food intakes differ widely from day to day, and estimates of the likely outcome of metabolic competitions among the four types of essential acids include an assumption that sufficient time with a consistent diet pattern has allowed a steady-state equilibration. There is a long history of unreliable dietary data due to errors in human recall, data base adequacy, etc. The successful fit of predicted and observed values for HUFA balance in Figure 1 makes the empirical relationship an objective clinical biomarker of HUFA balance to serve as a credible surrogate for inadequate dietary estimates.

The relationship between dietary n-3 HUFA and tissue HUFA composition found for most contemporary diets led some researchers to advocate n-3 HUFA biomarkers that neglect the influence of dietary n-6 PUFA. One proposed biomarker is the “Omega-3 Index”, defined as the sum of the percentage of EPA and DHA in total fatty acids of RBC (%EPA + DHA in RBC) [55]. Studies on subjects consuming Western diets have reported strong correlations between the %EPA + DHA in RBC and the %n-3 or the %n-6 in HUFA, suggesting that HUFA biomarkers might be useful in a limited range of dietary PUFA intake [5].

The %n-6 in HUFA parameter has some advantages over the %EPA + DHA in RBC, especially for use in low-cost, high-throughput health risk assessments. First, the %n-6 in HUFA focuses attention on the production of n-6 eicosanoids, a major determinant for many chronic diseases [2, 12]. Second, the value for 100 - %n-6 in HUFA equals the value for %n-3 in HUFA, which includes 22:5n-3, an acid often more abundant than 20:5n-3, EPA and associated with decreased eicosanoid production [56]. Sun et al. [57] concluded from data in the Nurses’ Health Study that 22:5n-3 alone, after total combined n-3 HUFA, was the best predictor of a lower incidence of nonfatal myocardial infarction among US women. Third, the proportions of n-6 or n-3 HUFA in total HUFA are comparable for liver, plasma, serum and RBC in rats [17] and for plasma, serum and RBC in humans [58, 59]. The comparable proportions likely reflect the continual remodelling of fatty acids in RBC phospholipids [60]. The rapid turnover makes proportions of HUFA in RBC respond to diet changes much faster than the overall 120-day lifespan of a RBC.

Moreover, the comparable HUFA proportions in whole blood make it possible to measure the %n-6 in HUFA economically without the need to centrifuge to separate blood fractions or prepare phospholipid fractions as done by others [58]. Fourth, the %n-6 in HUFA does not require integration of peaks that elute before methyl eicosanoate (20:0), resulting in a more rapid chromatographic analysis that is suitable for a high-throughput assays. Fifth, the similar chemical stabilities of HUFA allow each HUFA to serve as an internal recovery standard for other HUFA in the %n-6 in HUFA, regardless of the different stabilities of saturated, monounsaturated, and polyunsaturated acids. Furthermore, although studies have demonstrated that the %EPA + DHA in RBC can change with changed diets, there is no published equation to estimate quantitatively the response of RBC HUFA to a changed diet. In contrast, the Lands’ Equation provides estimates of likely values for the %n-6 in HUFA in multiple blood fractions for a wide range of diets and dietary changes.

The present study had several weaknesses, particularly the limited diversity and accuracy of dietary data available. As discussed above, the limited range of diets available for our analysis did not allow us to evaluate the full spectrum of possibly desirable human diets. Many of the diets analyzed did not report dietary intake of n-6 HUFA and/or n-3 HUFA, and they required the use of rough estimates to estimate tissue HUFA composition with the Lands’ Equation. In addition to incomplete dietary data, most studies did not report the amounts of all the HUFA present in the blood fraction analyzed. This weakness is particularly evident in reports that emphasize the %EPA + DHA in RBC and tend to neglect competing n-6 HUFA. 5.

5. CONCLUSIONS

This study shows that the Lands’ Equation reliably estimates the %n-6 in HUFA of plasma, serum, and RBC using the average dietary intake of n-6 and n-3 HUFA and their 18-carbon precursors. Future studies should confirm these findings and further assess the Lands’ Equation and its constants on a wider range of diets, particularly diets with low LA content. In addition, with the recognized causal role for n-6 arachidonate in chronic inflammatory and thrombotic disorders [61], it would be valuable to evaluate the %n-6 in HUFA as a biomarker for monitoring nutritional interventions in cardiovascular and other chronic inflammatory diseases.

Highlights.

• Blood compositions of omega-3 and omega-6 highly unsaturated fatty acids (HUFA) are biomarkers useful for estimating disease risks.

• Estimations of n-3 and n-6 HUFA compositions form dietary intakes typically do not include the contribution of dietary 18 carbon n-3 and n-6 precursor fatty acids.

• Circulating HUFA compositions can be reliably estimated from dietary intake data that not only includes n-3 and n-6 HUFA consumption, but also includes consumption of 18 carbon n-3 and n-6 precursor fatty acids.

List of abbreviations

AA

arachidonic acid

ALA

alpha-linolenic acid

DHA

docosahexaneoic acid

GLA

gamma-linolenic acid

en%

percent of energy

EPA

eicosapentaenoic acid

%EPA + DHA in RBC

percentage of EPA and DHA in total fatty acids of red blood cell

HUFA

highly unsaturated fatty acids

LA

linoleic acid

NR

not reported

n-3

omega-3

n-6

omega-6

%n-6 in HUFA

the percentage of omega-6 highly unsaturated fatty acids in total tissue HUFA

P-PL

plasma phospholipids

P-TL

plasma total lipids

PUFA

polyunsaturated fatty acids

RBC

red blood cell

RBC-PL

red blood cell phospholipids

S-PL

serum phospholipids

SDA

stearidonic acid

S-TL

serum total lipids