SUMMARY

1. Introduction

The health benefits of cardiovascular fitness are well known, and it has been shown that increased levels of exercise are inversely associated with morbidity and all-cause mortality. Extensive work has also shown that exercise decreases BMI, waist circumference, and adiposity [1]; however the beneficial effects of exercise on cardiovascular and metabolic disease [2], extend beyond changes in weight reduction [3] and involve modifications in immunity and inflammation [4,5]. This notion is supported by data showing that moderate intensity exercise prevents the development and delays the progression of chronic diseases by decreasing chronic low-grade inflammation [6]. Although exercise stimulates the production of inflammatory cytokines acutely [7], frequent exercise and physical activity are associated with lower levels of C-reactive protein (CRP) [8-11], a commonly used marker of systemic inflammation. This negative association is further enhanced by reductions in BMI [1].

Cardiovascular fitness, typically measured as maximal oxygen uptake (VO₂ max), is also inversely associated with CRP levels and inflammation [12]. After controlling for changes in traditional cardiovascular risk factors, beneficial changes in inflammatory markers [13], including CRP, have been shown to contribute the most to the cardiovascular risk-reducing effects of exercise [5]. Nonetheless, the beneficial effects of exercise vary among individuals with different health and nutritional status, and the biological factors responsible for such heterogeneity in responses to exercise and cardiovascular fitness remain largely unknown.

Omega-3 polyunsaturated fatty acids (PUFA), such as docosahexaenoic acid (DHA) and eicosapentaenoic acid (EPA), have been shown to have anti-inflammatory properties [14] leading to a lower risk for development of chronic inflammatory diseases [15]. This is supported by results from the REDUCE-IT trial, whereby treatment with EPA ethyl ester significantly reduced cardiovascular events in individuals on statin therapy but with residual hypertriglyceridemia [16]. Beneficial changes in cardiometabolic factors with reduced CRP levels have also been found in a double-blind randomized trial of individuals with metabolic syndrome receiving omega-3 PUFA supplementation and exercise [17]. Nonetheless, whether endogenous omega-3 PUFA levels contribute to the health benefits of cardiovascular fitness remains unknown.

Work from our laboratory and others has shown that omega-3 PUFA-derived specialized pro-resolving lipid mediators (SPMs) that actively resolve inflammation [18], decrease atherogenesis and prevent high fat diet-induced insulin resistance in rodent models [19-24]. Therefore, we postulated that higher intake of omega-3 PUFAs, by facilitating resolution of inflammation, may enhance the salutary effects of exercise. Using data from the from the National Health and Nutrition Examination Survey (NHANES) to test this hypothesis, we examined whether plasma levels of omega-3 PUFAs modify the association between cardiovascular fitness and inflammation, sought to determine the role of individual omega-3 PUFAs, and to identify populations most sensitive to the modifying influence of omega-3 PUFAs on cardiovascular fitness.

2. Materials and methods

The authors declare that all supporting data are available within the article and its Data Supplement.

3. Results

Table 1 presents the demographic and characteristics of the study population. Participants had an average age of 32.9 years, with a slightly higher percentage of male participants (55.4%), and were 49.2% were white, 19.6% black, and 31.3% other races. The average BMI of participants was 27.8, and 29.5% were considered current smokers. The mean CRP levels were 0.3 ± 0.5 mg/dL with mean VO₂ max of 39.2 ± 8.6 mL/kg/min.

Table 1

Demographic characteristics of the study population.

Characteristic	Total (n = 435)
continuous	Mean ± SD
Age (yrs)	32.9 ± 8.3
BMI	27.8 ± 6.0
Systolic BP (mmHg)	115.0 ± 12.4
Diastolic BP (mmHg)	70.0 ± 10.0
Total Cholesterol (mg/dL)	192.3 ± 37.8
HDL (mg/dL)	53.7 ± 15.7
LDL (mg/dL)	114.3 ± 33.6
Triglyceride (mg/dL)	127.4 ± 106.9
C-Reactive Protein (mg/dL)	0.3 ± 0.5
VO2 Max (ml/kg/min)	39.2 ± 8.6
categorical	n (%)
Sex (Male)	241 (55.4%)
Race
White	214 (49.2%)
Black	85 (19.6%)
Other (total)	136 (31.3%)
Mexican American	103 (23.7%)
Other Hispanic	18 (4.4%)
Other Race or multiracial	15 (3.5%)
Smoking status	128 (29.5%)
Education Level
<High School	86 (19.8%)
H.S. Diploma	119 (27.4%)
>High School	230 (52.9%)
Income
<$25,000	116 (26.9%)
$25,000-$74,999	191 (44.2%)
>$75,000	125 (29.0%)

Open in a separate window

Summary statistics for plasma levels of 24 fatty acids (FA), including saturated fatty acids (SFAs), monounsaturated fatty acids (MUFAs), and polyunsaturated fatty acids (PUFAs), and the summation of fatty acid groups are shown in Table 2. The sum of omega-3 PUFAs had the lowest mean value of 281.6 μmol/L, compared with a mean sum of 3660.1 μmol/L for SFAs.

Table 2

Summary statistics for fatty acid levels (μmol/L).

	Fatty Acid	Mean ± SD	n
SFA	Myristic Acid	130.2 ± 86.5	431
	Palmitic Acid	2688.4 ± 907.6	434
	Stearic Acid	680.1 ± 171.0	434
	Arachidic Acid	22.9 ± 5.0	423
	Docosanoic Acid	68.1 ± 15.7	421
	Lignoceric Acid	54.0 ± 13.0	420
MUFA	Myristoleic Acid	8.3 ± 8.4	434
	Palmitoleic Acid	237.3 ± 199.5	435
	cis-Vaccenic Acid	143.5 ± 56.0	425
	Oleic Acid	2088.3 ± 1140.5	433
	Eicosenoic Acid	13.7 ± 6.2	433
	Docosenoic Acid	4.4 ± 3.5	385
	Nervonic Acid	75.9 ± 18.8	418
PUFA
Omega-6	Linoleic Acid	3482.1 ± 834.4	435
	gamma-Linolenic Acid	48.0 ± 22.7	432
	Eicosadienoic Acid	22.3 ± 10.2	433
	homo-gamma-Linolenic Acid	155.5 ± 63.6	435
	Arachidonic Acid	780.7 ± 199.3	435
	Docosatetraenoic Acid	26.1 ± 8.6	435
	Docosapentaenoic-6 Acid	21.4 ± 8.8	435
Omega-3	alpha-Linolenic Acid	67.2 ± 40.0	435
	Eicosapentaenoic Acid	43.1 ± 28.6	434
	Docosapentaenoic-3 Acid	41.3 ± 13.3	435
	Docosahexaenoic Acid	129.7 ± 61.0	435
		Mean ± SD	n
∑ SFA		3660.1 ± 1160.3	409
∑ MUFA		2584.0 ± 1169.9	375
∑ PUFA		4829.4 ± 1058.9	429
∑ Omega-6		4544.5 ± 992.1	430
∑ Omega-3		281.6 ± 113.8	434
∑ Omega-6/∑ Omega-3		17.5 ± 4.8	429

Open in a separate window

Linear regression models were used to examine the relationship between VO₂ max and CRP levels (Fig. 1A). For every 10 mL/kg/min increase in VO₂ max, we observed 17.4% lower levels of CRP, suggesting that individuals with higher cardiovascular fitness have lower levels of systemic inflammation in this sample population, which is consistent with other published data.

Open in a separate window

Fig. 1.

Predicted mean of CRP and VO₂ max (A), Association between CRP and sums of fatty acid group tertiles (B). Multivariable linear regression models were used to examine the relationship between CRP with VO₂ max and fatty acid group tertiles. CRP was log transformed, and all models were adjusted for age, sex, BMI, and covarying fatty acid group based on Pearson's correlation coefficient (e.g., saturated fatty acids for all groups except saturated fatty acid model, which was adjusted for monounsaturated fatty acids). Beta represents change in log CRP per 1 unit of VO₂ max. Bars represent 95% confidence intervals. * Significantly different than reference (lowest tertile).

Next, to assess whether circulating levels of FAs are independently associated with levels of systemic inflammation, we performed linear regression models with tertiles of FAs, adjusting for covarying FA groups (Fig. 1B, Table S1, Fig. S1). Using the low tertile as the reference, we found that medium and high levels of both SFA and MUFAs were associated with higher CRP levels. Conversely, we found that higher levels of PUFAs are associated with significantly lower CRP (Fig. 1B). We then subdivided the PUFA class into omega-6 and omega-3 FAs and found lower levels of CRP (Fig. 1B), suggesting that omega-3 and omega-6 fatty acids contribute to reductions in inflammation. Given the possible clinical relevance of omega-6:omega-3 ratio in cardiovascular disease and its association with all-cause mortality [27], we also assessed and found higher CRP levels associated with higher tertiles of omega-6:omega-3 ratio when compared with low (Fig. 1B).

Given our observation that circulating levels of fatty acids affect the level of inflammation and that inflammation is also related to cardiovascular fitness, we next sought to determine whether the relationship between VO₂ max and CRP is modified by circulating levels of FAs (Fig. 2). In multivariable regression analysis, we found a significant negative interaction between tertiles of omega-3 PUFAs and VO₂ max, suggesting that the inverse association between VO₂ max and CRP is strengthened in individuals with higher omega-3 PUFA levels. No such interaction on VO₂ max with CRP was found with SFA, MUFA, omega-6 PUFA levels, or with the omega-6/omega-3 ratio.

Open in a separate window

Fig. 2.

Association between VO₂ max and CRP stratified by tertiles of fatty acid groups. Results are presented as percent difference and 95% confidence intervals per 10 mL/kg/min of VO₂ max. Solid black line indicates the percent difference between VO₂ max and CRP in the full sample. CRP was log-transformed for normality. Linear regression models were used to test for associations between log-CRP and VO₂ max or fatty acid group tertiles. All models were adjusted for age, sex, BMI, and covarying fatty acid group based on Pearson's correlation coefficient (e.g., saturated fatty acids for all groups except saturated fatty acid model, which was adjusted for monounsaturated fatty acids). * Represents significant interaction between VO₂ max and fatty acid tertile (ref = low).

When we stratified plasma levels of omega-3 PUFAs by tertiles, we found that the medium tertile showed the strongest association between VO₂ max and CRP, with a −32.0% difference (95% CI: −44.95, −15.9%) per 10 mL/kg/min of VO₂ max (Fig. 2). Conversely, within the low omega-3 PUFA category we found no association between VO₂ max and CRP (4.0%; 95% CI: −16.8%, 30.0%), which nullified the negative association of CRP and VO₂ max (Fig. 2). Although not statistically significant, we also found that those with higher levels of omega-3 PUFAs had stronger inverse associations between VO₂ max and CRP when compared with low tertile individuals. Taken together, these data suggest that the inverse relationship between VO₂ max and CRP depends upon circulating levels of omega-3 PUFAs.

Although we found clear effect modification of the relationship between VO₂ max and CRP by circulating levels of omega-3 FAs, this relationship could be affected by the presence of other fatty acids in the plasma. To assess the impact of other circulating fatty acids on the effects of omega-3 PUFAs, we performed an exploratory factor analysis of 24 individual fatty acids to identify how multiple Fas covary (Table 3). This analysis revealed five distinct FA factors, or dietary profiles (FA Profiles). FA Profile 1 accounted for 56.9% of the total variance in the data and was comprised of all fatty acid classes: SFA, MUFA, PUFA. FA Profile 2 accounted for 14.3% of the total variance, and was similarly comprised as FA Profile 1, however with less contribution from omega-6 and omega-3 PUFAs. FA Profile 3 was comprised predominantly of SFAs, while FA Profile 4 was predominantly MUFAs and omega-6 PUFAs. FA Profile 5 accounted for 5.3% of the total variance and consisted predominantly of omega-3 PUFAs.

We next assessed how these five FA profiles associated with CRP levels by grouping the profiles into tertiles and performing multivariable linear regression (Fig. 3). We found no significant associations between CRP and FA Profile 1, 3, or 4. However, for FA Profile 2, comprised of all FA groups but with low PUFAs, we found that participants in the highest tertile had significantly higher CRP levels than the lowest tertile. Interestingly, FA Profile 5, which is primarily comprised of omega-3 PUFAs without SFAs, we found a significant negative association, where the medium tertile had −31.9% lower CRP levels as compared with the lowest tertile (Fig. 3).

Open in a separate window

Fig. 3.

Association between CRP and exploratory factor analysis profiles (tertiles). Results are presented as percent difference and 95% confidence intervals as compared to the lowest tertile. P value for trend are included for each fatty acid profile. CRP was log-transformed for normality. Linear regression models were used to test for associations between log-CRP and factor analysis tertiles. All models were adjusted for age, sex, and BMI. * Significantly different than reference (lowest tertile).

We next sought to understand whether these FA Profiles modify the association between CRP and VO₂ max. Figure 4 presents the associations between VO₂ max and CRP stratified by tertiles of FA profiles. We found little evidence of effect modification for FA Profile 1 through 4. However, for FA Profile 5, largely comprised of omega-3 PUFAs without SFA, we found evidence of a negative interaction, which was strongest in the medium tertile of FAs with a −28.2% difference in CRP (95% CI: −43.4, −8.9) per 10 mL/kg/min of VO₂ max (p-value = 0.058; Fig. 4). Importantly, we found no significant association between VO₂ max and CRP in the lowest tertile for FA Profile 5; 1.0% difference in CRP (95% CI: −19.5%, 26.7%) per 10 mL/kg/min of VO₂ max, suggesting that the inverse association between cardiovascular fitness and inflammation was abolished in individuals with low levels of omega-3 fatty acids.

Open in a separate window

Fig. 4.

Associations between CRP and VO₂ max stratified by factor analysis tertiles. Results are presented as percent difference and 95% confidence intervals per 10 mL/kg/min of VO₂ max. Solid black line indicates the percent difference between VO₂ max and CRP in the full sample. CRP was log-transformed for normality. Linear regression models were used to test for associations between log-CRP and VO₂ max. All models were adjusted for age, sex, race, and BMI.

Given our results showing that the inverse relationship between VO₂ max and CRP is modified by omega-3 PUFA levels (Fig. 2) along with the complimentary results of the exploratory factor analysis (Table 3 and Fig. 3), we questioned the involvement of individual omega-3 PUFAs on this relationship. To address this, we assessed how individual omega-3 PUFAs interact with VO₂ max. This analysis indicated that the omega-3 PUFA alpha-linolenic acid significantly modified the negative association between VO₂ max and CRP levels (Fig. 5A). Interestingly, we also found that the omega-3 PUFA docosahexaenoic acid (DHA), but not eicosapentaenoic acid (EPA) nor docosapentaenoic acid (DPA_n-3), significantly modified the negative association between VO₂ max and CRP (Fig. 5A). Of note, none of the omega-6 PUFAs significantly altered the negative association between VO₂ max and CRP (Fig. 5B).

Open in a separate window

Fig. 5.

Associations between CRP and VO₂ max stratified by Individual omega-3 PUFA (A) and omega-6 PUFA (B) tertiles. Results are presented as percent difference and 95% confidence intervals per 10 mL/kg/min of VO₂ max. Solid black line indicates the percent difference between VO₂ max and CRP in the full sample. CRP was log-transformed for normality. Linear regression models were used to test for associations between log-CRP and VO₂ max. All models were adjusted for age, sex, and BMI. The alpha-linolenic acid model was additionally adjusted for docosahexaenoic acid. The docosahexaenoic acid model was additionally adjusted for alpha-linoleic acid. * Represents significant interaction between VO₂ max and fatty acid tertile (ref = low).

CBC data stratified by levels of DHA, but not alpha-linolenic acid (ALA) display significant differences in the percentage of circulating lymphocyte and segmented neutrophils (Tables S2 and S3), with higher levels of lymphocytes in with higher levels of DHA (Table S3). Conversely, segmented neutrophils were lower in participants with the highest levels of (DHA) as compared to the lowest tertile (Table S3). These results suggest that levels of ALA and DHA significantly modify the inverse relationship between VO₂ max and CRP, potentially through modifications in circulating lymphocyte and neutrophil cell populations.

We next sought to determine whether there were any specific subgroups that were more sensitive to the modifying effect of omega-3 PUFAs on the relationship between VO₂ max and CRP. Figure 6 shows the strength of the interaction between omega-3 PUFA and VO₂ max in participant subgroups. We found that the modifying effect of omega-3 PUFAs between VO₂ max and CRP was significant in males, White participants, participants of higher age (33–49 years) and in participants with low BMI (<30). Surprisingly, we found a near significant positive interaction between omega-3 PUFAs and VO₂ max in participants with high circulating levels of SFAs (p-value = 0.053). Participants with high plasma levels of SFAs also had higher omega-3 PUFAs compared with participants with low levels of SFAs (Fig. 6), suggesting that the benefits of omega-3 PUFAs may be decreased in those who also have high circulating levels of SFAs. Taken together, these findings suggest that the benefits of omega-3 PUFAs on the relationship between VO₂ max and CRP is more pronounced in males, Whites, older ages, individuals with lower BMI, and individuals with a low SFA diet.

Open in a separate window

Fig. 6.

Subgroup analysis of the interactions between omega-3 PUFA tertiles and VO₂ max in predicting CRP. The mean ± S.D. for the sum of omega-3 PUFA levels for each subgroup. Linear regression models were used to test for interactions between VO₂ max and omega-3 PUFA tertiles. All models were adjusted for age, sex, race, and BMI unless they were used as a stratification variable. Results are presented as percent difference and 95% confidence intervals for the interaction term in predicting CRP, using the low omega-3 group as a reference.

4. Discussion

In this study we found that cardiovascular fitness was inversely related to the levels of CRP, suggesting that those with higher levels of cardiovascular fitness have lower levels of inflammation. We found that higher level of omega-3 PUFAs modified the relationship between cardiovascular fitness and inflammation, whereas the relationship was not affected by circulating levels of SFAs, MUFAs, or omega-6 PUFAs. More specifically, we found that individuals who have low circulating levels of omega-3 PUFAs have a reduced inverse association with inflammation and higher levels of cardiovascular fitness. Looking at the contribution of different types of omega-3 PUFAs, we found that plasma levels of alpha-linolenic acid and docosahexaenoic acid, but not eicosapentaenoic acid or docosapentaenoic acid, significantly strengthen the negative association between cardiovascular fitness and chronic inflammation.

Previous work has shown that increased cardiovascular fitness and participation in exercise is associated with improvements in inflammatory markers and decreased levels of CRP [28]. These beneficial changes in CRP levels with increasing cardiovascular fitness persist even after controlling for factors such as obesity, blood pressure, and smoking [29], suggesting that exercise affects inflammation independently of its effects on blood pressure, obesity, and smoking. However, even though exercise has pervasive effects on both cardiovascular fitness and inflammation, individuals respond differently to exercise, even under controlled conditions. This disparity leads to variable outcomes following exercise in different individuals, and some individuals – “non-responders” – see little or no benefits of exercise [30]. Nonetheless, biological mechanisms that underlie this variability in individual responses to similar exercise protocols remains unclear but may relate to health and nutritional status. Therefore, a clear understanding of the factors that modify the salutary effects of exercise are needed to expand the potential health benefits of exercise across different demographic segments of the population. The results of our study reported here suggest that the intake of omega-3 PUFAs may be an important factor such that those with higher levels of omega-3 PUFAs are more likely to have lower levels of inflammation upon increasing cardiovascular fitness. Moreover, based on results from our factor analysis and subgroup analysis, we believe that higher levels of omega-3 PUFA in conjunction with low levels of SFA are important to obtain the anti-inflammatory benefit of exercise. Nonetheless, additional studies are needed to further understand how SFAs negatively affect the beneficial effect of omega-3 PUFAs.

Extensive work has shown that higher levels of omega-3 PUFAs are associated with lower levels of systemic inflammation [31,32]. This link may not be surprising because omega-3 PUFAs serve as precursors for proresolving lipid mediators that are agonists of resolution during acute inflammatory events [18], and therefore higher intake and higher circulating levels of these fatty acids are likely to promote the synthesis of products that lower and resolve inflammation. In this context, our study provides additional evidence for the association between omega-3 PUFAs and a decrease in the levels of inflammation markers such as CRP. These findings are consistent with results recently published by our laboratory in a pre-clinical murine model showing enhanced production of the DHA-derived proresolving mediator resolvin D1 and the resolution of acute inflammation following exercise training [33]. Consistent with this, human studies also show that dietary omega-3 PUFA supplementation decreases the production of pro-inflammatory cytokines TNFα and IL-6 [34]. Therefore, it is tempting to speculate that exercise promotes an anti-inflammatory effect by increasing the synthesis of proresolving mediators, where moderate to high levels of omega-3 PUFAs are required to fully redeem the anti-inflammatory potential of exercise.

The levels of ALA, EPA, DPA, and DHA in the blood are related to dietary consumption, while the essential fatty acid ALA also serves as the precursor for polyunsaturated fatty acids, including DHA, EPA, and DPA following the enzymatic activity of desaturases and elongases [35]. Previous work by Rocha-Rodrigues et al., show that with exercise there is an increase in the expression of FADS1 and Elongase 5 [36]. These findings support the notion that exercise training increases the formation of PUFA, likely to aid in the inhibition of acute inflammatory responses to exercise training. Our findings that circulating levels of omega-3 PUFAs, ALA and DHA, significantly modify the interaction between cardiovascular fitness and inflammation are consistent with this concept and further extend the importance of adequate dietary omega-3 PUFA intake.

In this study we used data from NHANES 2003–2004, which is a comprehensive study that has a random and generalizable study population, however because we combined two separate subsamples of participants with both FA levels and VO₂ max measurements, we did not perform survey weighted analyses which would allow our results to be generalizable to the United States population. Nonetheless, limitations of our study include that plasma fatty acid levels were only measured in select individuals thus limiting the size and scope of the population. Moreover, treadmill testing to assess VO₂ max in NHANES 2003–2004 was restricted to individuals between 12 and 49 years of age, thus limiting the extrapolation of our findings to individuals aged 50 or older. Additionally, our study is limited by the lack of measurements for additional inflammatory markers, such as measures of circulating inflammatory cytokines, including IL-6 or TNFα, or functional assessments of innate immunity, which are known to be altered by exercise [37]. These additional data would provide further insights into how omega-3 PUFAs modify the relationship between inflammation and VO₂ max. Nonetheless, other studies focus mainly on survey data for FA intake assessment, which allows for greater power and coverage, but is susceptible to exposure misclassification of data due to lack of absolute measurements. With the use of laboratory data in our study - VO₂ max, CRP levels, and plasma fatty acid concentrations - the uncertain reliability often seen with self-reported answers was eliminated, therefore increasing confidence in biological significance, nonetheless, the cross-sectional study design is susceptible to residual confounding.

To our knowledge, this study is the first to examine whether omega-3 PUFAs modify the inverse association between inflammation and cardiovascular fitness. Few studies have tested whether fatty acids modify the relationship between exercise and improved health outcomes. Nonetheless, it has been reported that individuals with metabolic syndrome supplemented with omega-3 PUFAs subjected to exercise exhibit improved serum CRP levels in comparison with control individuals [17]. Despite being limited to a small group of individuals with metabolic syndrome, these findings are consistent with the findings of our study showing that omega-3 fatty acids modify the relationship between cardiovascular fitness and inflammation. Interestingly, we found that high circulating levels of saturated fatty acids attenuate the interaction between omega-3 PUFA and VO₂ max on CRP levels. This observation may explain why we detected the strongest interactions in the medium tertiles as individuals in the highest tertile of omega-3 PUFAs also exhibited the highest levels of SFAs. These findings suggest that high levels of SFAs may attenuate the beneficial effect of omega-3 fatty acids on the relationship between cardiorespiratory fitness and inflammation. Nonetheless, future studies directly aimed at understanding how SFAs impinge on the beneficial effect of omega-3 FAs on inflammation with exercise are needed. Collectively, these findings however provide insights into the role of FA on the cardiovascular benefits of exercise and may help to explain why individuals respond differently to exercise.

Supplementary Material

Acknowledgements

Footnotes

References