Intuitive eating and biomarkers related to cardiovascular disease in older adults

Abstract

Objective:

Examine whether exercise and diet motivation are associated with 4 biomarkers related to cardiovascular disease (CVD).

Design:

Cross-sectional analysis. Data collection involved questionnaires, blood draws, body composition assessments, and accelerometry.

Setting:

Small, midwestern college town.

Participants:

Community older adults (≥ 58 years of age; n = 79) recruited through convenience sampling; the sample was representative of the population of interest for some demographic characteristics (e.g., age and sex) but not representative on other characteristics (e.g., level of activity).

Variables Measured:

Independent variables comprised self-reported intrinsic exercise motivation (Behavioral Regulation for Exercise Questionnaire-3) and intuitive eating (Intuitive Eating Scale-2). Dependent variables included inflammatory proteins (C-reactive protein [CRP] and Interleukin-6) and lipid levels (LDL/HDL and triglycerides) quantified from blood samples. Covariates included age, body mass index, sex, and objective physical activity measured by accelerometers worn for 7 days.

Analysis:

Multiple linear regression was used to assess the association between diet and exercise motivation and biomarker outcomes; we analyzed 4 regression models (one for each biomarker). Significance level p < .05.

Results:

Greater intuitive eating was associated with a lower LDL/HDL ratio (β = −.45, p = .001) and lower triglycerides (β = −.37, p = .003). Intrinsic exercise motivation was not associated with the biomarkers.

Conclusions and Implications:

Intuitive eating may be a key determinant of certain biomarkers and could be a viable target for interventions to help decrease risk of CVD among older adults.

Introduction

Cardiovascular disease (CVD) is a widespread problem, particularly among older adults who are becoming the largest sector of the population.¹ Older adults at risk for CVD are often encouraged to make lifestyle changes that include improvements to their physical activity and diet. Two parallel, and sometimes connected, motivating factors for these health behaviors include intrinsic exercise motivation and intuitive eating.² The purpose of this study was to examine how motivation for two key health behaviors relate to inflammatory and cholesterol biomarkers known to be related to CVD. Results could inform interventions to prevent and manage CVD among older adults.

The Centers for Disease Control and Prevention cites heart disease as the leading cause of death,³ with an estimated 121.5 million American adults (48%) having one or more types of CVD.⁴ For those aged 60–79 years-old, 77% of males and 78% of females have CVD, and for the 80+ year-old age group, 89% of males and 92% of females have CVD.⁵ In 2016, 64% of all CVD deaths occurred in people aged 75 and older.⁴

Older adults with known CVD risk factors are typically encouraged to make significant lifestyle changes, such as improving their physical activity and diet. Physical activity and diet are two health behaviors important for older adults since both have the potential to prevent, reduce, and reverse some diseases.^6,7 Evidence suggests that physical activity correlates with decreased risk of CVD,⁸ whereas sedentary lifestyles relate to increased risk of CVD.⁹ Poor nutrition also significantly contributes to many chronic health conditions, including CVD.¹⁰ Understanding older adults’ motivation for physical activity and diet is vital in helping them sustain healthy behaviors and ultimately reduce their risk for CVD.

Self-Determination Theory (SDT¹¹) has frequently been used to study motivation for physical activity behavior.^12,13 The SDT posits that motivation is multidimensional and exists on a continuum varying from highly controlled (external and introjected) to more self-determined (identified, integrated, and intrinsic) regulations that ultimately influence behavior.¹⁴ In comparing those whose motivation is authentic (i.e., self-endorsed) to those who are externally controlled for an action, studies have found that the former are more likely to sustain healthy behaviors.¹⁵ Intrinsic exercise motivation is consistently associated with higher levels of activity as well as positive psychological and health outcomes.^12,13,16 In a meta-analysis, SDT constructs predicted markers of physiological and molecular health, including healthier cholesterol levels.¹² Among older adults in particular, SDT constructs were significantly associated with high-density lipoprotein (HDL) levels¹⁷ and older adults who reported lower levels of exercise motivation had high levels of systemic inflammation.¹⁸ It is possible that physical activity may mediate the association between motivation and health markers (i.e., cholesterol, inflammation), but the studies above only assessed direct associations.

Some research also suggests that those who report intrinsic motivation to exercise are more likely to eat intuitively.² Intrinsic motivation for exercise mirrors the intuitive eating paradigm, which is one approach to understanding diet motivation. Traditional diet programs that focus on decreasing caloric intake are generally ineffective in terms of long-term health outcomes, and they often promote psychological distress.¹⁹ As a result, increasing attention has been given to non-diet approaches that are less focused on weight and restriction, such as intuitive eating. Intuitive eating is based on four main principles: unconditional permission to eat desired food when hungry; eating for physical rather than emotional reasons; reliance on internal hunger and satiety cues to determine when and how much to eat; and body-food choice congruence or choosing foods that feel good and energize the body.^20,21

Three systematic reviews suggested that non-diet approaches have consistent positive psychological effects but reported mixed findings regarding physical health markers such as blood pressure, blood glucose, and cholesterol.^19,22,23 In the general population, intuitive eating has correlated with lower weight and lower BMI.²³ The few studies that have examined the association between intuitive eating and non-weight related physical health markers have focused on clinical samples (e.g., overweight/obese), college students, and women.^24–30 For example, Hawks et al.²⁷ found that intuitive eating was related to lower triglyceride levels and improved cardiovascular risk among college women. Overall, the literature suggests that intuitive eating is negatively associated with BMI, positively associated with several psychological indicators (e.g., self-esteem), and possibly associated with various biomarkers,^{19, 22–30} although research has often been limited to clinical, female, and college-aged samples.

High levels of systemic inflammation and unhealthy cholesterol are two risk factors for CVD.^31–33 Among various inflammatory biomarkers, high levels of C-reactive protein (CRP) and interleukin-6 (IL-6) are frequently associated with higher risk for adverse heart conditions.^34,35 Unhealthy lipid levels (i.e., low levels of HDL cholesterol, high levels of low-density lipoprotein (LDL) cholesterol, and high levels of triglycerides) also increase risk of CVD among older adults.^36,37 Specifically, the LDL/HDL ratio has been shown to be a more accurate predictor of CVD risk than LDL or HDL alone³⁸ and elevated triglyceride levels are associated with increased CVD, sometimes independently of cholesterol levels.³⁹

Routine physical activity and dietary patterns can affect inflammatory markers^40,41 and cholesterol levels^7,42. One meta-analysis found a combined influence of physical activity and diet improvements, with beneficial effects on triglycerides and cholesterol levels.⁴³ Despite associations between these two health behaviors and CVD-related biomarkers, few studies have directly linked behavioral motivation for physical activity and diet to key biomarkers.

The purpose of the present study was to assess the link between intrinsic exercise motivation, intuitive eating, and biomarkers known to be important for CVD. See Figure 1 for the proposed conceptual framework. This study adds to the scant literature on the link between behavioral motivation (i.e., intrinsic exercise motivation and intuitive eating) and physical health markers. Given the paucity of research on behavioral motivation and biomarkers, we also explored the strength of associations to determine if intuitive eating or exercise motivation was a stronger predictor of CVD-related biomarkers. The focus of this study was on older adults, a population at high risk of CVD but for whom intuitive eating specifically has not been widely studied. A better understanding of the potential beneficial effects of intuitive eating and exercise motivation on cardiovascular risk factors among older adults could inform future interventions that ultimately decrease CVD.

Figure 1. Conceptual framework representing proposed associations between behavioral motivation, health behaviors, biomarkers, and cardiovascular disease.

Note. This study focused on the association between behavioral motivation (intuitive eating and intrinsic exercise motivation) and biomarkers. Measurement of health behaviors and actual cardiovascular disease was not examined. CRP = C-reactive protein; IL-6 = lnterleukin-6; LDL= Low-density lipoproteins; HDL = High-density lipoproteins; CVD = Cardiovascular disease

Methods

Participants

The Miami University Institutional Review Board approved this study through expedited review procedure, and participants provided written informed consent. Participants were recruited via flyers, social media, email, community events, Silver Sneakers classes, and word of mouth from a small, Midwestern college town and surrounding areas. Recruitment thus included convenience and snowball sampling. We first screened participants over the phone to ensure they met the study criteria. Inclusion criteria was being ≥ 58 years of age and willingness to participate. Exclusion criteria comprised history of falls (≥2/yr.); physical dependence; significant cardiovascular, metabolic, kidney or lung disease as determined by their primary care doctor; active cancer; recent treatment with anabolic steroids or corticosteroids; alcohol or drug abuse; tobacco use; or prescription anticoagulant use. Participants indicated their sex by selecting male or female. All data were cross-sectional.

Procedures

After agreeing to partake in the study, we asked participants to abstain from strenuous activity for three days prior to their lab visit. On the day prior to obtaining the blood sample, participants refrained from taking over-the-counter medications/supplements such as aspirin, ibuprofen, acetaminophen, fish oil, flax seed oil, and/or omega-3 fatty acid supplements as these influence inflammation. Participants fasted after 11 PM the night before their visit until after the visit the following day. Participants verbally confirmed their fasting upon arrival to the lab.

Participants reported to the Muscle Physiology Lab to undergo body composition testing and blood sampling. Participants completed questionnaires by themselves in a private room in the lab. After completing the questionnaires, participants had their blood drawn by faculty experienced in phlebotomy. We collected objective physical activity data and participants’ BMI, due to their associations with motivation, intuitive eating, inflammation, and cholesterol.^11,19,44,45

Measures

Intuitive eating.

We measured intuitive eating using the 23-item Intuitive Eating Scale-2 (IES-2).²¹ Prior studies have found support for the construct validity and internal consistency for the IES-2 among college students,²¹ community samples,⁴⁶ and older women.⁴⁷ Participants rated their agreement to each statement (e.g., “I trust my body to tell me how much to eat”; “I mostly eat foods that give my body energy and stamina”) on a 5-point scale (1 = “Strongly disagree”; 5 = “Strongly agree”). Higher scores indicated greater levels of intuitive eating. Responses were averaged for a total intuitive eating score that could range from 1 – 5. In the current sample, Cronbach α was .88.

Exercise motivation.

We used the Behavioral Regulation for Exercise Questionnaire-3 (BREQ-3)^48,49 to assess exercise motivation. Using a 0 to 4 scale (0 = “Not true for me”; 4 = “Very true for me”), participants indicated to what extent each of the 24 statements (e.g., “I exercise because it’s fun”; “It’s important to me to exercise regularly”) were true. BREQ-3 consists of six subscales: amotivation, external regulation, introjected regulation, identified regulation, integrated regulation, and intrinsic regulation. The BREQ-3 has demonstrated excellent internal consistency (α > 0.87) and criterion and factorial validity for middle-aged and older adults.^48,50 The BREQ-3 has been scored either as a multidimensional instrument with separate scores for each subscale, or as a unidimensional index known as the relative autonomy index that is calculated by weighting and aggregating each subscale score to indicate the degree of self-determination; both approaches are supported by the established validity of the scales.^48,51 The current study used only the intrinsic regulation subscale (4 items) to parallel the measure of intuitive eating.² Intuitive eating captures the more intuitive or intrinsic aspects of diet, rather than non-intuitive aspects included in the BREQ-3 (e.g., external regulation, introjected regulation). Moreover, prior research has linked intuitive eating to the intrinsic exercise motivation scale specifically,¹⁹ suggesting that these two measures most closely mirror each other. Higher scores represented greater intrinsic exercise motivation and possible scores ranged from 0 – 4. Cronbach α for the intrinsic regulation subscale was .95.

Objective physical activity.

Participants wore an accelerometer (Actical, Phillips Respironics, Murrysville, PA, 2003) on their waist during all waking hours (except when bathing or swimming) for 7 consecutive days after their visit to the lab. Participants kept a log to indicate the time when they put on and removed the device. We categorized accelerometer data by activity counts/minute (sedentary < 100, 100 < light < 431, moderate > 431).⁵² We used this data to calculate the ratio of sedentary to moderate minutes of physical activity per day, where lower values indicated more optimal levels of physical activity. Specifically, the number of sedentary minutes per day was calculated as the total sedentary minutes divided by the number of days the accelerometer was worn (i.e., 7); the number of moderate minutes per day was calculated in the same way. The ratio of sedentary to moderate minutes of physical activity per day thus represents the average daily sedentary minutes divided by the average daily moderate minutes of activity, where a lower ratio indicates greater moderate physical activity and less sedentary activity.

Inflammatory proteins and blood lipids.

After a 20-minute period of seated rest, participants had their blood drawn from an antecubital vein by venipuncture. The research assistant drew approximately 20 milliliters of blood into evacuated tubes using sterile procedures. We processed and stored the serum at −80°C until analyzed for inflammatory proteins. According to manufacturer’s instructions, we diluted the serum samples 1:100 in assay diluent before performing the assay. We quantified CRP using a commercially available, high-sensitivity, solid-phase sandwich Enzyme Linked Immunosorbant Assay (ELISA; R&D Systems, Minneapolis, MN) that is substantially sensitive to detect and differentiate between subclinical concentrations of CRP. We multiplied the concentration read from the standard curve by the dilution factor. Undiluted serum was used to determine the concentrations of IL-6 using an ELISA (R&D Systems, Minneapolis, MN). All plates were read using a microplate reader (BioTek). We quantified LDL, HDL, and triglycerides from blood samples using an Alere Cholestech LDX Analyzer. We estimated values that fell outside of the detectable range of the Cholestech analyzer using the Friedewald formula.⁵³

Body composition.

Undergraduate and graduate student researchers, who were trained in performing all assessments, measured participants’ weight and body composition using the InBody body composition analyzer (InBody 770, InBody, Cerritos, CA, 2014). The InBody test has demonstrated criterion validity and internal reliability in a sample of healthy adults.⁵⁴ The instrument provides quantitative measures of the following: intracellular, extracellular, and total body water; dry lean mass; segmental and total lean mass; skeletal muscle mass; total and segmental body fat mass; and visceral fat area. Height was measured by the same trained undergraduate and graduate researchers using a wall-mounted stadiometer,⁵⁵ and BMI was calculated by dividing weight (kilograms) by height squared (m²) to include as a covariate.

Analytic Strategy

We initially examined Pearson’s bivariate correlations among all key study variables. Our assessments confirmed linearity between predictors and outcomes and acceptable levels of multicollinearity as the variance inflation factors were all less than 2. Based on a criterion of 3 standard deviations from the mean, we identified 2 outliers for CRP, 1 outlier for IL-6, and 1 outlier for LDL/HDL. All outliers were winsorized to 3 standard deviations from the mean. We assessed normality using histograms and the Shapiro-Wilk test. Measures of CRP, IL-6, and triglycerides were highly positively skewed, so these three outcomes were log-transformed.

One participant had accelerometer data that did not match their activity log. Once this was discovered, the participant agreed to wear another accelerometer for an additional 7-day period. This accelerometer data also did not match the participant’s log, so we chose to code this participant’s accelerometer data as missing and use listwise deletion for the regression models. There were no other missing data.

Predictor variables included intrinsic exercise motivation and intuitive eating. Covariates comprised age, sex, BMI, and objective physical activity (counts/day). Outcome variables were LDL/HDL, triglycerides, CRP, and IL-6. We entered all predictors (intuitive eating and intrinsic exercise motivation) and covariates in one step into a multiple linear regression for each outcome, resulting in four separate regression models. P-values <.05 were considered statistically significant. Power analyses using G*Power 3.1.9.4 (2019; Franz Faul, Universität Kiel, Germany) indicated our sample size was sufficient to detect a medium effect (i.e., .15) for each outcome (assuming power = .80, alpha = .05, and 6 predictors including covariates).

We used statistical tests (i.e., the Breusch-Pagan test) and diagnostic plots of residuals to assess homogeneity of variance. The non-significant statistical tests and plots for LDL/HDL, triglycerides, and CRP provided evidence that the residual variance was homogeneous for each variable. The examination of IL-6 led to conflicting results. The Breusch-Pagan test was statistically significant (p = .04), suggesting the variance was not homogenous; however, the plot showed that the residuals were equally and normally spaced, with no identifiable patterns. Based on the residual plot, we determined the severity of heteroscedasticity for IL-6 was not sufficient to warrant a correction. We performed all analyses in Stata 16 (released 2019; StataCorp, College Station, TX).

Results

Participant characteristics are shown in Table 1. Participants primarily identified as female (72%). The average age of participants (n = 79) was 68.8 years (SD = 6.3, range 58 – 83). Based on BMI estimates, 1.3% of participants were classified as underweight, 54.4% as normal weight, 16.5% as overweight, and 27.8% as obese. The intrinsic exercise motivation scores were negatively skewed, suggesting this sample was highly intrinsically motivated for exercise. Participants in this sample were highly active, indicated by low sedentary/moderate activity ratio scores; the sedentary/moderate ratio was also positively skewed.

Table 1.

Descriptive Statistics of Older Adults ≥ 58 Years in a Midwestern Community (n = 79)

	Mean (SD)	Range
Age (years)	68.76 (6.26)	58 – 83
BMI (kg/m2)	26.21 (6.29)	16.5 – 45.9
Intuitive eatinga	3.47 (.59)	2.09 – 4.74
Intrinsic exercise motivationb	2.85 (1.13)	0 – 4
Sed/mod activity (Min/Day)	15.76 (11.46)	1.72 – 57.37
TRG (mg/dL)	88.73 (41.21)	45 – 204
LDL/HDL (mg/dL)	1.96 (.64)	.79 – 3.94
CRP (mg/L)	1.93 (2.25)	.10 – 8.93
IL-6 (ng/mL)	4.34 (1.31)	2.67 – 10.78

BMI = body mass index; Sed/mod activity = sedentary/moderate activity ratio; TRG = triglycerides; CRP = C-reactive protein; IL-6 = Interleukin-6

Winsorized values are presented for Sed/mod activity, LDL/HDL, CRP, and IL-6.

Higher scores for intuitive eating and intrinsic exercise motivation indicate higher levels of intuition and intrinsic motivation, respectively. Higher levels of TRG, LDL/HDL, CRP, and IL-6 represent greater risk of cardiovascular disease. A lower sed/mod activity ratio indicates higher levels of physical activity.

^aPossible score range 1 – 5

^bPossible score range 0 – 4.

Table 2 shows Pearson’s pairwise correlations for all variables. Age correlated negatively with BMI, r = −.48, p < .001, and positively associated with intuitive eating, r = .32, p = .004 and exercise motivation, r = .36, p = .001, suggesting that adults who were older reported more intrinsically motivated behaviors and lower BMIs. Additionally, BMI correlated negatively with both intuitive eating and intrinsic exercise motivation, r = −.52, r = −.53, respectively (both p < .001). Intuitive eating correlated positively with intrinsic exercise motivation, r = .44, p < .001.

Table 2.

Bivariate Pairwise Pearson’s Correlations Among Demographic, Behavioral Motivation, and Biomarker Variables (n = 79)

	Age (years)	Female	BMI (kg/m2)	Intuitive eating	Intrinsic exercise motivation	Sed/mod activity (Min/Day)	TRG (mg/dL)	LDL/HDL (mg/dL)	CRP (mg/L)	IL-6 (ng/mL)
Female	.02
BMI (kg/m2)	−.48***	.02
Intuitive eating	.32**	−.23*	−.52***
Intrinsic exercise motivation	.36**	−.01	−.53***	.44***
Sed/mod activity (Min/Day)	.13	.09	.06	−.01	−.11
TRG (mg/dL)	−.22	.09	.40***	−.51***	−.37***	−.01
LDL/HDL (mg/dL)	−.28*	−.12	.38***	−.47***	−.23*	−.04	.56***
CRP (mg/L)	−.06	.13	.51***	−.39***	−.17	.03	.30**	.33**
IL-6 (ng/mL)	.31**	−.04	.09	−.06	.04	.22*	.07	.10	.43***

BMI = body mass index; Sed/mod activity = sedentary/moderate activity ratio; TRG = triglycerides; CRP = C-reactive protein; IL-6 = Interleukin-6

^***p<.001

^**p< .01

^*p<.05. Significant p-values indicate that the two variables are significantly correlated using pairwise Pearson’s correlations.

Note: All correlations are presented for raw variables (i.e., not transformed or winsorized).

Higher scores for intuitive eating and intrinsic exercise motivation indicate higher levels of intuition and intrinsic motivation, respectively. Higher levels of TRG, LDL/HDL, CRP, and IL-6 represent greater risk of cardiovascular disease. A lower sed/mod activity ratio indicates higher levels of physical activity.

Standardized and unstandardized coefficients from regression analyses are in Table 3. Due to 1 participant with missing accelerometer data, regression analyses comprised 78 participants. Greater intuitive eating was associated with a lower LDL/HDL ratio (β = −.45, p = .001) and lower triglycerides (β = −.37, p = .003). However, intuitive eating was not related to CRP or IL-6. Moreover, intrinsic exercise motivation was not significantly associated with any biomarker. Overall, statistically significant effect sizes for intuitive eating were small-to-moderate (β = −.37, −.45), and the statistically non-significant effect sizes for intuitive eating (β = −.12, −.20) and intrinsic exercise motivation (β = −.15, .09, .10, .16) were relatively small. The covariate sex was significantly associated with LDL/HDL where women, on average, had an LDL/HDL ratio that was .31 lower than men’s (where lower ratios are considered more optimal). Age and BMI, also covariates, were significantly positively related to CRP (age: β = .24, p = .04; BMI: β = .61, p < .001). Age was also associated with IL-6 (β = .42, p = .001). The primary independent variables and covariates together explained the greatest amount of variance in CRP (30%) and the smallest amount of variance in IL-6 (13%).

Table 3.

Multiple Linear Regression Models for Variables Predicting Cardiometabolic Biomarkers in a Sample of Older Adults (n= 78)

	LDL/HDL (mg/dL)				TRG (mg/dL)				CRP (mg/L)				IL-6 (ng/mL)
	B (SE)	CI	β	P-value	B (SE)	CI	β	P-value	B (SE)	CI	β	P-value	B (SE)	CI	β	P-value
Behavioral motivation
Intuitive eating	−.48 (.13)	[−.75, −.22]	−.45	.001	−.27 (.09)	[−.44, −.09]	−.37	.003	−.38 (.22)	[−.83, .07]	−.20	.09	−.05 (.06)	[−.16, .06]	−.12	.38
Intrinsic exercise motivation	.05 (.07)	[−.09, .19]	.09	.49	−.06 (.05)	[−.15, .03]	−.15	.21	.15 (.12)	[−.08, .39]	.16	.19	.02 (.03)	[−.04, .08]	.10	.45
Covariates
Age (years)	−.01 (.01)	[−.09, .19]	−.06	.59	.00 (.01)	[−.02, .02]	.00	.99	.04 (.02)	[.00, .08]	.24	.04	.02 (.01)	[.01, .03]	.42	.001
Female	−.31 (.15)	[−.60, −.01]	−.22	.04	.04 (.10)	[−.15, .23]	.04	.68	.18 (.24)	[−.30, .67]	.07	.46	−.05 (.06)	[−.17, .07]	−.09	.41
BMI (kg/m2)	.02 (.01)	[−.01, .05]	.17	.22	.01 (.01)	[−.01, .03]	.14	.31	.11 (.02)	[.06, .15]	.61	<.001	.01 (.01)	[−.00, .02]	.28	.06
Sed/mod activity (Min/Day)	−.01 (.07)	[−.15, .12]	−.02	.82	−.01 (.04)	[−.10, .07]	−.03	.73	−.04 (.11)	[−.26, .18]	−.04	.71	.04 (.03)	[−.01, .10]	.17	.12
Adj. R 2	.24				.26				.30				.13

Unstandardized coefficients (B) are presented with standard errors (SE) in parentheses. Confidence intervals (CIs) correspond to the unstandardized coefficients.

BMI = body mass index; Sed/mod activity = sedentary/moderate activity ratio; TRG = triglycerides; CRP = C-reactive protein; IL-6 = Interleukin-6

^***p<.001

^**p< .01

^*p<.05. Significant p-values indicate that the independent variable is significantly associated with the outcome in the multiple linear regression. Note: All independent variables were entered in one step for each model/outcome. LDL/HDL was winsorized; CRP and IL-6 were both winsorized and log transformed; Sed/mod activity was centered and standardized.

Higher scores for intuitive eating and intrinsic exercise motivation indicate higher levels of intuition and intrinsic motivation, respectively. Higher levels of TRG, LDL/HDL, CRP, and IL-6 represent greater risk of cardiovascular disease. A lower sed/mod activity ratio indicates higher levels of physical activity.

Discussion

In the present study, we sought to examine the associations among intrinsic exercise motivation, intuitive eating, and biomarkers known to be important for CVD. Research on behavioral motivation and physical health markers, particularly among older adults, is lacking. Given the growing older adult population, a group for which CVD is extremely prevalent, understanding what influences CVD risk factors – and ultimately, how to prevent CVD – is a public health priority. We focused on four biomarkers known to be associated with CVD.

Participants’ average intrinsic exercise motivation scores were comparable with other studies.² Intuitive eating scores for the current sample were relatively similar to, although slightly lower than, scores from younger adults²¹ as well as older women.⁴⁷ Interestingly, participants in the current sample reported intuitive eating scores (mean = 3.47) closer to two samples of college students²¹ (means = 3.53; 3.50) than to a sample of older women⁴⁷ (mean = 3.66). This remained true when comparing intuitive eating scores for only females from the current sample (mean = 3.39) to the sample of older women.

Based on SDT^11,14 and prior research on intuitive eating,^23,27 we anticipated associations between behavioral motivation and each of the biomarkers. We found that intuitive eating was significantly associated with LDL/HDL and triglycerides in the expected direction but not with either of the inflammatory proteins. This is consistent with some research linking intuitive eating to cholesterol markers²⁴ but conflicts with other findings of significant associations between intuitive eating and CRP.²⁶ Importantly, these studies were conducted among college students and a clinical sample, respectively. In older women, one study found that higher intuitive eating was associated with fewer depressive symptoms but did not include biomarkers.⁴⁷

Neither of the behavioral motivation variables were significantly associated with the inflammatory biomarkers. The associations between health behaviors and IL-6 and CRP in particular may not be straightforward, since both IL-6 and CRP play important regulatory roles in the presence and absence of significant, chronic low-grade inflammation in ways that are designed to preserve health (through immune function and tissue repair). Overall, the general lack of association here may reflect more nuanced processes that differ among individuals and are masked by the analyses.

Intrinsic Exercise Motivation and Physical Activity

Although intrinsic exercise motivation was significantly associated with triglycerides and LDL/HDL using the Pearson’s correlation analysis, it was not associated with any biomarkers in the regression models. The correlation analysis showed that the correlation coefficient between intrinsic exercise motivation and intuitive eating was .44, indicating that the two variables shared about 19% of their variance (unadjusted). Due to the shared variance between intrinsic exercise motivation and intuitive eating, the correlations between intrinsic exercise motivation and the lipid biomarkers may be mitigated with the inclusion of intuitive eating in the regression models. However, the variance inflation factor values (all < 2) suggested that multicollinearity was not an issue for any of the models.

One possible explanation for the null findings relating to intrinsic exercise motivation could be due to the lack of variability in exercise motivation in this sample. Most of the participants reported high levels of intrinsic exercise motivation (median = 3.25, range = 0 – 4), whereas intuitive eating was normally distributed. Intrinsic exercise motivation was also not significantly correlated with objective physical activity. There are many possible reasons for this. Intrinsic exercise motivation was negatively skewed, and objective physical activity was positively skewed. Statistically, skew can truncate the range thereby limiting the variability in the sample. Conceptually, people may participate in physical activity for reasons other than intrinsic exercise motivation. For example, dual process models of exercise behavior posit that behavioral decisions are determined by both conscious and nonconscious processes.⁵⁶ Thus, our sample may consist of people who are regular exercisers that are driven by nonconscious processes not measured in this study. Objective physical activity was also not associated with any biomarkers. We would expect physical activity to be related to biomarkers, but because the sample was highly active, ceiling effects may have prevented the detection of an association.

Several limitations warrant caution in interpreting these findings. This study was cross-sectional so causation and directionality cannot be inferred. There was also an inherent selection bias in the sampling methods utilized due to convenience and snowball sampling. The participants volunteered for the study and were more likely to be actively involved in the community compared to those who were not reached by the recruitment methods utilized. The resulting sample was also relatively small and homogenous. The small sample size and relative homogeneity may have thwarted our ability to detect more associations between the independent variables and outcomes, particularly if our sample was at low risk of CVD. We also did not collect data on race/ethnicity, which limits the generalizability of the results. One additional limitation is that some of the effect sizes were smaller than anticipated leading some of the analyses to be underpowered. Moreover, the IES-2 has been validated in samples that include middle-aged and older adults^46,47 but has not been validated exclusively for older adults. Similarly, although examining objective sedentary and moderate activity as a ratio provides more fidelity than looking at either separately,^57,58 normative ranges have not yet been developed for the ratio measure. Finally, objective diet information could not be collected due to time and personnel constraints, but we did collect objective physical activity data using accelerometers.

Despite these limitations, one of the unique aspects of the present study was the method used for measuring inflammation. The measurement of gene expression has been a common approach to determining inflammation levels,⁵⁹ while fewer studies have directly measured inflammatory proteins.⁶⁰ Measuring inflammatory genes only shows the genetic potential for production of inflammatory proteins. Rather than measure inflammation indirectly by analyzing proinflammatory genetic expression as many of the previous studies have, a strength of this study is that inflammatory protein levels were directly measured.

Genetics studies have identified several genes (e.g., MC4R) that play a role in appetite, hunger cues, and satiety.⁶¹ It is plausible that these genes would also affect how people respond to intuitive eating questions, due to associations among appetite, BMI, and inflammation. Such genetic inquiries go beyond the scope of this study but could benefit from designs such as the co-twin control method, which help disentangle environmental and genetic effects.⁶²

Implications for Research and Practice

This study explored the association between behavioral motivation and CVD-related biomarkers among older adults and directly responded to calls from nutrition researchers to examine the association between intuitive eating and physical health during aging.⁴⁷ The fact that associations between intuitive eating and lipid biomarkers were detectable in a small, active, homogenous sample suggests that the associations are worthy of further exploration. Most of the prior research on intuitive eating has been conducted among college students^27,63 or specific clinical populations,^26,30 whereas these results extend findings to older adults.

Several future directions exist. Collecting objective diet information (e.g., 24-hour dietary recall) would permit the relationship between diet and intuitive eating to be tested. Additionally, one possible explanation for the lack of significant findings between intuitive eating and inflammatory markers could be due to older adults having undiagnosed food sensitivities,⁶⁴ as food allergies are often masked by symptoms associated with a general age-related decline of physiological and biological systems.⁶⁵ As a result, people may be unknowingly consuming foods that ultimately drive increases in inflammation. With objective diet information, future work may be able to explore the connection between diet, food sensitivities and inflammatory responses, and intuitive eating. Moreover, inclusion of additional inflammatory biomarkers (e.g., Interleukin 1 beta) could further our understanding of inflammatory processes in future work.

Overall, the results suggest a possible link between intuitive eating and cholesterol-based biomarkers. Additional longitudinal inquiries are needed to clearly discern patterns of causal influence between behavioral motivation and biomarkers related to CVD. These data also support future work testing these associations among a larger, more diverse sample to increase the generalizability of findings. From a practical perspective, if these findings are replicated and confirmed in longitudinal work, they would support the use of interventions targeting intuitive eating in older adults. Successful intervention work can ultimately help practitioners more effectively aid older adults in becoming more intuitive eaters.