Effect of Daily Avocado Intake on Cardiovascular Health Assessed by Life's Essential 8: An Ancillary Study of HAT, a Randomized Controlled Trial

Abstract

Background

The effect of single food‐based dietary interventions on the American Heart Association's cardiovascular health score, Life's Essential 8 (LE8), is unclear. The aim of this study was to examine the effect of daily avocado intake for 26 weeks on LE8 in adults with abdominal obesity.

Methods and Results

An ancillary analysis including participants (n=969; aged 51±14 years) from the HAT (Habitual Diet and Avocado Trial), a randomized controlled trial, was conducted. The Avocado‐Supplemented Diet Group was provided with 1 avocado per day, and the Habitual Diet Group was instructed to maintain their usual diet. LE8 component scores (diet, physical activity, nicotine exposure, sleep health, body mass index, blood lipids, blood glucose, and blood pressure) were calculated using a modified American Heart Association algorithm. The LE8 score was calculated as the unweighted average of each component (range, 0–100 points). Between‐group differences in the 26‐week change in LE8 were assessed using general linear models. No significant between‐group difference in the 26‐week change in the LE8 score from baseline was observed (0.79 points [95% CI, −0.41 to 2.00]). However, avocado intake increased the LE8 component scores for diet (3.53 points [95% CI, 1.38–5.68]), sleep health (3.20 points [95% CI, 0.38–6.02]), and blood lipids (3.46 points [95% CI, 1.03–5.90]) compared with the Habitual Diet Group.

Conclusions

Intake of 1 avocado per day for 26 weeks did not significantly affect the total cardiovascular health score in US adults with abdominal obesity. However, diet quality, sleep health, and blood lipids improved with daily avocado intake.

Registration

URL: https://www.clinicaltrials.gov; Unique identifier: NCT03528031.

Nonstandard Abbreviations and Acronyms

HAT

Habitual Diet and Avocado Trial

LE8

Life's Essential 8

VAT

visceral adipose tissue

Clinical Perspective.

What Is New?

• This randomized controlled trial is the first to assess the impact of a single food‐based dietary intervention on the change in the estimated cardiovascular health score assessed using a modified American Heart Association's Life's Essential 8 (LE8) algorithm.

• The addition of 1 avocado per day to a habitual diet for 26 weeks did not improve the LE8 score in US adults with abdominal obesity compared with a habitual intake low in avocados. Individual LE8 component scores for diet quality, sleep health, and blood lipids were significantly improved with daily avocado intake.

What Are the Clinical Implications?

• Dietary intervention represents an important strategy to prevent, manage, and treat cardiovascular disease risk; however, there is a paucity of research evaluating how single‐food dietary interventions impact the American Heart Association's LE8 components that contribute to cardiovascular health in adults at elevated risk for cardiovascular disease.

• These findings suggest that relatively large improvements across most, or all, LE8 components are needed for detectable improvements in the LE8 score, which is likely not achievable with incremental small behavior changes such as incorporation of a single food, and further research is needed to examine the dose and intensity of lifestyle interventions to improve the LE8 score and the usefulness of this metric for patients in clinical practice.

Cardiovascular disease (CVD) remains the leading cause of mortality worldwide, accounting for approximately one‐third of all deaths in 2021. ¹ In 2010, the American Heart Association (AHA) introduced the concept of cardiovascular health to emphasize health factors and behaviors that promote cardiovascular health and reduce the risk of CVD. ² Currently, ideal cardiovascular health is defined as the simultaneous presence of 8 favorable health factors (healthy body weight, blood pressure, blood lipids, and glucose) and behaviors (adequate physical activity, sleep, healthy diet, and avoidance of nicotine). ³ These health factors and behaviors are termed Life's Essential 8 (LE8). ³ The LE8 score, a composite cardiovascular health score, is scaled from 0 to 100 points and calculated as the unweighted average of all 8 component scores. A greater number of ideal cardiovascular health metrics is associated with lower risk of CVD, stroke, and cardiovascular mortality. ⁴ According to recent national statistics, the average LE8 score in the US adult population is 64.7 points (95% CI, 63.9–65.6), which is below the optimal level (≥80 points). Regardless of age and sex, the US population has a low prevalence (<1%) of ideal cardiovascular health. ⁵ By 2050, it is expected that the prevalence of CVD risk factors such as hypertension, diabetes, and obesity will increase by 10% to 20% over current levels. ⁶ Thus, given the presence of suboptimal cardiovascular health and expected increases in CVD risk factors in US adults, effective diet and lifestyle interventions are needed.

Diet quality is a behavioral health component of LE8 that directly affects other LE8 components such as body weight, blood pressure, and blood lipid and glucose concentrations. Few randomized controlled trials (RCTs) have evaluated the impact of lifestyle interventions on changes in the cardiovascular health score assessed by LE8 or the prior metric, Life's Simple 7. ⁷ , ⁸ , ⁹ Among the 3 studies identified, only 1 showed an improvement in the Life's Simple 7 metric after 6 months of the lifestyle intervention. ⁹ Given that lifestyle behaviors are interrelated and interdependent, effective dietary interventions may result in concomitant changes in other lifestyle behaviors. ¹⁰ In addition to modulating cardiometabolic health markers such as blood pressure, blood lipids, and glucose, it is possible that interventions improving dietary behaviors may have a positive influence on other health behaviors such as physical activity, sleep, and nicotine exposure. This has been recently demonstrated in 2 RCTs that showed that dietary supplementation of avocados improved diet quality, assessed using the Healthy Eating Index 2015 score, and physical activity, assessed via metabolic equivalent task minutes per week, after 6 months compared with low habitual intake of avocados. ¹¹ , ¹²

Currently, there is a paucity of RCTs examining whether single food‐based dietary interventions improve cardiovascular health trajectories evaluated via changes in the LE8 metric over time. ¹³ A greater understanding of the influence of single food‐based dietary interventions on cardiovascular health will inform nutritional and public health initiatives, such as Food is Medicine programs that implement food‐based interventions to prevent, manage, and treat diet‐related chronic disease. The aims of this ancillary study were to (1) examine the effect of daily avocado intake for 26 weeks on the total cardiovascular health score assessed using AHA's LE8 ³ metric in US adults with abdominal obesity and (2) evaluate whether changes in the total cardiovascular health score are associated with changes in cardiometabolic risk factors. Our primary hypothesis was that at the end of the intervention period, the addition of 1 avocado per day to a habitual diet would increase the LE8 score compared with a habitual diet without supplemental avocados. Furthermore, we hypothesized that an increase in the LE8 score in response to daily avocado intake would be associated with favorable changes in visceral adipose tissue (VAT) volume, hepatic fat fraction, CRP (C‐reactive protein), and waist circumference.

METHODS

Trial Design

HAT (Habitual Diet and Avocado Trial) was a multicenter, parallel‐arm, investigator‐blinded, 26‐week, rolling recruitment RCT conducted between June 2018 and October 2020 across 4 clinical sites in the United States. The COVID‐19 pandemic resulted in a shutdown of research activities and affected in‐person visits in mid‐March 2020, and thus, for those participants who had not completed the 26‐week prespecified study period, the study duration was extended to allow for final data collection. A detailed description of the trial protocol ¹⁴ and the results of the primary and secondary outcomes have been reported. ¹⁵ Briefly, participants were recruited using advertisements/flyers and were scheduled for an in‐person screening visit to assess eligibility. Eligible participants were randomly allocated in a 1:1 ratio to either the intervention or comparator group with permuted block randomization (varying block sizes of 4 and 8) and stratification by participant sex and clinical site using a web‐based system. Clinic staff and participants were not blinded to the allocation of randomization groups, but all HAT investigators, laboratory staff, and outcome assessors were blinded. HAT outcomes included in the current ancillary analysis were measured at baseline and 26 weeks, or in some cases as necessary thereafter. The coordinating center (Wake Forest University Health Sciences, Winston‐Salem, NC) served as the central institutional review board and administered the study protocol. At all 4 clinical sites, the study was reviewed by local institutional review boards. All participants provided written informed consent. The data that support the findings of this study are available from the Avocado Nutrition Center upon reasonable request. This trial is registered at clinicaltrials.gov (identifier: NCT03528031).

Participants

Participants were recruited from 4 clinical sites in the United States: Tufts University (Jean Mayer US Department of Agriculture Human Nutrition Research Center on Aging), Boston, Massachusetts; University of California at Los Angeles, Los Angeles, California; Loma Linda University, Loma Linda, California; and the Pennsylvania State University, University Park and Hershey, Pennsylvania. Eligible participants were aged ≥25 years, had elevated waist circumference (ie, ≥35 inches for women and ≥40 inches for men), and reported low usual avocado intake (≤2 avocados per month). Participants were excluded if they had a known avocado sensitivity or aversion to avocados; were pregnant, lactating, or had intention of pregnancy; unwillingness/inability to undergo magnetic resonance imaging scans for assessment of primary outcome (VAT volume); unstable medical conditions (eg, cardiac, gastrointestinal, or hepatic disease); weight loss or gain of ≥10 lb within the 12 months before enrollment; following a restricted or weight loss dietary pattern; and consumed >7 drinks per week for women and >14 drinks per week for men. ¹⁴

Intervention and Comparator Groups

Participants randomized to the Avocado‐Supplemented Diet Group (intervention) were instructed to follow their usual diet and lifestyle and were regularly provided with and asked to consume 1 large Hass avocado daily (≈168 g/d) during the 26‐week intervention period. The Avocado‐Supplemented Diet Group was instructed to pick up fresh Hass avocados from their study site every 2 weeks during the trial. To maximize adherence, participants were provided with written instructions on how to store, ripen, cut, peel, and pit avocados, and with avocado‐focused recipes along with examples of serving sizes. No additional guidance or dietary counseling was provided. During the COVID‐19 shutdown, for participants in the intervention group, avocados were delivered to their residences following safety protocols. Participants randomized to the Habitual Diet Group (comparator) were instructed to follow their usual diet and lifestyle and were asked to restrict avocado consumption to ≤2 avocados per month. No avocados or dietary counseling was provided to the Habitual Diet Group. Participants in the Avocado‐Supplemented Diet Group reported a high adherence throughout the 26‐week trial, with 88% to 95% of participants reporting intake of some avocado (<90% of study provided amount of 168 g/d) or 1 whole avocado (>95%) on the days 24‐hour recalls were conducted. In the Habitual Diet Group, 92% to 94% of participants maintained low/no avocado consumption. ¹⁵

Outcome Assessment

The HAT study was designed to achieve 85% power for detecting a clinically significant between‐group difference in the primary end point, 26‐week change in VAT volume from baseline, with a 2‐sided type 1 error rate of 5% and anticipated 10% attrition. ¹⁴ Prespecified secondary outcomes included hepatic fat fraction, serum CRP concentration, and components of the metabolic syndrome (fasting plasma concentrations of triglycerides, high‐density lipoprotein cholesterol, and glucose; systolic and diastolic blood pressure; and waist circumference). ¹⁴

For the present ancillary study, data collected at baseline and 26 weeks were used to calculate AHA's LE8 metric, ³ which includes the following 8 components: diet, physical activity, body mass index (BMI), blood lipids, blood glucose, blood pressure, nicotine exposure, and sleep health. Briefly, the participants’ medical history and information on physical activity and nicotine exposure were collected using in‐house health and demographic questionnaires. Sleep health was evaluated using the Pittsburgh Sleep Quality Index questionnaire. ¹⁶ BMI was calculated as body weight (in kilograms) divided by height (in meters squared). Body weight and height were measured using standardized protocols. Blood pressure was measured using automated devices with participants resting in the seated position. After a 5‐minute rest, brachial artery systolic and diastolic blood pressure were measured in triplicate (with 1‐minute intervals), and the final 2 measurements were averaged for analysis. Blood lipids and glucose were measured by the central laboratory, as previously described. ¹⁴ Non–high‐density lipoprotein cholesterol was calculated by subtracting high‐density lipoprotein cholesterol from total cholesterol concentrations. Dietary intake was assessed by interviewer‐administered 24‐hour dietary recalls using Nutrition Data System for Research versions 2017 and 2018. Nutrition Data System for Research data were used to calculate the Healthy Eating Index 2015 score, as previously described. ¹²

Each of the LE8 component scores were calculated by adapting the AHA's LE8 algorithm for adults ³ based on the data available (Table S1). The diet and BMI components were calculated according to the AHA's algorithm. ³ The physical activity component was calculated from self‐reported participation in regular physical activity (yes/no response options) and was scored as 100 points for “yes” or 0 points for “no,” because minutes of moderate or vigorous physical activity were not measured in HAT. The sleep health LE8 component was calculated based on self‐reported average hours of sleep per night and scored according to 4 levels (Table S1) instead of the 6 levels in the AHA's LE8 algorithm, and the score was not adjusted for sleep apnea because this was not collected in HAT. The LE8 component for nicotine exposure was defined as self‐reported use of cigarettes, excluding inhaled nicotine‐delivery systems, because these data were not collected in HAT, and was scored according to the AHA's algorithm except without adjustment for living with an active indoor smoker at home, because this was not available in the HAT data set. LE8 components for blood glucose, blood lipids, and blood pressure were measured and scored according to the AHA's algorithm, except without adjustment for medication use, because these data were not collected. For the LE8 blood glucose component, fasting plasma glucose concentrations >125 mg/dL were scored according to quintiles because hemoglobin A1c was not measured in the HAT study. All LE8 components were scored on a scale of 0 to 100 points, and the LE8 score was calculated as the unweighted average of the LE8 component scores. A higher score indicates a better total cardiovascular health score, following the AHA recommendations for high (80–100 points), moderate (50–79 points), and low (0–49 points) cardiovascular health. ³ All LE8 components were simultaneously measured at baseline and 26 weeks. Thus, the LE8 score was calculated only for these 2 time points.

Cardiometabolic risk factors including VAT volume and hepatic fat fraction, CRP, and waist circumference were measured at baseline and 26 weeks. Briefly, VAT volume and hepatic fat fraction were measured using magnetic resonance imaging scans, as outlined previously. ¹⁴ Serum CRP concentration was measured using a chemiluminescent immunoassay (IMMULITE 2000; Siemens Healthcare Diagnostics, Los Angeles, CA), as previously described. ¹⁴ Waist circumference was measured at baseline and 26 weeks using standardized protocols across all 4 clinical sites.

Statistical Analysis

In this HAT ancillary study, the total cardiovascular health score measured by AHA's LE8 metric was assessed as an exploratory outcome, and thus, no power calculation was performed. The analytical sample included participants with outcome data at both time points, baseline and 26 weeks. Data missingness was evaluated by testing for potential baseline differences between participants with and without missing data for study outcomes that were used to calculate the LE8 components. The Mann‐Whitney U test for independent sample means or χ² test for categorical variables was used to test for differences between missing and complete observations. The end point was defined as the 26‐week change in the LE8 score from baseline (calculated as the absolute difference between baseline and 26‐week LE8 scores). The primary analysis included participants who had data available for ≥1 LE8 component at baseline and 26 weeks. For this analysis, the LE8 score for each participant at each time point was calculated by taking an unweighted average of all available LE8 components. The secondary analysis was restricted to participants who had complete data for all 8 LE8 components at baseline and 26 weeks. For this analysis, the LE8 score for each participant at each time point was calculated by taking an unweighted average of all 8 LE8 components. A sensitivity analysis excluding physical activity from the LE8 score calculation was performed. This was done because physical activity was assessed using a single question, “Do you currently participate in regular physical activity?,” with a binary response option (ie, yes or no). Thus, data on self‐reported minutes of moderate/vigorous physical activity were not available to score this component according to the methods recommended by the AHA. ³ For this analysis, an unweighted average of all the LE8 components minus physical activity (ie, diet, BMI, blood lipids, blood glucose, blood pressure, nicotine exposure, and sleep health) was used. Furthermore, sensitivity analyses were conducted to determine whether the results of the primary analysis differed based on the participant inclusion cut point used for the minimum number of available LE8 components (ie, ≥2, ≥3, ≥4, ≥5, ≥6, or ≥7) at baseline and 26 weeks. For these sensitivity analyses, the LE8 score for each participant at each time point was calculated by taking an unweighted average of all available LE8 components. For all analyses, the between‐group difference in the change in the LE8 score and individual LE8 components from baseline was assessed using general linear models (SAS 9.4 PROC GLM). All models were adjusted for study site and baseline scores.

Between‐group differences in the 26‐week change in the LE8 score by subgroups were examined by including the subgroup of interest as a categorical covariate in the general linear models. The P values from the interaction term of subgroup and randomization group were used to test the significance of subgroup effects. Subgroups examined included: sex (men or women); self‐reported ethnicity (not Hispanic or Latino; Hispanic or Latino); self‐reported race (Black or Asian or White or Other); age (25–30 years, 31–59 years, ≥60 years); baseline BMI (<30 kg/m², 30–35 kg/m², >35 kg/m²), study site (Loma Linda University, University of California at Los Angeles, Pennsylvania State University‐University Park, Pennsylvania State University‐Hershey, Tufts University); and baseline cardiovascular health score category (low, 0–49; moderate, 50–79; high, 80–100).

The association between changes in the LE8 score and intervention effects on HAT primary and secondary end points (ie, 26‐week change in VAT volume), hepatic fat fraction, CRP, and waist circumference, were tested using general linear models (SAS 9.4 PROC GLM). First, a binary categorical covariate for the 26‐week change in the LE8 score from baseline was constructed based on whether there was a decrease/no change (≤0 points) or an increase (>0 points) in LE8. Between‐group differences in HAT primary and secondary end points were tested by including the levels of change in the LE8 score as a categorical covariate in the linear models. Models were adjusted for study site. The P values from the interaction term of the categorized change in the LE8 score and randomization group were used to test the significance of the effect of change in the LE8 score on primary and secondary HAT end points.

For all analyses, normality of the residuals produced from the linear regression model outputs was assessed using univariate analysis (SAS 9.4 PROC UNIVARIATE) and normal probability plots (Q–Q plots). Descriptive data (mean±SD) for the LE8 score and individual component scores were calculated by study time point and randomization group and by participant subgroups. Linear model‐based estimates for each group are presented as least squared mean±SEM. Estimated between‐group differences from the linear models are presented as least square means and 95% CI with nominal P values. P values were considered statistically significant at α<0.05. All analyses were conducted with SAS version 9.4 (SAS Institute, Cary, NC).

RESULTS

Baseline Characteristics of the Analytical Sample

A detailed description of participant selection and baseline characteristics of the entire study cohort (n=1008) has been previously summarized. ¹⁵ The analytical sample for this HAT ancillary study included a total of 969 participants (Avocado‐Supplemented Diet Group [n=484] and Habitual Diet Group [n=485]) who had data available for ≥1 LE8 component at baseline and 26 weeks (Figure S1). Of this primary analytical sample of 969 participants, 163 did not have complete data for all 8 LE8 components, and thus, the secondary analysis was restricted to 806 participants with complete data for all 8 LE8 components at baseline and 26 weeks (Avocado‐Supplemented Diet Group [n=400] and Habitual Diet Group [n=406]). For the primary analysis sample, approximately 4% of data were missing because of participant withdrawals. At baseline, no significant differences were found between participants with and without missing data for outcomes used to calculate the LE8 score (Table S2); however, participants who had missing data were somewhat younger. For the secondary analysis sample, approximately 20% of data were missing because of participant withdrawals and incomplete data, but there were no significant differences between participants with and without missing data (Table S2). All available data were included in this ancillary study without imputation for missing observations. Briefly, the primary analysis sample included 969 randomized participants (72.9% women) with a mean age of 51±14 years and BMI of 33.0±5.5 kg/m² at baseline (Table 1).

Table 1.

Baseline Characteristics of the Analytical Sample

Variables	Overall (n=969)	Study arms
		Avocado‐supplemented diet group (n=484)	Habitual diet group (n=485)
Age, y	51±14	50±14	51±14
Men, n (%)	263 (27.1)	137 (28.3)	126 (26.0)
Women, n (%)	706 (72.9)	347 (71.7)	359 (74.0)
Hispanic or Latino, n (%)	199 (20.6)	101 (20.9)	98 (20.2)
Race (self‐reported), n (%)
Black	157 (16.2)	71 (14.7)	86 (17.7)
Asian	60 (6.2)	27 (5.6)	33 (6.8)
White	673 (69.5)	351 (72.5)	322 (66.4)
Other*	76 (7.8)	34 (7.0)	42 (8.7)
Did not answer	3 (0.3)	1 (0.2)	2 (0.4)
Anthropometric measures
Weight, kg	93.0±19.1	92.9±19.1	93.2±19.1
Height, cm	167.5±9.5	167.8±9.4	167.3±9.5
Body mass index, kg/m2	33.0±5.5	32.8±5.4	33.2±5.6
Waist circumference, cm	109.4±13.0	109.3±12.9	109.5±13.1
Men	117.4±11.8	117.3±11.4	117.6±12.1
Women	106.3±12.1	106.0±12.0	106.6±12.2
Blood biochemical measures (fasting), mg/dL
Total cholesterol	188±39	186±40	190±39
LDL cholesterol	112±34	110±34	114±33
HDL cholesterol	52±13	52±13	53±13
Men	44±10	44±9	45±10
Women	55±13	55±13	56±13
Non‐HDL cholesterol	136±37	134±37	137±37
VLDL cholesterol	25±16	25±18	24±14
Triglycerides	124±82	126±90	122±72
Insulin, μIU/mL	17.6±17.7	17.6±18.7	17.6±16.6
Glucose	107±30	107±29	107±30
Blood pressure, mm Hg
Systolic	123±16	123±16	123±16
Diastolic	76±10	77±10	76±11
HEI 2015 score	53.7±14.7	53.4±15.0	54.0±14.5
Regular physical activity,† n (%)	650 (67.1)	338 (69.8)	312 (64.3)

Data are presented as mean±SD or number (percent). HDL indicates high‐density lipoprotein; HEI 2015, Healthy Eating Index 2015; LDL, low‐density lipoprotein; and VLDL, very low‐density lipoprotein.

^*Other includes American Indian, Alaskan Native, Native Hawaiian or other Pacific Islander, and multiracial.

^† Self‐reported participation in regular physical activity based on an affirmative response to the following question: Do you currently participate in regular physical activity?

Change in the LE8 Score and Individual LE8 Components

The LE8 score and individual component scores at baseline and 26 weeks are presented by randomization group for the primary analysis in Table 2 and secondary analysis in Table S3. The primary analysis in participants with ≥1 LE8 component at both study time points showed no significant between‐group difference (0.79 points [95% CI, −0.41 to 2.00]) in the 26‐week change in the LE8 score from baseline (Table 3). Restricting this analysis to participants with complete data for all LE8 components confirmed this result (Table S4; 0.80 points [95% CI, −0.42 to 2.02]). However, after removing physical activity from the LE8 metric calculation, the score increased (1.34 points [95% CI, 0.29–2.39]) from baseline to 26 weeks in the Avocado‐Supplemented Diet Group compared with the Habitual Diet Group (Table 3). Similar results were observed in participants who had complete data for all LE8 components (Table S4; 1.36 points [95% CI, 0.33–2.38]). Sensitivity analyses showed that the results of the primary analysis did not change based on inclusion of participants with ≥2, ≥3, ≥4, ≥5, ≥6, or ≥7 LE8 components at baseline and 26 weeks (Table S5).

Table 2.

LE8 Score and Individual Component Scores by Study Arm at Baseline and 26 Weeks

Variables	Study arms
	Avocado‐supplemented diet group			Habitual diet group
	n	Baseline	26 wks	n	Baseline	26 wks
LE8 score*	484	66.55±12.23	66.25±12.57	485	64.96±12.32	64.38±12.87
Diet	472	41.09±17.44	46.19±16.78	480	41.60±17.17	42.77±17.99
Physical activity†	438	70.55±45.63	62.79±48.39	438	64.38±47.94	62.79±48.39
Nicotine exposure	436	92.26±18.77	92.03±19.20	440	91.76±19.79	91.70±20.09
Sleep health	464	72.37±26.26	73.00±25.32	473	68.48±27.90	67.63±27.71
BMI	464	38.19±25.78	38.32±26.48	462	37.33±26.45	37.15±26.96
Blood glucose	469	75.50±27.21	73.50±27.19	475	74.51±26.96	73.09±27.47
Blood lipids	464	73.02±28.73	76.81±26.91	473	71.59±29.76	72.35±29.11
Blood pressure	463	70.52±27.38	72.35±27.12	463	70.79±27.45	72.19±27.77
LE8 score minus physical activity‡	484	66.01±11.01	66.84±11.43	485	64.98±11.21	64.72±12.23

Data are presented as mean±SD. Maximum score=100 points. The analytic sample includes participants who had data available for ≥1 LE8 component at baseline and 26 weeks. BMI indicates body mass index; and LE8, Life's Essential 8.

^*The LE8 score was calculated as the unweighted average of all available LE8 component scores at each time point.

^† Reported regular participation in physical activity, n (%): Avocado‐Supplemented Diet Group, baseline=309 (70.5), 26 week=275 (62.8); Habitual Diet Group, baseline=282 (64.4), 26 week=275 (62.8).

^‡ Unweighted average of all available observations for LE8 component scores, excluding physical activity, at baseline and 26 weeks.

Table 3.

Model‐Based Estimates of the 26‐Week Change in the LE8 Score and Individual Component Scores by Study Arm

Variables	Study arms				Estimated between‐group difference
	Avocado‐supplemented diet group		Habitual diet group		Mean (95% CI)	P value
	n	LSM±SEM	n	LSM±SEM
LE8 score*	484	−0.03±0.44	485	−0.82±0.44	0.79 (−0.41 to 2.00)	0.19
Diet	472	4.66±0.79	480	1.13±0.79	3.53 (1.38 to 5.68)	0.001†
Physical activity	438	−6.41±2.12	438	−3.58±2.13	−2.83 (−8.60 to 2.92)	0.33
Nicotine exposure	436	0.03±0.47	440	0.14±0.47	−0.11 (−1.40 to 1.18)	0.86
Sleep health	464	1.66±1.04	473	−1.54±1.04	3.20 (0.38 to 6.02)	0.02†
BMI	464	0.07±0.51	462	−0.29±0.51	0.36 (−1.02 to 1.75)	0.60
Blood glucose	469	−1.43±0.96	475	−1.16±0.96	−0.27 (−2.85 to 2.32)	0.84
Blood lipids	464	4.09±0.90	473	0.63±0.90	3.46 (1.03 to 5.90)	0.005†
Blood pressure	463	1.60±1.05	463	1.31±1.05	0.29 (−2.55 to 3.13)	0.83
LE8 score minus physical activity‡	484	1.01±0.38	485	−0.33±0.38	1.34 (0.29 to 2.39)	0.01†

Data are presented as least square means±SEM. The analytic sample includes participants who had data available for ≥1 LE8 component at baseline and 26 weeks. Linear models (SAS 9.4 PROC GLM) were used to estimate the between‐group difference in the 26‐wk change in the LE8 score and individual component scores from baseline with adjustment for study site and baseline scores. BMI indicates body mass index; LE8, Life's Essential 8; and LSM, least square means.

^*The LE8 score was calculated as the unweighted average of all available LE8 component scores at each time point.

^† Significant.

^‡ Unweighted average of all available observations for LE8 component scores, excluding physical activity, at baseline and 26 weeks.

Among the individual LE8 components, the primary analysis showed that the Avocado‐Supplemented Diet Group had a greater increase in scores for diet (3.53 points [95% CI, 1.38–5.68]), sleep health (3.20 points [95% CI, 0.38–6.02]), and blood lipids (3.46 points [95% CI, 1.03–5.90]) at 26 weeks compared with the Habitual Diet Group (Table 3). In participants with complete data for all LE8 components, the between‐group difference in the diet score remained (3.62 points [95% CI, 1.27–5.96]); however, for the sleep health (2.93 points [95% CI, −0.10 to 5.98]) and blood lipids (2.57 points [95% CI, −0.03 to 5.17]) component scores, the between‐group differences were attenuated to borderline nonsignificance (P=0.05 for both), likely because of the loss of statistical power with the reduced sample size (Table S4).

Change in the LE8 Score by Subgroups

The LE8 scores at baseline and 26 weeks are presented by randomization group across participant subgroups for the primary analysis in Table 4 and secondary analysis in Table S6. In the primary analysis, the 26‐week change in the LE8 score from baseline differed by self‐reported ethnicity between the randomization groups (Table 5). In participants who were not Hispanic or Latino, the 26‐week change in the LE8 score from baseline did not differ between groups after adjustment for study site and baseline scores, but among participants who were Hispanic or Latino, the LE8 score decreased (−2.97 points [95% CI, −5.84 to −0.10]) from baseline to 26 weeks in the Avocado‐Supplemented Diet Group compared with the Habitual Diet Group (Table 5). In a post hoc exploratory analysis conducted to determine which LE8 component(s) was driving this LE8 score reduction in Hispanic and Latino participants, there were no significant between‐group differences in the 26‐week change in individual LE8 components, but the magnitude of change was the greatest for the physical activity component, which was lower in the Avocado‐Supplemented Diet Group (−21.39 points [95% CI, −35.83 to −6.96]) compared with the Habitual Diet Group (Table S7). In the primary analysis, there was no between‐group difference in the 26‐week change in the LE8 score from baseline by sex, age group, baseline BMI, study site, and baseline cardiovascular health score category (Table 5). In the secondary analysis, there was no between‐group difference in the 26‐week change in the LE8 score from baseline by sex, self‐reported ethnicity and race, age group, baseline BMI, study site, and baseline cardiovascular health score category (Table S8).

Table 4.

LE8 Score by Participant Subgroups and Study Arm at Baseline and 26 Weeks

Subgroups	n	Study arms
		Avocado‐supplemented diet group		Habitual diet group
		Baseline	26 wks	Baseline	26 wks
		n=484	n=484	n=485	n=485
Sex
Women	706	67.79±12.36	67.43±12.73	66.40±12.09	65.61±12.88
		n=347	n=347	n=359	n=359
Men	263	63.41±11.36	63.29±11.68	60.85±12.09	60.87±12.24
		n=137	n=137	n=126	n=126
Self‐reported ethnicity
Not Hispanic or Latino	766	66.61±12.45	66.71±12.41	64.88±12.05	63.86±13.06
		n=382	n=382	n=384	n=384
Hispanic or Latino	199	66.33±11.50	64.52±13.13	65.74±13.01	66.94±11.50
		n=101	n=101	n=98	n=98
Self‐reported race
Black	157	63.02±14.42	63.26±13.70	63.38±13.45	62.07±14.33
		n=71	n=71	n=86	n=86
Asian	60	72.00±10.36	71.69±11.80	67.75±11.09	66.72±12.16
		n=27	n=27	n=33	n=33
White	673	66.96±11.52	66.50±12.32	65.13±12.04	64.96±12.50
		n=351	n=351	n=322	n=322
Other*	76	65.39±14.25	65.62±12.21	65.32±11.91	62.86±12.11
		n=34	n=34	n=42	n=42
Age group (y)
25–30	85	74.22±10.52	70.78±13.49	69.10±13.76	67.45±14.04
		n=38	n=38	n=47	n=47
31–59	620	65.95±11.80	65.68±12.44	64.18±12.36	64.27±12.64
		n=314	n=314	n=306	n=306
≥60	264	65.76±13.01	66.32±12.43	65.28±11.43	63.54±12.93
		n=132	n=132	n=132	n=132
Baseline BMI (kg/m2)
<30	313	73.83±10.02	73.08±11.03	72.51±10.33	71.30±12.33
		n=161	n=161	n=152	n=152
30–35	369	65.69±10.71	65.08±12.00	63.49±11.79	63.34±11.91
		n=185	n=185	n=184	n=184
>35	286	59.20±11.77	59.86±11.11	58.90±10.78	58.50±11.23
		n=138	n=138	n=148	n=148
Study site
Loma Linda University	245	64.39±11.33	64.75±12.53	62.77±12.74	64.41±12.45
		n=122	n=122	n=123	n=123
University of California at Los Angeles	236	68.93±12.28	67.64±13.30	69.13±11.78	66.95±12.78
		n=118	n=118	n=118	n=118
Pennsylvania State University‐University Park	131	69.25±11.81	68.89±11.70	65.19±11.65	66.32±13.41
		n=65	n=65	n=66	n=66
Pennsylvania State University‐Hershey	114	65.89±11.21	63.92±12.73	66.93±10.84	65.54±11.00
		n=57	n=57	n=57	n=57
Tufts University	243	65.27±13.24	66.10±12.02	62.05±12.30	60.25±13.08
		n=122	n=122	n=121	n=121
Baseline cardiovascular health score category
Low	110	42.52±6.19	48.82±10.48	43.73±5.10	52.16±10.77
		n=48	n=48	n=62	n=62
Moderate	733	66.46±7.72	66.12±10.55	65.59±7.80	64.19±11.47
		n=366	n=366	n=367	n=367
High	126	83.50±3.58	78.91±8.27	84.32±3.16	79.17±7.94
		n=70	n=70	n=56	n=56

Data are presented as mean±SD. Maximum score=100 points. Participants were included who had data available for ≥1 LE8 component at baseline and 26 weeks. The LE8 score was calculated as the unweighted average of all available LE8 component scores at each time point. BMI indicates body mass index; and LE8, Life's Essential 8.

^*Other includes American Indian, Alaskan Native, Native Hawaiian or other Pacific Islander, and multiracial.

Table 5.

Model‐Based Estimates of the 26‐Week Change in the LE8 Score From Baseline Across Participant Subgroups by Study Arm

Subgroups	n	Study arms		Estimated between‐group difference
		Avocado‐supplemented diet group (n=484)	Habitual diet group (n=485)	Mean (95% CI)	P values
					Interaction*	Subgroup†	Group‡
Sex					0.80	0.68	0.79
Women	706	−0.39±0.56	−0.77±0.56	0.38 (−1.15 to 1.91)
Men	263	−0.27±0.90	−0.28±0.94	0.01 (−2.50 to 2.52)
Self‐reported ethnicity					0.03§	0.92	0.43
Not Hispanic or Latino	766	0.05±0.54	−1.06±0.53	1.11 (−0.35 to 2.58)
Hispanic or Latino	199	−1.87±1.09	1.11±1.11	−2.97 (−5.84 to −0.10)
Self‐reported race					0.47	0.98	0.21
Black	157	0.55±1.27	−1.22±1.16	1.77 (−1.50 to 5.03)
Asian	60	−0.06±2.00	−0.77±1.81	0.71 (−4.56 to 5.98)
White	673	−0.66±0.56	−0.34±0.58	−0.32 (−1.89 to 1.24)
Other	76	0.48±1.80	−2.35±1.62	2.82 (−1.87 to 7.51)
Age group (y)					0.11	0.18	0.94
25–30	85	−3.28±1.68	−1.44±1.52	−1.84 (−6.27 to 2.59)
31–59	620	−0.38±0.59	0.00±0.60	−0.38 (−2.01 to 1.24)
≥60	264	0.53±0.91	−1.87±0.91	2.40 (−0.09 to 4.89)
Baseline BMI, kg/m2					0.64	0.49	0.58
<30	313	−0.80±0.83	−1.15±0.84	0.35 (−1.96 to 2.65)
30–35	369	−0.64±0.77	−0.25±0.78	−0.39 (−2.51 to 1.73)
>35	286	0.59±0.89	−0.54±0.86	1.13 (−1.28 to 3.53)
Study site					0.20	0.02‖	0.95
Loma Linda University	245	0.36±0.93	1.64±0.93	−1.28 (−3.86 to 1.31)
University of California at Los Angeles	236	−1.29±0.95	−2.18±0.95	0.89 (−1.75 to 3.53)
Pennsylvania State University‐University Park	131	−0.36±1.28	1.13±1.27	−1.49 (−5.03 to 2.05)
Pennsylvania State University‐Hershey	114	−1.96±1.37	−1.40±1.37	−0.57 (−4.36 to 3.22)
Tufts University	243	0.83±0.93	−1.81±0.94	2.64 (0.04 to 5.24)
Baseline cardiovascular health score category					0.36	<0.0001†	0.87
Low	110	6.36±1.43	8.16±1.27	−1.80 (−5.54 to 1.94)
Moderate	733	−0.37±0.53	−1.42±0.53	1.05 (−0.38 to 2.49)
High	126	−4.66±1.19	−4.98±1.34	0.32 (−3.18 to 3.81)

Data are presented as least square means±SEM. The analytic sample includes participants who had data available for ≥1 LE8 component at baseline and 26 weeks. The LE8 score was calculated as the unweighted average of all available LE8 component scores at each time point. Linear models (SAS 9.4 PROC GLM) were used to estimate the between‐group difference in the 26‐wk change in the LE8 score from baseline by subgroup with adjustment for study site. BMI indicates body mass index; and LE8, Life's Essential 8.

^* P value from the interaction term of subgroup and randomization group was used to test the significance of subgroup effects.

^† P value for main effect of the subgroup.

^‡ P value for main effect of randomization group.

^§ Remained significant after adjusting for baseline scores.

^‖ Tukey post‐hoc: Loma Linda University>University of California at Los Angeles; change scores between baseline cardiovascular health score categories were all different from each other.

Association Between Change in the LE8 Score and Intervention Effects on HAT Primary and Secondary End Points

No significant interaction effects of randomization by categorical change in the LE8 score were observed for the 26‐week change from baseline in VAT volume, hepatic fat fraction, and waist circumference for the primary analysis (Table 6) and secondary analysis (Table S9). Although there was an interaction between randomization group and change in the LE8 score for the 26‐week change in CRP (Table 6), this effect was no longer significant after adjusting for baseline CRP (P=0.09), suggesting that the effect may have been partly driven by differences in CRP concentrations at baseline.

Table 6.

Model‐Based Estimates of HAT Primary and Secondary End Points by Study Arm and by LE8 Score Change Category

Cardiometabolic risk outcomes	n	Study arms		Estimated between‐group difference
		Avocado‐supplemented diet group	Habitual diet group	Mean (95% CI)	P values
					Interaction*	Change category†	Group‡
Primary end point: Δ VAT volume, L
Δ LE8 score					0.77	0.01δ	0.43
Decrease ≤0	490	0.094±0.021	0.084±0.020	0.010 (−0.046 to 0.066)
		n=236	n=254
Increase >0	445	0.049±0.021	0.027±0.022	0.022 (−0.037 to 0.081)
		n=226	n=219
Secondary end points
Δ Hepatic fat fraction, %
Δ LE8 Score					0.60	<0.0001δ	0.24
Decrease ≤0	484	0.014±0.003	0.009±0.003	0.006 (−0.004 to 0.015)
		n=232	n=252
Increase >0	439	−0.003±0.004	−0.006±0.004	0.002 (−0.008 to 0.012)
		n=224	n=215
Δ hs‐CRP, mg/L
Δ LE8 score					0.03δ	0.72	0.91
Decrease ≤0	490	1.00±0.59	−0.18±0.58	1.18 (−0.42 to 2.78)
		n=237	n=253
Increase >0	443	−0.04±0.61	1.27±0.62	−1.30 (−2.98 to 0.38)
		n=228	n=215
Δ Waist circumference, cm
Δ LE8 score					0.57	0.20	0.25
Decrease ≤0	489	0.39±0.30	−0.12±0.30	0.51 (−0.30 to 1.33)
		n=238	n=251
Increase >0	437	−0.16±0.31	−0.34±0.32	0.17 (−0.69 to 1.04)
		n=226	n=211

Data are presented as least square means±SEM. The analytic sample includes participants who had data available for ≥1 LE8 component at baseline and 26 weeks. The LE8 score was calculated as the unweighted average of all available LE8 component scores at each time point. Linear models (SAS 9.4 PROC GLM) were used to estimate the between‐group difference in the 26‐wk change in HAT outcomes by LE8 score change category with adjustment for study site. HAT indicates Habitual Diet and Avocado Trial; hs‐CRP, high‐sensitivity C‐reactive protein; LE8, Life's Essential 8; and VAT, visceral adipose tissue.

^* P value from the interaction term of LE8 Score Change Category and randomization group.

^† P value for main effect of LE8 score change category.

^‡ P value for the main effect of randomization group.

^δ Significant.

DISCUSSION

In this HAT ancillary study, supplementing habitual diets with 1 avocado per day for 26 weeks had no significant impact on the LE8 score compared with a habitual diet without supplemental avocados. After the elimination of physical activity from the LE8 score calculation because of measurement limitations, a modest improvement in the LE8 score was observed in the Avocado‐Supplemented Diet Group compared with the Habitual Diet Group. Diet, sleep health, and blood lipid component scores increased with daily intake of 1 avocado for 26 weeks compared with habitual intake. The 26‐week change in the LE8 score in response to daily avocado consumption was consistent across sex, self‐reported race, age, baseline BMI, study site, and baseline cardiovascular health score. Changes in the LE8 score in response to daily avocado intake did not predict changes in cardiometabolic risk factors. The present study provides the first clinical assessment of the effect of dietary intervention on the LE8 metric. Findings from this work may inform future research evaluating the implementation of the LE8 score in clinical practice.

The hypothesis that daily avocado intake would improve the LE8 metric in US adults with abdominal obesity was primarily based on RCT evidence suggesting that avocado consumption improves CVD risk factors such as diet quality, total cholesterol, and low‐density lipoprotein cholesterol in adults, including those with overweight/obesity. ¹² , ¹⁷ However, these RCTs have also shown null effects of avocado consumption on other CVD risk factors, such as glycemic control, body weight and fat composition, blood pressure, and vascular health. ¹² , ¹⁷ , ¹⁸ Although the lack of a significant effect of avocado consumption on the LE8 score in the current study is consistent with prior HAT null findings for primary and most secondary end points, it is notable that the individual LE8 component scores for diet quality, blood lipids, and sleep health increased after 26 weeks of daily avocado consumption compared with habitual intake.

The average LE8 score of the HAT analytical sample at baseline was 65.7 points, indicating a moderate cardiovascular health score as defined by the AHA. This is close to US population estimates for adults (64.7 points) based on the National Health and Nutrition Examination Survey 2013 to 2018. ⁵ In the primary analysis, no between‐group difference in the LE8 score was observed. Notably, a nonsignificant trend toward blunting of the 26‐week decrease in the LE8 score was observed in the Avocado‐Supplemented Diet Group compared with the Habitual Diet Group. A sensitivity analysis was conducted, because physical activity was measured as either regular (score=100) or no participation (score=0) in physical activity, thus precluding scoring of this component based on self‐reported minutes of moderate/vigorous physical activity, as recommended by the AHA. ³ A small improvement in the LE8 score (1.35 points) was observed in the Avocado‐Supplemented Diet Group compared with the Habitual Diet Group when the physical activity score was excluded from the LE8 calculation. This inconsistency is likely driven by the baseline differences in the proportion of each group reporting regular physical activity, which resulted in larger score changes and greater variability compared with the other components because of the binary nature of the scoring system for the physical activity component. This likely attenuated the primary result toward the null. Overall, these findings and trends suggest potential for incorporating daily intake of avocados into the habitual diet to improve cardiovascular health.

Cardiovascular health trajectories over the 26‐week study were not statistically different across sex, race, age, study site, and baseline cardiovascular health score status subgroups. Interestingly, subgroup analyses showed that among Hispanic and Latino participants, the LE8 score decreased from baseline to 26 weeks in the Avocado‐Supplemented Diet Group compared with the Habitual Diet Group, but there were no between‐group differences in LE8 change scores in participants who were not Hispanic or Latino. Post hoc exploratory analyses suggest that the LE8 decline among Hispanic and Latino participants in the Avocado‐Supplemented Diet Group was driven by a decrease in the physical activity score, whereas the individual component scores for diet quality, nicotine exposure, blood glucose, and lipids tended to increase with avocado supplementation. This observation is in contrast to findings from a recent parallel‐arm RCT that showed that dietary supplementation of avocados (14 avocados per week) to Hispanic and Latino families for 6 months resulted in a significantly greater increase in physical activity (+1163 metabolic equivalent task minutes per week) compared with families assigned to the control arm (3 avocados per week). ¹¹ These discrepant findings are likely because of differences in the physical activity measurement methods between the studies. Hispanic and Latino populations have a high prevalence of physical inactivity relative to other non‐Hispanic ethnic groups. ¹¹ , ¹⁹ , ²⁰ Thus, future research is needed to investigate how the AHA's LE8 metric is impacted by tailoring evidence‐based healthy dietary patterns with affordable healthy foods that are both culturally relevant and accessible to diverse racial, ethnic, and socioeconomic populations in the United States.

Although no significant between‐group difference in the LE8 score was observed, improvements in the individual component scores for diet quality, sleep health, and blood lipids were observed in the Avocado‐Supplemented Diet Group compared with the Habitual Diet Group. As previously reported, in the HAT study, incorporating 1 avocado daily into the habitual diet for 26 weeks improved diet quality as assessed using the Healthy Eating Index 2015 score. Furthermore, the Avocado‐Supplemented Diet Group had modest reductions in total cholesterol and low‐density lipoprotein cholesterol concentrations after 26 weeks compared with the Habitual Diet Group. ¹⁵ These findings likely explain the observed improvements in the LE8 component scores for diet quality and blood lipids. We also observed a greater increase in the LE8 component score for sleep health after 26 weeks of daily avocado supplementation when compared with the Habitual Diet Group. Avocados are a nutritionally dense food rich in monounsaturated fatty acids, dietary fiber, potassium, folate, vitamin K, copper, and pantothenic acid. ²¹ , ²² It possible that the increase in the sleep health score observed in the Avocado‐Supplemented Diet Group is, in part, related to the improvement in diet quality reflected directly by avocado consumption or avocado‐mediated displacement of unhealthy food components; however, this is speculative and warrants further investigation.

Little heterogeneity in the effect of the intervention on VAT volume, hepatic fat fraction, CRP, and waist circumference was observed by the change in LE8 score (based on a decrease/no change [≤0 points] or an increase [>0 points]). It is likely that no effect modification was observed because of the relatively modest increase in the LE8 score in participants who had a 26‐week increase in the LE8 score (median change, 6.38 points). The magnitude of a clinically meaningful change in the LE8 metric remains unclear. A pooled analysis of 6 US‐based cohort studies showed that a 10‐point higher LE8 score is associated with a 22% to 40% lower risk of CVD among young, middle, and older age adults. ²³ However, a recent analysis using the UK Biobank data showed that a change in LE8 over a median of 10 years was associated with CVD risk, but an exposure–response relationship was not observed, suggesting a nonlinear relationship. ²⁴

As our findings demonstrate, small improvements in 3 of the 8 components did not result in an improvement in the LE8 score. Current guidance for lifestyle counseling in clinical practice is to prioritize small incremental changes to increase the likelihood of behavior change. ²⁵ These results suggest that the LE8 metric may be too insensitive to small changes in health and behavioral risk factors to be used as a tool for patient monitoring, tracking, and goal setting because of the potential for disillusionment when no change in the LE8 score is observed. Further research is needed to understand how the algorithm works in different contexts, including assessing changes at the individual level over time, and in different patient groups including those with and without obesity. Although our findings provide some insight, further research is warranted to examine the dose and intensity of intervention needed to improve LE8 at an individual level and the usefulness of this metric for patients in clinical practice.

A strength of this ancillary investigation is the inclusion of a large racially and ethnically diverse sample of adults recruited across 4 clinical sites in the United States. Furthermore, this study assessed the impact of a single food‐based dietary intervention on AHA's LE8 metric in adults with abdominal obesity, which affects almost 60% of US adults. ²⁶ A limitation of the investigation is that the AHA's definition and scoring approach ³ for physical activity could not be used, and the question we used to assess physical activity has not been validated. Furthermore, the AHA's definitions and scoring approach for LE8 components for blood lipids, blood glucose, blood pressure, nicotine exposure, and sleep health were, by necessity, modified according to the available HAT data set. Other limitations include the lack of data collected on social and psychological determinants of health. Future research should consider how socioeconomic constraints/barriers impact an individual's response to diet and lifestyle interventions that are aimed at improving cardiovascular health metrics.

CONCLUSIONS

In this ancillary study of HAT, intake of 1 avocado per day for 26 weeks did not significantly affect the total cardiovascular health score measured using AHA's LE8 metric compared with habitual intake in US adults with abdominal obesity. However, individual cardiovascular health component scores for diet quality, sleep health, and blood lipids were improved with supplemental avocado intake. Our findings suggest that single food‐based dietary interventions may not be sufficient to improve the LE8 score and that larger dietary pattern improvements, along with additional lifestyle modifications, may be needed to elicit detectable improvements in the LE8 metric.

Sources of Funding

This work was supported by the Avocado Nutrition Center. The funding agency had no role in data collection, analysis, or interpretation, and writing of the article.

Disclosures

None.

Supporting information

Tables S1–S9

Figure S1