Caloric restriction and time-restricted eating in older adults with overweight or obesity: The Health, Aging, and Later-Life Outcomes Pilot Study

Abstract

Background

In animal models, caloric restriction (CR) and time-restricted eating (TRE) extend lifespan and healthspan; however, the long-term benefits in humans are unknown. The goal of the Health, Aging and Later-Life Outcomes Pilot (HALLO-P) was to inform the design of a definitive trial to evaluate the long-term effects of CR and TRE in older adults with overweight or obesity.

Methods

HALLO-P randomized 90 older (≥60 years) adults with obesity or overweight to one of three 9-month interventions: (1) 20% CR delivered in-person; (2) 20% CR delivered remotely (RCR); and (3) 8-hour TRE with ad libitum caloric intake. The degree of sustained CR (by doubly labeled water), the sustainability of TRE, participant retention, and changes in body mass and composition, physical performance, and cardiometabolic risk factors were examined.

Results

Participants had a mean (SD) age of 67.2 (4.9) years and BMI of 31.7 (2.9) kg/m²; 62% were female and 83% White. Participant retention was 92%. The mean (SD) percent CR was 4.5% (11.0) in CR and 6.0% (10.3) in RCR. TRE participants reported eating within an 8.5-hour window a median of 84% of days. Mean change in body mass was −4.4, −6.7, and −1.0 kg in CR, RCR, and TRE, respectively. CR and RCR lost fat and lean soft tissue. Chair stand and 400-m walk times improved in RCR and TRE, and there were improvements in glucose and cholesterol levels in all 3 groups.

Conclusions

HALLO-P showed that the CR and TRE interventions were feasible and associated with improvements in health status over 9 months despite not achieving the target CR.

Introduction

Multimorbidity, the presence of 2 or more chronic health conditions, rises exponentially with age and is associated with greater healthcare utilization and costs, poorer physical and cognitive functioning and quality of life, and increased rates of mortality.^1–3 Obesity is a major risk factor for multimorbidity,⁴ as well as other common disabling conditions such as osteoarthritis, cardiovascular disease, and diabetes. The prevalence of obesity (body mass index [BMI] ≥ 30 kg/m²) among older adults in the United States has almost doubled over the last 3 decades, increasing from 22% in 1988-1994 to 38% in 2021-2023.⁵ In addition, the number of older adults is rapidly increasing in the United States—increasing from ∼35 million in 2000 to an estimated 71 million by 2030.⁵ These projections suggest a concomitant rise in the number of older adults with obesity and multimorbidity. Thus, interventions that reduce age-related multimorbidity and extend health span—either through the treatment of overweight/obesity or by slowing the aging process—are urgently needed.

Findings from animal models of caloric restriction (CR) show that CR prolongs lifespan and delays the onset of age-related health conditions.^6,7 However, the effects of CR in extending lifespan and healthspan in humans remain unknown. In the CALERIE trial of young and middle-aged adults without obesity who were randomized to CR or an ad libitum diet for 2 years, many health benefits were observed, including improvements in metabolic health and inflammation and reductions in biological aging, in the CR compared with the ad libitum group.⁸ Moderate CR in older adults with overweight or obesity has been shown to improve metabolic and functional health in short-term randomized controlled trials (eg, lower IL-6, blood pressure, and glucose and faster gait speed).^9,10 In the Look AHEAD trial, which randomized 5145 persons aged 45-76 years with type 2 diabetes and overweight or obesity to either an intensive lifestyle intervention (ILI) comprised of CR and physical activity or a diabetes support and education program over 9-11 years, ILI was associated with reduced rates of advancing multimorbidity, better mobility function, and lower health care utilization.^11–13

Studies of time-restricted feeding in rodent models also demonstrate extended lifespan and healthspan.¹⁴ Studies of time-restricted eating (TRE) in humans report short-term metabolic benefits, and a growing body of evidence suggests that TRE targets biological pathways affecting aging and may be an alternative to CR in adults with overweight and obesity.¹⁵ Most TRE studies with ad libitum caloric intake show a 2%-4% weight loss in individuals with overweight or obesity, although the magnitude of TRE’s effects on weight are small compared with CR, and beneficial effects on cardiometabolic outcomes are mixed.^16–18 Additionally, few studies of TRE have been conducted in older adults with overweight or obesity.^19–22

The Health, Aging, and Later-Life Outcomes Pilot (HALLO-P) study assessed the feasibility and acceptability of 3 interventions—20% CR delivered in-person; 20% CR delivered remotely via video conferencing (RCR); and 8-hour TRE with ad libitum caloric intake—to inform the design of a full-scale randomized controlled trial to evaluate the long-term effects of CR and/or TRE on the health of older adults with overweight or obesity. Here we report on the primary feasibility outcomes: (1) degree of sustained CR; (2) TRE sustainability; and (3) retention of study participants. Secondary outcomes included change in weight, fat mass and lean soft tissue, muscle mass, bone mineral density, physical performance, physical activity, energy expenditure, dietary intake, and cardiometabolic risk factors.

Methods

As previously described, HALLO-P was a 9-month, 3-group, single-blind, randomized pilot trial (NCT05424042).²³ HALLO-P recruited community-dwelling men and women aged 60 years and older from the Forsyth County, NC, area using direct mailing and advertisements between July 2022 and September 2023. The HALLO-P target recruitment goal was approximately 100 older adults with an indication for weight loss for whom participation was feasible and safe. The Institutional Review Board of the Wake Forest University School of Medicine approved all study-related procedures (IRB00072563), and all participants provided written informed consent.

Participants

Interested individuals were eligible for inclusion if they: were aged 60+ years; had a BMI of 30-37 kg/m² (originally a BMI of 30-35 kg/m² but expanded to 30-37 kg/m² in order to increase African American enrollment) or 27-<30 kg/m² with at least one obesity-related comorbidity; did not have a history of eating or nutritional disorders, comorbid conditions that could create a safety risk related to involvement in the interventions, or were taking medications that could interfere with the intervention or study outcomes; did not volunteer or do paid work during the late evening or night hours; were consuming calories over an eating window of at least 11 hours daily; and were willing to participate in a 9-month intervention study of CR or TRE. Eligibility was determined through the administration of a telephone screening and a subsequent in-person screening visit. A complete list of inclusion/exclusion criteria is in Table S1.

Study interventions

Participants were randomized individually to 1 of the 3 intervention groups, with stratification by sex, in waves consisting of 8-12 participants per group once at least 24 participants were eligible in a 1:1:1 ratio during the first 3 intervention waves and then in a 0:1:1 ratio, excluding the CR group, in the fourth and final intervention wave using a web-based randomization scheme for 9 months: (1) in-person 20% CR (n = 22); (2) remote (via video conferencing) 20% CR (RCR; n = 34); (3) in-person 8-hour TRE (n = 34). All 3 intervention groups met once a month individually and 3 times a month in a group setting with a behavioral coach and/or a registered dietitian during the first 6 months. During months 7 through 9 of the intervention, there was 1 individual and 1 group session each month.

The CR, RCR, and TRE interventions have been described elsewhere.²⁴ Briefly, each of the interventions combined specific nutritional and behavioral guidance for each participant to achieve adherence to treatment goals. All participants were encouraged to increase their daily steps by 20% from baseline levels, movement spread out across the day, and communicate with the intervention team via a study-specific “Companion App.” Participants in the CR and RCR groups were assigned a calorie level that would provide a daily caloric deficit of approximately 20% derived by subtracting 20% from each person’s estimated daily energy needs for weight maintenance determined by measuring baseline total energy expenditure (TEE) over 10-15 days using doubly labeled water (DLW). Participants in the CR and RCR interventions kept diet records using a study-provided tablet computer, and body weight was recorded daily using a cellular data collection-enabled electronic smart scale. Daily body weights were plotted within a “Zone of Adherence,” which displayed predicted weight loss given adherence to calorie goals generated by the National Institute of Diabetes and Digestive and Kidney Diseases (NIDDK) body weight planner.²⁵ Diet records and body weight were reviewed regularly with group leaders to revise calorie goals or address challenges as they arose. During group and individual sessions, participants in the TRE intervention received education on healthy eating and the benefits of TRE and discussed challenges associated with eating in a restricted window (eg, remaining sated during the evening, enjoying social interactions outside of the eating window) and potential solutions. They were not explicitly instructed to reduce their calorie intake. TRE participants logged the time they began and ended eating each day into the Companion app and were considered adherent if they ate within an 8.5-hour period. Daily step counts, body weight (CR groups only), macronutrient intake (CR groups only), timing of meals/snacks (TRE group only), and Companion App usage during the 9-month pilot were tracked, as these data were integral to the provision of each intervention. All participants stayed on the intervention until follow-up visits were completed.

Study measures

Study measures were collected at either screening (weight, body composition, and cardiometabolic biomarkers) or baseline (TEE, D₃-creatine, physical performance, physical activity, and dietary intake) prior to randomization and the start of intervention and again following the 9-month intervention by trained staff blinded to the intervention groups.

Weight and body composition

Weight (to the nearest hundredth of a kilogram) was measured on a balance beam scale and height (to the nearest tenth of a centimeter) using a stadiometer without shoes. BMI (kg/m²) was calculated using measured weight and height at baseline. Waist circumference (to the nearest tenth of an inch) was measured at the narrowest part of the torso between the rib cage and above the umbilicus.

Fat mass (FM) and whole body and appendicular lean soft tissue (ALST) mass were measured by dual-energy X-ray absorptiometry (DXA; Hologic Horizon A, APEX Software, version 5.6.x, Bedford, MA). Whole body scans were acquired with the participant supine and aligned with the scanner table. Bone mineral content was subtracted from total fat-free mass to determine whole body lean soft tissue (LST). ALST was calculated as the sum of LST in arms and legs, assuming that all nonfat and nonbone tissue is skeletal muscle. Areal bone mineral density (BMD) of the total hip, femoral neck, and lumbar spine was also assessed by DXA. All scans were performed and analyzed by a DXA technician certified by the International Society for Clinical Densitometry.

Muscle mass was estimated using the D₃-creatine (D₃Cr) method.²⁶ Participants were instructed to ingest a 30 mg oral dose of deuterated creatine and provide a fasted urine sample 3-6 days postdose. Urine samples were frozen at −70 °C until shipment to the University of California at Berkeley on dry ice for batch analysis of D₃Cr enrichment, as well as concentrations of unlabeled creatine and creatinine, by liquid chromatography mass spectrometry.²⁷ The amount of D₃Cr estimated to have been excreted (mg) and the mean steady-state D₃Cr enrichment ratio were input into a published algorithm to determine total body creatine pool size. Whole body muscle mass was then estimated by dividing the creatine pool size by 4.3 g/kg (which represents the average concentration of creatine in whole wet muscle).

Total Energy Expenditure (TEE)

TEE was assessed using DLW according to a protocol previously described.²⁸ Participants received a dose of DLW after providing a fasting urine sample; 2 additional fasting urine samples were collected 4.5 and 5 hours later. Participants returned to the clinic 10-15 days later, and 2 additional fasting urine samples were collected for DLW analysis. We assumed that baseline energy intake equaled baseline TEE. Energy intake at postintervention and during the intervention was calculated from TEE and change in body energy stores²⁹:

$$\mathit{Energy} \mathit{Intake}=\mathit{TEE}+\Delta \mathit{body} \mathit{energy} \mathit{stores}$$

Energy intake during the intervention was determined by combining pre- and postintervention DLW TEE measures and changes in FM and LST measured by DXA over the 9-month intervention:

$$\Delta \mathit{Energy} \mathit{Stores}=\Delta FM\times \mathit{energy} \mathit{equivalent}+\Delta \mathit{LST}\times \mathit{energy} \mathit{equivalent}$$

where energy equivalents of loss of 1 g FM = 9.3 kcal and 1 g LST = 1.1 kcal, while gain of FM = 13.2 kcal/g and LST = 2.2 kcal/g.^30,31

Percent CR over the 9-month intervention based on DLW was calculated as:

$$1-\left(\mathit{Energy} \mathit{Intake}/\mathit{Baseline} \mathit{TEE}\right)\mathrm{\hspace{0.17em}*\hspace{0.17em}}100$$

Physical performance

The Short Physical Performance Battery (SPPB) was administered to assess lower extremity physical performance.³² The SPPB consists of standing balance tasks (side-by-side, semi- and full-tandem stands), a 4-m walk to assess usual gait speed, and time to complete 5 repeated chair stands. Each of the 3 performance measures is assigned a score ranging from 0 (inability to perform the task) to 4 (the highest level of performance) and summed to create an SPPB score ranging from 0 to 12 (best). The SPPB was expanded to increase the holding time of the standing balance tasks from 10 to 30 seconds and added a single leg stand and a narrow 4-m walk test of balance (walking at a usual pace within lines of tape spaced 20 cm apart) to minimize ceiling effects of the SPPB.³³ The expanded SPPB scores are continuous and range from 0 to 4, with higher scores indicative of better performance. The fast-paced 400-m walk, a walking-based test of exercise tolerance and aerobic fitness, was also administered.³⁴ Participants were instructed to walk 400 m (10 laps of a 20-m course) on a flat indoor surface as quickly as possible at a maintainable pace, and the time to complete the walk was recorded in minutes and seconds. Finally, grip strength was measured twice in each hand using an isometric hydraulic hand dynamometer (Jamar, Bolingbrook, IL), with the mean of the 2 measurements on the dominant hand used in analyses. Participants were excluded from performing the test if they reported hand pain or recent hand or wrist surgery.

Physical activity

Participants wore an ActivPAL^TM 4 (PAL Technologies Ltd., Glasgow, UK) thigh-mounted triaxial accelerometer for 10-15 days. Data were analyzed using PALBatch software version 8.11.1.63. Behaviors, including daily stepping, were classified via the GHLA (version 2.2) algorithm following a 24-hour wear protocol such that days with at least 20 hours of wear were classified as valid and included in subsequent analyses. Average daily steps for participants with 4+ valid days are reported.

Cardiometabolic risk factors

Blood pressure was measured with an automatic sphygmomanometer after the participant had been seated for at least 5 min. Blood was drawn in the morning after an overnight fast. Fasting glucose, HbA1c, triglycerides, and total, HDL, and LDL cholesterol were measured at a clinical lab (LabCorp).

Dietary intake

Participants were asked to complete three 24-hour dietary recalls (2 weekdays and 1 weekend day) using the National Cancer Institute’s Automated Self-Administered 24-Hour (ASA24) Dietary Assessment Tool (https://epi.grants.cancer.gov/asa24). The ASA24 is based on the validated Multiple-Pass Method³⁵ and has been shown to estimate total energy and protein intakes to proximate recovery biomarkers.³⁶ The first 24-hour dietary recall at baseline and 9-month follow-up was completed during a clinic visit to orient participants; staff subsequently prompted participants to complete 2 additional 24-hour dietary recalls on their own using the study-provided tablet. Healthy Eating Index (HEI)-2020 scores (ranging from 0 to 100, with higher scores indicating better alignment with the US Dietary Guidelines) were calculated to assess diet quality.³⁷

Adverse events (AEs)

AEs were recorded during the monthly one-on-one intervention sessions, at each study visit, and by the voluntary reporting of participants at any time during the study. Adverse events were coded according to MedDRA categories and relatedness to the intervention. Participants also completed a signs and symptoms checklist at each monthly one-on-one intervention session and study visit that queried expected signs and symptoms related to the intervention: nausea/vomiting, headache, diarrhea, constipation, fatigue, irregular heartbeat, feeling faint, lightheadedness, dizziness, irritability, difficulty concentrating, low energy levels, low blood sugar, depression, joint pain/discomfort, muscle pain/discomfort, or skin irritation due to wearing study devices. Expected signs and symptoms were only recorded as an AE if they were severe (ie, interrupted the participant’s usual daily activity, required systemic drug therapy or other treatment, or potentially life-threatening or incapacitating).

Statistical analyses

Participants were stratified by the group to which they were randomized, and baseline characteristics were presented as mean (SD) or frequency (%). Consistent with recommendations from CONSORT for randomized trials, we did not test for balance among randomized groups. As a pilot study, we focused our analyses on estimation, rather than hypothesis testing of between-group differences, and present results in terms of means and 95% CIs within groups. Although there were only 2 measurements per individual (ie, baseline and 9 months) for most outcomes (ie, anthropometrics, body composition, physical performance, energy expenditure and self-reported dietary intake, cardiometabolic risk factors), we used contrasts with linear mixed-effects models to estimate group-specific least squares means at baseline, 9-month follow-up, and the change from baseline to follow-up. An unstructured covariance matrix was used to characterize the dependency between repeated measurements. This approach was chosen to more appropriately account for missing data and implement an intent-to-treat analysis, in which all participants were included in the analysis within the groups to which they were randomized. To account for potential dependency of responses for waves of participants receiving the interventions within intervention groups, sensitivity analyses were performed, adding random effects representing these groups of participants. For the majority of outcome measures, the variance component associated with this random effect was estimated as being zero or had an inconsequential effect on the estimated variances; thus, we present results from models without these random effects. AEs were presented in terms of counts and percentages by randomized group.

Results

Figure 1 provides the CONSORT diagram from prescreen contacts through 9-month follow-up. Of the 1753 prescreen contacts, 678 completed telephone screens, and 139 individuals attended an initial in-person screening visit and consented to participate. Of these 139 individuals, 135 were screened and 45 participants were excluded, leaving 90 participants who were randomized to CR (n = 22), RCR (n = 34), and TRE (n = 34). Seven participants withdrew (5 in TRE and 2 in CR), and 4 participants discontinued intervention but completed all study visits (2 in TRE, 1 in CR, and 1 in RCR) for an overall retention rate of 92%.

Figure 1.

HALLO-P CONSORT diagram.

The baseline characteristics of randomized participants by group are shown in Table 1. At baseline, participants’ mean (SD) age, BMI, and SPPB score were 67.2 (4.9) years, 31.7 (2.9) kg/m², and 11.4 (0.8), respectively; 62% were women and 83% were White. The most common comorbidities were arthritis (66%), hypertension (53%), and hyperlipidemia (52%). During the 9-month intervention, the median interquartile range (IQR) overall attendance at intervention sessions was 84% (70, 94): 83% (66, 90) in CR, 88% (75, 94) in RCR, and 80% (63, 94) in TRE. The median (IQR) percentage of days the Companion App was used was 96% (87, 100): 96% (90, 99) in CR, 98% (94, 100) in RCR, and 93% (64, 99) in TRE. In CR and RCR, the median percentage of days daily weights were provided was 61% (50, 75): 60% (34, 80) in CR and 61% (53, 72) in RCR. Based on the NIDDK body weight planner, daily body weights for 90% of participants in the CR group and 70% of participants in the RCR group were in the “Zone of Adherence” at least 75% of the time during the first 3 months of intervention; however, in the last 3 months of intervention, only 44% of participants in the CR group and 37% of participants in the RCR group had daily body weights in the “Zone of Adherence” at least 75% of the time. Participants increased their daily step count by an average (SD) of 17% (44%), 23% (40%), and 10% (25%) in CR, RCR, and TRE groups, respectively.

Table 1.

Participant characteristics at baseline overall and by intervention group: HALLO-P.

	Overall (n = 90)	CR (n = 22)	RCR (n = 34)	TRE (n = 34)
Age, mean (SD) (years)	67.2 (4.9)	66.6 (5.0)	68.7 (4.6)	66.1 (5.0)
Sex, n (%)
Male	34 (37.8)	9 (40.9)	13 (38.2)	12 (35.3)
Female	56 (62.2)	13 (59.1)	21 (61.8)	22 (64.7)
Race, n (%)
African American/Black	14 (15.6)	4 (18.2)	5 (14.7)	5 (14.7)
Caucasian/White	75 (83.3)	18 (81.8)	29 (85.3)	28 (82.4)
Other	1 (1.1)	0 (0.0)	0 (0.0)	1 (2.9)
Hispanic ethnicity, n (%)	2 (2.2)	0 (0.0)	0 (0.0)	2 (5.9)
Education, n (%)
High school or GED	9 (10.0)	0 (0.0)	4 (11.8)	5 (14.7)
Trade/vocational	1 (1.1)	1 (4.5)	0 (0.0)	0 (0.0)
Undergraduate degree	51 (56.7)	12 (54.5)	18 (52.9)	21 (61.8)
Graduate degree	29 (32.2)	9 (40.9)	12 (35.3)	8 (23.5)
Weight, mean (SD) (kg)	90.3 (12.9)	90.5 (11.9)	90.8 (14.0)	89.7 (12.7)
Height, mean (SD) (cm)	168.5 (8.7)	169.8 (8.9)	168.9 (9.1)	167.3 (8.3)
BMI, mean (SD) (kg/m2)	31.7 (2.9)	31.4 (2.9)	31.7 (2.9)	32.0 (2.9)
MoCA score, mean (SD)	26.1 (2.0)	26.0 (1.9)	26.2 (2.0)	26.0 (2.1)
CES-D score, mean (SD)	4.6 (3.7)	3.7 (3.3)	5.2 (4.1)	4.6 (3.5)
Chronic conditions, n (%)
CVDa	11 (12.2)	2 (9.1)	7 (20.6)	2 (5.9)
COPD/emphysema	2 (2.2)	1 (4.5)	0 (0.0)	1 (2.9)
Type 2 diabetes	3 (3.3)	2 (9.1)	1 (2.9)	0 (0.0)
Cancerb	14 (15.6)	3 (13.6)	9 (26.5)	2 (5.9)
Arthritis/joint pain	59 (65.6)	11 (50.0)	26 (76.5)	22 (64.7)
Hypertension	48 (53.3)	12 (54.5)	19 (55.9)	17 (50.0)
Hyperlipidemia	47 (52.2)	13 (59.1)	15 (44.1)	19 (55.9)
SPPB score, mean (SD)	11.4 (0.8)	11.7 (0.6)	11.3 (0.8)	11.4 (1.0)
Expanded SPPB score, mean (SD)	2.55 (0.34)	2.58 (0.35)	2.49 (0.31)	2.59 (0.36)
4-m gait speed (usual), mean (SD) (m/sec)	1.27 (0.22)	1.29 (0.26)	1.22 (0.21)	1.30 (0.21)
Chair stand time, mean (SD) (sec)	10.4 (2.6)	9.8 (2.0)	10.8 (2.7)	10.5 (2.8)
400-m walk time (fast), mean (SD) (sec)	299.8 (40.0)	292.4 (42.4)	305.2 (42.2)	298.7 (36.4)
Grip strength, mean (SD) (kg)	31.4 (9.4)	32.1 (8.8)	32.1 (11.3)	30.3 (7.8)
Step count, mean (SD)	6962 (2514)	7201 (2536)	6411 (1888)	7327 (2960)
Dietary intake, mean (SD)
Total energy (kcals/d)	2075 (685)	2152 (769)	2038 (632)	2063 (697)
Fat (g/d)	92.9 (35.2)	99.2 (40.5)	91.8 (30.4)	89.8 (36.5)
Carbohydrate (g/d)	223.5 (79.1)	221.2 (77.3)	212.4 (80.5)	236.4 (79.3)
Protein (g/d)	83.2 (33.6)	86.9 (40.0)	87.1 (31.9)	76.8 (30.4)
HEI score	58.0 (10.7)	56.1 (10.7)	59.2 (10.8)	57.9 (10.8)

Abbreviations: BMI, body mass index; CES-D, Center for Epidemiological Studies Depression scale; COPD, chronic obstructive pulmonary disease; CR, caloric restriction; CVD, cardiovascular disease; HALLO-P, Health, Aging, and Later-Life Outcomes Pilot; HEI, Healthy Eating Index; MoCA, Montreal Cognitive Assessment; RCR, remote caloric restriction; SPPB, Short Physical Performance Battery; TRE, time-restricted eating.

^aCVD: heart attack/myocardial infarction, stents, angioplasty, atrial fibrillation, congestive heart failure, stroke/transient ischemic attack.

^bCancer: excluding nonmelanoma skin cancers.

The mean (SD) and median (IQR) CR over the 9-month intervention based on DLW was 4.5% (11.0) and 7.5% (1.0, 11.4) in CR, 6.0% (10.3) and 8.2% (0.6, 13.1) in RCR, and 1.9% (7.2) and 1.4% (−1.3, 5.4) in TRE, respectively. Only 37% of CR and 40% of RCR participants maintained ≥10% CR at 9 months. The mean (SD) and median weight loss were 6.5% (5.9) and 7% in CR, 7.9% (5.8) and 8.5% in RCR, and 1.8% (4.4) and 1.3% in TRE, respectively. The association between percent of DLW-based CR and weight change over 9 months in CR and RCR combined: for a 1% decrease in calories over the 9 months, there was a 0.4% decrease in weight (β = −0.42, R² = 0.53, P < 0.0001; Figure S1). In the TRE group, the median (IQR) percentage of days participants ate within an 8.5-hour window was 84% (68, 95); however, only 63% of TRE participants ate within the window on at least 80% of days.

Table 2 shows the estimated baseline, follow-up, and change in anthropometrics and body composition from the mixed effects models over the 9-month intervention. On average, participants in the CR and RCR groups lost considerable weight (CR: –4.4 kg, 95% CI: −6.7, −2.1; RCR: −6.7 kg, 95% CI: −8.5, −4.8). The average weight loss among TRE participants was approximately 1 kg (95% CI: −2.9, 1.0). All 3 groups had a reduction in waist circumference at 9 months of approximately 1 inch or more. Among participants in the CR and RCR groups, total FM, LST, and ALST all decreased. The decrease in ALST in the TRE group (−0.3 kg, 95% CI: −0.7, 0.0) was approximately one-third of that observed in the RCR group (−1.0 kg, 95% CI: −1.3, −0.7). D₃Cr muscle mass decreased on average by 1.5 kg in the CR group (95% CI: −3.5, 0.6) but increased by a similar magnitude in the TRE group (1.9 kg, 95% CI: 0.1, 3.7). On average, total hip BMD decreased (−0.01 cm², 95% CI: −0.02, −0.00), but lumbar spine BMD increased (0.01 cm², 95% CI: 0.00, 0.02) in both the CR and RCR groups.

Table 2.

Change in anthropometrics and body composition over 9 months by intervention group: HALLO-P.

Measure	Baseline			Follow-up			Change (Follow-up—Baseline)
	CR	RCR	TRE	CR	RCR	TRE	CR	RCR	TRE
	LS means			LS means			LS means (95% CI)	LS means (95% CI)	LS means (95% CI)
Weight (kg)	90.5	90.8	89.7	86.1	84.1	88.8	−4.4 (−6.8, −2.1)	−6.7 (−8.5, −4.8)	−1.0 (−2.9, 1.0)
BMI (kg/m2)	31.4	31.7	32.0	30.1	29.5	31.7	−1.3 (−2.1, −0.5)	−2.2 (−2.9, −1.6)	−0.2 (−0.9, 0.5)
Waist circumference (in)	40.6	40.8	40.8	39.2	38.4	39.9	−1.4 (−2.3, −0.4)	−2.4 (−3.1, −1.6)	−0.9 (−1.7, −0.1)
Total fat mass (kg)	38.2	38.9	37.9	35.3	34.5	38.0	−2.9 (−4.7, −1.1)	−4.5 (−5.8, −3.1)	0.0 (−1.5, 1.5)
Percent fat mass (%)	42.0	42.5	42.3	40.6	40.5	42.5	−1.4 (−2.4, −0.4)	−2.0 (−2.8, −1.3)	0.3 (−0.6, 1.1)
Total LST (kg)	50.7	50.1	49.9	49.0	47.9	49.3	−1.7 (−2.5, −0.9)	−2.2 (−2.8, −1.6)	−0.6 (−1.3, 0.1)
Percent LST (%)	55.2	54.4	54.9	56.4	56.1	54.6	1.2 (0.2, 2.2)	1.7 (1.0, 2.5)	−0.3 (−1.1, 0.5)
ALST (kg)	21.0	21.1	20.8	20.5	20.1	20.4	−0.5 (−0.9, −0.1)	−1.0 (−1.3, −0.7)	−0.3 (−0.7, 0.0)
D3 muscle mass (kg)	31.1	27.7	28.7	29.6	28.4	30.6	−1.5 (−3.5, 0.6)	0.6 (−0.9, 2.2)	1.9 (0.1, 3.7)
Percent D3 muscle mass (%)	33.8	30.3	31.5	34.4	33.8	33.8	0.5 (−1.7, 2.7)	3.5 (1.8, 5.2)	2.3 (0.4, 4.2)
Total hip BMD (g/cm2)	0.96	0.91	0.93	0.95	0.90	0.93	−0.01 (−0.02, 0.00)	−0.01 (−0.02, −0.00)	−0.00 (−0.01, 0.01)
Femoral neck BMD (g/cm2)	0.80	0.74	0.76	0.80	0.74	0.76	0.00 (−0.01, 0.02)	−0.00 (−0.01, 0.01)	0.00 (−0.01, 0.01)
Lumbar spine BMD (g/cm2)	1.14	1.11	1.08	1.15	1.12	1.09	0.01 (−0.00, 0.02)	0.01 (0.00, 0.02)	0.01 (−0.01, 0.02)

Abbreviations: ALST, appendicular lean soft tissue; BMD, bone mineral density; BMI, body mass index; CR, caloric restriction; D3, D3-creatine; HALLO-P, Health, Aging, and Later-Life Outcomes Pilot; LST, lean soft tissue; RCR, remote caloric restriction; TRE, time-restricted eating.

Estimated mean baseline, follow-up, and change in physical performance and physical activity over the 9-month intervention are shown in Table 3. On average, SPPB score increased in participants in the TRE group (0.3 points, 95% CI: −0.0, 0.6), while the expanded SPPB score increased in participants in the RCR group (0.08 points, 95% CI: −0.01, 0.17). Usual 4-m gait speed declined, on average, by 0.15 (95% CI: −0.24, −0.06) and 0.12 (95% CI: −0.19, −0.04) m/sec in the CR and TRE groups, while 400-m walk time improved by approximately 20 seconds in both the RCR (−19.8, 95% CI: −29.3, −10.2) and TRE (−19.2, 95% CI: −29.8, −8.6) groups. Repeated chair stand time also improved by approximately 1 second in the RCR (−1.14, 95% CI: −1.81, −0.47) and TRE (−1.04, 95% CI: −1.76, −0.32) groups. On average, participants in the RCR group improved their step count by approximately 1500 steps (95% CI: 595, 2407).

Table 3.

Change in physical performance and physical activity over 9 months by intervention group: HALLO-P.

Measure	Baseline			Follow-up			Change (follow-up—baseline)
	CR	RCR	TRE	CR	RCR	TRE	CR	RCR	TRE
	LS means			LS means			LS means (95% CI)	LS means (95% CI)	LS means (95% CI)
SPPB score (range, 0-12)	11.7	11.3	11.4	11.7	11.5	11.7	−0.0 (−0.4, 0.3)	0.2 (−0.1, 0.4)	0.3 (−0.0, 0.6)
Expanded SPPB score (range, 0-4)	2.58	2.48	2.58	2.50	2.56	2.57	−0.09 (−0.20, 0.03)	0.08 (−0.01, 0.17)	−0.01 (−0.11, 0.09)
4-m gait speed (usual) (m/sec)	1.29	1.22	1.30	1.14	1.18	1.18	−0.15 (−0.24, −0.06)	−0.04 (−0.11, 0.03)	−0.12 (−0.19, −0.04)
Repeated chair stand time (sec)	9.77	10.82	10.48	9.42	9.68	9.44	−0.35 (−1.21, 0.52)	−1.14 (−1.81, −0.47)	−1.04 (−1.76, −0.32)
400-m walk time (sec)	292.25	305.18	298.71	285.20	285.41	279.52	−7.05 (−20.02, 5.92)	−19.76 (−29.28, −10.25)	−19.18 (−29.80, −8.56)
Grip strength (kg)	32.1	32.0	30.3	30.8	31.4	30.1	−1.4 (−3.6, 0.9)	−0.5 (−2.2, 1.2)	−0.2 (−2.0, 1.6)
Step count	7110	6453	7327	7790	7954	7701	680 (−496, 1856)	1501 (595, 2407)	374 (−615, 1362)

Abbreviations: CR, caloric restriction; HALLO-P, Health, Aging, and Later-Life Outcomes Pilot; RCR, remote caloric restriction; SPPB, Short Physical Performance Battery; TRE, time-restricted eating.

Estimated mean baseline, follow-up, and change in cardiometabolic risk factors are shown in Table 4. On average, changes from baseline to 9 months of systolic blood pressure were not observed in any group; however, changes in diastolic blood pressure were approximately 3 mm Hg lower in the RCR group (95% CI: −5.9, −0.5). Glucose improved on average by 5-8 mg/dL in all 3 groups (CR: −7.6, 95% CI: −11.8, −3.5; RCR: −5.8, 95% CI: −9.0, −2.6; TRE: −4.5, 95% CI: −7.9, −1.1). On average, total and LDL cholesterol decreased by approximately 16 mg/dL (95% CI: −26.7, −4.4 and −25.7, −6.9, respectively) and triglycerides by 18 mg/dL (95% CI: −35.7, −0.4) in the CR group, while HDL cholesterol improved by approximately 5 mg/dL in the RCR and TRE groups (95% CI: 1.5, 8.0, and 1.6, 8.6, respectively).

Table 4.

Change in cardiometabolic risk factors over 9 months by intervention group: HALLO-P.

Measure	Baseline			Follow-up			Change (Follow-up—baseline)
	CR	RCR	TRE	CR	RCR	TRE	CR	RCR	TRE
	LS means			LS means			LS means (95% CI)	LS means (95% CI)	LS means (95% CI)
Systolic blood pressure (mm Hg)	135.4	135.6	135.8	134.1	131.7	134.6	−1.2 (−8.8, 6.3)	−3.8 (−9.8, 2.1)	−1.2 (−7.4, 5.0)
Diastolic blood pressure (mm Hg)	79.0	76.7	77.0	76.7	73.5	77.3	−2.2 (−5.7, 1.2)	−3.2 (−5.9, −0.5)	0.4 (−2.5, 3.3)
Glucose (mg/dL)	102.9	98.1	97.1	95.3	92.4	92.6	−7.6 (−11.8, −3.5)	−5.8 (−9.0, −2.6)	−4.5 (−7.9, −1.1)
HbA1c (%)	5.7	5.6	5.7	5.8	5.6	5.8	0.0 (−0.1, 0.2)	−0.0 (−0.1, 0.1)	0.1 (−0.0, 0.2)
Cholesterol (mg/dL)	187.5	185.0	193.2	172.0	183.2	192.0	−15.5 (−26.7, −4.4)	−1.8 (−10.4, 6.8)	−1.2 (−10.4, 8.1)
HDL (mg/dL)	52.9	57.2	56.4	56.4	61.9	61.5	3.6 (−0.6, 7.8)	4.7 (1.5, 8.0)	5.1 (1.6, 8.6)
LDL (mg/dL)	115.3	109.4	116.2	99.0	104.7	111.2	−16.3 (−25.7, −6.9)	−4.7 (−12.0, 2.5)	−5.0 (−12.8, 2.8)
Triglycerides (mg/dL)	105.1	101.1	114.4	87.0	89.7	106.6	−18.1 (−35.7, −0.4)	−11.4 (−25.1, 2.3)	−7.7 (−22.3, 6.8)

Abbreviations: CR, caloric restriction; HALLO-P, Health, Aging, and Later-Life Outcomes Pilot; HbA1c, hemoglobin A1c; HDL, high-density lipoprotein; LDL, low-density lipoprotein; RCR, remote caloric restriction; TRE, time-restricted eating.

On average, there was little change in TEE by DLW at 9 months in all 3 groups (TEE: CR −27, 95% CI: −158, 104; RCR: −17, 95% CI: −121, 87; and TRE: −97, 95% CI: −207, 12). Change in self-reported energy intake at 9 months decreased on average between 244 and 371 kcals/d in all 3 groups (CR: −371, 95% CI: −612, −131; RCR: −244, 95% CI: −433, −55; and TRE: −256, 95% CI: −453, −58; Table S2). Average carbohydrate intake also decreased in all 3 groups, while fat intake decreased in the CR and RCR groups. On average, HEI scores improved by 7 points in the CR group (95% CI: 1.2, 13.0) and 5 points in participants in the RCR group (95% CI: 0.3, 9.4).

There were 187 AEs during the trial (43 in the CR group, 82 in the RCR group, and 62 in the TRE group; Table S3). Three AEs were definitely related to study procedures (2 in the CR group: rash from wearing the ActivPAL and syncope during blood draw; 1 in the TRE group: injury [bleeding] at the continuous glucose monitor site), and one AE was possibly related to study intervention (TRE group: dizziness). There were 6 serious adverse events reported during the trial (1 in the CR group, 2 in the RCR group, and 3 in the TRE group)—none of which were related to the study intervention or procedures. There were 232 expected signs/symptoms that were reported (53 in the CR group, 106 in the RCR group, and 73 in the TRE group)—all of which were reported as being mild or moderate in severity. The most common (>20 events) AEs and signs/symptoms reported were pain in the extremities, upper respiratory infections, fatigue, and myalgia.

Discussion

In this 9-month single-blind, randomized controlled trial of older adults with overweight or obesity, we assessed the feasibility of 3 interventions: 20% CR delivered in-person, 20% CR delivered remotely via video conferencing (RCR), and 8-hour TRE with ad libitum caloric intake. Participants in the CR and RCR groups sustained an average of 4.5% and 6.0% CR, respectively, over 9 months based on DLW, with at least 50% of participants achieving CR of 7.5% and 8.2% in the CR and RCR groups, respectively. This degree of CR resulted in an average of 6.5% and 7.9% weight loss in the CR and RCR groups, respectively, with at least 50% of participants achieving 7% and 8.5% weight loss in the CR and RCR groups, respectively. Although this degree of CR fell short of the 20% CR target, it still resulted in improvements in fat mass, physical function, and cardiometabolic risk factors. Participants in the TRE group had a median of 84% of days eating within an 8.5-hour window. Retention was excellent across all groups, ranging from 85% in TRE to 100% in RCR.

Similar to prior CR trials in older adults with overweight or obesity,^10,38–42 participants in CR and RCR lost DXA-derived total and ALST in addition to fat mass. There was also a concomitant decrease in D₃Cr muscle mass on average in the CR, but not RCR, group; however, the 95% CI included values consistent with an increase in D₃Cr in both the CR and RCR groups. This finding of decreased DXA-derived LST but preserved D₃Cr muscle mass has been observed in another small pilot study of CR in older adults.⁴³ Furthermore, the decrease in DXA-derived ALST in the CR and RCR groups was approximately one-third to almost one-half of the total LST lost, suggesting that a considerable portion of the LST lost during CR is due to losses in organ tissue (eg, kidney and liver), fibrotic and connective tissue, and water rather than skeletal muscle. As observed in prior CR trials in older adults with overweight or obesity,^10,38,44 total hip BMD also decreased in CR and RCR participants. Most prior CR trials in older adults with overweight or obesity have observed improvements in physical performance.^{10,38,40,45,46} Although lower extremity physical performance did not appreciably change in the CR and RCR participants and 4-m usual gait speed declined in CR participants, there were small but clinically meaningful improvements in the fast-paced 400-m walk and chair stand times in RCR participants,^47,48 which may be due in part to larger changes in both body mass and daily steps in this condition. Similar to prior CR trials,⁹ cardiometabolic risk factors also improved in CR and RCR, with glucose improving in both CR and RCR participants while cholesterol, LDL, and triglycerides improved in CR participants and HDL improved in RCR participants.

While prior CR trials have been conducted primarily using in-person contacts, HALLO-P included a remote CR arm to determine the feasibility of conducting CR trials remotely via videoconferencing. Few studies have examined remotely delivered CR interventions, particularly in older adults.⁴⁹ Delivery of an intervention remotely has the advantage of reducing both the cost of intervention delivery as well as the burden on participants who do not have to travel to a central location and can participate from their home. Reducing burden associated with participation was especially attractive for HALLO-P, as it was designed to plan for a future long-term behavioral intervention. In HALLO-P, the remote CR (RCR) group had the highest attendance at intervention sessions and achieved similar CR and weight loss as the in-person CR groups. Additionally, the RCR group had the highest satisfaction with their intervention.²⁴ Importantly, the RCR structure was developed hand-in-hand with older adults across several studies,^50,51 and this practice is critical for ensuring the intervention is perceived as both useful and usable by the target audience.⁵²

Prior TRE studies have been mixed, with some showing short-term metabolic benefits, including small amounts of weight loss and beneficial effects on cardiometabolic risk factors^16–18; however, few trials of TRE have been conducted in older adults. TRE studies in older adults with overweight or obesity have shown excellent adherence (>80%) to the prescribed TRE regimen; however, the studies were of short duration—only 6 weeks.^19–22 In HALLO-P, although the median percentage of days TRE participants ate within the 8-hour window was 84%, only 63% of TRE participants ate within the 8-hour window on at least 80% of days—possibly due to the longer 9-month duration of the trial. Prior TRE studies in older adults with overweight or obesity resulted in significant weight loss—though the magnitude of weight loss (∼2%) was small compared with CR.^19–22 Although TRE participants in HALLO-P did not lose appreciable weight (∼1%) over the 9-month intervention, their waist circumference did improve. Several prior TRE studies in older adults with overweight or obesity observed a decrease in fat mass, but not lean mass measured by bioelectrical impedance analysis.^20–22 In HALLO-P, the change in FM, LST, and BMD by DXA was minimal; however, D₃Cr muscle mass increased in TRE participants, possibly due to the increase in protein intake and daily steps. Only one study has examined the effects of TRE on physical performance in older adults with overweight or obesity and found an increase in the 6-minute walk distance.¹⁹ In TRE participants in HALLO-P, there was a trend toward improved lower extremity physical performance even though 4-m usual gait speed declined, and there were small but clinically meaningful improvements in the fast paced 400-m walk and chair stand times.^47,48 Some studies of TRE in middle-aged and younger adults have shown cardiometabolic risk factor improvement, including beneficial effects on blood pressure and glucose control.^16–18 In HALLO-P, both glucose and HDL cholesterol improved in TRE participants.

HALLO-P had several strengths, including the use of a randomized controlled trial design, use of DLW to assess degree of CR, the inclusion of a remotely delivered CR group, and excellent adherence to the interventions. In addition to measuring LST by DXA, HALLO-P also measured muscle mass by D₃CR. There were also important limitations of this pilot study. First, as a pilot trial focused on refining study procedures and the interventions to inform the design of a full-scale randomized controlled trial to evaluate the long-term effects of CR and/or TRE on the health of older adults with overweight or obesity,^23,24 the final sample size was relatively small, and preplanned analyses focused on estimating changes within intervention groups rather than testing between groups. We tended to see some improvement in adherence and response to the interventions by wave in response to refinements to the interventions. Second, most participants were White and recruited from a single community in North Carolina, limiting the generalizability of our findings. Recruiting older adults for a trial where TRE was one of the 3 interventions was also challenging, as many older individuals reported eating below the minimum eating window of at least 11 hours daily. Similar to other trials of lifestyle-based CR, adherence to the CR intervention waned over time. Finally, the interventions were conducted in older adults with overweight or obesity, making it difficult to distinguish between the effects of CR and TRE on aging with the benefits of treating obesity.

In conclusion, HALLO-P found all 3 interventions—20% CR delivered in-person, 20% CR delivered remotely via video conferencing (RCR), and 8-hour TRE with ad libitum caloric intake—to be feasible in older adults with overweight or obesity; however, the CR groups did not achieve the 20% CR target. The in-person CR and remote CR groups sustained 4.5% and 6.0% CR, resulting in 6.5% and 7.9% weight loss, respectively, over the 9-month intervention, indicating that remote delivery of CR interventions in this population may be a viable alternative to in-person delivery given the remote CR group’s achieved CR and weight loss and high satisfaction with their intervention.²⁴ While this pilot examined the feasibility of a lifestyle-based CR intervention, glucagon-like peptide-1 receptor agonists (GLP-1RAs), which reduce caloric intake by delaying gastric emptying and suppressing appetite and enhancing satiety by acting on the central nervous system, have shown greater weight loss efficacy than lifestyle-based interventions in controlled trials.⁵³ GLP-1RAs have also been shown to have pleiotropic effects across multiple organ systems⁵⁴; thus, future research should examine the role of GLP-1RAs, in conjunction with lifestyle-based approaches, on lifespan and healthspan in older adults with obesity. The TRE group had a median of 84% of days eating within an 8.5-hour window and, although the TRE group did not lose appreciable weight, they had similar improvements in physical performance and cardiometabolic risk factors to the CR groups. Further studies are needed to determine if TRE may be an alternative to CR in older adults with overweight and obesity. Larger and longer-term trials are needed to definitively assess the effect of these interventions on reducing age-related multimorbidity in older adults with overweight and obesity.

Supplementary material

Supplementary material is available at The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences online.

Funding

This work was supported by grants from the National Institute on Aging: U01AG073240 and P30AG021332 (Wake Forest Claude D. Pepper Older Americans Independence Center). The authors acknowledge the use of the services and facilities of the Clinical Research Unit, funded by the National Center for Advancing Translational Sciences (UL1 TR001420).

Conflicts of interest

D.K.H., M.A.E., and S.B.K. serve on the editorial board of the Medical Sciences section of The Journals of Gerontology, Series A: Biological Sciences and Medical Sciences. J.D.A. reports receiving research support from Eli Lilly, Boehringer Ingelheim, KVKTech, WW, Novo Nordisk, Regeneron, and Amgen. J.D.A. has served as a consultant or on advisory boards for Eli Lilly, Novo Nordisk, Regeneron, Amgen, Zealand Pharma, and Boehringer Ingelheim. The other authors declare no conflict of interest.

Data availability

The data presented herein will be made available upon reasonable request via contact to the research team.

Author contributions

Trial conceptualization: Denise K. Houston, Jason Fanning, Barbara J. Nicklas, Michael E. Miller, W. Jack Rejeski, Stephen B. Kritchevsky. Funding acquisition: Denise K. Houston, Jason Fanning, Barbara J. Nicklas, Mark A. Espeland, Jamy D. Ard, Michael E. Miller, W. Jack Rejeski, Stephen B. Kritchevsky. Developing the interventions: Jason Fanning, W. Jack Rejeski. Developing the “Companion app”: Jason Fanning. Providing the doubly labeled water doses and analyses: James P. Delany. Project administration: Cynthia L. Stowe, Kimberly Kennedy. Overseeing the implementation of the trial: Denise K. Houston, Jason Fanning, Barbara J. Nicklas, Michael E. Miller, W. Jack Rejeski, Stephen B. Kritchevsky. Medical oversight of the trial: Jamy D. Ard. Data curation: Michael Walkup, Rebecca H. Neiberg, Cynthia L. Stowe, Michael E. Miller. Data analyses: Fang-Chi Hsu, Shyh-Huei Chen, Michael Walkup, Rebecca H. Neiberg, Michael E. Miller. Writing the original draft: Denise K. Houston. All authors critically reviewed the manuscript and read and approved the final version.