Abstract
Background:
Long-term (≥12 months) randomized trials evaluating the efficacy of intermittent fasting (IMF) as a dietary weight loss strategy are limited. Furthermore, no studies have compared IMF versus daily caloric restriction (DCR) when both interventions are provided in the context of a guidelines-based behavioral weight loss program.
Objective:
To compare the effects of 4:3 IMF versus DCR on changes in weight at 12 months, with comprehensive behavioral support provided to both groups.
Design:
Randomized clinical trial. (ClinicalTrials.gov: NCT03411356)
Setting:
Denver, Colorado, and surrounding metropolitan area.
Participants:
Adults aged 18 to 60 years with body mass index (BMI) of 27 to 46 kg/m2.
Intervention:
The IMF group was instructed to restrict energy intake by 80% on 3 nonconsecutive days per week, with ad libitum intake (no restriction) the other 4 days (4:3 IMF). The DCR group was instructed to reduce daily energy intake by 34% to match the weekly energy deficit of 4:3 IMF. Both groups received a high-intensity comprehensive behavioral weight loss program that included group-based behavioral support and a recommendation to increase moderate-intensity physical activity to 300 minutes per week.
Measurements:
The primary outcome was change in body weight (in kilograms) at 12 months.
Results:
Of the 165 (4:3 IMF, n = 84; DCR, n = 81) randomly assigned participants (mean age, 42 years [SD, 9]; mean BMI, 34.1 kg/m2 [SD, 4.4]; 73.9% female), 125 completed the trial. In an intention-to-treat analysis, 4:3 IMF showed greater reductions in weight than DCR at 12 months (mean difference, 2.89 kg [95% CI, 5.65 to 0.14 kg]; P = 0.040).
Limitation:
Limited generalizability.
Conclusion:
Compared with DCR, 4:3 IMF resulted in modestly greater weight loss among adults with overweight or obesity enrolled in a 12-month, high-intensity, comprehensive behavioral weight loss program.
Primary Funding Source:
National Institute of Diabetes and Digestive and Kidney Diseases.
Daily caloric restriction (DCR) is the standard dietary approach for treating overweight and obesity and involves daily reduction in energy intake. However, adherence to DCR is difficult, and most adults who lose weight with DCR show significant weight regain 1 year later (1–4). Intermittent fasting (IMF) has been suggested as an alternative dietary weight loss strategy. This approach involves cycling between complete or near-complete (>75%) energy restriction on “fast” days and ad libitum energy intake on nonfast days. Various IMF paradigms have been proposed, including 2 nonconsecutive fast days per week (5:2 IMF), 3 nonconsecutive fast days per week (4:3 IMF), and fasting every other day (alternate-day fasting [ADF]). Some have hypothesized that IMF enhances dietary adherence because of its simplicity and because the periodic nature of energy restriction may mitigate the constant hunger associated with DCR (5).
Despite the growing popularity of IMF, few rigorously designed long-term (≥12 months) trials have compared weight loss resulting from IMF versus DCR in adults with overweight or obesity (6–10). A recent systematic review (11) identified only 4 long-term (≥12 months) randomized trials comparing IMF (all evaluating a 5:2 paradigm) and DCR in which the targeted weekly energy deficit was matched between groups (6–9). We identified 1 additional study that compared ADF and DCR over 1 year with a matched targeted energy deficit (10). All 5 studies (6–10) showed no difference in weight loss between the IMF paradigm studied and DCR, although there were some limitations. Studies comparing IMF and DCR have not evaluated a 4:3 IMF paradigm, which may produce greater weight loss than a 5:2 IMF paradigm (due to the addition of 1 fast day) and enable better adherence than an ADF paradigm (due to flexibility in the selection of fast days). Furthermore, current guidelines (12) suggest that the most effective behavioral treatment for obesity is a high-intensity (≥14 sessions in the initial 6 months, with a recommendation for continued support for an additional 6 months), comprehensive weight loss intervention provided in individual or group sessions by a trained interventionist. However, prior studies have not evaluated IMF compared with DCR in the context of a guidelines-based, comprehensive weight loss intervention. Provision of comprehensive behavioral support may enhance adherence to IMF and result in greater weight loss than has been observed in prior studies.
The objective of this 12-month randomized controlled trial was to compare weight loss resulting from 4:3 IMF versus DCR (matched for targeted weekly energy deficit) in the context of a guidelines-based, high-intensity, comprehensive behavioral weight loss intervention. We hypothesized that 4:3 IMF would result in greater weight loss (primary outcome) at 12 months than DCR.
Methods
Design Overview
DRIFT (Daily Caloric Restriction vs Intermittent Fasting Trial) was a 12-month prospective, individually randomized group treatment trial. Participants underwent block randomization in a 1:1 ratio (stratified by sex) to 4:3 IMF or DCR. Trial recruitment occurred from 5 January 2018 to 29 April 2021, with the last 12-month follow-up on 26 July 2022. The study protocol was approved by the Colorado Multiple Institutional Review Board (COMIRB) of the University of Colorado, Aurora, Colorado (protocol 17-0369). Participants provided written informed consent. The trial is registered at ClinicalTrials.gov (NCT03411356). Detailed methods have been published previously (13), and the trial protocol is available at Annals.org. The trial was affected by the COVID-19 pandemic; the intervention and outcome assessment time points were modified slightly but remained consistent between groups and honored the integrity of the original trial design (see the Supplement, available at Annals.org).
Setting and Participants
The trial was conducted at the University of Colorado Anschutz Medical Campus. Participants were recruited from the surrounding Denver metropolitan area using e-mails, flyers, print advertisements, and social media recruiting strategies. Inclusion criteria were age 18 to 60 years and body mass index (BMI) of 27 to 46 kg/m2. Briefly, exclusion criteria were history of diabetes mellitus, cardiovascular disease, chronic kidney disease (stages 4 to 5), major depression, eating disorders, bariatric surgery, use of antiobesity medications or other medications known to significantly affect body weight, and weight gain above 5 kg in the previous 3 months. Women who were pregnant, lactating, or trying to become pregnant were also excluded. Full inclusion and exclusion criteria have been published previously (13).
Randomization and Interventions
The randomization assignment was generated by the study biostatistician using a computer-generated block randomization (block size = 4), performed within strata defined by sex.
Although it was not plausible to blind participants, they were blinded to the study hypothesis. Due to the nature of the group-based behavioral intervention, it was also not plausible to blind study staff who enrolled participants and delivered the intervention or monitored its delivery. However, core laboratory staff performing analysis of blood samples, doubly labeled water samples, and 7-day diet diaries were blinded to randomized assignment.
The weekly targeted dietary energy deficit from baseline weight maintenance requirements was the same for both groups. Maintenance energy requirements were calculated as resting energy expenditure (measured by indirect calorimetry) (14) multiplied by an activity factor of 1.5. All participants received group-based behavioral support and instruction in calorie counting and were guided to target dietary macronutrient content of 55% carbohydrates, 15% protein, and 30% fat.
4:3 IMF Dietary Strategy
Participants in the 4:3 IMF group were prescribed a modified fast on 3 nonconsecutive days per week. The fast day calorie goal was prescribed to produce an 80% energy restriction from baseline maintenance energy requirements. The targeted weekly energy deficit was 34.3% (80% energy restriction × 3 days per week / 7 days). Participants were asked to count calories and log food intake only on fast days. On nonfast days, participants were not asked to restrict energy intake and were permitted to eat ad libitum but were encouraged to make healthy food and portion choices.
DCR Dietary Strategy
Participants in the DCR group were prescribed a daily calorie goal designed to produce a 34.3% energy deficit from baseline maintenance energy requirements to match the targeted weekly energy deficit prescribed to the 4:3 IMF group. Participants were asked to count calories and log food daily.
Physical Activity Prescription
Both groups received a recommendation to increase to 300 minutes of moderate-intensity aerobic physical activity (PA) per week over the initial 6 months and to maintain this level of PA for the duration of the intervention, consistent with current PA guidelines for weight management (15, 16). Participants were provided a 12-month fitness center membership.
Group-Based Behavioral Support
Both groups received a 12-month group-based behavioral weight loss program. Randomized groups met separately, and meetings were held weekly during months 0 to 3 and every 2 weeks during months 4 to 12 and were led by a registered dietitian. The DCR curriculum used a skills-based approach and cognitive behavioral strategies for lifestyle modification with a dietary focus on DCR (17, 18). The 4:3 IMF curriculum was similar to the DCR curriculum but was adapted for a dietary focus on 4:3 IMF. Both groups received information on portion control, incorporation of healthier food choices, and calorie counting, although as described earlier, participants in the 4:3 IMF group were asked to count calories and restrict intake on fast days only. Additional behavioral support details have been published previously (13).
Outcomes and Follow-up
The primary outcome was the change in body weight (in kilograms) at 12 months. Body weight was measured in the morning after an overnight fast with a calibrated digital scale (to the nearest ±0.1 kg). All other outcomes were prespecified secondary outcomes.
Fat mass and lean mass were measured with dual-energy x-ray absorptiometry (Hologic Discovery W, Apex Software Version 4.5.3 [Hologic Inc.]). Waist circumference was measured with a tape measure parallel to the floor, superior to the iliac crest. Systolic and diastolic blood pressures were measured with a manual sphygmomanometer as the average of 2 seated values after 5 minutes of rest. Absolute percentage caloric restrictions from baseline achieved over 6 and 12 months were calculated using the doubly labeled water intake balance method (see the Supplement for details) (19). Self-reported dietary energy and macronutrient intake were determined from 7-day diet records (19), analyzed using Nutrition Data System for Research software (University of Minnesota). The Binge Eating Scale (20) was used to assess binge eating behaviors. Total moderate to vigorous PA (MVPA) was assessed over 7 days using the activPAL micro device (activPAL4 [PAL Technologies]). Days were considered valid if there was at least 94.5% wear time and at least 4 days of valid wear with at least 1 weekend day. A 12-hour fasting blood sample was obtained to assess glucose level, insulin level, hemoglobin A1c level, and lipid panel. In the 4:3 IMF group, the fasting blood sample was obtained after a nonfast day to limit any acute effects of fasting on outcomes of interest. Homeostasis model assessment of insulin resistance (HOMA-IR) was calculated as ([insulin level in μIU/mL] × [fasting glucose level in mg/dL × 0.055] / 22.5).
Self-reported adverse events were collected using case forms by study staff and were adjudicated systematically regarding severity (mild, moderate, severe, or serious) and attribution (definitely related, probably related, unlikely to be related, or unrelated to the study interventions) by the study’s principal investigator. Collated adverse events and trial progress were reviewed annually by an independent safety officer and by COMIRB.
Statistical Analysis
An a priori power calculation was performed for the primary outcome (weight) using nQuery 7.0 (Statistical Solutions Ltd), based on assumptions reported previously (13). A sample size of 150 (75 per group) was estimated to provide 90% power at 5% significance (P < 0.05) to detect a between-group difference of 3 kg in change in body weight at 12 months from baseline using an intention-to-treat analysis. Given concerns about attrition related to the COVID-19 pandemic, we increased enrollment by 10%. Results were analyzed using SAS, version 9.4 (SAS Institute), by one of the authors (Z.P.). Baseline characteristics were summarized using descriptive statistics. Normality assumptions were evaluated, and transformations were used as appropriate. The primary analysis used the intention-to-treat principle. Each continuous outcome was analyzed using a linear mixed-effects model with an unstructured covariance structure. For the primary analysis, baseline and 12-month body weight were dependent variables. The main effects were coded as dummy variables and included randomized group (4:3 IMF or DCR), time (0 or 12 months, treated as a categorical variable), and the interaction between randomized group and time. To compare the efficacy of 4:3 IMF versus DCR for weight change, contrasts were used to assess the change from baseline to 12 months between the intervention groups. Analyses that included other time points (for example, 3, 6, and 9 months) used the same approach. Several post hoc sensitivity analyses were conducted to determine the robustness of the primary outcome results, including an analysis that removed the cohort that was most affected by the COVID-19 pandemic, 2 models that used multiple imputation (single-chain Markov-chain Monte Carlo and control-based [DCR-based] pattern-mixture model), and 2 models that used the actual number of days weight was measured from baseline (spline and polynomial) (additional details are provided in the Supplement). The 95% CIs were not adjusted for multiple comparisons and should not be used in place of hypothesis testing. Average percentage caloric restriction, percentage of participants achieving at least 5% or at least 10% weight loss, and incidence of attrition were calculated for both groups.
Role of the Funding Source
This work was supported by grants from the National Institutes of Health (R01 DK111622, R01 DK111622-02S1, P30 DK048520, UL1 TR002535, and F32 DK122652), which had no role in the study design; collection, management, analysis, or interpretation of the data; writing of the manuscript; or the decision to submit the manuscript for publication.
Results
Trial Participants
Of the 549 participants screened, 165 were randomly assigned (4:3 IMF, n = 84; DCR, n = 81) and 125 (76% of those who were randomly assigned) completed the trial (Figure 1). The incidence of attrition at 12 months was 19.0% in the 4:3 IMF group and 29.6% in the DCR group. Sample sizes for each variable by time point are shown in Supplement Table 1 (available at Annals.org). Baseline characteristics for both groups are shown in Table 1. There were more participants of Hispanic/Latino ethnicity in the DCR group than the 4:3 IMF group. Participants who did not complete the study were generally younger and/or more likely to be of Hispanic/Latino ethnicity compared with those who completed the study.
Figure 1.
Study flow diagram.
4:3 IMF = intermittent fasting with 3 nonconsecutive fast days per week; DCR = daily caloric restriction; ITT = intention-to-treat.
Table 1.
Baseline Characteristics of the Study Population*
| Characteristic | 4:3 IMF Group (n = 84) | DCR Group (n = 81) | All Randomly Assigned Participants (n = 165) | Participants Who Completed the Study (n = 125) | Participants Who Did Not Complete the Study (n = 40) |
|---|---|---|---|---|---|
|
| |||||
| Mean age (SD), y | 42 (10) | 42 (8) | 42 (9) | 43 (9) | 40 (8) |
| Mean anthropometric measures (SD) | |||||
| Weight, kg | 99.2 (16.0) | 95.5 (16.0) | 97.4 (16.1) | 97.1 (15.8) | 98.3 (17.1) |
| BMI, kg/m2 | 34.3 (4.4) | 33.9 (4.4) | 34.1 (4.4) | 33.9 (4.3) | 34.6 (4.7) |
| Waist circumference, cm | 108.7 (11.4) | 107.9 (10.9) | 108.3 (11.1) | 108.0 (10.8) | 109.2 (12.3) |
| Total body fat mass, % | 40.6 (7.8) | 41.1 (7.2) | 40.8 (7.5) | 40.6 (7.7) | 41.3 (6.9) |
| Systolic blood pressure, mm Hg | 120 (11) | 118 (12) | 119 (12) | 119 (12) | 118 (11) |
| Diastolic blood pressure, mm Hg | 75 (7) | 76 (9) | 75 (8) | 76 (8) | 74 (8) |
| Sex, % | |||||
| Female | 73.8 | 74.1 | 73.9 | 73.6 | 75.0 |
| Male | 26.2 | 25.9 | 26.1 | 26.4 | 25.0 |
| Race/ethnicity, % | |||||
| White | 85.7 | 86.4 | 86.1 | 84.8 | 90.0 |
| Asian | 7.1 | 2.5 | 4.8 | 5.6 | 2.5 |
| Black | 6.0 | 6.2 | 6.1 | 5.6 | 7.5 |
| Other | 1.2 | 4.9 | 3.0 | 4.0 | 0.0 |
| Ethnicity, % | |||||
| Hispanic or Latino | 17.9 | 29.6 | 23.6 | 16.8 | 45.0 |
| Not Hispanic or Latino | 82.1 | 70.4 | 76.4 | 83.2 | 55.0 |
| Education, % | |||||
| High school diploma | 2.5 | 6.2 | 4.2 | 3.2 | 7.5 |
| 1–3 y of college | 17.9 | 16.0 | 17.0 | 10.4 | 37.5 |
| College degree | 41.7 | 35.8 | 38.8 | 42.4 | 27.5 |
| Master's degree | 25.0 | 30.9 | 27.9 | 30.4 | 20.0 |
| Doctoral degree | 13.1 | 11.1 | 12.1 | 13.6 | 7.5 |
| Annual household income, % | |||||
| <$25 000 | 4.8 | 3.7 | 4.2 | 3.2 | 7.5 |
| $25 000-$45 000 | 11.9 | 9.9 | 10.9 | 10.4 | 12.5 |
| $45 001-$70 000 | 19.0 | 17.3 | 18.2 | 17.6 | 20.0 |
| $70 001-$110 000 | 15.5 | 27.2 | 21.2 | 20.8 | 22.5 |
| >$110 000 | 47.6 | 38.3 | 43.0 | 46.4 | 32.5 |
| Preferred not to answer | 1.2 | 3.7 | 2.4 | 1.6 | 5.0 |
| Marital status, % | |||||
| Single | 19 | 12.3 | 15.8 | 16.8 | 12.5 |
| Divorced | 6 | 9.9 | 7.9 | 6.4 | 12.5 |
| Committed relationship | 13.1 | 9.9 | 11.5 | 7.2 | 25.0 |
| Married | 60.7 | 67.9 | 64.2 | 69.6 | 47.5 |
| Widowed | 1.2 | 0 | 0.6 | 0.0 | 2.5 |
| Mean behavioral measures (SD) | |||||
| Self-reported energy intake, kcal/d | 1627 (475) | 1544 (397) | 1585 (437) | 1603 (433) | 1510 (456) |
| Total MVPA, min/d | 22 (12) | 22 (15) | 22 (14) | 22 (14) | 21 (13) |
| Mean metabolic/hormonal levels (SD) | |||||
| Total cholesterol, mg/dL | 179.7 (34.8) | 182.1 (32.6) | 180.9 (33.6) | 182.8 (34.7) | 174.9 (29.4) |
| LDL cholesterol, mg/dL | 106.8 (29.9) | 108.0 (28.8) | 107.4 (29.3) | 108.6 (30.0) | 103.3 (26.8) |
| HDL cholesterol, mg/dL | 45.1 (10.0) | 48.2 (11.9) | 46.6 (11.1) | 47.0 (11.5) | 45.3 (9.5) |
| Triglycerides, mg/dL | 137.4 (69.6) | 127.8 (78.1) | 132.7 (73.9) | 133.7 (76.2) | 129.3 (66.6) |
| Glucose, mg/dL | 93.7 (9.2) | 93.8 (9.4) | 93.7 (9.3) | 93.4 (8.7) | 94.9 (10.9) |
| Insulin, pmol/L | 84.7 (48.6) | 72.9 (41.7) | 79.2 (45.8) | 77.8 (41.7) | 81.9 (55.6) |
| HOMA-IR | 2.84 (1.85) | 2.45 (1.51) | 2.6 (1.7) | 2.59 (1.51) | 2.84 (2.20) |
| Hemoglobin A1c, % | 5.5 (0.3) | 5.5 (0.4) | 5.5 (0.3) | 5.5 (0.3) | 5.5 (0.4) |
4:3 IMF = intermittent fasting with 3 nonconsecutive fast days per week; BMI = body mass index; DCR = daily caloric restriction; HDL = high-density lipoprotein; HOMA-IR = homeostatic model assessment for insulin resistance; LDL = low-density lipoprotein; MVPA = moderate to vigorous physical activity.
Variables were analyzed using descriptive statistics. To convert cholesterol values to millimoles per liter, multiply by 0.0259. To convert triglyceride values to millimoles per liter, multiply by 0.0113. To convert glucose values to millimoles per liter, multiply by 0.0555.
Body Weight and Composition
For our primary end point, 4:3 IMF showed significantly greater weight loss than DCR (−7.7 kg [95% CI, −9.6 to −5.9 kg] vs. −4.8 kg [CI, −6.8 to −2.8 kg]; P = 0.040) at 12 months (Figure 2). Results from all sensitivity analyses showed consistent results (Supplement Figures 1 and 2 and Supplement Tables 2 to 4, available at Annals.org). At 12 months, the percentage change in body weight from baseline was −7.6% (CI, −9.5% to −5.7%) in the 4:3 IMF group and −5.0% (CI, −6.9% to −3.1%) in the DCR group. Fifty-eight percent (n = 50) of participants in the 4:3 IMF group achieved weight loss of at least 5% at 12 months versus 47% (n = 27) of DCR participants. Thirty-eight percent (n = 26) of 4:3 IMF participants achieved weight loss of at least 10% at 12 months versus 16% (n = 9) of DCR participants. Changes in body composition, BMI, and waist circumference also tended to favor 4:3 IMF, although estimates were imprecise (Figure 2; Supplement Table 5, available at Annals.org).
Figure 2.
Changes in absolute body weight (A), waist circumference (B), fat mass (C), and lean mass (D) from baseline, by randomized group.
Results are from a linear mixed-effects model with unstructured covariance using an intention-to-treat analysis. Error bars indicate 95% CIs. 4:3 IMF = intermittent fasting with 3 nonconsecutive fast days per week; DCR = daily caloric restriction.
Cardiometabolic Outcomes
At month 12, point estimates of change in systolic blood pressure, total and low-density lipoprotein cholesterol levels, triglyceride level, HOMA-IR, fasting glucose level, and hemoglobin A1c level favored 4:3 IMF, whereas point estimates of change in diastolic blood pressure and high-density lipoprotein cholesterol level favored DCR (Table 2; Supplement Table 5), although most estimates were imprecise and inconclusive.
Table 2.
Changes in Anthropometric, Cardiometabolic, and Behavioral Parameters, by Randomized Group*
| Variable | Month | 4:3 IMF Group | Change From Baseline in 4:3 IMF Group | DCR Group | Change From Baseline in DCR Group | Difference in Change From Baseline Between Groups |
|---|---|---|---|---|---|---|
|
| ||||||
| Weight, kg | 0 | 99.2 (95.8 to 102.6) | −7.7 (−9.6 to −5.9) | 95.5 (92.0 to 99.0) | −4.8 (−6.8 to −2.8) | −2.89 (−5.65 to −0.14) |
| 12 | 91.5 (87.9 to 95.0) | 90.7 (87.0 to 94.4) | ||||
| BMI, kg/m2 | 0 | 34.26(33.32 to 35.21) | −2.67 (−3.32 to −2.01) | 33.90 (32.94 to 34.86) | −1.67 (−2.38 to −0.96) | −1.00 (−1.96 to −0.03) |
| 12 | 31.60 (30.52 to 32.65) | 32.23 (31.13 to 33.34) | ||||
| Waist circumference, cm | 0 | 108.7 (106.3 to 111.1) | −8.6 (−10.7 to −6.6) | 107.9 (105.5 to 110.4) | −6.2 (−8.4 to −3.9) | −2.5 (−5.5 to 0.6) |
| 12 | 100.0 (97.3 to 102.7) | 101.8 (98.9 to 104.6) | ||||
| Fat mass, kg | 0 | 39.7 (37.5 to 42.0) | −6.0 (−7.4 to −4.5) | 38.6 (36.3 to 40.9) | −3.9 (−5.0 to −2.0) | −2.1 (−4.3 to 0.0) |
| 12 | 33.8 (31.2 to 36.3) | 34.7 (32.1 to 37.3) | ||||
| Lean mass, kg | 0 | 55.2 (52.9 to 57.6) | −1.8 (−2.5 to −1.2) | 52.7 (50.3 to 55.1) | −0.6 (−1.3 to 0.1) | −1.2 (−2.2 to −0.3) |
| 12 | 53.4 (51.0 to 55.8) | 52.0 (49.6 to 54.5) | ||||
| Systolic blood pressure, mm Hg | 0 | 119.65(117.17 to 122.14) | −2.26 (−5.23 to 0.72) | 118.06 (115.53 to 120.58) | −1.38 (−4.55 to 1.79) | −0.87 (−5.22 to 3.47) |
| 12 | 117.40(114.82 to 119.97) | 116.67 (113.88 to 119.46) | ||||
| Diastolic blood pressure, mm Hg | 0 | 75.16 (73.46 to 76.86) | −0.20 (−2.32 to 1.93) | 75.73 (74.00 to 77.46) | −1.50 (−3.77 to 0.77) | 1.30 (−1.81 to 4.41) |
| 12 | 74.96 (73.11 to 76.81) | 74.23 (72.22 to 76.23) | ||||
| Total cholesterol level, mg/dL | 0 | 179.70 (172.40 to 187.00) | −2.07 (−8.31 to 4.18) | 182.10 (174.71 to 189.49) | 0.62 (−6.35 to 7.58) | −2.68 (−12.04 to 6.67) |
| 12 | 177.63(169.76 to 185.50) | 182.72 (174.24 to 191.20) | ||||
| HDL cholesterol level, mg/dL | 0 | 45.06 (42.68 to 47.44) | 2.69 (10.3 to 4.34) | 48.17 (45.76 to 50.58) | 2.77(0.93 to 4.61) | −0.08 (−2.56 to 2.39) |
| 12 | 47.75 (45.36 to 50.13) | 50.94 (48.41 to 53.48) | ||||
| LDL cholesterol level, mg/dL | 0 | 106.76(100.40 to 113.12) | 0.06 (−5.75 to 5.87) | 107.97 (101.53 to 114.41) | 0.77 (−5.73 to 7.26) | −0.71 (−9.42 to 8.00) |
| 12 | 106.82(99.63 to 114.01) | 108.74 (100.96 to 116.52) | ||||
| Triglyceride level, mg/dL | 0 | 137.42(121.40 to 153.44) | −23.98 (−37.91 to −10.04) | 127.81 (111.60 to 144.03) | −14.97 (−30.24 to 0.30) | −9.01 (−29.68 to 11.66) |
| 12 | 113.45(99.89 to 128.01) | 112.85 (97.00 to 128.70) | ||||
| HOMA-IR | 0 | 2.84 (2.48 to 3.21) | −0.95 (−1.31 to −0.58) | 2.45 (2.08 to 2.82) | −0.37 (−0.77 to 0.02) | −0.57 (−1.11 to −0.03) |
| 12 | 1.90(1.55 to 2.24) | 2.07 (1.69 to 2.45) | ||||
| Fasting glucose level, mg/dL | 0 | 93.69 (91.67 to 95.70) | −3.46 (−5.45 to-1.47) | 93.80 (91.77 to 95.84) | −1.52 (−3.70 to 0.67) | −1.95 (−4.90 to 1.01) |
| 12 | 90.23(88.22 to 92.23) | 92.29 (90.08 to 94.49) | ||||
| Hemoglobin A1c level, % | 0 | 5.46 (5.39 to 5.53) | −0.13 (−0.19 to −0.07) | 5.48 (5.41 to 5.55) | −0.08 (−0.14 to −0.01) | −0.05 (−0.14 to 0.04) |
| 12 | 5.33 (5.25 to 5.41) | 5.40 (5.32 to 5.49) | ||||
| Total MVPA, min/wk | 0 | 154(133 to 174) | 50 (23 to 77) | 152(131 to 173) | 65 (34 to 97) | −15 (−56 to 27) |
| 12 | 204(171 to 237) | 217 (180 to 255) | ||||
| Steps, count/d | 0 | 7090 (6600 to 7581) | 1026 (430 to 1623) | 7540 (7036 to 8044) | 1477 (744 to 2176) | −451 (−1370 to 429) |
| 12 | 8116 (7397 to 8835) | 9017(8203 to 9831) | ||||
| Binge Eating Scale score | 0 | 11.68 (10.35 to 13.01) | −2.23 (−3.46 to-1.00) | 12.70 (11.35 to 14.06) | −0.25 (−1.64 to 1.13) | −1.98 (−3.83 to −0.12) |
| 12 | 9.45 (7.84 to 11.06) | 12.45 (10.71 to 14.19) | ||||
4:3 IMF = intermittent fasting with 3 nonconsecutive fast days per week; BMI = body mass index; DCR = daily caloric restriction; HDL = high-density lipoprotein; HOMA-IR = homeostatic model assessment for insulin resistance; LDL = low-density lipoprotein; MVPA = moderate to vigorous physical activity.
Results are estimates (95% CIs) from a linear mixed-effects model with time (0 or 12 months), randomized group (4:3 IMF or DCR), and the interaction between time and randomized group as dependent variables, with unstructured covariance using an intention-to-treat analysis. To convert cholesterol values to millimoles per liter, multiply by 0.0259. To convert triglyceride values to millimoles per liter, multiply by 0.0113. To convert glucose values to millimoles per liter, multiply by 0.0555.
Energy Intake and PA Adherence
Participants in the 4:3 IMF group showed greater percentage caloric restriction than those in the DCR group over the initial 6 months of the intervention (−14.25% [CI, −17.21% to −11.28%] vs. −10.20% [CI, −12.67% to −7.74%]) and over the entire 12-month intervention (−9.75% [CI, −12.83% to −6.68%] vs. −5.64% [CI, −8.04% to −3.24%]). However, both groups were below the targeted weekly energy deficit of 34.3%. Similarly, the 4:3 IMF group showed greater reductions in self-reported energy intake than the DCR group (Figure 3). Among 4:3 IMF participants, there was lower self-reported energy intake on fast versus nonfast days (Figure 3). Additional self-reported energy intake data are provided in the Supplement. There were no differences between the 4:3 IMF and DCR groups in changes in percentage of calories from fat (−1.5% [CI, −3.2% to 0.3%] vs. −1.6% [CI, −3.4% to 0.3%]), carbohydrates (−2.7% [CI, −4.8% to 0.5%] vs. −0.7% [CI, −3.0% to 1.6%]), or protein (3.6% [CI, 2.3% to 4.8%] vs. 3.6% [CI, 2.3% to 4.9%]) at 12 months. The 4:3 IMF group showed a greater improvement in Binge Eating Scale scores than the DCR group at 12 months (Table 2; Supplement Table 5). Both groups increased total MVPA and steps, with no between-group differences at 12 months (Table 2; Supplement Table 5).
Figure 3.
Changes in self-reported energy intake and prescribed vs. self-reported energy intake, by randomized group.
Results are from a linear mixed-effects model with unstructured covariance using an intention-to-treat analysis. Error bars indicate 95% CIs. 4:3 IMF = intermittent fasting with 3 nonconsecutive fast days per week; DCR = daily caloric restriction.
Adverse Events
There were no deaths and no protocol-related serious adverse events (Supplement Table 6, available at Annals.org). No participants withdrew from the study due to adverse events.
Discussion
We found that 4:3 IMF produced modestly superior weight loss versus DCR over 12 months in the context of a guidelines-based, high-intensity, comprehensive behavioral weight loss program. Objectively measured percentage caloric restriction was greater in the 4:3 IMF group than the DCR group, and there was no between-group difference in change in total MVPA, suggesting that differences in weight loss may have been due to greater adherence to 4:3 IMF versus DCR. The 4:3 IMF paradigm was well tolerated, and attrition was lower in this group (19.0%) than the DCR group (29.6%).
Alternative patterns of restricting dietary energy intake for weight loss are gaining attention due to the difficulty of adhering to a reduced-calorie diet on a daily basis. An appealing feature of IMF is that dieters do not have to focus on counting calories and restricting intake every day as they do with DCR (21). Furthermore, the periodic nature of fasting may mitigate the constant hunger associated with DCR (5). Although objectively measured adherence to targeted energy deficit (percentage caloric restriction from baseline) was below the target of 34.3% in both groups, 4:3 IMF showed greater percentage caloric restriction over the 12-month intervention compared with DCR. This suggests that, on average, 4:3 IMF may be more sustainable over 1 year than DCR. However, weight loss varied in both groups. Future studies should evaluate biological and behavioral predictors of response to both 4:3 IMF and DCR to provide insight for personalization of dietary recommendations for weight loss.
In contrast to the 2.89 kg greater weight loss observed in the 4:3 IMF group in the current study, prior randomized trials of isocaloric IMF versus DCR have shown no between-group differences in weight loss at 12 months (6–10). One factor that may explain this difference is the IMF dietary paradigm. Most previous long-term studies have evaluated a 5:2 IMF paradigm (6–9); however, paradigms involving 2 fast days per week may be insufficient to enhance weight loss over DCR. Of note, a prior 12-month randomized trial comparing isocaloric ADF versus DCR also found no difference in weight loss at 12 months (10). In that study, ADF alternated between 75% and 125% energy intake on “fast” and “feast” days, respectively, during a 6-month weight loss phase and between 50% and 150% energy intake on “fast” and “feast” days, respectively, during a subsequent 6-month maintenance phase. The diet diary data from that trial suggested that many ADF participants converted to de facto DCR, with little difference in energy intake between “fast” and “feast” days. In contrast, the pattern of energy intake from diet diaries in our trial showed adherence to distinct fast and nonfast days. The flexibility of the 4:3 IMF paradigm (in which participants can select the 3 nonconsecutive days on which to fast each week) may have enhanced dietary adherence and weight loss compared with a more rigid ADF paradigm (10).
Another reason for the difference between our findings and those of previous long-term (≥12 months) studies of IMF versus DCR may relate to the comprehensive behavioral support program. Headland and colleagues provided only initial advice from a dietitian with no ongoing support (8); Carter, Gray, and Sundfør and their respective colleagues provided 7 to 10 individual sessions with a dietitian over 12 months (6, 7, 9); and Trepanowski and colleagues provided no support in months 0 to 3, weekly individual dietary counseling in months 4 to 6, and monthly behavioral support in months 7 to 12 (10). In our study, all participants received a guidelines-based, high-intensity, comprehensive behavioral weight loss program that included a total of 34 group-based sessions focusing on support for the randomized dietary paradigm and increasing PA, which may have enhanced adoption and maintenance of 4:3 IMF compared with prior studies. Furthermore, information on portion control and healthy eating strategies provided to both groups may have led to the observed calorie reduction from baseline on nonfast days in the 4:3 IMF group. Finally, the group-based nature of the support sessions also allowed participants to share strategies to enhance adherence to the 4:3 IMF paradigm, which may have contributed to their success.
High levels of PA have been consistently associated with weight loss maintenance (22–24) and health benefits (15, 25). Thus, the effect of IMF on PA should be characterized before IMF is recommended more broadly. Only 1 prior long-term (≥12 months) study of isocaloric IMF versus DCR provided a PA prescription (10 000 steps per day); however, changes in PA were not measured (8). In other prior IMF studies, no PA prescription was provided (6, 7) or participants were specifically instructed not to change PA (9, 10). Our trial included a PA prescription consistent with current guidelines for weight management (12). Both groups increased total MVPA and steps in response to the PA prescription, suggesting that 4:3 IMF did not limit participants’ ability to increase PA.
Consistent with prior long-term studies of IMF versus DCR (6–10), both groups showed improvements in cardiometabolic risk factors. Although most point estimates favored 4:3 IMF to varying degrees, estimates were imprecise and inconclusive for many end points. These results should be interpreted with caution because the study was not designed to detect small but potentially clinically important differences in cardiometabolic end points. Future studies should compare changes in cardiometabolic parameters between IMF and DCR with a sample size designed to detect clinically meaningful differences.
Both groups showed improvements in binge eating scores, with greater improvements in the 4:3 IMF group than the DCR group. Of note, participants with binge eating disorders were excluded from the study. The 4:3 IMF intervention was safe, with only 4 mild diet-related adverse events over the 12-month intervention.
Our study has several limitations. Results from analyses of secondary outcomes should be interpreted with caution because the study was not powered to detect differences in the ranges observed for these clinical end points. In addition, these results should be viewed as hypothesis generating, given that type I error was not corrected for multiple comparisons. Similar to other 12-month behavioral weight loss intervention trials, our study had a 24% dropout rate. Only 26% of participants were male, and only 6% of participants identified as Black, limiting generalizability of the results to these populations. The generalizability of the results to older adults and adults with diabetes or cardiovascular disease is also limited. Finally, the results are not generalizable outside the context of a high-intensity, comprehensive behavioral weight loss program that may not be accessible to all people with obesity; however, this is the current recommended treatment approach.
Long-term adherence to DCR is challenging for many people. Our results suggest that 4:3 IMF is an alternative dietary weight loss strategy that may produce modestly superior weight loss compared with DCR at 12 months when provided in the context of a high-intensity, comprehensive behavioral weight loss program. Thus, 4:3 IMF should be considered within the range of evidence-based dietary weight loss approaches.
Supplementary Material
Acknowledgment:
The authors thank the DRIFT participants; Candace Baumgartner, Luciana Smith, and the Anschutz Health and Wellness Center Fitness Center; Wendy Kohrt, PhD, Jere Hamilton, and the Colorado NORC Energy Balance Assessment Core Laboratory; Janine Higgins, PhD, Tyson Marden, RD, and the Colorado Clinical Translational Science Institute Nutrition Core; W. Troy Donahoo, PhD; and Kevin Hall, PhD.
Grant Support:
This work was supported by grants from the National Institutes of Health (R01 DK111622, R01 DK111622-02S1, P30 DK048520, UL1 TR002535, and F32 DK122652).
Footnotes
Disclosures: Disclosure forms are available with the article online.
Author contributions are available at Annals.org.
Contributor Information
Victoria A. Catenacci, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, and Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Danielle M. Ostendorf, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, Aurora, Colorado; Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado; Department of Kinesiology, Recreation and Sports Studies, University of Tennessee Knoxville, Knoxville, Tennessee.
Zhaoxing Pan, Department of Biostatistics and Informatics, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Laura K. Kaizer, Department of Biostatistics and Informatics, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Seth A. Creasy, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, and Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Adnin Zaman, Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado; Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Rochester Medical Center, Rochester, New York.
Ann E. Caldwell, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, and Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Jared Dahle, Department of Biology, United States Air Force Academy, Colorado Springs, Colorado.
Bryan Swanson, School of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Matthew J. Breit, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, and Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus; Graduate School, Integrated Physiology Program, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Kristen Bing, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, and Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Liza T. Wayland, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, and Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Shelby L. Panter, Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Jared J. Scorsone, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, Aurora, Colorado
Daniel H. Bessesen, Anschutz Health and Wellness Center, University of Colorado, School of Medicine, Anschutz Medical Campus, and Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Paul MacLean, Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Edward L. Melanson, Division of Endocrinology, Metabolism, and Diabetes, Department of Medicine, University of Colorado Anschutz Medical Campus, and Division of Geriatric Medicine, Department of Medicine, University of Colorado Anschutz Medical Campus, Aurora, Colorado.
Data Sharing Statement:
The consent form allowed participants to opt out of sharing their deidentified data for future research, and not all participants consented to data sharing. Thus, the authors are unable to share a full data set underlying this manuscript. Analytic code for the primary outcome is available to approved persons through written agreements with Dr. Ostendorf (e-mail, dostendo@utk.edu).
References
- 1.Dansinger ML, Gleason JA, Griffith JL, et al. Comparison of the Atkins, Ornish, Weight Watchers, and Zone diets for weight loss and heart disease risk reduction: a randomized trial. JAMA. 2005;293:43–53. doi: 10.1001/jama.293.1.43 [DOI] [PubMed] [Google Scholar]
- 2.Das SK, Gilhooly CH, Golden JK, et al. Long-term effects of 2 energy-restricted diets differing in glycemic load on dietary adherence, body composition, and metabolism in CALERIE: a 1-y randomized controlled trial. Am J Clin Nutr. 2007;85:1023–1030. doi: 10.1093/ajcn/85.4.1023 [DOI] [PubMed] [Google Scholar]
- 3.Middleton KR, Anton SD, Perri MG. Long-term adherence to health behavior change. Am J Lifestyle Med. 2013;7:395–404. doi: 10.1177/1559827613488867 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Anderson JW, Konz EC, Frederich RC, et al. Long-term weight-loss maintenance: a meta-analysis of US studies. Am J Clin Nutr. 2001;74:579–584. doi: 10.1093/ajcn/74.5.579 [DOI] [PubMed] [Google Scholar]
- 5.Horne BD, Muhlestein JB, Anderson JL. Health effects of intermittent fasting: hormesis or harm? A systematic review. Am J Clin Nutr. 2015;102:464–470. doi: 10.3945/ajcn.115.109553 [DOI] [PubMed] [Google Scholar]
- 6.Carter S, Clifton PM, Keogh JB. Effect of intermittent compared with continuous energy restricted diet on glycemic control in patients with type 2 diabetes: a randomized noninferiority trial. JAMA Netw Open. 2018;1:e180756. doi: 10.1001/jamanetworkopen.2018.0756 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Gray KL, Clifton PM, Keogh JB. The effect of intermittent energy restriction on weight loss and diabetes risk markers in women with a history of gestational diabetes: a 12-month randomized control trial. Am J Clin Nutr. 2021;114:794–803. doi: 10.1093/ajcn/nqab058 [DOI] [PubMed] [Google Scholar]
- 8.Headland ML, Clifton PM, Keogh JB. Effect of intermittent compared to continuous energy restriction on weight loss and weight maintenance after 12 months in healthy overweight or obese adults. Int J Obes (Lond). 2019;43:2028–2036. doi: 10.1038/s41366-018-0247-2 [DOI] [PubMed] [Google Scholar]
- 9.Sundfør T, Svendsen M, Tonstad S. Effect of intermittent versus continuous energy restriction on weight loss, maintenance and cardiometabolic risk: a randomized 1-year trial. Nutr Metab Cardiovasc Dis. 2018;28:698–706. doi: 10.1016/j.numecd.2018.03.009 [DOI] [PubMed] [Google Scholar]
- 10.Trepanowski JF, Kroeger CM, Barnosky A, et al. Effect of alternate-day fasting on weight loss, weight maintenance, and cardioprotection among metabolically healthy obese adults: a randomized clinical trial. JAMA Intern Med. 2017;177:930–938. doi: 10.1001/jamainternmed.2017.0936 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 11.Ezzati A, Rosenkranz SK, Phelan J, et al. The effects of isocaloric intermittent fasting vs daily caloric restriction on weight loss and metabolic risk factors for noncommunicable chronic diseases: a systematic review of randomized controlled or comparative trials. J Acad Nutr Diet. 2023;123:318–329.e1. doi: 10.1016/j.jand.2022.09.013 [DOI] [PubMed] [Google Scholar]
- 12.Jensen MD, Ryan DH, Apovian CM, et al. ; Obesity Society. 2013 AHA/ACC/TOS guideline for the management of overweight and obesity in adults: a report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines and The Obesity Society. J Am Coll Cardiol. 2014;63:2985–3023. doi: 10.1016/j.jacc.2013.11.004 [DOI] [PubMed] [Google Scholar]
- 13.Ostendorf DM, Caldwell AE, Zaman A, et al. Comparison of weight loss induced by daily caloric restriction versus intermittent fasting (DRIFT) in individuals with obesity: study protocol for a 52-week randomized clinical trial. Trials. 2022;23:718. doi: 10.1186/s13063-022-06523-2 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Food and Agriculture Organization of the United Nations. Human Energy Requirements: Report of a Joint FAO/WHO/UNU Expert Consultation. Report no. 0379-5721. 26 March 2005. [Google Scholar]
- 15.Donnelly JE, Blair SN, Jakicic JM, et al. ; American College of Sports Medicine. American College of Sports Medicine Position Stand. Appropriate physical activity intervention strategies for weight loss and prevention of weight regain for adults. Med Sci Sports Exerc. 2009;41:459–471. doi: 10.1249/MSS.0b013e3181949333 [DOI] [PubMed] [Google Scholar]
- 16.U.S. Department of Health and Human Services. Physical Activity Guidelines for Americans. 2nd edition. 2018.
- 17.Peters JC, Wyatt HR, Foster GD, et al. The effects of water and non-nutritive sweetened beverages on weight loss during a 12-week weight loss treatment program. Obesity (Silver Spring). 2014;22:1415–1421. doi: 10.1002/oby.20737 [DOI] [PubMed] [Google Scholar]
- 18.Wyatt HR, Jortberg BT, Babbel C, et al. Weight loss in a community initiative that promotes decreased energy intake and increased physical activity and dairy consumption: Calcium Weighs-In. J Phys Act Health. 2008;5:28–44. doi: 10.1123/jpah.5.1.28 [DOI] [PubMed] [Google Scholar]
- 19.Racette SB, Das SK, Bhapkar M, et al. ; CALERIE Study Group. Approaches for quantifying energy intake and %calorie restriction during calorie restriction interventions in humans: the multicenter CALERIE study. Am J Physiol Endocrinol Metab. 2012;302:E441–E448. doi: 10.1152/ajpendo.00290.2011 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 20.Duarte C, Pinto-Gouveia J, Ferreira C. Expanding binge eating assessment: validity and screening value of the Binge Eating Scale in women from the general population. Eat Behav. 2015;18:41–47. doi: 10.1016/j.eatbeh.2015.03.007 [DOI] [PubMed] [Google Scholar]
- 21.Johnstone A Fasting for weight loss: an effective strategy or latest dieting trend? Int J Obes (Lond). 2015;39:727–733. doi: 10.1038/ijo.2014.214 [DOI] [PubMed] [Google Scholar]
- 22.Tate DF, Jeffery RW, Sherwood NE, et al. Long-term weight losses associated with prescription of higher physical activity goals. Are higher levels of physical activity protective against weight regain? Am J Clin Nutr. 2007;85:954–959. doi: 10.1093/ajcn/85.4.954 [DOI] [PubMed] [Google Scholar]
- 23.Jakicic JM, Marcus BH, Lang W, et al. Effect of exercise on 24-month weight loss maintenance in overweight women. Arch Intern Med. 2008;168:1550–1559. doi: 10.1001/archinte.168.14.1550 [DOI] [PMC free article] [PubMed] [Google Scholar]
- 24.Jakicic JM, Winters C, Lang W, et al. Effects of intermittent exercise and use of home exercise equipment on adherence, weight loss, and fitness in overweight women: a randomized trial. JAMA. 1999;282:1554–1560. doi: 10.1001/jama.282.16.1554 [DOI] [PubMed] [Google Scholar]
- 25.Fogelholm M Physical activity, fitness and fatness: relations to mortality, morbidity and disease risk factors. A systematic review. Obes Rev. 2010;11:202–221. doi: 10.1111/j.1467-789X.2009.00653.x [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.
Supplementary Materials
Data Availability Statement
The consent form allowed participants to opt out of sharing their deidentified data for future research, and not all participants consented to data sharing. Thus, the authors are unable to share a full data set underlying this manuscript. Analytic code for the primary outcome is available to approved persons through written agreements with Dr. Ostendorf (e-mail, dostendo@utk.edu).