A 2-Year Randomized Controlled Trial of Human Caloric Restriction: Feasibility and Effects on Predictors of Health Span and Longevity

Abstract

Background: Caloric restriction (CR), energy intake reduced below ad libitum (AL) intake, increases life span in many species. The implications for humans can be clarified by randomized controlled trials of CR.

Methods: To determine CR's feasibility, safety, and effects on predictors of longevity, disease risk factors, and quality of life in nonobese humans aged 21-51 years, 218 persons were randomized to a 2-year intervention designed to achieve 25% CR or to AL diet. Outcomes were change from baseline resting metabolic rate adjusted for weight change ("RMR residual") and core temperature (primary); plasma triiodothyronine (T3) and tumor necrosis factor-α (secondary); and exploratory physiological and psychological measures.

Results: Body mass index averaged 25.1 (range: 21.9-28.0 kg/m(2)). Eighty-two percent of CR and 95% of AL participants completed the protocol. The CR group achieved 11.7±0.7 %CR (mean ± standard error) and maintained 10.4±0.4% weight loss. Weight change in AL was negligible. RMR residual decreased significantly more in CR than AL at 12 months (p = .04) but not 24 months (M24). Core temperature change differed little between groups. T3 decreased more in CR at M12 and M24 (p < .001), while tumor necrosis factor-α decreased significantly more only at M24 (p = .02). CR had larger decreases in cardiometabolic risk factors and in daily energy expenditure adjusted for weight change, without adverse effects on quality of life.

Conclusions: Sustained CR is feasible in nonobese humans. The effects of the achieved CR on correlates of human survival and disease risk factors suggest potential benefits for aging-related outcomes that could be elucidated by further human studies.

Trial registration: ClinicalTrials.gov NCT00427193.

Keywords: Biomarkers; Caloric restriction; Metabolism; Nutrition; Risk factors.

Figure 1.

CONSORT diagram. After an extensive screening process (10), 238 individuals were eligible and 220 individuals were randomized. Two individuals, both assigned to the CR group, dropped out prior to starting the intervention, resulting in an ITT cohort of 218 participants; 75 in the AL control and 143 in the CR group (Table 1). Thirty participants were withdrawn or dropped from the intervention prior to completion including 4 (5.3%) in the AL control group and 26 (18.2%) in the CR group (p = .01). Three CR participants nevertheless continued their evaluations, and following ITT principles were included in all analyses. The 27 participants who failed to provide complete data tended to be younger and better educated (p = .01 for both); otherwise, there were no significant differences with respect to other demographic variables, body composition, dietary intake, metabolic parameters, core temperature, markers of inflammation, and safety parameters at baseline (data not shown). Across the time points, there were observations from 211 participants at month 6, 201 at month 12, 193 at month 18, and 191 at month 24. AL = ad libitum; CR = caloric restriction; ITT = intention-to-treat.

Figure 2.

Caloric restriction (A, only in the CR group) and weight change (B) over the course of the intervention in ad libitum (---) and CR (^___) groups (values are means ± standard error). Bars in A indicate mean values over the designated intervals. Points in B indicate values at individual time points. %CR in the AL group was measured for the intervals BL-M12 (1.3±1.1%) and M12-M24 (0.4±1.1%), p < .001 vs CR. AL = ad libitum; BL = baseline; CR = caloric restriction.

Figure 3.

Changes in resting metabolic rate (RMR) (Panel A) and total daily energy expenditure (TDEE) (Panel B) not attributable to changes in fat-free mass and fat mass at month 12 and month 24 in the AL group (black bars) and the CR group (gray bars). “Measured – Predicted” refers to the difference between measured values and the values predicted by our regression model based on baseline relationships of fat-free mass and fat mass to RMR and TDEE (often called residuals; see Methods section). Values are adjusted means ± standard error. p values refer to differences in change over time between the AL and CR groups. AL = ad libitum; CR = caloric restriction.

Figure 4.

Changes in total cholesterol (A), mean triglycerides (B), HOMA-IR (C), and mean blood pressure (D) at month 12 and month 24 in the AL control (black bars) and CR (gray bars) groups in the ITT analysis. Values are adjusted means ± standard error. p values refer to comparisons of changes over time between the AL control and CR groups. AL = ad libitum; CR = caloric restriction; HOMA-IR = homeostasis model assessment-estimated insulin resistance; ITT = intention-to-treat.