Caloric Restriction

Caloric restriction (CR) extends lifespan in yeast, worms, flies, and rodents, sometimes by 30-50%. In humans the picture is smaller and slower to build. The key trial is CALERIE-2, a 24-month, multi-center randomized trial of 218 healthy, non-obese adults run by Kraus and colleagues, with a 25% restriction target that participants actually sustained at close to 12% [1]. That gap between target and delivered restriction is itself a finding: free-living adults without an eating disorder or external enforcement rarely hold a quarter-calorie cut for two years.

What the trial showed: modest but real improvements in cardiometabolic risk markers, including lower LDL cholesterol, blood pressure, and insulin sensitivity, sustained through the full 24 months [1]. A secondary analysis of the CALERIE biobank by Belsky and colleagues found that CR slowed the pace of biological aging as measured by a DNA-methylation clock, one of the few human RCTs to show a measurable slowdown in an aging biomarker from a lifestyle intervention rather than a drug [1].

What the evidence does not show is that CR extends human lifespan. No trial has run long enough, or can ethically run long enough, to measure that endpoint directly. Everything in humans is a biomarker proxy, and biomarker proxies can mislead.

The counter-finding worth naming directly: CR carries a real cost in already-lean people. Villareal and colleagues, publishing in the Journal of Bone and Mineral Research, found that sustained caloric restriction reduced bone mineral density at clinically important fracture sites, specifically hip, in non-obese younger and middle-aged adults [2]. This is not a mechanistic worry, it is a measured effect in a controlled trial. An earlier six-month study found calorie-restricted young adults lost weight and fat without significant bone loss over that shorter window [3], so duration and starting body composition both matter: shorter restriction periods, or restriction started from a body with fat to lose, look safer than years of restriction in someone already lean. CALERIE-2 itself also documented reductions in lean mass alongside fat mass loss, a tradeoff the trial's authors were explicit about, not something buried in a subgroup analysis [4].

Who this does not work for, or where the evidence is weakest: CR is poorly suited to anyone already lean, anyone with a history of disordered eating, older adults at risk of sarcopenia, and pregnant or growing populations, none of whom were represented in CALERIE-2's healthy non-obese adult sample. The trial excluded people with obesity, so CR's effects in people carrying substantial excess fat, plausibly the population most likely to benefit metabolically, are extrapolated from other weight-loss literature rather than shown directly in this dataset. Bone and lean-mass loss make CR a poor fit for frail or older adults without a resistance-training countermeasure. There is also a real-world adherence problem: most people who attempt sustained restriction outside a supervised trial with (as CALERIE-2 had) behavioral counseling and meal-replacement support cannot sustain even the 12% level that CALERIE-2 achieved.

Biological plausibility for slower aging comes from CR's downstream effects on nutrient-sensing pathways, insulin/IGF-1 signaling, and mTOR activity, all named hallmarks-of-aging mechanisms; a mechanistic review situates CALERIE-2's findings within that broader framework rather than treating it as an isolated result [5][6]. The pattern across species, mouse to primate to the CALERIE-2 human data, is consistent enough that most aging researchers treat CR as a real signal, not noise, even without a mortality endpoint [4].

Plain takeaway: moderate CR (roughly 10-15% below maintenance, not the 25% headline target) improves measurable cardiometabolic and biological-aging markers in supervised trials over two years, with the tradeoff of measurable bone and lean-mass loss in people who start out lean. It has never been shown to extend human lifespan, and daily practice outside a trial setting rarely sustains the restriction levels that produced these results.

A proteomics analysis of CALERIE plasma samples, published in Nature Aging (online April 2026, Mishra et al.), measured over 7,000 proteins in samples from participants who sustained an average 14% calorie restriction for 2 years. Complement component C3 and its cleavage product C3a fell significantly. Visceral white adipose tissue macrophages were the main source of the age-related C3 rise, and the C3a decline tracked calorie restriction itself rather than the amount of weight lost. The authors frame complement deactivation as an immunometabolic checkpoint: a measurable mechanism by which moderate restriction damps inflammaging, separable from weight loss.