Three diet approaches dominate the longevity conversation right now: intermittent fasting, caloric restriction, and ketogenic eating. They're often discussed together as if they're variations of the same idea. They aren't. They work through different biological mechanisms, they produce different effects, and the head-to-head comparisons between them are messier than anyone selling one of them would like you to believe.
Let me pull them apart.
The thing that kicked off the research
In the 1930s, Clive McCay and his colleagues at Cornell noticed something strange. Rats fed a 30-40% reduced-calorie diet — nutritionally complete, just fewer calories — lived roughly 50% longer than control rats eating ad libitum. They also showed slower age-related decline.
This finding has been replicated in yeast, worms, flies, mice, rats, and some primates. It's one of the most robust findings in aging biology. The phenomenon is called caloric restriction (CR), and the field has been trying to understand *why* it works ever since.
The short version: caloric restriction appears to trigger a suite of cellular responses — autophagy (cellular cleanup), reduced mTOR signaling, increased stress resistance, shifts in gene expression — that collectively slow the biological markers of aging.
The question is: do you need to eat 30-40% less of everything, all the time, to get the benefit? Or are there other ways to trigger the same cellular responses?
What caloric restriction actually does biologically
When you consistently consume fewer calories than you need:
**mTOR signaling drops.** mTOR (mammalian target of rapamycin) is a cellular growth signal. High mTOR activity promotes protein synthesis and cellular proliferation; low mTOR activity promotes cellular cleanup and stress resistance. Cancer tends to involve overactive mTOR. Longevity research has tracked mTOR as one of the key mediators.
**IGF-1 drops.** Insulin-like growth factor 1 is another growth-promoting signal, highly sensitive to protein intake and overall caloric load. Lower IGF-1 is associated with longer life across species.
**Autophagy increases.** The cellular machinery that breaks down damaged proteins, dysfunctional mitochondria, and other cellular debris gets activated when nutrients are scarce. This is basically cellular spring cleaning.
**Insulin sensitivity improves.** Chronic caloric surplus impairs insulin signaling over time. Reducing calories, even modestly, reverses that.
**Gene expression shifts.** Sustained CR produces changes in the expression of hundreds of genes involved in stress response, DNA repair, and metabolic efficiency.
The problem with CR as a strategy for humans: living 30-40% under your maintenance calories is miserable, hard to sustain, and for some people medically risky. The compliance rates in long-term human CR studies are not good.
Hence the interest in alternatives.
What fasting does that's slightly different
Intermittent fasting (IF) isn't about eating less — it's about eating less often. The main forms:
**Time-restricted eating:** eating within an 8- or 10-hour window daily. This is the most tolerable form and the most studied in recent years.
**5:2 fasting:** eating normally five days, eating ~500-600 calories on two non-consecutive days.
**Alternate-day fasting:** alternating normal eating days with fasting days.
**Prolonged fasting:** 48-72 hours or longer, done occasionally rather than routinely.
What fasting does biologically overlaps with CR but has some distinct effects:
**Metabolic switching.** After roughly 12-36 hours without food, the liver depletes glycogen stores and starts mobilizing fatty acids. These get oxidized into ketone bodies, which the body (and brain) can use as fuel. This metabolic switch itself appears to trigger cellular cleanup processes.
**Stronger autophagy activation.** Extended fasting triggers autophagy more vigorously than modest daily CR does. Cells do more cleanup when they're actually hungry.
**Hormetic stress response.** Brief fasting stresses cells in ways they can recover from, and the recovery response is where most of the benefit lives. This is the same hormesis principle that makes exercise beneficial.
One of the direct CR-vs-fasting comparisons found that 5:2 IF produced similar weight loss to continuous CR in overweight women, but better insulin sensitivity improvements and more visceral fat loss. Not dramatically different, but not identical either.
What keto does, which is its own thing
Ketogenic diets work differently from the other two. You're not necessarily restricting calories. You're forcing the body into sustained ketosis by cutting carbohydrates dramatically (typically under 50g/day) and keeping protein moderate so the body runs on fat.
This triggers some of the same responses — ketone production, some degree of autophagy — but through a different door.
Ketones themselves have some interesting properties. Beta-hydroxybutyrate, the main circulating ketone, has been shown to act as a signaling molecule. It inhibits certain histone deacetylases (affecting gene expression) and may have direct anti-inflammatory effects. BHB crosses the blood-brain barrier easily and the brain uses it efficiently — this is part of why ketogenic diets have been therapeutic in treatment-resistant epilepsy for nearly a century.
Some of the aging-related claims for keto are extrapolated from the fasting literature (because both produce ketones), which is not quite the same thing as saying keto has been demonstrated to extend lifespan. The direct keto longevity data in humans is thinner than the marketing implies.
What keto does well:
- Weight loss, especially in people who came in insulin-resistant.
- Glycemic control.
- Possibly therapeutic in specific neurological conditions.
- Some reduction of inflammation markers.
What keto doesn't necessarily do:
- Match the breadth of cellular cleanup effects you get from actual fasting.
- Produce longer life than other good dietary patterns in head-to-head human studies (these studies barely exist).
- Work for everyone — a significant fraction of people don't tolerate it well long-term.
A practical note: taking exogenous ketone supplements while eating a normal diet is largely theater. The supplements raise blood ketone levels temporarily, but they don't trigger the underlying metabolic switch, and they actually *block* the body from mobilizing its own fat stores through a negative feedback loop. If you're going to benefit from ketones, they need to come from your own metabolism.
Circadian timing matters more than people think
One finding that keeps showing up across the literature: *when* you eat matters almost as much as what or how much.
Your metabolism isn't static throughout the day. Cortisol peaks in the early morning. Insulin sensitivity is higher earlier in the day and declines as the day goes on. Melatonin rises in the evening and suppresses insulin secretion.
Eating concentrated calories late at night puts food into a metabolically worse time slot. The same meal eaten at noon versus 10 PM produces different glucose and insulin responses.
This is why time-restricted eating works in part as a form of fasting: by compressing your eating window earlier in the day, you align food intake with your biology. A window of 8 AM to 6 PM is very different, metabolically, from a window of 2 PM to midnight, even if the total calories are identical.
What the direct comparisons show
The head-to-head studies between these approaches are fewer and smaller than you might expect. A few patterns that seem reasonably stable:
Weight loss across these approaches is comparable when calories are equated. If you're losing the same amount of weight, you're probably getting similar short-term benefits.
Intermittent fasting appears to have some metabolic advantages beyond simple calorie reduction — slightly better insulin sensitivity, more visceral fat loss.
Keto produces faster early weight loss (largely water) and may be more satiating for some people, making compliance easier.
Over the long term (1-2 years), differences between approaches tend to narrow. What you can *sustain* matters more than the specific approach.
What to actually do
If I were distilling the research into one sentence: *most of the benefit comes from not eating all day, not eating too much overall, and eating mostly in the first two-thirds of your day.*
A reasonable framework, supported across the literature:
Stop eating three hours before bed. This alone addresses most of the circadian misalignment.
Eat within a 10- or 12-hour window to start; you can tighten it later if you tolerate it.
Prioritize protein and produce. These are satiating, nutrient-dense, and support muscle maintenance, which matters more as you age than almost any other dietary factor.
Don't graze. Meals, not constant snacks. The cellular cleanup machinery needs time between fuel deliveries.
Once a year, consider a longer fast (48-72 hours) under sensible conditions. The evidence on the benefits of occasional prolonged fasts is promising. Don't do this if you have any eating-disorder history, medication interactions, or medical conditions that make it unsafe.
Keep the big picture in mind: the dietary pattern that extends your life is one you can actually live with for the next forty years. Perfection is not the goal. Consistency is.
You don't have to pick a camp. Most of the real-world benefits come from principles that all three approaches share: less frequent meals, earlier eating window, less total intake, more whole foods. The branding matters less than the practice.