Intermittent fasting: who benefits, who is harmed

Marina is 44, an architect, mother of two. She read Peter Attia, listened to three podcast episodes about autophagy, decided to do 18:6 — eating between noon and 6 p.m., fasting the rest. For two months, she felt sharp. Doctor, I never had this much mental clarity in the morning.

In the third month, her menstrual cycle was eleven days late. In the fourth, it disappeared. In the fifth, she came to the office with a different complaint: heavy afternoon fatigue, isolated hot flashes, sleep that did not restore. The labs showed FSH in an early menopausal-transition range, flattened morning cortisol, ferritin dropping from 78 to 34 ng/mL.

Marina is the common case. It was not lack of discipline. It was lack of indication.

Fasting is not a diet — it is a tool

Fasting is one of the oldest metabolic interventions humanity knows. It is part of religious traditions, part of the physiology of any mammal between one hunt and the next. What changed in the last twenty years is the volume of scientific evidence cataloging what happens in the human body in windows of 12, 16, 18, 24 hours without food.

The Rafael de Cabo and Mark Mattson review published in the New England Journal of Medicine in 2019 organized that map well. After 12 to 16 hours without caloric intake, hepatic glycogen runs out, and the liver begins to produce ketone bodies from mobilized fat. That “metabolic switch” — passing from glucose to ketone as the dominant fuel — activates signaling cascades linked to autophagy, mitochondrial biogenesis, reduction of low-grade inflammation, and improved insulin sensitivity.

These are real effects. Measurable. But the point that honest literature makes clear: the magnitude and direction of these effects depend on who is fasting.

Where the data show benefit

The most elegant study on the upside of fasting was conducted by Elizabeth Sutton and colleagues at the Pennington Biomedical Research Center, published in Cell Metabolism in 2018. Men with pre-diabetes were randomized to a 6-hour eating window starting early (breakfast at 8 a.m., last meal by 3 p.m.) or to a control 12-hour window. Everything isocaloric — same amount of food, no weight loss.

Result, in five weeks: improved insulin sensitivity, an 11 mmHg reduction in systolic pressure, a fall in oxidative-stress markers, reduction of appetite. All of this without losing a single kilogram. The effect came from the when, not the how much.

That profile — man 35–55, with incipient insulin resistance or pre-diabetes, increased waist circumference, fasting glucose at the limit — is the scenario in which intermittent fasting in an “early” window (early in the day) has consistent evidence of metabolic benefit.

Where the data show little — or harm

When the same protocol leaves the lab and becomes popular fashion, the result changes. The TREAT trial, conducted by Dylan Lowe and colleagues at UCSF and published in JAMA Internal Medicine in 2020, randomized 116 overweight adults to a 16:8 window (eating from noon to 8 p.m.) versus three normal meals for 12 weeks. Average weight loss in the fasting group: 1.17%. In the control: 0.75%. Not a significant difference. And more concerning: the fasting group lost more lean mass than the control.

The late window (lunch–dinner, no breakfast) is not the same thing as the early window. Eating in biological darkness — when your circadian clock expects digestive rest — wipes out much of the metabolic benefit that appears in light-aligned windows.

The most recent trial by Krista Varady and colleagues, published in Annals of Internal Medicine in 2023, compared a noon–8 p.m. window with a classical 25% caloric restriction. Modest weight loss in both groups, no difference in HbA1c, lipids, or pressure. Fasting, in that format, was not worse — but also not better.

And there is the chapter on women in hormonal transition. The literature is still scarce in randomized trials specific to this group, but the physiology is clear: the female hypothalamic-pituitary-gonadal axis is exquisitely sensitive to the signal of energy scarcity. GnRH pulses slow, kisspeptin falls, estradiol and progesterone fall. In women with regular cycles, repeated fasts of 16+ hours over weeks frequently result in menstrual irregularity, drop in libido, worse sleep in the luteal phase. In perimenopausal women — when the system is already oscillating — prolonged fasting tends to amplify the very symptoms the patient wants to attenuate.

Who probably benefits

The clinical reading we make today, based on available evidence:

• Man 35–60 with insulin resistance, fasting glucose 100–125 mg/dL, HbA1c 5.7–6.4%, increased abdominal circumference. Initial 12:12 window, evolving to 14:10 or 16:8 with breakfast preserved (eTRF — early time-restricted feeding). Last meal at 6–7 p.m.
• Adults with mild to moderate non-alcoholic fatty liver disease (NAFLD), preserved hormone function. Same logic — early window.
• People with uncontrolled nighttime snacking habits. Here fasting works less as a metabolic mechanism and more as a behavioral tool — limiting the window where the noise happens.

Who probably is harmed

• Perimenopausal women or women in early menopause. The estradiol oscillation already disorganizes sleep, mood, temperature. Adding a scarcity signal makes it worse.
• Woman with regular cycle and history of eating disorder, even subclinical. A rigid window can reorganize the restrictive pattern.
• Strength or endurance athlete in serious training. Post-workout muscle protein synthesis demands protein distribution every 3–4 hours. A 6-hour window compresses that distribution and reduces the anabolic stimulus.
• Adult >65 with reduced muscle mass. Sarcopenia + fasting = accelerated loss of lean mass. The literature is consistent.
• People with unstable glycemia, reactive hypoglycemia, decompensated anxiety disorder, pregnancy, or breastfeeding. Clear indications to avoid.

What to order before deciding

If the question in the visit is is it worth it for me?, the minimum panel is simple:

• Fasting glucose, fasting insulin, HbA1c (with HOMA-IR calculation).
• Lipid profile with ApoB.
• Morning cortisol (8 a.m., fasting), DHEA-S.
• In women: FSH, LH, estradiol, progesterone — dated relative to the cycle, or random in perimenopause.
• Body composition (DEXA or quality bioimpedance).
• Honest food history, with a 3-day diary.

These data, plus age, sex, life phase, goal, and history, say whether fasting is a tool or noise.

The reading the Continuum makes

Continuum Plenya does not treat fasting as a “healthy lifestyle” item for everyone. The nutritionist enters alongside the physician, and the question is never fasting yes or no. It is which window, at which life moment, with which protein, monitoring which marker. Marina, from the beginning of the story, today eats every four hours, with 30 g of protein per meal, in a loose 12:12 window. The cycle returned, ferritin climbed to 92, morning cortisol came back. I didn't know I had to eat more to feel less.

Intermittent fasting is not virtue. It is a scalpel — and a scalpel without indication harms more than it heals.