Refeeding Properly After Fasting

What is behind it

During fasting, the body switches to fatty acids and ketone bodies, insulin falls, and the intracellular stores of phosphate, potassium, magnesium and thiamine are depleted, without this necessarily showing up in blood values – the body keeps serum levels stable at the expense of the cells. As soon as carbohydrates arrive again, insulin rises abruptly and drives glucose into the cells together with phosphate, potassium and magnesium. Serum levels fall within hours to days. That is exactly the definition of refeeding syndrome.

At the same time, the need for thiamine rises because it is consumed as a cofactor in carbohydrate metabolism – which is why thiamine deficiency is part of the definition of the severe form. And the metabolism that the first meal meets is different from before the fast: after 60 hours of fasting there is acute insulin resistance, measured in an intervention study in premenopausal women with 48 hours of refeeding afterwards.

Why tolerance is a topic of its own

Prolonged fasting changes the intestinal mucosa histologically and ultrastructurally, and with it the absorptive capacity, and it changes the hormonal makeup of the enteroendocrine cells. In a prospective controlled observational study of 30 patients who had covered less than 50% of their requirements for at least one month, tissue expression of chromogranin A, GLP-1 and GIP in duodenal biopsies was significantly higher than in controls – chromogranin A 1.04% versus 0.41% stained area. This explains why the first meal after several days of fasting is received differently. It does not explain which foods are suitable in which order – there is no study on that.

How rare the syndrome really is

True refeeding syndrome is rare; refeeding electrolyte shifts are common. A systematic review of 28 studies with 10,412 intensive care patients found a reported incidence between 0% and 88% and a pooled incidence of 23% (95% CI 15 to 33) with a heterogeneity of I² = 99% – which in practice means that the studies did not measure the same thing. No consistent association with mortality was found (OR 1.51; 95% CI 0.79 to 2.88), and the certainty of evidence was low.

Even in the clinic, the risk criteria predict the event poorly. In a retrospective cohort of 86 adults with intestinal failure, 50% of those without risk criteria and 54.4% of those with risk criteria developed refeeding syndrome; the sensitivity of the ASPEN criteria was 0.8, the specificity 0.23.

What the studies show

ASPEN consensus: the verifiable definition

The authoritative consensus paper of the American Society for Parenteral and Enteral Nutrition establishes diagnostic criteria, risk stratification and approach. Significant risk exists with a single criterion – BMI below 16 kg/m², weight loss of 7.5% or more in 3 months or more than 10% in 6 months, negligible intake for more than 7 days, or less than 50% of requirements for more than 5 days in acute illness. Important for context: these recommendations are formulated for clinically malnourished patients, not for healthy people after voluntary fasting.

NICE CG32: the more conservative guideline

High risk exists with one strong criterion – BMI below 16 kg/m², weight loss of more than 15% in 3 to 6 months, little or no food intake for more than 10 days, or low potassium, phosphate or magnesium levels before starting – or with two weaker ones: BMI below 18.5 kg/m², weight loss of more than 10% in 3 to 6 months, little or no food for more than 5 days, a history of alcohol or drug use, or insulin, chemotherapy, antacids or diuretics. For voluntary fasting, the line with the 5 days is decisive – on its own it does not yet count.

The newest consensus corrects the direction

The consensus of the Australasian Society of Parenteral and Enteral Nutrition states: actual refeeding syndrome is rare, but all patients should be screened for the risk. It should only be diagnosed if at least 50% of the estimated requirement has been provided and electrolyte disturbances along with symptoms appeared only afterwards. It explicitly says there is no evidence for starting lower than is already recommended, and that target feeding rates should be reached within 24 to 72 hours.

The benefit only shows after the fast

In a randomized controlled trial with a five-day modified fasting program at home – about 600 kcal per day, 64 participants, 32 of them in the intervention group – the change in blood pressure at the end of the fasting phase was not significant. Blood pressure fell significantly only after food was reintroduced, both systolic and diastolic. This supports the core thesis of this page, albeit for blood pressure and not for tolerance.

What the studies do not deliver

There is no controlled study on how a healthy person should best start eating again after voluntary fasting. All recommendations on calorie amounts, on the order of foods and on waiting times come either from clinical guidelines for malnourished patients or from the protocols of individual fasting clinics; no randomized trial has compared two ways of breaking a fast. This applies explicitly to the widespread rules: neither starting with bone broth, soft egg yolk, fermented vegetables or avocado nor the 30 to 60 minutes of waiting time can be traced back to a study – these are values from experience. The transferability of the risk criteria is also limited, because both systems are tailored to people with pre-existing illness and malnutrition. And the recommendation not to eat large amounts of carbohydrates directly after fasting is physiologically plausible – the acute insulin resistance after 60 hours of fasting has been measured – but the claimed link to circulatory problems has never been studied in healthy people.

How to do it

What the programs actually did: after 4 to 21 days of Buchinger fasting, the reintroduction of food lasted 4 days on average with organic ovo-lacto-vegetarian food, increased from 800 to 1,600 kcal per day – the best-documented implementation, in more than 1,400 people. After pure water fasting for a median of 11 days, refeeding ran for a median of 5 days, range 3 to 17 days, as inpatients, with predominantly mild and transient adverse events. After 14 days of water fasting came 6 days of food reintroduction with a wholefood plant-based diet and a 45-day maintenance phase; in this phase lean mass was rebuilt while fat mass continued to fall. Clinically, in high-risk cases, NICE specifies 10 kcal/kg per day at the start with full requirements after 4 to 7 days, and ASPEN 10 to 20 kcal/kg in the first 24 hours with an increase of 33% every 1 to 2 days and electrolyte checks every 12 hours over 3 days. For healthy people after one to three days of fasting, a single robust line can be drawn from this: several days of build-up instead of one large meal.

Safety

Breaking the fast is the moment when the documented harm occurs. Refeeding syndrome mainly affects people with a BMI below 16 to 18.5 kg/m², with prior unintentional weight loss of more than 10 to 15%, with fasting periods of more than 5 to 10 days or with already low potassium, phosphate or magnesium; a history of alcohol or drug use as well as insulin, chemotherapy, antacids and diuretics count as additional risk factors. Anyone with one of these characteristics should break the fast under medical supervision – with electrolyte checks, not with dietary advice. The second pathway is thiamine deficiency: a 36-year-old man with no history of alcohol use developed Wernicke encephalopathy after 40 days of pure water fasting, with double vision, unsteady gait, confusion and bilateral abducens palsy; under high-dose intravenous thiamine and electrolyte correction, his condition improved markedly. Separate from this is refeeding edema, fluid retention after resuming nutrition, which is often mistaken for weight gain and is discussed in connection with hyperinsulinemia and capillary leak. At the same time, the opposite direction applies: underfeeding for too long is also a risk, and the most recent consensus recommends reaching the target rate within 24 to 72 hours – checking blood values is more important than going slowly. Those who should not fast in the first place should not break a fast on their own: underweight people, pregnant and breastfeeding women, children and adolescents, people with a history of eating disorders. In eating disorders, the refeeding risk is real and belongs in clinical hands.

Sources

• da Silva JSV et al., Nutrition in Clinical Practice 2020 (ASPEN)
• NICE Clinical Guideline CG32
• Matthews-Rensch K et al., Nutrition & Dietetics 2025
• Schneider L et al., Scientific Reports 2026
• Solar H et al., Intestinal Failure 2026
• Rossmeislová L et al., Heliyon 2024
• Wilhelmi de Toledo F et al., PLoS ONE 2019
• Grundler F et al., Genome Medicine 2026