Fasting-Mimicking Diet Side Effects and Tolerability

This page is a companion to the fasting-mimicking diet overview and the FMD research page. It does not add a new citation: every study below is already cited on those two pages. What is different here is the organizing question. Those pages are built around whether FMD works for a given outcome. This page gathers the same trials' own reporting on dropout and adverse events, by population and medication context, because that question has not had its own section anywhere on this site until now.

This page covers fasting-mimicking diet research in people without cancer or multiple sclerosis; trials of the diet alongside treatment for those conditions are a separate area and are not summarized here.

What Does an FMD Cycle Actually Feel Like?

None of the trials cited on this page was designed to measure how an FMD cycle feels as its main outcome; they were designed to measure weight, blood sugar, kidney markers, or skin, and recorded side effects as a secondary safety outcome. Four of them did record adverse events systematically on the Common Terminology Criteria for Adverse Events scale (1 mild to 5 fatal). The symptoms that recur across those reports are fatigue, weakness, headache, dizziness, muscle pain, and nausea during the 5-day cycle, graded mild or moderate (Brandhorst et al., 2015 [1]; Wei et al., 2017 [2]; van den Burg et al., 2024 [3]; Sulaj et al., 2022 [5]); the two trials that reported severity counts recorded nothing of grade 3 or higher (Wei et al., 2017 [2]; Sulaj et al., 2022 [5]). The 12-month trial notes these resolved in the weeks between cycles (van den Burg et al., 2024) [3], and the pilot found them more common after the first cycle than after the second or third (Brandhorst et al., 2015) [1]. Hunger itself was not reported as a graded outcome in any of these trials.

Dropout and Non-Completion Rates, Trial by Trial

A small pilot trial, the human portion of a larger mouse and yeast study, analysed 38 generally healthy adults: 19 who completed three 5-day FMD cycles and 19 controls. The paper reports that 5% of subjects were disqualified for non-compliance with the diet and that 14% of those enrolled withdrew for reasons unrelated to the diet, such as work and travel scheduling; it does not state the enrolled total those percentages are drawn from (Brandhorst et al., 2015) [1].

A larger trial randomized 100 generally healthy adults aged 20 to 70 to three monthly 5-day cycles or their usual diet. In the FMD arm, 13 of 52 participants (25%) were excluded or withdrew: six for scheduling conflicts, five for unspecified personal reasons, and two excluded for not following the diet. In the control arm, 5 of 48 (10%) withdrew (Wei et al., 2017) [2].

The longest trial in this set followed 100 people with type 2 diabetes, managed with metformin alone or diet alone, for 12 months. Eight of 51 FMD participants and 10 of 49 control participants were lost to follow-up, which the authors attribute to the inability to complete study visits rather than to the treatment; a further 13 FMD participants stopped the diet at some point but completed their follow-up visits. This trial excluded people on insulin or a sulfonylurea, so it carries no dropout or adverse-event data for them (van den Burg et al., 2024) [3].

In the 3-month trial in 56 South-Asian Surinamese adults with type 2 diabetes, 13 of the 19 assigned to the FMD completed the three cycles and the follow-up visit; two withdrew before or at the baseline visit (pregnancy; anxiety about blood draws), three discontinued for adverse events or non-adherence during or after the first cycle, and one was withdrawn by the investigators after a serious adverse event (van der Velden et al., 2024) [4]. In the 6-month trial in 40 people with type 2 diabetes and diabetic kidney disease, 38 completed three cycles, 33 completed the six months, and 31 were seen at the 3-month follow-up, with pandemic restrictions given as the main reason for withdrawal (Sulaj et al., 2022) [5]. In the 71-day trial in 45 healthy women, 22 of 24 FMD participants and all 21 controls completed (Maloh et al., 2023) [6].

These figures do not sort into one pattern. The trial with the largest share of FMD participants not completing was a 3-month trial in generally healthy adults, and most of its withdrawals were for scheduling or personal reasons; the 12-month trial's losses were spread evenly across its FMD and control arms. They should not be averaged into one sitewide number: each trial's own population, duration, and recorded reasons belong next to its own figure.

Reported Adverse Events in Medically Supervised Trials

Every trial in this set recorded adverse events. In the trials in generally healthy adults, those were the mild to moderate symptoms described above, with no serious adverse effects reported in the 100-person trial (Wei et al., 2017) [2] and no major adverse effects reported in the pilot (Brandhorst et al., 2015) [1]. In the 71-day trial in healthy women, the five recorded events (one each of hand tingling, leg pain, joint pain, a kidney infection, and a knee ligament rupture) were judged not serious and not related to the diet, and all resolved (Maloh et al., 2023) [6]. In the 12-month trial in people with type 2 diabetes on metformin or diet alone, eight serious adverse events occurred in five FMD participants and none in the control group; the authors judged none of them related to the study (van den Burg et al., 2024) [3].

Low blood sugar appears only in the two trials that enrolled people with type 2 diabetes on glucose-lowering medication, and both ran under active clinical supervision with dose adjustment written into the protocol.

A 3-month randomized trial enrolled 56 South-Asian Surinamese adults with type 2 diabetes, 19 of them assigned to the FMD. During each cycle, sulfonylureas and short-acting insulin were stopped and long-acting insulin was halved, with fasting glucose checks on the days after the cycle. Two low-blood-sugar episodes (below 4.0 mmol/L) were recorded in participants still using a sulfonylurea or long-acting insulin, one of whom left the trial for that reason. Three participants discontinued for adverse events or non-adherence during or after the first cycle, one was withdrawn by the investigators after a serious adverse event after the first cycle, and one participant's kidney function was found seriously reduced after the first cycle, which the authors attribute to dehydration and which fully recovered after intravenous fluids (van der Velden et al., 2024) [4].

A 6-month trial enrolled 40 people with type 2 diabetes and diabetic kidney disease, aged 50 to 75. During each cycle, short-acting insulin and oral glucose-lowering drugs were stopped and long-acting insulin was halved, participants checked their glucose at least four times a day, and a 24-hour telephone line was available. Mild or moderate weakness, muscle pain, dizziness, and headache were the most common symptoms, nothing of grade 3 or higher was reported, and low-blood-sugar episodes did not differ between the FMD group and the Mediterranean-diet comparison group (grade 1 in 10% versus 11% of participants, grade 2 in 5% versus 11%) (Sulaj et al., 2022) [5].

Both of these trials enrolled people already on diabetes medication, and both built dose adjustment and glucose monitoring into the protocol under active clinical supervision from the start. That is the context the low-blood-sugar episodes and the kidney-function episode happened in. Neither trial is evidence about what happens to someone on the same medication who tries an FMD cycle without that supervision in place, and this page does not treat it as such.

A Positive Tolerability Signal

Not every tolerability finding in this set is about adverse events or dropout. A trial in 45 healthy women, aged 35 to 60, followed three monthly FMD cycles over 71 days and found self-reported happiness and confidence improved, alongside separate skin-hydration findings (Maloh et al., 2023) [6]. This is a real signal worth stating plainly. It is also one small trial's self-reported outcome, not a general mood-benefit claim for FMD.

Who Is Most at Risk of a Side Effect

The trials above point to one clear pattern: low-blood-sugar risk concentrates in people already on diabetes medication, particularly insulin or a sulfonylurea, and the one recorded episode of reduced kidney function occurred in that same medically supervised diabetes trial and was attributed to dehydration. This page does not re-derive the full list of who should talk to a clinician before trying an FMD. See the overview's Who Should Not Try an FMD Without a Clinician section for the complete, already-reviewed list.

What the Trials Don't Tell Us

No trial in this set followed tolerability past about 12 months, the duration of the longest trial here (van den Burg et al., 2024) [3]. None measured whether side effects or dropout change across many repeated cycles over years of real-world use. The adverse-event and dropout reporting above is the secondary safety reporting of trials designed to measure metabolic, kidney, or skin outcomes, not a dedicated tolerability study, and the healthy-adult trials ran for about three months. Read this page as what these trials reported about tolerability over 71 days to 12 months, not as a complete safety profile.

Safety and Cautions

This is general information, not medical advice. Talk to a clinician before trying an FMD if you have a history of an eating disorder. Anyone taking medication for diabetes or blood pressure should talk to a clinician before trying a multi-day, low-calorie FMD cycle; low-blood-sugar risk is highest with insulin or a sulfonylurea. If you are already on this kind of medication and have questions about how it might interact with a change in eating pattern, talk to your clinician rather than adjusting your medication on your own. See the FMD overview for the full list of who should talk to a clinician first.

Authors and Funding of the Cited Studies

• Brandhorst et al., 2015 [1]: Longo is senior and corresponding author. From the paper: "VDL and TEM have equity interest in L-Nutra, a company that develops medical food. Neither author had any role in data analysis." The University of Southern California has licensed intellectual property to L-Nutra, with potential royalty payments to the university. Funding statement: "The human study was funded by the USC Edna Jones chair fund."
• Wei et al., 2017 [2]: Longo is senior and corresponding author. The paper's competing-interests statement: "The experimental FMD was provided by L-Nutra Inc." The University of Southern California has licensed intellectual property to L-Nutra, with potential royalty payments to the university. "V.D.L. and T.E.M., who have equity interest in L-Nutra, did not participate in the collection and analysis of the data. One-hundred percent of V.D.L.'s equity will be assigned to the nonprofit foundation Create Cures." U.S. patents related to the work have been filed. Funding statement: "Funding was provided by the USC Edna Jones chair fund to V.D.L."
• van den Burg et al., 2024 [3]: Longo is a co-author, not the senior author. Funding statement: the trial "was co-funded by Health~Holland, Top Sector Life Sciences & Health, the Dutch Diabetes Foundation and L-Nutra." The authors' relationships statement adds that Longo "is a founder and shareholder of L-Nutra (his shares are assigned to the Create Cures Foundation and other charitable and research organisations), owns patents licensed to L-Nutra, receives support for travel expenses from L-Nutra and is on the advisory board of L-Nutra."
• van der Velden et al., 2024 [4]: no Longo author. "Fasting mimicking diet boxes (Prolon®) were provided by L-Nutra Inc." A dietary supplement used in a comparison arm, Endocalyx, was provided by MicroVascular Health Solutions LLC; co-author Hans Vink (HV) lists MicroVascular Health Solutions LLC as an affiliation, and the paper's competing-interests statement reads: "HV is Chief Science Officer at GlycoCheck BV (Maastricht, the Netherlands). Other authors have nothing to disclose." The trial was funded by consortium grant LSHM16058-SGF (GLYCOTREAT).
• Sulaj et al., 2022 [5]: Longo is a co-author, not the senior author. Disclosure, from the paper: "L-Nutra, as funder of the fasting-mimicking diet used in this study, has no role in the design or conduct of the study... V.L. is founder and shareholder of L-Nutra; his shares are destined to the Create Cures Foundation and other charitable and research organizations." The trial was funded by the German Research Foundation (DFG) and the German Center for Diabetes Research (DZD); the paper also states: "The fasting-mimicking diet used in this study was funded by L-Nutra."
• Maloh et al., 2023 [6]: no Longo author. Several authors are identified as employees of L-Nutra, and the paper states L-Nutra supplied the study data. This research was funded by L-Nutra.
• Disclosures are listed where the full text of the paper was available to us; a study not listed here is not a statement that its authors have no ties.