Fasting and Blood Sugar: How Intermittent Fasting May Address Insulin Resistance at the Root
Your fasting glucose can stay unremarkable for years while your insulin keeps climbing in the background. Here is what the studies on time-restricted eating show, where they reach their limits and what continuous glucose monitoring can make visible.
Most people receive their diagnosis on the day their blood sugar tips over. But the story begins years earlier, and it begins with insulin, not with sugar. Whoever looks only at the fasting value sees the end of that development. Not its beginning.
You are sitting in your family doctor's office. Fasting glucose is fine. HbA1c too. Nothing remarkable, says the report.
And still something is off. After lunch you drop into a hole. At four in the afternoon you need something sweet, not out of pleasure but almost compulsively. Your waist is growing although you are not eating more than before. And in the morning you feel as if you had not slept at all.
I hear this description very often. And I understand the confusion. Because the values that are measured routinely are not wrong. They are only late.
What you can expect in this article
- Why insulin runs off course long before sugar does
- What HOMA-IR, fasting insulin and C-peptide really say
- What happens in your metabolism during a long fasting phase
- The four KPNI lenses on insulin and blood sugar
- The most important studies on time-restricted eating
- TREAT and the NEJM trial as honest counterweights
- What DiRECT shows about remission, and what it does not
- What continuous glucose monitoring can make visible
- The limits: weight as a confounder, muscle loss, short trials
- Who should not fast, or fast only with supervision
When the sensor shows what the lab report does not
Before we come to the numbers, a pattern I know very well. Anonymised and in parts composed, so that nobody can recognise a specific person in it.
“My sugar is normal. So why do I feel like this?”
Imagine a woman in her early fifties. She works a lot at a desk, sleeps badly and has gained about eight kilos in recent years, mostly around the abdomen. Her fasting glucose sits at the upper edge of the normal range, her HbA1c in the prediabetic range.
She has been given the advice that almost everyone gets: fewer carbohydrates, more movement, lose weight. She tried it. It changed little, and she felt quite guilty about that.
What we additionally looked at was not sugar alone, but fasting insulin, C-peptide and, derived from those, HOMA-IR. And we placed a glucose sensor for two weeks, so that she could see her own responses with her own eyes.
That was the moment something shifted for her. Her muesli in the morning produced a clearly higher rise than the same muesli in the evening. A twenty minute walk after eating visibly flattened the curve. And the night she had slept badly could be read in the trace the next morning.
She then moved her eating window forward and left out the late snacks. After several months, fasting insulin and HOMA-IR were lower than before, and she described the afternoon slump as having become less frequent.
I cannot claim causality, I am only documenting the temporal association. In parallel she had slept better, moved more and shopped differently. Which of those contributed what, I cannot separate out. This is a single case, not proof of cause and effect, and it is not transferable to other people.
The lesson: visibility changes behaviour more strongly than any well meant piece of advice.
Most people think prediabetes is a sugar problem. Physiologically it is first of all an insulin problem.
Sugar stays normal for so long because insulin keeps it normal with ever more effort. When sugar rises, that compensation is already exhausted. That is not bad news. It is a hint that there is a window of time in which a great deal can still move.
Why fasting glucose is the last marker to tip
Imagine insulin as a doorman who lets glucose into the cells. In insulin resistance the doorman hears less well. So the pancreas sends out more doormen. From the outside everything stays calm, because sugar still gets in. The price for that is a permanently high insulin level.
This phase can last years. And in exactly this phase fasting glucose is a poor detective.
Tabák and colleagues analysed the Whitehall II cohort, 6,538 British employees without diabetes at baseline. Over a median of 9.7 years, 505 of them developed type 2 diabetes.
Looking back, fasting glucose in the diabetes group at first rose only linearly. Only about three years before diagnosis did the curve turn steeply upwards, from 5.79 mmol/l to 7.40 mmol/l. The two hour value after a glucose load rose in the same period from 7.60 mmol/l to 11.90 mmol/l. Insulin sensitivity calculated with HOMA, by contrast, was already falling markedly during the five years before diagnosis. Beta cell function still rose between the fourth and third year before diagnosis, from 85.0 to 92.6 percent, and then fell to 62.4 percent by the time of diagnosis.
What this may mean for you: the body first compensates more, then the compensation runs out. Blood sugar does not follow this whole prehistory for a long time.
Tabák AG, Jokela M, Akbaraly TN, Brunner EJ, Kivimäki M, Witte DR. The Lancet, 2009. DOI: 10.1016/S0140-6736(09)60619-XWhat HOMA-IR, fasting insulin and C-peptide can do
HOMA-IR is a calculation model. It combines fasting glucose and fasting insulin into an estimate of how pronounced insulin resistance is. It was developed in 1985 by a group around Matthews in Oxford.
The estimate is remarkably usable. In the original paper it agreed with the elaborate euglycaemic clamp with a rank correlation coefficient of 0.88, and with fasting insulin alone at 0.81. At the same time the authors were honest: the coefficient of variation was 31 percent for insulin resistance. That is a compass, not a map.
| Marker | What it shows | Where its limit lies |
|---|---|---|
| Fasting glucose | The current result of the regulation. | According to the Whitehall data it moves markedly only about three years before diagnosis. |
| HbA1c | Average blood sugar load over roughly eight to twelve weeks. | An average smooths peaks away. It can be distorted when the lifespan of red blood cells is altered. |
| Fasting insulin | How much effort is needed to keep sugar normal. | No uniform reference range between laboratories, varies with time of day and with stress. |
| HOMA-IR | Estimate of insulin resistance from glucose and insulin. | A calculation model with limited precision. Better suited for following a course than for a single snapshot. |
| C-peptide | Released together with insulin and shows the body's own production. | Depends on kidney function. Says something about production, not directly about sensitivity. |
And now you know why an unremarkable fasting glucose neither explains your symptoms nor rules them out.
What happens in your metabolism when you do not eat for longer
The most important quantity in fasting is not hunger. It is the time during which your insulin level is allowed to be low.
Because insulin is not only a sugar key. It is at the same time your body's storage signal. As long as insulin is high, metabolism says: store. Only when it falls may the switch be flipped.
Insulin falls
Without food intake the insulin level sinks step by step. With it the insulin load also sinks, meaning the total amount of insulin your cells are exposed to across the day. A cell that receives less permanent signalling may respond more sensitively to the signal again.
The liver empties its glycogen store
The liver keeps a supply of storage sugar ready. After about twelve hours without food this supply is clearly reduced. A liver with a lower fill level takes up glucose more readily. That is a central building block of hepatic insulin sensitivity.
Lipolysis starts
Low insulin releases fat tissue. Fatty acids are mobilised and become available as fuel. This can also concern the fat stored in liver and muscle, where it disturbs insulin signalling pathways.
Ketone bodies appear
The liver builds ketone bodies from fatty acids. They are an alternative fuel, above all for the brain. In the review by de Cabo and Mattson in the New England Journal of Medicine this switching point is described as a metabolic switch.
Cellular clean-up processes ramp up
With low insulin and low energy availability, the balance between the sensors mTOR and AMPK shifts. Autophagy, the cell's own disposal of damaged components, may increase as a result. In humans the time windows for this are considerably less well documented than it often sounds online.
The internal clock has a say
Your insulin sensitivity is higher in the morning than in the evening. The pancreas, the liver and fat tissue have their own daily rhythms. A late dinner therefore meets a metabolism that has actually already powered down.
The four KPNI lenses on insulin and blood sugar
In clinical psychoneuroimmunology we do not look at one organ but at four interwoven systems. With blood sugar all four mesh together.
Nervous system
Chronic stress and lack of sleep raise sympathetic activity. That can increase glucose release from the liver and dampen insulin action at the muscle cell. This is why a bad night may become visible in the sensor trace the next morning.
Immune system
Visceral fat tissue is not a silent store. It releases messengers that maintain low grade inflammation. These inflammatory signals can interfere with the cell's insulin signalling chain. That is one of the reasons why abdominal fat and insulin resistance are so closely connected.
Metabolism
Fat that does not fit into fat tissue ends up in liver and muscle. There it can disturb the onward transmission of the insulin signal. This is exactly the point at which both weight loss and longer fasting phases take effect.
Hormonal system
Cortisol, thyroid hormones, growth hormone and sex hormones all intervene in glucose regulation. This is why women respond differently to the same meal across the cycle, and why a thyroid disorder can change the picture completely.
And now you know why insulin resistance is never only a question of nutrition.
“It is not the sugar in the blood that calls for help first. It is the amount of insulin needed to keep it looking unremarkable.”
Shukri Jarmoukli, ViveCura BerlinWhat the intervention trials actually show
Now to the data. I tell them in the order in which they changed the field, and I do not leave out the uncomfortable ones.
Sutton, Peterson and the group at the Pennington Biomedical Research Center asked a question nobody had answered cleanly before: does anything remain from fasting once you take weight loss out of the equation?
They took men with prediabetes through two phases of five weeks each. In one phase they ate within a window of six hours, with dinner before 3 pm. In the other within a window of twelve hours. All meals were provided, in an amount that kept weight stable. Under the early eating window, insulin sensitivity, beta cell responsiveness, blood pressure, oxidative stress and appetite improved.
What this may mean for you: part of the effect could depend on the time structure, not only on the calorie balance. That is a strong signal from a very small study. It ran only in men and only over five weeks.
Sutton EF, Beyl R, Early KS, Cefalu WT, Ravussin E, Peterson CM. Cell Metabolism, 2018. DOI: 10.1016/j.cmet.2018.04.010Wilkinson, Taub and the group around Panda examined 19 people with metabolic syndrome who had previously eaten spread over at least 14 hours a day. Most of them were already taking a statin or a blood pressure medication.
They shortened their eating window over twelve weeks to ten hours, which they could choose themselves. Weight, blood pressure and atherogenic blood lipids went down. The authors describe time-restricted eating as a lifestyle measure that could be used in addition to normal medical care.
What this may mean for you: ten hours is more practicable in everyday life than six. But the study had no control group and only 19 participants. It shows feasibility, not superiority.
Wilkinson MJ, Manoogian ENC, Zadourian A, et al. Cell Metabolism, 2020. DOI: 10.1016/j.cmet.2019.11.004Cienfuegos, Gabel and the group around Varady in Chicago compared two extreme eating windows. One group ate only between 3 pm and 7 pm, one between 1 pm and 7 pm, a third without any specification.
After eight weeks both fasting groups had lost about three percent of body weight, without counting calories. Energy intake fell in both groups by about 550 kilocalories per day. Insulin resistance and oxidative stress were lower than in the control group. A difference between four and six hours did not appear.
What this may mean for you: shorter is not automatically better. A large part of the effect is likely to come simply from the fact that less is eaten within a short window.
Cienfuegos S, Gabel K, Kalam F, et al. Cell Metabolism, 2020. DOI: 10.1016/j.cmet.2020.06.018Jamshed, Peterson and colleagues in Birmingham randomly divided 90 adults with obesity into two groups. Both received the same weight management support. One group ate between 7 am and 3 pm, the other over at least twelve hours.
After 14 weeks the group with the early window had lost 2.3 kg more, with a confidence interval from 3.7 to 0.9 kg. The difference in fat mass, at 1.4 kg, was not significant. Diastolic blood pressure was 4 mmHg lower, and mood and drive were better. The authors converted the effect into a daily calorie saving of about 214 kilocalories.
What this may mean for you: if there is to be an eating window at all, then rather early in the day. That fits the circadian logic and the results of Sutton and Peterson.
Jamshed H, Steger FL, Bryan DR, et al. JAMA Internal Medicine, 2022. DOI: 10.1001/jamainternmed.2022.3050And now the studies that make it uncomfortable
If I quoted only the four above, this article would be an advertising brochure. So here come the three that point in the other direction.
Lowe, Weiss and colleagues in San Francisco randomised 116 people with overweight or obesity, on average 46.5 years old. One group ate according to the 16:8 principle between 12 pm and 8 pm, the other had three structured meals per day.
The fasting group lost 0.94 kg, the comparison group 0.68 kg. The difference between the groups was 0.26 kg and was not significant. For fasting insulin, fasting glucose and HbA1c there were no significant changes. At the same time the appendicular lean mass index was lower in the fasting group.
What this may mean for you: a late eating window without any change in diet could deliver considerably less than the hype promises. And it could cost muscle mass.
Lowe DA, Wu N, Rohdin-Bibby L, et al. JAMA Internal Medicine, 2020. DOI: 10.1001/jamainternmed.2020.4153Liu, Zhang and colleagues in Guangzhou allocated 139 people with obesity to two groups. Both received the same calorie target. One group additionally ate only between 8 am and 4 pm.
After twelve months the eating window group had lost 8.0 kg, the comparison group 6.3 kg. The difference of 1.8 kg, with a confidence interval from 4.0 to plus 0.4 kg, was not significant. For waist circumference, body fat, blood pressure and the metabolic risk factors there were no meaningful differences either.
What this may mean for you: when calories are the same, the additional benefit of the eating window could be smaller than assumed. This is the longest trial of its kind so far.
Liu D, Huang Y, Huang C, et al. New England Journal of Medicine, 2022. DOI: 10.1056/NEJMoa2114833Trepanowski and the group around Varady compared alternate day fasting with daily calorie reduction and a control group over one year, 100 participants in total.
After twelve months the fasting group had lost 6.0 percent of weight, the group with daily calorie reduction 5.3 percent. For blood pressure, triglycerides, fasting glucose, fasting insulin and insulin resistance there were no significant differences between the intervention groups. The dropout rate was highest in the fasting arm at 38 percent. And LDL cholesterol after twelve months was 11.5 mg/dl higher in the fasting group than in the comparison group.
What this may mean for you: alternate day fasting is hard for many people to sustain. And it is not automatically superior metabolically.
Trepanowski JF, Kroeger CM, Barnosky A, et al. JAMA Internal Medicine, 2017. DOI: 10.1001/jamainternmed.2017.0936Two summaries put the overall picture in context. Cioffi and colleagues analysed eleven randomised trials over 8 to 24 weeks. The weight difference between intermittent and continuous energy restriction was minus 0.61 kg with a confidence interval from minus 1.70 to plus 0.47 kg, so without a robust advantage. For fasting insulin they found a small difference in favour of fasting, whose clinical relevance they themselves described as uncertain.
Patikorn and colleagues examined eleven meta-analyses with a total of 130 randomised trials in an umbrella review. The median study size was 38 participants, the median follow-up three months.
Of 104 associations examined, 28 were statistically significant, so 27 percent. Exactly one association, meaning 1 percent, was supported by high quality evidence: modified alternate day fasting over one to two months and a moderate reduction of body mass index. Six associations were of moderate quality, 72 percent were of very low quality. And: intermittent fasting was associated with a reduction in fat-free mass.
That is no reason to write fasting off. It is a reason to adjust expectations.
DiRECT: what remission shows, and what does not follow from it
When we talk about insulin resistance at the root, there is one trial we cannot get past. It has nothing to do with intermittent fasting. But it shows how far metabolism can move.
Lean, Taylor and colleagues in Scotland and northern England cared for 306 people with type 2 diabetes of at most six years' duration, 149 per group in the analysis. The intervention group received a structured formula diet, then a stepwise reintroduction of foods and long-term support.
After twelve months, 68 of 149 people in the intervention group reached remission, so 46 percent, compared with 6 of 149 in the control group. The extent of weight loss was decisive: of those who had gained weight, none of 76 was in remission. Of those who had lost 15 kg or more, 31 of 36 were. After 24 months remission still stood at 53 of 149, so 36 percent.
What this may mean for you: remission has been described in a proportion of participants, and it depends strongly on the extent and the durability of weight loss. That was a highly structured programme with medical supervision, not a do-it-yourself experiment.
Lean MEJ, Leslie WS, Barnes AC, et al. The Lancet, 2018. DOI: 10.1016/S0140-6736(17)33102-1The mechanistic companion study by Taylor and colleagues makes visible why that may be possible. Liver fat content in the intervention group fell from 16.0 percent to 3.1 percent, independently of whether blood sugar regulation recovered fully. Those who returned to the non-diabetic range differed above all through a recovery of the first phase insulin response. These people had had diabetes for a mean of 2.7 years, the others for 3.8 years.
The beta cells are not necessarily lost for good. That is perhaps the most important message from this work, and it contradicts the image of diabetes as an unstoppable one way street.
At the same time it shows a time axis. The earlier something happens, the more room there seems to be. That is the actual reason why I place so much weight on the early markers.
What continuous glucose monitoring can make visible
A lab value is a photograph. A sensor trace is a film. And some things only become visible in the film.
In continuous glucose monitoring, CGM for short, a small sensor sits in the tissue under the skin and measures every few minutes. From that, four levels of observation emerge: the height of values after eating, the time in range, the variability across the day and the nocturnal baseline.
Zeevi, Segal, Elinav and colleagues at the Weizmann Institute followed 800 people for a week with a glucose sensor and analysed 46,898 meals.
The result was the surprise of this work: participants responded very differently to identical meals. The researchers developed a prediction model from that, which calculates personal blood sugar responses, and confirmed it in an independent group of 100 people. In a blinded randomised dietary intervention, blood sugar responses were lower under the personalised recommendations.
What this may mean for you: the question is not only whether a food is healthy. It is also what it does in you. Universal nutrition rules could be less suitable for blood sugar than we thought.
Zeevi D, Korem T, Zmora N, et al. Cell, 2015. DOI: 10.1016/j.cell.2015.11.001Hall, Snyder and colleagues at Stanford analysed sensor traces and assigned people to different so-called glucotypes based on their patterns.
One finding stays with you: even people who count as normoglycaemic by the usual criteria spent 15 percent of the time at values in the prediabetic range and 2 percent in the diabetic range. The underlying physiology varied greatly between participants.
What this may mean for you: the category normal is coarser than the lab report suggests. That is an argument for looking at a course rather than a snapshot.
Hall H, Perelman D, Breschi A, et al. PLoS Biology, 2018. DOI: 10.1371/journal.pbio.2005143- It can show how your very personal curve responds to a particular meal.
- It can show what a walk after eating changes in your curve.
- It can show how a short night or a stressful day affects your baseline.
- It can show whether your trace is calm or restless, meaning how large the variability is.
- It cannot measure insulin resistance. For that you need insulin or C-peptide in blood.
- It is not validated as a diagnostic instrument for people without diabetes. The target ranges of the international consensus papers by Battelino and colleagues were developed for people with diabetes.
- It is not necessary for everyone. For some people it is a learning tool for a limited time, for others an additional source of unease.
In practice I therefore use CGM mostly for a limited period, over two to four weeks, with a clear question beforehand and a joint review afterwards. Not as permanent surveillance, but as a learning phase. And I always combine it with laboratory values that depict insulin itself, because the sensor shows only the result, not the effort behind it.
A functional view does not only ask: is the value within the normal range. It asks: how much regulation is needed for it to stay there, and how flexible is your metabolism still.
That is not a criticism of standard diagnostics. Fasting glucose and HbA1c are well validated and are rightly the standard for diagnosing diabetes. What a functional view can add is a look at the years before that.
Where the evidence ends, and why I say so
This is where it gets uncomfortable. And for exactly that reason it matters.
First, the confusion with weight. In most trials the fasting groups lose weight. Improved insulin sensitivity after weight loss is well documented. Whether the eating window contributes anything beyond that can only be tested in trials with stable weight. The work by Sutton and Peterson is so valuable for exactly that reason, and it is very small.
Second, the study sizes and the duration. In the umbrella review the median number of participants was 38 and the median follow-up three months. Metabolic changes need years. Three months are a beginning, not an answer.
Third, the loss of muscle. Muscle is the largest consumer of glucose in the body. If fasting costs fat-free mass, part of the benefit is lost again. In TREAT the appendicular lean mass index was lower in the fasting group, and in the umbrella review intermittent fasting overall was associated with a reduction in fat-free mass.
Fourth, the sensor question. CGM in people without diabetes is scientifically still a young field. There are no validated thresholds for this group, no evidence that sensor measurement prevents disease in the long run, and a real risk that it amplifies food anxiety in sensitive people.
My position on this, clearly named as a position: I consider time-restricted eating with an early window a sensible, low-threshold lever in early insulin resistance. I do not consider it superior to a good change of diet with enough protein and strength training. And I consider it a poor substitute for sleep.
Who should not fast, or fast only under medical supervision
Now the section I consider the most important. Fasting is a stimulus, and not every body currently has the reserves for stimuli.
- Pregnancy and breastfeeding. Here the body needs continuous supply. Fasting concepts do not belong in this phase of life.
- Being underweight. Whoever already has too few reserves loses, under fasting, preferentially what they can least spare.
- A history of an eating disorder. Rules around eating times can set old patterns going again. That also applies to the combination of a sensor and a food diary.
- Children and adolescents. Growth and development need reliable intake, not restriction.
- Type 1 diabetes. Here there is a real risk of hypoglycaemia and of a derailment towards ketoacidosis.
- Insulin and sulfonylureas. These medications lower blood sugar regardless of whether you have eaten. During a fasting phase that can turn into a dangerous hypoglycaemia.
- Other glucose lowering medications. With SGLT2 inhibitors and other drug classes too, the conditions change under fasting. That belongs in a conversation beforehand.
- Medication taken with meals. A shifted eating window also shifts the intake. That can change tolerability and effectiveness.
- High stress load, exhaustion, lack of sleep. Fasting is an additional stress stimulus. On an exhausted system it can amplify symptoms instead of easing them.
- Thyroid disease. The thyroid responds sensitively to energy shortage. Taking thyroid hormone is also tied to a time interval away from meals.
- Gout, kidney disease, migraine with fasting triggers. Here an individual assessment is worthwhile before you start.
Very important: an adjustment of medication may be made exclusively by the physician treating you. Never on your own, never on the basis of a blog article, this one included.
Grajower and Horne have written a review on exactly this. Their conclusion is twofold: the data on fasting in diabetes are very thin, there is essentially one randomised trial and a few case reports. And even so, fasting in diabetes can be arranged safely under medical supervision and with appropriate blood sugar monitoring. Both belong together.
Three levers you can look at this week
Enough theory. What follows from this in concrete terms?
Lever 1: move the eating window forward, do not only shorten it
- The clearest difference between the positive and the sobering trials was the time of day, not the length.
- A realistic entry point is a window of ten to twelve hours that begins early and ends early.
- Keep the last meal at a clear distance from bedtime where possible. Your insulin sensitivity is lower in the evening.
- If your work means you eat late, a regular rhythm matters more than a short window.
Lever 2: think about muscle, not only about the scales
- Enough protein in every meal within the window. That is the best protection against a loss of fat-free mass.
- Strength training two to three times per week. Muscles are the largest glucose consumer you can influence.
- Ten to twenty minutes of walking after the largest meal. That can visibly flatten the blood sugar peak.
- If you become weaker at the same weight, the balance between stimulus and recovery is not right.
Lever 3: look at the early markers before sugar does it for you
- At your next blood draw, ask whether fasting insulin can be measured in addition to fasting glucose.
- HOMA-IR can be calculated from both values. As a measure of change over time it says more than as a single value.
- If a sensor seems sensible, then for a limited period, with a clear question beforehand and a joint review afterwards.
- If you take medication that affects blood sugar, speak with the practice treating you before any change.
And now you know why the question is not whether intermittent fasting lowers blood sugar. It is: in whom, in which time window, with which diet behind it and measured with which marker.
Frequently asked questions about fasting, blood sugar and insulin resistance
What exactly is insulin resistance?
Insulin resistance means that your cells respond more weakly to the signal insulin than they used to. Insulin is the key that brings sugar out of the blood and into the cells. When the lock becomes blunter, the pancreas produces more keys. Blood sugar therefore stays within the normal range for a long time, while the insulin level rises in the background. That is exactly why insulin resistance can look unremarkable for years when only fasting glucose is measured. It is not a diagnosis with a sharp boundary, it is a continuum.
Why is fasting glucose not enough as an early warning sign?
Because it moves late. In the Whitehall II analysis by Tabák and colleagues, 6,538 people without diabetes were followed over a median of 9.7 years, and 505 developed type 2 diabetes. At first fasting glucose rose only slowly. Only about three years before diagnosis did it climb steeply, from 5.79 mmol/l to 7.40 mmol/l. Insulin sensitivity calculated with HOMA, by contrast, had already fallen markedly during the five years before that. So fasting glucose is more the end of a long development than its beginning.
What does HOMA-IR say and how is it derived?
HOMA stands for Homeostasis Model Assessment. The value is calculated from fasting glucose and fasting insulin and estimates how pronounced insulin resistance is. Matthews and colleagues described the model in 1985 in Diabetologia. The estimate correlated with the elaborate euglycaemic clamp with a rank correlation coefficient of 0.88, and with fasting insulin alone at 0.81. The authors themselves named a low precision, though: the coefficient of variation was 31 percent for insulin resistance. HOMA-IR is therefore a good orientation value, not a precision instrument.
Can intermittent fasting improve insulin sensitivity even without weight loss?
There is a very interesting indication for that. Sutton, Peterson and the group at the Pennington Biomedical Research Center carried out a controlled feeding study in men with prediabetes. All of them were given enough food to keep their weight stable. For five weeks they ate within a six hour window with dinner before 3 pm, then they switched to a twelve hour window. Under the early eating window, insulin sensitivity, beta cell responsiveness, blood pressure and oxidative stress improved. That was the first study to show that part of the effect does not depend on weight. It is small, short and was carried out exclusively in men.
Which form of fasting is best studied in insulin resistance?
The best documented approach is time-restricted eating. Windows of ten, eight, six and four hours have been examined. Wilkinson and the group around Panda tested ten hours in 19 people with metabolic syndrome over twelve weeks. Cienfuegos and the group around Varady compared four against six hours over eight weeks and found about three percent weight loss in both groups and lower insulin resistance than in controls. Jamshed and colleagues examined an early window from 7 am to 3 pm over 14 weeks. Whether the shortest window is the best cannot be derived from these data.
Is intermittent fasting better than classic calorie restriction?
For weight, probably not. Cioffi and colleagues pooled eleven randomised trials over 8 to 24 weeks. The weight difference between intermittent and continuous energy restriction was minus 0.61 kg with a confidence interval from minus 1.70 to plus 0.47 kg, so without a robust advantage. For fasting insulin there was a small difference in favour of the intermittent approach, and the authors themselves called its clinical relevance uncertain. In the NEJM trial by Liu and colleagues with 139 people over twelve months, the difference was minus 1.8 kg and was not significant. The possible advantage of fasting lies more in everyday practicability than in superiority.
What does the TREAT trial show and why does it matter?
TREAT is the most important counterweight to fasting enthusiasm. Lowe and colleagues randomised 116 people with overweight or obesity over twelve weeks either to a 16:8 window from 12 pm to 8 pm or to three structured meals. The weight difference between the groups was minus 0.26 kg and was not significant. For fasting insulin, fasting glucose and HbA1c there were no significant changes. There was also a striking difference in the appendicular lean mass index to the disadvantage of the fasting group. The context matters: the eating window lay late in the day, and there was no change in diet. That could be exactly what explains the difference from the trials with an early window.
Can type 2 diabetes go into remission?
In trials, remission has been described in a proportion of participants. In the DiRECT trial by Lean and colleagues, 68 of 149 people in the intervention group reached remission after twelve months, which is 46 percent, compared with 6 of 149 in the control group. After 24 months remission still stood at 53 of 149, so 36 percent. The extent of weight loss was decisive: among those who lost 15 kg or more, 31 of 36 were in remission, and among those who gained weight, none of 76. The accompanying study by Taylor and colleagues showed in parallel a fall in liver fat from 16.0 to 3.1 percent. That was, however, a highly structured programme with a formula diet and medical supervision, not intermittent fasting.
What is the point of continuous glucose monitoring if I do not have diabetes?
It can make individual patterns visible that a single fasting measurement does not show. Zeevi and colleagues followed 800 people for a week with a sensor and analysed 46,898 meals. Participants responded very differently to identical meals. Hall and colleagues found that even people who count as normoglycaemic by standard criteria spent 15 percent of the time at values in the prediabetic range. That is informative, but it has limits: CGM is not validated as a diagnostic instrument in people without diabetes, the target ranges of the international consensus papers apply to people with diabetes, and not everyone needs a sensor.
Who should not fast, or fast only under medical supervision?
Pregnant and breastfeeding women, people who are underweight and people with a history of an eating disorder should not fast. Children and adolescents do not belong in fasting concepts either. In type 1 diabetes, on insulin and on sulfonylureas there is a real risk of hypoglycaemia, and here fasting is conceivable only under medical supervision. Grajower and Horne emphasise in their review that the data in diabetes are very thin and that an adjustment of medication may be made exclusively by the physician treating you. Caution also applies with a high stress load, with thyroid disease and with medication whose intake is tied to meals.
Will I lose muscle mass with intermittent fasting?
The risk is real and is named too rarely. In the umbrella review by Patikorn and colleagues, which analysed eleven meta-analyses with 130 randomised trials, intermittent fasting was associated with a reduction in fat-free mass. In the TREAT trial the appendicular lean mass index was lower in the fasting group than in the comparison group. Muscle is the largest consumer of glucose in the body. So whoever loses muscle also loses part of the capacity they actually want to improve. That is why enough protein and strength training belong, in my view, to every fasting concept.
Read on in the guide Fasting
Blood sugar is only one of the levels on which fasting phases can change something. Cellular clean-up, the cycle and movement belong to the same picture.
Fasting and blood sugar: insulin resistance at the root
Early markers, mechanisms, studies and the limits of the evidence.
You are hereIntermittent fasting 16:8: what happens hour by hour
The time course in metabolism, from the last meal to the sixteenth hour.
Autophagy: when does cellular clean-up really begin?
Why the famous hour threshold is more weakly documented scientifically than it sounds.
Fasting in women: the cycle has a say
Why the same form of fasting lands differently in women than in men.
Movement as medicine: the effect at cell level
How muscle work can bring glucose into the cell even without insulin.
All articles in the guide Fasting
From the first fasting hour to the limits of fasting research.
Sources
All details were cross-checked against title, journal, year, PMID and DOI via PubMed. The study type is given in square brackets.
- Sutton EF, Beyl R, Early KS, Cefalu WT, Ravussin E, Peterson CM. Early Time-Restricted Feeding Improves Insulin Sensitivity, Blood Pressure, and Oxidative Stress Even without Weight Loss in Men with Prediabetes. Cell Metab. 2018;27(6):1212-1221.e3. DOI: 10.1016/j.cmet.2018.04.010 (PMID 29754952) [RCT, controlled feeding study, crossover, men with prediabetes]
- Lowe DA, Wu N, Rohdin-Bibby L, Moore AH, Kelly N, Liu YE, Philip E, Vittinghoff E, Heymsfield SB, Olgin JE, Shepherd JA, Weiss EJ. Effects of Time-Restricted Eating on Weight Loss and Other Metabolic Parameters in Women and Men With Overweight and Obesity: The TREAT Randomized Clinical Trial. JAMA Intern Med. 2020;180(11):1491-1499. DOI: 10.1001/jamainternmed.2020.4153 (PMID 32986097) [RCT, n=116, 12 weeks]
- Liu D, Huang Y, Huang C, Yang S, Wei X, Zhang P, Guo D, Lin J, Xu B, Li C, He H, He J, Liu S, Shi L, Xue Y, Zhang H. Calorie Restriction with or without Time-Restricted Eating in Weight Loss. N Engl J Med. 2022;386(16):1495-1504. DOI: 10.1056/NEJMoa2114833 (PMID 35443107) [RCT, n=139, 12 months]
- Jamshed H, Steger FL, Bryan DR, Richman JS, Warriner AH, Hanick CJ, Martin CK, Salvy SJ, Peterson CM. Effectiveness of Early Time-Restricted Eating for Weight Loss, Fat Loss, and Cardiometabolic Health in Adults With Obesity: A Randomized Clinical Trial. JAMA Intern Med. 2022;182(9):953-962. DOI: 10.1001/jamainternmed.2022.3050 (PMID 35939311) [RCT, n=90, 14 weeks]
- Cienfuegos S, Gabel K, Kalam F, Ezpeleta M, Wiseman E, Pavlou V, Lin S, Oliveira ML, Varady KA. Effects of 4- and 6-h Time-Restricted Feeding on Weight and Cardiometabolic Health: A Randomized Controlled Trial in Adults with Obesity. Cell Metab. 2020;32(3):366-378.e3. DOI: 10.1016/j.cmet.2020.06.018 (PMID 32673591) [RCT, 8 weeks]
- Wilkinson MJ, Manoogian ENC, Zadourian A, Lo H, Fakhouri S, Shoghi A, Wang X, Fleischer JG, Navlakha S, Panda S, Taub PR. Ten-Hour Time-Restricted Eating Reduces Weight, Blood Pressure, and Atherogenic Lipids in Patients with Metabolic Syndrome. Cell Metab. 2020;31(1):92-104.e5. DOI: 10.1016/j.cmet.2019.11.004 (PMID 31813824) [Real-world study, single arm, n=19, 12 weeks]
- Trepanowski JF, Kroeger CM, Barnosky A, Klempel MC, Bhutani S, Hoddy KK, Gabel K, Freels S, Rigdon J, Rood J, Ravussin E, Varady KA. Effect of Alternate-Day Fasting on Weight Loss, Weight Maintenance, and Cardioprotection Among Metabolically Healthy Obese Adults: A Randomized Clinical Trial. JAMA Intern Med. 2017;177(7):930-938. DOI: 10.1001/jamainternmed.2017.0936 (PMID 28459931) [RCT, n=100, 12 months]
- Cioffi I, Evangelista A, Ponzo V, Ciccone G, Soldati L, Santarpia L, Contaldo F, Pasanisi F, Ghigo E, Bo S. Intermittent versus continuous energy restriction on weight loss and cardiometabolic outcomes: a systematic review and meta-analysis of randomized controlled trials. J Transl Med. 2018;16(1):371. DOI: 10.1186/s12967-018-1748-4 (PMID 30583725) [Meta-analysis, k=11 trials]
- Patikorn C, Roubal K, Veettil SK, Chandran V, Pham T, Lee YY, Giovannucci EL, Varady KA, Chaiyakunapruk N. Intermittent Fasting and Obesity-Related Health Outcomes: An Umbrella Review of Meta-analyses of Randomized Clinical Trials. JAMA Netw Open. 2021;4(12):e2139558. DOI: 10.1001/jamanetworkopen.2021.39558 (PMID 34919135) [Systematic review, k=11 meta-analyses, 130 RCTs]
- Tabák AG, Jokela M, Akbaraly TN, Brunner EJ, Kivimäki M, Witte DR. Trajectories of glycaemia, insulin sensitivity, and insulin secretion before diagnosis of type 2 diabetes: an analysis from the Whitehall II study. Lancet. 2009;373(9682):2215-2221. DOI: 10.1016/S0140-6736(09)60619-X (PMID 19515410) [Real-world cohort, n=6,538, median 9.7 years]
- Matthews DR, Hosker JP, Rudenski AS, Naylor BA, Treacher DF, Turner RC. Homeostasis model assessment: insulin resistance and beta-cell function from fasting plasma glucose and insulin concentrations in man. Diabetologia. 1985;28(7):412-419. DOI: 10.1007/BF00280883 (PMID 3899825) [Pathophysiology, method validation against clamp]
- Zeevi D, Korem T, Zmora N, Israeli D, Rothschild D, Weinberger A, Ben-Yacov O, Lador D, Avnit-Sagi T, Lotan-Pompan M, Suez J, Mahdi JA, Matot E, Malka G, Kosower N, Rein M, Zilberman-Schapira G, Dohnalová L, Pevsner-Fischer M, Bikovsky R, Halpern Z, Elinav E, Segal E. Personalized Nutrition by Prediction of Glycemic Responses. Cell. 2015;163(5):1079-1094. DOI: 10.1016/j.cell.2015.11.001 (PMID 26590418) [RCT component within a cohort, n=800 plus 100 validation]
- Hall H, Perelman D, Breschi A, Limcaoco P, Kellogg R, McLaughlin T, Snyder M. Glucotypes reveal new patterns of glucose dysregulation. PLoS Biol. 2018;16(7):e2005143. DOI: 10.1371/journal.pbio.2005143 (PMID 30040822) [Real-world observational study with CGM]
- Battelino T, Danne T, Bergenstal RM, Amiel SA, Beck R, Biester T, et al. Clinical Targets for Continuous Glucose Monitoring Data Interpretation: Recommendations From the International Consensus on Time in Range. Diabetes Care. 2019;42(8):1593-1603. DOI: 10.2337/dci19-0028 (PMID 31177185) [Review, international consensus recommendation]
- Lean MEJ, Leslie WS, Barnes AC, Brosnahan N, Thom G, McCombie L, Peters C, Zhyzhneuskaya S, Al-Mrabeh A, Hollingsworth KG, Rodrigues AM, Rehackova L, Adamson AJ, Sniehotta FF, Mathers JC, Ross HM, McIlvenna Y, Stefanetti R, Trenell M, Welsh P, Kean S, Ford I, McConnachie A, Sattar N, Taylor R. Primary care-led weight management for remission of type 2 diabetes (DiRECT): an open-label, cluster-randomised trial. Lancet. 2018;391(10120):541-551. DOI: 10.1016/S0140-6736(17)33102-1 (PMID 29221645) [RCT, cluster randomised, n=306, 12 months]
- Lean MEJ, Leslie WS, Barnes AC, Brosnahan N, Thom G, McCombie L, et al. Durability of a primary care-led weight-management intervention for remission of type 2 diabetes: 2-year results of the DiRECT open-label, cluster-randomised trial. Lancet Diabetes Endocrinol. 2019;7(5):344-355. DOI: 10.1016/S2213-8587(19)30068-3 (PMID 30852132) [RCT, cluster randomised, 24 month analysis]
- Taylor R, Al-Mrabeh A, Zhyzhneuskaya S, Peters C, Barnes AC, Aribisala BS, Hollingsworth KG, Mathers JC, Sattar N, Lean MEJ. Remission of Human Type 2 Diabetes Requires Decrease in Liver and Pancreas Fat Content but Is Dependent upon Capacity for β Cell Recovery. Cell Metab. 2018;28(4):547-556.e3. DOI: 10.1016/j.cmet.2018.07.003 (PMID 30078554) [RCT substudy, mechanistic investigation in humans]
- de Cabo R, Mattson MP. Effects of Intermittent Fasting on Health, Aging, and Disease. N Engl J Med. 2019;381(26):2541-2551. DOI: 10.1056/NEJMra1905136 (PMID 31881139) [Mechanistic review]
- Grajower MM, Horne BD. Clinical Management of Intermittent Fasting in Patients with Diabetes Mellitus. Nutrients. 2019;11(4):873. DOI: 10.3390/nu11040873 (PMID 31003482) [Review, safety and medication]