Anyone who has repeatedly lost weight and watched it gradually return will recognise the frustration. The first diet appears to work, normal life eventually resumes, some or all of the weight comes back and another attempt begins. After enough repetitions, it is understandable that people start wondering whether they have somehow damaged themselves in the process. Perhaps all those diets have made the body increasingly resistant to weight loss. Perhaps the metabolism has been slowed permanently. Perhaps each round leaves behind slightly more fat and slightly less muscle until, eventually, the body has been metabolically wrecked by years of dieting.

This idea has become so deeply embedded in popular nutrition culture that “yo-yo dieting ruins your metabolism” is now repeated almost as established physiology. It also sounds biologically plausible because several things really do happen during weight loss that make maintaining a lower weight difficult. Energy expenditure falls, appetite frequently increases and hormones involved in energy balance change. When weight subsequently returns, the whole experience can feel like fairly compelling evidence that the metabolism has fought back and won.

The difficulty is that several different biological phenomena have become bundled together under one convenient explanation. Metabolic adaptation after weight loss is real. Weight regain after weight loss is extremely common. Repeated cycles of losing and regaining weight are also common. What is much less certain is whether those cycles progressively damage metabolism, make somebody increasingly prone to obesity or create metabolic harm beyond the consequences of regaining the weight itself.

That distinction has become particularly important in the era of GLP-1-based obesity medicines. Millions of people may now experience substantial pharmacological weight loss, while withdrawal studies have already demonstrated that considerable weight regain can occur when treatment stops. If repeated loss and regain were independently metabolically damaging, this would have enormous implications for the long-term use of these medicines.

Two important scientific assessments published in 2026 have therefore revisited the question. After examining evidence from human and animal research, the emerging conclusion is rather different from the popular story. Weight cycling is not necessarily desirable, and repeatedly regaining lost weight can clearly remove many of the health improvements produced by losing it. However, current evidence does not convincingly demonstrate that intentional weight loss followed by regain progressively “breaks” the metabolism in people with obesity.

To understand why, we first need to separate the normal metabolic response to weight loss from the much bigger claim that repeated dieting causes permanent metabolic damage.

Your metabolism should fall when you lose weight

One of the most commonly misunderstood consequences of weight loss is a reduction in energy expenditure.

A smaller body generally requires less energy to maintain than a larger one. If somebody loses 20 kilograms, their resting metabolic rate would therefore be expected to decrease even if nothing unusual had happened physiologically. There is less tissue to maintain, less mass to move around and usually some reduction in both fat mass and fat-free mass.

Total daily energy expenditure can fall for several additional reasons. The energetic cost of physical movement decreases because a lighter body requires less energy to move. Food intake is lower, reducing the thermic effect of food. People may also subconsciously alter spontaneous physical activity and other components of non-exercise activity thermogenesis.

None of this represents a damaged metabolism. It represents a body that has become smaller and is receiving less energy.

The interesting part begins when energy expenditure falls more than would be predicted from those changes.

This phenomenon is usually called adaptive thermogenesis or metabolic adaptation. Researchers can estimate how much energy a person should expend based upon variables such as fat-free mass, fat mass, age and sex, then compare that prediction with what is actually measured. During or following weight loss, measured expenditure can sometimes be lower than predicted.

This is a genuine physiological phenomenon, but its magnitude and persistence have been the subject of considerable debate.

Metabolic adaptation is real, but its size has been exaggerated

A 2022 systematic review examined 33 studies involving 2,528 adults and found evidence of adaptive thermogenesis in many of them. However, there was enormous variation in how it was measured, and studies using more rigorous methodology tended to find considerably smaller effects. Importantly, metabolic adaptation appeared to diminish or disappear after periods of weight stabilisation and neutral energy balance (Nunes et al., 2022).

This matters because measuring metabolism while somebody is actively losing weight can produce a very different result from measuring it once body weight and energy balance have stabilised.

During active energy restriction, the body is responding not merely to being smaller but to an ongoing energy deficit. Once energy intake and expenditure have returned to equilibrium, some of that additional metabolic suppression can diminish.

A carefully controlled study following 171 women after weight loss illustrates the point. Participants lost an average of approximately 12 kilograms. Immediately after weight loss, resting metabolic rate was around 54 kcal per day lower than predicted, indicating modest metabolic adaptation. Among women measured under weight-stable conditions at one and two years, however, the difference had fallen to approximately 18–19 kcal per day and was no longer statistically significant. Most importantly, the degree of metabolic adaptation did not predict subsequent weight regain (Martins et al., 2020).

A difference of 20 kcal per day is physiologically interesting. It is not the metabolic catastrophe sometimes described online.

Then there was The Biggest Loser

No discussion of metabolic adaptation can avoid the study that probably did more than any other to create the idea of a permanently damaged metabolism.

Researchers followed contestants from the American television programme The Biggest Loser, where participants underwent an extraordinarily intensive programme of severe energy restriction and enormous amounts of exercise. The intervention was about as representative of ordinary weight management as Formula One is of driving to Tesco, but it created a remarkable natural experiment in extreme weight loss.

The 16 original participants lost an average of approximately 58 kilograms during the 30-week competition. Resting metabolic rate fell by around 610 kcal per day. Six years later, 14 participants returned for follow-up. They had regained an average of approximately 41 kilograms, yet their resting metabolic rate remained substantially below baseline. After adjustment for changes in body composition and age, the researchers estimated metabolic adaptation at approximately 500 kcal per day (Fothergill et al., 2016).

The findings understandably attracted enormous attention and were frequently interpreted as proof that extreme dieting permanently destroys metabolism.

There are several reasons to be careful with that conclusion. The study involved only 14 people at long-term follow-up and examined an unusually extreme intervention. More importantly, the participants who maintained the greatest amount of weight loss actually displayed the greatest metabolic adaptation six years later. Metabolic adaptation measured at the end of the competition did not predict who subsequently regained the most weight.

This creates a more complicated interpretation. Rather than functioning simply as a scar left behind by dieting, some of the persistent adaptation may reflect the body's ongoing response to maintaining a weight below its previous level. In other words, metabolic adaptation may be partly proportional to the degree to which somebody is successfully maintaining weight loss.

The study remains important because it demonstrates that substantial metabolic adaptation can persist under extreme circumstances. What it does not demonstrate is that every diet progressively reduces resting metabolism or that repeated attempts at weight loss inevitably produce cumulative metabolic damage.

Appetite may be the bigger biological problem

The intense focus on resting metabolic rate may also distract from another powerful component of weight regain: appetite.

Following weight loss, the body does not simply conserve some energy. It can also increase the biological drive to replace what has been lost.

Leptin provides one part of this response. Produced predominantly by adipose tissue, circulating leptin concentrations broadly reflect energy stores. When fat mass falls, leptin concentrations decline. The hypothalamus interprets this reduction as a signal that stored energy has become less abundant, influencing appetite and energy expenditure in ways that favour restoration of body weight.

Other appetite-related hormones change as well.

In a landmark study, Sumithran and colleagues placed adults with overweight or obesity on a very-low-energy diet for ten weeks and examined a range of hormones involved in appetite regulation. After weight loss, leptin, peptide YY, cholecystokinin, insulin and amylin changed in directions that would generally be expected to favour increased eating, while ghrelin increased. Importantly, several of these alterations were still detectable one year later, and participants reported greater hunger than they had at baseline (Sumithran et al., 2011).

This provides a much richer explanation for weight regain than the idea that somebody's metabolic engine has simply become broken. A person attempting to maintain substantial weight loss may be living in a body that requires slightly less energy while simultaneously generating stronger signals encouraging them to eat.

That combination can make maintenance extraordinarily difficult without requiring any permanent metabolic damage whatsoever.

Weight regain is not evidence that a diet damaged metabolism

This distinction is worth dwelling on because it changes how we interpret what happens after a diet.

Imagine somebody loses 15 kilograms. Their resting metabolic rate decreases partly because they now have a smaller body and partly because some metabolic adaptation may occur. Hunger increases. Food becomes more salient. Portion sizes gradually increase and eventually much of the weight returns.

From the individual's perspective, the conclusion that dieting slowed their metabolism is entirely understandable. They were lighter, they felt hungrier, maintaining the new weight became increasingly difficult and eventually they regained it.

Physiologically, however, the process does not require a progressively damaged metabolic system. It can be explained largely through the normal biological defence of energy stores combined with environmental and behavioural factors that make maintaining an energy deficit extremely difficult over long periods.

This is also why describing weight regain as a failure of willpower is equally unsatisfactory. The body is actively participating in the process.

The important question for weight cycling is whether repeating this sequence creates additional damage each time.

Does each diet make your resting metabolism progressively slower?

If the traditional yo-yo dieting story were correct, we might expect repeated cycles of weight loss and regain to produce a ratchet effect. Resting metabolic rate would fall during the first diet, fail to recover completely when weight returned, fall further during the next diet and gradually become lower with each successive cycle.

Human experimental evidence has not demonstrated this convincingly.

One particularly useful study dates back to 1991. Eleven women with obesity underwent three consecutive cycles consisting of two weeks of a very-low-energy diet followed by four weeks of unrestricted eating. Basal metabolic rate fell during dieting, as expected, but returned towards normal during the regain periods. By the end of the experiment, the researchers found no evidence that repeated weight cycling had progressively reduced basal metabolic rate or increased the proportion of body fat (Jebb et al., 1991).

Another study examined women after substantial weight loss followed by complete weight regain. Resting energy expenditure fell during weight loss but returned to baseline after the weight was regained, while body-fat distribution was essentially unchanged (Wadden et al., 1996).

These are small studies and cannot settle every question about decades of repeated dieting. They do, however, directly test the proposed mechanism and fail to show the progressive metabolic deterioration that the popular narrative predicts.

The comprehensive 2026 appraisal by Magkos and Stefan reached the same general conclusion after examining the broader human and animal literature. Current evidence does not provide strong support for a causal effect of weight cycling itself on resting metabolic rate in people with obesity (Magkos and Stefan, 2026a).

What about progressively gaining more fat and losing more muscle?

A second version of the weight-cycling hypothesis concerns body composition rather than metabolic rate.

During weight loss, both fat mass and lean tissue can be lost. If weight is subsequently regained primarily as fat, then repeated cycles could theoretically produce a progressively fatter and less muscular body even if total body weight eventually returned to the same place.

This would matter considerably because fat-free mass is metabolically active and skeletal muscle contributes to glucose disposal, physical function and long-term metabolic health. Repeated disproportionate loss of lean tissue followed by preferential fat regain could therefore create exactly the kind of metabolic deterioration attributed to yo-yo dieting.

The mechanism is plausible. Evidence that it routinely happens in people with obesity is much less convincing.

When weight is regained following moderate intentional weight loss, both fat and fat-free tissue tend to be restored. The exact proportions vary according to age, sex, initial adiposity, physical activity, protein intake and the amount and speed of weight change, but human studies do not consistently demonstrate that repeated cycles inevitably produce a progressively higher body-fat percentage.

The 2026 Lancet Diabetes & Endocrinology analysis concluded that current evidence does not support a causal link between weight cycling itself and adverse changes in body composition in people with obesity (Magkos and Stefan, 2026a).

There is, however, an important caveat involving people who begin relatively lean.

Weight cycling may not mean the same thing in a lean person

The physiology of losing and regaining weight depends partly upon how much fat and lean tissue somebody has before the process begins.

A person with substantial adipose stores can derive a greater proportion of the energy required during weight loss from fat. A lean individual undergoing aggressive weight reduction has less stored fat available and may consequently deplete fat-free tissue to a greater relative extent.

During subsequent refeeding, fat can sometimes be restored more rapidly than fat-free tissue. This phenomenon has been described as preferential catch-up fat. If eating remains elevated until lean tissue has also recovered, body fat may temporarily overshoot its original level.

Classic observations from the Minnesota Starvation Experiment provide an extreme example. Healthy, lean men underwent prolonged semi-starvation and subsequently displayed intense hyperphagia and rapid restoration of fat during rehabilitation. The biology of recovery from semi-starvation in lean individuals is clearly not identical to moderate therapeutic weight loss in somebody with obesity.

This distinction has led researchers to propose that repeated dieting may carry different consequences according to starting adiposity. The October 2026 Nature Reviews Endocrinology commentary argues that weight cycling could plausibly contribute to future obesity in susceptible lean individuals repeatedly attempting to lose weight, while appearing to have relatively little effect on the progression of obesity in people who already have it (Magkos and Stefan, 2026b).

This has practical relevance in a culture where dieting is hardly restricted to people with a medical indication for weight loss. Normal-weight individuals, athletes repeatedly “making weight”, and young people pursuing repeated aggressive diets for aesthetic reasons may represent a very different physiological situation from someone treating clinically significant obesity.

It is therefore too crude either to declare weight cycling universally harmless or to insist that it universally destroys metabolism.

The starting context matters.

Observational studies appear to tell a more worrying story

This is where the evidence becomes particularly interesting.

Numerous observational studies have found associations between fluctuations in body weight and poorer health outcomes. Meta-analyses have reported associations between greater weight variability and cardiovascular disease, cardiovascular mortality, hypertension and all-cause mortality.

A 2019 meta-analysis including more than 440,000 participants found that people with greater body-weight fluctuation had higher observed risks of all-cause mortality, cardiovascular mortality and cardiovascular disease (Zou et al., 2019).

Read in isolation, that sounds like strong evidence that weight cycling is harmful.

The problem is determining what caused the weight fluctuations.

People do not lose weight only because they intentionally go on diets. Weight can fall because of cancer, chronic respiratory disease, frailty, depression, gastrointestinal disease, smoking-related illness and numerous other conditions. These illnesses can independently increase mortality.

This creates the problem of reverse causality. Instead of weight fluctuation causing disease, underlying disease may cause weight fluctuation.

There are additional confounders. People with more severe obesity may make more frequent attempts to lose weight, meaning that a history of repeated dieting can simply identify people who have experienced greater or longer exposure to obesity. Ageing itself alters body composition. Smoking can influence body weight. Medication, socioeconomic circumstances and physical activity can affect both weight trajectory and disease risk.

Even defining weight cycling consistently has proved surprisingly difficult. Studies have used different thresholds for the amount of weight lost and regained, different numbers of cycles and different time periods. Some measure intentional weight loss; others simply calculate variability from repeated body-weight measurements and assume that fluctuations represent dieting.

Those are not interchangeable exposures.

Intentionality changes the interpretation considerably

The distinction between intentional and unintentional weight loss is one of the most important methodological problems in this field.

In the 2019 cardiovascular meta-analysis, only a very small proportion of the included reports could distinguish clearly between intentional and unintentional weight fluctuation. Interestingly, the subgroup containing intentional weight cycling did not show a statistically significant association with either all-cause or cardiovascular mortality, although the small number of studies limits how confidently that finding can be interpreted (Zou et al., 2019).

This problem has been recognised for decades. Epidemiological studies showing associations between weight variability and mortality have often been interpreted as evidence that repeated dieting is dangerous even when researchers did not actually know whether participants were dieting.

That is quite a leap.

If somebody loses ten kilograms because of undiagnosed disease, regains some weight during recovery and later loses weight again as illness progresses, classifying that pattern alongside somebody intentionally losing ten kilograms through dietary intervention creates an obvious problem.

The weight graph may look similar while the underlying biology could scarcely be more different.

This is one reason observational associations between weight fluctuation and mortality cannot establish that intentional dieting caused the increased risk.

Regaining weight can still reverse the benefits of losing it

None of this means that weight regain is metabolically irrelevant.

When somebody with obesity loses clinically meaningful amounts of weight, blood pressure can improve, insulin sensitivity can increase, liver fat can decrease, triglycerides can fall and glycaemic control can improve. Mechanical load on joints decreases and physical function may improve.

If much of that weight returns, many of those benefits can diminish.

That is undesirable, but it is not the same thing as the body becoming metabolically worse than it was before the person lost weight.

This distinction is surprisingly important.

Suppose somebody with obesity loses 15% of their body weight and experiences substantial improvements in HbA1c, blood pressure and liver fat. Five years later they have regained the weight and those measures return towards their previous values. The intervention has not produced permanent benefit, but neither does that automatically mean it has damaged them.

During the period at lower weight, they may have spent several years with improved metabolic health and reduced mechanical burden.

The 2026 Lancet Diabetes & Endocrinology review argues that the benefits associated with periods of intentional weight reduction — including improvements in metabolic markers, cardiovascular health and quality of life — are likely to outweigh the hypothetical risks attributed to weight fluctuation in people with obesity (Magkos and Stefan, 2026a).

This is a considerably more useful message than telling somebody who has regained weight that they have ruined their metabolism and should perhaps never attempt weight loss again.

GLP-1 medicines make this question newly relevant

The arrival of highly effective obesity pharmacotherapy has given weight cycling a very modern context.

Semaglutide, tirzepatide and the next generation of multi-agonist drugs can produce weight reductions previously difficult to achieve outside bariatric surgery. However, discontinuation of treatment commonly leads to substantial regain because the pharmacological suppression of appetite and other physiological effects diminish when treatment stops.

This has prompted understandable concern about what might happen if people repeatedly start and stop treatment.

At present, there is little evidence that pharmacologically induced weight cycling creates some unique form of metabolic damage. The much more immediate problem is that stopping an effective treatment can remove the physiological assistance that helped maintain the lower weight.

This supports the increasingly accepted view of obesity as a chronic condition that may require long-term management rather than a temporary intervention followed by an assumption that biology will politely remain where treatment left it.

There may be many legitimate reasons for stopping medication, including adverse effects, cost, availability, pregnancy or personal preference. Understanding how best to minimise weight regain after discontinuation is therefore an important research priority.

But fear of “ruining the metabolism” should not be substituted for evidence.

Why does losing weight repeatedly sometimes seem to become harder?

There is still a very real experience behind the myth.

Someone who has attempted weight loss repeatedly over twenty years may genuinely find the process harder at 50 than they did at 30. It is tempting to blame the diets themselves, but several things have changed during those twenty years.

Ageing alters body composition. Without appropriate resistance exercise, skeletal muscle can gradually decline, reducing one component of daily energy expenditure. Physical activity may decrease. Sleep may deteriorate. Menopause can alter fat distribution and influence body composition in women. Medications may change. Injuries can reduce movement. Metabolic disease may progress. Food environments and lifestyle circumstances can become very different.

The individual may also begin each new attempt at a higher body weight because previous regain exceeded previous loss, but that does not prove that dieting caused the overshoot. People predisposed to obesity are already living within biological and environmental conditions that promote long-term weight gain.

Repeated dieting can therefore become correlated with increasing body weight without necessarily being its cause. The people who struggle most with weight are also the people most likely to make repeated attempts to lose it.

This is a classic example of why correlation becomes particularly treacherous when human behaviour and physiology interact.

There are still sensible ways to reduce the downside of weight regain

Saying that weight cycling has not been shown to destroy metabolism does not mean we should be indifferent to it. Repeatedly undertaking extreme diets, losing large amounts of weight and regaining them is hardly an ideal long-term strategy.

The more useful objective is improving the durability and quality of weight loss.

Resistance training deserves particular attention because it helps preserve skeletal muscle during energy restriction. Adequate protein becomes increasingly important for the same reason, particularly in older adults. Avoiding unnecessarily aggressive energy restriction may also make weight loss more tolerable and reduce the physiological and behavioural pressure created by severe deprivation.

Dietary patterns need to be maintainable at the lower body weight. A programme that somebody can tolerate for twelve weeks but would never voluntarily follow for twelve years may produce impressive short-term photographs without solving the longer-term biological problem.

Physical activity also becomes particularly valuable during weight maintenance. Exercise may not always produce enormous weight loss by itself, but higher levels of activity are repeatedly associated with better maintenance and provide benefits to cardiovascular fitness, insulin sensitivity, muscle, bone and mental health that extend well beyond the scales.

Pharmacotherapy and bariatric surgery should be considered legitimate components of obesity treatment when clinically appropriate rather than evidence that lifestyle has somehow failed. If biological pressure to regain weight is substantial, using effective medical treatment to help manage that biology is no more morally suspect than treating hypertension pharmacologically when lifestyle alone is insufficient.

The objective should be sustained health improvement, not winning a purity contest over how it was achieved.

Metabolism is adaptive

The phrase “damaged metabolism” is appealing because it provides a simple explanation for an extremely frustrating experience. Unfortunately, it also suggests that the body has become defective when much of what we observe after weight loss is better understood as normal adaptive physiology.

A smaller body requires less energy. During active energy restriction, expenditure may fall somewhat further than changes in body composition alone would predict. Appetite-regulating hormones can change in ways that encourage food intake, and those signals can persist while a lower weight is maintained. These adaptations make long-term weight control more difficult, sometimes considerably so.

That biology deserves to be taken seriously. It helps explain why telling people with obesity simply to eat less and maintain sufficient discipline is such an inadequate model of treatment.

What the evidence does not currently support is the additional claim that each intentional cycle of weight loss and regain progressively drives resting metabolism lower, permanently worsens body composition and leaves people with obesity metabolically more damaged than before.

The newest assessments of the evidence make that distinction increasingly clear. In people with obesity, weight cycling itself appears to contribute far less to long-term metabolic harm than has often been assumed. The more important problems are the difficulty of maintaining weight loss, the return of obesity-related risk when weight is regained and the need to preserve muscle and nutritional quality throughout the process.

There may be a different conversation to have about repeated aggressive dieting in people who begin lean, where disproportionate depletion and restoration of fat and lean tissue could plausibly encourage fat overshooting. That nuance matters, particularly in athletes and in a culture that encourages perfectly healthy people to repeatedly force their bodies below weights they can comfortably maintain.

For someone with obesity who has previously lost weight and regained it, however, the evidence offers a much more constructive message. A previous unsuccessful attempt has not condemned them to a progressively failing metabolism, and fear of weight cycling is not a good reason to avoid an appropriately managed attempt to improve metabolic health.

The body certainly remembers weight loss in the sense that it mounts physiological responses designed to defend energy stores. That is one of the reasons obesity can be so difficult to treat. But adaptation is not the same thing as damage, and weight regain is not proof that the metabolism has been ruined.

Understanding that difference allows us to stop blaming either the person or their supposedly “broken” body and concentrate instead on the much more useful question: how can we make the benefits of weight loss easier to maintain?

References

Fothergill, E., Guo, J., Howard, L., Kerns, J.C., Knuth, N.D., Brychta, R., Chen, K.Y., Skarulis, M.C., Walter, M., Walter, P.J. and Hall, K.D. (2016) ‘Persistent metabolic adaptation 6 years after “The Biggest Loser” competition’, Obesity, 24(8), pp. 1612–1619. doi:10.1002/oby.21538.

Jebb, S.A., Goldberg, G.R., Coward, W.A., Murgatroyd, P.R. and Prentice, A.M. (1991) ‘Effects of weight cycling caused by intermittent dieting on metabolic rate and body composition in obese women’, International Journal of Obesity, 15(5), pp. 367–374.

Magkos, F. and Stefan, N. (2026a) ‘Is weight cycling clinically harmful?’, The Lancet Diabetes & Endocrinology, 14(7), pp. 594–607. doi:10.1016/S2213-8587(26)00037-9.

Magkos, F. and Stefan, N. (2026b) ‘Effects of weight cycling in the presence and absence of obesity’, Nature Reviews Endocrinology. doi:10.1038/s41574-026-01311-3.

Martins, C., Gower, B.A., Hill, J.O. and Hunter, G.R. (2020) ‘Metabolic adaptation is not a major barrier to weight-loss maintenance’, American Journal of Clinical Nutrition, 112(3), pp. 558–565. doi:10.1093/ajcn/nqaa086.

Nunes, C.L., Casanova, N., Francisco, R., Bosy-Westphal, A., Hopkins, M., Sardinha, L.B. and Silva, A.M. (2022) ‘Does adaptive thermogenesis occur after weight loss in adults? A systematic review’, British Journal of Nutrition, 127(3), pp. 451–469. doi:10.1017/S0007114521001094.

Sumithran, P., Prendergast, L.A., Delbridge, E., Purcell, K., Shulkes, A., Kriketos, A. and Proietto, J. (2011) ‘Long-term persistence of hormonal adaptations to weight loss’, New England Journal of Medicine, 365(17), pp. 1597–1604. doi:10.1056/NEJMoa1105816.

Wadden, T.A., Bartlett, S., Letizia, K.A., Foster, G.D., Stunkard, A.J. and Conill, A. (1996) ‘Effects of weight cycling on the resting energy expenditure and body composition of obese women’, International Journal of Eating Disorders, 19(1), pp. 5–12. doi:10.1002/(SICI)1098-108X(199601)19:1<5::AID-EAT2>3.0.CO;2-T.

Zou, H., Yin, P., Liu, L., Liu, W., Zhang, Z., Yang, Y., Li, W., Zong, Q. and Yu, X. (2019) ‘Body-weight fluctuation was associated with increased risk for cardiovascular disease, all-cause and cardiovascular mortality: a systematic review and meta-analysis’, Frontiers in Endocrinology, 10, 728. doi:10.3389/fendo.2019.00728.