Dr. Yulia Lurye
← All articles Obesity

Remission of obesity: lifelong therapy, or restored body composition and metabolic flexibility?

20 Oct 2026

Can you imagine a time when obesity is in remission? And if so, what would that remission look like?

Would it be pharmacologically controlled, with patients on some form of lifelong therapy to maintain a normal weight? Or would repeated courses of treatment be sufficient to reset energy balance, with long treatment holidays in between?

The newest incretin-based medicines address excessive caloric intake through appetite suppression, and some of the new co-agonists are also claimed to increase energy expenditure. Imagine appetite inhibition tailored to quieten food noise without inducing anhedonia, and energy expenditure raised just enough to offset the metabolic adaptation that accompanies weight loss. Would anything remain for metabolic agents to treat?

These are the questions I am curious about. To categorise the effects of metabolic drugs, I divided energy balance into four quadrants.

Does each quadrant deserve a distinct therapeutic approach, that is, a targeted drug, and could this become the foundation of a four-pillar therapy as in heart failure? Or will a more precisely chosen monotherapy, matched to an individual’s biology, prove the better fit? I suspect most patients would prefer to take one drug, and ideally to stop it one day.

Energy balance is often presented as a two-arm mechanism in which energy storage must be matched by energy expenditure. I divide each arm into an upper and a lower quadrant (Figure 1).

The upper quadrants describe disorders of quantity: too much energy stored, or too much dissipated. The lower quadrants describe disorders of regulation: energy stored in the wrong compartment, or an organism that can no longer raise its expenditure on demand.

  • 1. Excessive energy storage: high total fat mass, the mechanical burden of obesity, or “fat mass disease”.
  • 2. Impaired energy partitioning: ectopic and visceral fat deposition, or “sick fat disease”.
  • 3. “Fast metabolism”: increased energy expenditure, through energy dissipation, thermogenesis and hypermetabolic states.
  • 4. “Slow metabolism”: reduced energy expenditure, through a lowered resting metabolic rate or a lost capacity to accelerate.

At the centre of the diagram sits metabolic flexibility itself, which I define as the inducibility of both arms of energy balance, together with the tonic and episodic control of energy state. Metabolic flexibility is a central part, as it is a sort of a steering wheel and am accelerator, regulating the inducibility of both arms of energy balance and the tonic and episodic control of energy state.

1. Excessive energy storage: “fat mass disease”

Part of the morbidity of obesity is a direct consequence of the amount of adipose tissue. Bray, Kim and Wilding termed this “fat mass disease” and contrasted it with “sick fat disease”, the morbidity that arises from dysfunctional adipose tissue, also known as adiposopathy.

Whether fat mass carries risk independently of metabolic dysfunction is best tested in people with so-called metabolically healthy obesity: obesity without diabetes, hypertension or treated dyslipidaemia. In a cohort of 3.5 million adults in the UK, this group still had a higher incidence of heart failure (HR 1.96) and coronary heart disease (HR 1.49) than metabolically healthy individuals of normal weight (Caleyachetty 2017).

The complications driven predominantly by the quantity of fat fall into three groups.

  1. Mechanical. Obstructive sleep apnoea and osteoarthritis of weight-bearing joints are the classical examples. Excess visceral fat adds raised intra-abdominal pressure, which contributes to gastro-oesophageal reflux, abdominal hernia, elevation of the diaphragm with reduced lung volumes, and venous congestion. Expansion of organ-associated fat within confined spaces acts in the same way: epicardial fat increases pericardial restraint in the obese phenotype of heart failure with preserved ejection fraction (Obokata 2017), and renal sinus fat compresses hilar vessels and lymphatics, raising interstitial pressure and promoting sodium retention; in the Framingham cohort it was associated with hypertension and chronic kidney disease (Foster 2011).
  2. Haemodynamic. A larger body requires a larger circulating volume and higher cardiac output. Over time this volume load remodels the left ventricle and increases the risk of heart failure, a picture described as obesity cardiomyopathy. In the kidney, the same load produces glomerular hyperfiltration, which progresses to obesity-related glomerulopathy.
  3. Endocrine. Adipose tissue is an endocrine organ in its own right. Aromatase in adipose stromal cells converts androgens to oestrogens, which contributes to endometrial cancer, the malignancy with the strongest association with BMI (Lauby-Secretan 2016), and to postmenopausal breast cancer. In men, the same conversion, together with the effects of leptin on the hypothalamic–pituitary–gonadal axis, causes obesity-related secondary hypogonadism. Hypogonadism in turn lowers resting energy expenditure, so excessive storage in quadrant 1 feeds directly into quadrant 4: one example of how the quadrants, and their pathomechanisms, are interconnected.

Excess visceral fat also contributes to systemic insulin resistance and compensatory hyperinsulinaemia, which sets off several vicious cycles. This is where quadrant 1 begins to merge into quadrant 2.

2. Impaired energy partitioning: “sick fat disease”

The classical metabolic complications of obesity, namely type 2 diabetes, atherogenic dyslipidaemia and metabolic dysfunction-associated steatotic liver disease (MASLD), depend more on where fat is stored than on how much of it there is. People of Asian ancestry develop these diseases at a lower BMI. In familial partial lipodystrophy (for example FPLD2, caused by LMNA variants), subcutaneous fat of the limbs and gluteal region is lost while visceral fat is preserved, and patients develop severe insulin resistance, MASLD and hypertriglyceridaemia at a normal BMI. Conversely, “favourable adiposity” alleles are associated with higher body fat but a lower risk of type 2 diabetes, hypertension and heart disease, because the additional fat is stored subcutaneously (Yaghootkar 2016).

Three mechanisms of impaired partitioning can be distinguished.

2a. Lipotoxic: lipid overload in parenchymal cells

Lipid accumulates inside cells beyond their capacity to store and oxidise it: hepatocytes, skeletal myocytes, cardiomyocytes, pancreatic β-cells, renal tubular cells and podocytes. Triglyceride itself is a relatively inert storage form; the injury is mediated by bioactive intermediates. Diacylglycerols activate protein kinase Cε in the liver and PKCθ in skeletal muscle, impairing insulin signalling (Samuel and Shulman 2012), while ceramides and free cholesterol induce endoplasmic reticulum stress, mitochondrial oxidative stress and apoptosis. Esterification into triglyceride droplets is in fact partly protective (Listenberger 2003), which is consistent with the “athlete’s paradox” of high intramyocellular lipid with excellent insulin sensitivity.

De novo lipogenesis contributes to this overload, but differently by organ. In the liver it is a normal pathway, yet in MASLD it supplies about 26% of hepatic triglyceride (Donnelly 2005). In the human diabetic kidney, by contrast, lipogenic genes are downregulated, and lipid accumulates because uptake increases while fatty acid oxidation and cholesterol efflux fail (Herman-Edelstein 2014; Kang 2015).

2b. Adipogenic: adipocytes derived from local progenitors

Some organs generate adipocytes from their own progenitor cells. In skeletal muscle, PDGFRα⁺ fibro-adipogenic progenitors differentiate into intermuscular adipocytes (Uezumi 2010; Joe 2010); intermuscular fat expands with obesity, ageing and inactivity and correlates with insulin resistance and reduced strength. In the myocardium, fibro-adipocyte progenitors produce the fibrofatty replacement of arrhythmogenic cardiomyopathy (Lombardi 2016). Marrow adipose tissue and pancreatic steatosis follow the same principle. This route causes harm mainly through paracrine inflammation and replacement of functional parenchyma, rather than direct lipotoxicity.

2c. Redistributive: visceral and organ-associated fat

Visceral adipose tissue is part of normal anatomy. It expands with a positive energy balance (quadrant 1), but it also gains fat at the expense of subcutaneous depots when partitioning fails. A positive energy balance is required to increase total fat mass; in its absence, the distribution of fat can still change.

Hypercortisolism is the classical example: central and visceral adiposity increases while limb fat and skeletal muscle are lost. Omental fat expresses more glucocorticoid receptors and more 11β-hydroxysteroid dehydrogenase type 1, which regenerates active cortisol locally, prompting Bujalska, Kumar and Stewart to describe central obesity as “Cushing’s disease of the omentum” (1997). The same shift accompanies the menopausal transition (Lovejoy 2008), low testosterone and androgen deprivation therapy, ageing, and older antiretroviral regimens for HIV. Thiazolidinediones are the classical counter-example, promoting adipogenesis predominantly in subcutaneous fat.

Once redistributed, visceral and organ-associated fat causes harm through two routes, in addition to the mechanical compression described in quadrant 1.

Can adipose tissue be reduced more selectively?

Should we target the more harmful visceral fat specifically and spare subcutaneous fat? Or is the proportion of subcutaneous, visceral and truncal fat loss achieved with GLP-1 receptor agonists already good enough? In other words, shall we be fussy about distinguishing lightness and leanness, and figuring out which of the two matters more for overall survival?

3. “Fast metabolism”: increased energy expenditure

Drugs and diseases increase energy expenditure through two distinct mechanisms: increased ATP demand, and reduced efficiency of ATP production.

3a. Increased demand

When more work is performed, more substrate is oxidised. Thyrotoxicosis is the classical hypermetabolic state, increasing ATP consumption by Na⁺/K⁺-ATPase and SERCA. Burns, major trauma and sepsis are the classical examples in critical care; after extensive burns, hypermetabolism can persist for months. In fever, expenditure rises by approximately 10–13% for each 1 °C, and phaeochromocytoma, with its catecholamine excess, is a natural model of sympathetic drive.

Skeletal muscle, the principal effector of the inducible component of energy expenditure, can itself be overactivated: shivering is the physiological example, and malignant hyperthermia, in which calcium is released uncontrollably through the RYR1 channel, the pathological extreme.

Sympathomimetic drugs act on this route. They do increase energy expenditure, but the effect is modest and invariably accompanied by higher heart rate and blood pressure, a trade-off that has repeatedly ended their use in obesity. Ephedra was banned in the United States in 2004 on cardiovascular grounds, and sibutramine, partly thermogenic, was withdrawn in 2010 after the SCOUT trial showed an excess of non-fatal myocardial infarction and stroke. Caffeine and nicotine are the everyday examples; the loss of nicotine’s thermogenic effect is one reason for weight gain after smoking cessation.

3b. Reduced efficiency: mitochondrial uncoupling

In uncoupling, protons leak back across the inner mitochondrial membrane without passing through ATP synthase, and the energy of the proton gradient is released as heat, much as water bypassing the turbines of a dam generates no electricity.

The physiological uncoupler is uncoupling protein 1 (UCP1) in brown and beige adipose tissue, the thermogenic counterpart of the storage adipocyte, activated by cold and by sympathetic stimulation. A single 200 mg dose of mirabegron, a β3-adrenoceptor agonist, activated human brown adipose tissue and raised resting metabolic rate by about 200 kcal/day, though it also increased heart rate and systolic blood pressure (Cypess 2015). Skeletal muscle has an analogous mechanism: sarcolipin uncouples SERCA and generates heat without shivering (Bal 2012). Luft disease, described in 1962 in a euthyroid patient with severe hypermetabolism and “loosely coupled” mitochondria, was the first mitochondrial disease to be characterised and remains nature’s model of uncoupling. Thyrotoxicosis also increases proton leak, so it acts through both 3a and 3b.

Pharmacological uncouplers have a troubled history. 2,4-dinitrophenol (DNP) was widely prescribed for weight loss in the 1930s and abandoned after deaths from hyperthermia and cases of cataract; it is still sold illicitly, and fatalities continue to be reported (Grundlingh 2011). Salicylate poisoning offers a familiar example from clinical toxicology. The modern clinical example is HU6, a DNP prodrug described by its developer as a “controlled metabolic accelerator”. In the phase 2a HuMAIN trial in obesity-related heart failure with preserved ejection fraction, HU6 reduced body weight, total fat mass and visceral fat over 19 weeks without a significant loss of lean mass (JAMA Cardiology, 2025).

Among the newest agents, glucagon receptor agonism increases energy expenditure in human infusion studies, which forms part of the rationale for co-agonists such as retatrutide and survodutide. The extent to which their weight loss reflects increased expenditure rather than reduced intake has not been established.

This history explains why the expenditure arm came to be regarded as the less safe therapeutic path: successive attempts to accelerate metabolism in order to “burn” excess fat encountered hyperthermia, tachycardia and hypertension.

4. “Slow metabolism”: reduced energy expenditure

Energy expenditure has two components, each of which can decline independently: a tonic floor and an inducible ceiling.

4a. A lowered floor

The tonic floor is the resting metabolic rate. It is determined mainly by fat-free mass, and in particular by the small but metabolically very active organs (brain, liver, heart and kidneys), and is modulated by thyroid hormone and sympathetic tone; individuals with low resting muscle sympathetic nerve activity may be predisposed to weight gain through a lower metabolic rate (Spraul 1993).

A lowered floor represents true metabolic slowing. Hypothyroidism and hypogonadism are the endocrine examples. Adaptive thermogenesis after weight loss is the most relevant example for obesity therapy: energy expenditure falls below the value predicted from body composition (Rosenbaum and Leibel 2010). Physical inactivity, by contrast, reduces the total amount of energy expended but not the metabolic rate per unit of tissue.

4b. A lost ceiling

The inducible ceiling is the extent to which expenditure can rise on demand: exercise capacity, diet-induced thermogenesis and cold-induced thermogenesis. Its principal effector is skeletal muscle, and its loss is above all a loss of muscle: sarcopenia, disuse atrophy, heart failure myopathy, and muscle wasting in chronic kidney disease. Brown adipose tissue contributes too; its activity declines with ageing and is lower in obesity (van Marken Lichtenbelt 2009), although its quantitative contribution to adult energy expenditure remains debated.

Bed rest shows how quickly the ceiling falls and then draws the floor down with it. Activity-related expenditure declines immediately; fat-free mass follows, reducing resting expenditure over time. Physical inactivity alone is sufficient to induce metabolic inflexibility, with impaired switching between fat and glucose oxidation, ectopic lipid accumulation and insulin resistance (Bergouignan 2011).

Heart failure shows that the two components can move in opposite directions. Resting metabolic rate is about 18% higher in patients with heart failure than in healthy controls (Poehlman 1994), reflecting the increased work of breathing, sympathetic activation and circulating inflammatory cytokines. Total daily energy expenditure measured by doubly labelled water is nevertheless lower, because physical activity and peak oxygen consumption are markedly reduced (Toth 1997). Heart failure therefore belongs to quadrant 3 at rest and to quadrant 4 overall, as an example of the lost ceiling.

Muscle wasting in chronic kidney disease requires a distinction between two terms that often appear together. A hypermetabolic state is one of increased energy expenditure; a catabolic state is one of net tissue breakdown, predominantly of protein. Burns, sepsis and cancer cachexia are both. Renal acidosis is catabolic but not hypermetabolic. Metabolic acidosis activates muscle proteolysis through the ATP-dependent ubiquitin–proteasome pathway, by a glucocorticoid-dependent mechanism (May 1986; Bailey 1996), yet resting energy expenditure in non-dialysed chronic kidney disease is lower than in matched controls, approximately 1,325 versus 1,448 kcal/day, even after adjustment for lean body mass (Avesani 2004).

Can lost function be restored?

Is there, then, a loss of function in this arm of energy balance, and can it be safely restored? In many chronic diseases the answer to the first question appears to be yes. The therapeutic goal should not be to accelerate metabolism beyond normal, but to restore a lowered floor and a lost ceiling: to rebuild skeletal muscle, correct acidosis and hormonal deficiencies, and recover the thermogenic response to food, cold and exercise. That is a different approach from the historical use of uncoupling agents, and potentially a far safer one.

Remission as restored metabolic flexibility

The term “metabolic inflexibility” was introduced to describe skeletal muscle: the failure to switch from fatty acid to glucose oxidation in response to insulin (Kelley and Mandarino 2000). I use it more broadly, for the inducibility of both arms of energy balance and the tonic and episodic control of energy state. The upper quadrants represent disorders of quantity and the lower quadrants disorders of regulation and responsiveness; metabolic flexibility, at the centre of the diagram, connects them all.

This returns me to the questions with which I began. If incretin-based therapies control energy intake, and perhaps part of the adaptive response to weight loss, the remaining therapeutic territory lies in the lower quadrants: preferential reduction of visceral adiposity with preservation of skeletal muscle mass, and restoration of the capacity to increase energy expenditure on demand. Remission of obesity might then be defined by restored metabolic flexibility, not merely by a normal body weight.

References