The study enrolled 12 women with anorexia nervosa and 13 healthy female controls, all aged between 18 and 30 years. Participants with anorexia had a body mass index below 17.5 kilograms per square metre, a disease duration ranging from one to eleven years, and no lifetime history of binge eating. Each participant attended three visits within 28 days: a screening examination, a combined PET/CT scan, and a whole-body MRI. The centrepiece of the investigation was the hyperinsulinemic, euglycemic clamp, a technique considered the gold standard for measuring insulin sensitivity. Insulin was infused at a carefully calibrated dose while glucose was delivered at a variable rate to keep plasma glucose steady at 5.0 millimoles per litre. Roughly 75 minutes into the clamp, once the metabolic steady state had been reached, the researchers injected 150 megabecquerels of the fluorinated glucose tracer and captured dynamic images of the thorax and brain, followed by static frames of the neck, abdomen, and thighs.

By comparing the rate at which radioactive glucose accumulated in each tissue with the concentration of tracer in the plasma, the team calculated tissue-specific glucose uptake in micromoles per 100 grams of tissue per minute. The volumes of interest covered the right lobe of the liver, the quadriceps femoris muscles, the left ventricular myocardium, subcutaneous adipose tissue at the waistline and mid-thigh, supraclavicular brown adipose tissue, humeral shaft bone marrow, and the duodenal wall. Whole-body MRI with a Dixon sequence then quantified visceral and subcutaneous fat volumes and tissue fat fractions, allowing the metabolic data to be interpreted in the context of body composition.

The whole-body results confirmed what earlier clamp studies had suggested: anorexia nervosa is not associated with improved insulin sensitivity, despite the characteristically low fasting insulin levels and strikingly favourable HOMA-IR values seen in these patients. When expressed relative to total body weight, insulin-stimulated glucose disposal was actually higher in the women with anorexia, but this advantage disappeared when the same values were referenced to fat-free mass, indicating that the apparent improvement simply reflects the smaller metabolic machinery of a lighter body. Endogenous glucose production during hyperinsulinaemia, a measure of hepatic insulin action, did not differ between the groups. Fasting insulin was less than half the level seen in controls, and the Matsuda insulin sensitivity index derived from an oral glucose tolerance test was markedly higher in the anorexia group, yet these proxies did not translate into genuinely enhanced insulin responsiveness at the tissue level.

The most unexpected tissue-level finding emerged in the intestine. Glucose uptake in the duodenal wall was significantly higher in the women with anorexia nervosa, at 3.0 versus 2.6 micromoles per 100 grams per minute, and this was the only tissue in which a statistically significant difference between the groups was observed. Intriguingly, duodenal glucose uptake correlated negatively with disease duration, suggesting that the intestinal adaptation may weaken as chronic undernutrition progresses, possibly through a reduction in mucosal surface area. The authors speculate that prolonged fasting upregulates nutrient transporters on enterocytes, and that altered enteroendocrine activity, particularly the incretin hormones GLP-1 and GIP, may link intestinal nutrient sensing to central appetite regulation and peripheral insulin sensitivity. Gastrointestinal complications such as delayed gastric emptying and constipation are common in anorexia, and this study provides the first direct evidence that the gut wall is also metabolically active in an altered way.

Perhaps the most striking individual-level observation came from the post hoc analysis of PET images. In five of the twelve women with anorexia, insulin-stimulated glucose uptake in subcutaneous adipose tissue exceeded that in skeletal muscle, a pattern that is essentially never seen in healthy metabolism and that contradicted a larger reference dataset of 326 individuals previously studied at the same centre. In these five participants, adipose tissue glucose uptake reached 5.0 micromoles per 100 grams per minute while skeletal muscle uptake fell to 1.8, and muscle uptake was significantly lower than in the other women with anorexia. Across the entire study group, higher adipose tissue glucose uptake correlated with lower plasma insulin levels during the clamp, which the researchers interpret as evidence of increased insulin clearance. Prolonged fasting is known to induce insulin resistance in skeletal muscle while potentiating insulin binding to adipocytes, and animal studies have shown that repeated cycles of starvation and refeeding upregulate lipogenesis in white adipose tissue, a process that demands a higher glucose supply to the fat depot.

Hormonal profiling added another layer of complexity. Fasting growth hormone levels were similar between the groups, but 60 minutes into the clamp, growth hormone remained elevated in the women with anorexia while it was suppressed in the controls. In the healthy controls, higher growth hormone during hyperinsulinaemia correlated strongly with classic markers of insulin resistance: it was associated with lower whole-body glucose disposal, lower skeletal muscle glucose uptake, higher endogenous glucose production, and higher brain glucose uptake. In the women with anorexia, all of these associations were absent. The researchers attribute this dissociation to the growth hormone resistance and disrupted pulsatile growth hormone secretion that are well established in anorexia nervosa, which drive down IGF-1 levels and blunt the hormone’s peripheral metabolic actions. The finding raises the possibility that a regulatory loop linking cerebral glucose demand to growth hormone secretion, which in turn induces peripheral insulin resistance to spare glucose for the brain, is fundamentally rewired in the disorder.

Brain glucose metabolism, measured during hyperinsulinaemia rather than the fasting state used in earlier PET studies, showed no significant differences between the groups in whole brain, white matter, gray matter, or specific anatomical regions. Previous fasting-state studies had reported globally lowered cerebral glucose uptake in the acute phase of anorexia, with normalization after weight restoration. The authors suggest that elevated free fatty acids and ketone bodies may compete with glucose for brain uptake in the fasted state, and that because hyperinsulinaemia suppresses lipolysis and ketogenesis, these differences diminish under clamp conditions. The slightly older control group, since aging reduces brain glucose uptake during hyperinsulinaemia, may also have masked a true difference in this small sample.

The MRI data revealed distinct patterns of tissue fat distribution. Women with anorexia had dramatically smaller volumes of abdominal and femoral subcutaneous adipose tissue, visceral adipose tissue, and thigh skeletal muscle. Subcutaneous adipose tissue fat fractions were lower in the anorexia group, indicating a higher water content within the fat depots, while femoral bone marrow fat content was higher, a finding consistent with previous reports linking marrow adiposity to impaired bone health in anorexia. Visceral adipose tissue volume was reduced far less than subcutaneous volume, and visceral fat fraction was unchanged, echoing earlier observations that visceral fat is relatively preserved in the disorder. Liver and skeletal muscle fat fractions, markers of ectopic lipid accumulation, were similar between the groups.

The authors acknowledge that the small cohort, recruited initially with a target of 30 participants per group, limits the statistical power and prevents subgroup analyses based on feeding behaviour or exercise patterns. The cross-sectional design cannot establish causality between hormonal changes and tissue glucose metabolism, and because only women were studied, the results may not generalize to men. Nevertheless, the study is the first to examine glucose metabolism in multiple tissues within the same individuals with anorexia nervosa, and it provides a foundation for understanding the interplay between glucose metabolism, hormonal regulation, and organ crosstalk in the disorder. The researchers propose that anorexia nervosa should be understood as a metabo-psychiatric condition in which genetic, epigenetic, behavioural, and hormonal factors produce varying metabolic phenotypes, with some patients showing heightened and others impaired insulin sensitivity in muscle and fat. Of all the factors examined, growth hormone levels during steady hyperinsulinaemia, and their lost association with glucose metabolism in muscle, liver, and brain, were the most distinctive difference between patients and controls, and these insights may ultimately inform more targeted approaches to managing the metabolic consequences of the disorder.

Subject of Research: Physiology

Article Title: Tissue‐specific insulin sensitivity in anorexia nervosa: A cross‐sectional positron emission tomography (PET) study in female adults

Article References: Latva‐Rasku, A., Pak, K., Rebelos, E., Honka, M.-J., Dadson, P., Koffert, J., Tuisku, J., Pekkarinen, L., Laurila, S., Hieta, J., Laitinen, K., Kirjavainen, A. K., Hirvonen, J., Karlsson, H. K., Karukivi, M., Nummenmaa, L., & Nuutila, P. (2026). Tissue‐specific insulin sensitivity in anorexia nervosa: A cross‐sectional positron emission tomography ( PET ) study in female adults. Physiological Reports, 14(19), Article e71142. https://doi.org/10.14814/phy2.71142

Image Credits: AI Generated

DOI: 10.14814/phy2.71142

Keywords: anorexia nervosa, insulin sensitivity, PET imaging, glucose metabolism, growth hormone, metabolism, Physiological Reports, Tissue, specific, insulin, sensitivity, anorexia