15 October 2026 · 14 min read

Slim but Metabolically Unhealthy? Why BMI Misses Hidden Fat and Which Markers Pick It Up

Executive summary

  • In 1981, the concept of the metabolically obese, normal-weight individual was introduced to refer to those who appear slim according to standard weight charts but exhibit metabolic irregularities typically associated with obesity (Ref 1, Ref 2).
  • US survey data from cross-sectional studies indicate that roughly one-quarter of normal-weight adults exhibit metabolic abnormalities according to a specific criterion (Ref 3).
  • When examining aggregated data from observational studies, those with normal weight who exhibited these abnormalities faced a comparable increased risk of death and cardiovascular issues to obese individuals with the same abnormalities, whereas metabolically healthy obesity maintained a lower, yet genuine, long-term risk (Ref 4).
  • The hypothesis of adipose tissue expandability, which is partially based on findings from animals and cell studies, suggests that when an individual's capacity to store fat is exceeded, lipids might accumulate in the liver and muscle (Ref 5).
  • An MRI study revealed that body mass index is an unreliable indicator of abdominal or liver fat, and identified individuals who appear slim externally but have internal fat (Ref 6).
  • Physiological insulin resistance, acting as a protective mechanism where cells resist glucose uptake, redirects carbohydrate from muscle storage towards liver fat production, as indicated by a small study involving young, lean adults (Ref 7).
  • The triglyceride-to-HDL ratio, waist-to-height ratio and fasting insulin can screen for issues without imaging, although the established thresholds are largely based on findings from overweight subjects and provide limited discrimination (Ref 8, Ref 9, Ref 10).
  • In a particular study, 70 per cent of individuals with type 2 diabetes and a body mass index ranging from 21 to 27 achieved remission through weight loss, with the focus on weight reduction rather than limiting carbohydrates (Ref 11).

I need to disclose my initial position and financial stake right away. I operate a private clinic specialising in cosmetic, metabolic, and anti-ageing treatments, and I provide consultations on a paid basis. My usual dietary choice and recommendation is an animal-sourced, carnivore approach, while considering carbohydrate-induced liver fat production as key to the issue. The way I interpret the literature is influenced by that perspective.

I attempted to incorporate the findings that challenge my perspective. There is no controlled study, to my knowledge, examining a strict animal-based diet in lean individuals who are metabolically unhealthy. The data available leans mainly on observational groups, cross-sectional scans, and limited mechanistic research, with almost all of it indicating correlation instead of causation.

The metabolically unhealthy normal-weight phenotype

Metabolic health isn't assured by having a normal body mass index. The term was first introduced in a 1981 hypothesis paper that suggested certain non-obese individuals, as defined by standard weight tables, may exhibit elevated insulin levels and could benefit from reduced food intake (Ref 1). The article provided no clinical information. The same first author’s 1998 review indicated that these individuals likely sit at one extreme of the spectrum termed the insulin resistance syndrome by the authors. The condition I refer to is physiological insulin resistance, which acts as a protective mechanism where cells no longer allow glucose entry upon receiving insulin's signal. The authors suggested that individuals in this group might explain the increased occurrence of type 2 diabetes and cardiovascular disease in adults with a body mass index of 20 to 27 who accumulate a moderate amount of fat, ranging from 2 to 10 kg, as adults, also identifying central fat distribution, inactivity, and low aerobic fitness as contributing factors (Ref 2).

A study of the United States National Health and Nutrition Examination Survey revealed that 23.5 per cent of adults with a normal weight exhibited two or more metabolic irregularities, including high triglyceride levels, elevated fasting glucose or low HDL (Ref 3). The survey's findings, based on a cross-sectional design, indicate that approximately one-quarter of normal-weight adults in the studied population exhibited at least two metabolic irregularities. The number is contingent upon the chosen criteria for abnormality, and the research does not demonstrate that these markers result in disease.

The review systematically assessed all-cause mortality and cardiovascular incidents across eight studies involving 61,386 participants through a meta-analysis. Metabolically unhealthy normal-weight individuals showed a relative risk of 3.14 when contrasted with metabolically healthy normal-weight counterparts, akin to the 2.65 observed in the metabolically unhealthy obese group, featuring overlapping confidence intervals (Ref 4). An analysis combining data from five U.S. adult cohorts aged 40 and above with newly diagnosed diabetes showed that the 12 per cent who were of normal weight at diagnosis exhibited roughly twice the adjusted total mortality rate compared to those who were overweight or obese, with a hazard ratio of 2.08 (Ref 12). These results are based on observation; factors such as smoking, prior illness, or unintentional weight loss may account for the increased mortality in the normal-weight group.

The same meta-analysis further demonstrated that metabolically healthy obese individuals maintained a relative risk of 1.24 for adverse events when only studies with ten or more years of follow-up were included (Ref 4). This mainstream evidence supports the notion that even with normal blood markers, excess weight poses some risk.

Where the fat is stored rather than how much

The risk might not be entirely accounted for by total weight alone, so the specific areas where fat is stored could be significant. In the United Kingdom Prospective Diabetes Study cohort, which included individuals diagnosed in the 1970s and 1980s, the median body mass index was 28, with a third of participants having a body mass index under 25 at the time of diagnosis (Ref 13). The authors suggested that an unmeasured personal fat threshold, unrelated to body mass index, exists beyond which type 2 diabetes is probable.

The adipose tissue expandability hypothesis is another related suggestion. The hypothesis posits that genetic and environmental influences determine a person's upper limit for adipose tissue expansion. When this threshold is exceeded, lipids start to build up in additional tissues, where the authors suggest that harm results, including the condition termed insulin resistance by the authors, which I interpret as physiological insulin resistance (Ref 5). The review is based in part on studies involving animals and cells but does not provide any measurements of human effects. Imagine a suitcase varying in size for each individual; when it reaches capacity, surplus fuel could be stored in the liver, the muscles, and surrounding the abdominal organs.

Magnetic resonance imaging can pinpoint the location of the fat, while it does not evaluate the hypothesis. A cross-sectional study involving 477 white participants at Imperial College London revealed significant differences in intra-abdominal fat and liver lipid levels, which were not entirely accounted for by standard obesity indicators. The researchers identified and named a new subphenotype they called thin on the outside, fat on the inside (Ref 6). The research documented no clinical incidents and was restricted to white participants, yet it indicates that a normal body mass index might coexist with significant internal fat.

The meta-analysis of 93 studies, involving over ten million individuals, revealed that 19.2 per cent of those with non-alcoholic fatty liver disease had a lean body composition (Ref 14). The study's authors argued that screening for fatty liver disease should not rely solely on obesity. The studies and countries used different definitions of lean, leading to high heterogeneity, yet the results imply that fatty liver is not uncommon among slim individuals.

Carbohydrate, muscle and the liver

The way in which muscle and liver process carbohydrate might contribute to the answer. A study at Yale, controlled and physiological in nature, contrasted young, lean adults whom the authors categorised as insulin resistant (a condition I refer to as physiological insulin resistance) with a comparable group in terms of age, weight, BMI, and activity level. The insulin-resistant group exhibited a net muscle glycogen synthesis roughly 60% lower following two high-carbohydrate mixed meals, whereas hepatic de novo lipogenesis and hepatic triglyceride synthesis were each more than double (Ref 7). The insulin-resistant group exhibited elevated plasma triglyceride levels alongside reduced HDL, a lipoprotein indicator, whereas intra-abdominal fat volume showed no difference between the groups.

Muscle tissue resists holding onto the carbohydrate, so it moves to the liver and is converted into fat, according to the authors. This small study, which only included young, lean subjects and two meals instead of a habitual diet, does not indicate the effects of a different diet. This study indicates how carbohydrate could transform into liver fat when muscle refuses to store it. This is my interpretation of the process, and I am open to revising it if conflicting evidence arises.

Those with elevated liver fat in a distinct isotope study involving 24 individuals with metabolic syndrome or elevated liver enzymes exhibited a rate of new fatty acid production over three times higher than those with lower liver fat, whereas the release of free fatty acids from adipose tissue showed no difference across the groups (Ref 15). This small study, which looked at associations rather than causation, omitted body mass index details in its summary and did not examine carbohydrate or fructose consumption, yet aligns with the previously described mechanism. In my opinion, fructose-rich foods pose significant issues in this context, despite the absence of testing in this paper.

Physical activity alters the situation. In a randomised cross-over study involving 12 young, lean adults with physiological insulin resistance, researchers provided a carbohydrate-rich meal both at rest and following a single 45-minute exercise session. Following physical activity, the rate of net muscle glycogen synthesis in the postprandial state increased more than threefold, while hepatic de novo lipogenesis decreased by approximately 30% (Ref 16). This trial had an extremely limited sample size and demonstrated just a short-term impact. This study does not demonstrate whether the findings apply to endurance athletes who regularly consume high amounts of carbohydrates over extended periods.

Skin signs as a possible early clue

The skin occasionally appears to offer an initial indication of metabolic shifts in my clinic. The scientific literature does not strongly endorse my observed pattern. A suggested indicator is the acrochordon, also known as a skin tag. An Indian hospital case-control study matched 110 patients with at least two skin tags against 110 controls of similar age and sex. Among the participants, 58 of the patients received a diagnosis of overt diabetes, compared to 12 of the controls, while the patients exhibited notably elevated levels of fasting glucose, triglycerides, and blood pressure (Ref 17).

The paper proposed that skin tags could indicate metabolic syndrome in the skin, although the control group was only matched for age and sex, and the patient group had a higher body mass index, meaning the study cannot establish a link separate from weight or prove that skin tags serve as a marker for slim individuals specifically.

The topic of acne arises when talking about insulin and carbohydrate intake. In a randomised dietary study involving 43 male acne patients aged 15 to 25, researchers contrasted the effects of a low-glycaemic-load diet against a control diet that focused on carbohydrate-dense foods. In the low-glycaemic-load group, total lesion counts dropped by 23.5 over 12 weeks, compared with a reduction of 12.0 in the control group, and this difference was statistically significant (Ref 18).

The study involved a limited group of young males who weren't chosen for their normal weight, and since the experimental group experienced greater weight reduction, the diet's impact cannot be distinguished from the weight loss. The diet tested in the experiment had moderate carbohydrate content and a low glycaemic load, rather than being low-carbohydrate or a carnivore diet. The findings might suggest that reducing glycaemic load helps with acne, but they don't demonstrate that cutting carbohydrates eliminates acne in individuals who are slim.

Screening without a scan

Routine use of imaging is cost-prohibitive. A meta-analysis and systematic review of 31 studies involving over 300,000 adults showed that the waist-to-height ratio enhanced the identification of cardiometabolic risk factors by 4 to 5 percentage points relative to body mass index (Ref 8). The same research team previously identified an average threshold of 0.50 for both genders and proposed the public health guideline "keep your waist circumference to less than half your height" (Ref 9).

In that previous review, the average area under the curve was slightly higher for the waist-to-height ratio (0.704) compared to body mass index (0.671). The results from these cross-sectional studies don't prove that all individuals with a normal weight but a waist-to-height ratio exceeding 0.5 have fatty liver.

Another dimension is provided by blood markers. In a cross-sectional analysis involving 258 overweight participants without diabetes, researchers employed a direct insulin suppression test to detect physiological insulin resistance. The fasting insulin concentration, along with the triglyceride concentration and the triglyceride-to-HDL ratio, proved to be the most helpful indicators (Ref 10). The published ideal thresholds were 1.47 mmol/L for triglycerides, 1.8 mmol/L (3.0 mg/dL) for the triglyceride-to-HDL ratio, and 109 pmol/L for fasting insulin.

Significant conditions apply to these thresholds. The study involved overweight volunteers from one research centre, meaning the cut-offs might not apply to individuals who are slim. The markers identified between 57% and 67% of affected individuals correctly but missed many, while 68% to 85% of unaffected individuals were accurately excluded, leading to incorrect flags for some. In my own practice, I aim for a triglyceride-to-HDL ratio (in mmol/L) under 1, with fasting insulin significantly below the published cut-off. My clinical preferences are not established thresholds derived from this paper or any other cited here.

Reversibility through weight loss

The possibility of reversing this phenotype hinges primarily on findings from individuals diagnosed with type 2 diabetes. The Newcastle research programme was the subject of a narrative review examining the testing of the personal fat threshold hypothesis. Weight loss achieved in stages among individuals with type 2 diabetes and a body mass index ranging from 21 to 27 resulted in remission for 70 per cent (Ref 11).

This indicates that the condition might be reversible, although the evidence is highly limited. The review's author heads the programme as the chief investigator while also participating as a co-author of the personal fat threshold study, linking the two as part of a single research line rather than separate validation. This study involved participants with diabetes, thus it does not provide direct evidence for lean individuals without diabetes. Weight loss was the intervention, not limiting carbohydrates. The study did not evaluate my personal diet plan, with remission determined through glucose levels instead of long-term cardiovascular results.

What this means in practice

This review indicates that fat accumulated in the liver and abdominal region might have metabolic implications, with physiological insulin resistance potentially redirecting consumed carbohydrates towards liver fat deposition instead of muscle storage, despite much of the supporting evidence being observational or derived from limited mechanistic studies.

According to my perspective, straightforward and inexpensive indicators like the waist-to-height ratio, the triglyceride-to-HDL ratio, liver enzymes, and a fasting insulin level offer a more comprehensive understanding than relying solely on weight. Your individual situation may prompt a lifestyle reassessment with your clinician if test results are abnormal. In lean individuals, weight loss trials have been the sole focus, limited to those with type 2 diabetes, yet adopting practices like emphasising animal protein and fat, exposure to morning sunlight, safeguarding sleep, and engaging in resistance exercise can be sensible measures to explore. Long-term dietary trials in this slim phenotype have not been conducted, and I am unable to see you through the screen.

Disclosures

I operate a private practice specialising in cosmetic, metabolic, and anti-ageing medicine, providing consultations for a fee. I thus stand to gain financially if readers benefit from this analysis. I prefer and typically advise consuming animal-based, higher-protein meals, which inclines me to support carbohydrate restriction. I have attempted to mitigate that bias through pointing out that the studies in this evidence base focused on weight loss instead of carbohydrate restriction, incorporating the data indicating that metabolically healthy obesity still poses long-term risk, and highlighting instances where two references originate from the same research group.

The content here should not be considered tailored medical guidance, and it cannot take the place of personal medical advice from your doctor. If consultations are something you're interested in for applying this to your personal blood work or situation, note that I provide them, though numerous skilled clinicians already cover this area extensively and there's no obligation to visit me.

References

  1. Ruderman NB, Schneider SH, Berchtold P. The "metabolically-obese," normal-weight individual. Am J Clin Nutr, 1981. PMID 7270486. https://doi.org/10.1093/ajcn/34.8.1617
  2. Ruderman N, Chisholm D, Pi-Sunyer X, et al. The metabolically obese, normal-weight individual revisited. Diabetes, 1998. PMID 9588440. https://doi.org/10.2337/diabetes.47.5.699
  3. Wildman RP, Muntner P, Reynolds K, et al. The obese without cardiometabolic risk factor clustering and the normal weight with cardiometabolic risk factor clustering: prevalence and correlates of 2 phenotypes among the US population (NHANES 1999-2004). Arch Intern Med, 2008. PMID 18695075. https://doi.org/10.1001/archinte.168.15.1617
  4. Kramer CK, Zinman B, Retnakaran R. Are metabolically healthy overweight and obesity benign conditions? A systematic review and meta-analysis. Ann Intern Med, 2013. PMID 24297192. https://doi.org/10.7326/0003-4819-159-11-201312030-00008
  5. Virtue S, Vidal-Puig A. Adipose tissue expandability, lipotoxicity and the Metabolic Syndrome: an allostatic perspective. Biochim Biophys Acta, 2010. PMID 20056169. https://doi.org/10.1016/j.bbalip.2009.12.006
  6. Thomas EL, Parkinson JR, Frost GS, et al. The missing risk: MRI and MRS phenotyping of abdominal adiposity and ectopic fat. Obesity (Silver Spring), 2011. PMID 21660078. https://doi.org/10.1038/oby.2011.142
  7. Petersen KF, Dufour S, Savage DB, et al. The role of skeletal muscle insulin resistance in the pathogenesis of the metabolic syndrome. Proc Natl Acad Sci U S A, 2007. PMID 17640906. https://doi.org/10.1073/pnas.0705408104
  8. Ashwell M, Gunn P, Gibson S. Waist-to-height ratio is a better screening tool than waist circumference and BMI for adult cardiometabolic risk factors: systematic review and meta-analysis. Obes Rev, 2012. PMID 22106927. https://doi.org/10.1111/j.1467-789X.2011.00952.x
  9. Browning LM, Hsieh SD, Ashwell M. A systematic review of waist-to-height ratio as a screening tool for the prediction of cardiovascular disease and diabetes: 0.5 could be a suitable global boundary value. Nutr Res Rev, 2010. PMID 20819243. https://doi.org/10.1017/S0954422410000144
  10. McLaughlin T, Abbasi F, Cheal K, Chu J, Lamendola C, Reaven G. Use of metabolic markers to identify overweight individuals who are insulin resistant. Ann Intern Med, 2003. PMID 14623617. https://doi.org/10.7326/0003-4819-139-10-200311180-00007
  11. Taylor R. Aetiology of type 2 diabetes: an experimental medicine odyssey. Diabetologia, 2025. PMID 40316731. https://doi.org/10.1007/s00125-025-06428-0
  12. Carnethon MR, De Chavez PJD, Biggs ML, et al. Association of weight status with mortality in adults with incident diabetes. JAMA, 2012. PMID 22871870. https://doi.org/10.1001/jama.2012.9282
  13. Taylor R, Holman RR. Normal weight individuals who develop type 2 diabetes: the personal fat threshold. Clin Sci (Lond), 2015. PMID 25515001. https://doi.org/10.1042/CS20140553
  14. Ye Q, Zou B, Yeo YH, et al. Global prevalence, incidence, and outcomes of non-obese or lean non-alcoholic fatty liver disease: a systematic review and meta-analysis. Lancet Gastroenterol Hepatol, 2020. PMID 32413340. https://doi.org/10.1016/S2468-1253(20)30077-7
  15. Lambert JE, Ramos-Roman MA, Browning JD, Parks EJ. Increased de novo lipogenesis is a distinct characteristic of individuals with nonalcoholic fatty liver disease. Gastroenterology, 2014. PMID 24316260. https://doi.org/10.1053/j.gastro.2013.11.049
  16. Rabol R, Petersen KF, Dufour S, Flannery C, Shulman GI. Reversal of muscle insulin resistance with exercise reduces postprandial hepatic de novo lipogenesis in insulin resistant individuals. Proc Natl Acad Sci U S A, 2011. PMID 21808028. https://doi.org/10.1073/pnas.1110105108
  17. Shah R, Jindal A, Patel NM. Acrochordons as a cutaneous sign of metabolic syndrome: a case-control study. Ann Med Health Sci Res, 2014. PMID 24761238. https://doi.org/10.4103/2141-9248.129040
  18. Smith RN, Mann NJ, Braue A, Makelainen H, Varigos GA. A low-glycemic-load diet improves symptoms in acne vulgaris patients: a randomized controlled trial. Am J Clin Nutr, 2007. PMID 17616769. https://doi.org/10.1093/ajcn/86.1.107