Diabetes, Metabolic Syndrome & Insulin Resistance

Type 1 diabetes begins with the immune system attacking the insulin-producing beta cells in the pancreas. In most cases this happens in childhood or young adulthood, though late-onset Type 1 (called LADA, latent autoimmune diabetes in adults) is more common than most general practitioners realize. Without insulin, glucose cannot enter cells, the body begins burning fat in an uncontrolled way, and the resulting ketoacidosis can be fatal within days. Type 1 requires insulin replacement for life. It is not caused by diet or lifestyle, and patients with Type 1 deserve to stop hearing that implication.

Type 2 is a different pathology entirely. It begins with insulin resistance: cells, particularly in muscle, liver, and fat tissue, stop responding normally to insulin. The pancreas compensates by producing more insulin. For years, sometimes decades, this compensation keeps blood glucose in an acceptable range while the metabolic damage quietly accumulates. Eventually the beta cells exhaust their reserve, insulin production falls, and blood glucose rises above the diagnostic threshold. Type 2 accounts for approximately 90 to 95 percent of diabetes cases in the United States (American Diabetes Association, Diabetes Care 2024, DOI: 10.2337/dc24-S002).

The practical distinction for a patient in my clinic: Type 1 is managed primarily with insulin dose titration and requires ongoing specialist involvement. Type 2 is, in its early stages, often reversible with metabolic intervention, and in its later stages manageable with an increasingly powerful array of medications. The conversation about Type 2 should almost never stop at diagnosis. It should start with the fifteen years before diagnosis.

The word "prediabetes" was coined partly as a motivational tool. The logic was that telling someone they had a condition might spur behavior change, whereas telling them they had a borderline lab value would not. The problem is that the word implies a waiting room, a harmless antechamber before the real disease. Clinical research shows that is wrong.

Cardiovascular risk begins rising well before the A1c crosses 6.5 percent, the formal diabetes threshold. In a large observational analysis of UK Biobank data, cardiovascular events and mortality were significantly elevated in individuals with A1c values in the prediabetes range compared to those with normal glucose metabolism (Emerging Risk Factors Collaboration, JAMA 2010, DOI: 10.1001/jama.2010.1060). The relationship between glucose and cardiovascular risk is continuous, not binary. Every increment of A1c above the normal range carries incremental risk, regardless of what we call the zone.

In my clinic, I have stopped using the word prediabetes with patients who are at high cardiovascular risk. I say instead: "Your glucose metabolism is impaired. You are not diabetic yet, but your arterial walls cannot tell the difference." That is not alarmist. It is accurate. The Diabetes Prevention Program, the landmark NIH trial, showed that lifestyle intervention in people with impaired fasting glucose reduced progression to frank diabetes by 58 percent over three years (Knowler et al, NEJM 2002, DOI: 10.1056/NEJMoa012512). The word "prevention" in that trial name is itself a mild misnomer. What the DPP showed is that metabolic reversal is possible, and that the window for intervention is during the stage most clinicians call "pre" and most patients hear as "not yet serious."

The hemoglobin A1c is an elegant test because red blood cells live roughly 90 days and accumulate glycation in proportion to ambient glucose exposure. It is not, however, a complete picture of glucose metabolism. Two people with A1c 5.7 can have very different patterns: one with stable, mildly elevated glucose throughout the day, another with normal fasting glucose and sharp post-meal spikes that are invisible on the A1c. Continuous glucose monitoring studies have shown that post-meal glucose excursions, even when the A1c appears normal or only mildly elevated, correlate with endothelial dysfunction and oxidative stress (Ceriello et al, Diabetologia 2008, DOI: 10.1007/s00125-008-1086-6).

For the heart, the mechanism is multi-pronged. Glucose attaches to proteins in arterial walls through advanced glycation end products (AGEs), making them stiffer and less elastic. Hyperglycemia generates reactive oxygen species that oxidize LDL particles and damage endothelial cells. Insulin resistance, which is almost always present when A1c reaches 5.7, drives triglyceride elevation and HDL reduction, the lipid signature most directly linked to small dense LDL particle formation. Small dense LDL particles are more atherogenic per particle than large buoyant LDL.

A patient with A1c 5.7 and normal cholesterol by standard lipid panel may nonetheless have a cardiovascular risk profile that a CAC score, fasting insulin measurement, and ApoB level would reveal as concerning. The A1c tells one part of the story. The story is bigger than one test.

The word "reverse" deserves precision. What we mean clinically is returning fasting glucose, A1c, and fasting insulin to ranges consistent with normal metabolic function. This is achievable for most people in the prediabetes range, though it is not guaranteed, and some individuals with significant beta cell impairment will find the ceiling lower than they expected.

The DPP lifestyle intervention achieved its results with a 7 percent reduction in body weight and 150 minutes per week of moderate physical activity, primarily walking (Knowler et al, NEJM 2002, DOI: 10.1056/NEJMoa012512). That is not a punishing protocol. What the DPP findings do suggest is that the weight loss matters more than the specific dietary approach used to achieve it. Mediterranean, low-carbohydrate, and plant-based approaches have all shown improvement in metabolic markers; the trial evidence does not conclusively favor one over another for prediabetes reversal specifically.

The ten-year follow-up of the DPP showed that participants who had reverted to normal glucose metabolism during the trial had significantly lower rates of diabetes at year ten than those who had not reverted, even accounting for subsequent weight regain. This implies that the window of reversal, once achieved, confers durable biological benefit beyond what the maintained weight loss alone would predict.

Timeline in clinical practice: I typically see meaningful A1c improvement within three months of consistent dietary change combined with resistance exercise. Full reversal to A1c below 5.7, when it happens, usually occurs between six and twelve months. Some patients achieve it faster with more aggressive carbohydrate restriction. The pace matters less than the direction.

Insulin is the key that opens the cellular door to glucose. In insulin resistance, the lock becomes sticky. The pancreas, sensing that glucose is not clearing properly, manufactures more keys. For years this works. Blood glucose stays in an acceptable range because the pancreas is compensating. But the high circulating insulin itself causes harm: it promotes fat storage in the liver, drives triglyceride synthesis, contributes to hypertension through sodium retention, and stimulates the sympathetic nervous system. You can have insulin resistance for a decade with a normal fasting glucose.

The signs that suggest insulin resistance in a clinical consultation: waist circumference above 35 inches in women and 40 inches in men, triglycerides above 150 mg/dL, HDL below 40 in men or 50 in women, blood pressure drifting above 130/80, fasting glucose above 95 (even if below 100, which is technically "normal"), and fatigue after meals, particularly carbohydrate-heavy meals. None of these individually is diagnostic. Together they tell a story.

The best single test in standard practice is the HOMA-IR calculation, which uses fasting glucose and fasting insulin. A HOMA-IR above 2.5 is concerning; above 4 in a non-diabetic patient, in my clinical judgment, warrants intervention. The more sophisticated oral glucose tolerance test with insulin levels at 30 and 120 minutes can identify early insulin resistance before HOMA-IR becomes abnormal, but it is not yet standard in most primary care settings.

HOMA-IR was developed in the 1980s by Matthews and colleagues as an approximation of insulin resistance from simple fasting measurements. It correlates reasonably well with the gold-standard hyperinsulinemic-euglycemic clamp, which is too cumbersome for routine clinical use (Matthews et al, Diabetologia 1985, DOI: 10.1007/BF00280883). The formula gives a number that, while not perfect, is clinically useful and costs almost nothing to obtain alongside a standard fasting lipid panel and metabolic panel.

The limitation of HOMA-IR is that it captures only the fasting state. A person can have normal fasting insulin and HOMA-IR but mount an exaggerated insulin response to carbohydrates that is invisible on this test. Ideally, HOMA-IR is part of a panel that includes fasting triglycerides, HDL, and A1c, interpreted together rather than in isolation.

In my practice I order fasting insulin on almost every patient over 40 who comes in for cardiovascular risk assessment, particularly those who have never had it drawn. The number of patients I have found to have elevated HOMA-IR with a completely normal standard metabolic panel is substantial. These patients often come in believing their metabolic health is fine, and the HOMA-IR changes the conversation in a way that the standard panel never would.

The test requires that the patient be fasting for at least eight hours. Many labs do not include fasting insulin in a standard panel; you have to specifically request it. This is, I will say plainly, an oversight in standard preventive care.

Think of three different ways to assess how hard an engine is working. Fasting glucose tells you the output, the current RPM reading. A1c tells you the average RPM over the last three months. Fasting insulin tells you how much fuel the engine is burning to maintain that RPM. A car running at normal speed while burning three times the expected fuel has a mechanical problem that the speedometer alone cannot detect.

Fasting glucose is most useful for detecting overt diabetes (above 126 mg/dL fasting) and tracking gross metabolic deterioration. Its limitation is sensitivity: a person can have significant insulin resistance with a fasting glucose between 90 and 100, in the "normal" range, for years before it rises above the diagnostic threshold.

A1c is more stable than a single fasting glucose measurement because it averages three months of exposure. It is useful for diagnosis and for monitoring glycemic control in people already known to have diabetes. Its limitations include variation related to red cell turnover (falsely low in hemolytic anemia, falsely high in iron deficiency), and inability to capture glucose variability or post-meal excursions.

Fasting insulin is the most informative of the three for detecting early insulin resistance, yet it is the least commonly ordered. A fasting insulin above 10 to 12 microunits per milliliter in a non-diabetic adult warrants attention. Above 15 is consistently associated with metabolic syndrome components. Above 20, in my practice, typically triggers a broader metabolic workup including liver imaging for non-alcoholic fatty liver disease, more aggressive cardiovascular risk stratification, and a serious conversation about intervention.

This is the question I get most often from patients who have read something in a newsletter. The honest answer is: it depends on what you are trying to learn, and whether you will use the information well or spend two weeks anxious about numbers that are normal.

For people with frank diabetes or prediabetes, CGM has documented value in improving glycemic control. A 2021 randomized trial showed that CGM use in Type 2 diabetes patients not on insulin improved A1c meaningfully compared to conventional monitoring (Beck et al, JAMA 2017, DOI: 10.1001/jama.2017.4115). The evidence for non-diabetic individuals is much weaker. Most published studies in this population are observational, small, or industry-affiliated.

What a CGM can genuinely reveal in a non-diabetic: which foods cause unexpectedly sharp glucose spikes in that individual, how stress and sleep affect glucose, and whether post-meal glucose returns to baseline within two hours. These are informative. The risk is misinterpretation: many people see glucose spikes to 130 or 140 mg/dL after a meal, which are physiologically normal, and interpret them as pathological.

My clinical position: for a motivated patient with cardiovascular risk factors, a two-week CGM trial combined with a structured food log and a follow-up conversation is a reasonable, though not guideline-supported, approach to identifying dietary triggers and motivating behavior change. For a healthy person without metabolic risk, the information-to-anxiety ratio is less favorable.

The most consistent finding from CGM studies in non-diabetic individuals is the substantial person-to-person variation in glucose response to identical meals. A slice of white bread that spikes one person to 145 mg/dL may push another person to only 105 mg/dL. This was demonstrated compellingly in a 2015 Weizmann Institute study of 800 participants, which showed that postprandial glucose responses to identical foods varied enormously between individuals and were better predicted by the individual's gut microbiome composition and baseline metabolic parameters than by the glycemic index of the food (Zeevi et al, Cell 2015, DOI: 10.1016/j.cell.2015.11.001).

This personalization insight is genuinely useful. If you wear a CGM for two weeks and learn that white rice causes a three-hour glucose excursion in your particular metabolism while sweet potato does not, that is specific, useful dietary information about your own metabolism. The limitation is that this insight does not require indefinite CGM use. A structured two-to-four-week trial with a specific learning objective is very different from wearing a sensor continuously as a lifestyle indefinitely.

What CGM in a healthy person cannot tell you: it cannot diagnose cardiovascular disease, measure insulin sensitivity directly, or replace a fasting insulin and lipid panel. A glucose curve that looks beautiful on a CGM does not exclude insulin resistance; a person can have elevated fasting insulin and a HOMA-IR above 3 with glucose curves that look entirely normal to a CGM.

The American Diabetes Association defines postprandial hyperglycemia as glucose above 180 mg/dL two hours after a meal in diabetic patients, but this threshold is set for monitoring established disease, not for defining what is normal in a healthy person. In clinical research on non-diabetic adults, postprandial glucose consistently returns to pre-meal levels within 90 to 120 minutes in metabolically healthy individuals (Ceriello et al, Diabetologia 2008, DOI: 10.1007/s00125-008-1086-6).

The significance of postprandial spikes has been debated, but the evidence for cardiovascular harm from recurrent high glucose excursions is accumulating. The DECODE study, a large European cohort analysis, showed that two-hour postprandial glucose was a stronger predictor of cardiovascular mortality than fasting glucose, even after adjusting for A1c (DECODE Study Group, Lancet 1999, DOI: 10.1016/S0140-6736(98)12141-7). This makes physiological sense: endothelial cells are exposed to oxidative stress during each glucose spike, and the cumulative burden of years of daily excursions is atherogenic.

Practically: a post-meal spike to 130 mg/dL returning to 90 mg/dL within 90 minutes is normal and not concerning. A spike to 165 mg/dL that takes four hours to return to baseline, repeated at every meal, in a person who is told their A1c is "fine," is a metabolic story worth investigating. The A1c can average out repeated excursions and still register below the diagnostic threshold for prediabetes.

I have checked fasting insulin on hundreds of patients who were told they were metabolically healthy. The number of times that test has come back elevated, in a patient with a completely normal basic metabolic panel and A1c, is not small. These are the patients who are heading toward a diagnosis of diabetes in five to eight years but who currently have no findable code for their clinical state.

The reason fasting insulin is not standard is partly historical. When routine metabolic testing was systematized in the 1960s and 1970s, insulin assays were not widely available and the cost was prohibitive. The basic metabolic panel was built around glucose, and insulin was inferred by implication. That convention has persisted largely unchanged, even though insulin assays are now inexpensive and widely available.

The clinical value of fasting insulin: it identifies insulin resistance a decade before A1c rises above the prediabetes threshold. It guides treatment decisions. A patient with HOMA-IR of 5 and A1c 5.4 is not "pre-diabetic" by guidelines, but in my practice I treat that combination with the same urgency as a formal prediabetes diagnosis, because the metabolic trajectory is identical.

If your physician has never checked your fasting insulin, ask for it by name at your next visit. Request it alongside your fasting lipid panel, A1c, and fasting glucose. If the lab return shows insulin above 10, the next conversation is about why, and what to do.

The five-criteria framework, established by the National Cholesterol Education Program Adult Treatment Panel III in 2001 and subsequently adopted by the International Diabetes Federation with minor modifications, represents a clinical attempt to identify individuals whose cluster of metabolic abnormalities confers more cardiovascular risk than any single component would predict. The five criteria share a common pathophysiological root: insulin resistance and its downstream consequences.

Metabolic syndrome affects approximately one-third of American adults, with prevalence rising sharply above age 40 (Aguilar et al, JAMA 2015, DOI: 10.1001/jama.2015.4287). The cardiovascular risk associated with metabolic syndrome is mediated through multiple simultaneous mechanisms: the atherogenic dyslipidemia pattern (high triglycerides, low HDL, elevated small dense LDL), the low-grade inflammatory state driven by visceral adipose tissue, the endothelial dysfunction driven by hyperinsulinemia and impaired nitric oxide synthesis, and the prothrombotic state driven by elevated fibrinogen and plasminogen activator inhibitor-1.

The diagnostic limitation of the five-criteria framework is that it is a binary threshold model for what is actually a continuous spectrum. A person with four borderline criteria who does not meet three thresholds is not less at risk than someone who technically qualifies; they are simply classified differently. This is why I use the criteria as a starting conversation, not an ending one.

Subcutaneous fat, the fat you can pinch, is metabolically relatively inert. It stores energy and provides some insulation. It contributes to body weight and BMI, but large amounts of subcutaneous fat in the absence of visceral excess, a pattern seen more commonly in some African populations, does not carry the same cardiometabolic risk as visceral fat.

Visceral fat behaves differently because it is anatomically positioned to drain its metabolic products directly into the portal circulation, giving the liver first exposure to free fatty acids, inflammatory cytokines, and adipokines released by visceral adipocytes. This drives hepatic insulin resistance, promotes triglyceride synthesis, and fuels systemic low-grade inflammation. Visceral fat also secretes less adiponectin (a protective adipokine that improves insulin sensitivity) and more resistin and interleukin-6 than subcutaneous fat.

Imaging studies using CT or MRI to measure visceral fat directly have consistently shown that visceral fat area is a stronger predictor of cardiovascular events and metabolic syndrome than total body fat or BMI (Despres et al, Nature 2006, DOI: 10.1038/nature04917). The limitation of visceral fat measurement is that CT and MRI are expensive for routine screening. Waist circumference is the clinical surrogate, imperfect but accessible. In my practice I use waist circumference, fasting triglycerides, and HDL together as a simple visceral fat signal. The pattern of waist above 38 inches with triglycerides above 150 and HDL below 40 has a reasonably high specificity for elevated visceral fat.

BMI was created in the 1830s by the Belgian mathematician Adolphe Quetelet for population-level statistical analysis, not individual clinical assessment. It was adopted by medicine because it is easy to calculate from two measurements available in any office. Its limitations are well-documented: it classifies lean, muscular athletes as overweight or obese, and it classifies individuals with low muscle mass and significant visceral fat, the so-called TOFI phenotype, as normal weight while missing their cardiometabolic risk entirely.

In large prospective cohort studies, waist circumference predicts cardiovascular mortality and diabetes incidence independently of BMI. The INTERHEART study, which enrolled 27,000 patients across 52 countries, found that abdominal obesity measured by waist-to-hip ratio was a stronger predictor of myocardial infarction risk than BMI in every region studied (Yusuf et al, Lancet 2004, DOI: 10.1016/S0140-6736(04)17018-9). A man with BMI 24 and waist circumference 40 inches has a different cardiovascular risk profile than a man with BMI 28 and waist circumference 34 inches, and BMI alone would suggest the reverse.

The clinically useful thresholds for waist circumference: for men of European ancestry, above 40 inches (102 cm) is the widely cited threshold for elevated cardiometabolic risk; for men of South Asian, East Asian, or African ancestry, the threshold may be lower (around 35 to 37 inches) because visceral fat accumulation at lower overall body weight appears to carry equivalent metabolic risk. This is not always reflected in the guidelines patients are handed.

The TOFI concept emerged from magnetic resonance imaging studies in the early 2000s that systematically compared internal fat distribution to external appearance. Thomas et al. studied a cohort of normal-BMI individuals and found that 30 percent of those classified as "healthy weight" had visceral fat and liver fat levels associated with metabolic syndrome despite appearing lean (Thomas et al, Obesity 2012, DOI: 10.1038/oby.2012.83).

The TOFI phenotype is more common than appreciated among individuals of South and East Asian ancestry, among older adults who have lost muscle mass with preserved or redistributed fat (sarcopenic obesity), and among former athletes who have become sedentary in middle age. The last category is something I see in clinic: the former college football player or competitive cyclist who remains lean by the scale but whose waist has expanded modestly, whose fasting triglycerides are 190, and whose fasting insulin is 18. He does not look like a metabolic risk on paper. He is one.

The clinical catch is that standard risk calculators, including the Pooled Cohort Equations used to estimate 10-year cardiovascular risk, do not adequately capture TOFI physiology. BMI-based inputs will systematically underestimate risk in these patients. A CAC score, fasting insulin, and HOMA-IR are more informative than the calculated risk score for individuals in this phenotype.

The mechanisms by which a lean person develops insulin resistance are distinct from but overlapping with those in obesity. Ectopic fat deposition, particularly intramyocellular lipid accumulation in muscle fibers and hepatic steatosis, can impair insulin signaling in the absence of overt obesity. In lean insulin-resistant individuals, the fat is stored in the wrong places at relatively normal total amounts, disrupting insulin signaling within affected tissues.

Population data support this. In the HERITAGE Family Study cohort and similar exercise intervention studies, insulin resistance as measured by glucose clamp was found in a subset of lean individuals, and this lean-insulin-resistant phenotype was associated with dyslipidemia, elevated cardiovascular risk markers, and worse fitness responses to training (Bouchard et al, Med Sci Sports Exerc 1994, DOI: 10.1249/00005768-199409000-00041).

Genetic predisposition contributes meaningfully. Polymorphisms in genes involved in insulin signaling, adipokine production, and mitochondrial function can create insulin resistance in lean individuals in the absence of the adiposity that typically accompanies the condition. This is seen clinically in patients with strong family histories of Type 2 diabetes who present lean, young, and with A1c values in the prediabetes range.

Practically: if you are lean but have a strong family history of diabetes, carry fat centrally, have triglycerides above 150 with HDL below 45, or feel metabolically sluggish without an obvious explanation, a fasting insulin and HOMA-IR is warranted regardless of your BMI.

The "metabolically healthy obese" versus "metabolically unhealthy lean" comparison has been examined in several large cohort studies. A meta-analysis of 65,000 individuals found that metabolically unhealthy individuals, regardless of weight, had cardiovascular risk approximately twice that of metabolically healthy normal-weight individuals. Metabolically unhealthy lean individuals had cardiovascular risk comparable to metabolically unhealthy obese individuals, despite having lower body weight (Camhi et al, Obesity 2017, DOI: 10.1002/oby.21950).

The mechanism driving this risk in lean metabolically unhealthy individuals includes the same pathways as in obese individuals: atherogenic dyslipidemia, hyperinsulinemia, low-grade inflammation, endothelial dysfunction, and a prothrombotic state. The difference is that in a lean person, these mechanisms operate without the external signal that typically prompts investigation. Their physician does not reach for a metabolic panel with the same urgency because the body habitus does not suggest metabolic disease.

This represents a genuine gap in standard risk stratification. The Pooled Cohort Equations used to calculate 10-year ASCVD risk will assign lower absolute risk to a lean person simply because weight and BMI are not direct inputs. A lean 52-year-old man with fasting insulin of 20, HOMA-IR of 4.8, HDL of 38, and triglycerides of 210 may have a calculated 10-year risk of 7 percent when his actual biological risk may be substantially higher.

Glucose and fructose are metabolized very differently despite being structural isomers. Glucose is taken up by cells throughout the body in an insulin-dependent fashion, distributing its metabolic burden widely. Fructose is absorbed and delivered to the liver, where it bypasses the key regulatory enzyme phosphofructokinase and proceeds to lipogenesis with less feedback inhibition. The liver converts fructose to triglycerides with high efficiency, particularly in excess.

Controlled feeding studies, including work by Stanhope and colleagues, showed that isocaloric substitution of glucose with fructose significantly increased visceral fat, liver fat, and fasting triglycerides over ten weeks in overweight adults, even with total caloric intake held constant (Stanhope et al, J Clin Invest 2009, DOI: 10.1172/JCI37385). This is the mechanistic foundation for clinical concerns about sugar-sweetened beverages, which deliver large fructose loads rapidly.

The clinical implication is that not all calories are metabolically equivalent when it comes to visceral fat accumulation. A 200-calorie serving of almonds and a 200-calorie serving of cola do not impose the same hepatic or visceral fat burden. This does not mean fructose from whole fruit is a cardiovascular threat; the fiber, water content, and lower fructose concentration in whole fruit attenuate the hepatic load. The concern is concentrated, liquid fructose, primarily sugar-sweetened beverages, fruit juices (even "natural" ones), and foods with high-fructose corn syrup as a prominent ingredient.

A 2023 study in Nature Medicine found that plasma erythritol levels were associated with significantly elevated risk of major adverse cardiovascular events in a large cardiovascular phenotyping cohort, and that erythritol promoted platelet aggregation and thrombosis in animal and in vitro models (Witkowski et al, Nat Med 2023, DOI: 10.1038/s41591-023-02223-9). This study generated considerable attention and some appropriate pushback. The limitations include the observational design, inability to establish causation, and the possibility that erythritol was a marker of dietary patterns or metabolic state rather than a direct cause.

Saccharin, sucralose, and aspartame have been studied for decades with no definitive cardiovascular harm signal in randomized trials, but none of those trials were powered for cardiovascular endpoints, and most were short-term. The question of whether regular artificial sweetener use modifies gut microbiome composition, glucose regulation, or appetite signaling in ways that are clinically meaningful over decades is genuinely unresolved.

My clinical position: for a patient with Type 2 diabetes or obesity trying to reduce sugar-sweetened beverage consumption, a short-term transition to diet beverages or low-calorie sweetened alternatives is a reasonable harm-reduction step. For long-term use, particularly of erythritol and other sugar alcohols that are being added to a growing range of processed foods, I would recommend restraint until better data exists. Water remains the default.

The cephalic-phase insulin response is a conditioned reflex: the body anticipates incoming calories when it detects sweetness and releases a small amount of insulin preemptively. Research has shown this response with saccharin and other sweeteners in some studies, though the magnitude is modest and variable across individuals. A 2020 meta-analysis of 29 randomized trials found no consistent or clinically significant effect of non-nutritive sweeteners on fasting insulin or insulin response to a subsequent glucose challenge (Meyers & Brindal, Obes Rev 2020, DOI: 10.1111/obr.12956).

However, the gut microbiome story is more concerning. The 2014 Weizmann Institute study found that saccharin, sucralose, and aspartame all altered gut microbiome composition in mice and in a human pilot experiment, and that the microbiome changes were associated with glucose intolerance that was transferable to germ-free mice (Suez et al, Nature 2014, DOI: 10.1038/nature13793). This is an Early-level finding: plausible, concerning, not yet definitive in large human trials.

The practical point is that the insulin response question, while interesting, may not be the most important question. If artificial sweeteners maintain cravings for sweet taste, promote overconsumption of calories elsewhere, or alter metabolic regulation through gut-mediated pathways, the net effect on insulin sensitivity and cardiovascular risk could be meaningful even without a direct insulin-stimulating effect.

GLP-1 (glucagon-like peptide-1) receptors are expressed not only in the pancreas and gut but also in cardiomyocytes, vascular smooth muscle cells, and macrophages in the arterial wall. The direct cardiac effects of GLP-1 receptor agonism include reduced inflammation in atherosclerotic plaque, decreased macrophage foam cell formation, improved endothelial function through nitric oxide upregulation, modest reductions in blood pressure, and direct effects on cardiac conduction and myocardial contractility.

The LEADER trial of liraglutide enrolled 9,340 patients with Type 2 diabetes and high cardiovascular risk. Liraglutide reduced the composite of cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke by 13 percent compared to placebo over a median follow-up of 3.8 years (Marso et al, NEJM 2016, DOI: 10.1056/NEJMoa1603827). Importantly, this cardiovascular benefit appeared before meaningful weight separation between the arms, suggesting mechanisms beyond weight reduction alone.

Subsequent trials with semaglutide (SUSTAIN-6, PIONEER-6) confirmed the cardiovascular benefit of this drug class in diabetic patients. The next question, answered by the SELECT trial discussed in Q22, was whether the benefit extends to non-diabetic patients with obesity and cardiovascular disease.

The cardiovascular case for GLP-1s solidified with the SELECT trial, which enrolled 17,604 non-diabetic adults aged 45 or older with obesity (BMI at or above 27) and established cardiovascular disease. Semaglutide 2.4 mg weekly reduced the composite of cardiovascular death, nonfatal myocardial infarction, and nonfatal stroke by 20 percent compared to placebo over a mean follow-up of 39.8 months (Lincoff et al, NEJM 2023, DOI: 10.1056/NEJMoa2307563). This was the first cardiovascular outcomes trial to demonstrate benefit in non-diabetic individuals, and it changed the clinical framing of semaglutide from "diabetes drug" to cardiovascular drug used in obesity.

The magnitude of the SELECT benefit, 20 percent relative risk reduction, is clinically meaningful and compares favorably to statin benefit in secondary prevention populations. The absolute risk reduction in SELECT was approximately 1.5 percent over 39 months, which translates to a number needed to treat of approximately 67 patients for 3.5 years to prevent one cardiovascular event. For context, high-intensity statin therapy in secondary prevention has an NNT of approximately 50 to 70 over five years.

The mechanism of cardiovascular benefit in the SELECT population appears to involve both weight-related improvements (improved blood pressure, lipid profiles, inflammatory markers) and the direct vascular effects of GLP-1 receptor agonism described in Q21. Disentangling the two remains an active area of research.

This question requires careful precision because "healthy obese" is a phenotype the SELECT trial specifically excluded. All 17,604 participants had existing cardiovascular disease, prior myocardial infarction, stroke, or symptomatic peripheral arterial disease. The stunning 20 percent reduction in events was demonstrated in a secondary prevention population.

What this means practically: a 48-year-old man with obesity and no prior cardiovascular events, no diabetes, and a 10-year ASCVD risk of 8 percent is not yet represented in the SELECT evidence base for GLP-1 prescription. His physician cannot yet point to a randomized trial that shows cardiovascular benefit from semaglutide in his specific clinical context.

What this does suggest: given that the LEADER, SUSTAIN-6, and SELECT trials collectively show GLP-1 benefit across a spectrum of cardiovascular risk, it is reasonable to hypothesize that primary prevention benefit exists. The SELECT trial gives a signal, not a prescription, for the primary prevention space. Ongoing trials, including SOUL and FLOW, will provide additional data on semaglutide in populations with varying cardiovascular risk.

For a clinician advising a healthy obese patient today, the weight loss benefit and the metabolic improvements from GLP-1 therapy are sufficient reason to discuss the medication. The cardiovascular benefit in primary prevention is biologically plausible but not yet proven.

The clinical decision framework for a non-diabetic with high cardiac risk and obesity today: established CVD plus obesity places a patient in the SELECT evidence base, where semaglutide has an NNT of 67 over 3.5 years. This is comparable to benefit seen with statins and antihypertensives in high-risk populations and is sufficient, in my practice, to discuss GLP-1 therapy alongside standard secondary prevention medications.

For non-diabetics with high calculated cardiovascular risk but no prior events (primary prevention), the evidence base is thinner. The strongest case rests on the indirect chain: GLP-1 therapy reduces weight significantly in most patients, weight reduction in obese individuals improves blood pressure, reduces triglycerides, raises HDL, and reduces inflammatory markers, all of which reduce cardiovascular risk through established pathways. Whether this risk reduction translates to the magnitude seen in SELECT remains unknown.

The practical considerations for the conversation with a patient: GLP-1 medications have a meaningful side effect profile (nausea, vomiting, gastroparesis risk, potential pancreatitis, the ongoing thyroid signal not yet resolved in humans), cost barriers (semaglutide can exceed $1,000 per month without insurance coverage), and the question of duration (discussed in Q48). These are not disqualifying considerations, but they are part of an honest informed consent conversation.

Metformin is one of the most studied medications in existence. It has been used for Type 2 diabetes since the 1950s and has an exceptional long-term safety record. Its mechanisms relevant to aging include activation of AMPK (AMP-activated protein kinase, a cellular energy sensor), inhibition of complex I of the mitochondrial electron transport chain, reduction of hepatic glucose production, and modest effects on mTOR signaling, a pathway implicated in cellular senescence.

Observational data have repeatedly suggested that diabetic patients on metformin live longer and have lower cancer rates than diabetic patients on other glucose-lowering agents, and in some analyses, longer than matched non-diabetic patients not on metformin. These findings are striking but carry all the limitations of observational epidemiology: confounding by indication, healthy user bias, and the challenge of comparing drug users to non-users across populations.

The TAME trial (Targeting Aging with Metformin), sponsored by the National Institute on Aging, is a phase III randomized, placebo-controlled trial enrolling 3,000 non-diabetic adults aged 65 to 79. It is testing whether metformin delays the onset of aging-related conditions including cardiovascular disease, cancer, and cognitive decline (Barzilai et al, Cell Metab 2016, DOI: 10.1016/j.cmet.2016.05.011). Results are expected in the latter half of the decade.

Until TAME reports, prescribing metformin for longevity in non-diabetics is off-label, not guideline-supported, and rests on a biologically plausible but unproven hypothesis.

TAME is a conceptually significant trial because it treats aging as a modifiable biological process rather than as a background condition. The rationale is that by targeting aging biology, specifically cellular senescence, mitochondrial dysfunction, and the chronic inflammation associated with aging (sometimes called "inflammaging"), a single intervention might delay the onset of multiple age-related diseases simultaneously.

The primary composite endpoint is cancer, cardiovascular event, or dementia. Secondary endpoints include physical function, frailty, and mortality. The trial dose is 1,500 mg of metformin daily, a dose at the lower end of the standard diabetes dosing range, chosen to minimize gastrointestinal side effects in this population (Barzilai et al, Cell Metab 2016, DOI: 10.1016/j.cmet.2016.05.011).

What TAME will and will not answer: if positive, it will establish that metformin delays specific aging-related diseases in older non-diabetic adults. It will not answer whether the same benefit applies to younger adults, whether the benefit is specific to metformin or is a class effect of AMPK activation, or whether the benefit is additive to existing cardiovascular medications. These are follow-up questions for subsequent trials.

The broader significance is methodological: TAME pioneered the concept of a regulatory pathway for "anti-aging" trials at the FDA, which now recognizes "delaying aging" as a legitimate therapeutic target for clinical trial purposes. That regulatory development may prove as important as the trial result itself.

Intermittent fasting influences insulin resistance through several mechanisms: caloric restriction (in most real-world implementations, fasting protocols reduce total caloric intake), reduction in insulin secretion during fasting periods allowing receptor upregulation, enhancement of hepatic insulin sensitivity through liver glycogen depletion, and activation of autophagy and AMPK pathways that improve cellular metabolic efficiency.

A 2020 meta-analysis of 27 trials found that intermittent fasting reduced fasting insulin by a mean of 14 to 20 percent and improved HOMA-IR significantly compared to control conditions. These effects were largely, though not entirely, mediated by the associated weight loss (Harris et al, JAMA Intern Med 2020, DOI: 10.1001/jamainternmed.2020.4836). The trials that controlled for total caloric intake while comparing intermittent versus continuous restriction showed similar metabolic improvements, suggesting that the pattern of eating matters independently of total calories to some degree.

The cardiovascular evidence for intermittent fasting is less developed than the metabolic evidence. No large cardiovascular outcomes trial has tested intermittent fasting. A 2024 observational study raised a signal that time-restricted eating patterns of less than 8 hours per day were associated with higher cardiovascular mortality, though the methodology was criticized and the findings have not been replicated in randomized data.

Time-restricted eating is distinct from intermittent fasting protocols that involve alternating fasting days or multi-day fasts. TRE is a daily practice: eating begins at, for example, 10 am and ends by 6 pm, with no caloric intake outside that window. The rationale draws from circadian biology: metabolic processes including insulin sensitivity, gut motility, and hepatic lipid metabolism follow circadian rhythms that are partly food-entrained. Eating in alignment with the active phase of the circadian day (daytime) may improve metabolic efficiency relative to eating at night.

A well-conducted 2022 randomized trial by Lowe et al. in the New England Journal of Medicine found that time-restricted eating (8-hour window) compared to standard three-meal eating did not produce significantly greater weight loss over 12 months when caloric intake was controlled (Lowe et al, NEJM 2022, DOI: 10.1056/NEJMoa2114833). This dampened some of the enthusiasm for TRE as a strategy distinct from caloric restriction. The trial did find modest improvements in fasting glucose and blood pressure in the TRE group.

The 2024 observational study associating short eating windows with cardiovascular mortality used dietary recall data from a nationally representative sample, which is a methodologically limited approach for inferring causation. People who spontaneously eat in a very short window may do so due to illness, socioeconomic constraints, or shift work, all of which are independent cardiovascular risk factors. This finding should be monitored but does not change clinical guidance.

"Low-carb" encompasses a spectrum from moderate carbohydrate restriction (100 to 130 grams per day) to ketogenic diets (below 20 to 50 grams per day). The metabolic effects differ somewhat across this spectrum, and most published trials use different definitions, making direct comparison difficult.

Across the consistent findings: reducing refined carbohydrates and added sugars reduces fasting triglycerides reliably and substantially (often 30 to 50 percent reduction in patients with elevated baseline), raises HDL (typically 5 to 15 percent), reduces fasting glucose and fasting insulin, and produces modest blood pressure improvement in those with hypertension. These are all cardiovascular risk-favorable changes (Bueno et al, Br J Nutr 2013, DOI: 10.1017/S0007114513000548).

The LDL question is more complicated. Low-carb diets typically produce small to modest LDL reductions in most people. A subset of individuals, estimated at 5 to 10 percent, experience dramatic LDL elevations on low-carb or ketogenic diets, sometimes LDL rising above 200 to 300 mg/dL. This is the "lean mass hyper-responder" pattern discussed in Q31. Whether elevated LDL in this context carries the same cardiovascular risk as elevated LDL from other causes is an open and actively debated question.

My clinical approach: I use low-carb dietary patterns for patients with metabolic syndrome, prediabetes, and significantly elevated triglycerides. I monitor LDL and ApoB before and after any significant dietary shift. If ApoB rises substantially, the dietary benefit on triglycerides and HDL does not automatically outweigh the atherogenic particle burden.

The ketogenic diet, defined as carbohydrate intake below 20 to 50 grams per day with resulting ketosis, produces a distinctive metabolic profile. Triglycerides typically fall dramatically, often by 40 to 60 percent in hypertriglyceridemic patients. HDL rises, usually by 10 to 20 percent. Fasting glucose and fasting insulin improve substantially in individuals with insulin resistance. Blood pressure often decreases modestly.

These changes sound uniformly favorable. The complication is the atherogenic particle story. In a proportion of individuals on keto, LDL cholesterol rises significantly, sometimes dramatically. More importantly, ApoB (which measures the total number of atherogenic particles and is a stronger predictor of cardiovascular risk than LDL-C) rises in some keto practitioners even when LDL-C appears stable. This is because keto can increase the production of large buoyant LDL particles that raise LDL-C but also increase the total particle number measured by ApoB.

The 2023 American College of Cardiology Scientific Session presentation by Norwitz and colleagues described the lean mass hyper-responder phenotype in detail and showed that some individuals on keto develop LDL above 300 mg/dL with triglycerides below 70 and HDL above 80, a pattern not seen in standard dyslipidemia and whose cardiovascular significance is debated but concerning.

For a patient starting a ketogenic diet: I obtain a baseline lipid panel with ApoB, repeat at three months, and if ApoB rises substantially, we have a frank conversation about whether the metabolic gains outweigh the atherogenic particle signal.

The LMHR pattern was characterized in detail by Dave Feldman and subsequently studied by Nick Norwitz and colleagues. The proposed mechanism involves the "lipid energy model": in lean, active individuals on a ketogenic diet, the liver exports very large quantities of VLDL particles to supply peripheral tissues (particularly muscle) with fatty acids for fuel. When insulin is low and the liver is in full fat-burning mode, LDL-C measured by standard Friedewald equations can reach very high levels while the metabolic picture otherwise looks excellent.

The critical clinical question is whether high LDL in the LMHR context carries the same cardiovascular risk as high LDL in an insulin-resistant individual with elevated triglycerides. Standard risk models assume that elevated LDL-C always means elevated atherogenic particle burden. The LMHR proponents argue that in the context of very low insulin, low triglycerides, and high HDL, the atherogenic biology of elevated LDL may be modified.

The KETO-CTA trial, a coronary CT angiography study enrolling LMHR individuals on long-term keto diets, is assessing coronary plaque burden as a surrogate endpoint. Results are pending as of mid-2026. Until those data are available, I treat significant ApoB elevation on keto as a cardiovascular signal regardless of the metabolic context, because ApoB is the most direct measure of atherogenic particle number and the trial evidence for its predictive value is robust.

The carnivore diet, consisting exclusively of animal products with zero plant foods, is a further restriction of the ketogenic diet. Its proponents claim benefits including elimination of food sensitivities, resolution of autoimmune symptoms, and metabolic improvement. Its critics point to the complete absence of fiber, the potential for accelerated atherosclerosis from saturated fat and high ApoB, and the disruption of gut microbiome diversity.

What limited data exists: a 2021 survey study of 2,029 self-reported carnivore diet adherents found high satisfaction rates and self-reported improvements in multiple health outcomes, including weight and glucose control (Lennerz et al, Am J Clin Nutr 2021, DOI: 10.1093/ajcn/nqab337). This is survey data with all the limitations thereof; people who are doing well on a diet are more likely to participate in surveys about it than people who are not.

The cardiovascular concern I hold most seriously is the ApoB question, the same concern as with keto. If a carnivore diet significantly raises ApoB, the plaque consequences of that particle burden over five to ten years could negate whatever metabolic improvements were achieved in year one. Without long-term coronary imaging data, I cannot tell a patient that the carnivore diet is safe for the arteries.

My clinical position: if a patient reports feeling dramatically better on a carnivore approach and their metabolic markers have improved, I do not argue them out of it. I check their ApoB, LDL-C, hsCRP, and renal function at baseline and again at six months, and I monitor the cardiovascular signal closely.

The Mediterranean diet is characterized by high intake of extra-virgin olive oil, nuts, legumes, vegetables, fruits, fish, and whole grains, with low to moderate consumption of red meat and dairy, and low consumption of processed foods and added sugars. It is not a single nutrient; it is a dietary pattern that appears to work through multiple simultaneous mechanisms.

The PREDIMED trial (Prevención con Dieta Mediterránea) enrolled 7,447 adults at high cardiovascular risk and randomized them to a Mediterranean diet supplemented with extra-virgin olive oil, a Mediterranean diet supplemented with nuts, or a control low-fat diet. After a median of 4.8 years, both Mediterranean diet arms had approximately 30 percent fewer major adverse cardiovascular events than the control arm (Estruch et al, NEJM 2018, DOI: 10.1056/NEJMoa1800389). The trial was replicated by PREDIMED-Plus in a cohort with caloric restriction added to the Mediterranean diet, showing additional benefit.

When compared to other named dietary patterns, the Mediterranean approach consistently outperforms or equals alternatives in cardiovascular outcomes data. Low-carb diets have not been tested in comparable cardiovascular outcomes trials. The DASH diet has strong blood pressure evidence but less cardiovascular event data. Plant-based approaches have observational support and mechanistic plausibility but fewer large RCTs.

The Mediterranean diet is the most evidence-supported dietary pattern for cardiovascular risk reduction and is my first-choice recommendation for most patients regardless of metabolic status.

The PREDIMED trial (published in NEJM 2013, then republished in corrected form in 2018 after randomization irregularities at one site were identified and corrected without changing the overall conclusions) is the most important dietary trial in cardiovascular medicine. Its findings matter because they are based on a hard endpoint, cardiovascular events, not just surrogate markers like cholesterol levels or blood pressure.

The participants were Spanish adults aged 55 to 80 with Type 2 diabetes or at least three major cardiovascular risk factors. They were assigned to: Mediterranean diet plus at least four tablespoons per day of extra-virgin olive oil, Mediterranean diet plus 30 grams of mixed nuts per day, or a low-fat control diet. The Mediterranean groups were provided with the supplemental foods free of charge and received quarterly dietary counseling.

At 4.8 years, the combined endpoint rate was 3.4 percent per year in the olive oil group and 3.3 percent per year in the nuts group, compared to 4.4 percent per year in the control group. The benefit was seen for cardiovascular death and stroke specifically; myocardial infarction trends did not reach statistical significance in isolation.

Several mechanistic pathways likely contributed: the anti-inflammatory effects of polyphenols in olive oil and nuts, improved lipid profiles (particularly HDL increase and triglyceride reduction), blood pressure lowering, and reduced platelet aggregation from omega-3 fatty acids in nuts and fish.

The mechanisms by which fiber influences glucose metabolism are multiple. Soluble fiber (found in oats, legumes, apples, and psyllium) forms a gel in the gut that slows the absorption of glucose from the small intestine, reducing the height and extending the duration of postprandial glucose curves. This directly reduces the glycemic load of a meal regardless of total carbohydrate content.

The gut microbiome pathway may be equally important. Fermentable fiber is metabolized by gut bacteria into short-chain fatty acids, primarily butyrate, propionate, and acetate. Butyrate is the primary energy source for colonocytes and has anti-inflammatory effects on the gut epithelium. Propionate reaches the liver and suppresses hepatic glucose production. These microbially mediated effects on glucose metabolism are biologically distinct from the physical slowing of glucose absorption.

Population data are consistent and striking. A 2019 dose-response meta-analysis of 185 prospective studies and 58 clinical trials found that people consuming 25 to 29 grams of dietary fiber per day had 15 to 30 percent lower rates of all-cause mortality, cardiovascular disease, Type 2 diabetes, and colorectal cancer compared to people consuming below 15 grams per day (Reynolds et al, Lancet 2019, DOI: 10.1016/S0140-6736(18)31809-9). Average American fiber intake is approximately 17 grams per day, well below the recommended 25 to 38 grams.

Resistant starch exists in four forms. Type 1 is physically enclosed in intact plant cell walls (whole grains, legumes). Type 2 is found in raw, unprocessed foods (raw potatoes, green bananas). Type 3, retrograde resistant starch, forms when starchy foods are cooked and then cooled, a transformation that converts some digestible starch to resistant starch through retrogradation. Type 4 is chemically modified starch used in food manufacturing.

The type 3 pathway is practically important because it means that cooked-and-cooled rice, potatoes, or pasta has a different glycemic effect than the freshly cooked version of the same food. Reheating does not reverse the retrogradation completely. This is not a large effect, but it is real: cooling cooked white rice overnight reduces its glycemic index by approximately 10 to 20 percent, as shown in small but well-designed metabolic studies.

The insulin sensitizing effects of resistant starch are mediated primarily through short-chain fatty acid production, particularly propionate's suppression of hepatic glucose output and butyrate's improvement of intestinal barrier function. A 2018 meta-analysis of 17 randomized trials found that resistant starch supplementation significantly improved fasting insulin and insulin sensitivity measures compared to control, with effects most pronounced in individuals with baseline insulin resistance (Kwak et al, Nutr J 2018, DOI: 10.1186/s12937-018-0361-3).

The mechanism is direct and well-established. Muscle contraction activates AMPK and releases calcium from the sarcoplasmic reticulum, triggering translocation of GLUT4 transporters to the muscle cell surface independent of insulin. This means that during and after exercise, glucose enters muscle cells through a pathway that does not require insulin signaling. For an insulin-resistant individual, exercise essentially bypasses the broken lock and opens the door directly.

The magnitude of improvement is clinically significant. A 2020 meta-analysis of 106 randomized trials found that combined aerobic and resistance training reduced HOMA-IR by an average of 0.73 units (approximately 25 percent of baseline) in people with Type 2 diabetes or prediabetes, with effects seen across all training modalities (Colberg et al, Diabetes Care 2016, DOI: 10.2337/dc16-1728). The combination of aerobic exercise and resistance training was consistently more effective than either alone.

The dose-response relationship is reasonably linear up to approximately 150 to 300 minutes per week of moderate-intensity exercise. Beyond that range, incremental metabolic benefit per additional hour of exercise diminishes, though cardiovascular fitness continues to improve.

The speed of response is clinically useful for patient motivation: measurable improvement in fasting insulin can occur within three to four weeks of initiating a consistent exercise program. This is one of the fastest-acting metabolic interventions available.

This is one of the most consistently reproducible findings in exercise physiology. The mechanism involves two temporal phases: the immediate effect during exercise (AMPK-mediated, insulin-independent GLUT4 translocation) and the post-exercise effect (enhanced insulin signaling in recovering muscle through increased GLUT4 expression and improved glycogen synthase activity). The second phase, the 24 to 72-hour window of improved insulin sensitivity, means that the timing of the next exercise session relative to the expiration of the previous session's benefit matters metabolically.

A study by Mikines et al. showed that insulin sensitivity, measured by hyperinsulinemic euglycemic clamp, was significantly enhanced for 48 hours following a single bout of cycling in healthy trained men. The enhancement was greater with higher exercise intensity and longer duration, up to a ceiling (Mikines et al, Am J Physiol 1988, DOI: 10.1152/ajpendo.1988.254.3.E248).

The practical implication is that exercise frequency, not just weekly duration, matters for metabolic outcomes. Walking for 90 minutes on Saturday and doing nothing the rest of the week is metabolically inferior to walking for 20 minutes every day, even if the total weekly duration is similar. The continuous maintenance of GLUT4 upregulation through regular sessions maintains a metabolic floor that single long sessions do not.

This is perhaps the most under-emphasized fact in metabolic medicine: skeletal muscle is the largest glucose disposal organ in the body. When you eat a meal and insulin is released, approximately 80 percent of the glucose that insulin drives into cells goes into muscle. Not fat tissue. Not the liver. Muscle. This means that the total mass of your functional skeletal muscle is the primary determinant of how quickly and efficiently your body clears postprandial glucose.

As muscle mass declines with aging (sarcopenia, discussed in Q40), glucose disposal capacity falls proportionally. The same carbohydrate meal that a 30-year-old man with 80 kilograms of lean mass can clear efficiently may drive prolonged postprandial glucose elevation in a 65-year-old man with 55 kilograms of lean mass, even if the older man has no other metabolic disease. This age-related decline in muscle-mediated glucose disposal is a major contributor to the rising A1c that many men experience in their 50s and 60s even when dietary habits have not changed.

The intervention is resistance training. Building and maintaining muscle mass through progressive resistance exercise is not primarily about aesthetics or strength in the context of metabolic medicine. It is about maintaining the body's primary glucose disposal system. A meta-analysis of 42 trials showed that resistance training improved insulin sensitivity and glycated hemoglobin in patients with Type 2 diabetes independently of aerobic exercise (Colberg et al, Diabetes Care 2016, DOI: 10.2337/dc16-1728).

Sarcopenia affects approximately 10 percent of adults over 60 and up to 50 percent of those over 80 (Cruz-Jentoft et al, Age Ageing 2019, DOI: 10.1093/ageing/afz046). The loss begins earlier: muscle mass typically peaks in the late 20s to early 30s and begins declining gradually from the 40s, accelerating after 60. The rate of loss averages approximately 1 to 2 percent of muscle mass per year after age 50 in sedentary individuals, with parallel losses in strength and power.

The metabolic consequences of sarcopenia go beyond glucose disposal. Muscle tissue is the primary site of resting fat oxidation; reduced muscle mass decreases basal metabolic rate, making fat accumulation more likely with the same caloric intake. Muscle secretes myokines, including irisin and IL-6, that have systemic anti-inflammatory and metabolic regulatory effects; reduced muscle mass means reduced myokine secretion and reduced metabolic signaling.

The cardiovascular connection is bidirectional: sarcopenia is associated with higher rates of heart failure, coronary artery disease, and cardiovascular mortality, and cardiovascular disease accelerates sarcopenia through inflammation, impaired exercise capacity, and poor nutrition. The two conditions feed each other.

Prevention and treatment center on resistance training (the most evidence-based intervention for preserving and building muscle mass at any age) combined with adequate protein intake. Protein requirements for muscle protein synthesis are higher in older adults than standard guidelines suggest: most exercise physiology research supports 1.6 to 2.2 grams of protein per kilogram of body weight per day in active adults trying to maintain or build muscle.

The skinny fat phenotype is the body composition version of the TOFI concept. In terms of metabolic mechanism, what matters is the ratio of lean mass to fat mass and the distribution of fat (visceral versus subcutaneous), not total body weight. A 175-pound man with 40 percent body fat and very low muscle mass is at higher cardiometabolic risk than a 195-pound man with 20 percent body fat and substantial lean mass, even though the scale favors the leaner-appearing man.

Population data confirm this. The NHANES study found that normal-weight individuals with elevated body fat percentage had significantly higher rates of metabolic syndrome and cardiovascular risk factors than normal-weight individuals with lower body fat (Romero-Corral et al, Int J Obes 2010, DOI: 10.1038/ijo.2010.103). The atherogenic mechanisms are identical to those in obese individuals: visceral fat-mediated inflammation, atherogenic dyslipidemia, and insulin resistance.

The practical implication for clinical risk assessment: BMI and weight alone are insufficient. Body composition measurement, either through DXA (dual-energy X-ray absorptiometry), bioelectrical impedance, or even the simple clinical estimate of waist circumference relative to height, adds meaningful risk information. A man who looks thin but has a waist above 36 inches, triglycerides above 150, and HDL below 40 deserves a full metabolic workup regardless of what the scale says.

The sleep-metabolism connection is one of the most consistent findings in metabolic research, and one of the most ignored in clinical practice. Adults require seven to nine hours of sleep per night for metabolic normality. This is not a lifestyle preference; it is a biological requirement whose violation has measurable physiological consequences within days.

Cortisol is the primary mediator. Sleep deprivation activates the HPA axis, elevating morning cortisol and maintaining higher cortisol throughout the day. Cortisol is a glucocorticoid: it promotes hepatic glucose production, inhibits insulin signaling in peripheral tissues, and promotes visceral fat deposition over time. One week of sleep restriction to five hours per night in healthy young adults produced insulin resistance measurable by glucose clamp in a study by Spiegel et al. (Spiegel et al, Sleep 1999, DOI: 10.1093/sleep/22.3.407).

Slow-wave sleep, the deepest stage of non-REM sleep, is specifically associated with metabolic restoration through growth hormone secretion and suppression of cortisol. Conditions that fragment sleep and reduce slow-wave sleep, including obstructive sleep apnea and poor sleep hygiene, impair these restorative processes directly.

Population-level data consistently show that habitual sleep below six hours per night is associated with 30 to 50 percent higher risk of developing Type 2 diabetes, with a dose-response relationship between sleep duration and metabolic risk (Cappuccio et al, Diabetes Care 2010, DOI: 10.2337/dc09-1124).

This finding always surprises patients. They accept intuitively that years of sleep deprivation might affect metabolism, but that one night has measurable effects the next morning seems disproportionate. The biology is straightforward: growth hormone, which is secreted primarily during slow-wave sleep, plays a key role in maintaining insulin sensitivity, particularly in muscle. A single night with suppressed slow-wave sleep blunts growth hormone secretion and produces detectable insulin resistance by morning.

The Tasali et al. study at the University of Chicago selectively suppressed slow-wave sleep for three nights in healthy young adults using acoustic stimulation that reduced deep sleep without fully waking participants. After three nights, insulin sensitivity fell by 25 percent, comparable to levels seen in adults 10 years older or in impaired glucose tolerance (Tasali et al, PNAS 2008, DOI: 10.1073/pnas.0706446105).

The inflammatory pathway compounds the insulin effect. A single night of short sleep (below 6 hours) increases levels of interleukin-6 and tumor necrosis factor-alpha, both of which impair insulin receptor signaling. These inflammatory mediators are the same ones elevated in obesity-related insulin resistance; sleep deprivation produces a biochemically similar, if less severe, state.

For a patient trying to reverse insulin resistance, this means that even perfect diet and exercise are partially undermined by consistent short sleep. Sleep is not a lifestyle add-on to a metabolic plan. It is foundational.

The relationship between stress and glucose is one of evolutionary design. The cortisol stress response was designed to mobilize glucose for acute physical threat, fight or flight. Liver glycogen is broken down, new glucose is synthesized, and peripheral insulin signaling is suppressed to preserve glucose for the brain and contracting muscles. This makes sense for a 30-second sprint away from a predator.

It makes considerably less sense for a ten-year period of executive stress, financial anxiety, or relationship difficulty. Chronically elevated cortisol maintains a state of physiological glucose mobilization that was designed to be brief. The liver is continuously generating glucose it was not asked for. Peripheral tissues are continuously being told to resist insulin. Visceral fat, which has high cortisol receptor density compared to subcutaneous fat, accumulates preferentially under chronic cortisol exposure.

Prospective data from the Whitehall II study, a longitudinal cohort of British civil servants, showed that chronic work stress and adverse psychosocial environment were independently associated with metabolic syndrome development over follow-up, after adjustment for diet, activity, and socioeconomic status (Chandola et al, BMJ 2006, DOI: 10.1136/bmj.38693.435301.80). The mechanisms included cortisol-driven glucose dysregulation, HPA axis hyperactivity, and sympathetic nervous system activation that raised blood pressure and promoted visceral fat accumulation.

The clinical implication is that treating insulin resistance without addressing the stress burden is incomplete. A patient eating well, exercising, and sleeping appropriately but living in a state of chronic psychological stress may still have elevated fasting cortisol, impaired overnight insulin sensitivity, and glucose metabolism that responds partially but not fully to lifestyle intervention.

The kidney normally filters approximately 180 grams of glucose per day and reabsorbs almost all of it through SGLT2 transporters in the proximal tubule, returning glucose to the bloodstream. SGLT2 inhibitors block this reabsorption, forcing the kidney to excrete 50 to 100 grams of glucose in the urine per day. The result is modest blood glucose lowering accompanied by glycosuria and a mild osmotic diuresis.

The clinical revolution with this drug class came not from their glucose-lowering but from their cardiovascular trial results. The EMPA-REG OUTCOME trial showed empagliflozin reduced cardiovascular death by 38 percent in diabetic patients with established cardiovascular disease, a stunning result that was initially met with considerable skepticism (Zinman et al, NEJM 2015, DOI: 10.1056/NEJMoa1504720). The CREDENCE, DAPA-HF, and EMPEROR-Reduced trials subsequently showed benefit extending to heart failure patients regardless of diabetes status.

The mechanism of cardiovascular and renal benefit from SGLT2 inhibitors is only partially understood. Proposed mechanisms include preload and afterload reduction from osmotic diuresis and natriuresis, suppression of myocardial sodium-calcium exchange that reduces myocardial fibrosis, anti-inflammatory effects, and improvement in renal hyperfiltration that slows progression of diabetic nephropathy. The benefit in non-diabetic heart failure patients suggests the mechanisms are largely glucose-independent.

The DAPA-HF trial enrolled 4,744 heart failure patients with ejection fraction at or below 40 percent; 55 percent did not have diabetes. Dapagliflozin reduced the composite of worsening heart failure or cardiovascular death by 26 percent compared to placebo, with nearly identical benefit in the diabetic and non-diabetic subgroups (McMurray et al, NEJM 2019, DOI: 10.1056/NEJMoa1911303). The EMPEROR-Reduced trial confirmed this finding with empagliflozin in a similar population.

This was the rare and satisfying moment in cardiology when a drug developed for one indication turns out to be therapeutic for a different condition through a mechanism that only becomes clear after the trial. The cardiovascular mechanisms of SGLT2 inhibitor benefit in heart failure now appear to include reductions in ventricular preload and afterload through natriuresis, anti-fibrotic effects on myocardial remodeling, improvement of mitochondrial energetics in the stressed myocardium, and reduction in myocardial inflammation.

The 2022 ACC/AHA Heart Failure Guidelines now include SGLT2 inhibitors as a Class IIa recommendation for heart failure with reduced ejection fraction regardless of diabetes status, alongside angiotensin-converting enzyme inhibitors, beta-blockers, and mineralocorticoid receptor antagonists. In clinical practice, this means that a non-diabetic patient admitted for heart failure decompensation should be considered for SGLT2 inhibitor therapy before discharge.

The pattern of rapid weight regain after GLP-1 discontinuation was clearly documented in the STEP 1 Extension trial. Participants who had lost a mean of 17.3 percent body weight over 68 weeks and then stopped semaglutide regained two-thirds of that weight within 52 weeks. Most cardiometabolic improvements, including blood pressure, lipids, and glucose, reverted toward or past the pre-treatment baseline (Wilding et al, Diabetes Obes Metab 2022, DOI: 10.1111/dom.14725).

The cardiovascular risk question is whether rapid weight regain imposes an additional risk beyond simply returning to the pre-treatment metabolic state. There is biological reason for concern. Rapid weight cycling, including regain after significant loss, is associated with increased sympathetic nervous system activation, adverse cardiac remodeling, and elevations in inflammatory markers. The Look AHEAD trial showed that weight fluctuation was associated with higher cardiovascular risk than stable obesity in some subgroup analyses.

Whether the SELECT trial population, who showed 20 percent cardiovascular event reduction with semaglutide, would experience rapid cardiovascular risk reversal or rebound upon stopping is a critical unanswered question. The SELECT trial did not have a discontinuation arm.

The practical clinical implication is that GLP-1 therapy, if started for cardiovascular or metabolic indications, appears to require long-term or indefinite continuation to maintain benefit, similar to how statins must be continued to maintain LDL reduction.

GLP-1 receptor agonists address a chronic biological condition: impaired GLP-1 secretion, dysregulated appetite signaling, and excess adiposity. When the medication is stopped, the underlying condition remains. This is the same logic as antihypertensive therapy: no one expects to take a blood pressure medication for six months and then stop indefinitely while maintaining the effect. The biological factors driving the elevated blood pressure have not been cured; they have been managed.

The safety data from SUSTAIN, STEP, and SELECT trials extend to 3.5 to 4 years, with no new safety signals emerging in longer-term observational data. Concerns that received considerable attention include thyroid C-cell tumor risk (seen in rodent models at very high doses, not yet replicated in human data from trial follow-up), pancreatitis (modestly elevated rate in early studies, attenuated in larger trials), and gastroparesis with prolonged high-dose use.

The enduring question is what happens at year ten or twenty of continuous GLP-1 therapy. This will not be knowable from existing trial data for some years. The practical clinical approach: for patients with established cardiovascular disease or high obesity-related metabolic risk, the benefits of continuation substantially outweigh the unknown long-term risks based on current evidence. For lower-risk patients, the duration question warrants an annual re-evaluation of risk-benefit.

This is a question I genuinely enjoy because it forces clinical prioritization. If I am in a resource-limited setting, or if a patient can only add one test to their annual labs, what do I choose?

Not the A1c, which will stay normal for years while insulin resistance advances silently. Not a lipid panel, which is already standard and will miss metabolic syndrome components. Not a thyroid panel or a testosterone level, both of which are common requests but less immediately informative about cardiometabolic trajectory than insulin.

Fasting insulin, combined with fasting glucose already present on most standard labs, gives the HOMA-IR. That number tells me the metabolic story in a way no single standard test can. It tells me whether the pancreas is compensating for insulin resistance, how long it has likely been doing so, and how urgent the intervention conversation is.

A second choice would be ApoB for the atherogenic particle story. A third choice would be lipoprotein(a) for the genetic cardiovascular risk component. But for metabolic assessment specifically, fasting insulin is the gap in standard preventive care that returns the highest clinical yield.

The irony is that the test is cheap, widely available, and rarely ordered. When I trained in internal medicine, I did not routinely check fasting insulin. I learned to do so later, in clinical practice, when I started seeing the divergence between patients' standard lab results and their actual metabolic trajectories. That is a foible of medical education that I have tried to correct in how I practice now.

An A1c of 5.6 is one point below the formal prediabetes threshold of 5.7, but the number is not the clinical point. The clinical point is that you have a measurable signal of metabolic drift and an intervention window that is genuinely open. The DPP trial showed that lifestyle intervention produces its most dramatic diabetes prevention at exactly this stage, before the pancreas has been stressed into significant beta cell loss.

Tomorrow, literally: arrange fasting labs for glucose and insulin. If your last physical did not include both, these can often be added to a repeat draw at your lab with a physician order, or requested at your next visit. While waiting for those results, take a twenty-minute walk after dinner. This is not a symbolic gesture; as discussed in Q38, it activates GLUT4 uptake in muscle cells and reduces postprandial glucose for the next eighteen hours.

This week: review what you are eating at breakfast and lunch. The first meal of the day sets the glucose and insulin pattern for the following hours. A breakfast of refined carbohydrates with no protein or fat will produce a glucose spike and compensatory insulin surge that increases hunger by late morning and sets an insulin-resistant metabolic tone for the rest of the day. A breakfast with protein and fiber, or no breakfast with a morning walk, produces a different pattern.

Over the next three months: prioritize resistance training three days per week, prioritize seven to eight hours of sleep, and monitor how you feel metabolically. When you return for follow-up labs, if the fasting insulin is normal and the A1c has held or improved, you have evidence that your intervention is working. If the insulin is elevated, the conversation about medications, dietitian referral, and more intensive monitoring begins.

An A1c of 5.6 is not a crisis. It is a message from your metabolism that it has been trying to maintain something it is finding increasingly effortful. The message is legible. The response is clear. The window is open.