Biology: Insulin & Glucose

Under euglycemic hyperinsulinemic clamp conditions, roughly 80% of glucose uptake occurs in skeletal muscle, and adipose tissue takes up less than 5% of an infused glucose load. In the fasted state the picture inverts: about 70 to 75% of glucose uptake happens in tissues that do not need insulin at all, chiefly brain, red cells and splanchnic tissue. Measured in: Synthesis of euglycemic clamp and limb-catheterization studies in adults with and without type 2 diabetes. The 80% figure describes a laboratory state, insulin infused to a steady high level with glucose held constant, which is not the same as a mixed meal eaten at a table. It sets the proportions, not the exact share of any given dinner.

Contraction raises muscle glucose uptake by up to 50-fold through signaling that is distinct from the insulin pathway, involving AMPK, calcium and nitric oxide upstream of GLUT4 movement to the membrane. That route is preserved in insulin-resistant muscle, which is why exercise lowers glucose in people whose insulin has stopped working well. Measured in: Review of human and rodent muscle physiology, biopsy and tracer work. Both sources are narrative reviews rather than primary measurements, and their own position is that the pathway attribution is unsettled and redundant by design. The strong and quotable part is the preservation of the contraction route in insulin-resistant muscle; the molecular detail beneath it is the soft part.

Under two 8-day laboratory protocols separating circadian phase from behavior, identical meals produced 17% higher postprandial glucose in the biological evening than in the biological morning, with early-phase insulin 27% lower. Circadian misalignment on top of that added a further 6%. Measured in: Healthy adults in a randomized crossover with a 12-hour behavioral inversion, so the same meals were eaten at opposite circadian phases. A behavioral inversion rather than a forced-desynchrony protocol; the authors contrast this design with their own earlier forced-desynchrony work. It isolates circadian phase from meal content, and it does not isolate it from the disruption of inverting a day.

Insulin release to an intravenous glucose load is biphasic: a rapid burst of stored insulin, then a slower sustained phase. Abolishing the early phase experimentally in healthy people produces impaired glucose tolerance and excessive glucose excursions, and restoring an early insulin rise in people with type 2 diabetes improves the post-meal glucose profile. Measured in: Review of human and animal physiological studies of early-phase insulin secretion. Most of the clean phase separation comes from intravenous glucose, which is not how anyone eats. After a mixed meal the two phases blur into one another, and the early rise is smaller and slower than the intravenous experiment shows.

After an overnight fast, whole-body glucose production was higher in people with type 2 diabetes than in controls, 11.1 against 8.9 micromol per kg per minute. Net hepatic glycogenolysis was lower, 1.3 against 2.8, so the excess came from gluconeogenesis, which accounted for 88% of glucose production against 70% in controls. Measured in: 7 people with type 2 diabetes and 5 controls, studied with 13C magnetic resonance spectroscopy across 23 hours of fasting. Twelve people in total, measured in one laboratory in 1992 with a method that was new at the time. The direction has held up in later work; the exact percentages come from a very small sample.

Using a labeled meal and an oral glucose tolerance test, post-meal hyperglycemia was driven by lower rates of glucose disappearance from the blood, not by glucose arriving faster from the gut and not by failure to switch off the liver. Endogenous glucose production was promptly suppressed and meal glucose appeared at a comparable rate in both groups. Measured in: 32 people with impaired fasting glucose and 28 with normal fasting glucose, each studied on two occasions. Sixty people at one center, and the tracer method partitions glucose fluxes by model rather than measuring each tissue directly. It tells you that disposal was the limiting step without saying which muscle bed it happened in.

After 45 to 60 minutes of ergometer exercise at 60 to 70% of VO2max, plasma-membrane GLUT4 rose 74 ± 20% above resting values in people with type 2 diabetes and 71 ± 18% in controls, measured on open muscle biopsy. It rose in every participant. Measured in: 10 people, 5 with type 2 diabetes (2 men, 3 women) and 5 controls (all men), each giving two open biopsies of the same muscle 3 to 6 weeks apart from opposite legs, one at rest and one after exercise. This trial had no insulin arm. The insulin-stimulated defect it is usually contrasted against was reported in earlier separate studies rather than measured in these ten people, so the contrast is assembled across papers rather than observed within one. Five people per group, one exercise bout, and the outcome is transporter protein in a membrane fraction rather than glucose leaving the blood. Resting membrane GLUT4 was about 32% lower in the diabetes group, which did not reach significance at this size.

During reproductive life, women store fat subcutaneously in preference to visceral and ectopic sites, have higher insulin sensitivity than men, and have greater capacity for insulin secretion and a larger incretin response. Those advantages disappear as glucose tolerance deteriorates toward diabetes. In most of the world diabetes is more prevalent in men, most markedly in middle age. Measured in: Review of human and animal work on sex differences in energy balance and glucose metabolism. A narrative review, and much of the mechanistic support is from rodent models where males reliably develop insulin resistance and hyperglycemia more readily than females. The human clamp studies behind the sensitivity difference are individually small.

Measured up to eight times per cycle with visits timed by fertility monitor, HOMA-IR rose from 1.35 in the mid-follicular phase to 1.59 in the early luteal phase and 1.55 in the late luteal phase. The change came from insulin rather than glucose. HOMA-IR tracked estradiol and progesterone positively and FSH and SHBG inversely. The authors describe the variation as minor. Measured in: 257 healthy premenopausal women, mean age 27, mean BMI 24, followed across one or two cycles. What could explain it instead: Diet, sleep and activity also shift across the cycle and were not controlled, so part of the change attributed to hormones may be behavior that travels with them.. The swing is small: about 0.2 HOMA-IR units, well inside the range a single person's fasting insulin moves for other reasons. It is a reason to standardize cycle phase in research, not a reason to plan a week around it.

Followed annually for four years, all the women gained subcutaneous abdominal fat with age, and only those who became postmenopausal gained visceral fat and total body fat. Sleeping energy expenditure fell 1.5 times further in that group (-7.9% against -5.3%), and fat oxidation fell 32% in the women who became postmenopausal and did not change in those who did not. Physical activity fell significantly two years before menopause and stayed low. Measured in: 156 healthy initially premenopausal women (103 Caucasian, 53 African-American), 51 of whom became postmenopausal during follow-up; 24-hour energy expenditure measured by whole-room calorimeter in 34. What could explain it instead: Activity dropped two years before menopause, so behavior was already changing before the hormonal transition completed, and part of the fat gain attributed to menopause may follow the activity change instead.. Observational, and aging and menopause happen together, so separating them relies on comparing women who crossed the transition against women of similar age who did not. The energy expenditure measurements come from a subset of 34.

The dip below baseline 2 to 3 hours after a standardized meal predicted hunger and later eating better than the peak did. Larger dips tracked more hunger at 2 to 3 hours (r = 0.16), a shorter gap to the next meal (r = -0.14), greater energy intake at 3 to 4 hours (r = 0.19) and greater intake over 24 hours (r = 0.27). Direction was consistent in a US validation cohort. Measured in: 1,070 adults across a UK exploratory and a US validation cohort, eating 8,624 standardized meals followed by 71,715 free-living meals with continuous monitoring. What could explain it instead: Hunger and subsequent intake were self-reported under free-living conditions, so people whose glucose dips more may also differ in habitual meal composition, sleep and activity, all of which independently affect appetite.. The correlations are small, between 0.14 and 0.27, which at this sample size is a reliable signal explaining a small share of the variation. Several authors are employed by the company that sells the monitoring program the data came from.

In age and BMI-matched severely obese men and women studied with a two-step clamp and glucose tracer, men had lower hepatic insulin sensitivity: insulin suppressed endogenous glucose production by 61.7% in men against 72.8% in women (p = 0.028). Adipose tissue insulin sensitivity, peripheral glucose disposal, basal glucose production and liver fat content did not differ. Measured in: 46 severely obese adults, mean age 48 and 46, mean BMI 41 in both groups; liver fat measured in a subset of 27. Forty-six people, all severely obese with a mean BMI of 41, so it does not describe a lean population. It is one study and the difference it found was confined to the liver.

In the metabolic substudy of the DiRECT remission trial, liver fat fell from 16.0% to 3.1% immediately after weight loss, and pancreas fat and plasma triglyceride fell whether or not glucose control normalized. What separated the people who reached non-diabetic glucose control was recovery of the first-phase insulin response, from 0.04 to 0.11 nmol/min/m2, still present at 12 months. Responders had shorter diabetes duration than non-responders, 2.7 against 3.8 years. Measured in: Metabolic substudy of the Diabetes Remission Clinical Trial: 64 in the intervention group and 26 controls. Comparing responders with non-responders inside a trial is an observational comparison, not a randomized one, so the duration difference and the beta cell recovery could both be markers of how far the disease had gone rather than a cause and its effect.

Using a sequential multi-step euglycemic clamp on each of three occasions, at rest, immediately after 60 minutes of ergometer exercise at 150 W, and 48 hours later, both insulin sensitivity and responsiveness were still improved at 48 hours. Apparent Km fell from 52 ± 3 to 43 ± 4 and then 40 ± 3 µU/ml; Vmax rose from 9.5 ± 0.8 to 10.9 ± 0.7 and 10.7 ± 0.8 mg/min/kg. In three further participants nothing remained at five days. Measured in: 7 untrained men, with 3 more studied at 5 days. Seven untrained men in one laboratory in 1988, with the five-day null resting on three of them. What persisted at 48 hours was specifically the conversion of glucose to glycogen, still 7.2 against 5.7 at rest, while maximal glucose OXIDATION went the other way and was lower after exercise than at rest. So "insulin sensitivity improves" is a compression of a more specific finding. No glucoregulatory hormone or metabolite measured could explain the effect.

After adjustment for age, ethnicity, sex and both generalized and central obesity, each 10% higher skeletal muscle index, meaning muscle mass as a share of body weight, was associated with an 11% lower HOMA-IR (95% CI 6 to 15%) and a 12% lower prevalence of prediabetes or diabetes (95% CI 1 to 21%). The associations were stronger in people without diabetes. Measured in: 13,644 participants in the third US National Health and Nutrition Examination Survey. What could explain it instead: Reverse causation is the obvious one: insulin resistance and the inactivity that accompanies it reduce muscle mass, so low muscle may be a consequence rather than a cause. Physical activity itself is not adjusted for and independently affects both sides.. Cross-sectional, so the arrow could run either way, and muscle mass was estimated by bioelectrical impedance rather than measured by scan. The outcome is HOMA-IR, a fasting-blood index, not a clamp.

Over 7.1 years, treated diabetes occurred in 8.3% on conjugated equine estrogen against 9.3% on placebo, HR 0.88 (95% CI 0.77 to 1.01, p = 0.072). HOMA-IR fell significantly against control in the first year, a between-group difference of -0.53, and the difference was gone at 3 and 6 years. Measured in: 10,739 postmenopausal women aged 50 to 79 who had previously had a hysterectomy, randomized in the Women's Health Initiative estrogen-alone trial. The primary comparison did not reach significance, and diabetes was ascertained by self-reported treatment rather than by testing everyone. The authors state that estrogen should not be used to prevent diabetes, because its other effects rule out long-term use for that purpose.

On 600 kcal a day, fasting plasma glucose normalized within one week, from 166 mg/dL (9.2 mmol/l) to 106 mg/dL (5.9 mmol/l). Insulin suppression of hepatic glucose output improved from 43% to 74% over the same week, against 68% in non-diabetic controls. Liver triacylglycerol fell from 12.8% to 2.9% by week 8, and the first-phase insulin response rose from 0.19 to 0.46 nmol/min/m2. Measured in: 11 people with type 2 diabetes, 9 men and 2 women, mean age 49.5, mean BMI 33.6, studied before and after 1, 4 and 8 weeks; 8 weight-matched non-diabetic participants studied once as a reference. Eleven people, no control arm receiving the same intervention, and mostly men. The reference group was measured once rather than followed, so it anchors the numbers without testing them.

In people with type 2 diabetes and a BMI under 27, repeated 5% weight loss cycles took BMI from 24.8 to 22.5 over 12 months and total body fat from 32.1% to 27.6%. Liver fat content, liver fat export and fasting insulin all fell to normal. Sustained remission, meaning HbA1c under 48 mmol/mol off all glucose-lowering drugs, was reached by 14 of 20, after initial weight loss of 6.5%. Post-meal insulin secretion rose but stayed below the matched controls. Measured in: 20 people with type 2 diabetes and BMI under 27 (13 women, 7 men, mean age 59), with 20 matched normoglycemic controls studied once. Twenty people, no randomization and no control arm undergoing the same weight loss, so the remission rate is not comparable with a randomized trial's. The matched controls were measured once, which anchors what normal looks like without testing the intervention.

Impaired glucose tolerance and impaired fasting glucose identify overlapping but distinct populations, and the site of insulin resistance differs. People with impaired glucose tolerance have marked muscle insulin resistance with only mild hepatic insulin resistance. People with impaired fasting glucose have severe hepatic insulin resistance with normal or near-normal muscle insulin sensitivity. Both show reduced early-phase insulin secretion; only impaired glucose tolerance also shows impaired late-phase secretion. Measured in: Synthesis of clamp and tracer studies in people with impaired fasting glucose, impaired glucose tolerance and normal glucose tolerance. A review by one research group, drawing on their own clamp work as well as others', and the categories it describes are thresholds drawn on a continuous distribution rather than distinct diseases.

Iron deficiency, with or without anemia, raised HbA1c against controls with no matching rise in glucose. Non-iron-deficiency anemias showed a possible fall in HbA1c. So the same blood glucose can read high or low on HbA1c depending on what the red cells are doing. Measured in: 12 studies from 544 screened, in non-pregnant adults not known to have diabetes. Most of the included work is on iron deficiency; the evidence on other erythrocyte abnormalities is thin, and the review says so. It does not quantify how large the shift is or at what degree of deficiency it starts to matter. The population figures here describe the systematic review (12 of 544 studies); the second citation is a narrative clinical review that corroborates the direction rather than being one of those studies.

After adjustment for plasma glucose and other characteristics that correlate with HbA1c, HbA1c was higher in Black than in White participants at every level of glycemia: 0.13 and 0.21 percentage points at normal glucose tolerance in the two samples, 0.26 and 0.30 at prediabetes, and 0.47 in both at diabetes. The gap widened as glucose tolerance worsened. Measured in: 1,581 non-Hispanic Black and White adults aged 18 to 87 in the SIGT study, and 1,967 aged over 40 in NHANES III, none with known diabetes. What could explain it instead: Self-identified race is a social category standing in for unmeasured biology, including red cell lifespan and glycation rate, and for unmeasured differences in access to care, diet and measurement setting. The adjustment covers measured glucose, not those.. The mechanism is unknown, which the authors state as a limitation. The consequence is practical rather than theoretical: the same HbA1c threshold does not correspond to the same glucose in everybody.

Repeating a 75 g oral glucose tolerance test in the same people, 95% of test-retest differences in those with normal glucose tolerance fell within 15% of the median for fasting glucose, 46% for the 2-hour glucose, 61% for fasting insulin and 125% for 2-hour insulin. The variation was overwhelmingly biological rather than analytical. Measured in: 524 adults aged 50 to 74 from a general Caucasian population without known diabetes, each tested twice: 246 with normal tolerance, 198 with impaired tolerance and 80 with newly detected diabetes. What could explain it instead: Prior days' carbohydrate intake, activity, sleep and intercurrent illness all shift a glucose tolerance test and were not controlled between the two visits, so part of what looks like measurement noise is real short-term physiology.. One population, one age band, and repeat tests done under research conditions. That makes these figures a floor on real-world variability rather than a ceiling.

The authors of the HOMA model set out its appropriate use as cohort and epidemiological studies, and name inappropriate uses including measuring beta cell function in isolation, adding that the primary input data have to be robust. Those inputs are not standardized: across 12 commercial insulin immunoassays from 9 manufacturers, among-assay coefficients of variation ran from 12% to 66%, median 24%, and a common reference preparation did not fix it. Measured in: The model paper is a methodological review by the group that built HOMA; the assay data come from an American Diabetes Association workgroup evaluation of commercial insulin methods. Neither paper says HOMA-IR is useless, and it correlates respectably with clamp measurements at group level. The limitation is what happens to a group-level index when one person compares their own number against someone else's, measured on a different assay.

Wearing a blinded current-generation sensor for up to 10 days, mean glucose was 98 to 99 mg/dl (5.4 to 5.5 mmol/l) in every age group except those over 60, where it was 104 mg/dl (5.8 mmol/l). Median time between 70 and 140 mg/dl (3.9 to 7.8 mmol/l) was 96%, median time above 140 mg/dl was 2.1% of the day, about 30 minutes, and mean within-person coefficient of variation was 17%. Measured in: 153 healthy, non-obese, non-pregnant children and adults aged 7 to 80 without diabetes, across 12 US centers. What could explain it instead: The device reads above blood glucose by a variable margin, so the small amount of time recorded above range is partly a property of the measurement rather than of the person.. A non-obese, screened, healthy sample, so it describes what a sensor reads in people selected to be well rather than in the general population. It is also one sensor model, and sensors differ.

Pooled prevalence of isolated impaired glucose tolerance was 8% in women against 5% in men, and isolated impaired fasting glucose 15% in women against 21% in men. Compared with men, women had higher odds of isolated impaired glucose tolerance (OR 1.42, 95% CI 1.23 to 1.65) and lower odds of isolated impaired fasting glucose (OR 0.65, 0.44 to 0.96), with similar odds of the combined state. Measured in: 8 studies suitable for meta-analysis, 52,256 participants of whom 25,263 were women. Eight studies, with moderate certainty for the prevalence estimates and high certainty only for the two between-sex comparisons. The practical consequence, that a fasting-glucose-only screen finds proportionally fewer women, is the authors' own reading rather than something the analysis tested directly.

CGM-estimated fasting and postprandial glucose ran 16 mg/dL (0.9 ± 0.6 mmol/L) and 16 mg/dL (0.9 ± 0.5 mmol/L) above capillary estimates (both p<0.001). The size of the bias varied by test food and by individual. CGM overestimated time above 140 mg/dL (7.8 mmol/L) roughly fourfold, falling to roughly twofold after adjusting for the baseline difference. Measured in: 15 healthy adults, each completing seven laboratory visits with randomized carbohydrate challenges including glucose, whole fruit, blended fruit and commercial smoothies, sampled every 15 minutes for 120 minutes. Fifteen people, one sensor type, one laboratory. The finding is about a specific device against capillary sampling, and it should not be read as a general property of every monitor on the market.

Acarbose, a drug whose action is to blunt the post-meal glucose rise specifically, produced no reduction in major adverse cardiovascular events over a median of five years: 470 events (14%) on acarbose against 479 (15%) on placebo, HR 0.98, 95% CI 0.86 to 1.11. It did reduce progression to diabetes, 13% against 16%, rate ratio 0.82 (0.71 to 0.94). Measured in: 6,522 Chinese adults with coronary heart disease and impaired glucose tolerance, randomized to acarbose or placebo and followed a median of 5.0 years. One trial, in one country, in people who already had coronary disease, and a drug acting at the gut wall is not the same intervention as a muscle taking glucose up during contraction. What it constrains is how much a lower post-meal number by itself can be read to promise.