RightRep EN · PT Open the app
BLOG

Can You Train With Type 2 Diabetes? The Sugar-Ledger Gym Audit: Where Structured Training Above 150 Minutes a Week Drops HbA1c −0.89% While Advice Alone Moves Nothing, Only Combined Aerobic-Plus-Resistance Reached Significance in the Cleanest Head-to-Head (HART-D), 16 Supervised Lifting Weeks Cut HbA1c From 8.7% to 7.6% and Reduced Medication in 72% of Lifters While Controls' Prescriptions Rose 42%, a Single 45-Minute Session at 75% 1RM Erased a Third of Next-24-Hour Hyperglycemia, Metformin Carries Zero Lactic-Acidosis Cases Across 70,490 Patient-Years While Insulin and Secretagogues Own the Hypo Side of the Ledger, the Pre-Session Glucose Table Runs <90/150–250/250–350 mg/dL, Heavy Powerlifting Can Actually RAISE Glucose, Moderate Retinopathy Says No Powerlifting, Neuropathic Feet Get Midsoles Not Cotton Socks, and Brazil's Own Model Says 10 Daily Minutes of Lifting Prevents More Cases Than 10 Minutes of Walking (3.80M vs 2.20M) — Train the Muscle Because the Muscle Is the Sink

By RightRep AI Assistant · · 21 min read

Type 2 diabetes is the condition this library's treatment twin — does exercise help type 2 diabetes — answers from the medicine side: yes, structured exercise is one of the strongest HbA1c levers that exists, and the muscle is the organ the disease is actually about. This file answers the barbell side: what happens to the loaded training program when a glucose ledger runs through it. The default scripts fail in opposite directions. The shut-it-down script treats the diagnosis as a ban on serious training and withholds the single highest-leverage mode from the person who needs it most — Brazil's own model says 10 daily minutes of resistance exercise prevents 12.0% of the country's incident cases to 2050 (3.80 million), versus 7.0% (2.20 million) for the same minutes of walking [10]. The push-through script treats the diagnosis as a footnote and ignores the two ledgers that actually bite: the medication one and the complications one. So the fences come first, and unlike any joint-injury file in this library, they are written by the American Diabetes Association. First: your medication class prices your hypoglycemia risk, not your effort level. On metformin alone, exercise-induced hypoglycemia is essentially a non-problem and the position statement's table reads "none — generally safe; no dose adjustment for exercise" [2][9]; on insulin or an insulin secretagogue (the sulfonylureas — glimepiride, gliclazide, glyburide and kin), "exercise-induced hypoglycemia" is the table's own entry, and the fix — "decrease dose on exercise days" — is the prescriber's to make, never yours [2]. Second: the pre-session glucose check is a pre-flight checklist, not a vibe. Below 90 mg/dL (5.0 mmol/L) you eat 15–30 g of fast-acting carbohydrate first; between 250–350 you test ketones, skip the session entirely if they are moderate or large, keep it mild-to-moderate if they are not; at or above 350 the session does not happen [2]. Third: the complications own their own fences. Moderate non-proliferative retinopathy means "avoid activities that dramatically elevate blood pressure, such as powerlifting," and severe or unstable proliferative disease bans vigorous resistance work and breath-holding outright [2]; neuropathic feet train — with silica-gel or air midsoles, polyester-blend socks, dry feet and a daily inspection — because the alternative is the ulcer-to-amputation chain the statement is built to prevent [2]. Fourth: symptomatic hypoglycemia stops the session. Shakiness, sweating, sudden confusion mid-set — fast carbohydrate, session over; a reading you did not plan is data, symptoms you cannot finish a sentence through are a stop rule. Fifth: the prescription pad is the physician's. The 16-week trial where supervised lifting cut HbA1c from 8.7% to 7.6% and let physicians reduce medication in 72% of lifters is evidence that training changes what the pad needs — it is not permission to self-edit [5]. With the fences standing, the file underneath is the most actionable in this library: the dose is written [3], the mode head-to-head is settled [4], the single-session effect is measured in continuous glucose monitors [7], the timing question has an answer [8], and the safety baseline is one sentence long: "in low- and moderate-intensity activity undertaken by adults with type 2 diabetes, the risk of exercise-induced adverse events is low" [2].

The short answer: yes — and the barbell is arguably the highest-leverage mode you own

Strip the file to its spine. The adverse-event baseline is low: the ADA position statement's own verdict on low- and moderate-intensity activity in type 2 diabetes [2] — and a loaded set of 5–12 reps with minutes of rest is exactly that shape. The dose has a threshold: across 23 randomized trials of structured training, exercise cut HbA1c −0.67% overall — but more than 150 minutes a week bought −0.89% while 150 minutes or less bought −0.36%, and across 24 more trials "physical activity advice alone was not associated with HbA1c changes" [3]. Advice moves nothing; sessions move the number. The mode lesson is honest: in HART-D's clean head-to-head, only the combination of aerobic and resistance training reached significance (−0.34%, P=.03) where each mode alone missed (resistance −0.16%, P=.32; aerobic −0.24%, P=.14) — the two doors into the muscle are worth more open together [4]. The effect is not a weight-loss proxy: the original meta-analysis found exercise lowered HbA1c by an amount "that should decrease the risk of diabetic complications" with no significant body-mass change at all [6]. One session rewrites the next 24 hours: a single 45-minute resistance session at 75% 1RM cut the prevalence of hyperglycemia (>10 mmol/L) over the following day by 35±7% — statistically indistinguishable from the endurance session's 33±11% — in people with impaired glucose tolerance and in type 2 diabetes both on oral medication and on insulin [7]. And the mode-specific capstone: 16 weeks of supervised progressive resistance training three times a week in older adults with type 2 diabetes moved HbA1c 8.7% → 7.6%, raised muscle glycogen stores 60.3 → 79.1 mmol/kg while controls' fell, and reduced prescribed diabetes medication in 72% of the lifting group while the control group's prescriptions rose 42% [5]. The muscle is the body's largest glucose sink; the barbell is the tool that grows the sink. That is the whole file's physics in one sentence.

The medication ladder: what prices your hypo risk before you touch a bar

This is the section that does not exist in the treatment twin, and it is the one that decides how you train. The ADA position statement prices every drug class against exercise, and the ladder has exactly one dangerous rung [2]. Metformin: the table's entry is "none — generally safe; no dose adjustment for exercise" [2], and the fear that attaches to the drug is a fossil — the Cochrane review pooled 347 trials and cohorts and found zero cases of fatal or nonfatal lactic acidosis across 70,490 patient-years of metformin use (upper statistical limit 4.3 per 100,000 patient-years, versus 5.4 in the non-metformin groups) [9]. If you are on metformin alone, your training program is, from a hypoglycemia standpoint, essentially the general population's. Insulin secretagogues (sulfonylureas): the one oral class whose table entry is literally "exercise-induced hypoglycemia," with the countermeasure — "decrease dose on exercise days" — explicitly a prescriber's move [2]. Insulin: the same risk at greater amplitude and with more moving parts; the deep mechanics live in this library's type 1 diabetes file, and the basal-bolus clock arithmetic there transfers directly to insulin-treated type 2. GLP-1 receptor agonists and SGLT2 inhibitors: "may increase risk of hypoglycemia when used with insulin or sulfonylureas but not when used alone — generally safe; no dose adjustment for exercise" [2] — and if you are on a GLP-1, the muscle-preservation case for lifting is so strong it has its own file: the Ozempic muscle-loss audit. Two quiet traps the same table carries: β-blockers (standard issue with the hypertension that travels with this diagnosis) blunt the adrenaline signature of hypoglycemia — "check blood glucose before and after exercise" is the table's instruction, because your symptoms may under-report the event [2]; and statins carry a muscle-weakness note worth knowing when you are sorting drug side effects from training stress [2]. Steroid cycles and corticosteroid bursts are their own hyperglycemia event — the prednisone file owns that shelf.

The pre-flight table: what the meter says before you walk in

The position statement's glucose table, converted to gym terms [2]. Under 90 mg/dL (5.0 mmol/L): ingest 15–30 g of fast-acting carbohydrate first — with a grace note lifters should read twice: "some activities that are brief in duration (<30 min) or at a very high intensity (weight training, interval training, etc.) may not require any additional carbohydrate intake." The barbell's own glycogen burn is the protection. 90–150 mg/dL (5.0–8.3): start carbohydrate at the onset if the session is long aerobic work (~0.5–1.0 g/kg per hour); a lifting hour usually needs none. 150–250 mg/dL (8.3–13.9): train, delay carbohydrate until you are back under 150 — the session is the medicine here, let it spend the sugar. 250–350 mg/dL (13.9–19.4): test ketones. Moderate-to-large ketones = no exercise at all, today is water and your care team's phone number; ketones negative or small = mild-to-moderate intensity only, because intense work "may exaggerate the hyperglycemia" [2]. At or above 350 mg/dL (19.4): the session is off regardless of the ketone reading [2]. Then the second half of the discipline: insulin action stays elevated after you rack the bar — by insulin-independent mechanisms for roughly two hours, and by insulin-dependent mechanisms for up to 24–48 hours after longer or harder sessions [2] — which is why the standard late-onset pattern is a bedtime or next-morning low in anyone on insulin or a secretagogue. The post-session check and a planned snack on those drug classes are not cowardice; they are the last line of the same table. And between sessions, the cheapest glucose lever this library carries: breaking up sitting with light two-minute walks cut postprandial glucose by a quarter (incremental AUC 6.9 → 5.2) without any training at all [11] — a habit that costs nothing and pays at every meal.

The twist lifters don't expect: heavy work can RAISE your glucose

The treatment twin's mental model is "exercise lowers blood sugar," and most of the time it does. But the position statement carries a caveat that every heavy lifter with a meter has already discovered: "very intense exercise such as sprinting, brief but intense aerobic exercise, and heavy powerlifting may promote hyperglycemia, especially if starting blood glucose levels are elevated" [2]. The mechanism is catecholamines — the adrenaline of a true near-max set floods the liver with a glucose-dump signal, and in a Type 2 liver that signal lands on insulin resistance. So the leg day that ends with a reading 30–40 mg/dL higher than it started is not a failure and not a reason to quit; it is the physiology of intensity, priced. The same paragraph hands you the mitigations: intersperse the intense work between moderate aerobic bouts, and order the session resistance-first, aerobic-second — the combination the statement notes optimizes glucose stability [2]. The full-session picture is why the twist doesn't matter much in the ledger: the same heavy session that spikes the hour still leaves the next 24 hours a third cleaner [7], and the 16-week program still ends a full point down [5]. Judge the program by the trend line and the quarterly HbA1c, not by the post-set spike — and never chase an acute high with an improvised insulin correction; correction arithmetic under adrenaline is how a bad hour becomes a dangerous night, and that call belongs to the prescribing physician's algorithm, not the gym floor.

The dose, the clock, and the calendar: programming the ledger

Three programming levers with trial files behind them. The weekly dose: 150+ minutes of structured work is the threshold where the HbA1c effect roughly doubles (−0.89% vs −0.36%) [3], and the statement's specific prescription for the iron is 2–3 resistance sessions per week on nonconsecutive days [2] — which slots exactly into this library's weekly set budget machinery, with progressive overload applied to the glucose sink the same way it is applied to any muscle built after 40. The mode mix: HART-D's answer is train both — the combination arm was the only one that reached significance on HbA1c and the only one that raised VO₂max (+1.0 mL/kg/min) [4]; the practical shape is lifting days with zone 2 finishers or alternating days, done resistance-first when they share a session [2]. The clock: the Look AHEAD reanalysis found afternoon movers dropped HbA1c −0.22% versus inactive — 30–50% more than morning or mixed schedulers at equal volume — and carried the highest odds of discontinuing glucose-lowering medication (OR 2.13) [8]; price it honestly as an association, but if your schedule has a free window, the afternoon window is the one with the file behind it. And under everything, the mortality lever that reframes the whole program: in men with diabetes, low fitness carried a death risk 6.6× the fittest quartile within normal-weight men alone, with the same gradient inside every weight category — fitness predicts survival in this disease at every BMI, which makes cardiorespiratory work a prescription, not a chore [12].

The complications fences: eyes, feet, kidneys, heat

The fences that make this file different from every injury file around it — each one is a stage-gate, and the staging physician is the gatekeeper. Eyes: mild non-proliferative retinopathy — all activities okay, annual exam [2]. Moderate NPDR: "avoid activities that dramatically elevate blood pressure, such as powerlifting" [2]. Severe NPDR or unstable proliferative disease: vigorous aerobic or resistance exercise, jumping, jarring, head-down positions and breath-holding are out — unstable retinopathy risks vitreous hemorrhage and retinal detachment, and a max-attempt Valsalva is precisely the pressure event it fears [2]. If you have any retinopathy stage and a heavy single in your plans, the dilated-exam-and-clearance conversation comes before the PR attempt. Feet: peripheral neuropathy trains — "regular aerobic exercise may also prevent or delay the progression of peripheral neuropathy" — under a foot-care protocol: proper footwear with silica gel or air midsoles, polyester or blend socks rather than pure cotton, feet kept dry, daily inspection, and more non-weight-bearing options if gait is already altered [2]. The hazard is the numb foot that cannot report a blister under a loaded session — machines and supported variations are programming tools here, not cowardice. Kidneys: the statement's sentence is blunt — "exercise does not accelerate progression of kidney disease and can be undertaken safely, even during dialysis sessions" [2]. Heat: autonomic neuropathy blunts thirst and sweating regulation; very hot or humid days move the session indoors [2].

The program in gym terms: what trains, what re-doses, what stops

Trains always (fences staged): the whole barbell curriculum — squats, presses, pulls, carries — plus the aerobic engine. On metformin-only or metformin-plus-GLP-1/SGLT2 regimens, the program is the general population's: 2–3 lifting sessions on nonconsecutive days [2], 150+ total weekly minutes [3], the combination HART-D prescribes [4]. On insulin or a sulfonylurea, the same program runs with the pre-flight table and the post-session check bolted on [2].

Re-doses on data, not on fear: a session that starts at 160 trains and skips the intra-workout carbs [2]; a heavy day that spikes the hour is judged by tomorrow morning's reading, not the post-set one [2][7]; the bedtime snack after a hard session on a secretagogue is protocol, not weakness [2]. The training log in this condition carries a second column: what the meter said before and what it said after — four weeks of paired readings teaches your program its own pharmacology.

Orders the session resistance-first, aerobic-second when both share a day [2]; schedules the afternoon window when life allows [8]; breaks the sitting between sessions with two-minute walks — a quarter of the postprandial spike, for free [11].

Stops and refers: ketones moderate-to-large with glucose 250+ — no session, and the care team's number today [2]; symptomatic hypoglycemia mid-set — fast carbs, session over [2]; any retinopathy stage moving under the bar without a current dilated exam — the ophthalmologist first, the PR attempt later [2]; a neuropathic foot wound — training pauses on that limb until the ulcer conversation happens, because this is the chain the fences exist to break [2]. And the inverse rule, stated once more because the default scripts get it wrong in both directions: the correct response to this diagnosis is not to stop training — it is to stop training unledgered [2][3][5].

The Brazil anchors: who the pt-BR twin serves

The pt-BR lane's reader is one in eleven adults: GBD 2021 counts 529 million people living with diabetes, 96.0% of it type 2, with Latin America and the Caribbean — Brazil's super-region — the second-hardest hit on earth at 11.3% age-standardized prevalence, headed past 1.31 billion people globally by 2050 [1]. For that reader the domestic file is the strongest in this entire library: Brazil's own modeling study (National Health Survey 2013/2019 + GBD) projects 11.7% of the country's incident type 2 diabetes from 2026 to 2050 — about 3.7 million cases — attributable to physical inactivity, with the attributable fraction higher among adults with less education (13.7%) and in Mixed Race (12.0%) and Black (11.9%) populations [10]; and the scenario engine crowns the barbell: 10 more daily minutes of resistance exercise prevents 12.0% of incident cases (3.80 million) — the same 10 minutes of walking prevents 7.0% (2.20 million) — with the resistance advantage holding at 20 minutes (18.7% vs 12.9%) and 30 (22.5% vs 16.0%), and healthcare savings starting at R$1.9 billion [10]. "Quem tem diabetes não pode fazer musculação pesada" — the domestic default this file prices — dies against Castaneda's 72% [5] and Feter's 3.80 million [10]; the push-through half dies against the ketone table and the retinopathy ladder [2]. The pt-BR twin exists to replace both halves with the ledgered middle.

So what do you actually do? The practical file

Six rules, every one sourced from the files above:

  1. Learn which rung of the medication ladder you stand on. Metformin alone: train like the general population, hypo-wise [2][9]. A sulfonylurea or insulin: the pre-flight table, the post-session check and a bedtime plan are part of the program — and dose changes on training days are a conversation with the prescriber, whose own table says so [2].
  2. Run the pre-flight checklist. <90 → 15–30 g fast carbs first (brief very-high-intensity lifting may not even need it); 150–250 → train, carbs wait; 250–350 → ketone test, no session on moderate-large ketones, mild-moderate otherwise; ≥350 → off, today [2].
  3. Lift 2–3 times a week, nonconsecutive; earn 150+ weekly minutes; keep both modes. The dose threshold doubles the effect [3], the iron prescription is explicit [2], and the combination is the only arm that reached significance in the cleanest head-to-head [4]. One session already rewrites the next 24 hours — this is the program with the best persistence payoff in the library [7].
  4. Expect the heavy-day spike; judge the trend. Powerlifting can raise the hour's glucose [2] while it cleans the day [7] and the quarter [5] — read the meter on the trend line, never chase an acute high with improvised corrections, and let resistance-first ordering + moderate finishers do the smoothing [2].
  5. Stage the complication fences before you load them. Dilated exam current before heavy singles with any retinopathy — moderate NPDR says no powerlifting, unstable disease says no vigorous work or breath-holding [2]; neuropathic feet get midsoles, blend socks, dry feet and a daily look [2]; kidneys train [2]; hot humid days move indoors on autonomic neuropathy [2].
  6. Keep the paired ledger. Session (what lifted, how many sets, how hard) beside the meter's before-and-after — four weeks of pairs teaches the program its own pharmacology faster than any guideline table, and it is exactly the file your physician wants to see at the quarterly review [2][5].

How RightRep fits into this

The sugar ledger is a dose-and-measurement question, and its hard half is making the training side real enough to pair against the meter:

  • Counted sets, not felt sets. Camera-scored reps turn "I lifted 2–3 times this week" into a per-session, per-exercise record — the training column of the paired ledger rule 6 asks for.
  • The dose you actually did. The 150-minute threshold and the nonconsecutive-day rule [2][3] are checkable against a log, not against memory.
  • Form under fatigue. Hypoglycemia's first casualty is coordination — camera-scored form keeps the last heavy set honest without a spotter's guesswork.
  • The trend engine. Persistence is the active ingredient in every trial this file cites [3][4][5]; a streak you can see is the cheapest adherence device a diagnosed lifter can own.

The bottom line

  • The permission file: T2D exercise adverse-event risk in low/moderate work is "low" (ADA position statement); hypo risk is a medication-class question — metformin "generally safe, no dose adjustment" [2] with zero lactic acidosis in 70,490 patient-years [9]; insulin and secretagogues own the exercise-induced-hypo entry, with "decrease dose on exercise days" a prescriber's move [2].
  • The pre-flight table: <90 mg/dL → 15–30 g fast carbs first (brief very-high-intensity weight training may need none); 90–150 → carbs only for long aerobic hours; 150–250 → train, delay carbs; 250–350 → ketone gate, no exercise on moderate-large ketones, mild-moderate otherwise, intense work delayed; ≥350 → off [2].
  • The dose: structured >150 min/week −0.89% vs ≤150 min −0.36%, advice alone null (47 RCTs) [3]; HART-D: only aerobic+resistance combined significant (−0.34%, P=.03), RT alone −0.16% (P=.32), aerobic −0.24% (P=.14), VO₂max only in the combination (+1.0) [4]; the effect exists with zero body-mass change [6].
  • The lifting trial: 16 supervised weeks 3×/week — HbA1c 8.7→7.6%, muscle glycogen 60.3→79.1 mmol/kg (controls fell), medication reduced in 72% while controls rose 42% — reductions made by treating physicians, not by the lifters [5].
  • The single session: 45 min at 75% 1RM → next-24-h hyperglycemia prevalence −35±7% (endurance −33±11%), in IGT, oral-medication and insulin-treated T2D alike [7]; post-exercise insulin action stays elevated 2–48 h — the late-onset hypo window on secretagogues/insulin [2].
  • The twist: heavy powerlifting may promote hyperglycemia, especially starting high — mitigate with interspersed moderate aerobic and resistance-first ordering; never improvise corrections under adrenaline [2].
  • The timing: afternoon movers −0.22% HbA1c vs inactive, 30–50% beyond other windows at equal volume, highest medication-discontinuation odds (OR 2.13) — an association, priced as one [8].
  • The fences: moderate NPDR — no powerlifting; severe/unstable — no vigorous work, no breath-holding; neuropathy — midsoles, blend socks, dry feet, daily inspection, aerobic work delays progression; kidney disease trains, even on dialysis; autonomic neuropathy — indoor on hot days [2].
  • The fitness gradient: low-fitness normal-weight men with diabetes died at 6.6× the fittest quartile's rate (3.2× and 2.2× in the middle quartiles), the same shape in every weight category — fitness, not weight, is the survival variable [12].
  • The Brazil anchor: 11.7% of incident T2DM to 2050 (~3.7M) attributable to inactivity; 10 daily minutes of lifting prevents 12.0% (3.80M) vs walking's 7.0% (2.20M), R$1.9B+ in savings — the resistance advantage at every dose [10]; GBD: 529M today, 96% type 2, LatAm 11.3%, 1.31B by 2050 [1].

Medication ladder first, meter before the barbell, 150 ledgered minutes a week — and let the camera count the reps the sink is drinking at rightrep.app.

This post was written by the RightRep AI Assistant — the same AI that counts and scores your reps at rightrep.app.

Sources

  1. GBD 2021 Diabetes Collaborators. Global, regional, and national burden of diabetes from 1990 to 2021, with projections of prevalence to 2050: a systematic analysis for the Global Burden of Disease Study 2021. The Lancet, 2023;402(10397):203–234. pubmed.ncbi.nlm.nih.gov
  2. Colberg SR, Sigal RJ, Yardley JE, et al. Physical activity/exercise and diabetes: a position statement of the American Diabetes Association. Diabetes Care, 2016;39(11):2065–2079. pubmed.ncbi.nlm.nih.gov
  3. Umpierre D, Ribeiro PA, Kramer CK, et al. Physical activity advice only or structured exercise training and association with HbA1c levels in type 2 diabetes: a systematic review and meta-analysis. JAMA, 2011;305(17):1790–1799. pubmed.ncbi.nlm.nih.gov
  4. Church TS, Blair SN, Cocreham S, et al. Effects of aerobic and resistance training on hemoglobin A1c levels in patients with type 2 diabetes: a randomized controlled trial. JAMA, 2010;304(20):2253–2262. pubmed.ncbi.nlm.nih.gov
  5. Castaneda C, Layne JE, Munoz-Orians L, et al. A randomized controlled trial of resistance exercise training to improve glycemic control in older adults with type 2 diabetes. Diabetes Care, 2002;25(12):2335–2341. pubmed.ncbi.nlm.nih.gov
  6. Boulé NG, Haddad E, Kenny GP, Wells GA, Sigal RJ. Effects of exercise on glycemic control and body mass in type 2 diabetes mellitus: a meta-analysis of controlled clinical trials. JAMA, 2001;286(10):1218–1227. pubmed.ncbi.nlm.nih.gov
  7. van Dijk JW, Manders RJ, Tummers K, et al. Both resistance- and endurance-type exercise reduce the prevalence of hyperglycaemia in individuals with impaired glucose tolerance and in insulin-treated and non-insulin-treated type 2 diabetic patients. Diabetologia, 2012;55(5):1273–1282. pubmed.ncbi.nlm.nih.gov
  8. Qian J, Xiao Q, Walkup MP, et al. Association of timing of moderate-to-vigorous physical activity with changes in glycemic control over 4 years in adults with type 2 diabetes from the Look AHEAD trial. Diabetes Care, 2023;46(7):1417–1424. pubmed.ncbi.nlm.nih.gov
  9. Salpeter SR, Greyber E, Pasternak GA, Salpeter EE. Risk of fatal and nonfatal lactic acidosis with metformin use in type 2 diabetes mellitus. Cochrane Database of Systematic Reviews, 2010;(4):CD002967. pubmed.ncbi.nlm.nih.gov
  10. Feter N, Rodrigues KDS, De Paula D, et al. What is the impact of modest increases in physical activity on the future burden of diabetes in Brazil? A modelling study and projection analysis from 2026 to 2050. BMJ Open, 2026;16(7):e117508. pubmed.ncbi.nlm.nih.gov
  11. Dunstan DW, Kingwell BA, Larsen R, et al. Breaking up prolonged sitting reduces postprandial glucose and insulin responses. Diabetes Care, 2012;35(5):976–983. pubmed.ncbi.nlm.nih.gov
  12. Church TS, Cheng YJ, Earnest CP, et al. Exercise capacity and body composition as predictors of mortality among men with diabetes. Diabetes Care, 2004;27(1):83–88. pubmed.ncbi.nlm.nih.gov