Part 3 — Lifestyle
Chapter 10
Sleep, Stress, and Hormonal Balance
You have followed the plate, taken your walk, and eaten the same breakfast as yesterday. Yet this morning the number on your meter or sensor is higher than you expected. Before blaming yourself or the breakfast, ask what else happened in the last 24 hours. Perhaps you slept badly, or a family worry or a night shift kept your body on alert. Perhaps a medicine, or a stage of life, is changing how your hormones behave.

Food and movement are the levers most people learn first. Sleep, stress, and hormones are the background conditions in which those levers work. They are not moral tests, and many are not fully in your control: a night-shift nurse in Manila, a taxi driver in Lagos, and a new parent in Berlin cannot simply "sleep more". What you can do is understand how these factors act on glucose, recognise the ones that need medical attention, and make small, realistic changes to the rest.

Section 10.1 looks at sleep: what a short night does to insulin, how sleep apnoea and shift work fit in, and what sleep habits can and cannot do. Section 10.2 explains the stress hormones and how much stress-reduction methods change HbA1c. Section 10.3 covers other hormonal influences, from thyroid conditions to steroid medicines, with one rule throughout: if you suspect one applies to you, tell your care team rather than treating it yourself.


10.1 Sleep and Glucose

What a short night does to insulin

Researchers have measured this directly by limiting volunteers' sleep in a laboratory and then testing how well their bodies respond to insulin. The less glucose the body needs to be given to stay steady, the more insulin resistance (Chapter 1) there is.

Study Who Sleep pattern Main result
Donga 2010, healthy adults[@donga2010] 9 healthy adults One night of about 4 hours vs a normal night Glucose infusion needed during the clamp fell by about 25%; peripheral glucose disposal fell by about 20%
Donga 2010, type 1 diabetes[@donga2010t1] 7 adults with type 1 diabetes One night of 222 vs 469 minutes Glucose infusion rate fell by about 21%
Buxton 2010[@buxton2010] 20 healthy young men 5 hours a night for 7 nights, after 3 nights of 10 hours Insulin sensitivity fell by 20% (IVGTT method) and 11% (clamp method)

Percentages are as reported in the abstracts of the three laboratory studies cited; IVGTT is an intravenous glucose tolerance test. Each study had 7 to 20 participants.

You may see claims that insulin sensitivity falls by 25 to 30 percent after four hours of sleep. Studies like these are the source of such headlines, but their own results range from about 11 to 25 percent depending on the design and the measure, and the volunteers were few, mostly young, and studied in a laboratory. The fair reading is that a short night, and more so a run of short nights, makes the body somewhat less responsive to insulin the next day. It is one reason readings can drift up without any change in food, and not a reason to panic after one bad night. If you use insulin or medicines that can cause lows, share any pattern you notice after short nights with your care team, and do not adjust doses yourself.

Science Corner: How these experiments work Most of these studies used the euglycaemic clamp: insulin is infused at a fixed rate, and glucose is infused at whatever rate keeps blood glucose steady. That glucose infusion rate measures insulin action. The method is precise but needs tubes, tracers, and hours of monitoring, so only a handful of volunteers can take part. Its strength is cause and effect, because the researchers decided how long people slept; large population studies offer the opposite trade, with many people but only associations.

Sleep and diabetes risk over the years

Population studies follow many people for years and ask who develops type 2 diabetes. A meta-analysis of 10 studies with 107,756 participants and 3,586 new cases found higher risk in people reporting short sleep (relative risk 1.28), long sleep (1.48), difficulty falling asleep (1.57), and difficulty staying asleep (1.84).[@cappuccio2010] A larger dose-response analysis of 482,502 participants and 18,443 new cases found a U-shaped pattern, with the lowest risk at 7 to 8 hours a night. Each hour of sleep below 7 hours was associated with a relative risk of 1.09, and each hour above 7 hours with 1.14.[@shan2015]

These are associations. People who sleep poorly may differ in other ways, and long sleep may sometimes signal illness rather than cause it. The link also runs both ways: the ADA notes that disrupted sleep is a risk factor for type 2 diabetes and is also common in people who already have it.[@ada2026s5] For most adults, roughly 7 to 8 hours is a sensible aim, although needs vary.

Sleep apnoea: the sleep problem worth asking about

In obstructive sleep apnoea (OSA), the airway repeatedly narrows or closes during sleep, breathing pauses, oxygen levels dip, and sleep is broken into fragments. The ADA explains that this fragmentation and low oxygen activate the stress (sympathetic) nervous system, which contributes to raised glucose and insulin resistance.[@ada2026s5]

OSA is common in people with type 2 diabetes. The ADA reports that it is four to ten times more frequent with obesity, especially central obesity, and that more than 80 percent of participants in the Look AHEAD trial (people with obesity and type 2 diabetes) had it.[@ada2026s4] A meta-analysis of 18 studies covering 10,754 people with type 2 diabetes found a pooled prevalence of 56 percent, with country estimates from 15 to 90 percent.[@andayeshgar2022] The exact figure matters less than the message: a large share of people with type 2 diabetes may have some degree of apnoea.

When to ask for screening. The ADA advises evaluating for OSA in people with suggestive symptoms: excessive daytime sleepiness, snoring, and witnessed apnoeas (a partner sees you stop breathing).[@ada2026s4] Because you are asleep when it happens, a partner's observation is often the best clue. If any of these apply, tell your doctor; diagnosis is by a sleep study, at home or in a clinic.

What treatment adds. Treatment options include lifestyle change, continuous positive airway pressure (CPAP, a mask that keeps the airway open with gentle air pressure), oral appliances, and surgery, which the ADA says improve quality of life and blood pressure.[@ada2026s4] For glucose, a meta-analysis of 11 randomised trials with 964 people with OSA and type 2 diabetes found that CPAP lowered HbA1c by 0.24 percentage points compared with no active treatment, and the reduction was larger with more hours of nightly use.[@herth2023] That is a modest change; the ADA points to quality of life and blood pressure as the clearer gains. Some newer weight- and glucose-lowering medicines have also reduced apnoea severity in trials of adults with obesity,[@ada2026s5] which is a conversation for your care team (Chapter 11).

Shift work and irregular hours

A meta-analysis of 12 studies (226,652 participants, 14,595 with diabetes) found that people who had ever done shift work had 9 percent higher odds of diabetes (odds ratio 1.09). The association was stronger in men (1.37) than in women (1.09), and in rotating shifts.[@gan2015] This is association, and shift work also affects meal times, activity, and stress. The ADA notes that disrupted circadian rhythms in people with diabetes are associated with higher HbA1c, and suggests that clinicians ask about shift schedules and days off.[@ada2026s5]

If you work shifts, you are not failing at sleep; you are working against your body clock. Protect one main sleep block in a dark, quiet room, keep it as regular as your roster allows, and tell your care team your shift pattern, because the timing of meals, activity, and any glucose-lowering medicine may need planning around it.

Sleep habits: what the guidelines say, and what trials found

The ADA recommends counselling people with diabetes to practise sleep-promoting routines (its strongest evidence grade, A): a regular bedtime and rising time; a dark, quiet, cool bedroom; a wind-down routine; phones and tablets on silent or off (diabetes devices excepted); daytime exercise; no daytime naps; less caffeine and nicotine in the evening; no spicy food at night; and no alcohol before bed. It advises referral to a sleep specialist for significant difficulties, and notes that cognitive behavioural therapy for insomnia (CBT-I) has improved sleep and glucose measures in studies.[@ada2026s5]

Keep expectations realistic. A systematic review of 22 studies found that CBT-I and sleep education improved sleep quality, but the reduction in HbA1c (0.35 percentage points) was not statistically significant. Extending sleep in short sleepers raised sleep time and, in five of six studies, was linked with measures of insulin resistance, but the authors called the effects on glucose "inconclusive".[@kothari2021] Good sleep is worth having for its own sake and supports the other habits in this book; it is not a substitute for them.

The night shift of your hormones: night-time lows and the early-morning rise

Two things can push a morning reading up, and telling them apart matters. The dawn phenomenon is a rise in glucose in the early hours, when growth hormone bursts during the night and cortisol climbs before waking. In 12 people with type 1 diabetes whose insulin was infused to match their needs, insulin sensitivity was lower at dawn than in the early night: liver glucose output was about 30 percent higher and glucose use by tissues about 25 percent lower. In seven of them, blocking the night-time growth hormone bursts removed the increase in liver output.[@perriello1990] In 248 people with type 2 diabetes not on insulin who wore sensors, the median rise from the overnight low to the pre-breakfast value was 16 mg/dL (interquartile range 0 to 31.5), and the dawn rise raised HbA1c by about 0.4 percentage points on average.[@monnier2013]

{{fig:F10-A}}

The second possibility is a rebound: a night-time low followed by a morning high (the Somogyi effect). Studies disagree on how often it happens. In 4,705 nights of sensor data from 2,600 people with type 2 diabetes on stable insulin, morning glucose was lower, not higher, after nights with a low, and the authors found no support for the effect.[@huang2022] In 755 sensor users with type 1 diabetes over 14 days, about one in three had at least one night-time low followed by a high before 6 a.m.[@gonzalezvidal2025] Sensors can also reveal lows that cause no symptoms.[@ada2026s6]

Do not change your evening medicine or food because of one high morning reading. Chapter 12 explains how sensors show the overnight pattern, and Chapter 13 covers treating lows. If you use insulin or a sulfonylurea and see repeated morning highs or suspect night lows, ask your care team to review the pattern before any change.


10.2 Stress Hormones and Blood Glucose

Two speeds of stress

When you face a threat, real or imagined, the body prepares to act. Two hormones respond at different speeds, and a third joins them.

Adrenaline (epinephrine) is the fast one, released within seconds. It increases glucose production by the liver and reduces insulin release and insulin-driven glucose uptake by tissues.[@nonogaki2000] Cortisol is slower, acting over hours. It promotes glucose production in the liver and reduces glucose uptake in muscle and fat, so it opposes insulin; one review concludes that its main effect during stress is to preserve glucose for the brain.[@kuo2015] Glucagon, made by alpha cells in the pancreas, tells the liver to release glucose by breaking down glycogen and making new glucose.[@amin2025]

{{fig:F10-B}}

For a person without diabetes, insulin rises to answer this, and glucose returns to normal. When insulin is deficient or cells are resistant (Chapter 1), the answer is weaker, so glucose rises higher and stays up longer.

Science Corner: A survival response, not a defect Rising glucose during stress is not a malfunction. In acute illness, "stress hyperglycaemia" appears to be an evolutionary survival response that supplies fuel to vital organs, although in hospital it is also a marker of how sick someone is.[@marik2013] In diabetes, this normal response lands on a system that cannot fully compensate, which is why illness, injury, and surgery so often push readings up.

When the stress is illness

An infection, injury, or operation is a physical stress. The ADA warns that such events raise the risk of both high and low glucose in people with diabetes, and that in severe cases they can lead to hyperglycaemic crises that are life-threatening and need immediate care.[@ada2026s6] Ask your care team, before you are ill, for written "sick-day" advice covering how often to check glucose, what to drink and eat, and when to call. For people at risk of diabetic ketoacidosis, the ADA advises measuring ketones during illness, especially if glucose exceeds 200 mg/dL (11.1 mmol/L). If you use insulin, do not stop your long-acting (background) insulin just because you are not eating; call your team for instructions.[@ada2026s6] Seek urgent medical attention for persistent vomiting, marked thirst and passing large amounts of urine, signs of dehydration, or drowsiness or confusion (Chapter 13).

When the stress is chronic

Constant pressure from caring duties, insecure work, conflict, debt, or displacement is different from a single alarm. A review describes psychological stress as a predictor of new type 2 diabetes and as a prognostic factor in people who already have it.[@hackett2017] One cohort followed 7,251 Swedish men for 35 years: the 15.5 percent who reported permanent stress at home or work had a higher risk of diabetes (hazard ratio 1.45 after adjustment for age, social class, physical inactivity, BMI, and blood pressure).[@novak2013] That study was of men only and measured stress with a single self-reported question, so it shows an association, not a proven cause. Stress can also change sleep, eating, and activity, which is worth noticing in yourself.

None of this means you are at fault for being stressed. Many stressors are real and cannot be removed; the useful question is which small parts you can influence.

Do stress-reduction methods lower HbA1c?

Several methods have been tested in people with diabetes. Participants cannot be kept unaware of what they are doing, which makes bias hard to exclude.

Approach Evidence reviewed Change in HbA1c vs comparison group Comment
Mindfulness-based programmes[@ni2021] 8 randomised trials, 841 people with type 1 or type 2 diabetes 0.25 percentage points lower (95% CI 0.07 to 0.43) Also lowered stress, depression, and diabetes distress
Yoga[@li2026yoga] 28 randomised trials, 2,241 adults with type 2 diabetes 0.64 percentage points lower (95% credible interval 0.41 to 0.87) Certainty rated low; evidence outside India limited
Psychological therapies, mainly CBT and counselling[@winkley2020] 96 studies overall; 49 in 12,009 adults with type 2 diabetes About 0.33 percentage points lower in type 2 diabetes Authors: "minimal clinical benefit"; no benefit for type 1 diabetes

Each row is a separate meta-analysis of different trials, so the figures are not directly comparable. Reductions are shown as positive numbers of percentage points as reported in the abstracts; 0.33 percentage points is about 3.5 mmol/mol.

Three lessons follow. First, HbA1c benefits are modest, and the certainty of the evidence is limited. Second, mindfulness also improved stress, depression, and diabetes-related distress,[@ni2021] worthwhile outcomes in themselves. Third, yoga involves movement, so its effect may partly reflect physical activity (Chapter 9). For breathing exercises and relaxation practised alone there is no diabetes-specific meta-analysis to quote for HbA1c; the ADA lists mindful breathing among brief coping strategies for care teams to teach.[@ada2026s5] Treat these methods as add-ons for well-being, sleep, and coping, chosen by what you will enjoy and keep doing, not as replacements for medicine, food, or activity.

Social support and when to seek more help

The ADA notes that social support promotes better health, and that lack of it is associated with poorer outcomes in people with diabetes.[@ada2026s1] Practical forms include a family member who joins your walk, a peer group in person or online, or a diabetes educator. If you feel overwhelmed by the daily demands of diabetes, that has a name, diabetes distress, and the ADA recommends asking about it at least once a year and referring people to a behavioural health professional when it is not eased by routine care.[@ada2026s5] It is different from burnout and depression, both covered in Chapter 16. If you feel persistently low or hopeless, or have thoughts of harming yourself, tell your doctor promptly or seek emergency help; you do not have to manage that alone.


10.3 Other Hormonal Influences

Insulin is only one of many hormones. A few conditions and medicines change the balance and make glucose harder to manage, and their treatment sits with a clinician, not with diet or willpower. If any of the following sounds like you, tell your care team, and do not stop or change any medicine on your own.

Thyroid conditions

Thyroid disorders and diabetes are closely linked: a review found each is more common in people with the other, and that untreated thyroid dysfunction can impair metabolic control.[@biondi2019] In type 1 diabetes, the ADA recommends screening for autoimmune thyroid disease soon after diagnosis and at intervals if clinically indicated, because an underactive thyroid has non-specific symptoms and a gradual onset.[@ada2026s4] Screening is a blood test. If you feel unusually tired, cold, or overheated, or notice unexplained weight change, ask whether your thyroid has been checked.

Polycystic ovary syndrome (PCOS)

PCOS is a common hormonal condition that affects the ovaries and can cause irregular periods, acne, and excess hair growth. It is a recognised risk factor for type 2 diabetes, and the ADA lists it among the reasons to test for diabetes and prediabetes.[@ada2026] The 2023 international guideline highlights metabolic risk factors, sleep apnoea, and a very high prevalence of psychological features, and keeps healthy lifestyle and emotional wellbeing central to care, while describing its evidence as generally low to moderate in quality.[@teede2023] If you have PCOS, ask how often your glucose should be checked, and mention snoring or daytime tiredness (see also Chapter 2).

Menopause

A meta-analysis of 19 studies found that both early menopause (odds ratio 1.24) and late menopause (1.14) were associated with higher odds of type 2 diabetes, compared with menopause at a typical age; the authors noted that more research is needed to be certain.[@yazdkhasti2024] For women with type 1 diabetes, a 2025 narrative review reports that glycaemic variability and changing insulin needs are frequently described during the menopause transition, but that evidence to guide treatment, including hormone therapy, is limited.[@courtney2025] Keep track of your glucose patterns through this stage and discuss symptoms and any hormone therapy with a clinician who knows your diabetes. Night sweats can be mistaken for night-time lows, so if you have a sensor or meter, check rather than guess.

Testosterone in men

Testosterone levels are on average lower in men with diabetes, but obesity is a major confounder. The ADA advises asking men with diabetes or prediabetes about sexual health and, if signs or symptoms of low testosterone are present, testing morning total testosterone.[@ada2026s4] Treatment is a prescription decision that follows proper blood tests.

Steroid medicines and other drugs that raise glucose

Glucocorticoids ("steroids", taken as tablets or injections) mimic cortisol and are used for autoimmune conditions such as rheumatoid arthritis, after organ transplants, and alongside some cancer treatments. The ADA reports that up to 18 to 32 percent of people treated with higher-than-physiological doses developed raised glucose, depending on the drug, dose, and duration.[@ada2026] Because steroid-related highs come mainly from insulin resistance, a fasting reading can miss them; the ADA advises checking one to two hours after meals or at random times instead, since relying on fasting glucose can delay diagnosis.[@ada2026] If you are prescribed steroids, ask your prescriber how often to check, and report thirst or frequent urination early. Never stop or reduce a steroid on your own; after longer courses it usually must be tapered under supervision.

The ADA also advises glucose screening for people taking thiazide diuretics, some HIV medicines, and second-generation antipsychotics, which can raise glucose.[@ada2026] Do not stop any of these because of this chapter; your team can check your glucose and adjust the plan.

Rarer conditions, and pregnancy

Excess cortisol (Cushing syndrome) or excess growth hormone (acromegaly) is associated with insulin resistance and can cause diabetes.[@defano2024] These are rare and are diagnosed by specialists. Pregnancy is a different, common hormonal state: placental hormones make cells more resistant to insulin, which is why gestational diabetes can appear (Chapter 3). Anyone with diabetes who is pregnant, or planning pregnancy, should have specialist care.


Key Takeaways

Action Points

  1. Notice your sleep for two weeks. Write down bed and rising times, and whether you snore, wake often, or feel sleepy by day. Bring the notes to your next appointment.
  2. Change one sleep habit. Pick one item from the ADA list, such as a fixed rising time, a dark room, or no phone in bed, and keep it for a month.
  3. Ask about sleep apnoea if you have the signs. Loud snoring, witnessed pauses, or heavy daytime sleepiness are reasons to request screening.
  4. Choose one stress practice you will actually do. A ten-minute walk with a friend, a breathing exercise, a yoga class, or a mindfulness programme all count. Try it for eight weeks and note how you feel and what your readings do, as an add-on to your plan.
  5. Tell your care team about hormones and medicines. Mention thyroid, PCOS, menopause symptoms, and any steroid tablets or injections, and ask for a written sick-day plan before you need it.

This book is intended for education and does not replace personal medical advice. If you have diabetes or take glucose-lowering medication, please consult your healthcare team before changing your diet, exercise, or treatment.

All chapters