Every source this book rests on, the exact claim it carries, and the grade I gave it. Find your paper. If the sentence beside it is a fair reading of what you found, that is all I need to know. If it is not, that is worth more to me than anything else on this site.
You do not need to read the book. One line, two minutes. The question is not whether the evidence is good, which you already know. It is whether I have used your work for something it does not support.
That is the error I cannot catch from the inside: the citation checks out, the number is right, and the conclusion is wrong anyway. It is the one thing only you can settle.
Reply to [email protected] with the line and what is wrong with it.
Where you are right, the page changes and you are named in the edition that publishes, unless you
would rather not be.
The system: how the pillars connect · Sleep · Eat · Exercise & recovery · Data and support · Across genders · Habits & individual variation · The studies behind what I left out · Guidance and formularies
| The source | What the book says it shows, and where |
|---|---|
| 1 · Nutrients → Fitness. Morton et al., BJSM 2018 (49 RCTs, n≈1,900).Strong | Protein augments resistance-training gains. Full entry under Eat.Used on pages 26, 29, 32, 40 |
| 2 · Deep sleep → Recovery. Van Cauter, Leproult & Plat, JAMA 2000.Established | Growth hormone is released in slow-wave sleep. Van Cauter is correlational and in men, but suppressing slow-wave sleep does lower the overnight pulse and enhancing it raises it, so the direction has been tested.Used on pages 20, 21, 43 |
| 3 · Nutrients ↔ Gut. Wastyk et al., Cell 2021.Emerging | Fermented food shifted the microbiome. Primary immune endpoint null; n≈18 per arm.Used on pages 33, 34 |
| 4 · Deep sleep → Nutrients. Tasali, JAMA Intern Med 2022 (RCT, n=80).Established | Extending sleep cut intake by ~270 kcal a day. The behavioral result is solid; the ghrelin and leptin story told alongside it is not, and later pooled trials, not entries here, found no change.Used on pages 20, 21, 29 |
| 5 · The day’s practice → Light·N2. Laventure et al., PLOS Biol 2016.Emerging | N2 spindles consolidate motor-skill learning. A small lab experiment on finger-tapping, not on fitness.Used on page 19 |
| 6 · Deep ↔ Light·N2. Ng et al., eLife 2025.Emerging | Slow-wave and spindle coupling (small effect); the map link is the claim.Used on pages 19, 70 |
| 7 · Fitness → Deep sleep. Kredlow et al., J Behav Med 2015 (meta).Established | Exercise increases slow-wave sleep. Confidence is good; the gain is single-digit minutes, largest in poor sleepers.Used on pages 20, 21, 40 |
| 8 · Nutrients → Deep sleep. St-Onge et al., JCSM 2016.Emerging | Fiber predicts more slow-wave sleep; the map link is the claim.Used on page 26 |
| 9 · Sleep ↔ Gut. Smith, PLOS One 2019.Emerging | Sleep tracks with microbiome diversity (cross-sectional).Used on pages 33, 34 |
| 10 · Fitness ↔ Gut. Allen et al., MSSE 2018.Emerging | Exercise enriches fiber-fed bacteria in 32 humans over six weeks, reversing on detraining; the fatty acids they make rose in lean participants only. A gut-to-strength trial was null; not an entry here.Used on pages 33, 34 |
| 11 · Light·N2 → Nutrients. Vallat et al., Cell Rep Med 2023 (n>600, replicated in >1,900).Emerging | Spindle coupling predicts next-day fasting glucose. Observational, a surrogate outcome, and no human trial has manipulated spindles to test it.Used on page 19 |
| 12 · Deep → REM. Ohayon et al., Sleep 2004 (meta, 65 studies, n=3,577).Established | Normative stage percentages across the lifespan: deep sleep 10 to 20% of the night and falling with age, REM 20 to 25%. That distribution is settled sleep-lab measurement. The circuit-level account of why rests partly on rodent work, graded Emerging under Park & Weber below.Used on page 15 |
| The source | What the book says it shows, and where |
|---|---|
| Scheer et al., PNAS 2009 (n=10).Emerging | With the body clock deliberately pushed out of step with the day, glucose, appetite hormones, and blood pressure all worsened. Tiny, and the reason misalignment is treated separately from short sleep.Used on page 64 |
| Leary et al., JAMA Neurol 2020 (MrOS n=2,675; WSC n=1,386).Moderate | Every 5% less REM tracked with ~13% higher all-cause mortality, replicated. Observational, one night of recording, and low REM is also caused by illness, antidepressants and apnea.Used on pages 22, 23 |
| van der Helm et al., Curr Biol 2011 (n=18 with sleep recording).Emerging | Low prefrontal gamma during REM, not REM duration, tracked with an overnight drop in amygdala response. Small, single lab, and descriptive only.Used on pages 22, 23 |
| Park & Weber, Front Psychol 2020 (review). Also cited under link 12.Emerging | Deep sleep and REM alternate through the night. Descriptive, and mostly rodent-derived.Used on pages 22, 56 |
| Buxton et al., Diabetes 2010 (single-arm, n=20 men).Emerging | A week of five-hour nights cut insulin sensitivity ~11%, measured directly. No control arm, no recovery arm, men only.Used on pages 17, 46, 47, 64 |
| Beals et al., Diabetes Care 2026 (RCT, n=29).Emerging | Six weeks of sleep extension did not improve insulin sensitivity. A small null: n=29, a surrogate endpoint.Used on pages 17, 46 |
| Benjafield et al., Lancet Respir Med 2019.Moderate | Modeled global sleep apnea prevalence, ~936m aged 30 to 69. A modeled estimate, not a count, and funded by a maker of sleep apnea devices.Used on page 17 |
| Angarita et al., Addict Sci Clin Pract 2016.Moderate | Cannabis lowers REM; quitting brings a vivid-dream rebound.Used on page 23 |
| Abernethy & Todd, Eur J Clin Pharmacol 1985.Emerging | Low-dose combined pills lengthen caffeine half-life from about 5.4 to 7.9 hours by blocking the enzyme that clears it. A small pharmacokinetic study, graded accordingly.Used on page 23 |
| Hilli et al., J Clin Pharmacol 2008 (n=29).Emerging | The combined pill raises melatonin exposure four- to five-fold by the same route.Used on page 23 |
| Jaehne et al., Sleep Med Rev 2009 (review, 52 studies).Moderate | Nicotine use: longer sleep onset, more fragmentation, less slow-wave sleep, and REM suppression in most studies. A narrative review of 52 studies with partly inconsistent methods and results, which is why it sits here rather than higher. Read the direction as the reliable part.Used on page 23 |
| Stutz, Eiholzer & Spengler, Sports Med 2019 (meta, 23 studies).Established | Evening exercise raised slow-wave sleep by about 1.3 percentage points. Vigorous work close to bed may delay sleep onset instead.Used on page 21 |
| Härtter et al., Clin Pharmacol Ther 2000 (n=5 healthy men).Emerging | Fluvoxamine raised melatonin exposure roughly seventeen-fold. Tiny sample, men only, and the basis for the avoid advice.Used on page 23 |
| Erland & Saxena, J Clin Sleep Med 2017 (31 products).Emerging | Melatonin content ran from far below the label dose to several times above it, and some products contained serotonin. A laboratory assay of what is on the shelf, not a trial.Used on page 15 |
| Haghayegh et al., Sleep Med Rev 2019 (17 studies).Established | A bath at 40 to 42.5°C for at least ten minutes, one to two hours before bed, shortens sleep onset. Pooled from small studies, so the direction is firmer than the size.Used on page 21 |
| Downie et al., Cochrane 2023 (blue-light filtering lenses).Moderate | No benefit for eye strain or visual acuity. On sleep, which is what I cite it for, the review is indeterminate rather than negative, so the direction is not settled and I grade it there.Used on page 17 |
| Pasquier et al., J Therm Biol 2025 (9 RCTs); Riedy, Sleep Med Rev 2021 (38 articles).Moderate | Cooling bedding moved no sleep measure. White-noise evidence graded very low, which the reviewers note contradicts its use. Graded on the weaker of the two.Used on page 17 |
| Wang et al., Sleep Med 2025.Moderate | Exercise and slow-wave sleep, in people who already have a sleep disorder.Used on pages 20, 21, 48, 49, 71 |
| Gardiner et al., Sleep Med Rev 2024 (meta, 27 studies).Established | REM delayed and shortened dose-dependently. Total sleep time could not be pooled, so fragmentation rests on Ebrahim.Used on pages 15, 23, 57, 60 |
| Gardiner et al., Sleep Med Rev 2023 (meta).Strong | Caffeine harms sleep continuity and duration; REM share largely unaffected.Used on pages 15, 23, 57, 60 |
| Carskadon & Dement, Principles and Practice of Sleep Medicine, 6th ed. 2017.Background | Sleep architecture: deep front-loaded, REM builds toward morning. A textbook chapter.Used on pages 15, 19, 21, 22, 23 |
| Windred et al., Sleep 2024 (UK Biobank, n=60,977).Moderate | Regularity out-predicted duration; top four quintiles 20 to 48% lower. Observational.Used on pages 15, 16, 18, 56, 60, 65 |
| Cappuccio, Sleep 2010; Yin, JAHA 2017 (meta of cohorts).Moderate | Duration U-curve, lowest risk near seven hours. Self-reported duration, and the long-sleep arm is partly reverse causation.Used on pages 15, 16 |
| Khalsa et al., J Physiol 2003 (phase-response curve, n=21).Emerging | Light in the late night and early morning advances the clock; light in the evening delays it. This is the phase-response curve the morning-daylight instruction rests on. It measures clock timing under bright laboratory light, not ten minutes of daylight outside a front door. Graded on this one study, not the wider literature. The ten minutes is my dose, not theirs.Used on page 59 |
| Chang et al., PNAS 2015 (crossover, n=12).Emerging | Evening screen light delayed melatonin and the clock. Twelve people, at a deliberately strong light dose.Used on pages 15, 16, 34, 48, 53 |
| Ebrahim et al., Alcohol Clin Exp Res 2013 (review).Moderate | Alcohol suppresses REM and fragments the second half of the night.Used on pages 21, 22, 23 |
| Qaseem et al. (ACP), Ann Intern Med 2016.Guideline | CBT-I is first-line for chronic insomnia, ahead of medication.Used on pages 16, 18, 55 |
| Circadin SmPC (electronic Medicines Compendium).Guideline | Estrogen, fluvoxamine and quinolones raise melatonin levels. Not for pregnancy or breastfeeding.Used on page 59 |
| Herxheimer & Petrie, Cochrane 2002 (10 RCTs).Strong | 0.5 to 5 mg at the destination bedtime reduces jet lag, most clearly flying east across five or more zones. Above 5 mg is no better.Used on page 59 |
| DVLA, Assessing Fitness to Drive (UK, current edition).Guideline | Do not drive while sleepiness is uncontrolled. The notification rules themselves differ by country, and in the US by state, so the book sends you to ask rather than printing one.Used on page 69 |
| WHO Housing and Health Guidelines, 2018.Guideline | 18°C as a minimum indoor temperature. Not a sleep target.Used on pages 27, 31, 37, 66, 71 |
| The source | What the book says it shows, and where |
|---|---|
| Te Morenga et al., BMJ 2012 (30 RCTs, 38 cohorts).Strong | Cutting sugar ad libitum, −0.80 kg. Swapping it for the same calories of other carbohydrate, +0.04 kg: no effect.Used on page 32 |
| SACN 2015; NDNS 2019 to 2023 (UK); NHANES (US).Guideline | The 30 g fiber recommendation, and the surveys behind the shortfall: mean intake about 16 g in UK surveys and 17 g in US NHANES, with roughly one adult in twenty reaching the target on either side.Used on pages 26, 31, 32 |
| Hall et al., Cell Metab 2019 (inpatient RCT, n=20).Emerging | On matched menus, people ate about 500 kcal a day more on the ultra-processed arm. Small, tightly controlled, and the strongest causal evidence in this area.Used on page 27 |
| Sacks et al., Circulation 2017 (AHA advisory).Guideline | The basis for the line that with fat, the type matters more than the amount.In the list; the map link is the claim |
| McDonald et al., mSystems 2018 (American Gut, cross-sectional).Emerging | The thirty-plants-a-week figure. Citizen science, a small contrast within it, and no trial.Used on page 34 |
| Liu et al., NEJM 2022 (RCT, n=139).Established | Time-restricted eating added nothing to the same calorie cut over 12 months.Used on pages 26, 27 |
| Dicken et al., Nat Med 2025 (crossover, n=55).Moderate | Minimally processed beat ultra-processed by ~1 percentage point of body weight over eight weeks. All had overweight or obesity and already ate mostly ultra-processed food. No penalty on lipids or glucose.Used on page 27 |
| Martinez Steele et al., BMJ Open 2016 (NHANES).Emerging | Ultra-processed food supplied ~57.9% of US dietary energy. Cross-sectional survey data.Used on page 28 |
| Zmora, Suez et al., Cell 2018.Moderate | Probiotic colonization of the gut mucosa is person-specific and often transient. Stool testing does not detect it.Used on page 34 |
| Morton et al., BJSM 2018 (49 RCTs, n≈1,900).Strong | Protein plus training; benefit flattens above ~1.6 g/kg.Used on pages 26, 29, 32, 40 |
| Reynolds et al., Lancet 2019 (185 cohorts, ~135m person-years; 58 trials).Moderate | 15 to 30% lower mortality and less major disease. The authors’ own GRADE certainty is moderate, in people without chronic disease.Used on pages 26, 27, 32 |
| Lane et al., BMJ 2024 (umbrella review, 45 pooled analyses, n≈9.9m).Moderate | Worse outcomes across 32 of 45 endpoints, but GRADE quality low or very low for 41. Observational, very broad exposure.Used on pages 26, 27 |
| Mamerow et al., J Nutr 2014 (crossover, n=8).Emerging | Even distribution raised 24-hour muscle protein synthesis 25%. Eight people, a surrogate endpoint, and later trials found no difference.Used on pages 28, 32 |
| Zhao et al., JAMA Netw Open 2023.Moderate | The moderate-drinking ‘benefit’ is mostly a study artifact. 107 observational cohorts, whose point is that the corpus is confounded.Used on pages 7, 26, 28 |
| The source | What the book says it shows, and where |
|---|---|
| Thorogood et al., Am J Med 2011 (14 RCTs, n=1,847).Strong | Aerobic exercise alone, −1.6 kg at six months. The authors conclude it is not an effective weight-loss therapy.Used on page 39 |
| Lee et al., Circulation 2011 (n=14,345 men).Moderate | Each 1-MET gain in fitness, 15% lower all-cause mortality. Change in weight added nothing once fitness change was in the model. Prospective cohort.Used on page 39 |
| Xiao & Ren, Life 2025 (78 studies); Bonafiglia, Front Physiol 2021 (149 studies).Strong | Once measurement error is modeled, no consistent evidence of true differences in trainability. The non-responder is mostly noise.Used on page 40 |
| Leong et al., Lancet 2015 (PURE, n=139,691).Moderate | Grip strength predicted mortality more strongly than systolic blood pressure. Prospective cohort.Used on page 39 |
| Shailendra et al., Am J Prev Med 2022 (meta of cohorts).Moderate | Resistance plus aerobic training tracked with lower mortality than either alone. Observational.Used on page 38 |
| Han et al., BJSM 2022.Moderate | VO₂max and mortality, supporting the Mandsager finding. Observational.Used on page 41 |
| Balachandran et al., JAMA Netw Open 2022 (20 RCTs, n=566, mean age 70).Moderate | Power training gave a modest edge over strength training for physical function. Small trials and the authors rated it low-certainty.Used on page 66 |
| Roberts et al., J Physiol 2015 (RCT).Established | Cold water immersion after lifting blunted long-term strength and hypertrophy gains.Used on pages 43, 49 |
| Currier et al., ACSM Position Stand, Med Sci Sports Exerc 2026 (137 reviews, >30,000 people).Guideline | Load, full range, 2 to 3 sets and twice a week improve strength. Training to failure, time under tension, equipment and periodization did not consistently matter.Used on page 41 |
| Mandsager et al., JAMA Netw Open 2018 (retrospective, n=122,007).Moderate | Higher VO₂max, lower mortality, no ceiling: ~40% between below- and above-average fitness, roughly fivefold from least fit to the top 2%. Patients referred for testing, not a random sample, so confounding by indication is the caveat.Used on pages 37, 38, 40, 41 |
| Ding et al., Lancet Public Health 2025; Paluch, Lancet Public Health 2022.Moderate | ~47% lower all-cause mortality from 2,000 to 7,000 steps, then a plateau, age-dependent. Authors graded certainty moderate for most outcomes, low for several.Used on pages 37, 38, 42 |
| Sherrington et al., Cochrane 2019 (108 RCTs, n=23,407).Strong | Balance and functional exercise cut falls 24% (39 trials, n=7,920, high certainty); adding resistance, 34% (11 trials, n=1,374, moderate certainty).Used on pages 39, 42, 66 |
| NHIS 2020 (US National Health Interview Survey).Moderate | About 32% of adults did muscle-strengthening activity on two or more days a week. A national participation rate, not a trial.Used on page 40 |
| Momma et al., BJSM 2022 (meta of cohorts).Moderate | Muscle-strengthening activity and lower mortality, strongest at modest volumes.Used on pages 37, 38, 42 |
| Bull et al., BJSM 2020 (WHO guidelines).Guideline | 150 to 300 minutes of moderate activity a week, plus strength on two days. Over 65, three or more days combining strength, balance and moving around.Used on page 37 |
| Buffey et al., Sports Med 2022 (meta of trials).Moderate | Light walks of two to five minutes break up sitting and blunt the glucose rise against staying seated. Acute crossovers, mostly in adults with overweight.Used on pages 42, 57 |
| Saner et al., J Physiol 2020.Emerging | Five nights at four hours cut the muscle-building response ~18%. Three interval sessions across the same nights held it at control levels. Eight men per group.Used on pages 20, 21, 40, 43, 44 |
| Stamatakis et al., Nat Med 2022 (n=25,241).Moderate | Brief vigorous daily bursts track with lower mortality, near-linear dose-response. Prospective cohort.Used on pages 37, 38 |
| The source | What the book says it shows, and where |
|---|---|
| Abdelhamid et al., Cochrane 2020 (45 RCTs, n=143,693).Strong | Omega-3 supplements: all-cause mortality RR 0.97 and cardiovascular events RR 0.96, both high certainty. A small reduction in coronary deaths is possible, low certainty.Used on page 49 |
| Kazeminasab et al., Nutrients 2025 (69 RCTs, n=1,937).Strong | Squat +5.64 kg; leg press, women and over-50s all null. Larger and newer than Wang, and the lower estimate. The disagreement with the postmenopausal review is confined to that subgroup, not the main direction.Used on pages 48, 49, 71 |
| Naddafha et al., J Int Soc Sports Nutr 2026 (7 RCTs, n=608, postmenopausal).Emerging | Leg-press 1RM +7.5 kg (CI +2.2 to +12.8), lean mass +0.37 kg, bone unchanged. Only at 5 g a day with lifting. Seven trials, short and on stand-in outcomes, which is where the rubric would put it. It also disagrees with the larger review above, so it takes the one-step downgrade the smaller of two disagreeing reviews gets here. The participant count could not be re-verified at source, so check it before relying on it.Used on page 71 |
| NIH Office of Dietary Supplements fact sheets.Guideline | Iron and micronutrient limits.Used on pages 31, 32, 72 |
| Liao et al., Eur J Med Res 2026 (systematic review).Moderate | Continuous glucose monitoring showed no glycemic benefit in normoglycemic people; the benefit was confined to prediabetes.Used on pages 47, 63 |
| Plews et al., Sports Med 2013 (narrative review).Emerging | HRV read against your own baseline rather than a norm. Derived from elite endurance athletes managing training load, not from health outcomes.Used on page 47 |
| Welch, Schwartz & Woloshin, Overdiagnosed, 2011.Background | Overtesting harms. Background reading rather than a study.Used on pages 47, 63 |
| AHA/ATVB scientific statement 2022.Guideline | Lp(a) is 70 to 90% inherited and high in roughly one in five people, more often in African ancestry and less often in East Asian. Measure at least once; repeat if borderline or after menopause.Used on pages 32, 47, 63 |
| Lopez-Jaramillo et al., Lancet Healthy Longev 2023 (PURE, n=141,243).Moderate | TyG flags insulin resistance from a standard lipid panel.Used on page 47 |
| Wang et al., Nutrients 2024 (23 RCTs, adults under 50).Established | Creatine plus training adds strength: lower body +11.35 kg, upper body +4.43 kg. Almost all male, null in the female participants. Graded down one step as the higher of two disagreeing estimates; I lean on the larger review.Used on pages 20, 21, 48, 49, 71 |
| Laukkanen et al., JAMA Intern Med 2015 (cohort).Moderate | Sauna use and lower mortality. Finnish men, and healthy-user bias is not excluded.Used on page 49 |
| Sesso et al., Am J Clin Nutr 2022 (COSMOS, n=21,442).Established | No benefit on cancer, cardiovascular events or mortality in the already-fed. A small cognitive benefit, with no reduction in dementia.Used on page 49 |
| The source | What the book says it shows, and where |
|---|---|
| Kravitz et al., Sleep 2008 (SWAN, n=3,045).Moderate | Sleep problems rose through the transition; vasomotor symptoms tracked the worse nights.Used on pages 21, 71, 73 |
| Greendale et al., J Bone Miner Res 2012 (SWAN cohort, DXA).Moderate | Loss accelerated ~1 year before the final period: ~2.5% a year at the spine, ~1.8% at the neck.Used on pages 41, 71 |
| El Khoudary et al., Circulation 2020 (AHA scientific statement).Guideline | Lipids, blood pressure and visceral fat rise across the transition. A clinical body reading observational data, so I treat it as I treat the other scientific statements here.Used on page 71 |
| Black & Matkin-Hussey, Physiologia 2024 (narrative review).Emerging | Concludes at least the RDA, 0.8 g/kg a day. The women-specific evidence is sparse and low quality.Used on pages 29, 72 |
| Bauer et al., JAMDA 2013 (PROT-AGE).Guideline | The older-adult protein consensus behind the higher target.Used on pages 29, 72 |
| Patel et al., Sports Med Open 2021 (scoping review, 250 studies).Moderate | Women were ~18% of participants; over half the studies excluded them.Used on pages 39, 72 |
| Leproult & Van Cauter, JAMA 2011 (single-arm, n=10 men).Emerging | A week at five hours cut daytime testosterone 10 to 15%. One lab, unreplicated.Used on pages 43, 44, 70 |
| The source | What the book says it shows, and where |
|---|---|
| Wood & Rünger, Annu Rev Psychol 2016.Moderate | Habits are cue-driven and context-dependent.Used on pages 52, 53 |
| Keller et al., Br J Health Psychol 2021 (RCT, n=192).Moderate | Habit stacking was no better than a plain time-of-day cue. A null result, not proof of equivalence.Used on pages 52, 53, 58 |
| Dai, Milkman & Riis, Manage Sci 2014; Milkman, Minson & Volpp, Manage Sci 2014.Moderate | Fresh-start dates lift starts after a landmark; temptation bundling raised gym attendance in a controlled study. Both real, both modest, and short-lived.Used on pages 11, 30, 53 |
| Harkin et al., Psychol Bull 2016 (138 studies).Established | Progress monitoring improved goal attainment; writing it down worked best. The behavior-change literature carries publication-bias risk, so read d=0.40 as the top of a range. Capped at Established for that reason.Used on pages 52, 53 |
| Kravitz & Duan (AHRQ) 2014.Background | N-of-1 trial methods guide. A methods authority rather than a study.Used on pages 21, 71, 73 |
| Gollwitzer & Sheeran, 2006 (meta, 94 tests).Moderate | If-then plans close the intention-action gap, pooled d≈0.65. Largest for discrete one-off actions, smaller for sustained change, and not corrected for publication bias. Capped at Established for that, then down one further step because the larger 2024 re-pooling below corrects it downward and finds the bias substantial.Used on pages 52, 53, 58 |
| Sheeran, Listrom & Gollwitzer, Eur Rev Soc Psychol 2024 (meta, 642 tests).Established | The 2006 finding re-pooled. Publication bias was substantial, and the two standard corrections put the effect at d≈0.35 and d≈0.15. Real, and smaller than the headline.Used on pages 52, 53, 58 |
| Lally et al., Eur J Soc Psychol 2010.Moderate | Habit formation ran a median ~66 days in 39 usable curves from 96 volunteers, over 12 weeks.Used on pages 51, 58 |
| Hagger et al., 2016 (Registered Replication).Strong | The ‘willpower muscle’ failed to replicate across 23 labs.Used on pages 51, 58 |
| Zeevi et al., Cell 2015 (n=800).Moderate | Differing glucose responses to identical meals. A short-term marker.Used on page 73 |
| Bouchard et al., J Appl Physiol 1999 (HERITAGE, n=481).Moderate | Trainability varies widely and runs in families. Measures familial clustering of observed change, which is a different quantity from true trainability; see Xiao and Bonafiglia under Exercise. A supervised family study, not randomized.Used on page 73 |
| Koemel et al., eClinicalMedicine 2026 (UK Biobank, n=59,078).Moderate | Modest concurrent gains in sleep, activity and diet tracked with longer lifespan and healthspan.Used on page 7 |
| Li et al., Circulation 2018 (n≈123,000).Moderate | Five habits tracked with ~12 to 14 more years at 50. Observational; sleep is not one of them.Used on page 7 |
| The source | What the book says it shows, and where |
|---|---|
| Lincoff et al. (SELECT), NEJM 2023 (RCT, n=17,604).Established | Semaglutide cut cardiovascular events in people with obesity and established heart disease. Named on the what-I-left-out page; outside this book’s scope, not below its evidence bar.Used on page 74 |
| SURMOUNT-OSA, NEJM 2024 (RCTs).Established | Tirzepatide reduced obstructive sleep apnea severity, and is licensed on that basis. Named for the same reason.Used on page 74 |
| Holt-Lunstad et al., PLoS Med 2010 (meta).Moderate | Strong social ties predict ~50% higher survival odds; associational. Named on the what-I-left-out page.Used on page 74 |
| The source | What the book says it shows, and where |
|---|---|
| AASM restless legs guideline, J Clin Sleep Med 2025 (Winkelman).Guideline | Check ferritin in everyone with clinically significant restless legs; treat iron at a much higher threshold than for anemia. A consensus good-practice statement, not a trial-derived cutoff.Used on page 69 |
| WHO sodium guideline 2012; SACN 2003 (UK); Dietary Guidelines for Americans.Guideline | Under 5 g of salt a day (WHO), 6 g (UK), or under 2,300 mg of sodium (US). The three differ, so I print the range.Used on pages 27, 31, 37, 66, 71 |
| SACN, Vitamin D and Health, 2016; NHS.Guideline | 10 µg daily in autumn and winter, year-round with little sun. Upper limit 100 µg (4,000 IU).Used on pages 26, 31, 32 |
| NHS, FSA and NICE NG201 (antenatal care).Guideline | Caffeine under 200 mg a day, cleared far more slowly by the third trimester; avoid unpasteurized and mold-ripened cheese. 400 µg of folic acid daily from before conception to week 12; 5 mg where risk is higher, including diabetes and epilepsy medicine, on clinical consensus rather than trials.Used on page 72 |
| EFSA, caffeine scientific opinion, 2015.Guideline | Up to 400 mg a day for adults, and 200 mg in pregnancy.Used on page 11 |
| ESC/EAS dyslipidaemia guideline 2019.Guideline | ApoB goals are set by overall risk. The 1.0 g/L (100 mg/dL) figure on the warning-signs page is the moderate-risk goal, and the usual upper limit of normal in low-risk adults; a higher-risk person is given a lower one.Used on page 69 |
| NICE NG133 (hypertension in pregnancy).Guideline | New hypertension at or above 140/90 after 20 weeks needs secondary-care assessment. The threshold on the warning-signs page.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| NICE NG136 (hypertension in adults).Guideline | 180/120 with symptoms is a same-day referral; without symptoms, organ-damage checks and review within a week. A home average at or above 135/85 is stage 1 hypertension.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| NICE CG100 and CG115 (alcohol-use disorders).Guideline | Withdrawal in a dependent drinker can cause seizures. The caution on stopping abruptly comes from here.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| NICE NG238 (lipid modification).Guideline | Muscle symptoms on a statin: measure creatine kinase rather than stopping on your own.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| NICE NG203 (chronic kidney disease).Guideline | The eGFR bands behind the magnesium advice, extrapolated from CKD practice.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| NICE NG246 (overweight and obesity management).Guideline | Waist-to-height under 0.5, in place of a single waist figure.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| British Thyroid Association; RCOG Green-top 76.Guideline | Levothyroxine dose usually rises by 25% to 40% in pregnancy, early.Used on pages 31, 32, 72 |
| British Menopause Society guidance; WHI reanalyses.Guideline | HRT risk-benefit for healthy women under 60 within ten years of menopause.Used on pages 31, 32, 72 |
| NICE MTG70 (2022).Guideline | Digital CBT-I recommended for NHS use in adults with insomnia.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| NICE NG28 (type 2 diabetes); MHRA drug safety updates; ESPEN clinical nutrition guidance.Guideline | The diabetes, metformin and clinical-nutrition points on the adjustment pages.Used on pages 12, 18, 21, 32, 43, 49, 69, 72 |
| BNF and the electronic Medicines Compendium.Guideline | The intervals on the adjustment pages.Used on pages 31, 32, 72 |