24 evidence-first reviews in this topic, built on 338 peer-reviewed citations. Null and negative trials are reported alongside the positive ones — that is the point of the format.
In the 204-person LEAP trial, eight weeks of load-management education plus hip-abductor exercise left 77.3% of people with gluteal tendinopathy rating themselves at least moderately…
Where the evidence lands: Pain over the greater trochanter is, on imaging, usually a problem of the gluteal tendons and iliotibial band rather than bursitis: in 877 ultrasound examinations 20.2% showed bursitis while 49.9% showed gluteal tendinosis and 28.5% a thickened iliotibial band, and in one series of consecutive pelvic MRIs half of the hips without trochanteric pain showed tendinopathy, so imaging cannot make the diagnosis alone. The syndrome is common (15.0% unilateral and 8.5% bilateral in women aged 50 to 79; 6.6% and 1.9% in men), often persistent (36% still in pain at 1 year and 29% at 5 in a Dutch general-practice cohort) and appears to confer disability similar to end-stage hip osteoarthritis. The LEAP randomized trial (n=204) found education plus exercise beat both a corticosteroid injection and wait and see on global improvement at 8 weeks (77.3% vs 58.5% vs 29.4%; NNT 2.0 vs waiting, 5.0 vs injection) and at 52 weeks (78.6% vs 58.3% vs 51.9%), in a trial where participants were not blinded and contact time with practitioners was unequal; at a year the injection no longer beat waiting on global rating (difference 6.4 points, CI -10.7 to 23.6), though it still did on pain (2.3 vs 3.2), and exercise and injection had converged on pain (2.1 vs 2.3). Strength did not differ between the exercise and injection arms at 8 weeks, and an exploratory mediation analysis suggested the benefit may run through function, pain self-efficacy and less constant pain rather than torque. The program as designed: education in avoiding postures and movements that compress the tendons against the trochanter, 14 sessions over 8 weeks, 4 to 6 daily home exercises taking no more than 15 to 20 minutes, hip-abductor strengthening and adduction control, functional work at Borg 13 to 15 and heavy slow strengthening at 14 to 17, customary stretches left out, pain allowed to 5 of 10 during heavy work if it eased afterwards and was not worse that night or next morning. A Glasgow pilot (30 randomized, 23 analysed, no untreated control) found VISA-G rose by about 10 points in both the isometric and isotonic groups, a difference of -0.1 (CI -13.8 to 13.5), too imprecise to show equivalence; GLoBE (n=94) found no added effect of loading over sham when everyone received compression-avoidance education; a 2009 home program of strengthening and stretching had 7% success at 1 month but 80% at 15. Injections give a real short benefit (55% vs 34% recovered at 3 months in a primary-care RCT; 61% vs 60% at 12). PRP beat corticosteroid at 12 weeks in one 80-patient double-blind trial and again at 24 weeks, after open-label crossover had opened at 3 months, which leaves no randomized 2-year comparison; no placebo or exercise arm was included, and two of four PRP trials found no difference from control. Shockwave gives short-term pain relief (SMD -0.431 at 2 to 4 months) on high-bias trials. A 2024 meta-analysis (6 RCTs, 733) recommends exercise as first-line treatment on few trials; a 2025 network meta-analysis (19 RCTs, 1,701) ranks exercise first for pain and VISA-G function while calling comparisons between treatments challenging. Over a third of completers in the Glasgow pilot had not improved on VISA-G by 12 weeks.
The Achilles complaint that ends running seasons turned out not to be an inflammation at all, which may help explain why rest and anti-inflammatories so often disappoint. What has held up…
Where the evidence lands: Chronically painful tendons show no inflammatory cells on histology (the basis for renaming tendinitis as tendinopathy), which helps explain why rest and anti-inflammatories underperform. The loading evidence: Alfredson's prospective 1998 study returned all 15 athletes to full running after 12 weeks of heavy eccentric calf training while all 15 conventionally treated comparisons went to surgery (nonrandomized comparison); Beyer's RCT (58 patients) found eccentric and heavy slow resistance training equally effective at 52 weeks with higher compliance under HSR (92% vs 78%); Kongsgaard's patellar RCT (39 men) found cortisone's 12-week benefit deteriorated by half-year while both loading programs held, with HSR elevating collagen turnover. Silbernagel's RCT (38 patients) demonstrated no negative effects from continued running under a pain-monitoring model versus 6 weeks of rest (57→85 vs 57→91 on the VISA-A-S score, no significant difference), and her 5-year case series found 80% fully recovered on exercise alone, with fear of movement inversely correlated with heel-rise work recovery (−0.590). De Vos's double-blind JAMA trial found PRP no better than saline (+21.7 vs +20.5 points; difference 0.9, CI excluding the predefined 12-point margin) with eccentrics in both arms. Risk-factor evidence is limited and high-bias: nine factors with limited support (including quinolones, alcohol, plantar-flexor weakness); twenty-six with none, including body weight, foot posture and activity level.
Creatine's muscle evidence in adults over 50 comes entirely from trials where everyone was lifting, and dietary nitrate's dose-response study found 4.2 mmol did nothing while 8.4 mmol…
Where the evidence lands: Creatine monohydrate at 3-5 g daily and roughly 400 mg of dietary nitrate are each well evidenced on their own, but creatine's benefit in older adults is inseparable from resistance training and nitrate's benefit is time-dependent, peaking two to three hours after it is swallowed. Buy nitrate by the printed milligram, keep antibacterial mouthwash away from it, and treat the combination as reasonable rather than proven: no randomized trial has tested the two together.
In the SWAN cohort, LDL and apolipoprotein B jump, and arterial stiffening accelerates from 0.9 to 7.5 percent a year, specifically in the year on either side of the final menstrual…
Where the evidence lands: In SWAN's piecewise analyses, only total cholesterol, LDL and apolipoprotein B jumped within the year around the final menstrual period (1,054 women; pattern consistent across ethnic groups), while glucose, insulin, blood pressure, fibrinogen and CRP tracked chronological age — and carotid-femoral pulse-wave velocity accelerated from 0.9 to 7.5 percent per year in that same window before flattening (339 women). Earlier menopause carries a graded, observational risk ladder for a first non-fatal cardiovascular event versus menopause at 50–51: HR 1.55 before age 40, 1.30 at 40–44, 1.12 at 45–49, strongest before age 60 and persisting in never smokers (301,438 pooled women); a separate 310,329-woman meta-analysis puts menopause before 45 at RR 1.50 for coronary disease and 1.12 for all-cause mortality. The hormone trials ended in a precise place: WHI's estrogen+progestin arm was stopped early (absolute harms of roughly 7–8 excess events per 10,000 person-years each for CHD, stroke, pulmonary embolism and breast cancer), the 13-year overview found estrogen-alone more favorable in women who started in their 50s (19 fewer global-index events per 10,000 women annually, vs 51 excess at 70–79) with no effect on all-cause mortality for either regimen, and ELITE showed estradiol slowing carotid intima-media thickening only when started within 6 years of menopause — a surrogate, not events, with cardiac CT (calcium, stenosis, plaque) null in both strata. Trial investigators' own bottom line: not for chronic-disease prevention, appropriate for symptom management in some women — which is a clinician-level decision.
Home monitoring earns its effect only when the readings feed decisions; on its own it moved pressure a statistically meaningless 1 mm Hg. What does move the number, in randomized trials:…
Where the evidence lands: In randomized trials, blood-pressure self-monitoring lowered systolic pressure only when the readings drove treatment decisions (−3.5 to −4.7 mm Hg in TASMINH4; −6.1 with intensive support in pooled data) — monitoring alone moved it a non-significant 1.0 mm Hg. The DASH eating pattern lowered systolic pressure 5.5 mm Hg overall and 11.4 mm Hg in hypertensives with sodium and weight held constant, and adding low sodium widened the gap versus a typical high-sodium American diet to 7.1 mm Hg in those without hypertension and 11.5 in those with it. Increased potassium cut systolic pressure 3.49 mm Hg overall, with the effect seen only in those with hypertension. A 270-trial network meta-analysis ranked isometric exercise (−8.24 mm Hg) above aerobic, resistance and interval training, with the wall squat the top submode. Weight loss returned roughly 1 mm Hg per kilogram. SPRINT showed why the number matters: a sustained ~15-point separation produced 25 percent fewer major cardiovascular events and 27 percent lower all-cause mortality, alongside more hypotension, syncope, electrolyte and kidney adverse events — target-setting belongs with a clinician.
Zone 2 is a naming convention, not a physiological constant, and most of the evidence behind it comes from trained endurance athletes. What the polarized-training trials, the HIIT…
Where the evidence lands: Zone 2 means low-intensity work below the lactate threshold, but the label shifts between zone systems: in the polarized-training literature, zone 2 is the between-thresholds band that model deliberately minimises to 3–5% of training time. Identification is imprecise — six lactate-threshold methods applied to one dataset of elite skiers showed no agreement at a ±0.5 mM criterion. Polarized training beat other distributions for VO₂peak by SMD 0.24 (95% CI 0.01–0.48), only in trials under 12 weeks and only in highly trained athletes, with no advantage for time trial (SMD −0.01), time to exhaustion or threshold speed. In 30 recreational runners, 10 km improvement was 5.0% vs 3.6%, not significant. Head-to-head for VO₂max, interval training leads continuous training by about 1.2 mL·kg⁻¹·min⁻¹ in healthy adults aged 18 to 45 and 1.74 in older adults; in women the difference was null. A PGC-1α meta-analysis found large effects for both interval (g = 1.29) and continuous (g = 1.01) training, with no significant difference between them, in a pooled estimate with substantial heterogeneity. The longevity evidence — HR 0.47 for high vs low fitness, HR 0.53 at 7,000 steps — is entirely observational and assigned no one an intensity.
At least one in three adults over 65 falls each year, and the trials are unusually clear about what helps. Balance-specific exercise cuts the fall rate by about a quarter, tai ji quan beat…
Where the evidence lands: Balance and functional exercise reduces the rate of falls by 24 percent in high-certainty Cochrane evidence (34 percent, at moderate certainty, when combined with resistance training), and a tailored tai ji quan program cut the fall rate 58 percent versus stretching and 31 percent versus multimodal exercise in a 670-person head-to-head trial. Home fall-hazard reduction cuts falls 38 percent in people at elevated risk but showed no benefit in unselected people. Common prescriptions carry some of the largest risk multipliers in observational data (SSRIs OR 2.02, opioids 1.60, polypharmacy 1.75) — one reason geriatric practice emphasizes regular medication review; expedited first-eye cataract surgery reduced the rate of falling 34 percent in a randomized trial; an annual 500,000 IU vitamin D dose increased falls 15 percent and fractures 26 percent. Gait speed plus age and sex predicts survival about as well as multi-factor risk models — and you can measure it yourself.
Aerobic capacity falls faster with every decade, in people who exercise as well as those who do not — yet controlled trials raise the number in older adults by a few mL/kg/min. Five years…
Where the evidence lands: The decline is real and it accelerates: in the Baltimore cohort it ran from 3%-6% per 10 years in the 20s and 30s to more than 20% per 10 years in the 70s and beyond, in every activity quartile. Training has not been shown to flatten that slope; trained men are measured higher at every age, but only controlled trials speak to what training does to an individual. Pooled trials in sedentary older adults show a net VO2max gain of 3.78 mL/kg/min, largest at 66%-73% of heart rate reserve, 40-50 minutes, 3-4 days a week; head-to-head, intervals came out ahead by 1-2 mL/kg/min. Fitness-mortality findings are associations, and non-response is real.
Most night workers’ clocks never fully adapt, and the trials show they do not need to. Timed light and darkness, a fixed sleep anchor, a short nap, early caffeine and daytime meals…
Where the evidence lands: Fewer than 3% of permanent night workers fully adapt their body clock, but partial realignment is enough to bring night-shift alertness and performance close to daytime levels, and it comes from bright light on the shift, dark glasses on the commute, a fixed dark sleep window, and a 30-minute nap. Melatonin adds about 24 minutes of daytime sleep with no dose-response from 1 to 10 mg; caffeine six hours before sleep still shortens it; eating at night impaired glucose tolerance in three laboratory studies. The long-term cardiovascular and diabetes associations are modest and appear after about five years of exposure.
Lean mass can fall during semaglutide or tirzepatide treatment, but a DXA number is not muscle function—and direct preservation trials are just beginning.
Where the evidence lands: GLP-1-based treatment usually removes much more fat than lean tissue, but absolute lean mass can decline. Track strength and physical function—not DXA alone—and treat exercise and adequate nutrition as important care components whose precise protective effects during semaglutide or tirzepatide therapy are still under study.
A viral longevity lesson says to sleep earlier. The stronger—and less glamorous—case is to protect enough sleep on a schedule your body can keep.
Where the evidence lands: Sleep regularity predicts mortality more strongly than duration in large observational cohorts, but it has not been shown to extend life. Aim for enough sleep and a sustainable schedule; an earlier bedtime is not universally better.
The NMN + resveratrol + pterostilbene stack pairs an NAD+ precursor with sirtuin activators to echo caloric-restriction biology — a coherent rationale built mostly from mechanism and mouse…
Where the evidence lands: The design logic is scientifically literate — fuel NAD+, activate SIRT1 — but the combined stack is untested in humans, and resveratrol's caloric-restriction-mimetic potential appears limited even in mice.
Social media has collapsed careful dose reduction, intermittent maintenance, and unapproved compounded ‘microdoses’ into one seductive idea. Human trials support ongoing treatment for…
Where the evidence lands: Evidence supports ongoing GLP-1 treatment for many people with obesity and suggests that supervised dose reduction can sometimes preserve weight loss. It does not show that tiny compounded doses benefit healthy-weight adults, reduce generalized inflammation, or extend human life.
Mortality benefits begin well below 10,000 steps and the curve often flattens with age. More movement helps, but the exact target is not magic.
Where the evidence lands: Health benefits begin below 10,000 steps and often flatten around 6,000-8,000 in older adults. The best target is a safe, sustainable increase from baseline, not a magic threshold.
Coffee drinkers often have lower mortality in cohorts, including decaf drinkers. Mendelian-randomization results are less decisive, keeping causality and optimal dose uncertain.
Where the evidence lands: Coffee has one of nutrition epidemiology's more consistent favorable associations, but observational evidence does not prove that starting coffee extends life. Tolerance, sleep, preparation, and individual risk still matter.
Short-lived cognitive complaints have been reported, but randomized and observational syntheses do not show that statins cause dementia; some cohorts associate treatment with lower risk.
Where the evidence lands: Statins have not been shown to cause dementia. Population evidence is neutral to reassuring, although new individual symptoms should be evaluated rather than dismissed.
Mouse regeneration experiments and small chemotherapy studies became a universal fasting prescription. Human immune rejuvenation has not been demonstrated.
Where the evidence lands: Prolonged fasting affects nutrient-sensing pathways, but no human trial establishes that 72 hours resets immunity or comprehensively clears damaged cells. The claim is an extrapolation from animal and early translational work.
Sleep-duration studies show a U-shaped mortality curve, but long sleep can be a marker of illness and self-reports cannot define one ideal number for everyone.
Where the evidence lands: Sleep and mortality form a U-shaped observational association, not a precise causal prescription. Quality, regularity, symptoms, and individual need matter alongside duration.
Older cohorts suggested protection, but abstainer bias and genetic studies weaken the causal story. Starting to drink is not a longevity strategy.
Where the evidence lands: The classic alcohol J-curve is vulnerable to former-drinker and healthy-user bias. Genetic studies do not support starting alcohol for cardiovascular protection or longevity.
Self-reported success stories and nutrient calculations cannot establish long-term safety, cardiovascular effects, or lifespan benefit from an all-animal-food diet.
Where the evidence lands: No trial shows that a carnivore diet extends life or prevents chronic disease. Short-term symptom reports are hypothesis-generating, while long-term safety and cardiovascular effects remain unanswered.
Three mechanisms drive the afternoon crash, and only one of them answers to more coffee. An evidence-graded day structure for knowledge workers.
Where the evidence lands: Your afternoon crash is at least three separate problems: attention degrading within a long task, sleep pressure building through the day, and any actual sleep deficit you are carrying. Build the day around that, keep caffeine deliberate and well clear of the hours before bed, and treat naps, light and movement as small helps rather than fixes. Nothing in the supplement aisle in this evidence set beats the schedule itself. If two weeks of a fixed structure changes nothing, stop optimizing and get assessed.
You bought the chair and the standing desk and it still aches. Here is what has trial support in office workers, and what is a waste of money.
Where the evidence lands: The link between sitting and back pain is real but modest and almost entirely cross-sectional, which is why buying furniture did not fix your neck. What has trial support in office workers is unglamorous: about ten minutes of high-intensity neck and shoulder resistance work on most days, walking accumulated toward a real weekly total, and enough consistency to hit roughly seven sessions in ten for three months. Skip the early scan unless you have red flags, and treat saddle numbness or any new bladder or bowel change as an emergency department visit today, not a wait-and-see.
What the loading trials actually show about rebuilding a cranky Achilles, patellar tendon, knee or shoulder after 45, and how long it really takes.
Where the evidence lands: Tendon responds to heavy, slow, progressive load, and it still responds after 50, but on a timescale of months rather than weeks. Pick one loading exercise, run it three times a week, judge it by how you feel the next morning, and give it 12 weeks before you judge it at all. Isometrics and blood flow restriction cuffs are useful tools for keeping you loading when pain or joint tolerance gets in the way, and no supplement in this literature outperforms the training itself.
What the trials actually show about heavy training, protein, supplements and broken sleep in midlife, including the ones that found nothing.
Where the evidence lands: Heavy, supervised, technique-first resistance training around three times a week is the intervention with the best evidence for bone density, strength and function in the postmenopausal women these trials studied, and no supplement in this evidence base replaces it, though none of those trials measured fractures, which is what makes a treatment-range scan a clinician conversation rather than a training problem. Protein and creatine are small modifiers around that training, and calcium or vitamin D supplements did not prevent fractures in the largest pooled analysis of adults not being treated for osteoporosis. Sleep improves with a structured cognitive behavioral program rather than a schedule tweak, and feeling too hot at night is worth treating as a primary variable. Anything here about hormone therapy is reporting from the review papers, not a recommendation.
330 distinct peer-reviewed papers, published 1981–2026, across 174 journals. Each links to its PubMed record; where Magellan has published a full study write-up, that is linked too.
Strong evidence
Across meta-analyses and randomized trials, creatine monohydrate combined with resistance training…
Moderate evidence
Randomized trials of methyl-salicylate/menthol patches and menthol-based topical products show…
Moderate evidence
Meta-analyses of randomized trials show that low-load resistance training with blood flow…
Prefer the interactive version? Open the article feed in the Magellan app →