Applies to a specific group
This appraisal is written for the obese older adult, ~70, with sarcopenic obesity (excess visceral fat together with muscle loss), in whom hypertension, dysglycemia and fatty liver (MASLD) are common, bone density is already reduced, and several medications are in play. It is not written for the frail or institutionalised elderly (for whom the exercise levers must be dosed down and some fall evidence weakens), and not for the lean, fit 70-year-old (whose big rocks are already pulled). Which big rocks are still unpulled for this person decides which levers matter — read the order below as the ranking for this stratum, not a universal list.
The two big rocks here are blood pressure and body fat; the binding constraint is muscle. For a person of this description, most of the achievable benefit sits in a short list, and the ordering is not the same as it would be at 40.
- If this person still smokes, that is the whole first paragraph. No lever below competes with stopping, and every estimate here is dwarfed by it.
- Lowering blood pressure is the best-proven way to prevent death and cardiovascular events at this age — the effect is large, certain, and lands on outcomes that matter.
- Drawing down visceral and liver fat attacks the root of the dysglycemia and fatty liver — but the evidence that it prevents heart attacks is weak, so pull it for diabetes remission, glucose and function, not on a promise about the heart.
- Every gram of fat lost must be paired with resistance training and enough protein, or a large share of what comes off is muscle and bone this person cannot spare.
- Balance-focused exercise is the single highest-certainty lever on the page — it keeps people off the floor, and falls are where injury and lost independence actually happen.
- The big rocks also protect the brain. Dementia is a critical outcome at this age, and the levers that move it are mostly the ones already ranked — lowering blood pressure, staying active, drawing down the metabolic load — so the top of this list is also the top of the dementia-prevention list. That is a second reason to pull the same levers, not a new lever to add.
- Several treatments that help younger adults turn neutral or harmful here — testosterone, calcium-plus-vitamin-D, and the most aggressive weight-loss and blood-pressure targets among them.
If this person smokes, quitting outranks everything else on the page
Nothing modifiable at 70 carries an effect this large. Smoking roughly triples all-cause mortality (hazard ratio ~3.0) and costs on the order of a decade of life, and cessation reverses much of the excess in an age-graded way — the earlier the larger, but the benefit is still substantial at this age (Jha et al., 2013). This is stratum-gated: it belongs at the top only if the person smokes, and drops off the list entirely if they do not — which is exactly why a universal ranking is the wrong object -> Layer 1 - Ranking Interventions for a Stratum, Smoking and Mortality.
Getting blood pressure down is the best-warranted lever for preventing death and cardiovascular events here
Among the always-relevant levers, blood-pressure lowering has the strongest combination of effect size and certainty on hard outcomes. Each 5 mmHg reduction in systolic pressure lowers major cardiovascular events by about 10% (HR 0.91, 95% CI 0.89-0.94), and — the load-bearing point for this stratum — the relative benefit is broadly constant across baseline pressure and across primary vs secondary prevention (Blood Pressure Lowering Treatment Trialists Collaboration, 2021). In the high-risk older population of SPRINT, a more intensive target cut all-cause mortality (HR 0.73) with a number-needed-to-treat of about 61 over 3.26 years (SPRINT Research Group, 2015).
The decision rule is decide on absolute risk, not on the blood-pressure number: because the relative effect is constant, the absolute benefit scales with baseline cardiovascular risk, which is high in this stratum -> Blood Pressure Lowering and Cardiovascular Events, Baseline Risk and the Relative-Absolute Split. How aggressive the target is, however, can itself change the lever’s sign — see the sign-flip section below.
Lowering blood pressure also lowers dementia risk, and here the evidence is randomized rather than observational. An individual-patient meta-analysis pooled five double-blind placebo-controlled antihypertensive trials (28,008 people, mean age 69) and found treatment cut incident dementia — «an adjusted odds ratio 0.87 (95% confidence interval: 0.75, 0.99) in favour of antihypertensive treatment reducing risk of incident dementia with a mean BP lowering of 10/ 4 mmHg» (Peters et al., 2022). The absolute gain is small — dementia struck 403 (2.9%) on treatment against 458 (3.3%) on placebo over a median 4.3 years (Peters et al., 2022). That works out to roughly four fewer cases per thousand treated, and it is a conservative floor: the trials stopped early on their cardiovascular endpoints before slow-accruing dementia could fully show.
Randomization also erased the observational U-shape — no excess dementia appeared at low pressure or in the oldest patients. So blood-pressure lowering buys a second patient-important outcome at this age, and it is the one lever here where the drug, not a lifestyle change, carries the randomized dementia evidence -> Dementia Prevention and Modifiable Risk Factors.
Drawing down visceral and liver fat reverses diabetes and fatty liver — but do not expect it to prevent heart attacks
For a sarcopenic-obese person, the fat is the root that feeds the dysglycemia, the hypertension and the MASLD, so weight loss is the highest-leverage structural move — but its evidence is outcome-specific, and that specificity is the whole point.
Where it is strong (intermediate and patient-important-but-not-hard outcomes). Total diet replacement produced type-2-diabetes remission in 46% at one year (odds ratio ~19.7 vs control), with a steep dose-response on weight lost — from ~0% remission at no loss to ~86% at >=15 kg (Lean et al., 2018). Liver fat responds on a graded ladder: roughly >=5% body-weight loss clears steatosis, 7-10% reduces inflammation, and >=10% can regress fibrosis (European Association for the Study of the Liver, 2024). The mechanism underneath both is the same: negative energy balance draws fat back out of liver and pancreas once a person drops below their personal fat threshold — and intra-organ (liver and pancreas) fat drives the dysglycemia, with visceral fat only a marker for it: «Extent of visceral fat accumulation is a surrogate marker for intra-organ fat excess, but is not pathophysiologically related to adverse metabolic consequences» (Taylor & Holman, 2014) — which is why remission works even at lower BMI -> Ectopic Fat and Depot-Specific Risk, Total Diet Replacement and Type 2 Diabetes Remission, Fatty Liver MASLD and Weight Loss.
Where it is weak (hard cardiovascular events). Look AHEAD, the large lifestyle-weight-loss trial in type-2 diabetes, was null for cardiovascular events (HR 0.95) (Look AHEAD Research Group, 2013). A meta-analysis of weight-loss interventions found a modest all-cause mortality reduction (RR 0.82, ~6 fewer deaths per 1000) but a null for cardiovascular events (RR 0.93) (Ma et al., 2017). So the honest framing: weight loss is pulled here for remission, glycemia, liver fat and function, with a possible small mortality benefit — not on a claim that it prevents heart attacks -> Does Weight Loss Reduce Cardiovascular Events.
A metabolically-‘healthy’ obese profile is not a free pass, but it is not an emergency either — the excess risk is real, slow (emerging after ~10 years) and small in absolute terms (RR ~1.24) (Kramer et al., 2013).
The fat-loss lever touches the brain too, but weakly and second-hand. Diabetes and obesity are among the modifiable dementia risk factors, so drawing down the metabolic load plausibly lowers dementia risk — but that benefit is observational and runs mostly through the cardiometabolic channels already counted, not a separate dividend to bank -> Dementia Prevention and Modifiable Risk Factors. One stratum-level fact is worth naming: a large share of this population already takes an antidiabetic drug, and several classes — metformin, GLP-1 receptor agonists — track lower dementia risk in observational umbrellas (GLP-1 RA RR 0.35, but I2 98.5% and low certainty), while sulphonylureas run the other way (RR 1.39, 1.04-1.87) (Kuate Defo et al., 2023). Read that as what an existing prescription may incidentally do, not a reason to select an agent — that choice is a prescriber’s.
Lift weights and eat enough protein so the weight you lose is fat, not muscle and bone
This is the lever that makes the previous one safe. In a sarcopenic-obese 70-year-old, an unmodified weight-loss diet strips lean mass and bone the person is already short of — so the fat-loss and muscle-protection levers are a pair, not a sequence.
Resistance training is the driver; protein is the adjunct. Training adds strength and lean mass (about +2.49 kg one-rep-max, +0.30 kg lean mass in pooled trials), with protein supplementation augmenting the gain up to a break-point in the region of 1.6 g/kg body weight per day — a break-point whose confidence interval is wide and whose knee is not statistically firm (Morton et al., 2017) -> Protein and Resistance Training for Muscle and Strength, and the dose-response of the training itself is graded rather than all-or-nothing (Currier et al., 2023).
Once training, movement velocity beats heavy load for function. For someone already resistance-training, favoring power (explosive effort) over heavy strength buys more on function tests: against strength training, power training gives SMD 0.36 to 0.43 on activity-based function tests — graded high certainty — and SMD 0.99 on muscle power (el Hadouchi et al., 2022). It is a rare protocol choice that moves a patient-relevant outcome and does not contradict Currier’s second-order-differences result above: Currier graded strength and mass, while power’s edge shows on function, what matters at this age. The honest limit: the review measured no daily-life outcome (el Hadouchi et al., 2022) and, by its outcome menu, no falls — read this as a well-supported steer on function-test capacity, not a demonstrated fall or independence benefit -> Power Training and Physical Function in Older Adults.
The protein target is higher at this age, and it is a floor. Ageing blunts the muscle-protein response (anabolic resistance), so older adults need more protein per meal (~25-30 g, ~1.0-1.2 g/kg/day at minimum, and ~1.6 g/kg is a floor not a ceiling under training) (Bauer et al., 2013) -> Anabolic Resistance, Protein and Resistance Training for Muscle and Strength.
Protein specifically protects fat-free mass during a deficit. Higher protein intake during energy restriction preserves fat-free mass that would otherwise be lost with the fat (Refalo et al., 2025), and exercise loses less muscle per unit of energy deficit than diet alone while targeting visceral fat slightly better (a small effect, ES ~-0.18) (Recchia et al., 2023).
How much the protein adds over the training depends on baseline status. In healthy, well-nourished older adults, adding a nutritional supplement to resistance training showed «no significant differences between groups in muscle mass, muscle strength, or physical functional performance» (creatine on lean mass the lone exception) (Choi et al., 2021) — where intake is already adequate, the training does the work. In the diagnosed-sarcopenic older adult — this stratum — a compound protein-plus-vitamin-D supplement may nudge grip strength on top of training (WMD 1.87, 95% CI 0.01-3.74, P=0.049), though mass stays unchanged and the effect is borderline and heterogeneous (Song et al., 2023). So the floor still holds — protect protein while losing weight — but a supplement does not substitute for the training, and it adds little once intake is adequate.
Read the muscle numbers as surrogates with a prognostic shadow, not as targets. Low grip strength predicts all-cause mortality (HR ~1.16-1.20 per 5 kg lower; ~1.48 per SD) (Celis-Morales et al., 2018) and low muscle mass predicts it too (SMD -0.18), most strongly in the overweight/obese older adult — the sarcopenic-obesity stratum (de Santana et al., 2021). But these are predictors; no trial shows that raising mass or grip lowers mortality, so train for strength, function and muscle-preservation, and measure grip to track it -> Grip Strength and Mortality, Low Muscle Mass and Mortality, Sarcopenia Definition and Diagnosis, Surrogate Outcomes. The one caveat: sarcopenic obesity has no agreed operational cut-off, so the stratum this whole page centres on is the least well-defined one (EWGSOP2 sets none) (Cruz-Jentoft et al., 2018).
Balance-focused exercise, not weights, keeps this person off the floor
Falls are where fractures, hospitalisation and lost independence actually occur, and here the evidence is unusually good. Exercise reduces the rate of falls in community-dwelling older adults by about 23% (rate ratio 0.77, 95% CI 0.71-0.83, high-certainty), and the active ingredient is balance and functional training (RaR 0.76, high-certainty) (Sherrington et al., 2019). The crucial specificity: resistance training alone does not reduce falls (RaR 1.14) — it builds the muscle, but balance training prevents the fall, so the two exercise levers are not interchangeable . This is the highest-certainty exercise outcome on the page and, for many at this age, the one weighted above a marginal change in length of life -> Exercise for Preventing Falls in Older Adults.
Concretely, balance training in these trials means transferring bodyweight from one part of the body to another and challenging the balance systems directly, alongside gait work (changing pace, level and direction) and task-specific functional practice, «assessment-based, tailored and progressed» (Sherrington et al., 2019). The active ingredient is balance that is challenged, not exercise volume — the trials credit «exercises that safely challenge balance» (Sherrington et al., 2019) — so the programme has to keep getting harder as it gets easier rather than settle into a fixed routine. What the evidence does not fix is the exact drill progression for a given person: that is a coaching or prescriber call this page does not make.
A network meta-analysis sharpens the dose Sherrington could not give. Pillay’s analysis for the Canadian Task Force (219 RCTs, 167,864 participants) ranked the whole fall-prevention menu — vision treatment, home-hazard checks, multifactorial programmes and the rest — and balance work still lands on top: «Fourteen of the 21 (67%) interventions with some mod- erate certainty evidence for benefit had a focus on exer- cise», the top tier being supervised, long-duration balance/resistance training and group tai chi (Pillay et al., 2024).
The dose is the upgrade: supervised means «> two sessions, not including ini- tial instruction» and long-duration means «> 3 months», so the effective programme for this person is supervised and sustained, not a one-off class (Pillay et al., 2024). Pillay «included 125 of 283 studies included in the previous review» (Pillay et al., 2024), so this sharpens Sherrington (type-F refinement), it is not independent (type-E) corroboration.
Two adjacent cautions. Vitamin D and calcium are not the fall/fracture lever here — see the sign-flip section. And in the already-frail (outside this stratum), the fall-prevention signal weakens and the programme must be dosed down to avoid causing injury.
But frailty is a lever, not a verdict: activity programmes in pre-frail and frail older adults improve mobility (SMD 0.60), activities of daily living (SMD 0.50), cognition (SMD 0.35) and quality of life (SMD 0.60) at moderate certainty, and lower frailty itself (RR 0.58) (Racey et al., 2021). Read that number with a caveat that is definitional rather than motivational: WHO’s evidence profile notes «A lack of consensus regarding the definition of frailty, and an absence of core measures to assess this means any attempt to create an optimal intervention will be impeded» (World Health Organization, 2020) — so train the measurable components and read frailty as the composite that follows -> Frailty.
Keep moving — most of the mortality benefit banks in the first steps off the couch
General physical activity carries a real mortality benefit, but its shape matters for how hard to push. The gradient is steepest at the bottom: moving from the least-active quartile to the next roughly halves mortality (HR ~0.48), with the most-active reaching HR ~0.27, and the curve flattens at a modest daily dose of moderate-to-vigorous activity (Ekelund et al., 2019) -> Physical Activity Dose and Mortality (where the plateau dose is stated). The decision consequence: the first increment off a sedentary baseline is worth far more than the last, so the target is “get off the floor of inactivity,” not “maximise.”
Activity also lowers dementia risk, and the signal survives the check that usually sinks it. A meta-analysis of 58 cohorts found physical activity associated with «a decreased risk of all- cause dementia (pooled relative risk 0.80, 95% CI 0.77 to 0.84, n=257 983)» (Iso-Markku et al., 2022), and the association held in cohorts followed 20 years or more — so it is not merely early, undiagnosed dementia lowering activity — and did not depend on APOE genotype. It stays observational (higher baseline cognitive reserve is the confound it cannot remove), but it adds a second patient-important outcome to a lever no drug replicates -> Physical Activity Dose and Mortality, Dementia Prevention and Modifiable Risk Factors.
Past these, the remaining levers are smaller and less certain — and that ceiling is itself the finding
Below the big rocks, the levers are genuinely present but smaller, lower-certainty, or aimed at quality-of-life rather than length of life. Reporting that they are small is a decision-change in its own right: it licenses this person to stop optimising once the rocks above are handled -> Layer 1 - Ranking Interventions for a Stratum.
Muscle-strengthening activity and glycemia/mortality. Muscle-strengthening activity is associated with ~15% lower all-cause mortality (RR 0.85) and similar reductions in cardiovascular disease and diabetes — but the certainty is very low (observational), and the upper arm of its J-shape should not be believed (Momma et al., 2022) -> Muscle-Strengthening Activity and Mortality, The U-Shaped Association Artifact.
Weight loss plus exercise for knee osteoarthritis pain. Combined intensive diet and exercise beats exercise alone for knee-osteoarthritis pain and function (Messier et al., 2013) — and it shares the weight-loss lever already ranked, so it is close to free here.
Exercise for chronic pain and for arthritis symptoms. Exercise is safe and probably gives a small improvement in physical function in chronic pain, at low certainty (Geneen et al., 2017), with arthritis-specific programmes supported for symptom and function benefit (Rausch Osthoff et al., 2018) -> Chronic Pain and Physical Activity.
Exercise and diet quality for mood. Exercise has a moderate antidepressant effect (standardized mean difference ~-0.4 to -0.6, low certainty) (Noetel et al., 2024) and better diet quality is associated with less depression (OR ~0.77, observational and fragile) (Molendijk et al., 2018). These are peripheral levers, admitted mainly through their physical intersection -> Depression and Modifiable Exposures.
Diet for cognition — a candidate lever, weaker than its cardiovascular arm. Older adults who eat more fruit and vegetables have fewer cognitive disorders (OR 0.82, 95% CI 0.75-0.90), the signal landing on cognitive impairment and dementia but null for Alzheimer’s specifically (Zhou et al., 2022); the flavonoid component of the same foods carries a matching but smaller association (adverse cognitive events OR 0.90, 0.83-0.98) (Peng et al., 2026). Both are observational and measurement-error-laden, and they read as one dietary-pattern signal — not two independent witnesses — most likely riding the cardiometabolic route already counted. Eat the whole foods for their better-evidenced reasons; chasing a flavonoid cognition target is not a big rock -> Fruit and Vegetable Intake and Health, Flavonoid Intake and Cognitive Function.
Sleep and cognition — screen a real problem, do not over-read the numbers. Sleep disorders track higher dementia risk in a large cohort meta-analysis, but the arms differ. Insomnia and sleep-disordered breathing carry modest, better-footed associations; the long-sleep -> Alzheimer’s signal (RR 1.66, 95% CI 1.44-1.91) is most likely a preclinical marker of incipient disease rather than a cause (Zhang et al., 2025), and no trial shows that treating a sleep disorder prevents dementia. Treat a genuine sleep problem on its own merits and as a sign of higher baseline risk; do not tell a healthy long sleeper to cut sleep to save their brain -> Sleep and Cognitive Decline.
Alcohol and the brain — no protective dose to chase. Heavy drinking (above ~38 g/day) raises dementia risk; the apparent protection of light drinking is unadjudicated and most likely an artifact of counting former heavy drinkers among the abstainers, with Mendelian randomization finding no protective effect (Xu et al., 2017) -> Alcohol and Mortality and Vascular Disease. There is no evidenced case for taking up or keeping up drinking for cognitive protection.
Dairy for the brain — a null wrapped around a confounded curve. The only dose-response meta-analysis of dairy and dementia (15 cohorts, 312,580 people) is null highest-versus-lowest: RR 0.94 (95% CI 0.82-1.07). Its lone dip near 150 g/day is a between-population artifact, low-intake Asian cohorts reading protective and high-intake European cohorts null at non-overlapping doses; milk and cheese are each null on the matched highest-versus-lowest contrast (Villoz et al., 2024). The pooled evidence is primary-prevention in relatively healthy cohorts — Villoz excluded diabetes and metabolic disease — so it reaches this comorbid stratum only as a conservative null-extrapolation. No dairy lever to add or subtract for cognition -> Dairy, The U-Shaped Association Artifact.
The aging eye — the same rocks, one more organ. The modifiable evidence for age-related macular degeneration adds no new lever. Physical activity is insufficient-evidence rather than protective: high-versus-low activity gives RR 0.92 (95% CI 0.84-1.01), an interval that crosses 1 (Aune et al., 2026). A Mediterranean pattern is protective for AMD progression but not clearly for its development (Marques-Couto et al., 2025), and the risk factors that do carry weight (smoking, hypertension, cardiovascular disease) are the big rocks already ranked (Babaker et al., 2025). So AMD is another organ-specific reason to pull the same levers, not a lever of its own -> Age-Related Macular Degeneration and Modifiable Exposures, Shared Modifiable Levers Across Age-Related Diseases.
Some treatments that help younger adults change sign at 70 — the levers to not pull, or to pull carefully
These are the traps: interventions with a positive or neutral profile earlier in life whose balance tips toward net harm, or toward “not worth it,” in this stratum.
Testosterone for the low-testosterone, obese man: the muscle lever here is training, not a hormone. Testosterone raises lean mass on the scale (+1.6-3.6 kg) but function lags, and in the TRAVERSE trial it increased fractures (HR 1.43, 95% CI 1.04-1.97) — the opposite of the intended effect — alongside more atrial fibrillation, acute kidney injury and pulmonary embolism, with no mortality benefit (Snyder et al., 2024), (López Hilario et al., 2026). And much of the low testosterone in an obese man is secondary to the obesity itself and reverses with weight loss (Corona et al., 2013) — so the fat-loss and resistance-training levers already ranked address the cause -> Testosterone Adiposity and Muscle.
Vitamin D plus calcium is not the fracture lever in the non-deficient. In community-dwelling older adults who are not deficient, supplementation did not reduce fractures and raised kidney-stone risk (RR ~1.18) (Kahwati et al., 2018). The fall/fracture lever is balance exercise, not the supplement.
Intensive weight loss, unmodified, strips muscle and bone. The sign-flip is not weight loss itself but weight loss without the resistance-training-and-protein pairing above — in a sarcopenic-obese person with already-reduced bone density, an aggressive diet alone can worsen the muscle and bone deficit that most threatens independence -> Protein and Resistance Training for Muscle and Strength. Muscle-preserving weight loss is the version to run; the rapid-loss-then-regain pattern also carries its own metabolic-adaptation cost -> Weight-Loss Maintenance and Metabolic Adaptation.
GLP-1 weight-loss drugs (semaglutide, tirzepatide) shed muscle along with the fat — the same sign-flip as unmodified dieting above, now a dominant weight-loss lever in exactly this population.
Across seven obesity-dose trials, rapid GLP-1 weight loss cut absolute lean mass by −1.74 kg (95% CI −3.04 to −0.45, high certainty) — roughly 30% of the weight lost, at or just inside the ordinary 20-30% diet-induced band, so on average the drug is not special: rapid weight loss loses muscle whatever drives it (Laverde et al., 2026). Lean mass is a surrogate for muscle and for function, and Laverde measured neither strength nor capacity — so this is a directional muscle-mass signal, not proven function loss.
Which face governs flips with the stratum. For a young, well-muscled obese adult the composition ratio improves and the loss is minor; for the sarcopenic-obese older adult the absolute loss becomes a patient-important harm (falls, fractures, lost independence) — and Laverde’s young, low-comorbidity trials explicitly leave this high-risk stratum to «future studies» (Laverde et al., 2026). Worse, the drug suppresses intake exactly when defending older muscle needs more per-meal protein to clear the anabolic-resistance threshold — cutting the very protein the muscle now needs more of. Per agent, tirzepatide and subcutaneous semaglutide are «among the most harmful» for lean mass while liraglutide and oral semaglutide «had little or no effect» (Nong et al., 2026) — so a frail patient is worst served by the strongest agents -> Comparing Obesity Drugs.
The same drugs carry a low-certainty observational signal running the other way on the brain: GLP-1 receptor agonists track lower incident dementia in two umbrella reviews (all-cause dementia OR 0.55, 95% CI 0.35-0.87 (Yang et al., 2026); RR 0.35, 0.16-0.78 (Kuate Defo et al., 2023)). Both are observational and hypothesis-generating — a possible offset to log, not a benefit to bank, and it does not soften the muscle-loss caution above -> Dementia Prevention and Modifiable Risk Factors.
The rule is not «don’t use it»: pair any GLP-1 weight loss in this stratum with resistance training and deliberate protein, and judge the complete strategy (drug + training + protein) on function, not the drug’s naive weight number — the same pairing the unmodified-weight-loss bullet above demands -> GLP-1 and Lean Mass, Protein and Resistance Training for Muscle and Strength.
Glycaemic control matters here; tight glycaemic control matters less. Lowering glucose is a baseline-risk (route-a) lever like blood pressure — the higher this stratum’s complication risk, the more a given relative reduction is worth. But two things reweight tight targets downward at 70: the shorter runway buys fewer complication-free years for the slow microvascular endpoints, and the harm of tight control — hypoglycaemia and its downstream falls and cognitive cost — rises with age and frailty. So the target loosens with age; agent and glucose goal are a prescriber’s call, not a number to read off here .
The most aggressive blood-pressure target trades benefit against real harms. The SPRINT benefit is real, but the intensive arm caused more hypotension, syncope and acute kidney injury — so at this age, titrate the target against those harms rather than push it to a number (SPRINT Research Group, 2015).
Statins for primary prevention sit near the edge of the evidence at ~70. The absolute benefit is small (all-cause mortality RR ~0.92), and the U.S. Preventive Services Task Force reaches no recommendation at age >=76 (insufficient evidence) (US Preventive Services Task Force, 2022). Guidance confidence tracks the same edge: at >=70 the European cardiology guideline drops lipid-lowering to Class IIb («may be considered») even in the very-high-risk band (European Society of Cardiology, 2021) — a longevity lever buys fewer disease-free years once competing mortality shortens the runway. That is a claim about horizon, not the drug: where absolute risk is high the lipid lever still pays, since «even a small absolute reduction in LDL-C may be beneficial in a high- or very-high-risk patient» (European Society of Cardiology, 2021) -> Statins for Primary Prevention and the Power of Zero CAC, SCORE2 Baseline Risk and the ESC Treatment Thresholds.
Two statin harms weigh most heavily in exactly this stratum — older adults report the most muscle symptoms, face the most deprescribing pressure, and carry the most polypharmacy — yet neither is a reason to stop an indicated statin, and the two fail as reasons for opposite kinds of reason. The muscle aches are a perceived harm the drug mostly does not cause: across double-blind statin-versus-placebo trials the causal excess is tiny (RR 1.03), and in a blinded n-of-1 rechallenge 90% of the symptom burden a statin provoked was also provoked by an inert placebo (Reith et al., 2022), (Wood et al., 2020) — so a new ache, especially after the first year, is very unlikely the statin, and a blinded rechallenge separates the rare true myopathy from nocebo.
New-onset diabetes is the mirror image: a real pharmacological effect, but small and dose-dependent (RR 1.10 on low- or moderate-intensity statins, 1.36 on high-intensity), concentrated in people whose glycaemia already sits near the diagnostic threshold, driven by a tiny HbA1c shift, and already netted out of the vascular benefit — so it warrants no routine post-initiation glucose monitoring to catch a statin effect (standard diabetes screening continues unchanged) and does not change the start-or-continue decision (Reith et al., 2024). The two harms differ in kind: the muscle harm is mostly not real, the diabetes harm real but outweighed — and both point to the same rule, do not stop an indicated statin, for opposite reasons -> Statin Muscle Symptoms and the Nocebo Effect, Statins and New-Onset Diabetes.
What this ranking cannot yet settle
Three gaps bound the confidence above, and naming them is part of the appraisal.
- No trial has tested muscle- and bone-preserving weight loss against hard outcomes in this exact stratum. The fat-loss-plus-resistance-training-plus-protein pairing is assembled from separate lines of evidence — remission trials, training meta-analyses, fat-free-mass-preservation trials — and no single RCT runs the combined protocol in sarcopenic-obese 70-year-olds to a mortality or fracture endpoint.
- The obesity ‘paradox’ in the oldest old is unresolved. Whether a higher BMI is protective, neutral, or an artifact of reverse causation and smoking confounding shifts with age, and this stratum sits where that argument is least settled -> The U-Shaped Association Artifact.
- The trajectory this person may care about most is the least measured. Trials count deaths and events, not the shape of decline — whether a lever compresses late-life disability or merely prolongs it — so the quality-of-life-trajectory outcome that can outweigh length of life here is systematically under-measured.
- The dementia evidence is mostly borrowed or observational, not stratum-specific. Blood-pressure lowering is the one lever with a randomized dementia effect; the activity, diet, sleep and alcohol arms are observational, and much of the risk-factor evidence is measured in midlife, not at 70. No trial has shown that pulling any of these levers in this stratum prevents dementia, and whether treating a sleep disorder or raising diet quality lowers dementia incidence is untested -> Dementia Prevention and Modifiable Risk Factors.
The bottom line
- If the person smokes, treat quitting as the entire first step — it dwarfs everything below.
- Get blood pressure down, deciding on absolute cardiovascular risk rather than the number, and titrate the target against dizziness, syncope and kidney function at this age.
- Draw down visceral and liver fat for diabetes remission, glucose and fatty liver — and for function — not on a promise it will prevent heart attacks.
- Never run a fat-loss effort without resistance training and adequate protein (~1.0-1.2 g/kg/day minimum, ~25-30 g per meal, higher under training), so what comes off is fat, not the muscle and bone this person cannot spare.
- Do balance and functional training for falls — the highest-certainty lever here; weights build muscle but do not, by themselves, prevent falls.
- Bank the easy activity gains first (the first daily walk buys most of the benefit) and accept that past these levers the remaining returns are small — that ceiling is a result, not a failure.
- The same rocks protect the brain. Getting blood pressure down, staying active and drawing down the metabolic load each lower dementia risk as well — the strongest dementia evidence sits on the levers already at the top, so there is no separate brain programme to run.
- Leave the sign-flipped levers alone or handle them with care: testosterone, calcium-plus-vitamin-D in the non-deficient, unmodified aggressive weight loss, the most intensive BP target, and — at the edge of the evidence — primary-prevention statins.
Four things this ranking cannot tell you
- The loop is open. This page grades whether the reasoning is coherent and faithful to its sources — it cannot grade whether following it makes anyone better off. No operation here checks a recommendation against a realised outcome.
- It appraises, it does not prescribe. Selecting, dosing, titrating, screening for drug interactions and managing this person’s specific medications and labs are prescriber acts that need materials this page does not hold.
- The default is population-level; the person is not the population. These are estimates for a stratum. Which outcomes matter most, what is achievable, and where a contraindication fires are elicited per person — the ranking is the starting point, not the answer.
- One axis only: health. This page weighs longevity, cardiometabolic and musculoskeletal function, falls, pain and independence. It holds no data on cost, environmental load, or anything else, and does not price the health findings against them — that weighting is the person’s.
Evidence box
Question ’For an obese older adult (~70, sarcopenic obesity, hypertension/dysglycemia/MASLD common, reduced BMD, polypharmacy), which modifiable exposures carry the largest expected effect on the patient-important outcomes that matter at this age ranked by effect x certainty, where does the remaining marginal return flatten, and where (if anywhere) does a lever that helps a younger adult change sign for this stratum?‘ Evidence included 55 sources — 36 gold, 17 high, 2 moderate Overall certainty Low (see Rating Certainty of Evidence) Source-selection note 2 source(s) below the gold evidence bar feed this page: Taylor (mechanism, moderate); Song (meta-analysis, moderate). Each labelled by tier; none load-bearing for the core claims. Last updated 2026-09-09 · Independently reviewed: No · Full edit history