Physical activity and cardiorespiratory fitness are among the most reproducible modifiable determinants of cardiovascular and all-cause mortality. Two new papers this month sharpen the picture. A stair-climbing meta-analysis puts a large number on an everyday activity. A human proteomic study gives exercise intensity a molecular signature. Set against the existing evidence, exercise clears a bar most observational lifestyle claims do not.

The evidence in brief:

• Stair climbing. 9 studies, 480,479 people. 39 percent lower cardiovascular mortality (RR 0.61) and 24 percent lower all-cause mortality (RR 0.76). Almost entirely observational, and the cardiovascular figure is the soft one (see below).
• Cardiac rehabilitation (randomized). 85 trials, 23,430 patients. Cardiovascular mortality RR 0.74 (NNT 37) and lower reinfarction, but no significant effect on all-cause mortality at one year.
• Fitness as a marker. Each 1-MET higher cardiorespiratory fitness tracks with roughly 11 to 17 percent lower all-cause mortality, graded and modifiable in both directions.
• Type. In a one-year randomized trial, aerobic and combined training improved the cardiovascular risk profile. Resistance training alone did not.
• Intensity. Sprint-interval exercise remodeled the plasma proteome roughly twentyfold more than moderate exercise, and the intensity-sensitive proteins mapped preferentially onto cardiometabolic protection. Association plus mechanism, small male cohort.

The pattern is what makes exercise unusual. The observational mortality numbers leak where such numbers usually do, at all-cause mortality, while the randomized trials, the dose-response gradient, and now a mechanism converge on a real cardiovascular benefit. The rest of this piece works through the evidence by age, type, and intensity, and marks where the observational figures should be discounted.

Start with the honest weakness. The stair-climbing meta-analysis is built almost entirely on cohort studies, and its headline all-cause mortality figure is a 24 percent reduction. That is a very large number for a single behavior, and largeness is the tell. When an everyday habit appears to lower your risk of dying from everything at once, from sepsis and cancer and car accidents, the habit is usually a marker of who you already are rather than a cause of what happens next.

The tell shows up again when you compare the population studies to the randomized ones. In cohorts, active people have roughly 31 percent lower all-cause mortality than inactive people. Then you run the actual trials, and the giant all-cause number gets smaller and, in the best data, disappears. Hold that thought. It is the whole argument.

The stair-climbing number deserves its own asterisk, and it cuts against the grain of what comes next. The pooled 39 percent reduction in cardiovascular death is larger than the best single cohort can support. The UK Biobank stair-climbing analysis, the largest of its kind, found lower all-cause and cancer mortality but no association with cardiovascular death (HR 1.08). So the figure in the headline that looks most cardiovascular is the least stable one. Keep the direction and discount the magnitude. The durable cardiovascular evidence in this piece comes from the randomized trials, not the stairwell.

The relative benefit holds across the lifespan, and in absolute terms it often grows, because older people start closer to the events you are trying to prevent. A pooled analysis of four multinational megacohorts found benefit in every age band, with older adults reaching maximum benefit at higher volumes, around four to five times the guideline minimum, versus roughly the guideline dose in younger adults.

The reassuring part for anyone counseling an eighty-year-old is that the intervention gets simpler with age, not harder. In the Cardiovascular Health Study, older adults who walked faster than three miles per hour had roughly half the coronary and stroke risk of those who walked under two. Pace alone carried signal. What shifts with age is the mix. The emphasis moves from aerobic capacity in early adulthood toward strength, balance, and neuromuscular work in the seventh and eighth decades, where a prevented fall is its own cardiovascular event.

Resistance training earns its own protection, and it saturates early. Observational data put the independent association at roughly 15 percent lower all-cause mortality and 17 percent lower cardiovascular risk, with most of the benefit reached at a strikingly low dose, about 30 to 60 minutes per week. More lifting past that does not buy much more mortality benefit, which is not what the gym marketing implies.

When you actually randomize the modalities, the ranking is clean. The CardioRACE trial put 406 overweight or obese adults into aerobic, resistance, combined, or control for a year. Aerobic training improved the composite cardiovascular risk score (change of −0.15, p=0.01). Combined training improved it (−0.16, p=0.009). Resistance training alone did not (−0.02, p=0.69). Lift for your muscles and your glucose. Do not count on the barbell by itself to fix your blood pressure and your lipids. Observationally, combined training shows the strongest mortality associations of the three, roughly 40 to 46 percent lower all-cause and cardiovascular mortality against 18 to 29 percent for either mode alone, though those figures carry the usual cohort caveat.

For a fixed total volume, a higher share of vigorous work adds a little more benefit. In a cohort of 403,681 people, moderate and vigorous activity lowered mortality similarly per unit, but a larger fraction of vigorous-to-total activity was independently associated with lower all-cause mortality. Accelerometer data point the same way. Intensity adds to volume rather than replacing it, and interval work raises fitness faster for less time, which matters because fitness is the marker that tracks mortality best.

This is where the new mechanistic paper is genuinely interesting, and where discipline matters most. Olsen and colleagues give intensity a plausible molecular signature. The short, brutal dose does not just burn more; it releases a different and apparently more protective set of signals into the blood, from muscle, liver, and fat. That reframes intensity as a possible active ingredient rather than only a way to accumulate dose faster. It is a strong hypothesis with a small, male, mechanism-level dataset underneath it. Promising. Not yet an outcome.

The dose-response does not reverse at the top. Even at very high volumes, up to something like seven to ten times the guideline minimum, cardiovascular and all-cause mortality stay low. The curve plateaus rather than turning back on itself. The one real exception is rhythm. Lifelong high-volume endurance athletes, overwhelmingly middle-aged men, carry a two to tenfold higher risk of atrial fibrillation, along with more coronary calcium and patches of myocardial fibrosis. The signal is muted or absent in women. The AHA frames it correctly. The absolute risk is small and the survival benefit of a lifetime of training still wins. It is an asterisk, not a warning label.