Ironman series, part one. How endurance performance is built by the system rather than by a single number.
VO2 max has quietly become shorthand for aerobic fitness. It is a single figure, it moves in response to training, and a number that behaves this obligingly is easy to trust. Within limits, that trust is earned, because VO2 max does describe something real: the highest rate at which the body can take up and use oxygen under hard effort. The trouble is that those limits are tighter than the reputation, and for anything that lasts several hours the result is decided mostly by properties the figure never records.
This is an educational and strategic perspective, not personal medical advice.
A ceiling you rarely touch
VO2 max is defined at maximum effort, and a long endurance event almost never visits that intensity. The test drives an athlete to exhaustion and reads off the top of oxygen uptake, a genuine trait reflecting how much blood the heart moves and how well muscle pulls oxygen from it. It is also the trait that shifts fastest under short blocks of hard intervals, which is a large part of why people fixate on it. A number that climbs inside a couple of months feels like proof of progress, and in the narrow sense of lifting the ceiling, it is.
What that ceiling does not tell you is how fast someone can travel for hours, because the sustainable working level is run by partly separate machinery. Holding seventy to seventy-five percent of maximum heart rate across an afternoon asks how efficiently the aerobic system runs below its peak, how long that efficiency lasts as fatigue builds, and where the fuel comes from over time. A maximal test is close to blind to all three, which is why two people with an identical VO2 max can finish a long race far apart.
Why the number stops predicting past ninety minutes
Below a certain level of aerobic capacity, lifting VO2 max clearly helps, because the ceiling itself is the constraint. Above it, and for trained endurance athletes that level often sits around the low fifties in millilitres per kilogram per minute, extra headroom adds little to long-distance pace, because the athlete is no longer limited by the maximum. They are limited by how much of it they can hold, for how long, and how steadily.
That capacity now has a name. Researchers have formalised durability as the resistance of an athlete's physiological profile to decay over prolonged exercise, and proposed it as a fourth determinant of endurance performance (Maunder et al., 2021). A rested test, VO2 max included, describes the body at the start line. Durability describes how much of that profile is still standing in the fourth or fifth hour, and it is the survival, not the starting value, that separates a strong finish from a fade. A rested measurement cannot see it, which is much of why the number on the report and the time on the clock can disagree.
What actually carries the hours
Several slower adaptations do the real work, and each answers to a different stimulus than the one that lifts a maximal test. One is mitochondrial: endurance training raises both the volume and the oxidative enzyme activity of muscle mitochondria, shown clearly when Holloszy found the oxidative capacity of trained muscle roughly doubling (Holloszy, 1967). Later human work adds that the density of the internal cristae membranes rises too, running higher in endurance-trained muscle and tracking with whole-body oxygen uptake (Nielsen et al., 2017). Between them they set how much sustained aerobic work the muscle can produce at a moderate pace.
Another is delivery. Sustained low-intensity volume thickens the capillary network around each muscle fibre; the classic human study recorded roughly a twenty percent rise in capillary density after eight weeks of training (Andersen and Henriksson, 1977). Denser capillaries move oxygen in and metabolites out more effectively at submaximal effort, and since peak power is capped by neural recruitment rather than oxygen supply, this helps the long steady effort far more than the maximal one.
The third is metabolic, and it is the one a maximal test is least able to see. Over months the muscle grows better at burning fat and clearing lactate, so that at a given submaximal pace more energy comes from fat and less from a finite glycogen store. Comparisons of professional endurance athletes with less-fit people show exactly that: more fat oxidation and lower blood lactate at matched intensities, the signature of metabolic flexibility (San-Millán and Brooks, 2018). Across many hours this is close to decisive, since glycogen runs out and fat does not, so the athlete who spares glycogen arrives at the final hour with reserves. A maximal test measures carbohydrate use at full effort, nearly the opposite of what the race relies on.
These systems also lean on one another. More mitochondria without the capillaries to feed them, or better fat oxidation without the recovery to sustain the training that builds it, simply meets whichever component fell behind. They rise together, over months, on sustained aerobic work, which is the honest reason they are less gratifying to pursue than a figure that jumps in a fortnight.
Where the training pattern comes from
The distribution that grows these adaptations is well documented and remarkably stable across sports. Examined closely, the training of highly trained endurance athletes tends to settle near eighty percent of sessions at low intensity and the rest concentrated in genuinely hard work, with little time in the moderate zone that feels productive but sits in between (Seiler, 2010). It keeps reappearing because it matches what the body responds to: easy volume builds the base, a small dose of hard work holds the top end, and the middle is avoided as a default because it costs a lot of fatigue for adaptations it does not specifically drive. A controlled nine-week comparison in well-trained athletes pointed the same way, with a polarised model outperforming a threshold-centred one on the variables that matter for endurance (Stöggl and Sperlich, 2014).
A decline that is not a loss
Set out like this, the counterintuitive part is simple. A block built on easy volume plus a little hard work is not the stimulus that keeps a maximal test at its peak, since holding VO2 max there needs regular near-maximal efforts that an endurance block deliberately keeps scarce. So the maximum can drift down over such a block while the systems that carry a race grow stronger. Ordinary things like accumulated fatigue or measurement noise can nudge it too, but either way the lesson holds: a falling number alongside improving sustained performance is evidence that VO2 max was never the limiting factor.
My own case is only an illustration, not proof. In the months before a long triathlon in Cascais my measured VO2 max fell by around three points, which reads on paper as lost fitness, yet the race itself, close to eleven hours, held at a steady heart rate across all three disciplines with almost no drift from start to finish. The steadiness was the signal worth reading, because it showed the aerobic and metabolic systems doing their job, while the falling number described a capacity the day never asked for.
Measuring the thing that decides the race
The practical shift is in what you choose to track. A maximal test reports a ceiling you rarely reach, worth having as one input and no more. For a long event the more useful signals describe sustained capacity: how stable heart rate stays over a long effort, whether pace at a fixed heart rate improves across a block, how efficiently the body burns fat at race intensity, and how well output holds into the later hours. These resist collapse into one tidy figure, which is precisely why they get passed over in favour of one that is easy to publish and compare. The reasonable posture is to treat VO2 max as a threshold check, confirm it is high enough, then stop optimising it and spend the attention on the adaptations that decide a long day.
Where Atlas Cove fits
The same logic reaches well past sport. Durable change in health, like durability in a race, comes less from one sharp intervention than from the right distribution of work and recovery held steady over months, with measurement chosen to match the decision rather than the number that is easiest to show. A short reset can begin the process; what carries it is the same patient accumulation that builds a long-distance engine, and the same willingness to judge progress by what a system can sustain rather than by what it can peak at once. Read that way, a maximal figure that slips while endurance improves is not a warning but a price paid knowingly for the capacity the distance actually rewards.
Tom Wuerden · Co-Founder
Engineer turned Ironman