Maximizing Your VO2Max Gains from Interval Training: Breathe Harder, Gain More

#1

Twenty-two well-trained cyclists performed the same interval prescription for nine weeks: 21 sessions of five by eight minutes, with the power output individually matched to each rider’s own 40-minute maximal effort. Because the power was matched, everyone was doing an equally hard workout on paper. But the researchers measured oxygen uptake breath by breath during all 21 sessions, something no previous training study has done, and found the riders were not doing the same workout physiologically. Average oxygen consumption during the work intervals ranged from 74 percent to 91 percent of each rider’s maximum. That spread turned out to predict who improved. Riders in the top half for oxygen uptake gained roughly three times more VO2max than the bottom half, and roughly twice as much peak power and lactate-threshold power. Percentage of VO2max was a better predictor than heart rate, and more reproducible session to session.

Interval training advice has rested for forty years on a plausible but never directly tested assumption: that the more time you spend near your ceiling for oxygen consumption, the more your aerobic system adapts. Everyone from coaches to cardiologists repeats it. Nobody had actually measured it across a full training block.

The problem has always been logistics. Measuring oxygen uptake means wearing a mask connected to a metabolic cart, and researchers typically manage this for one or two sessions out of a twelve-session study. A group at Inland Norway University of Applied Sciences did it for all 21 sessions in a nine-week intervention, which is the reason this paper matters more than its modest size suggests.

Twenty-two competitive cyclists, average VO2max of 67 mL per kg per minute, did the identical protocol. Five work intervals of eight minutes, three minutes of recovery between them, twice weekly, with the power output for each rider set to their own best 40-minute effort measured beforehand. This is the standard way to equalize training load. It failed to equalize the physiological stimulus.

Some riders spent fifteen minutes per session above 90 percent of their VO2max. Others spent two minutes. Same relative power, same duration, same protocol, wildly different internal load. Over the nine weeks the high-uptake group accumulated close to four extra hours above that 90 percent threshold compared with the low-uptake group.

The gains tracked the oxygen. Cyclists in the upper half for oxygen uptake improved VO2max by 8.0 percent against 2.7 percent in the lower half. Peak incremental power rose 8.6 percent against 4.2 percent. Power at the four millimolar lactate threshold, the number that most closely tracks race performance in endurance cycling, rose 7.8 percent against 3.0 percent. Across the group as a whole, statistical modelling suggested that raising average interval intensity from 80 to 90 percent of VO2max would be worth about 2.5 mL per kg per minute of extra VO2max over nine weeks.

Heart rate, the metric almost everyone actually uses, performed poorly. Percentage of maximum heart rate correlated with fewer outcomes and was substantially less reproducible from session to session. The authors are blunt about the implication: heart rate and oxygen uptake should not be treated as interchangeable measures of interval quality.

The obvious caution is that riders were not assigned to high or low oxygen uptake. They sorted themselves. Cyclists with a higher baseline lactate threshold relative to their VO2max tended to end up in the high group, and the high group also logged more total training and considerably more easy riding. So the paper establishes a real and quantitatively large association, not a proven cause. What it does settle is that a power-matched interval session is not an intensity-matched interval session, and that the difference is large enough to change the outcome of a training block.

For anyone tracking VO2max as a longevity marker rather than a race number, the practical message is the same. The protocol is not the stimulus.

Actionable Insights

The headline finding is that two people can do the identical interval workout and get very different results, because the same wattage drives different oxygen consumption in different bodies.

Effect sizes, translated. Comparing the top and bottom halves for oxygen uptake during intervals, VO2max improved 5.23 versus 1.78 mL per kg per minute. Cohen’s d is 1.50, meaning the average person in the high group beat roughly 93 percent of the low group. Peak power improved 0.51 versus 0.23 watts per kilogram, d = 1.10, which for a 75 kg rider is 38 watts versus 17. Lactate threshold power improved 0.31 versus 0.12 watts per kilogram, d = 1.15, or 23 watts versus 9. Large effects by any convention, though the confidence intervals are wide with only 11 people per group.

What to do with it. Long intervals of four to eight minutes at a hard but sustainable effort drive oxygen uptake toward maximum better than short bursts, but only if you actually get there. Do not assume heart rate confirms it; heart rate was the weaker and less repeatable signal here. If you can get one instrumented session with a metabolic cart, it tells you more about your interval quality than a year of heart rate files. Absent that, the working proxy is that a productive set feels breathing-limited by minute three and stays there.

Everyone in this study rode at the same relative power from their own 40-minute max. Nobody was slacking. The 74 to 91 percent spread came from phenotype, not effort, and baseline lactate threshold predicted who landed where. For a typical person training for healthspan the gap isn’t 84 versus 91 percent, it’s never approaching 80 percent at all. So the lever is protocol design rather than measurement.

On targets , the session number is 10 to 15 minutes above 90 percent of VO2max, twice weekly, which is what separated the groups (roughly 318 versus 78 accumulated minutes across the block). But the larger point for a longevity audience is proportion: in Mandsager’s 122,007-patient cohort, moving from the bottom fitness quartile to above average is a hazard ratio of 0.52, and that step requires no interval optimization whatsoever. The fraction-of-VO2max question is a refinement for people already training seriously, not the main event.

Context and Source