HR vs VO₂ as marker of HIT aerobic effectiveness

The aerobic effectiveness of a HIT session is conventionally described by *time spent at a high fraction of VO₂max* (≥90% VO₂max is the usual convention, not an outcome-validated threshold in this base; REF-041 below related the mean %VO₂max of the work intervals, not time ≥90%, to adaptation). REF-041 (n=22 well-trained cyclists, 21 sessions of 5×8 min at 40-min power over 9 weeks, VO₂ measured in every session) found that the %VO₂max sustained during the work intervals predicted the gains: VO₂max (R² adj = 0.54), maximal aerobic power (R² = 0.44), a performance index (R² = 0.36) and power at 4 mmol/L (R² = 0.25).

But VO₂ is not measurable in routine training without a metabolic cart.

Athletes and coaches default to % HRmax + power-time-at-target as field-side surrogates. REF-041 also evaluated the %HRmax surrogate against the same adaptation outcomes — it was related to fewer outcome variables and showed poorer session-to-session repeatability than %VO₂max; the authors conclude %VO₂max is the measure that best reflects the magnitude of adaptation.

A direct within-session demonstration: in REF-194 (11 well-trained cross-country skiers, 5×5 min, abstract-checked 2026-09-27) fast-start intervals produced a higher peak and average VO₂ than constant-speed intervals at the same mean speed (98.2 vs 95.4% and 87.6 vs 86.1% VO₂peak) with a lower RPE — and no difference in mean heart rate.

Same HR, slightly more oxygen uptake (about 1.5 points of VO₂peak on average; one acute session in skiers — a transfer for cyclists and runners).

Inference from the measured dissociation above; the mechanisms in this table are physiological reasoning (the REF-041 entry carries stroke volume, day-to-day noise and HR plateau as interpretation or inference, not measurement).

Tier 1 (most robust — the tier ranking in this matrix is the base's own practice, not from a study; no entry here tested power-time-at-target against adaptation): Power-based dose. Time at or above a defined %PO40min / %MAP target (e.g. time >90% MAP, time at ≥VO₂max-power). Cycling-specific and unaffected by SV, hydration, T, caffeine. Use as primary intra-session metric.

Caveat: durability-decoupling (REF-423; REF-426 in that range is unverified) means late-session power at the same %FTP costs more — read alongside cumulative duration.

⚠ Tier 1 invalid sub-metric: "work above CP/FTP" in kJ. Although this number is prominently displayed by WKO5, TrainingPeaks, intervals.icu and similar tools, it does NOT track time>90% VO2max or mean %VO2max across HIT formats — they dissociate (REF-130 Vaccari 2020: the abstract reports time >90% VO₂peak of 312 s for decreasing-length intervals vs 182 s for short 30/20 and 179 s for long 3/2 intervals, at similar time to exhaustion, lactate and RPE; the work-above-CP comparison is full-text data carried by the REF-130 entry (not in the abstract; the entry checked it against the full text 2026-09-27): mean %VO₂max 84.2 / 81.5 / 79.6% and work above CP 22.0 / 18.7 / 19.3 kJ for decreasing / short / long — the format with by far the most time >90% VO₂peak had only slightly more work above CP than the long format, and the short format had less work above CP than the long one at the same time >90%).

VO2 during HIT depends on the interaction of work intensity, recovery intensity, and work:rest ratio — not on total work-above-threshold integral (inference).

Short formats keep VO2 elevated during brief recoveries (slow off-kinetics) and accumulate %VO2max minutes despite modest work-above-CP totals (inference, not tested in REF-130 — there the short format did not beat the long one on time >90% VO₂peak). Use time-at-or-above target power, NOT total kJ above CP/FTP, for session-quality evaluation (the kJ dissociation is REF-130's full-text data; time-at-target-power as the replacement is the base's own practice, not from REF-130).

Same logic-class as the %HRmax caveat below — a popular convenience metric that doesn't track the underlying aerobic signal.

Tier 2: %HRmax time-in-zone, interpreted with confounder control. Useful for:

Tier 3: Subjective markers — RPE per interval, ability to hit target power on the last work-bout, breathing quality. Cheap, fast, surprisingly informative when combined with Tier 1/2.

Tier 4 (optional, expensive): Direct VO₂. Cosmed K5 or similar portable metabolic systems are emerging in 2026 but remain impractical for routine training. Reserve for lab benchmarking.

Not from a study; the base's own practice (the tiers, the three-number readout, the 4-6 week trend and the 3-5 session rule are house procedure built on the evidence above).