REF-700 · Rønnestad BR, Bakken TA, Thyli V, Hansen J, Ellefsen S, Hammarstrøm D — 2022 · International Journal of Sports Physiology and Performance · Paper · PubMed
Fast-start/decreasing-speed (DEC) and varied-intensity (VAR) long intervals both accumulate more time ≥90% VO₂peak than speed- and duration-matched constant-speed intervals (CON) in well-trained XC skiers
And are equivalent to each other
Participants
n=10 well-trained cross-country skiers, double-poling VO₂peak 69.6 ± 3.5 ml/kg/min (age is not in the abstract)
Protocol
Roller-ski double poling; three sessions on separate days in randomized order, speed- and duration-matched: 5 × 5-min work intervals, 3-min recovery. Only the within-interval pacing differed:
DEC — fast start followed by decreasing speed
VAR — systematic variation in intensity within the interval (over-under)
CON — constant speed
Exact sub-interval structure (full text; not checked against the abstract): DEC = 2 min at 100% maximal aerobic speed (MAS) then 3 min at LT2 + 20% of (MAS − LT2); VAR = three 40-s surges at 100% MAS interspersed with three 1-min blocks at LT2 + 20% of (MAS − LT2).
Not from this paper: the cycling translation used in this base (DEC ≈ 2 min @ ~120% FTP → 3 min @ ~105% FTP; VAR ≈ 40 s @ ~120% / 1 min @ ~105% FTP; CON ≈ ~110% FTP) is an approximation, not the study's units.
Key Findings
DEC and VAR led to longer time ≥90% VO₂peak than CON (P = 0.016 and P = 0.033) and higher mean %VO₂peak (P = 0.036 and P = 0.009)
DEC ≈ VAR — no differences between them (P = 0.930 and P = 0.759)
Time ≥90% HRpeak, mean HR, breathing frequency, ventilation and blood lactate were similar across all three formats
No differences in perceptual responses — mean RPE, session RPE, pain score (all P > 0.147)
Per-format values (full text; not checked against the abstract): mean %VO₂max 86.3 ± 2.9 (DEC) / 87.2 ± 2.9 (VAR) / 85.2 ± 2.8 (CON); time ≥90% VO₂max 19.5 ± 6.6 / 19.0 ± 6.6 / 17.6 ± 6.3 min; RPE 15.8 ± 1.5 / 16.2 ± 1.2 / 16.3 ± 1.5
Authors' conclusion: DEC and VAR are a good alternative for accumulating more time at a high %VO₂peak without excessive perceptual effort, and they mimic the intensity variation of XC-ski competition
Practical Application
In 10 well-trained XC skiers (VO₂peak 69.6) doing speed- and duration-matched 5 × 5-min / 3-min roller-ski double-poling intervals, a fast-start-then-decreasing interval (DEC) and an over-under interval (VAR) both produced longer time ≥90% VO₂peak and a higher mean %VO₂peak than constant-speed intervals, with no difference between DEC and VAR and identical HR, lactate, ventilation and RPE — so in a VO₂max block, structure the interior of long intervals (open hard, or alternate over/under) rather than holding a flat pace: more time near VO₂max at the same effort.
Over-unders are validated here as a long-interval VO₂max format, not only a threshold / lactate-clearance tool — equal to fast-start.
Choose by preference and terrain: fast-start suits a single hard opener; over-under suits rolling terrain or ERG variation. Perceived effort was the same, so pick the format the athlete executes cleanly.
Tested in XC skiing (double poling); transfer to cycling is inference — the sibling fast-start finding in cyclists is REF-194.
Limitations
Small (n = 10), well-trained cross-country skiers; acute physiological responses (VO₂ time-in-zone), not a training-adaptation or performance trial. Effort-matched acute VO₂-time superiority is a strong proxy for aerobic stimulus (REF-045 and REF-046 establish that link) but is not a direct adaptation outcome.
Intensities are anchored to MAS / LT2 in roller-ski double poling; the FTP approximations above are indicative, not the study's own units.
Per-format minutes and percentages come from the full text and were not checked against the abstract.
How this entry was checked and changed
Why this level
regraded 2026-09-27 from B+: crossover, n=10
Content correction 2026-09-27
Tier 4-5 -> Tier 3 (abstract: "well-trained", VO₂peak 69.6); Sport "xc-skiing, cycling" -> xc-skiing (cycling transfer is inference).
Format names aligned to the abstract (DEC / VAR / CON); per-format table values, sub-interval structure and "age ~25" marked as full text; FTP conversions marked as not from the paper.
Added the abstract's null results (HR, ventilation, lactate, RPE, session RPE, pain all similar) and P-values; Level B+ -> B (crossover, n=10).