REF-1210 · Neary, Bhambhani & McKenzie, 2003 · Canadian Journal of Applied Physiology · Paper · PubMed

How much should I cut volume in a one-week taper?

In 11 trained cyclists, halving volume for 7 days with intervals kept sped a 20-km TT by 5.4%; 30% and 80% cuts were not significant. Tiny arms; a companion report had p = 0.057.

Blimited evidenceMeasured on trained to highly trained athletes

Neary, Bhambhani and McKenzie compared three 7-day step tapers in trained male cyclists. After 3 weeks of baseline training, 11 cyclists were "randomly assigned" to tapers that cut training volume by 30% (T30, n = 5), 50% (T50, n = 6) or 80% (T80, n = 6) while keeping intensity at 85% VO2max, and rode simulated 20-km time trials with cardiorespiratory measurements every 5 km.

Performance improved significantly only after T50 (5.4%), with higher VO2 and O2 pulse; T30 and T80 showed no significant differences. The authors conclude that a moderate (50%) cut in weekly volume appeared optimal and that keeping intensity while placing the volume properly in the taper can improve performance. The arm sizes add to 17, not 11.

A companion paper by the same group on a cohort with the same design (Neary, McKenzie & Bhambhani 2005, Dynamic Medicine, DOI 10.1186/1476-5918-4-4; not in the base) has 11 cyclists and 5 per taper, with some doing more than one taper (its abstract says each did two or three; its Methods say two did all three and three did two, which would fill 18 arm places, not 15), and reports a 4.53% T50 improvement at p = 0.057; its Discussion describes the 2003 study as earlier work "in a slightly larger group (n = 6)" with "a similar protocol".

Our reading: a small cycling dose-response that fits the 41-60% band of REF-107 and REF-3492, with intensity kept; it shows that the 50% taper reached significance and the others did not, not that 50% beat 30% or 80%, and the reporting of the sample is inconsistent.

Protocol

Who: n = 11 male cyclists (VO2max 4.78 ± 0.66 L/min) per the abstract; taper allocations T30 n = 5, T50 n = 6, T80 n = 6.
Baseline: 3 weeks of baseline training before the taper.
Tapers: 7 days, stepwise reduction of training volume to 70%, 50% or 20% of baseline (cuts of 30%, 50%, 80%), intensity kept at 85% VO2max.
Outcome: simulated 20-km time trial before and after the taper, with cardiorespiratory measurements every 5 km.
Not given in the abstract: baseline training volume and content, whether cyclists repeated tapers (and the washout between them), the statistical test, absolute time-trial times, and funding.

Key Results

Taper (7 days, intensity kept)n20-km time trialPhysiology
T30: volume −30%5no significant change—
T50: volume −50%6improved 5.4% (significant; printed "5.4%, 0.05" in the abstract)higher VO2 and O2 pulse
T80: volume −80%6no significant change—
No between-taper test is reported in the abstract; the T30 and T80 changes are not given. In the 2005 companion report the taper sessions (60/55/50/45, 45/40/35/30 and 30/15/10/5 min) against a 4 × 60 min baseline week are cuts of about 12.5%, 37.5% and 75%, not 30/50/80% (our arithmetic from its Table 1).
The sample does not add up: 11 cyclists, but arms of 5 + 6 + 6 = 17.The companion 2005 report (same three tapers, same 20-km time trial, 11 cyclists, VO2max 4.68 ± 0.57 L/min) says cyclists were randomly assigned with 5 per taper and some completed more than one (abstract: each did two or three; Methods: two did all three and three did two, i.e. 18 arm places, not 15), and gives the T50 gain as 4.53% (p = 0.057, Mann-Whitney U).If both describe the same cyclists (our inference), the arms are partly repeated measures, and the two papers disagree on the arm size and on whether the T50 gain was significant.
In a pooled analysis: the Wang 2023 meta-analysis full text (REF-3492, Table 1) lists this study and the 2005 report as two separate studies with identical age (22.6 ± 4.7), sample (11 men) and arms (−30, −50 and −80%), so the same cyclists may have been counted twice (our inference).In its time-trial forest plot (Fig 2) the two reports enter as six data sets with different numbers (n = 6/5/6 and 5/5/5), together 20.3% of the weight, the 2005 three 9.5%; they sit in the 21-40%, 41-60% and ≥60% volume bands, the ≤7-day band and the step-taper band (our reading of Wang's Fig 2 and Table 1).

Mechanisms

Measured: the 50% taper went with higher VO2 and O2 pulse during the time trial.
Authors' interpretation: a moderate volume cut with intensity kept removes enough fatigue to improve performance; a 30% cut removes too little and an 80% cut too much training.
Inference: with 5-6 per arm, "no significant change" after T30 and T80 is no evidence of a difference, not proof of none; this 80%-cut arm of 6 does not outweigh small runner trials in which an 85% cut with intervals kept (REF-1211) and an 80% cut of interval volume with daily training (REF-1200; the group resting every third day did not improve) improved performance within groups; Wang's pooled ≥60% band was not significant either (SMD −0.21, −0.58 to 0.16, REF-3492).

Practical Application

How much should I cut volume in a one-week taper? In 11 trained cyclists, halving volume for 7 days with intervals kept sped a 20-km TT by 5.4%; 30% and 80% cuts were not significant. Tiny arms; a companion report had p = 0.057.
The default stays: about 40-60% off, intensity kept (SYNTH-34; REF-107, A; REF-3492, A-). This is one of the few cycling trials behind the band.
Bigger cuts are not ruled out: 85% with intervals kept improved a 5-km run (REF-1211, B) and Mujika & Padilla put the effective range at 60-90% (REF-1196, B+). Inference, not the base's position (SYNTH-34 keeps 41-60%): an 80% cut that keeps the hard reps may be an option in a short taper, but the pooled ≥60% band in REF-3492 was not significant, so it is not the default.
Keep intensity: every taper here kept work at 85% VO2max; dropping intensity is what the taper review and trials warn against (REF-110, B+; Shepley 1992, REF-1199).
Shape: these were one-step cuts; whether a progressive taper does better is not tested here (SYNTH-34 keeps progressive as the default from REF-107's pooled optimum; REF-3492 found step and progressive both worked).

Limitations

Abstract only. The full text (NRC Research Press) is not open access; the training programme, statistics and absolute times could not be checked.
Tiny arms of 5-6, men only; within-group tests only.
Inconsistent sample reporting between the abstract (11 cyclists, arms of 17) and the companion 2005 report (11 cyclists, 5 per arm, some repeating tapers in an account that does not add up, T50 p = 0.057).
Possible double counting downstream: the cohort appears twice in the REF-3492 table.
One-step tapers of one week after 3 weeks of baseline training; not a test of longer or progressive tapers.
Funding and conflicts: not given in the abstract.
How this entry was checked and changed
Why this level

from the abstract; small (n = 11 cyclists; taper arms of 5, 6 and 6), significance tested within each taper, not between them