REF-1261 · Bini & Priego-Quesada, 2022 · Journal of Sports Sciences · Paper · PubMed

How should I set my saddle height?

From your knee angle measured while pedalling, not sitting still (graded strong in a 41-study review). Changes under ~4% of leg length barely change joint loads or efficiency; injury effect unknown.

Main caveat: From the abstract only. No PMC copy; the publisher version is closed.
moderate evidenceMeasured on recreational to highly trained athletes

Bini (La Trobe University; AUT, Auckland) and Priego-Quesada (University of Valencia) set out to update and expand the 2011 narrative review of saddle height (REF-882, same first author): how saddle height should be determined, and what changing it does to performance and injury-risk outcomes.

They searched Medline, Scopus, CINAHL, OVID and Google Scholar, screened 29,398 records, read 66 studies in full and included 41, appraising them with the Appraisal tool for Cross-sectional Studies.

They rate the evidence as strong that saddle height should be configured using dynamic measurements of the knee angle, and that limb kinematics change with saddle height; moderate that changes of less than 4% of leg length produce trivial to small changes in lower-limb loads and no effect on oxygen uptake or efficiency; and limited on supramaximal performance and on injury risk.

Our reading: the method message changes the base — the knee angle that sets the saddle should be read while the rider pedals, not with the rider still on the bike (the base's 25-30° comes from the 2011 review and does not say how it was measured) — while the injury message does not: no study here shows that a saddle-height change prevents or relieves pain.

From the abstract, so the dynamic target range, the size of the static-dynamic difference and which studies carry each grade could not be read.

Protocol

Design: systematic review; aim stated as updating and expanding the 2011 narrative review.
Sources: Medline, Scopus, CINAHL, OVID and Google Scholar; all documents (dates not given in the abstract).
Screening: 29,398 records; full-text screening of 66; n = 41 studies included.
Appraisal: the Appraisal tool for Cross-sectional Studies (AXIS) for methodological quality and risk of bias; how the evidence grades (strong, moderate, limited) were assigned is not in the abstract.
Outcome families: methods to determine saddle height; lower-limb kinematics and loads; oxygen uptake and efficiency; supramaximal performance; injury risk.
Funding and interests: the Crossref and PubMed funder field carries only the fragment "associated with the work featured", probably the end of a statement that no funding was associated with the work (our reading); conflict-of-interest statement not available (full text not read).The first author wrote the 2011 review this one updates (REF-882) and appears to be the author of a published letter on the 3D bike-fit trial (REF-1262); the second author is first author of the bike-fit survey REF-1263. No industry affiliation is printed.

Key Results

QuestionGrade (authors, abstract)Finding
How to set saddle heightStrongConfigure it from dynamic measurements of the knee angle
Saddle height and limb kinematicsStrongLimb kinematics change with saddle height
Change under 4% of leg length: lower-limb loadsModerateTrivial to small changes
Change under 4% of leg length: oxygen uptake and efficiencyModerateNo effect
Supramaximal performanceLimited—
Injury riskLimited—
Search yield: 29,398 records screened, 66 read in full, 41 included.
Scale of "4% of leg length" (our arithmetic): for a leg length of 85-95 cm it is about 3.4-3.8 cm; adjustments of a few millimetres sit far inside it. The abstract does not say which leg-length measure (inseam, trochanteric height) is meant; REF-882 used trochanteric length for the same 4% threshold.
Not in the abstract: the dynamic knee-angle range recommended, the static-dynamic difference, formula methods (inseam percentages, heel method) and how they compared, numbers of participants, and effect sizes.

Mechanisms

Measured (as summarised by the review): saddle height changes lower-limb joint kinematics; changes under 4% of leg length change joint loads trivially to slightly and leave oxygen uptake and efficiency unchanged.
Authors' interpretation: saddle height should be configured from dynamic measurements of the knee angle (graded strong); the reason, for example how far static and pedalling angles differ, is not given in the abstract.
Inference: the ankle moves while pedalling (the foot drops at the bottom of the stroke), so a knee angle read with the rider still and a knee angle read during pedalling are different numbers; a target set for one method cannot be applied with the other. The size of the difference is not in this abstract.
Inference: joint-load changes are a mechanism for pain, not a pain outcome; that a large change in saddle height alters knee loading makes it a plausible contributor to knee pain, as REF-882 said, but no included study is reported to show that changing it alters pain.

Practical Application

How should I set my saddle height? From your knee angle measured while pedalling, not sitting still (graded strong in a 41-study review). Changes under ~4% of leg length barely change joint loads or efficiency; injury effect unknown.
Measure while pedalling (B+, from the abstract): the review grades dynamic knee-angle measurement strong; that a static goniometer reading and a pedalling reading differ is our inference (no size is given in the abstract), so any target angle must name its method.The base's 25-30° of knee flexion (REF-882; used in SYNTH-73, SYNTH-55 and SYNTH-81) is the 2011 recommendation and does not state a measurement method; this abstract gives no dynamic range to replace it. Not from this paper; the base's own practice: do not quote an angle without saying how it was measured.
Small tweaks: under about 4% of leg length (3-4 cm for most adults, our arithmetic) joint loads change only trivially to slightly and oxygen uptake and efficiency not at all; a few millimetres up or down is not a performance or loading decision. Whether a few millimetres changes a symptom was not tested.
Knee pain: saddle height stays a candidate, not a proven cause — injury evidence is limited here, and conflicting in REF-1260. Not from this paper; the base's own practice: ask about the load change first, then one adjustment at a time with a re-check after about two weeks (SYNTH-73).

Limitations

From the abstract only. No PMC copy; the publisher version is closed; the University of Valencia repository copy listed by OpenAlex sits behind a bot check and was not read. The dynamic target, the methods compared, the grading rules and each study's size could not be checked.
Evidence type. Cross-sectional and laboratory studies of mechanics and physiology; strong and moderate are the authors' grades for measurement and mechanical questions, not for pain or injury, where the evidence is limited.
An appraisal tool for cross-sectional studies does not assess crossover order effects or measurement validity in depth; how "strong" was defined is not in the abstract.
Same author as the review it updates. Agreement with REF-882 on the 4% threshold is partly the same group reading the same literature; the shift to dynamic measurement is the change.
Interests. No funding (our reading of a truncated funder field); no conflict statement could be read; no industry affiliation is printed.
How this entry was checked and changed
Why this level

from the abstract; a systematic search of five sources (Medline, Scopus, CINAHL, OVID, Google Scholar) with quality appraisal, but the included studies are cross-sectional and laboratory studies of kinematics, joint loads and physiology, the strong and moderate grades are the authors' own and concern how to measure and what changes mechanically, the injury evidence is limited, and the number of studies and participants behind each grade is not in the abstract. Same level as REF-882, which it updates with a systematic method