REF-1725 · Barrajón et al., 2026 · Cureus, · Paper · PubMed

Lower-back pain or discomfort fell after an individual bike fit in all three studies a systematic review could find (312 cyclists: one randomised trial of a 3D motion-capture fit and two uncontrolled before-after studies, follow-up 30 days to six months), but the review pooled nothing and prints no effect sizes, so the size of the benefit and how much of it is the fit rather than time or expectation are unknown; nothing here tests injury prevention. Read in full

Main caveat: Three studies, two without a control group; the uncontrolled improvements may reflect time, natural recovery or expectation.
limited evidenceMeasured on recreational to trained athletesRead in full

Barrajón and colleagues (Zaragoza, Lleida and Alicante) set out to collect every study that changed a cyclist's bike set-up and measured lower-back pain, and to compare static with dynamic fitting. Five databases searched to October 2025 yielded three eligible studies.

A 1999 study tilted the saddle nose down by 10-15° in 80 recreational cyclists and followed them for six months: 72% reported their back pain gone, 20% much better and 7% unchanged, with no control group.

A 2021 study fitted 160 amateur mountain bikers with a 3D motion-capture system and found pain, comfort and fatigue scores better 30 days later (p < 0.001), again without a control group.

A randomised trial in 72 trained cyclists, as the review describes it, compared a 3D fit with the riders' usual set-up and found less back discomfort in the fitted group at 30 days (p < 0.001), with less spinal flexion and more symmetrical leg movement.

The review rated the trial 8 of 10 on PEDro and the two uncontrolled studies 12 and 11 of 16 on MINORS, pooled nothing for lack of comparable data, and concluded that individual, preferably dynamic, fitting appears to reduce lower-back pain and should be part of injury prevention. In its own discussion it calls the certainty of the evidence moderate to low.

Our reading: every study moved in the same direction, but two of the three have no comparison group, the outcomes are self-rated by riders who knew they had been fitted, and the review's prevention conclusion is not something any included study measured.

Participants

3 studies, 312 cyclists: 80 recreational (30 F / 50 M, aged 17-72), 160 amateur mountain bikers (120 M / 40 F, mean age 38.7 ± 8.0) and 72 trained cyclists (49 M / 23 F, mean age 36.5 ± 7.9).
Whether riders were recruited because they had back pain is not stated for the two later studies; the 1999 result (resolution in 72%) implies its riders had back pain.

Protocol

Search: Boolean string of cycling terms, bike-fit terms and back-pain terms in five databases, inception to October 2025; reference lists screened; two independent screeners with a third to settle disagreements.
Inclusion: recreational or competitive cyclists of any age or sex; any bike-fit or set-up change (saddle height or setback, bar position, a full static or dynamic fit); a comparison (before vs after, usual position or another fit); a back-pain outcome or related posture measure; randomised, quasi-experimental, cohort or case-control designs.
Quality: PEDro for the randomised trial, MINORS for the non-randomised studies, two reviewers by consensus; summarised narratively. No GRADE rating.
Synthesis: qualitative only, grouped by intervention and outcome; no meta-analysis, no heterogeneity statistic, no sensitivity or reporting-bias analysis.
Funding and interests: no financial support received; all authors declare no relevant financial or other relationships.

Key Results

Study (design, quality)RidersIntervention, follow-upResult as reported in the review
Salai 1999 (before-after, MINORS 11/16)80 recreationalSaddle tilted 10-15° nose down, set from fluoroscopic analysis; 6 monthsBack pain gone in 72%, much better in 20%, unchanged in 7% (p < 0.01); no control
Scoz 2021 (before-after, MINORS 12/16)160 amateur mountain bikers3D kinematic fit (Retül) to recommended joint ranges; 30 daysPain (VAS), comfort (FEEL) and fatigue (OMNI) all improved (p < 0.001); back pain improvement called moderate; no control
Dias Scoz 2022 (RCT, PEDro 8/10)72 trained cyclists3D dynamic fit (Retül) vs usual set-up; 30 daysLess back discomfort in the fitted group (p < 0.001); less spinal flexion and better lower-limb symmetry
No mean differences, confidence intervals or effect sizes are given for any study; the text speaks of moderate-to-large effect sizes for the two 3D-fit studies without printing them.
Quality detail (Table 1): only the trial randomised, concealed allocation and blinded its assessors; no study blinded participants or fitters; both uncontrolled studies lacked unbiased assessment of endpoints.
The mechanisms named in the abstract (pelvic tilt, spinal alignment, lower-limb kinematics) are the authors' interpretation; only the trial reports kinematic measures.

Practical Application

A bike fit may ease a back that hurts on the bike, but the evidence is thin: one 30-day randomised trial and two uncontrolled studies, all self-rated. That fits the base's position that a fit is reasonable for comfort or a complaint that persists (SYNTH-92); it does not show by how much, and nothing here tests prevention.
Where it differs from the guide: the review concludes that bike fitting should be a key part of cycling injury prevention. SYNTH-92 says a fit has never been shown to prevent injury; this review adds no prevention outcome, so the guide's position stands, and the review's conclusion goes beyond its data.
Saddle tilt is a new lead, not a rule. In 1999, 80 recreational riders whose saddle was tilted 10-15° nose down mostly reported their back pain gone after six months, with no comparison group. Inference, not tested here: a small forward tilt is one change a rider with back pain could try, one change at a time and judged after about two weeks as SYNTH-92 says; that procedure is the base's practice.
Ask first what the guide asks first. In a survey of 1,274 amateurs, back pain before taking up cycling went with four-fold odds of back pain in the past year, and a past bike-fit visit dropped out of the model (REF-1265); load and earlier episodes come before geometry (SYNTH-92).
Not treatment advice. Back pain that travels down a leg, numbness or weakness, or pain that does not settle goes to a physician or physiotherapist (SYNTH-73).

Limitations

Three studies, two without a control group; the uncontrolled improvements may reflect time, natural recovery or expectation.
All pain outcomes are self-rated by riders who knew they had been fitted; follow-up 30 days in both 3D-fit studies.
No pooled estimate and almost no numbers: no mean differences, intervals or effect sizes are reported, so the size of any benefit cannot be read.
The review cites its randomised trial as Dias Scoz et al., 2022 and describes 72 trained cyclists (49 M / 23 F), a Retül fit against the usual set-up, VAS discomfort and FEEL comfort, and 30 days.The base's entry for a trial by the same first author, first published online in November 2022, describes 162 amateurs, a posture handout as control, FEEL, OMNI and a numeric pain scale, and 15 days (REF-1262, read from its abstract). First author, year and the FEEL scale match; sample, control, pain scale and follow-up all differ.The review's reference list is not in the text read, so which paper it cites cannot be checked: it may be the same trial misdescribed or a different report, and neither is assumed here.
Table 1 credits the uncontrolled 1999 study with between-group comparisons and mixes PEDro and MINORS items in one list, so the appraisal is loosely reported.
PROSPERO registration was retrospective; only English and Spanish studies; the authors themselves note moderate to low certainty.
No funding and no conflicts declared.
How this entry was checked and changed
Why this level

a systematic review whose quality appraisal we read, but it rests on three studies, two of them uncontrolled before-after designs, synthesised narratively with no pooled estimate; the one controlled result is a single trial of 72 riders as the review describes it; the base holds a trial by the same first author at B+ (REF-1262), but whether the review means that trial cannot be settled (Limitations)