REF-1271 · Gan et al., 2021 · Sexual Medicine Reviews · Paper · PubMed

Does cycling cause erectile dysfunction?

Unclear: 6 pooled studies showed no difference unadjusted and twice the odds (OR 2.00) only after adjusting for age and health. Observational; it cannot show cause.

Main caveat: Observational inputs: cross-sectional comparisons pooled; associations only.
moderate evidenceMeasured on non-athletes to trained athletes

Gan and colleagues (University of North Carolina urology, with the North Carolina Translational and Clinical Sciences Institute and the UNC Health Sciences Library) searched several major databases from inception through 2018 for English-language studies with original data that compared erectile dysfunction (ED) between cyclists and non-cycling controls using a validated measure — the International Index of Erectile Function (IIEF) or its short form, the Sexual Health Inventory for Men (SHIM).

From each group they extracted age, SHIM score and comorbidities; ED was defined as SHIM ≤ 21. Six of 843 screened studies qualified, with 3,330 cyclists and 1,524 controls. Without adjustment, cyclists and controls did not differ in the odds of ED or in mean SHIM score.

With adjustment for age and comorbidities in a generalized linear mixed-effects model, cyclists had higher odds of ED (OR 2.00, 95% CI 1.57 to 2.55). The authors conclude that limited evidence supports a positive correlation between cycling and ED after adjustment, and that heterogeneity suggests some groups of cyclists may be more vulnerable.

Our reading: an honest counterweight to the "cyclists are fine" message of REF-1268, but the signal exists only after an adjustment made at the level of study groups, the inputs are cross-sectional comparisons, and the six studies are not named in the abstract.

The fair summary for an athlete is that the link between cycling and ED is uncertain, and that numbness and pain on the saddle are the signals to act on.

Protocol

Search: "several major databases" from inception through 2018 (names not given in the abstract), terms on cycling and erectile dysfunction.
Eligibility: written in English; original data; ED compared between cyclists and non-cycling controls; a validated ED measure such as the IIEF or SHIM.
Extraction: age, SHIM score and comorbidities for every group.
Outcomes: mean SHIM score and presence of ED (SHIM ≤ 21) per group; unadjusted odds ratio and mean SHIM difference; both adjusted for age and comorbidities.
Analysis: generalized linear mixed-effects model.
Included: 843 studies evaluated; 6 included; n = 4854 men (3,330 cyclists and 1,524 non-cycling controls). The six studies, their designs and countries are not listed in the abstract.
Affiliations, funding and interests: University of North Carolina School of Medicine (Gan also University of Pennsylvania; Coward also UNC Fertility LLC, Raleigh), North Carolina Translational and Clinical Sciences Institute, UNC Health Sciences Library (PubMed record).No funding is deposited in the Crossref record and no conflict-of-interest statement is in the PubMed record; the paper's own statements could not be read. No saddle or bicycle company appears among the affiliations.A letter (McManus A, Connor MJ, Desai A, Miah S, "The Hard Truth Regarding Cycling and Erectile Dysfunction?", Sex Med Rev 2020;8(4):635-636) and the authors' reply (636-637) were published; neither was read.

Key Results

Cyclists vs non-cycling controlsUnadjustedAdjusted for age and comorbidities
Odds of ED (SHIM ≤ 21)no significant difference (value not given in the abstract)OR 2.00 (95% CI 1.57 to 2.55)
Mean SHIM scoreno significant difference (value not given)not given in the abstract
Heterogeneity: described by the authors; no I² or between-study variance in the abstract.
Another figure for the same paper: the systematic review REF-1270 cites this meta-analysis (as "Gao et al.") with an odds ratio of 1.55 "when controlling for age"; that figure is not in this abstract, and whether it is an age-only model in the full text could not be checked.
The likely largest input: the men's survey REF-1268 (2,774 cyclists and 1,158 controls) meets the stated inclusion criteria and would be most of the pooled 3,330 and 1,524 if it is one of the six; its own multivariable SHIM means favoured cyclists (20.7 and 19.9 vs 19.5). The abstract does not list the six studies, so this is our inference.

Mechanisms

Measured: no unadjusted difference in ED or SHIM between cyclists and non-cyclists; higher adjusted odds of ED in cyclists.
Authors' interpretation: limited evidence of a positive correlation after adjustment; heterogeneity points to subgroups of cyclists who may be more vulnerable and deserve study.
Inference: an estimate that moves from no difference to OR 2.00 on adjustment means the cyclists in these studies were younger or healthier than the controls, so the crude comparison hid a difference — or that the adjustment over-corrects. Adjusting with group-level age and comorbidity values (the abstract says these were extracted for each group) is prone to ecological bias, and the direction of the correction cannot be checked from the abstract.
Inference: the route usually proposed is compression of the pudendal neurovascular bundle on the saddle, which the laboratory studies in REF-1270 support as an acute effect on penile perfusion; this meta-analysis does not test any route and cannot say that cycling causes ED.

Practical Application

Does cycling cause erectile dysfunction? Unclear: 6 pooled studies showed no difference unadjusted and twice the odds (OR 2.00) only after adjusting for age and health. Observational; it cannot show cause.
Give the honest two-sided answer (B+, from the abstract, observational): the largest survey in the base found cyclists' SHIM scores no worse than swimmers' and runners' (REF-1268); this pooled analysis found higher adjusted odds of ED. Neither shows cause.
Act on saddle symptoms: standing regularly and a bar level with the saddle went with less genital numbness in men (REF-1268); standing and no-nose saddles protect penile perfusion in the lab (REF-1270).
See a doctor. Not from this paper; the base's own practice: a new or persistent problem with erections is a reason to see a doctor whatever the suspected cause; the assistant does not diagnose and does not attribute it to cycling.

Limitations

Observational inputs: cross-sectional comparisons pooled; associations only.
Adjustment carries the result: no difference unadjusted, OR 2.00 adjusted; the adjustment used values extracted per group, which can mislead when individuals differ within groups.
Heterogeneity acknowledged by the authors; the six studies and their quality are not given in the abstract.
Inconsistent citation: another review quotes this meta-analysis as OR 1.55 controlling for age (REF-1270).
English only, search to 2018.
Abstract only; a published letter and reply were not read.
Funding and conflicts: not available in the Crossref or PubMed records; one author also lists UNC Fertility LLC.
How this entry was checked and changed
Why this level

(from the abstract) — one step below the rubric's A- line (an SR/MA "with serious heterogeneity") because the six inputs are observational comparisons, not trials, the only positive estimate appears after an adjustment for age and comorbidities built from values extracted per group, the studies were heterogeneous, and only the abstract was read (Elsevier; no open copy listed by OpenAlex)