Results

Umbrella Review

Study Selection

A total of 817 systematic reviews were identified from the database searches. After removing duplicates and screening titles and abstracts, 46 systematic reviews remained for further full-text screening. Following a careful full-text screening, 14 systematic reviews were included in the umbrella review [14–19, 21–24, 27, 41, 42] (Fig. 1). A full list of excluded studies with reasons is available in Appendix 2 of the ESM.

Fig. 1: Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA) flow diagram illustrating the number of records identified, screened, assessed for eligibility and included in the review. A full list of excluded studies with reasons is available in Appendices 2 and 5 of the ESM. UI urinary incontinence, WOS Web of Science

Quality Assessment

AMSTAR 2 classified 12 of 14 included reviews as critically low confidence, and two were rated low confidence (Appendix 3 of the ESM). The most frequent critical issues were the absence of a pre-registered protocol or a priori design (Item 2), and the failure to provide a list of excluded studies with justification (Item 7), which was absent in all reviews. Additional recurring limitations included inadequate or non-validated risk-of-bias assessment of primary studies (Item 9) [14–18, 41, 42], failure to consider risk of bias when interpreting pooled results (Item 13) [14, 17, 18, 41], and lack of publication bias assessment when a meta-analysis was conducted and appropriate (Item 15) [14, 20, 24, 27]. Across 14 systematic reviews comprising 99 unique primary studies, the corrected covered area was 6.5%, indicating a moderate overlap and supporting the retention of all eligible systematic reviews. The citation matrix used to assess overlap is provided in Appendix 4 of the ESM.

Study Characteristics

The umbrella review included eight systematic reviews and six meta-analyses published between 2017 and 2025, encompassing populations across a wide age spectrum and frequently combining adolescent and adult samples, with heterogeneous parity status reported (Table 1). Only three reviews restricted inclusion to nulliparous women [14, 15, 19]. Competitive level was inconsistently reported and commonly aggregated across amateur, high-level and professional athletes. Only two reviews specifically addressed professional athletes [41, 42]. The scope of the included reviews varied considerably, with qualitative syntheses that ranged from < 5 to > 30 primary studies. Sport classification across reviews was heterogeneous, with most reviews examining mixed sports. Some reviews focused on specific disciplines, particularly volleyball [42], CrossFit [17, 22, 24] and strength-based sports such as powerlifting and weightlifting [23], whereas others adopted broader classifications based on mechanical impact [18, 20, 21]. Outcomes also varied, with a small number of reviews distinguishing between UI subtypes [16, 20, 22, 41, 42].

First author, year | Design | Age range (years) | Parity status | Level of competition | n studiesb | nparticipants | AMSTAR 2 ratingc | Sport classification | Outcome | UI prevalence (%)
Almousa, 2019 [14] | SR | 12–45 | Nulliparous | Mixed | 23 | 2459 | Critically low | Mixed sports | UI | 5.7–80.0
Álvarez-García, 2022 [22] | SRMA | 12–45 | Mixed | Mixed | 13 (12) | 2187 | Critically low | CrossFit | UI | 32.1 (95% CI 22.2–43.8)
 |  |  |  |  |  |  |  | CrossFit | SUI | 35.8 (95% CI 19.4–56.4)
Alves, 2025 [23] | SR | 20–79 | Mixed | Mixed | 5 | 1809 | Low | Powerlifting | UI | 41.0–48.8
 |  |  |  |  |  |  |  | Weightlifting | UI | 36.6–54.1
Cerruto, 2020 [16] | SRMAa | 15–> 45 | Mixed | Mixed | 31 (5) | 8335 | Critically low | Mixed sports | UI | 58.7
 |  |  |  |  |  |  |  | Mixed sports | UI (daily life) | 32.8
 |  |  |  |  |  |  |  | Mixed sports | EXUI | 36.3
 |  |  |  |  |  |  |  | Mixed sports | SUI | 23.0
 |  |  |  |  |  |  |  | Mixed sports | UUI | 11.0
 |  |  |  |  |  |  |  | Mixed sports | MUI | 11.9
Culleton-Quinn, 2022 [41] | SR | 12–89 | Mixed | Professional | 32 | NR | Critically low | Mixed sports | UI, SUI, UUI, MUI | NR
de Mattos Lourenco, 2017 [18] | SR | 12–69 | Mixed | Mixed | 22 | 7507 | Critically low | High-impact | UI | 42.2–61.2
 |  |  |  |  |  |  |  | Medium-impact | UI | 6.3–44.4
 |  |  |  |  |  |  |  | Low-impact | UI | 5.6–15.3
Domínguez-Antuña, 2023 [17] | SRMAa | 18–71 | Mixed | Mixed | 13 (13) | 4823 | Critically low | CrossFit | UI | 44.5
 |  |  |  |  |  |  |  | CrossFit | SUI | 81.2
García-Perdomo, 2022 [27] | SRMAa | 17–43 | Mixed | Mixed | 3 (3) | 1018 | Critically low | High-impact | UI | 12.2–52.9
 |  |  |  |  |  |  |  | Low-impact | UI | 3.8–37.5
Martins, 2017 [15] | SR | 12–45 | Nulliparous | Mixed | 10 | 2272 | Critically low | Mixed sports | UI | 12.0–80.0
Pires, 2020 [21] | SRMA | 18–45 | Mixed | Mixed | 9 | 1254 | Low | High-impact | UI | 25.9 (95% CI 23.5–28.3)
 |  |  |  |  |  |  |  | High-impact | SUI | 20.7 (95% CI 18.5–22.9)
Romero Parra, 2024 [24] | SR | 18–48 | Mixed | Mixed | 7 | 1333 | Critically low | CrossFit | UI | 32.1–44.5
Silva Pereira, 2017 [42] | SR | NR | Mixed | Professional | 5 | 237 | Critically low | Volleyball | EXUI | 9.0–30.0
 |  |  |  |  |  |  |  | Volleyball | UI (daily life) | 17.0–18.0
Syeda, 2024 [19] | SR | 18–45 | Nulliparous | Mixed | 9 | NR | Critically low | Mixed sports | UI | 5.7–80.0
Teixeira, 2018 [20] | SRMA | NR | Mixed | Mixed | 8 (8) | 1714 | Critically low | Mixed sports | UI | 36.1 (95% CI 26.5–46.8)
 |  |  |  |  |  |  |  | Mixed sports | SUI | 43.6 (95% IC 24.4–44.7)
 |  |  |  |  |  |  |  | High-impact | UI | 39.9 (95% IC 34.5–45.2)
 |  |  |  |  |  |  |  | Low-/moderate-impact | UI | 44.1 (95% IC 36.1–52.3)

Urinary Incontinence (UI) Prevalence in Women Athletes Across Systematic Reviews and Meta-analyses

Characteristics of the systematic reviews and meta-analyses are detailed in Table 1. Urinary incontinence was consistently identified as a prevalent condition among sportswomen, although reported prevalence varied widely from 5.7 to 80.0% according to sport modality, mechanical impact profile, population composition and outcome definition. Across UI subtypes, SUI was consistently reported as the predominant, whereas UUI and MUI were less frequently observed.

Stratified findings by sport impact [14, 18, 20, 21, 27] generally reported higher UI prevalence in high-impact sports (12.2–61.2%) compared with low-impact sports (3.8–37.5%). However, meta-analyses pooling single-arm UI prevalence estimates reported overlapping pooled values across impact categories, including 36.1% (95% CI 26.5–46.8) in mixed-sport female athlete populations [20], 44.1% (95% CI 36.1–52.3) in low-impact sports [20], and 25.9% (95% CI 23.5–28.3) and 39.9% (95% CI 34.5–45.2) in high-impact sports [20, 21].

Four reviews were focused on particular sports disciplines. In CrossFit, the overall UI and SUI prevalence of 44.5% and 81.2% [17] and pooled prevalences of UI and SUI of 32.1% (95% CI 22.2–43.8) and 35.8 (95% CI 19.4–56.4) [22], with leakage commonly associated with jump-based exercises (e.g. jump rope single/double unders, box jumps), running and high-load lifts, including deadlifts and squats. In powerlifting and weightlifting, reported UI prevalences were 41.0–48.8% and 36.6–54.1%, with leakage frequently associated with high-load lifts such as squats and deadlifts [23]. In professional volleyball, overall UI prevalence was 9–30% during sports practice and identified urinary leakage as occurring particularly during jumping actions, effort-based and abdominal training, and around competition periods, with higher training and competition exposure associated with more frequent leakage episodes [42].

Conclusions from some reviews were strongly influenced by population composition, particularly age, parity status and competitive exposure. Reviews restricted to nulliparous athletes reported wide prevalence ranges (up to 80%) [14, 15, 19]; however, the highest estimates derived from adolescent and young elite trampolinists in whom involuntary UI was reported during trampoline training [43]. When attention was restricted to adult nulliparous cohorts, UI prevalence estimates were substantially lower and broadly comparable to those reported in parous female athletes, typically with a range from approximately 10–40%. Reviews including mixed parity profiles showed similarly heterogeneous prevalence estimates, reflecting the combined influence of sport-specific mechanical demands and age-related exposure patterns, rather than parity status alone.

Across the included systematic reviews and meta-analyses, several recurrent methodological limitations were consistently highlighted: (1) substantial heterogeneity in sport exposure, outcome definitions and population characteristics, limiting the attribution of UI prevalence to specific sport characteristics; (2) athlete samples frequently pooled multiple disciplines with different impact and load profiles, while sport classification, parity status, age range, competitive level and training volume were inconsistently defined or reported; and (3) the predominance of cross-sectional designs and reliance on self-reported UI further constrained causal inference and increased susceptibility to reporting bias.

Meta-analysis

Study Selection

A total of 1384 primary studies were identified (804 from the reference list and 580 from the forward citation searching). After duplication and title/abstract screening, 114 underwent a full-text review, of which 32 studies [44–76] met the inclusion criteria and were included for quantitative synthesis (Fig. 1). A full list of excluded studies with reasons is available in Appendix 5 of the ESM.

Quality Assessment

AXIS classified 14 of 32 original studies as high quality and 18 as moderate quality (Appendix 6 of the ESM). Egger’s regression test did not indicate small-study effects for competitive level (p = 0.206), sport impact (p = 0.592) or sport modality (p = 0.216). However, evidence of funnel plot asymmetry was observed for sport discipline (p = 0.043), suggesting that smaller studies within specific sport disciplines tend to report systematically different UI prevalence estimates compared with larger studies. Egger’s regression test and funnel plots for all subgroup meta-analyses are shown in Appendix 7 of the ESM.

Study Characteristics

Characteristics of the primary studies are detailed in Table 2. Detailed sport characteristics, UI outcomes and training exposure for each cohort are provided in Appendix 8 of the ESM. Urinary incontinence was primarily assessed using validated self-report questionnaires, most commonly the ICIQ-UI Short Form. Only four studies relied on International Continence Society Criteria-aligned single-item measures and were examined in sensitivity analyses [44, 62, 63, 67]. The included studies comprised a total of 4649 women athletes from 120 cohorts, predominantly nulliparous (93.9% in average). The evidence base was unevenly distributed across sport contexts. Most participants were drawn from ball games and weight-based sports, and the majority competed in high-impact or medium-impact modalities, with low-impact sports under-represented. Competitive level was variably reported, with most cohorts comprising amateur or professional athletes. Overall, the distribution of disciplines and exposure categories highlights substantial clustering of evidence in a limited number of sports, which should be considered when interpreting subgroup analyses.

First author, year | Country | Nulliparous (%) | Age, years, meana | n (with UI) | UI assessment (recall window) | AXIS ratingb
Araujo, 2019 [59] | Brazil | 80.2 | 30.2 (4.6) | 127 (45) | ICIQ-UI SF (4 weeks) | Moderate
Cardoso, 2018 [60] | Brazil | 100 | 21.6 (2.7) | 118 (82) | ICIQ-UI SF (4 weeks) | High
Caylet, 2006 [44] | France | NR | 23.4 (4.5) | 126 (32) | ICS-aligned item (days to months) | High
Clapin, 2025 [45] | France | 96 | 24.0 (4.3) | 159 (79) | ICIQ-UI SF (4 weeks) | High
Culleton-Quinn, 2024 [46] | Ireland | 95 | 23.4 (2.6) | 159 (98) | ICIQ-UI SF (4 weeks) | High
Cygańska, 2025 [47] | Poland | 100 | 24.7 (4.0) | 48 (9) | APFQ (unspecified) | High
da Silva Borin, 2013 [48] | Brazil | 100 | 24.0 (8.5) | 30 (5) | BFLUTS (unspecified) | Moderate
Da Silva Pereira, 2021 [50] | Brazil | 76 | 25.5 | 75 (53) | UDI-6SF (unspecified) | Moderate
Dakic, 2025 [61] | Australia | 89.3 | 26.4 (5.0) | 28 (13) | ICIQ-UI SF (4 weeks) | High
de Souza Pereira 2022 [49] | Brazil | 77.8 | 30.1 | 189 (73) | ICIQ-UI SF (4 weeks) | High
Dockter, 2007 [62] | USA | 100 | 19.2 (1.0) | 109 (51) | ICS-aligned item (lifetime) | Moderate
dos Santos, 2009 [63] | Brazil | 100 | 21.4 (1.7) | 91 (10) | ICS-aligned item (lifetime) | Moderate
Faulks, 2021 [64] | Australia | 86.0 | 26.0 (5.1) | 65 (39) | QUID (2 weeks) | Moderate
Fozzatti, 2012 [51] | Brazil | 100 | 25.7 (5.3) | 600 (89) | ICIQ-UI SF (4 weeks) | Moderate
Gan, 2025 [65] | USA | 100 | 20.0 | 29 (18) | LURN SI-29 (week to year) | Moderate
Gill, 2025 [66] | Australia | 100 | 25.0 | 81 (44) | QUID (lifetime) | High
Hagovska, 2018 [74] | Slovak Republic | 100 | 21.1 (3.9) | 278 (33) | ICIQ-UI SF (4 weeks) | High
Jácome, 2011 [67] | Portugal | 90.6 | 23.0 (4.4) | 74 (34) | ICS-aligned item (lifetime) | Moderate
Kelečić, 2023 [52] | Croatia | 97 | 22.6 (4.6) | 88 (35) | ICIQ-UI SF (4 weeks) | Moderate
Lopes, 2020 [53] | Brazil | 94 | 28.6 (4.5) | 50 (10) | ICIQ-UI SF (4 weeks) | High
Machado, 2021 [54] | Brazil | 100 | 27.4 (3.7) | 20 (12) | ICIQ-UI SF (4 weeks) | Moderate
McCarthy-Ryan, 2024 [73] | UK | 65.7 | 28.0 (8.0) | 396 (250) | ICIQ-UI SF (4 weeks) | Moderate
Middlekauff, 2016 [55] | USA | 100 | 26.8 (3.8) | 35 (9) | EPIQ (not reported) | Moderate
Parr, 2023 [56] | USA | 100 | 20.2 (1.5) | 202 (68) | ICIQ-UI SF (4 weeks) | High
Patrizzi, 2014 [68] | Brazil | 100 | 23.9 (3.8) | 108 (46) | ICS-aligned item (unspecified) | Moderate
Poświata, 2014 | Poland | 76 | 28.1 (5.2) | 112 (67) | UDI-6 (unspecified) | Moderate
Pisani, 2022 [75] | Brazil | 100 | 28.3 (5.9) | 616 (200) | ICIQ-UI SF (4 weeks) | Moderate
Sandwith, 2021 [57] | Canada | 100 | 19.9 (1.8) | 95 (51) | UDI-6 (unspecified) | High
Sebastián-Rico, 2024 [69] | Spain | 100 | 25.7 (4.7) | 235 (82) | ICIQ-UI SF (4 weeks) | High
Tarczewska, 2024 [70] | Poland | 100 | 22.0 | 61 (27) | UDI-6SF (unspecified) | Moderate
Valastro, 2025 [71] | Belgium | 100 | 25.0 | 37 (10) | USP (4 weeks) | High
Winder, 2023 [58] | USA | 87.5 | 25.4 (5.2) | 208 (60) | ICIQ-UI SF (4 weeks) | Moderate

Meta-analysis and Meta-regression of UI Prevalence in Women Athletes

Overall pooled UI prevalence in women athletes was 39% (95% CI 33–46; Fig. 2), with substantial heterogeneity (_I_2 = 94.3%). Pooled prevalences by UI subtype were 33% for SUI (95% CI 27–41; Fig. 3), 14% for UUI (95% CI 7–27; Fig. 4), 12% for MUI (95% CI 6–24; Fig. 5) and 39% for EXUI (95% CI 28–50; Fig. 6), with similar substantial heterogeneity (_I_2 > 93%). Sensitivity analyses revealed slight changes when considering only nulliparous samples (UI: 34%, SUI: 33%, UUI: 16%, EXUI: 28%) and only validated screening tools (UI: 42%, SUI: 32%, UUI: 13%, EXUI: 38%). Unexplained heterogeneity persists despite methodological restrictions (_I_2 = 81–96%). No sport-label classification meaningfully reduced between-cohort heterogeneity in meta-regression analyses.

Fig. 2: Forest plot of the pooled prevalence of urinary incontinence (UI) in women athletes aged 18–45 years. CI confidence interval

Fig. 3: Forest plot of the pooled prevalence of stress urinary incontinence (SUI) in women athletes aged 18–45 years. CI confidence interval

Fig. 4: Forest plot of the pooled prevalence of urgency urinary incontinence (UUI) in women athletes aged 18–45 years. CI confidence interval

Fig. 5: Forest plot of the pooled prevalence of mixed urinary incontinence (MUI) in women athletes aged 18–45 years. CI confidence interval, UI urinary incontinence

Fig. 6: Forest plot of the pooled prevalence of exercise-related/exertional urinary incontinence (EXUI) in women athletes aged 18–45 years. CI confidence interval, UI urinary incontinence

Across sport disciplines, rugby showed the highest pooled prevalence of UI (50%, 95% CI 40–59), exceeding the overall estimate, whereas swimming exhibited the lowest (27%, 95% CI 18–37). However, differences between sport disciplines were statistically inconclusive in the meta-analysis (p = 0.058, _I_2 = 78%; Fig. 7) and not significant in the meta-regression (QM(10) = 11.71; p = 0.304; Table 3). This was consistent with sensitivity analyses in only nulliparous samples (QM(8) = 4.85, p = 0.773, n = 1606) and only validated screening tools (QM(8) = 9.18, p = 0.327, n = 2729).

Fig. 7: Forest plot of the pooled prevalence and subgroup differences of urinary incontinence (UI) in women athletes aged 18–45 years by sport discipline. CI confidence interval

Comparison | Reference | OR (95% CI) | p-value | Comparison | Reference
n UI (total) | k | n UI (total) | k
Basketball | Athletics or track and field | 0.53 (0.21–1.30) | 0.164 | 57 (170) | 10 | 87 (171) | 8
CrossFit | Athletics or track and field | 0.52 (0.21–1.29) | 0.160 | 349 (1037) | 6 | 87 (171) | 8
CrossFit | Basketball | 0.99 (0.41–2.37) | 0.974 | 349 (1037) | 6 | 57 (170) | 10
Gymnastics | Athletics or track and field | 0.19 (0.03–1.06) | 0.058 | 3 (36) | 3 | 87 (171) | 8
Gymnastics | Basketball | 0.35 (0.06–1.97) | 0.235 | 3 (36) | 3 | 57 (170) | 10
Gymnastics | CrossFit | 0.36 (0.06–2.01) | 0.243 | 3 (36) | 3 | 349 (1037) | 6
Handball | Athletics or track and field | 0.49 (0.18–1.34) | 0.165 | 43 (118) | 6 | 87 (171) | 8
Handball | Basketball | 0.93 (0.35–2.46) | 0.886 | 43 (118) | 6 | 57 (170) | 10
Handball | CrossFit | 0.94 (0.35–2.52) | 0.910 | 43 (118) | 6 | 349 (1037) | 6
Handball | Gymnastics | 2.65 (0.45–15.7) | 0.283 | 43 (118) | 6 | 3 (36) | 3
Rugby | Athletics or track and field | 0.94 (0.39–2.26) | 0.881 | 447 (796) | 7 | 87 (171) | 8
Rugby | Basketball | 1.77 (0.76–4.13) | 0.187 | 447 (796) | 7 | 57 (170) | 10
Rugby | CrossFit | 1.80 (0.76–4.24) | 0.181 | 447 (796) | 7 | 349 (1037) | 6
Rugby | Gymnastics | 5.03 (0.91–27.9) | 0.065 | 447 (796) | 7 | 3 (36) | 3
Rugby | Handball | 1.90 (0.73–4.94) | 0.187 | 447 (796) | 7 | 43 (118) | 6
Soccer | Athletics or track and field | 0.34 (0.13–0.90) | 0.030 | 108 (342) | 8 | 87 (171) | 8
Soccer | Basketball | 0.65 (0.25–1.66) | 0.364 | 108 (342) | 8 | 57 (170) | 10
Soccer | CrossFit | 0.66 (0.25–1.70) | 0.386 | 108 (342) | 8 | 349 (1037) | 6
Soccer | Gymnastics | 1.84 (0.32–10.7) | 0.497 | 108 (342) | 8 | 3 (36) | 3
Soccer | Handball | 0.70 (0.25–1.96) | 0.492 | 108 (342) | 8 | 43 (118) | 6
Soccer | Rugby | 0.37 (0.15–0.92) | 0.032 | 108 (342) | 8 | 447 (796) | 7
Softball | Athletics or track and field | 0.62 (0.17–2.27) | 0.471 | 21 (55) | 3 | 87 (171) | 8
Softball | Basketball | 1.17 (0.33–4.21) | 0.805 | 21 (55) | 3 | 57 (170) | 10
Softball | CrossFit | 1.19 (0.33–4.30) | 0.789 | 21 (55) | 3 | 349 (1037) | 6
Softball | Gymnastics | 3.34 (0.47–23.7) | 0.228 | 21 (55) | 3 | 3 (36) | 3
Softball | Handball | 1.26 (0.33–4.86) | 0.737 | 21 (55) | 3 | 43 (118) | 6
Softball | Rugby | 0.66 (0.19–2.35) | 0.524 | 21 (55) | 3 | 447 (796) | 7
Softball | Soccer | 1.81 (0.48–6.83) | 0.379 | 21 (55) | 3 | 108 (342) | 8
Swimming | Athletics or track and field | 0.37 (0.13–1.05) | 0.063 | 24 (98) | 6 | 87 (171) | 8
Swimming | Basketball | 0.69 (0.25–1.94) | 0.488 | 24 (98) | 6 | 57 (170) | 10
Swimming | CrossFit | 0.71 (0.25–1.99) | 0.509 | 24 (98) | 6 | 349 (1037) | 6
Swimming | Gymnastics | 1.97 (0.32–12.1) | 0.461 | 24 (98) | 6 | 3 (36) | 3
Swimming | Handball | 0.75 (0.24–2.28) | 0.608 | 24 (98) | 6 | 43 (118) | 6
Swimming | Rugby | 0.39 (0.14–1.08) | 0.070 | 24 (98) | 6 | 447 (796) | 7
Swimming | Soccer | 1.07 (0.36–3.19) | 0.899 | 24 (98) | 6 | 108 (342) | 8
Swimming | Softball | 0.59 (0.15–2.38) | 0.460 | 24 (98) | 6 | 21 (55) | 3
Tennis | Athletics or track and field | 0.42 (0.12–1.41) | 0.160 | 14 (60) | 5 | 87 (171) | 8
Tennis | Basketball | 0.79 (0.24–2.61) | 0.698 | 14 (60) | 5 | 57 (170) | 10
Tennis | CrossFit | 0.80 (0.24–2.67) | 0.718 | 14 (60) | 5 | 349 (1037) | 6
Tennis | Gymnastics | 2.24 (0.33–15.1) | 0.408 | 14 (60) | 5 | 3 (36) | 3
Tennis | Handball | 0.85 (0.24–3.04) | 0.799 | 14 (60) | 5 | 43 (118) | 6
Tennis | Rugby | 0.45 (0.14–1.46) | 0.181 | 14 (60) | 5 | 447 (796) | 7
Tennis | Soccer | 1.22 (0.35–4.26) | 0.757 | 14 (60) | 5 | 108 (342) | 8
Tennis | Softball | 0.67 (0.15–3.08) | 0.608 | 14 (60) | 5 | 21 (55) | 3
Tennis | Swimming | 1.14 (0.30–4.25) | 0.851 | 14 (60) | 5 | 24 (98) | 6
Volleyball | Athletics or track and field | 0.72 (0.30–1.71) | 0.452 | 113 (242) | 11 | 87 (171) | 8
Volleyball | Basketball | 1.35 (0.58–3.14) | 0.479 | 113 (242) | 11 | 57 (170) | 10
Volleyball | CrossFit | 1.37 (0.59–3.22) | 0.465 | 113 (242) | 11 | 349 (1037) | 6
Volleyball | Gymnastics | 3.85 (0.70–21.2) | 0.122 | 113 (242) | 11 | 3 (36) | 3
Volleyball | Handball | 1.45 (0.56–3.75) | 0.439 | 113 (242) | 11 | 43 (118) | 6
Volleyball | Rugby | 0.76 (0.34–1.74) | 0.522 | 113 (242) | 11 | 447 (796) | 7
Volleyball | Soccer | 2.09 (0.84–5.22) | 0.114 | 113 (242) | 11 | 108 (342) | 8
Volleyball | Softball | 1.15 (0.33–4.06) | 0.824 | 113 (242) | 11 | 21 (55) | 3
Volleyball | Swimming | 1.95 (0.71–5.33) | 0.193 | 113 (242) | 11 | 24 (98) | 6
Volleyball | Tennis | 1.72 (0.53–5.58) | 0.369 | 113 (242) | 11 | 14 (60) | 5

Across sport modalities, ball games showed the highest pooled prevalence of UI (43%, 95% CI 35–52), whereas technical sports reported the lowest (18%, 95% CI 10–31). Meta-analysis yielded significant differences (p = 0.018, n = 4,458; Fig. 8) but under a high heterogeneity (between-group _I_2 = 89.0%; within-group _I_2 range = 0–91.2%). Meta-regression did not identify sport modality as a significant moderator (QM(4) = 6.76; p = 0.149; Table 4). This was consistent with sensitivity analyses in only nulliparous samples (QM(4) = 2.81, p = 0.590, n = 2602) and only validated screening tools (QM(4) = 4.64, p = 0.326, n = 3902).

Fig. 8: Forest plot of the pooled prevalence and subgroup differences of urinary incontinence (UI) in women athletes aged 18–45 years by sport modality. CI confidence interval

Comparison | Reference | OR (95% CI) | p-value | Comparison | Reference
n UI (total) | k | n UI (total) | k
Ball games | Aesthetic | 2.16 (1.02–4.59) | 0.044 | 998 (2146) | 21 | 104 (415) | 9
Endurance | Aesthetic | 1.76 (0.75–4.11) | 0.195 | 175 (532) | 12 | 104 (415) | 9
Endurance | Ball games | 0.81 (0.42–1.56) | 0.531 | 175 (532) | 12 | 998 (2146) | 21
Technical | Aesthetic | 0.63 (0.14–2.77) | 0.539 | 10 (56) | 3 | 104 (415) | 9
Technical | Ball games | 0.29 (0.07–1.15) | 0.079 | 10 (56) | 3 | 998 (2146) | 21
Technical | Endurance | 0.36 (0.08–1.50) | 0.161 | 10 (56) | 3 | 175 (532) | 12
Weight | Aesthetic | 1.42 (0.62–3.26) | 0.406 | 413 (1309) | 14 | 104 (415) | 9
Weight | Ball games | 0.66 (0.35–1.23) | 0.187 | 413 (1309) | 14 | 998 (2146) | 21
Weight | Endurance | 0.81 (0.39–1.70) | 0.576 | 413 (1309) | 14 | 175 (532) | 12
Weight | Technical | 2.26 (0.54–9.42) | 0.261 | 413 (1309) | 14 | 10 (56) | 3

Across sports, medium-impact disciplines showed a higher pooled prevalence of UI (43%, 95% CI 35–52) than high-impact (34%, 95% CI 27–42) and low-impact (33%, 95% CI 19–50). However, sport impact showed no significant effect on UI pooled prevalences both in the meta-analysis (p = 0.257, n = 4649, _I_2 = 91.2%; Fig. 9) and meta-regression (QM(2) = 2.75; p = 0.253; Table 5). This was consistent with sensitivity analyses in only nulliparous samples (QM(2) = 1.83, p = 0.400, n = 2793) and only validated screening tools (QM(2) = 1.69, p = 0.428, n = 4093).

Fig. 9: Forest plot of the pooled prevalence and subgroup differences of urinary incontinence (UI) in women athletes aged 18–45 years by sport impact. CI confidence interval

Comparison | Reference | OR (95% CI) | p-value | Comparison | Reference
n UI (total) | k | n UI (total) | k
Low impact | High impact | 0.93 (0.45–1.95) | 0.857 | 690 (2140) | 22 | 121 (545) | 8
Medium impact | High impact | 1.47 (0.88–2.46) | 0.137 | 690 (2140) | 22 | 921 (1964) | 21
Medium impact | Low impact | 1.58 (0.76–3.29) | 0.225 | 121 (545) | 8 | 921 (1964) | 21

Across competitive levels, professional athletes showed the highest pooled prevalence of UI (49%, 95% CI 39–59), exceeding the overall estimate, whereas amateurs exhibited the lowest (32%, 95% CI 23–42). Although the subgroup meta-analysis did not reach conventional statistical significance in the meta-analysis (p = 0.074), the observed direction and magnitude of the estimates suggest a potential gradient across competitive levels, which could not be confirmed given substantial between-study heterogeneity (_I_2 = 92.9%). This trend was confirmed in meta-regression analysis (QM(2) = 5.19, p = 0.075; Table 6), where pairwise contrasts suggested higher odds of UI in professional athletes compared with amateur athletes. This pattern was not observed in analyses restricted to nulliparous samples (QM(2) = 2.69, p = 0.260; n = 2756), but approached statistical significance when restricted to studies using validated screening tools (QM(2) = 6.10, p = 0.050; n = 3660) (Fig. 10).

Comparison | Reference | OR (95% CI) | p-value | Comparison | Reference
n UI (total) | k | n UI (total) | k
High level | Amateur | 1.51 (0.75–3.02) | 0.246 | 595 (2253) | 14 | 320 (709) | 8
Professional | Amateur | 2.05 (1.10–3.83) | 0.024 | 595 (2253) | 14 | 557 (1254) | 11
Professional | High level | 1.36 (0.66–2.80) | 0.399 | 320 (709) | 8 | 557 (1254) | 11

Fig. 10: Forest plot of the pooled prevalence and subgroup differences of urinary incontinence (UI) in women athletes aged 18–45 years by competitive level. CI confidence interval

Meta-regression of training volume and intensity showed that greater weekly training hours were associated with higher odds of UI (per + 1 h·week⁻1: OR = 1.06, 95% CI 1.01–1.11, p = 0.011, n = 2362 athletes, Fig. 11). Using the sample SD, increments of 5.24 h·week⁻1 correspond to OR = 1.38 (95% CI 1.07–1.76). There was no evidence that the association differed by sports impact (p = 0.454) or competitive level (p = 0.750). This was consistent with sensitivity analyses in only nulliparous samples (per + 1 h·week⁻1: OR = 1.06, 95% CI 1.01–1.11; sample SD = 5.5 h·week⁻1, OR = 1.49, 95% CI 1.12–1.99, p = 0.011, n = 1308) and only validated screening tools (per + 1 h·week⁻1: OR = 1.08, 95% CI 1.02–1.13; sample SD = 5.56 h·week⁻1; OR = 1.40, 95% CI 1.05–1.85, p = 0.007, n = 2188).

Fig. 11: Association between weekly training volume (h·week⁻1) and urinary incontinence (UI) prevalence in women athletes aged 18–45 years. Each point represents an individual cohort, plotted as the observed UI prevalence (UI cases/total sample). Point size is proportional to cohort sample size (n). The solid line shows the predicted UI prevalence from a random-effects meta-regression (logit-transformed proportions; random-effects model estimator), and the shaded band represents the 95% confidence interval around the model prediction. Rug marks along the x-axis indicate the distribution of cohort training volumes