Article Text

Performing high-level sport is strongly associated with urinary incontinence in elite athletes: a comparative study of 372 elite female athletes and 372 controls
  1. Alice Carvalhais1,2,3,
  2. Renato Natal Jorge3,4,
  3. Kari Bø5
  1. 1Department of Physiotherapy, CESPU, North Polytechnic Institute of Health, Paredes, Portugal
  2. 2Faculty of Sport, University of Porto, Porto, Portugal
  3. 3Institute of Science and Innovation in Mechanical and Industrial Engineering, Porto, Portugal
  4. 4Faculty of Engineering, Department of Mechanical Engineering, University of Porto, Porto, Portugal
  5. 5Department of Sports Medicine, Norwegian School of Sport Sciences, Oslo, Norway
  1. Correspondence to Alice Carvalhais, Department of Physiotherapy, CESPU, North Polytechnic Institute of Health, Rua Central de Gandra, 1317, 4585-116 Gandra, Paredes, Portugal; acarvalhais{at}


Objective To evaluate the prevalence of urinary incontinence (UI) in female elite athletes compared with controls and to investigate potential risk factors for UI among elite athletes.

Methods This cross-sectional study included 372 elite athletes (athletes group, AG) and 372 age-matched controls (control group, CG). The median age was low (19 years) and the vast majority were nulliparous. Potential risk factors, including clinical, demographic and sports practice characteristics, were collected by using a questionnaire. The International Consultation on Urinary Incontinence Questionnaire-Urinary Incontinence-Short Form was applied to estimate the prevalence of UI. OR with 95% CIs were used to estimate the association with UI. The final model was adjusted for constipation, family history of UI and history of urinary infection.

Results The prevalence of UI was 29.6% and 13.4% in AG and CG, respectively (p<0.001). The following prevalences were obtained: AG: 19.6% and CG: 3.5% (p<0.001) for stress UI, AG: 3.8% and CG: 5.4% (p=0.292) for urgency UI and AG: 5.9% and CG: 0.8% (p<0.001) for mixed UI. After adjustment, performing high-level sport (adjusted (adj) OR=3.31; 95% CI 2.20 to 4.97), family history of UI (adj OR=1.54; 95% CI 1.04 to 2.29), history of urinary infection (adj OR=1.53; 95% CI 1.05 to 2.23) and constipation (adj OR=1.79; 95% CI 1.07 to 2.98) were associated with UI.

Conclusion The prevalence of UI among Portuguese female elite athletes is high and the odds of UI were three times higher than in controls. Also, constipation, family history of UI and history of urinary infections were significantly associated with UI.


Statistics from


Urinary incontinence (UI) is a common problem among females, with a negative impact on quality of life.1 2 UI is defined as ‘the complaint of any involuntary leakage of urine’.3 Previous studies have reported that the prevalence rates of UI in young women aged 15–44 years in the general population range from 6.2% to 12.4%.4–9 Stress urinary incontinence (SUI) is defined as ‘the complaint of involuntary loss of urine on effort or physical exertion, or on sneezing or coughing’3 and is the most common type of female UI, with prevalence rates varying from 3% to 58.4%.10 The wide variation in prevalence rates is explained by the different definitions used and populations studied.11 12

Although moderate physical activity has been reported to decrease the odds of having a risk of developing UI, strenuous work/exercise has been listed as a possible risk factor for the condition.13 14 A systematic review by Bø15 showed that the highest prevalence of UI during sports involved high-impact activities, such as gymnastics and ball games. Regarding the prevalence of UI in different sports, the rates vary between 0% in golf players16 and 88.9% in artistic gymnastics and trampoline athletes.17

Two opposing hypotheses have been suggested regarding the effect of exercise on the pelvic floor muscles (PFMs). One proposes that the increased impact of abdominal pressure and ground-reaction forces may lead to simultaneous or precontraction of the PFM, resulting in a training effect of the muscles. The other suggests that strenuous physical activity without simultaneous co-contraction of the PFM may overload, stretch and weaken the pelvic floor.15

Until now, several risk factors for SUI in the general population have been identified such as vaginal birth, age, diabetes mellitus, obesity, recurrent urinary tract infections, neurological diseases and chronic bronchitis,1 but few epidemiological studies on the prevalence and risk factors of UI in female elite athletes have been published.18–21 Hence, the aim of the present study was to evaluate the prevalence of UI in female elite athletes of different sports compared with age-matched controls and to investigate possible risk factors for UI, including sports practice and anthropometric characteristics, in addition to medical history.


This is a cross-sectional study of Portuguese female elite athletes. Data on the athletes were collected between November 2014 and October 2015, whereas data on a control group (CG) were collected between April 2015 and February 2016. Presidents of the larger national sports federations in Portugal were contacted by email and asked to participate in the study. Five of 18 presidents did not respond. A total of 416 elite athletes were identified from the different sports federations. Of this number, the authors were able to contact 373 athletes and 372 (89.4%) agreed to participate in the research. For comparison, a CG of 372 individuals was included. Controls were contacted in high schools, universities, malls and other public areas. Women were approached by one of the researchers (AC) and invited to participate in the study and if they met the selection criteria, they fulfil the questionnaire’. The participation rate was 92.5%. Matching procedure was performed in a 1:1 ratio, taking the year of birth as reference.

Female elite athletes aged 15–48 years who had been part of their respective national teams for at least 1 year and who had reached an international level of competition were included in the study. Age-matched controls were recruited from the general population. The exclusion criteria were pregnancy at the time of the study or during the last year, any illnesses or inability to understand the Portuguese language. A specific exclusion criterion for the control participants was practice of exercise more than twice weekly.

Data were collected with the use of hand-delivered questionnaires. Each questionnaire was divided into three sections. Section 1 included data on general background variables, such as age, weight, height and parity. Section 2 contained questions related to medical history, such as common illnesses, medications, constipation, urinary infection, family history of UI, gynaecological and obstetric history and whether UI had been reported to a physician. Section 3 covered questions on sports practice characteristics, such as type of sport, years of sports participation and volume of training (hours/week). In addition, the athletes were asked about UI in relation to sports practice and, if present, whether this occurred during training and/or competition, as well as the impact of UI on sports performance and whether any type of protection was worn to prevent leakage.

UI was assessed by using the International Consultation on Incontinence Questionnaire-Urinary Incontinence-Short Form (ICIQ-UI-SF),22 which is a reliable and valid questionnaire for assessing the prevalence, severity, impact on quality of life and type of UI.22 The ICIQ-UI-SF has been translated into Portuguese language, and its Portuguese translation has been validated.23 The questionnaire includes three scored items. The first item asks: ‘‘How often do you leak urine?’’ (0, never; 1, about once a week or less often; 2, two or three times a week; 3, about once a day; 4, several times a day and 5, all the time). The second item asks: ‘‘How much urine do you usually leak (whether you wear protection or not)?’’ (0, none; 2, a small amount; 4, a moderate amount and 6, a large amount). The third item asks: ‘‘Overall, how much does leaking urine interfere with your everyday life?’’ (a Visual Analogue Scale is used, ranging from 0, not at all, to 10, a great deal). A fourth non-scored item asks about the patient's perception of the cause and type of leakage. UI was defined as any UI (one or more positive responses to the fourth question of ICIQ-UI-SF). Frequency of UI was assessed by the first item; the second item was used to evaluate the amount of leakage and the third item determined the overall impact of UI on quality of life. When the involuntary loss of urine was associated with coughing, sneezing, physical activity or exercise, UI was classified as SUI. Involuntary loss of urine before reaching the toilet was classified as urgency urinary incontinence (UUI). Cases of both SUI and UUI were classified as mixed urinary incontinence (MUI). The remainder was classified as other types of UI.

The sports included in the present study were classified into seven groups as previously described19 24 (table 1).

Table 1

Sports disciplines under the different sports groups and the corresponding number of participants

The study was carried out according to the principles of the Declaration of Helsinki and was approved by the Ethics Committee of the Faculty of Sports of the University of Porto (CEFADE 17.2014). Written informed consent was obtained from all participants.

Statistical analysis

Statistical analyses were done by using the SPSS, version 23 (IBM, Chicago, Illinois, USA). Quantitative variables were described as medians and IQRs, and categorical variables were described as counts and percentages (%). The medians between athletes and the CG were compared by using the Mann-Whitney test and the proportions by applying the χ2 test or Fisher's exact test. The associations were assessed by binary logistic regression analysis with the use of crude and adjusted ORs (OR and adj OR) and respective 95% CIs. The associations were adjusted for constipation, family history of UI and history of urinary infection, and the model was based on the significance of crude associations. Significance level was set at p<0.05.


A total of 744 women were included in this study: 372 in the elite athletes group (AG) and 372 in the CG. The vast majority of the women had normal body mass index (BMI) and did not report constipation or family history of UI. Few were parous or had undergone gynaecological surgery. The controls presented higher values regarding history of urinary infection. The athletes reported less constipation, had lower parity and were less likely to have a family history of UI and a history of urinary infection compared with the controls (table 2).

Table 2

Sociodemographic, anthropometric, medical history and sports practice characteristics of athletes and controls

Table 3

Prevalence of UI and characteristics of leakage in athletes and controls

Athletes who practised gravity sports (G7) reported the highest prevalence of UI (84.4%) compared with all the others sports practice groups (G1: 14.3%, G2: 28.6%, G3: 21.4%, G4: 44.0%, G5: 25.8% and G6: 33.3%). The differences were statistically significant if all groups were considered (p<0.001), but no differences were found between the other groups after exclusion of G7 (p=0.263).

The prevalence rates of SUI in the different sports disciplines ranged from 0% (figure roller-skating, swimming and weightlifting) to 82.4% (trampolining) (figure 1).

Figure 1

Prevalence (%) of stress urinary incontinence within each sport. *Tennis/Beach tennis/Table tennis.

Athletes had three times higher odds of UI than controls and this remained significant after adjusted for constipation, family history of UI and history of urinary infection. Also, constipation, family history of UI and history of urinary infection had a significant effect (table 4). In athletes, a weak association was found between UI and hours of practice per week (crude OR=1.04; 95% CI 1.00 to 1.08). No association was found between UI and years of sports practice (crude OR=1.02; 95% CI 0.98 to 1.07).

Table 4

Potential risk factors of UI in athletes and controls: multivariate analysis

Most of the athletes with UI (n=110) reported urinary leakage during sports practice (82; 74.5%), often in the middle/at the end of training/competition (69; 84.1%). In 45 (55.6%) athletes, UI occurred only during sports practice. Twelve (14.6%) of the athletes with UI applied strategies to reduce visible leakage, most often the use of pads (9; 75%); 32 (39.1%) considered that the leakage affected their sports performance. None of the athletes with UI representing synchronised swimming, horse jumping, roller hockey, football/futsal, judo/karate or volleyball reported that UI affected their sports performance.


Our study shows a higher prevalence of UI in elite athletes than in controls, with SUI being the most prevalent type of UI. High-level sports practice was associated with three times the odds for UI compared with controls. Additionally, constipation, family history of UI and history of urinary infections were associated with UI. These results correspond with the findings of other studies.1 5 7 25–27 Because the controls presented higher prevalence rates of these risk factors, one would expect them to have a higher prevalence of UI compared with the elite athletes; however, this was not the case.

Differences between groups regarding BMI, constipation, parity, family history of UI and history of urinary tract infections were found, but these differences were small and may be considered of no clinical relevance. However, as BMI, constipation and to a certain degree, urinary tract infections are modifiable risk factors for UI, prophylactic measures can be implement in order to promote continence. Other studies carried out in young female athletes, either in recreational or competitive sports practice,28 are in agreement with results reporting high prevalence rates of both UI and SUI.29 Only a few studies included a CG,17 19 30 31 and in all except one,19 a higher prevalence was found in athletes compared with controls. The different prevalence rates obtained may be due to environmental, sociocultural and genetic factors; the inclusion of a CG is important to rule out the influence of sports practice on the pelvic floor and UI symptoms.

How the prevalence of UI compares in different sports?

In the present study, differences in the prevalence of UI were also found across sports groups, with the rates ranging from 14.3% in technical sports to 84.4% in gravity sports. Contradictory to our findings, in other studies, the highest rates were obtained in technical sports19 31 and the lowest in gravity sports.31 Differences in the sports included in these studies may explain this finding. Consistently with former research,17 18 21 32 female trampoliners showed the highest prevalence of SUI. One would think that the impact leading to leakage would be the landing phase. However, Eliasson et al32 found that leakage occurred mainly when doing double somersaults and when rehearsing new, strenuous and difficult exercises.

Regarding horse riding, to our knowledge, this is the first study to report on the prevalence of UI in horse riding female elite athletes competing in horse jumping events; the rate is among the lowest values obtained for the different sports modalities in this study.33 Alanee et al33 found that female horse riders did not present an increased risk of UI and suggested that equestrian sports might rather decrease the risk of female UI. Riders contract the hip adductors to maintain their balance, and a synergistic contraction of the striated urethral wall muscle and the PFM has been found during hip adductor muscle contraction;34 thus, horse riding might strengthen the PFM. To our knowledge, no previous studies have evaluated the muscle strength of the PFM in horseback riders. Further clinical studies are needed to clarify this hypothesis.

Regarding ball games, the volleyball athletes in our study reported a significantly lower prevalence of SUI compared with, for example, basketball and roller hockey. There are inconsistencies across studies in the reported prevalence rates of UI in volleyball, which vary between 4.3% and 44.4%.17 35 36 Although volleyball players perform leaps in the execution of a shot, they have a more stationary stance because they do not move across the field throughout the game. Thus, the impact of ground reaction forces on the pelvic floor may be less in volleyball than in other high-impact sports. However, one study found that female volleyball players had significantly weaker PFM strength compared with a non-exercising CG and that low PFM strength was correlated with SUI.37

Another interesting difference found in the present study was between athletes in aquatic sports disciplines, in which none of the swimmers reported SUI, in contrast with those in water polo and synchronised swimming athletes. Other studies have a wide prevalence range in swimmers, varying from 6%16 to 84.6%.17 Although in aquatic sports the body is partially immersed in water and therefore ground reaction forces have no impact on the pelvic floor, some differences could exist between modalities in terms of abrupt and repetitive increases in IAP due to twisting and other specific sport body movements. However, the prevalence of UI may also be underestimated in water, as the leakage may be more difficult to perceive; that is, the leakage is hard to see, and there is no smell of urine.

The finding that more than half of the athletes reported UI only during sports practice, more frequently in the middle/at the end of training/competition sessions and that training volume was found to be associated with UI, indicates that both PFM strength and endurance may be required to stay dry. A decrease in the maximum voluntary contraction of the PFM after strenuous exercise has been observed in young nulliparous women with SUI.38 Thus, there is a need to develop valid methods of measuring the PFM function during different sports activities.

Effect on sports performance

More than one-third of the athletes considered UI to have implications on sports performance; however, less than 15% reported using strategies to hide the leakage, and significantly fewer athletes than controls had told a physician about the condition. Not telling healthcare professionals about UI is common.16 20 39 Because this behaviour seems to have persisted in the last two decades, there is a need to implement educational programs for healthcare providers, fitness instructors and coaches.15 Some studies indicates that presence of UI at a young age40 41 and strenuous activity during teen years42 may be a risk factor for later UI. Hence, interventions to prevent and treat the condition at an early stage is important.5 There is strong evidence that PFM training (PFMT) is an effective treatment for SUI in the general female population, and there is international consensus that PFMT should be a first-line treatment for SUI and MUI.43 To date, no randomised controlled trials (RCTs) of PFMT to prevent or treat SUI in female elite athletes have been carried out, and there is little knowledge about strenuous exercise as a risk factor for the development of UI.15 Three small case studies on PFMT in female athletes showed promising results.44–46 However, this observed effect has to be evaluated in RCTs of high methodological quality.

As our AG comprised athletes who were on the respective national teams for at least 1 year, some athletes might have quit competition due to UI. Consequently, prevalence rate of UI may be under-reported.

Strengths and limitations

The strengths of the present study are the large sample size, which included sports that had not been previously investigated; the high response rate, the comparison with an age-matched CG and the use of a reliable and validated questionnaire to assess UI.22 The limitations of this research are the use of questionnaires only, with no additional urodynamic or other clinical assessment methods to verify UI. In addition, the sample sizes for some of the sports groups were small. Although some sports were merged into larger groups,19 the sample sizes may still have been too small to register statistically significant differences between some groups.


The present study found a higher prevalence of UI among Portuguese elite athletes compared with age-matched controls and high-level sports practice was identified as an independent risk factor for UI. Besides high-level sports practice, constipation, familiar history of UI and history of urinary infections were also associated with high odds of UI.

What are the findings?

  • The prevalence of urinary incontinence (UI) among elite athletes is high, and the reported prevalence rates vary widely between sport disciplines.

  • High impact sports are associated with the highest prevalence of UI.

How might it impact on clinical practice in the future?

  • The current study compared the prevalence of UI between a young population of elite athletes and an age-matched control group.

  • The prevalence rates of UI in sports modalities that had not been previously reported were presented, for example, roller hockey, figure roller-skating and water polo.

  • Some modifiable risk factors were identified, which could help in the development of prevention and treatment strategies for this specific group.


We thank Daniela Simões, Department of Physiotherapy, CESPU, North Polytechnic Institute of Health, for valuable help with statistical analysis.


View Abstract


  • Contributors AC was involved in planning the study, reviewed and analysed the literature, wrote the paper, draft and approved the final version. RNJ critically revised the article and approved the final version. KB was involved in planning the study, critically revised the article and approved the final version.

  • Competing interests None declared.

  • Ethics approval Ethics Committee of the Faculty of Sports of the University of Porto (CEFADE 17.2014).

  • Provenance and peer review Not commissioned; externally peer reviewed.

  • Correction notice This paper has been amended since it was published Online First. Owing to a scripting error, some of the publisher names in the references were replaced with 'BMJ Publishing Group'. This only affected the full text version, not the PDF. We have since corrected these errors and the correct publishers have been inserted into the references.

Request permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.