International Urogynecology Journal

, Volume 23, Issue 9, pp 1231–1237

High prevalence of pelvic floor muscle dysfunction in hospitalized elderly women with urinary incontinence


    • Department of Internal and Geriatric MedicineHochzirl Hospital
  • Stephan C. Jansen
    • Department of Internal and Geriatric Medicine and Palliative CareDeggendorf Hospital
  • Markus Kofler
    • Department of NeurologyHochzirl Hospital
  • Monika Lechleitner
    • Department of Internal and Geriatric MedicineHochzirl Hospital
Original Article

DOI: 10.1007/s00192-011-1628-4

Cite this article as:
Talasz, H., Jansen, S.C., Kofler, M. et al. Int Urogynecol J (2012) 23: 1231. doi:10.1007/s00192-011-1628-4


Introduction and hypothesis

The purpose of this study was to determine pelvic floor muscle (PFM) function in hospitalized elderly women with urinary incontinence (UI).


A cross-sectional study was performed using data of 704 patients, routinely collected by means of a clinical UI assessment.


Only 25.5% of the patients were able to perform normal PFM contractions (Oxford grading scale score ≥3); 74.5% were unable to contract their PFM or showed weak PFM activity without circular contraction or elevation of the vagina. Vulvovaginal mucosal dystrophy was noted in 84% of the patients. A significant positive correlation of PFM function was found to cognitive status (MMSE score), mobility (Tinetti performance score), and history of previous PFM training; a negative correlation of PFM function was found to patients’ age and vulvovaginal mucosal dystrophy, and no significant correlation to body mass index, parity, or history of hysterectomy.


Targeted clinical UI assessment including digital vaginal palpation should be performed in all incontinent elderly women in order to detect PFM dysfunction and to optimize therapeutic measures.


Elderly incontinent womenPelvic floor muscle functionUrinary incontinenceUrinary incontinence assessment


Epidemiologic and clinical studies reveal the highest prevalence of urinary incontinence (UI) in the elderly [1, 2]. The occurrence of UI is increased in the presence of comorbidities and age-related functional decline [1]. In older patients, UI is considered to be not a single diagnosis but rather a consequence of different underlying conditions and age-related factors, such as, cognitive decline, psychological changes, neurological disorders, lower genitourinary diseases, other comorbidities, and side effects of drugs [1, 35]. The use of a multidimensional clinically based UI assessment offers an adequate diagnostic tool for identification of conditions that might contribute to and aggravate incontinence and to facilitate proper UI management when a patient with symptoms of UI is presented initially [3, 6, 7]. Physical evaluation is generally considered as important part of UI assessment [3, 7] and should include visual inspection of the urogenital region, vaginal palpation, and testing pelvic floor muscle (PFM) function [7].

According to the Pelvic Floor Clinical Assessment Group of the International Continence Society (ICS) [8], normal function of the PFM is defined as “the ability to perform a voluntary muscle contraction, resulting in a circular closing of vagina, urethra, and anus and in a cranio–ventral movement of the perineum and upward movement of the pelvic organs.” An involuntary contraction is defined as “muscular contraction before and during a rise in intra-abdominal pressure in order to prevent incontinence” [8]. Although different methods exist to verify PFM function, digital vaginal palpation is considered to be the gold standard in primary clinical assessment [8, 9]. For grading PFM contraction strength, the modified Oxford grading scale by Laycock [10] is one of the frequently used assessment tools [8].

PFM dysfunction is considered to be a fundamental pathophysiological mechanism in development of female UI [8]. Numerous studies about PFM function in younger women have been published [8, 11]; however, less information about the elderly is available, and these data have often been obtained by questionnaires only [1, 12]. Behavioral interventions such as PFM and bladder training are recommended as first-line conservative treatment in elderly incontinent women [1, 3, 5, 7]. Both interventions in particular require awareness of the pelvic floor and the ability to voluntarily perform a PFM contraction. The aims of PFM training are to improve PFM contraction strength and to learn timed PFM contraction before and during increase in intra-abdominal pressure [8, 1315]. The biological rationale for the use of voluntary PFM contractions as possible deferral technique to inhibit urgency to void during bladder training is based on Godec’s observation that a detrusor muscle contraction can be inhibited by PFM contractions [16, 17].

Since clinical information about PFM function in the elderly population is scarce, we performed a cross-sectional study using data of 704 hospitalized incontinent patients, routinely collected by means of a clinical UI assessment, in order to determine PFM function and to identify possible correlations between PFM dysfunction and age-related conditions, common medical symptoms, and subjective burden of disease.

Patients and methods

The study was performed in a general hospital in Austria with departments of internal and geriatric medicine and neurolgy. Between November 2000 and June 2009, 1,100 women older than 65 years routinely underwent basic clinical UI assessment according to current geriatric guidelines [3, 5, 7, 18]. Patients were admitted to the hospital for cardiovascular, inflammatory, metabolic, or neurological disease, as well as for rehabilitative treatment of arthritis and remobilization after fractures or joint replacement. Most patients showed a number of comorbidities and polypharmacy. UI assessment was performed within 1 week after admission in patients who complained about UI with high associated subjective burden of disease or in patients with symptoms of UI based on the nurses’ documentation. Assessment data were routinely documented in the patient chart. After approval by the ethics committee of Innsbruck Medical University, Austria, the data were retrospectively entered off-line into a spreadsheet (Microsoft Excel) for statistical analysis, when all of the following items of UI assessment were properly documented: PFM functional status (evaluated by digital vaginal palpation and graded according to the modified Oxford grading scale), age, body mass index, subjective burden of UI, predominant UI symptoms, number of childbirths, history of hysterectomy, history of previous PFM training, vulvovaginal mucosal status, Tinetti score, and Mini-Mental State Examination (MMSE) score. Complete data sets of 704 patients were available for statistical analysis (Table 1).
Table 1

Patients’ characteristics (n = 704)




Mean ± SD


Age (years)a



78.2 (±6.3)


Body mass index (kg/m2)a



26.9 (±5.4)


MMSE scorea



26.5 (±3.3)


Tinetti scorea



17.2 (±3.3)


Number of childbirths



2.5 (±2.1)


 Nulliparous women




 1–3 childbirths




 >3 childbirths




Patients reporting previous PFM training




Patients reporting history of hysterectomy




Subjective burden of diseaseb




No burden of UI




Little burden of UI




High burden of UI




Predominant UI symptomsc




 Predominant urgeny UI




 Predominant stress UI




 Symptoms of mixed UI




 Postvoid residual volume >100 ml




aAge, body mass index, MMSE, Tinetti scores, and number of childbirths are expressed as mean ± SD

bSubjective burden of UI was evaluated by means of a three-point scale (no, little, high negative impact of UI on quality of life)

cPredominant UI symptoms were evaluated by means of questionnaires and patients’ subjective comments

PFM function testing

PFM function testing was carried out by the same person in all patients—a female physician experienced in pelvic floor assessment and re-education. After positioning the index finger into the distal part of the vagina, the investigator asked patients to contract the PFM, to lift inward, and to squeeze around the finger. When necessary, instructions were repeated up to three times, and the women were also asked to squeeze the PFM as they would when feeling acute urinary urgency. Constriction and elevation of the vaginal wall and strength of PFM contraction were graded according to the modified Oxford grading scale by Laycock [10] (Table 2). To determine involuntary PFM contractions, women were asked to cough without changing position. Each circular constriction of PFM before or during coughing was defined as involuntary PFM contraction, regardless of additional cranial movement of the vagina or not.
Table 2

Modified Oxford grading scale


No discernible PFM contraction


Minor muscle “flicker”


Weak PFM contraction without any circular contraction and elevation of the vagina


Moderate PFM contraction with a circular contraction and elevation of the vagina


Good PFM contraction


Strong PFM contraction

Graduation of maximum PFM contraction, according to the six point Oxford scale modified by Laycock [10]

Vulvovaginal mucosal status

Vulvovaginal mucosal status was subjectively classified by the investigator (grade 0 = normal mucosal status; grade 1 = mucosal dystrophy with paleness, dryness, or redness; grade 2 = mucosal dystrophy with additional urethral caruncles, ulcerations, hemorrhage, and/or abnormal vaginal discharge). Patients with advanced pathologies were referred for further gynecological evaluation [7].

Body mass index

Body mass index was calculated from the women’s body weight and height (kilograms per square meter).

Subjective burden of UI

Subjective burden of UI was estimated by asking the women to rate the subjective negative impact of UI on quality of life (0 = no burden; 1 = little burden; 2 = high burden).

Predominant UI symptoms

Predominant UI symptoms were evaluated by means of a non-validated questionnaire for screening UI symptoms (Table 3) and of focused personal history obtained by the examiner.
Table 3

Questionnaire for screening UI symptoms

Postvoid residual volume

More than 100 ml—determined by ultrasonography—was estimated to be relevant and noted in the patient’s chart.

Number of childbirths

Number of childbirths was noted as given by the patients. No differentiation was made between vaginal deliveries or abdominal Cesarean sections.

History of hysterectomy

History of hysterectomy was based on patients’ response only. No differentiation was made between the types of surgery.

History of previous PFM training

History of previous PFM training was based on patients’ comments. Women were asked whether they have ever performed PFM exercises. Possible answers were “yes” or “no.”

Tinetti performance test

The Tinetti performance test [19] was used to assess patients’ mobility, balance, and gait and to predict falls. With a maximum score of 16 points for evaluating balance, 12 points for evaluating gait, and a cut-off score of 20 points; this test is recommended as gold standard for testing mobility in the elderly.

Mini-mental state examination

Mini-mental state examination (MMSE) [20] was used to evaluate cognitive function. This brief screening test with a maximum score of 30 points and a cut-off score of 24 points is widely used for detecting presence and global severity of cognitive impairment in the elderly.

Statistical analysis

Statistical analysis was performed by use of the SPSS 18.0 statistic program. Descriptive statistics was performed on the entire group of patients. Oxford grading scores, age, body mass index, MMSE scores, Tinetti scores, and number of childbirths are expressed as mean ± SD. Correlation coefficients were calculated by Pearson testing. An analysis of subgroups determined by predominant symptoms of UI was performed by Kolmogorov–Smirnov testing to confirm normal data distribution and applying unpaired Student’s t test. A p value <0.05 was considered statistically significant.


Terminology relating to PFM function and dysfunction used in this article conforms to the definitions recommended by the International Urogynecological Association and the International Continence Society [8, 21].


Patient characteristics are shown in Table 1. Mean age was 78.2 ± 6.3 years (range, 65–96 years). Among the 704 elderly women, 280 presented with predominant urge UI symptoms, 269 with mixed stress/urge UI symptoms, and 55 with predominant stress UI symptoms. In 100 patients, more than 100 ml postvoid residual volume was registered by ultrasound.

Only 25.5% (n = 180) of the patients were able to perform normal or strong PFM contractions with noticeable contraction and elevation of the vagina (Oxford Grades 3 and 4); 39.2% (n = 276) showed only weak PFM activity without circular contraction and elevation of the vagina (Oxford Grades 1 and 2), and 35.2% (n = 248) were not aware of their pelvic floor and could not voluntarily influence PFM (Oxford Grade 0). None of the patients was able to perform a strong PFM contraction (Oxford Grade 5) (Fig. 1). Before and during coughing, palpable involuntary PFM contraction was noted in only five patients. Most patients presented with considerable downward displacement of the pelvic floor during coughing.
Fig. 1

Ability to contract the pelvic floor muscles (Oxford Grades 0–5) in the whole group—expressed in percent of the whole group (n = 704) of incontinent female elderly patients

Only 16% of the women (n = 111) showed normal vulvovaginal mucosal status on visual inspection, 17% (n = 116) showed moderate mucosal dystrophy with paleness, dryness, or redness, and 67% (n = 475) presented pronounced mucosal alteration with ulceration, hemorrhage, urethral caruncles, or abnormal vaginal discharge.

Significant positive correlation of PFM function was found to higher MMSE score (r = 0.177; p < 0.001), higher Tinetti score (r = 0.110; p < 0.05) and history of PFM training (r = 0.155; p < 0.001). Negative correlation was found between PFM function and patients’ age (r = −0.139; p < 0.001) and extent of mucosal dystrophy (r = −0.090; p < 0.05). No significant correlation was found between PFM function and body mass index (r = 0.067; p = 0.081, parity (r = −0.041; p = 0.281) and history of hysterectomy (r = 0.001; p = 0.994).

Mean Oxford grading score was significantly lower (p < 0.05) in patients with increased residual volume (mean Oxford score, 1.2 ± 1.3 SD) as compared to those without residual volume (mean Oxford score, 1.5 ± 1.3 SD). No significant difference (p > 0.1) was noted between patients with predominant symptoms of stress UI (mean Oxford score, 1.5 ± 1.4 SD) and those with predominant symptoms of urge UI (mean Oxford score 1.7 ± 1.4 SD).

Notably, subjective burden did not correlate to PFM function (r = 0.017; p = 0.685), nor to mucosal status (r = −0.014; p = 0.721), nor to MMSE (r = −0.069; p = 0.084), while the latter showed significant inverse correlation to mucosal status (r = −0.094; p = 0.019).


The main findings of this study in a large number of hospitalized elderly incontinent women were (1) high percentage of inability to contract PFM voluntarily or involuntarily when requested during clinical UI assessment; (2) positive correlation of PFM function to cognitive and mobility status; (3) negative correlation of PFM function to age and extent of vulvovaginal mucosal dystrophy; (4) no correlations of PFM function to BMI, number of childbirths, and history of hysterectomy; and (5) poor awareness of the pelvic region in the investigated women.

The study reflects daily clinical practice in a geriatric setting. UI assessment was routinely performed according to an exactly defined diagnostic protocol. The primary aim of the assessment was not to define the type of UI, but rather to identify possible underlying causes and aggravating factors for UI, treatable in a geriatric setting or requiring referral to specialized gynecological, urological, or neurological departments for further investigation or treatment. Some of the applied assessment tools are not validated, but proved to be very simple to apply. As many of the investigated patients showed a large number of comorbidities and age-related functional decline, brief questionnaires—easy to understand and respond to—were used to evaluate the patients’ predominant UI symptoms and to estimate their subjective burden of disease. Main focus of the UI assessment was on targeted physical examination and functional testing.

PFM function was assessed according to Laycock et al. [10], an easily applied method without negative effects for the women, yet adequate for planning treatment strategies [9, 22]. Even though not perfectly suitable for research purposes due to its subjective quantification, intra- and intertester reliability of the modified Oxford grading are reportedly acceptable [22]. We also tested involuntary PFM contraction before and during coughing by digital vaginal palpation, nonetheless being aware of its more variable performance, more difficult assessment, and known low interobserver reliability [23]. However, as also discussed by Slieker ten Hove et al. [23], the results of involuntary PFM testing may serve to contribute to planning functional PFM training programs.

High prevalence of PFM dysfunction in the investigated patients is remarkable and has not been reported to date. Correlation of PFM function to cognition and mobility status, but inverse correlation to age, indicate that age-related changes may play an important role in the pathogenesis of PFM dysfunction. However, definition as to what is normal in older patients is lacking [24], and only few reliable investigations are available on the influence of age on PFM function [25] and on its pathophysiological mechanisms related to UI. Previous questionnaire-based studies could not demonstrate an association between UI and PFM dysfunction in older women after controlling for obesity, parity, menopause, and hormone use [12]. However, based on our clinical experience and on present results, history-based UI evaluation and questionnaires are not ideal instruments for PFM evaluation in older patients with a high prevalence of morphological changes, comorbidities and age-related functional decline. One of the reasons may be reduced body awareness exhibited by these patients. Although not formally assessed, many women were completely unaware of their pelvic region. They could not influence or contract PFM voluntarily or involuntarily. Pelvic floor and pelvic organs showed unhindered caudal displacement during every cough. We did not, however, further evaluate these results because of the known low intertester reliability [23].

In the present study, predominant UI symptoms reported by the patients rather than the type of UI were correlated to PFM function, yielding the poorest performance in those patients with increased postvoid residual volume. It would be interesting to confirm these results in women with urodynamically tested UI.

One would expect patients’ burden to be high when PFM function is low and mucosal status is poor, but this was not the case. In fact, mucosal dystrophy was inversely correlated to MMSE, consistent with reduction of body awareness due to cognitive impairment. Altogether, these results indicate that subjective burden is not a good indicator for PFM assessment in the elderly.

The influence of postmenopausal estrogen deficiency on vulvovaginal and urethral mucosa and its aggravating role on UI is well-known [26]. Atrophic vaginitis itself does not cause UI [5] but is considered to be an aggravating factor for UI especially in older postmenopausal women. Leibovitz et al. [27] found vulvovaginal dystrophy in 30% of elderly female residents of nursing homes. The higher prevalence of pathologies in our patients—84% showed mucosal dystrophy—may be attributed to the selection of patients suffering from UI and hospitalized due to acute illness or functional decline. Notably, urethral caruncles were present in a high number of our patients. These soft, friable, red outgrowths along the posterior edge of the urethra are scarcely addressed in geriatric literature or in guidelines for screening and treatment of UI in older women. Although little is known about the clinical relevance of these caruncles, their scarce appearance in literature may indicate a trend towards neglecting physical investigation in favor of symptom- and questionnaire-based UI evaluation in these patients [4, 12].

The present findings confirm previous data [11], indicating that number of childbirths does not further reduce PFM function in elderly incontinent women, and that pregnancy and vaginal delivery are significant risk factors for UI, but become less important with increasing age [5]. Although body weight concerns are controversial in geriatric patients, greater weight is considered to offer some protective benefits. This may be associated with higher bone mineral density and with relatively greater isometric muscle strength [28], which could be a reason why higher body mass index did not correlate to PFM dysfunction in the present study.

Although reported PFM training in the past correlated positively to PFM function, in agreement with Sherburn et al. [29], a high percentage of the older women did not sufficiently benefit. We did not evaluate the exact duration and quality of previous PFM training, but considering local practice, women had almost certainly trained in groups supervised by a physiotherapist. Digital vaginal PFM function testing is not routinely performed before commencing PFM exercises. In our opinion, only women with normal or perhaps mildly weak PFM contraction are likely to be able to carry out PFM training instructions and thus able to train muscles and correct coughing patterns by themselves. However, even a high proportion of younger women were shown to be unable to control their PFM muscles properly [11], rendering it even more unlikely that elderly women could achieve efficient PFM training by themselves. Therefore, especially in older women with poor PFM function, routine digital vaginal palpation is mandatory to confirm correct PFM contraction before initiating and during PFM training programs [9, 29].

Limitations of the present study are its uncontrolled design, the highly selective population, and the use of several non-validated assessment tools. Additionally, as medical history was predominantly based on patients’ comments only, some results pertaining to personal histories could have been biased by inaccuracies due to certain patient’s impaired cognitive status.


Targeted clinical UI assessment focusing mainly on physical examination and functional testing revealed a high prevalence of PFM dysfunction and reduced awareness of the pelvic region in hospitalized incontinent elderly women. These findings should be taken into account when advising behavioral interventions such as PFM and bladder training in older patients with UI.


The authors would like to acknowledge Mira Ljubinkovic, Hochzirl, Austria, for consistently collecting data and assisting the project during the whole study period, and to express their gratitude to Ellen Quirbach, Hochzirl, Austria, for providing editorial help with the manuscript.

Conflicts of interest


Copyright information

© The International Urogynecological Association 2011