Regression Model for the Prediction of Risk of Pelvic Floor Muscle Weakness Among Older Adults

Pelvic floor muscle (PFM) weakness is a common presentation among older adults, assessment of which is complicated by privacy issues, availability of perineometer, and lack of trained therapists. Several factors seem to influence PFM strength. Hence, this study developed an alternate method using a regression model to determine PFM strength among institutionalized older women using 27 independent variables. A total of 102 participants were assessed for their PFM strength using peritron perineometer. Age, years of menopause, pelvic surgical procedures, strength of hip muscles, and degenerative disc disorders were found to have a significant odds ratio and were included in the regression model (r2 = 0.484). It may be used as a simple and easy-to-administer method for the evaluation of PFM weakness.

dwelling older women. The penalties being more on institutionalized older women could be due to poor nutritional status, depression, lack of communication, embarrassment, fear and trust issues with the caretakers, and reduced physical and leisure activities. 7,8 Hence, pelvic floor muscle strength (PFMS) is an important determinant of geriatric women's health; the evaluation of the PFMs plays a significant role in the management of its dysfunction. Perineometry is the standard and most reliable technique to evaluate PFMS and endurance. 9 Nevertheless, it might not be possible to use perineometer in the older care settings because of the cost factor, lack of availability, 10 need for trained personnel, medicolegal issues, and privacy concerns, so there is a need to develop a substitute method to predict the PFM strength.

METHODS
This case-control study was carried out between March 2016 and February 2017. The study protocol was approved by the institutional ethics committee, Kasturba Hospital, Manipal, India (IEC 111/2016). Participants older than 60 years were recruited from institutions for older adults with similar sociodemographic background located in Udupi and Mangalore taluk, India. Residents were screened and excluded if they had stoma in situ, congenital anorectal malformations, pelvic inflammatory disease, neurological dysfunction, a score of less than 24 in Mini-Mental State Examination, pelvic or abdominal surgery in the last 6 months, and current low back pain. All the participants were explained about the study procedure and signed written informed consent.
The dependent variable was the maximum voluntary contraction of PFM, assessed using digital Peritron 9300 Perineometer (Peritron; Cardio-Design, Australia) with a vaginal probe (molded 23 × 17.5 × 5 cm) covered with latex sleeves. Participants were instructed not to hold their breath during the contractions to avoid any influence from the abdominal muscles. Each contraction was held for 5 to 10 seconds with a rest period of 5 seconds in between the 2 consecutive contractions. The average reading in centimeter of H 2 O of the first 3 contractions was recorded. 11 The participants who had PFMS of 18 cm of H 2 O and less with urinary incontinence were considered as cases and more than 18 with no urinary leakage were categorized as controls. This cutoff was decided on the basis of the research by Gameiro et al 12 and Chevalier et al 13 and also by expert opinion. The correct recruitment of PFMs was confirmed by therapist's finger insertion method prior to assessment using perineometer.
The independent variables assessed were age, parity (number of births), number of miscarriage, history of pelvic surgery, years postmenopause, body mass index, core muscle strength, peripheral muscle strength (6 muscles), functional comorbidity index (18 comorbidities), and functional mobility. Weight and height were measured using a calibrated weighing scale and an inch tape, respectively, and body mass index (kg/m 2 ) was calculated. Core muscle strength was assessed as the maximum time that a prone plank position could be sustained (duration in seconds).
Muscle strength of hip adductors, hip external rotators, and hip extensors was assessed using a handheld isometric dynamometer. 14,15 For the assessment of hip adductors, the subject was positioned in the side lying on the lower limb to be assessed. The leg to be evaluated was placed with the knee extended and the hip in neutral rotation; the contralateral limb was stabilized by the examiner at 90° of hip and knee flexion supported by pillows. The participant had to perform a maximum isometric contraction of adductors, with resistance to movement applied just superior to the medial malleolus.
Evaluation of lateral hip rotators was carried out with the participant in high sitting with hip and knee flexed to 90°. The participant held his or her arms against the body. With the hip in slight lateral rotation and the medial malleolus aligned with the midline of the body, the participant performed a maximal isometric contraction of external rotators, with resistance to movement applied just superior to the medial malleolus. The strength of hip extensors was evaluated in prone lying with the knee flexed to 90° in hip slight lateral rotation. The participant had to perform an isometric contraction of hip extensors, with resistance to movement applied to the distal thigh posteriorly.
Functional comorbidity index is a diagnosis-based index, which will rule out the age-related comorbidities of the participants. 16 It consists of assessment of 18 comorbidities such as arthritis, osteoporosis, asthma, chronic obstructive pulmonary disease, angina, heart failure, myocardial infarction, neurological disease, stroke, peripheral vascular disease, diabetes, gastrointestinal disease, depression, anxiety, visual impairment, hearing impairment, disc disease, and obesity. Each item scores 1 for each diagnosis and the sum of all the 18 items is the final score. Functional mobility was assessed using the Timed Up and Go test. 17 It calculates the time in seconds and milliseconds that participant takes to rise from a chair, walk 3 m, turn around, walk back to the chair, and sit down. During this test, the participants were made to wear their daily footwear, were allowed to use their walking aid, and were instructed to walk at their normal walking speed.
All statistical analysis was performed with SPSS version 15.0. Descriptive statistics were used to summarize demographic data. Binary logistic regression with stepwise backward technique was used to determine whether the independent variables were associated with pelvic floor weakness, to determine the odds ratio, and to develop the prediction model. The level of significance was set at P ≤ .05.

RESULTS
One hundred seventy-three institutionalized older women were screened and 71 were excluded, of whom 24 did not give written informed consent, 22 had low backache at the time of the study, 14 had a score of less than 24 in Mini-Mental State Examination, and 11 had undergone abdominal surgery within the last 6 months (Figure). So 102 participants (46 cases and 56 controls) were included in the study and completed the assessment. Table 1 depicts the demographic characteristics of participants among cases and controls. The mean age and mean years postmenopause showed a statistically significant difference between cases and controls, where the mean age was 3.60 ± 1.19 years and mean years postmenopause were 3.54 ± 1.12 years more among cases when compared with controls. Table 1 represents the muscle strength profile of participants in pounds. The mean isometric peripheral muscle strength was less among cases when compared with controls. The mean PFMS (in centimeter of H 2 O ) was shown to have a difference of 10.46 ± 1.14 between cases and controls, whereas mean prone plank time was 1.23 ± 0.88 seconds less among cases when compared with controls. Only 17 participants among the controls and 10 among the cases could perform prone planks and were scored more than 0. A statistically significant difference was observed in right and left hip adductors, left hip external rotators, left hip extensors, and PFMS between cases and controls. Table 2 represents the percentage of comorbidities among participants in cases and controls. Only 2.2% of the participants among cases and 3.6% among controls did not report any diagnosed comorbidity. Around 63.9% of the participants in cases and 66.0% among controls had multimorbidities. Diagnosed diabetes mellitus was reported by the maximum number of participants with 67.5% and 62.5%, respectively, among cases and controls. Arthritis was the next common problem among cases and controls with 65.2% and 62.5%, respectively.
The adjusted and unadjusted odds ratio with the exposed factors is shown in Table 3. The variables that showed statistically significant odds ratio were age, years postmenopause, pelvic surgery, degenerative disc disorders, and strengths of right and left hip adductors, left hip external rotators, and left hip extensor muscles; hence, those were taken for developing the model.  Previous studies have reported that with advancing age, the PFM strength reduces persistently. 3,7,18 This age-related decline could be hormonal, neural, or the architectural changes in the muscles. Estrogen deficiency accounts to be the major contributing factor for the decline in PFM strength. 19 As one ages, the actin-myosin sliding mechanism is disturbed during muscle contraction due to the deficiency of β-estrogen receptors in type 1 muscle fibers. In addition, it corresponds to a vicious cycle in which the symptom (subdued PFM strength) and the consequences of PFD such as immobility, embarrassment, and social anxiety 20 occur collectively. Other causes such as agerelated neuropathic changes, sarcopenia of the muscle fibers, and neuromuscular changes with age could also be attributed to reduced muscle strength.
In the studies done by Marianna et al 5 and Heidi and Alejandra, 21 it is distinct that PFM strength is associated with peripheral (hip adductors, hip extensors, and hip external rotators) and core muscle strength. It has been reported that these muscles are in close proximity to the PFM, thereby causing a synergic contraction with strenuous activities. 22 The current study also reports a positive correlation between peripheral muscle strength and PFMS.
A positive relationship was seen between PFM and degenerative disc disorders. Several studies acknowledge the association between lower lumbar and sacroiliac joint dysfunction with PFM weakness and urinary incontinence. 23 Nicholas et al 24 quoted that in older women, denervation of pelvic sphincter muscles takes place due to the degeneration of the lumbosacral canal; age-related decline in the bone and muscle architecture also adds to the problem, which influences PFMS. 25 The study has potential significance in clinical scenarios where there are high chances of violation of patient privacy. The model developed contributes toward addressing the issues related to PFM strength assessment such as high cost, lack of availability of equipment, need for trained personnel, medicolegal issues, and privacy concerns. In a typical clinical scenario, a clinician can use the model and estimate PFM weakness by putting the values of 4 continuous variables (age, hip adductor, external rotator, and extensor muscle strength) and 2 categorical variables (pelvic surgery and disc disorders; value 1 for present and 0 for absent). Hence, it may be used as a simple and easy-to-administer method for determining PFM weakness. The majority of the participants complained of shoulder discomfort during the prone plank test and thus could not complete the test procedures. A few factors that might influence PFMS, namely, duration of labor, history of incontinence after childbirth, physical activity levels of the participants, and PFM tone were not evaluated and are some of the limitations of this study. Another limitation could be the recall bias, as information of many of the independent variables was collected from older individuals themselves, and the researcher bias, as all the assessments were carried out only by 1 investigator. Future research may focus on validating the model, the development of a similar regression model for community-dwelling elderly, and address the limitations of this study by incorporating other independent variables.

CONCLUSION
A regression model for determining PFM weakness among institutionalized older women has been developed, which may be used as a simple and easy-to-administer method. The variables in the model are age, years postmenopause, pelvic surgery, disc disorders, and hip muscle strength.