The Persistency Index: a novel screening tool for identifying myofascial pelvic floor dysfunction in patients seeking care for lower urinary tract symptoms

Abstract

Background:

Patients with myofascial pelvic floor dysfunction often present with lower urinary tract symptoms (LUTS) such as urinary frequency, urgency, and bladder pressure. Often confused with other lower urinary tract disorders, this constellation of symptoms, recently termed myofascial frequency syndrome (MFS), is distinct from other LUTS and optimally responds to pelvic floor physical therapy. A detailed pelvic floor myofascial exam performed by a skilled provider is currently the only method to identify MFS. Despite a high impact on quality of life, low awareness of this condition combined with no objective diagnostic testing leads to the frequent mis- or under-diagnosis of MFS.

Objective:

To develop a screening measure to identify patients with MFS (bothersome LUTS secondary to myofascial pelvic floor dysfunction) from patient-reported symptoms.

Study Design:

A homogenous population of patients with MFS was identified by provider diagnosis from a tertiary urology practice and verified by standardized pelvic floor myofascial exam and pelvic floor electromyography. Least Angle Shrinkage and Selection Operator was employed to identify candidate features from the OAB-q, fGUPI, and PFDI questionnaires predictive of MFS in a pooled population also containing subjects with overactive bladder (OAB, n=42), interstitial cystitis/bladder pain syndrome (IC/BPS, n=51), and asymptomatic controls (n = 54) (Derivation cohort). A simple, summated score of the most discriminatory questions using the original scaling of the PFDI5 (0 to 4) and GUPI5 (0 to 5) and modified scaling of fGUPI2B (0 or 3) had an Area Under the Curve of 0.75. As MFS was more prevalent in younger subjects, inclusion of an age penalty (3 points added if under age 50) improved the Area Under the Curve to 0.8. This score, defined as the Persistency Index (PI) (possible score 0–15), performed well in the evaluation of a Validation cohort of 719 patients with various LUTS, including OAB (n=285), IC/BPS (n=53), MFS (n=111), minimally bothered controls (n=209) and unknown diagnoses (n=61), exhibiting a similar Area Under the Curve of 0.79. Youden’s Index was used to identify the optimal cut point PI score for maximizing sensitivity and specificity.

Results:

The severity (PFDI5) and persistent nature (fGUPI5) of the sensation of incomplete bladder emptying as well as dyspareunia (fGUPI2b) were the most discriminatory characteristics of the MFS group, which were combined with age to create the PI. A PI score greater than or equal to 7 accurately identified patients with MFS from an unselected population of individuals with LUTS with 80% sensitivity and 61% specificity (Table 1). Combination of the PI with the previously defined bladder pain composite index and urge incontinence composite index separated a population of women seeking care for LUTS into groups consistent with OAB, IC/BPS, and MFS phenotypes with an overall diagnostic accuracy of 82%.

Conclusions:

We propose a novel screening method for patients presenting with LUTS to identify patients with myofascial frequency syndrome. As telemedicine becomes more common, this index provides a way of screening for MFS and initiating pelvic floor physical therapy even prior to a confirmatory pelvic exam.

Introduction

Lower urinary tract symptoms (LUTS), such as urinary frequency, urgency, and bladder discomfort, are common and affect most individuals during their life.¹ Despite this high prevalence, diagnostic tools are sparse and rely heavily on clinician judgement. This lack of objective criteria leads to inadequately defined diagnostic schema, which classify patients based on symptom clusters such as overactive bladder (OAB) and interstitial cystitis/bladder pain syndrome (IC/BPS) that do not specify underlying pathophysiology. Such contemporary allocation strategies for LUTS often do not effectively differentiate subsets of patients that require different treatment approaches. As a result, most patients with LUTS fail to respond to initial treatment approaches and are frequently lost to care.² Better tools to recognize more homogeneous LUTS phenotypes would dramatically improve both the recognition and management of patients with LUTS by primary care and specialist providers

We recently described a novel LUTS diagnosis termed Myofascial Frequency Syndrome (MFS).³ This prevalent condition⁴ is characterized by the dominant sensation of incomplete bladder emptying (without true urinary retention) in concert with urinary frequency and urgency prompted more by discomfort than fear of incontinence; a symptom complex we dubbed persistency. Affected subjects display myofascial pelvic floor dysfunction (hypertonicity with or without pelvic floor trigger points) on examination and pelvic floor electromyography/ urodynamic testing and improve with pelvic floor-directed therapies such as myofascial release-based physical therapy or biofeedback.

Despite being exceedingly common,⁴ due to the vague mix of uncomfortable bladder sensations and urinary complaints, MFS is often mistaken for other symptom complexes, such as IC/BPS, OAB, pelvic organ prolapse (POP), or even urinary tract infection (UTI).^4–6 No specific ICD-10 code specifies this diagnosis, making epidemiologic assessment and surveillance difficult. Given the non-specific constellation of symptoms, MFS can be challenging for providers to recognize, as this diagnosis requires a detailed pelvic exam, sufficient understanding of the global symptom complex, and ruling out confounding conditions.⁶ Given these challenges, which are compounded by increasing reliance on telemedicine and limited visit times, improved tools are required to assist providers in recognizing MFS. A symptoms-based measure associated with the MFS phenotype could greatly assist providers in suspecting a myofascial origin to a patient’s urinary symptoms, thus focusing the physical assessment and possible treatments.

As initial characterization of MFS revealed a consistent pattern of associated symptoms across multiple independent cohorts,³ we hypothesized it would be feasible to construct a symptoms-based measure to identify MFS subjects. With a patient-derived indicator identifying these patients, earlier interventions with appropriate treatment (i.e., pelvic floor physical therapy) can be achieved, even when providers are less facile at making the diagnosis.

We therefore sought to develop a screening index to identify patients with possible MFS for use in telemedicine and by providers unskilled with a discriminate pelvic exam. By comparing a population of MFS subjects with patients unequivocally categorized as OAB or IC/BPS, we developed a novel measure, which we term the Persistency Index (PI), to screen for this underrecognized type of LUTS. We further propose a modified diagnostic nomogram incorporating this measure that is capable of differentiating this cohort of patients from classical OAB and IC/BPS.

Materials and Methods

Study Inclusion

After local Institutional Review Board (IRB#00040261) approval, female subjects presenting for care in a specialized urogynecology clinic were included (Fig. 1). At initial consultation, all subjects were administered three validated, written questionnaires: the 1) female Genitourinary Pain Index (fGUPI)⁷, 2) Overactive Bladder Questionnaire (OABq)⁸, and 3) Pelvic Floor Distress Inventory (PFDI-20).⁹ The fGUPI, which contains subscales assessing pain, urinary symptoms, and quality of life, measures the nature and severity of genitourinary pain.⁷ In the OABq, symptom questions (1–8) assess both continent and incontinent OAB symptoms.⁸ The PFDI-20 measures a range of pelvic floor symptoms and includes three domains measuring: 1) urinary [Urinary Distress Inventory (UDI-6)], 2) defecatory [Colorectal-Anal Distress Inventory (CRADI-8)], and 3) prolapse [Pelvic Organ Prolapse Distress Inventory (POPDI-6)] symptoms.⁹

Figure 1. Study Design.

We derived a screening measure associated with myofascial frequency syndrome (MFS), dubbed the Persistency Index. This measure was comprised of the most salient features of MFS identified by comparison of a group of subjects with LUTS, abnormal pelvic floor findings on exam and EMG findings of a tonically contracted pelvic floor to asymptomatic subjects and patients with OAB and IC/BPS (Derivation Cohort). We then determined the real-world performance of this measure in classifying subjects with MFS from a large population of individuals presenting to a urogynecology clinic (Validation Cohort).

Study Cohorts

A Derivation cohort of 215 subjects from a single-center urogynecology practice between January and December 2018 was employed to develop the PI. This population contained 68 subjects with urinary symptoms and demonstrable myofascial dysfunction on physical exam confirmed on pelvic floor electromyography. The myofascial dysfunction was secondarily verified as the source of their symptomatology by demonstrable improvements in their urinary symptoms after myofascial release-based, pelvic floor physical therapy (PFPT). This group was compiled with two other groups with LUTS: 1) 42 subjects diagnosed with OAB who endorsed significant urgency incontinence, displayed detrusor overactivity on urodynamic evaluation, and lacked bladder pain on exam or questionnaire assessment, and 2) 51 subjects with a clinical diagnosis of IC/BPS with significant bladder pain on physical examination, who endorsed pain with bladder filling on validated questionnaires, and who lacked any incontinence. An additional 54 subjects with asymptomatic questionnaire responses served as controls.

A Validation cohort of 719 subjects evaluated consecutively between January and December of 2019 served to assess the performance of candidate symptomatic measures in an unselected population. This cohort contained 111 subjects with diagnosed MFS (subjects with pelvic floor hypertonicity or trigger points on exam, presenting with urinary frequency, persistent bladder pressure, and a sensation of incomplete bladder emptying who bore a primary diagnosis of high-tone pelvic floor dysfunction), 285 subjects with OAB, 53 with IC/BPS, 209 subjects with minimal bother, and 61 subjects with multiple diagnoses. This cohort was additionally administered the O’Leary–Sant Indices, which include the Interstitial Cystitis Symptom and Problem Indices (ICSI/ICPI), that measures the severity and bother of urinary frequency, urgency, nocturia, and bladder pain.¹⁰

Derivation of the Persistency Index

We used the Derivation cohort to create a measure of urinary symptoms associated with a myofascial origin, termed the Persistency Index (PI). We used a Least Angle Shrinkage and Selection Operator (LASSO) to identify candidate predictors from the OABq symptom subdomain (8 questions), fGUPI (15 questions), and PFDI-20 (20 questions). Ten-fold cross-validation selected the appropriate shrinkage parameter (λ=0.01307) which identified 20 of the 41 variables as potential predictors (i.e., non-zero coefficients). From these, we selected 5 variables with positive standardized shrunken coefficients > 0.4 that had acceptable face validity for capturing the underlying clinical phenotype (e.g., that were in agreement with previously defined phenotypic characteristics³). These five variables were then used in a multivariable logistic regression model. The statistically significant, positive model coefficients (PFDI5, fGUPI5, GUPI2B) were retained resulting in a reduced model consisting of three predictors. As reduced age was significantly associated with myofascial dysfunction, age was added into the multivariable regression model. To increase usability of this model, we generated a simple summated score of these questions weighted approximately by their coefficients from the multivariable model. This score consisted of the original scaling of the PFDI5 (0 to 4) and GUPI5 (0 to 5) with a modified scaling of GUPI2B (0 or 3 if positive) and an age penalty for subjects <50 (3 points added if under age 50), resulting in a maximal possible score of 15.

Modification of Diagnostic Nomogram

The Persistency Index was added to the previous diagnostic nomogram¹¹ to classify OAB vs. IC/BPS to generate the phenotyping Comprehensive Lower Urinary Symptoms (p-CLUS) nomogram. In this nomogram, the PI, bladder pain composite index (BPCI), urge incontinence composite index (UICI), and Bother Index in combination served to classify patients as non-bothered subjects, IC/BPS, OAB, and MFS. Inaccuracy was calculated as the percentage of subjects incorrectly classified by the nomogram to a diagnosis different from the physician-assigned diagnosis.

Statistical analysis

Bivariate differences between groups were examined using Welch’s t and chi-square tests in the Derivation and Validation cohorts. All analyses were conducted in Stata version 16.1, Stata Corp (College Station, Texas, USA).

Results

MFS subjects are a distinct subset of subjects with LUTS.

The Derivation cohort (Table 1) consisted of patients presenting for care in a tertiary urogynecology practice in 2018. Subjects with confirmed MFS (n=68), in whom myofascial dysfunction was confirmed on physical exam and pelvic floor electromyography. MFS subjects were compared to a population of controls (n=54) and cases with OAB (n=42) or IC/BPS (n=51). MFS subjects exhibited similar levels of urinary frequency and urge as both OAB and IC/BPS subjects, but lacked urgency incontinence (OABq8, UDI-6 2) or pain with bladder filling (fGUPI 2c). This group specifically exhibited scores elevated over the other groups relating to a sensation of incomplete emptying (fGUPI5, POPDI-6 5), dyspareunia (fGUPI 2b), and defecatory dysfunction (CRADI-8 1,2).

Table 1:

Symptomatic features of the derivation cohort by diagnosis

Variable	Question	Scale	MUFS (n=68)	Asymptomatic control (n=54)	OAB (n=42)	IC or BPS (n=51)	Adjusted P value (MUFS vs asymptomatic control, OAB, or IC or BPS)
							Asymptomatic control	OAB	IC or BPS
Agea		0–95	43.62 (14.89)	54.22 (15.26)	64.03 (11.40)	49.05 (13.70)	<.001	<.001	.054
OAB-q2	An uncomfortable urge to urinate	1–6	3.72 (1.34)	1.69 (1.16)	4.08 (1.56)	3.65 (1.57)	<.001	.214	.784
OAB-q3	A sudden urge to urinate	1–6	2.29 (1.50)	1.39 (0.81)	4.43 (1.19)	2.08 (1.38)	<.001	<.001	.423
OAB-q4	Accidental loss of small amounts of urine	1–6	1.91 (1.47)	1.63 (0.78)	4.29 (1.38)	1.47 (0.67)	.206	<.001	.049
OAB-q5	Nighttime urination	1–6	2.87 (1.64)	2.09 (1.33)	4.00 (1.86)	2.98 (1.63)	.006	.001	.712
OAB-q6	Waking at night because you had to urinate	1–6	3.29 (1.65)	2.38 (1.44)	4.44 (1.55)	3.29 (1.59)	.002	<.001	1.000
OAB-q8	Urine loss associated with a strong desire to urinate	1–6	1.53 (1.13)	1.30 (0.57)	4.36 (1.39)	1.31 (0.65)	.168	<.001	.224
fGUPI1ab	Pain or discomfort at the entrance to the vagina	0–1	0.51	0.07	0.05	0.43	<.001	<.001	.372
fGUPI1bb	Pain or discomfort in the vagina	0–1	0.49	0.07	0.07	0.43	<.001	<.001	.563
fGUPI1cb	Pain or discomfort in the urethra	0–1	0.44	0.04	0.05	0.59	<.001	<.001	.114
fGUPI1db	Pain or discomfort below the waist or in the pubic or bladder area	0–1	0.42	0.08	0.10	0.90	<.001	<.001	<.001
fGUPI2ab	Pain or burning during urination	0–1	0.44	0.02	0.10	0.62	<.001	<.001	.055
fGUPI2ba,b	Pain or discomfort during or after sexual intercourse	0–1	0.52	0.09	0.12	0.47	<.001	<.001	.581
fGUPI2cb	Pain or discomfort as your bladder fills	0–1	0.09	0.00	0.00	0.92	.025	.048	<.001
fGUPI2db	Pain or discomfort relieved by voiding	0–1	0.23	0.06	0.05	0.73	.009	.011	<.001
fGUPI3	Frequency of pain or discomfort over the last week	0–5	3.00 (1.43)	0.37 (0.66)	0.68 (0.91)	3.00 (1.31)	<.001	<.001	1.000
fGUPI4	Number that best describes average pain or discomfort	0–10	3.85 (2.23)	0.39 (1.08)	1.43 (1.74)	5.91 (2.31)	<.001	<.001	<.001
fGUPI5a	Frequency of sensation of incomplete emptying	0–5	2.76 (1.25)	0.39 (0.60)	1.39 (1.18)	2.12 (1.48)	<.001	<.001	.011
fGUPI6	Need to urinate <2 h after last urinating	0–5	3.19 (1.21)	1.04 (1.06)	3.19 (1.40)	3.06 (1.52)	<.001	.998	.598
fGUPI7	Have your symptoms kept you from doing the kinds of things you would usually do?	0–3	1.49 (1.03)	0.04 (0.19)	1.45 (1.15)	1.51 (1.12)	<.001	.877	.902
fGUPI8	How much did you think about your symptoms?	0–3	2.50 (0.70)	0.37 (0.52)	2.43 (0.70)	2.51 (0.76)	<.001	.605	.942
fGUPI9	Satisfaction with current symptoms	0–6	4.59 (1.22)	1.59 (1.11)	4.62 (1.10)	5.04 (1.02)	<.001	.894	.035
POPDI-1	Pressure in the lower abdomen	0–4	1.83 (1.42)	0.34 (0.83)	0.85 (1.21)	1.80 (1.37)	<.001	<.001	.887
POPDI-2	Heaviness or dullness in the lower abdomen	0–4	1.46 (1.45)	0.23 (0.64)	0.62 (1.11)	1.84 (1.43)	<.001	.002	.156
POPDI-3	A bulge or something falling out that can be seen or felt in the vaginal area	0–4	0.36 (0.99)	0.20 (0.74)	0.29 (0.72)	0.16 (0.50)	.326	.690	.175
POPDI-4	A need to push on the vagina or around the rectum to have a complete bowel movement	0–4	0.42 (0.89)	0.31 (0.89)	0.68 (1.25)	0.45 (1.10)	.527	.218	.857
POPDI-5a	A feeling of incomplete bladder emptying	0–4	2.31 (1.27)	0.33 (0.75)	1.38 (1.41)	1.29 (1.40)	<.001	.001	<.001
POPDI-6	A need to push up in the vagina area to start or complete urination	0–4	0.16 (0.61)	0.02 (0.14)	0.07 (0.35)	0.12 (0.52)	.095	.399	.679
CRADI-8-1	A need to strain too hard to have a bowel movement	0–4	1.31 (1.43)	0.20 (0.63)	1.10 (1.34)	0.80 (1.28)	<.001	.437	.048
CRADI-8-2	A feeling that you have not completely emptied your bowels after a bowel movement	0–4	1.33 (1.37)	0.36 (0.83)	0.83 (1.10)	0.80 (1.20)	<.001	.050	.031
CRADI-8-3	Losing stool without control when stools are well formed	0–4	0.12 (0.56)	0.11 (0.60)	0.32 (0.85)	0.10 (0.50)	.938	.149	.831
CRADI-8-4	Losing stool without control when stool is loose or liquid	0–4	0.27 (0.77)	0.15 (0.76)	0.83 (1.25)	0.24 (0.74)	.392	.004	.813
CRADI-8-5	Losing gas from the rectum without control	0–4	0.71 (1.22)	0.31 (0.73)	0.98 (1.33)	0.31 (0.73)	.045	.279	.045
CRADI-8-6	Pain with passing stools	0–4	0.34 (0.91)	0.00 (0.00)	0.27 (0.71)	0.16 (0.64)	.007	.654	.217
CRADI-8-7	A strong sense of urgency and have to rush to the bathroom to have a bowel movement	0–4	0.70 (1.18)	0.30 (0.85)	0.81 (1.15)	0.41 (0.85)	.040	.640	.141
CRADI-8-8	Stool passes through the rectum and bulges outside during or after a bowel movement	0–4	0.24 (0.87)	0.13 (0.58)	0.22 (0.61)	0.04 (0.20)	.444	.919	.116
UDI-6-1	Bothered by frequent urination	0–4	2.16 (1.40)	0.74 (1.12)	2.95 (1.27)	2.46 (1.18)	<.001	.004	.225
UDI-6-2	Bothered by leakage related to feeling of urgency	0–4	0.41 (0.78)	0.31 (0.72)	3.12 (0.92)	0.69 (0.79)	.482	<.001	.060
UDI-6-3	Bothered by leakage related to physical activity, coughing, or sneezing	0–4	0.79 (1.13)	0.83 (1.16)	2.00 (1.47)	0.63 (1.00)	.863	<.001	.404
UDI-6-4	Bothered by small amounts of leakage (drops)	0–4	0.76 (1.27)	0.42 (0.84)	2.26 (1.48)	0.31 (0.73)	.090	<.001	.027
UDI-6-5	Bothered by difficulty emptying bladder	0–4	1.61 (1.37)	0.15 (0.53)	0.74 (1.27)	1.45 (1.39)	<.001	.001	.546
UDI-6-6	Bothered by pain or discomfort in the lower abdominal or genital area	0–4	1.65 (1.52)	0.19 (0.62)	0.51 (1.05)	2.10 (1.43)	<.001	<.001	.110

Data are presented as mean (standard deviation) for all interval variables, except for the binary variables, which document the population proportions scoring positive for each individual feature. Pairwise comparison of interval variables was performed using the Welch t test, and binary variables (designated with “^b”) were examined using the chi-square test.

BPS, bladder pain syndrome; CRADI-8, Colorectal-Anal Distress Inventory 8; fGUPI, Female Genitourinary Pain Index; IC, interstitial cystitis; ICPI, Interstitial Cystitis Problem Index; ICSI, Interstitial Cystitis Symptoms Index; MUFS, myofascial urinary frequency syndrome; NS, not significant; OAB, overactive bladder; OAB-q, Overactive Bladder Questionnaire; POPDI-6, Pelvic Organ Prolapse Distress Inventory 6; UDI-6, Urinary Distress Inventory 6.

Derivation of the Persistency Index.

We sought to generate a measure derived from individual, symptomatic questions that was associated with the MFS group using the Derivation cohort (Fig. 2). To develop the PI, Least Angle Shrinkage and Selection Operator (LASSO) was used to identify candidate predictors from the administered validated questionnaires (Supplemental Table 1). Of the 20 (of 41) potential predictor variables with non-zero coefficients, five had positive standardized shrunken coefficients > 0.4, each of which reflected the common symptoms previously described for this condition (thus conveying good face validity).³ These were then used in a multivariable logistic regression model to predict MFS. Use of the three statistically significant model coefficients (PFDI5, fGUPI5, GUPI2B) resulted in a reduced model with an in-sample area under the curve (AUC) of 0.75 (Supplemental Table 2).

Figure 2. The Persistency Index describes the unique features of subjects with LUTS of a myofascial origin.

A) Least Angle Shrinkage and Selection Operator (LASSO) with 10-fold cross-validation (λ=0.01307) identified 20 of the 41 variables from the OABq, fGUPI and PFDI as potential predictors. Five of these variables with a positive standardized shrunken coefficient > 0.4 as well as face validity for capturing the underlying clinical phenotype were used in a multivariable logistic regression model. Variables with statistically significant model coefficients (PFDI5, GUPI5, GUPI2B) were retained resulting in a reduced model consisting of these three predictors weighted based on these coefficients, which together comprise the Persistency Composite Index (PCI). Box and whisker plots compare the distribution of scores on the PFDI20 question 5 (B), fGUPI question 2b (C), fGUPI question 5 (D) and age (E) for the population of patients with myofascial frequency syndrome (MFS) in comparison to subjects with any other bothersome LUTS.

Age is a valuable predictor of MFS

As MFS subjects were noted to be significantly younger than OAB subjects with a similar severity of urinary symptoms, we evaluated the utility of including age as a predictor of MFS. Inclusion of age in the multivariable regression model significantly improved the AUC to 0.8 (Supplemental Table 2). To simplify this model for ease of clinical use, regression coefficients were scaled to generate a simple, summated score of these questions. A total possible score of 15 included the original scaling of the PFDI5 (0 to 4) and fGUPI5 (0 to 5), a modified scaling of GUPI2B (0 or 3 if positive), and an age factor for subjects <50 (3 points added if under age 50). This simplified, summated score also showed an in-sample AUC of 0.79 (Supplemental Table 2). Given greater ease of use to the simple, summated score with minimal loss of performance, this simple score was defined as the Persistency Index (PI).

The Persistency Index can identify the MFS phenotype from unselected LUTS patients.

The PI was then applied to a Validation cohort consisting of 719 patients who consecutively sought care in the 2019 calendar year. This cohort included subjects with MFS (n=111), OAB (n=285), IC/BPS (n=53), controls (n=209), and those with LUTS of unclear etiology (n=61). The out-of-sample AUC for the simple, summated PI in the Validation cohort was 0.74 (95% CI 0.67, 0.83) (Supplemental Table 3). Sensitivity and specificity were calculated for each PI score in the Validation cohort (Supplemental Table 4). Youden’s Index was used to identify the optimal cut point for maximizing sensitivity and specificity, which was found to be a simple, summated score ≥ 7 which had 80% sensitivity and 61% specificity. Overall accuracy of diagnosis at this score was 65% (Supplemental Table 4, Fig. 3).

Figure 3. Performance of the Persistency Composite Index.

A) Distribution of Persistency across the pooled population of 215 subjects by diagnosis (Derivation Dataset). Positivity on the PCI was defined as >7, which best defined the population with myofascial frequency syndrome. B) Sensitivity, Specificity, and Overall Accuracy of the PCI for identifying MFS in the Validation Dataset. A summated score of 7 denotes the cut point threshold that maximizes both sensitivity and specificity. IC/BPS: interstitial cystitis/bladder pain syndrome; MFS: myofascial frequency syndrome; OAB: Overactive Bladder.

p-CLUS nomogram can distinguish different etiologies of LUTS

Application of these four indices (Bother, BPCI, UICI, and PI]) to this unselected population was able to separate the population into phenotypic groups, which correlated well with specialist-assigned diagnoses of OAB, IC/BPS, and MFS (Fig. 4). Subjects identified as MFS patients using PI>7 (n=125) were highly bothered subjects (median Bother Index: 8.8±2.0) with low scores on the UICI (1.6±0.4) and BPCI (1.1±0.5) and elevated PCI scores (8.4±1.0). This independent population of MFS subjects exhibited the same pattern of symptomatology on their patient-reported questionnaires as observed in the initial description of MFS;³ questions describing pelvic pressure and heaviness (POPDI-6 1 and 2), straining to defecate (CRADI-8 1), urinary frequency (UDI-6 1, ICSI2, ICPI1, fGUPI 6), incomplete emptying (UDI-6 5), and bladder/pelvic discomfort (UDI-6 6, fGUPI 1, fGUPI 4, ICPI4) were significantly increased over controls in this population (Supplemental Table 5). The algorithm categorized subjects with a known MFS diagnosis with good overall accuracy; only 18% of subjects overall had a discordant nomogram classification to the physician-assigned diagnosis (15% inaccuracy for IC/BPS, 13% inaccuracy for MFS, and 23% inaccuracy for OAB).

Figure 4. BPCI, UICI and PI define distinct symptomatic populations with LUTS.

A) Subjects from the Validation Dataset were plotted in a 3D plot with each index on a different axis. Bother is indicated by the size of the sphere indicating each subject. The colors represent the diagnosis given by the consulting subspecialty physician. Diagnosis correlated well with the position in the 3D plot designated by the ellipses for each LUTS phenotype designated by the same color.

Comment

Principal Findings

MFS, a symptom complex of urinary frequency, bladder pressure, sensation of incomplete emptying, muscle laxity/weakness and persistent desire to urinate, is common in patients presenting for urogynecologic care but remains underrecognized. Currently, most diagnostic approaches to LUTS only marginally address pelvic floor myofascial dysfunction as an independent contributor to urinary symptomatology. However, the pelvic floor plays a central role in pelvic, visceral function and dysregulation can provoke pelvic pain syndromes and urinary complaints.¹² While muscle laxity is widely acknowledged as a contributor to POP or SUI, muscle hypertonicity or discoordination is rarely recognized in urologic disorders. While understood to contribute to dysfunctional voiding, dysynergia of the pelvic floor can demonstrate a vast spectrum of clinical presentations. Widely fluctuating estimates of its prevalence are indicative of a high rate of misdiagnosis and poor understanding of the contribution of the pelvic floor to urinary complaints.¹³ A growing body of evidence, however, suggests that increased pelvic floor tone underlies many urinary, gastrointestinal, and sexual complaints, even in the absence of pain.¹⁴

Results in the Context of What is Known

In the colorectal literature, a high-tone or hypertonic pelvic floor is well understood to result in symptoms of fecal urgency, stool frequency, and obstructive defecation, which can often present as chronic, idiopathic constipation.¹⁵ As in defecation, normal urinary function similarly requires appropriate relaxation and coordination of the pelvic floor muscles, including urinary sphincters. Thus, a growing body of evidence suggests that a high-tone pelvic floor (“non-relaxing” or fixed) can result in analogous symptoms in the urinary tract, embodied in the clinical phenotype of MFS.^{3, 4, 6, 16–18}

MFS is prevalent in individuals highly bothered by their LUTS. In our unselected population of subjects seeking care in a pelvic medicine specialty clinic, MFS comprised more than 20% of subjects presenting with LUTS, a similar prevalence to that seen in the initial cohort used to describe this condition.³ This is similar to the prevalence of pelvic floor dysfunction (dyssynergic defecation) in approximately 22–27% of patients presenting with chronic, idiopathic constipation.^{19, 20} Increasing awareness of this prevalent cause of pelvic symptomatology, however, may not be sufficient to improve care; additional tools and standards for diagnosis are needed.

Clinical Implications

The screening measure developed in this study, the PI, performs well in a real-world, unselected population of patients seeking care for LUTS to identify MFS with an overall accuracy of 65%. This accuracy reflects the fact that the chosen threshold for diagnosis tends to overpredict the likelihood of MFS (Supplemental Fig. 1), but misses few with true MFS, as evidenced by the 80% sensitivity. As the likely outcome of obtaining a diagnosis of MFS would be a referral to pelvic floor physical therapy, a treatment that carries very few side effects and the potential for symptomatic improvements even in patients with other diagnoses, the risks of over diagnosing MFS are minimal. This risk-benefit ratio reflects the optimal features of a screening measure, especially for an underrecognized condition which could assist providers in identifying patients that would benefit from consideration of myofascial-directed therapies and perhaps avoid unnecessary pharmacologic or even surgical interventions.

The addition of the PI to our previously described LUTS diagnostic nomogram provided a new diagnostic algorithm, the phenotyping Comprehensive Lower Urinary Symptoms (p-CLUS) nomogram. Application of this classification method provided an overall accuracy of LUTS classification in an unselected population of care-seeking women of 82% (Supplemental Fig. 2), using only 11 symptomatic questions and patient age. As this information can be obtained without an in-person assessment, this may be a valuable clinical tool that could be used by providers across a range of disciplines, even when assessing patients by telemedicine, to help with initial diagnostic evaluation and treatment assignment.

Research Implications

As this dataset included only women, utilizing validated questionnaires that are specific to pelvic symptomatology in women, it is not clear how these symptoms manifest in men, nor the prevalence of myofascial dysfunction as a cause for LUTS. A comprehensive myofascial pelvic exam is more challenging in men, as an internal assessment can only be performed during a digital rectal exam, which is frequently perceived as uncomfortable and may provoke reflexive pelvic floor contraction confounding the utility of examination.

Strengths and Limitations

Such discomfort on exam confounding assessment may also manifest in women. The accuracy of the PI at diagnosing MFS was determined without an objective gold standard for diagnosis; the comparative measure determining whether a subject’s urinary symptoms derived from a myofascial origin was based on symptom assessment and subjective pelvic floor exam by a board-certified urogynecologist. This method, while the current diagnostic standard, is itself an inconsistent measure.^21–23 Thus, it is possible that the real accuracy of the PI in identifying MFS is greater than anticipated, as some patients with MFS may be misclassified by this subjective approach. In a 3-dimensional graphic representation of patient symptoms (Fig. 4), a significant number of patients diagnosed with OAB are noted within the region of the graph with a high PI, low BPCI (bladder pain) and low UICI (urgency incontinence). While it is possible the PI is not accurate in such subjects, it remains equally plausible that these patients represent misdiagnosed cases of MFS which could have been recognized by using a screening method such as the PI.

Conclusions

In summary, we provide a screening measure helpful in identifying patients with MFS, which would likely benefit from myofascial-directed therapies. While the diagnostic utility of the PI will require further prospective evaluation, early identification of such patients will facilitate targeted therapy, expedite recovery, and minimize unnecessary pharmacologic and procedural interventions.

Tweetable Statement:

The Persistency Index is a new screening measure to identify patients with urinary symptoms related to myofascial pelvic floor dysfunction based on validated survey data.

AJOG at a Glance:

• We developed an algorithmic screening measure based on patient reported symptoms to identify patients with urinary symptoms related to myofascial pelvic floor dysfunction.

• This algorithm separated populations of women, a total of 1,084, in both derivation and validation cohorts who were seeking care for lower urinary tract symptoms into groups consistent with Overactive Bladder, Interstitial Cystitis/Bladder, and Myofascial Frequency Syndrome phenotypes with an overall diagnostic accuracy of 82%.

• This screening approach will help identify patients with myofascial pelvic floor dysfunction who would likely benefit from myofascial-directed therapies. Early identification of such patients would facilitate targeted therapy, expedite recovery, and minimize unnecessary pharmacologic and procedural interventions

Abbreviations:

AUC

Area Under Curve

BP

Bladder Pain

BPCI

Bladder Pain Composite Index

CRADI-8

Colorectal-Anal Distress Inventory

DN

Double Negative

DP

Double Positive

fGUPI

female Genitourinary Pain Index

FPMRS

Female Pelvic Medicine and Reconstructive Surgery

IC/BPS

Interstitial Cystitis/Bladder Pain Syndrome

ICSI

Interstitial Cystitis Symptom Index

ICPI

Interstitial Cystitis Problem Index

LASSO

Least Angle Shrinkage and Selection Operator

LUTS

Lower Urinary Tract Symptoms

MFS

Myofascial Frequency Syndrome

MPP/PFM

Myofascial Pelvic Pain/Pelvic Floor Myalgia

OAB

Overactive Bladder

OABq

Overactive Bladder questionnaire

PI

Persistency Index

PFDI

Pelvic Floor Distress Index – 20

PFPT

Pelvic Floor Physical Therapy

POPDI-6

Pelvic Organ Prolapse Distress Inventory 6

PCI

Persistency Composite Index

p-CLUS

Phenotyping of Comprehensive Lower Urinary Symptoms

PA

Predictive Accuracy

UI

Urgency Incontinence

UICI

Urge Incontinence Composite Index

UDI-6

Urinary Distress Inventory 6

UTI

Urinary Tract Infection