Correlation of symptoms with location and severity of pelvic organ prolapse

Transcription

1 Correlation of symptoms with location and severity of pelvic organ prolapse R. Mark Ellerkmann, MD, a,b Geoffrey W. Cundiff, MD, a Clifford F. Melick, PhD, b Mikio A. Nihira, MD, b Kenneth Leffler, MD, a and Alfred E. Bent, MD b Baltimore, Md OBJECTIVE: The purpose of this study was to compare the symptoms that are related to pelvic floor dysfunction with the location and severity of the coexisting prolapse. STUDY DESIGN: Two hundred thirty-seven consecutive patients with symptomatic pelvic organ prolapse came to Johns Hopkins Medicine during a 24-month period beginning in July 1998 and completed a symptom-specific Likert scale questionnaire that included standardized questions that were compiled from commonly used validated instruments. All questionnaires were completed by the patients before they were seen by a physician. Further evaluation included a standardized physical examination that included the International Continence Society s system for grading uterovaginal prolapse. Symptoms were categorized according to both severity and associated anatomic compartment. Symptoms that were related to urinary and anal incontinence and voiding, defecatory, sexual, and pelvic floor dysfunction were analyzed with respect to location and severity of pelvic organ prolapse with the use of the nonparametric correlation coefficient, Kendall s tau-b. RESULTS: The mean age of the women was 57.2 years (range, years); 109 of the women (46%) had undergone hysterectomy. Overall, stage II was the most common pelvic organ prolapse (51%) that was encountered. In 77 patients (33%), anterior compartment pelvic organ prolapse predominated; 46 patients (19%) demonstrated posterior compartment prolapse, whereas 26 patients (11%) had apical prolapse. In 88 patients (37%), no single location was more severe than another. Voiding dysfunction that was characterized by urinary hesitancy, prolonged or intermittent flow, and a need to change position was associated with the increasing severity of anterior and apical pelvic organ prolapse. Pelvic pressure and discomfort along with visualization of prolapse were strongly associated with worsening stages of pelvic organ prolapse in all compartments. Defecatory dysfunction characterized by incomplete evacuation and digital manipulation was associated with worsening posterior compartment pelvic organ prolapse. Impairment of sexual relations and duration of abstinence were strongly associated with worsening pelvic organ prolapse. An inverse correlation was observed between increasing severity of pelvic organ prolapse and urinary incontinence and enuresis. CONCLUSION: Women with pelvic organ prolapse experience symptoms that do not necessarily correlate with compartment-specific defects. Increasing severity of pelvic organ prolapse is weakly to moderately associated with several specific symptoms that are related to urinary incontinence and voiding, defecatory, and sexual dysfunction. ( 2001;185: ) Key words: Pelvic organ prolapse, symptoms, defecatory and sexual dysfunction, urinary incontinence Women with pelvic organ prolapse may have a plethora of symptoms that are commonly attributed to the anatomic compartments that are involved. Urinary incontinence, voiding dysfunction, and irritative symptoms are From the Department of Gynecology and Obstetrics, Johns Hopkins Medicine a ; and the Department of Gynecology, The Greater Baltimore Medical Center. b Presented at the Twenty-seventh Annual Meeting of the Society of Gynecologic Surgeons, Orlando, Fla, March 5-7, Reprint requests: R. Mark Ellerkmann, MD, Department of Gynecology, The Greater Baltimore Medical Center, 6569 N Charles St, Suite 307, Baltimore, MD Copyright 2001 by Mosby, Inc /2001 $ /6/ doi: /mob commonly related to anterior compartment prolapse. Symptoms that are related to defecatory dysfunction (including the feeling of incomplete rectal emptying, the need for digital manipulation to facilitate defecation, and dyschezia) have been correlated with posterior wall prolapse. 1,2 Symptoms of pelvic floor herniation (including vaginal or pelvic pressure, fullness or discomfort, and/or protruding tissues) are attributed to all compartments. With the exception of Chou et al, 3 who reported a lack of significant differences in symptoms that are related to bowel function in women with and without enteroceles, few studies address the association between symptoms and enterocele. Although numerous studies suggest a common cause between the development of pelvic organ pro- 1332

2 Volume 185, Number 6 Ellerkmann et al 1333 lapse and urinary and fecal incontinence, coexisting symptoms that are related to dysfunction may vary and may not necessarily be specific to any one compartment. Furthermore, because pelvic floor weakness is infrequently localized to one anatomic compartment, numerous symptoms that are attributed commonly to specific compartments may coexist. The National Institute for Child Health and Human Development recently released a document that standardizes terminology for researchers who specialize in female pelvic floor dysfunction. In the report, they noted an absence of research that defines the associations between specific symptoms and support defects and identified this deficit as a research priority. 4 The objective of our study was to characterize and compare the symptoms that are related to pelvic floor dysfunction with the location and severity of the coexisting prolapse. We wanted to determine whether symptoms were compartment specific and whether more severe stages of prolapse were associated with increasing severity of symptoms and the degree of patient discomfort. Material and methods All patients whose cases were included for analysis in this study were seen and examined at The Johns Hopkins Medicine Urogynecology Clinic during a 24-month period commencing in July Most of the patients were ultimately examined for problems that were associated with pelvic organ prolapse and dysfunction (including but not limited to urinary and fecal incontinence, voiding, and defecatory and sexual dysfunction). New patients were asked to complete 2 questionnaires before their initial scheduled appointment. If a patient failed to complete the questionnaire before her scheduled appointment, she was asked to do so before being seen. Ancillary support personnel provided nondirective assistance to patients who were unable to complete the questionnaires by themselves. All questionnaires were completed before the patient was seen by a physician. The first questionnaire was designed to provide the practitioner with a summary of the patient s medical, surgical, family, and social history and included questions that are relevant to current medications and allergies, underlying medical conditions, surgical, obstetric, gynecologic, family, and social history. The second questionnaire is a condition-specific symptom survey and quality of life instrument for women with pelvic floor dysfunction. This questionnaire is comprised of questions adapted and modified from previously validated instruments that are used to access patients with urinary incontinence and pelvic floor dysfunction. 5-7 The Pelvic Floor Dysfunction Questionnaire is designed to evaluate both the presence and degree of patient distress with respect to symptoms and the impact that these symptoms have on day-to-day activities. The Pelvic Floor Dysfunction Questionnaire is divided into 8 domains, each with questions that are limited to specific areas of dysfunction. These domains include urinary incontinence, urinary irritative symptoms, voiding dysfunction, pelvic organ prolapse symptoms, fecal incontinence, defecatory dysfunction, pelvic pain, and sexual dysfunction. Likert scale assignment was used to respectively quantify severity (none, 0; minimally, 1; moderately, 2; severely, 3) and duration of symptoms (never, 0; <25% of time, 1; <50% of time, 2; <75% of time, 3; 100% of time, 4) and impact on quality of life. Two hundred thirty-seven consecutive patients completed the questionnaires and underwent standardized physical examination that was performed by the same attending urogynecologist (G.W.C.) or by urogynecology fellows under his direct supervision. Physical examination included an assessment of neuromuscular function with an evaluation of cranial nerves, deep tendon reflexes, anal and bulbocavernosus reflexes, and pelvic muscle strength. Pelvic examinations were performed in the dorsal lithotomy position. An empty supine stress test, a postvoid urine measurement, a cotton swab urethral mobility test, and a bimanual pelvic examination were performed for all patients. A Graves speculum and ring forceps were used to independently access pelvic floor support in each anatomic compartment (anterior and posterior vaginal walls and vaginal apex). All points for the pelvic organ prolapse quantification examination, except for total vaginal length, were recorded at maximal protrusion with Valsalva maneuver. The staging of pelvic organ prolapse conformed to the standards and terminology set forth by the International Continence Society. 8 Based on the pelvic organ prolapse quantification examination values, stages 0 to 4 were assigned to each pelvic floor compartment (anterior, apical, posterior) in each patient. If possible, an overall stage was assigned to each patient, according to the most severely prolapsing compartment. Statistical analysis was performed, with the severity of symptoms and stages of prolapse as ordinal variables. Because the variables analyzed are, for the most part, derived from ordinal scales in which the numbers represent rank orders but do not give any information regarding the differences between adjacent ranks, correlation analysis was performed with Kendall s tau-b, a nonparametric measure of association. Kendall s tau-b is not an inferential statistical test but rather a descriptive statistical measure that represents the degree of relationship between 2 variables. As with most correlation coefficients, the closer the value is to 1, the stronger is the correlation. Conversely, as the coefficient approaches 1, the stronger an inverse correlation becomes. Values that range between 0.1 and 0.3 are considered to be weakly correlated; values that range between 0.31 and 0.6 are considered to be moderately correlated, and values that are greater

3 1334 Ellerkmann et al December 2001 Table I. Distribution of prolapse stage with respect to anatomic compartment Prolapse location Anterior Apical Posterior Prolapse stage n % n % n % Stage Stage Stage Stage Stage Totals than 0.61 are considered to be strongly correlated. Probability values for this analysis evaluate the hypothesis that the correlation is zero. A probability value of.05 or less was considered statistically significant. Results Two hundred thirty-seven women completed the questionnaires and subsequently underwent urogynecologic examination. The mean age of the women was 57.2 years (range, years). One hundred eighty-six women (79%) were postmenopausal; of these, 151 women (64%) were receiving hormone replacement therapy. Mean parity was 2.2 (range, 0-9). Forty-one women (17%) were nulliparous; 41 women (17%) reported 1 delivery, and the remaining 155 women (65%) reported more than 1 delivery. Of the 515 deliveries reported in this patient population, 20 deliveries (4%) were performed by cesarean delivery. The use of forceps was recalled in 96 deliveries (18%), and nonspecified vaginal tears were associated with 155 deliveries (31%). Previous gynecologic operation was performed in 182 women (77%), including hysterectomy in 109 women (46%) and operations for urinary incontinence or prolapse in 39 women (16%). An analysis of our pelvic organ prolapse quantification examination data revealed 26 women (11%) had no demonstrable prolapse. Twenty-four women (10%) were found to have an overall stage 1 prolapse; 121 women (51%) had stage 2 prolapse; 44 women (19%) had stage 3 prolapse, and 22 women (9%) had stage 4 prolapse. In attempting to assign a most severe location to each patient, no single location predominated in 88 women (37%). When a predominate compartment could be assigned, 77 women (33%) had anterior compartment prolapse. Posterior compartment prolapse predominated in 46 women (19%); apical prolapse was found in 26 women (11%). When we evaluated demonstrable prolapse within a specific compartment, stages 1 and 2 predominated in all 3 locations (Table I). One hundred seventy-two women (73%) reported urinary incontinence. Of these women with incontinence, 23 women (13%) reported stress incontinence only; 9 women (5%) reported urge incontinence only; 10 women (5%) reported unconscious leakage; and the remaining 130 women (76%) reported a combination of both urge and stress incontinence. Thirty-two women (19%) reported episodes of enuresis. Irritative urinary symptoms were quite common, with more than two thirds of the women reporting some degree of frequency and urgency (Table II). Symptoms of voiding dysfunction were also commonly reported. Nearly one half of our patients experienced difficulty emptying their bladders, and more than 50% of our patients experienced symptoms of incomplete emptying and postvoid dribbling (Table II). Symptoms that were attributable to obstructive voiding (including hesitancy, weak or prolonged flow, intermittent flow, and the need to change position to facilitate micturition) were also quite common in our patient population (Table II). A large percentage of the women also complained of symptoms that were suggestive of pelvic floor herniation that included the sensation of lower abdominal pressure, pelvic heaviness, pelvic discomfort with standing or activity, and visualization of a bulge or protrusion in the vaginal area (Table II). Defecatory dysfunction was also commonly expressed. Two thirds of the women reported constipation, although dyschezia, incomplete evacuation, and the need to place a finger in the rectum or vagina to facilitate defecation were also common complaints. All but 5 women answered questions regarding the use of over-the-counter laxatives and cathartics: 86 women (36%) reported using either fiber supplements or stool softeners on a regular basis to facilitate defecation; 42 women (18%) reported using enemas, and 73 women (31%) used laxatives. Seventy-three patients reported experiencing episodes of fecal incontinence; of these, 46 women (63%) thought this interfered with their normal activities, and 41 women (56%) thought their fecal incontinence was getting worse. Pelvic pain was reported in 44% of women (102/237 women). Of those experiencing some degree of pain, 70 women (69%) reported that the pain interfered with their quality of life, and 61 women (60%) considered their pain to be getting worse.

4 Volume 185, Number 6 Ellerkmann et al 1335 Table II. Frequency of symptoms (n = 237) Symptom n % Irritative symptoms Frequency Urgency Voiding dysfunction Difficulty emptying Sensation of incomplete voiding Hesitancy Weak/prolonged flow Intermittent flow Postvoid dribbling Require position change Defecatory dysfunction Constipation Dyschezia Incomplete evacuation Digital manipulation Fecal incontinence Pelvic floor herniation Lower abdominal pressure Pelvic heaviness Pelvic discomfort when standing Visualization of prolapse Table III. Correlation between symptoms and worsening prolapse with respect to location using Kendall s tau-b correlation coefficient Symptoms Severity of prolapse (tau-b) Anterior Apical Posterior Stress incontinence Enuresis Urinary incontinence Hesitancy Prolonged flow Intermittent flow Change position Difficulty voiding Pelvic pressure Visualize bulge Pelvic discomfort Pelvic heaviness Incomplete evacuation Digital manipulation Impairment of sex life Duration of abstinence Only statistically significant correlation coefficients are shown (P <.05). With respect to sexual dysfunction, 105 women (44%) reported that they were sexually active; 128 women (54%) reported that they were not sexually active, and 4 women (2%) did not answer the question. In those patients who reported sexual activity, 69% (72/105 women) reported dyspareunia (45 women with penile insertion and 62 women with deep penetration). Fifty-seven percent (60/105 women) reported that dyspareunia had adversely affected their frequency of intercourse. Other factors that adversely affected sexual relations to some degree in sexually active patients included fecal incontinence (15%), urinary incontinence (27%), pelvic organ prolapse (28%), spousal limitations (37%), and pelvic pain (41%). The severity and duration of pelvic floor symptoms were analyzed with respect to each vaginal compartment (anterior, apical, posterior) and its respective stage of prolapse with the use of the nonparametric correlation coefficient, Kendall s tau-b. Symptoms were also analyzed with respect to overall stage. A total of 62 variables were systematically evaluated; of these, 16 variables demonstrated statistically significant (P <.05) weak to moderate correlations with specific compartments and increasing severity of prolapse. An inverse correlation was observed between the increasing severity of anterior compartment prolapse and urinary incontinence (stress incontinence and enuresis). Urinary incontinence was weakly inversely correlated with worsening prolapse in all 3 compartments (Table III). Two symptoms that were suggestive of voiding dysfunction, the sensation of hesitancy and the need to change position to facilitate urination, were weakly correlated with worsening prolapse in all compartments. Prolonged micturition and intermittent flow were associated with worsening anterior and apical compartment prolapse, whereas difficulty emptying the bladder was correlated with anterior prolapse only (Table III). Moderate positive correlations were also found with 4 symptoms of pelvic floor herniation and worsening prolapse in all 3 compartments: lower abdominal pressure, pelvic heaviness and discomfort, and visualization of actual prolapse (Table III). The sensation of incomplete defecation was weakly correlated with worsening posterior compartment prolapse. The need to place a finger in the vagina or rectum or on the perineum to facilitate defecation was weakly correlated to both worsening apical and posterior compartment prolapse. The impairment of sexual relations was moderately associated with worsening prolapse in all 3 compartments, apical being the most pronounced. Worsening anterior compartment prolapse was weakly correlated with the increasing time of sexual abstinence. Comment The most important finding in this study is that very few correlations were encountered between symptoms of pelvic floor dysfunction and the presence of pelvic organ prolapse. Although there were weak to moderate correlations with respect to several symptoms that are typically thought to be compartment specific, it was not possible to determine a specific stage of prolapse at which these symptoms became more pronounced. Symptoms that correlated with worsening prolapse in more than one compartment (urinary incontinence and hesitancy, pro-

5 1336 Ellerkmann et al December 2001 longed and intermittent micturition, the need to change position to facilitate micturition, visualization of prolapse, pelvic pressure, heaviness and discomfort, digital manipulation, and impairment of sexual relations) did not differentiate between compartments. Given this, our findings may have limited clinical significance. Patients who experience worsening prolapse, for example, may have more pelvic discomfort or be more easily able to visualize their prolapse, but the compartment responsible for these symptoms remains elusive. The correlations that we found, however, intuitively make sense. The weak inverse relationship between the worsening anterior compartment and stress incontinence may be a result of mechanical obstruction or kinking of the urethra. Bump et al 9 reported higher transmission ratios and maximum urethral closure pressures in women with severe degrees of uterovaginal prolapse, which suggests that the stress continence mechanism in women with severe prolapse results from mechanical obstruction of the less mobile urethra. Richardson et al 10 reported similar findings and conclusions. Irritative voiding symptoms and detrusor instability may also be found concurrently with cystocele. 11 Romanzi et al 12 reported a significant association between voiding dysfunction and detrusor instability and pelvic organ prolapse. Interestingly, our analysis did not demonstrate a correlation between worsening anterior compartment prolapse and urge incontinence. Our findings are consistent with the literature, however, with respect to voiding dysfunction. The fact that we found moderately strong correlations between the worsening degrees of prolapse in all compartments and the symptoms of pelvic discomfort and the visualization of a bulge or protrusion should come as no surprise. With respect to defecatory dysfunction, our findings that show weak correlations with only 2 symptoms (sensation of incomplete evacuation and digital manipulation) and posterior compartment prolapse are fairly consistent with previously published studies. Capps 13 and Yoshioka et al 14 found no correlation between symptoms and rectocele size. This lack of correlation has also been confirmed radiographically with evacuation proctography. 15 Weber et al 16 compared symptoms of bowel dysfunction with respect to stage of posterior compartment prolapse and reported no clinically significant associations between the symptoms and the extent of prolapse. They did, however, find a weakly positive correlation between more advanced posterior vaginal prolapse and severity of bowel symptoms. Sexual dysfunction has been attributed to prolapse and incontinence. 17 More recently, Weber et al 18 reported that the increasing stage of prolapse predicted interference with sexual activity but did not affect the description of satisfaction with the sexual relationship or the frequency of intercourse. Our findings suggest a mild to moderate correlation between the impairment of sexual activity and worsening prolapse in all 3 compartments. We could not determine from our data whether this had a specific impact on coital frequency or relationship satisfaction. Patients with worse anterior compartment prolapse were also more likely to have been abstinent the longest. Another interesting finding is the distribution of pelvic organ prolapse stage in our patient population. Although representative of a large metropolitan referral practice, the vast majority of patients (51.1%) had stage 2 prolapse. This is not very different from the 47.7% prevalence of stage 2 prolapse that was found by Swift 19 in a general population of female patients who were seen for routine gynecologic health care. Our trend, however, toward more severe prolapse (stage 3 and 4) is likely to be representative of our generally older population and its known risk factors. Many practices make use of questionnaires to gather demographic information and patient history. Questionnaires that attempt to document patients symptoms and the quality of life related to those symptoms are more difficult to formulate because they are usually specific to the practice and a given population and because they ideally require validation for internal consistency, test-retest reliability, and construct validity. One of the weaknesses of this study is that our questionnaire, albeit comprised of many questions compiled from previously validated condition-specific quality of life instruments for women with pelvic floor dysfunction, was itself not validated. This fact undermines the strength of our conclusions. Additionally, the use of questionnaires introduces inherent error not only in terms of recall bias but also in terms of missing information, which in itself impairs the power of our correlation coefficient. Another weakness in this study is that the examiner was not blinded to the results of the questionnaire before examining the patients, which could lend itself to selection bias. Given our tertiary care referral practice, it is clear that our findings may not be generalizable to other patient populations. Our goal was simply to analyze the presence and severity of symptoms that could be attributed to pelvic floor dysfunction and to see whether any of these symptoms were compartment specific. If they were, did they become more pronounced with a worsening stage of prolapse? We found that women with pelvic organ prolapse may experience symptoms that do not necessarily correlate with compartment specific prolapse. Based on our data, we found that increasing severity of prolapse is only weakly to moderately associated with several specific symptoms that are related to urinary incontinence, voiding, and defecatory and sexual dysfunction. A study that incorporates a larger population of patients and a more stringent protocol for questionnaire completion and physical examination documentation is underway to both confirm and clarify these findings.

6 Volume 185, Number 6 Ellerkmann et al 1337 REFERENCES 1. Spence-Jones C, Kamm MA, Henry MM, Hudson CN. Bowel dysfunction: a pathogenic factor in uterovaginal prolapse and urinary stress incontinence. Br J Obstet Gyneacol 1994;101: Weber AM, Walters MD, Ballard LA, Booher DL, Piedmonte MR. Posterior vaginal prolapse and bowel function. Obstet Gynecol 1998;179: Chou Q, Weber AM, Piedmonte MA. Clinical presentation of enterocele. Obstet Gynecol 2000;96: Weber AM, Abrams P, Brubaker L, Cundiff GW, Davis G, Dmochowski RR, et al. The standardization of terminology for researchers in female pelvic floor dysfunction. Int Urogyn J 2001; 12: Ubersax JS, Wyman JF, Shumaker SA, McClish DK, Fantl JA. Short forms to assess life quality and symptom distress for urinary incontinence in women: the incontinence impact questionnaire and urogenital distress inventory. Neurourol Urodyn 1995;24: Rockwood TH, Church JM, Fleshman JW, Kane RL, Mavrantonis C, Thorson AG, et al. Patient and surgeon ranking of the severity of symptoms associated with fecal incontinence. Dis Colon Rectum 2000;43: Agachan F, Chen T, Pfeifer J, Reissman P, Wexner SD. A constipation scoring system to simplify evaluation and management of constipated patients. Dis Colon Rectum 1996;39: Bump RC, Mattiasson A, Bo K, Brubaker LP, DeLancey JO, Klarskov P, et al. The standardization of terminology of female pelvic organ prolapse and pelvic floor dysfunction. Am J Obstet Gynecol 1996;175: Bump RC, Fantl JA, Hurt WG. The mechanism of urinary continence in women with severe uterovaginal prolapse: result of barrier studies. Obstet Gynecol 1988;72: Richardson DA, Bent AE, Ostergard DR. The effect of uterovaginal prolapse on urethrovesical pressure dynamics. Am J Obstet Gynecol 1983;146: Friis E, Hjortrup A, Nielsen JE, Sanders S, Walter S. Urinary incontinence and genital prolapse: a prospective blind study of the value of urodynamic evaluation. J Urol 1982;128: Romanzi LJ, Chaikin DC, Blaivas JG. The effect of genital prolapse on voiding. J Urol 1999;161: Capps WR Jr. Rectoplasty and perineoplasty for the symptomatic rectocele: a report of fifty cases. Dis Colon Rectum 1975; 18: Yoshioka K, Matsui Y, Yamada O, Sakaguchi M, Takada H, Hioki K, et al. Physiologic and anatomic assessment of patients with rectocele. Dis Colon Rectum 1991;34: Kelvin FM, Maglinte DDT, Hornback JA. Dynamic cystoproctography of female pelvic floor defects and their interrelationships. Am J Radiol 1997;169: Weber AM, Walters MD, Ballard LA, Booher DL, Piedmonte MR. Posterior vaginal prolapse and bowel function. Obstet Gynecol 1998;179: Field SM, Hilton P. The prevalence of sexual problems in women attending for urodynamic investigation. Int Urogynecol J 1993;4: Weber AM, Walters MD, Schover LR. Sexual function in women with uterovaginal prolapse and urinary incontinence. Obstet Gynecol 1995;85: Swift SE. The distribution of pelvic organ support in a population of female subjects seen for routine gynecologic health care. 2000;183: Discussion DR R. EDWARD VARNER, Birmingham, Ala. I commend and congratulate Drs Ellerkmann, Cundiff, Melick, Nihira, Leffler, and Bent for undertaking such a project and for their hard work. Such a database is very important to help us better understand pelvic organ prolapse and to better decide outcome measurements for our interventions. In summary, the purpose of the project was to correlate various symptoms and their severity with various stages of pelvic organ prolapse, specifically with sites of prolapse. The symptom and quality of life questionnaire was modified from several previously validated instruments, and a Likert Scale was used to quantitate severity and duration of symptoms. Patient evaluations for prolapse were performed in lithotomy with and without Valsalva, and defects were documented by standard pelvic organ prolapse quantification terminology. Data presented on the total group of patients included a description of the prolapse stage and the overall frequencies of each symptom that was evaluated. In addition, statistical analyses were performed with Kendall s tau-b analysis to correlate the severity of the specific symptoms and the stages of prolapse in the anterior, posterior, and apical compartments. Kendall s tau-b is a nonparametric statistic used to compare 2 or more distributions of variables with the use of ranks of individual variables. Such a method is useful when the distributions cannot meet the assumptions for parametric tests, such as normality. For this study, comparisons of all the symptoms and their severity with all stages and sites of pelvic organ prolapse in one or more tables would be very cumbersome. The Kendall tau-b analysis allowed for a more concise interpretation of these data. Only the symptoms that correlated statistically with sites of prolapse were presented in Table III. In addition to these data, I would find it useful to see frequencies of the more site-specific symptoms as related to stage 3 and 4 prolapse at those sites. The study found that, in general, the presence of symptoms and the severity of symptoms did not correlate well with advancing stages of site-specific pelvic organ prolapse and that many common symptoms did not differentiate between compartments. I suspect that most clinicians will not be surprised at these findings. Why might patients with prolapse have varying degrees of symptoms? The variability of pathologic conditions within specific anatomic compartments may explain some of these differences. For example, some individuals with anterior defects have additional micturition defects; some individuals do not. Some individuals have recurrent cystitis; other individuals do not. In addition, patient perception of the symptoms may vary. For example, an anxious 35-year-old woman with stage 2 to 3 prolapse may perceive her symptoms differently than a debilitated, partially demented elderly patient. The authors well describe possible weaknesses of their questionnaire and the failure to blind the examiner to questionnaire results. They are now undertaking an expansion of the study with a more stringent protocol for questionnaires. This will be labor intensive and produce a large amount of data that, I suspect, could be confusing if presented en masse. I urge the authors to narrow their reports to the examination of specific types of pelvic support defects rather than a group of defects together. In addition to the examination of apical, anterior, and pos-

7 1338 Ellerkmann et al December 2001 terior prolapse, it would be useful to see a correlation of some measure of pelvic floor dysfunction with symptoms. DR ELLERKMANN (Closing). In response to your question about statistical power, we did not perform a power calculation before data analysis. This study was designed to be a preliminary look at pelvic floor symptoms. The question that we really wanted to answer was whether there was a relationship between compartment-specific pelvic organ prolapse and specific symptoms. Moreover, was there a point of prolapse at and beyond which symptoms became more pronounced? From our research, this became very difficult to define, because we observed no strong relationships. Dr Weber and her colleagues at the Cleveland Clinic have published extensively on pelvic floor dysfunction that is related to prolapse and have made several insightful correlations. We are in the process of reexamining our data to see whether we can draw any other conclusions. As you know, one of the biggest problems we have with questionnaires is not only the issue of validation but also the problem that is associated with patients failing to completely answer all the questions. Missing data can certainly compromise the validity and strength of any conclusions. We are working on a computer-generated questionnaire that will compel patients to answer all questions.