Geller, Elizabeth J. MD; Hankins, Kelly J. MD; Parnell, Brent A. MD; Robinson, Barbara L. MD; Dunivan, Gena C. MD
Postoperative voiding dysfunction has been defined as the inability to void with fluid in the bladder during the postoperative period.1 This condition commonly affects women after gynecologic surgery, including surgery for pelvic organ prolapse and urinary incontinence. Undetected voiding dysfunction can lead to retention of urine, which in turn can lead to urinary tract infections, bladder distention, and damage to the detrusor muscle.2
Rates of postoperative voiding dysfunction have been reported to range from 5% to 71%.3,4 A recent study reported a 43% rate of voiding dysfunction after pelvic organ prolapse surgery that included placement of a midurethral sling.5 The wide variation is the result of the many types of surgeries performed as well as the many definitions of urinary retention that exist.6
Methods of evaluating voiding function vary greatly based on practice patterns with no current consensus as to the best method. Some common methods include retrograde filling, spontaneous filling, and bladder scanning through ultrasonography. The retrograde method may have advantages as a result of faster performance with fewer catheterizations, while allowing for accurate measurement of postvoid residual. A recent study in women undergoing transvaginal outpatient surgery found that women were more likely to pass the retrograde method.7 The primary outcome, however, was not diagnostic accuracy but time to hospital discharge.
The primary aim of our study is to compare the diagnostic accuracy of retrograde and spontaneous techniques in detecting postoperative voiding dysfunction. Secondary aims include the ability of both methods to predict prolonged retention as well as patient satisfaction with both techniques.
MATERIALS AND METHODS
We conducted a prospective crossover trial assessing the retrograde and spontaneous methods of voiding trials. Because the objective of the study was to assess the diagnostic accuracy of two testing methods, the Standards for Reporting of Diagnostic Accuracy checklist was our primary resource.8 The study was initiated after approval by the University of North Carolina at Chapel Hill institutional review board. Participants included women undergoing operations for correction of pelvic organ prolapse, urinary incontinence, or both between November 2009 and March 2010 at the University of North Carolina. This population was chosen as a result of its relatively high expected rate of voiding dysfunction. These women were recruited from the outpatient urogynecology clinic and were enrolled consecutively during either the surgical preoperative evaluation or in the preoperative holding area before surgery. Participation in the study did not alter their surgical treatment plan, which had already been predetermined before enrollment by their clinician. Women were excluded if they were unable to give informed consent, younger than 18 years of age, non-English-speaking, planning for suprapubic catheter placement at the time of surgery, planning for same-day discharge, or an intraoperative cystotomy occurred.
Group assignment was determined by randomization through computer-generated random number blocks of six. A sequentially numbered, opaque, sealed envelope was opened for each patient after the completion of the operation, and women were randomly assigned to one of two groups: retrograde first or spontaneous first. If a cystotomy occurred, randomization did not take place and that woman was excluded from the study because she would not be undergoing a voiding trial before hospital discharge. All participants underwent both techniques of voiding trials and therefore served as their own comparison group with randomization determining the order of the two techniques. We designed the study with a randomized crossover intervention to avoid any influence of one intervention on the other. By having half the patients undergo retrograde followed by spontaneous and the other half doing the opposite, in a randomized fashion, we sought to overcome any influence of either the order of performance or the effect of a first-morning void on the actual voiding trial results.
Voiding trials were performed on the morning of postoperative day one either by the subject's nurse or by the physician in the event that the nurse was unavailable. A standardized protocol for each method was reviewed with all healthcare providers to ensure consistency. Because neither method of voiding trial is considered the gold standard, both methods were performed on all patients to compare them head to head. Patients were divided into two groups: retrograde first or spontaneous first. All patients had an indwelling Foley catheter placed intraoperatively. In one group, the retrograde technique was performed first; after confirming that all urine was drained from the bladder with the indwelling Foley catheter, a 300-mL bolus of saline was instilled into the bladder through the indwelling catheter. After removing the catheter, the patient was asked to void within 30 minutes, and the voided volume was recorded. Postvoid residual was indirectly determined by subtracting the voided volume from the 300 mL of instilled fluid. If the patient voided less than 200 mL, catheterization was performed. The spontaneous technique was then performed by allowing the patient's bladder to fill spontaneously over no more than 4 hours. Women were then instructed to void on desire into a measuring container after which straight catheterization was performed to assess the postvoid residual. Two consecutive spontaneous tests were performed for complete assessment using this technique (as is the protocol at our institution). Both must be passed to pass the spontaneous method. In the second group, the indwelling Foley catheter was removed and the spontaneous method was performed first (two consecutive times) followed by the retrograde method. Failure for either method (retrograde or spontaneous) was considered a postvoid residual greater than 100 mL. This reflects the most commonly used clinical value and meets the definition established by the International Continence Society.9 Additionally, the spontaneous trial was considered inadequate if the total voided volume was less than 200 mL. Inadequate voiding was treated as a failed spontaneous trial, because voiding function could not be fully assessed. Those who failed either method were discharged home performing intermittent self-catheterization until they had two consecutive postvoid residuals less than 100 mL. Patients completed a voiding diary at home throughout the period of catheterization and mailed it back to the investigators on completion in a self-addressed, stamped envelope. Time to return of normal voiding function was recorded for all patients.
Satisfaction with both voiding trial techniques was assessed by a brief questionnaire given to each patient after completion of both methods of the voiding trial and before hospital discharge. Participants were asked which method was easier to undergo, which method was more physically comfortable, and which of the two methods was preferred. There was a dichotomous choice for each answer, either retrograde or spontaneous.
The electronic medical record was reviewed to obtain baseline demographic data as well as information regarding postoperative outcomes such as urinary tract infection (UTI), ileus, pneumonia, deep venous thromboembolism, or pulmonary embolism. Intraoperative data were also obtained, including type of surgery performed, estimated blood loss, and any complications.
Statistical analysis was performed using SPSS 16.0 including Student's t test for continuous variables, χ2 and Fisher's exact for categorical variables, Mann-Whitney U for ranked variables, and logistic regression analyses as well as sensitivity, specificity, and positive and negative predictive values with 95% confidence intervals (CIs) for these diagnostic tests. The clinical outcome for these tests was postoperative voiding dysfunction lasting greater than 7 days. An a priori power analysis determined that 48 patients (24 per group) would be needed to have an 80% power to detect a 30% difference in pass rates between the two techniques of voiding trials (60% for retrograde and 30% for spontaneous) with a two-sided α of 0.05.7
A total of 60 women were enrolled and 50 completed the study—25 in each group (Fig. 1). The groups had similar baseline demographic characteristics (Table 1). There were no significant differences in the types of pelvic organ prolapse and/or urinary incontinence surgery performed between the two groups, including the placement of midurethral slings. In addition, preoperative Pelvic Organ Prolapse Quantification measurements and preoperative urodynamic voiding function were similar between groups.
All 50 patients completed both methods of voiding trials. Failure rates were 62% for retrograde and 84% for spontaneous (P=.001). Those who failed retrograde were in retention for an average of 12.6±14.5 days, whereas those who failed spontaneous were in retention for an average of 10.2±3.0 days. Patients who failed both methods were in retention for an average of 12.6±14.4 days, whereas those who failed only one method were in retention for 2.5±2.1 days (P=.004). There was a significant difference in mean number of days in retention among each of these four scenarios (P=.016). Retrograde failures were all the result of a high postvoid residual. Spontaneous failures were the result of either a high postvoid residual (68%) or an inadequate voided volume (16%). There was no difference in the number of days in retention based on reason for failure (high postvoid residual compared with inadequate volume). Of the 42 patients who failed the spontaneous method, only three (7%) had discrepant results on the two void attempts; two of these patients failed the first attempt and had an inadequate volume on the second attempt, and one patient failed the first attempt and passed the second attempt. All other patients either passed both or failed both attempts. Sensitivity, specificity, positive predictive value, and negative predictive value were determined for both methods of voiding trials (Table 2). The outcome of interest was defined as postoperative urinary retention lasting greater than 7 days. A Fisher's exact test was used to create a 2×2 table for each voiding trial method comparing retention lasting greater than 7 days with failure of the voiding trial. These values were used to calculate the data presented in Table 2. In assessing the diagnostic validity of these tests, the retrograde method had a composite score of sensitivity plus specificity equal to 1.53, whereas the spontaneous method had a composite score of 1.26. A composite score greater than 1.5 is an indicator of good validity and shows that the retrograde method can more accurately discriminate between the presence and absence of voiding dysfunction.10
Postoperative UTI was similar between groups (16.0% compared with 12.5%, P=.79). Rates of UTI were not increased in those who were discharged home self-catheterizing (13.8% compared with 17.6%, P=.99). Patients sent home on a catheterization method were not given antibiotics prophylactically. Other postoperative complication rates were low and not statistically significant: one postoperative transfusion in the group undergoing spontaneous first and one postoperative fever in each group.
We also assessed for cofactors affecting voiding trial pass or fail rates. Variables found to be significant in bivariate analysis included hysterectomy, anterior repair, posterior repair, mesh sling, and preoperative stress incontinence. In a logistic regression model controlling for these factors, anterior and posterior repair were found to significantly decrease the odds of passing either method of voiding trial (anterior repair: P=.038, odds ratio 0.08, 95% CI 0.01–0.86; posterior repair: P=.029, odds ratio 0.11, 95% CI 0.02–0.80).
Review of the preference questionnaire found that patients preferred the retrograde method (51.1% compared with 44.4%). This preference was not affected by randomization order (P=.40). Those who passed both methods were more likely to prefer the retrograde method because no catheterization was required with passing. The majority of healthcare providers (75%) also preferred the retrograde technique as a result of ease and rapidity of performance.
Women undergoing surgery for pelvic organ prolapse, urinary incontinence, or both often have voiding dysfunction in the immediate postoperative period. Documentation of spontaneous voiding is not sufficient to preclude voiding dysfunction.11 Several methods of voiding trials exist, but there are few studies comparing these various methods. We found that the retrograde method is a more accurate predictor of postoperative voiding dysfunction with a higher positive predictive value, an excellent negative predictive value, and useful sensitivity and specificity.
A recently published study by Pulvino et al12 assessing the retrograde and spontaneous methods also found that the retrograde method was a better predictor of postoperative voiding dysfunction. However, they did not assess the diagnostic accuracy of either method, which is lacking from the literature. Thus, the goal of our study was to determine diagnostic accuracy for each method of voiding trial. We found that both methods have a relatively low positive predictive value, but the retrograde method has more clinical relevance. Women who failed the retrograde method were in retention for an average of 12.6 days, whereas those who failed the spontaneous method but passed the retrograde test were only in retention for 2.5 days. Thus, a woman who passes the retrograde method can be reassured that she will not have prolonged voiding dysfunction, no matter what the spontaneous method shows. The spontaneous method has a very low positive predictive value, leading to more false-positive diagnoses of voiding dysfunction and more women sent home self-catheterizing. This is likely the result of the strict definition used for failure, ie, either an elevated postvoid residual or an inadequate voided volume. One could argue that an inadequate trial is not actually a failure, although it would not allow assessment of voiding function and would necessitate discharge on a self-catheterization regimen nonetheless. There was no difference in the number of days in retention between those who failed the spontaneous method as a result of a high postvoid residual and those who failed as a result of an inadequate voided volume.
The relatively low positive predictive value of both tests indicates that they are overly conservative in advocating postoperative catheterization. One must weigh the potential risk of UTI associated with catheterization against the risk of incomplete bladder emptying, which can also lead to UTI as well as detrusor damage, remembering that the rate of UTI was not higher in those who catheterized after discharge. The negative predictive value for both methods is very high, ensuring that patients who pass either method will likely have normal voiding function in the postoperative period.
The fact that the spontaneous method can produce inadequate results is a disadvantage, because some patients may be labeled with voiding dysfunction who actually are not being fully evaluated as a result of low urine output. The retrograde method circumvents this problem by filling the bladder rather than waiting for actual urine output. Another benefit of the retrograde method is the speed of performance and the potential for fewer catheterizations. In addition, this method did not delay hospital discharge as was found in some cases with the spontaneous method. A delayed discharge can increase costs to the hospital and burden the nursing staff and other healthcare providers. This method also lends itself to the same day surgery setting as a result of its rapid performance. For these reasons, retrograde was preferred by patients and staff. Given that the retrograde method provides a conclusive result, has higher specificity and positive predictive value, and is faster to perform, we recommend this method over the spontaneous method and have adopted it at our institution as the primary voiding trial method.
Our overall rate of postoperative voiding dysfunction was high with 84% failing at least one method of voiding trial (including cases of inadequate results). Rates in the literature range from 5% to 71%.3,4 For the purposes of this investigation, our definition of failure was very strict with patients who had an elevated postvoid residual or an inadequate trial being considered a failure. If we exclude inadequate results, the rate would decrease to 62% for the spontaneous method, which is consistent with a 58.7% reported rate of short-term voiding dysfunction after pelvic organ prolapse, urinary incontinence surgery, or both.13 The findings appear clinically relevant, because those who failed both methods had a longer period of retention than those who failed only one method. We also found that anterior and posterior repair decrease the odds of passing either method of voiding trial. However, these findings are not robust based on the wide CIs and small number of patients involved (15 underwent anterior repair in each group and 11 underwent posterior repair in each group).
Strengths of this investigation include the prospective crossover design with each patient serving as her own comparison. Currently there is no gold standard to assess postoperative voiding dysfunction. We therefore chose to have all patients undergo both tests to compare the two methods head to head. In addition, our study was adequately powered to detect a difference between the two methods of voiding trial in predicting postoperative voiding dysfunction. We had a low dropout rate for our primary outcome as a result of the fact that the intervention took place before hospital discharge. We also had a high follow-up rate for assessing days in retention with data available on 49 of 50 patients. A weakness of the study is the use of nonvalidated preference questionnaires. In addition, the difference in voiding trial preference was not significant, which is reasonable given that the study was not powered to formally assess this. Furthermore, neither patients nor the staff who performed the voiding trials were blinded to the intervention. This was not possible as a result of the nature of the intervention but would not likely affect the outcome because the results were the objective measurement of urine output to determine voiding trial pass or fail rates. Finally, one might argue that we should have performed a straight catheterization to directly determine postvoid residual for the retrograde method. We chose to determine postvoid residual indirectly to test the accuracy of the retrograde method as it would be performed in a clinical setting. Although this may be a limitation of this method, it outperformed the spontaneous method in terms of diagnostic accuracy.
In evaluating the diagnostic accuracy of two methods of voiding trials, the retrograde method performs better than the spontaneous method in discriminating which patients have postoperative voiding dysfunction. This method may be preferable as a result of the speed of performance, potential for fewer catheterizations, and greater diagnostic accuracy. It is important for voiding dysfunction to be assessed after surgery for pelvic organ prolapse and/or urinary incontinence. Understanding the accuracy of these tests will aid in interpretation of their results. As a result, our findings may help clinicians make a more informed decision as to which test to choose when evaluating postoperative voiding dysfunction.
1. Tammela T, Kontturi M, Lukkarinen O. Postoperative urinary retention. I. Incidence and predisposing factors. Scand J Urol Nephrol 1986;20:197–201.
2. Rosseland LA, Stubhaug A, Breivik H. Detecting postoperative urinary retention with an ultrasound scanner. Acta Anaesthesiol Scand 2002;46:279–82.
3. Pavlin DJ, Pavlin EG, Fitzgibbon DR, Koerschgen ME, Plitt TM. Management of bladder function after outpatient surgery. Anesthesiology 1999;91:42–50.
4. Petros JG, Rimm EB, Robillard RJ, Argy O. Factors influencing postoperative urinary retention in patients undergoing elective inguinal herniorrhaphy. Am J Surg 1991;161:431–33.
5. Partoll LM. Efficacy of tension-free vaginal tape with other pelvic reconstructive surgery. Am J Obstet Gynecol 2002;186:1292–5.
6. Darrah DM, Griebling TL, Silverstein JH. Postoperative urinary retention. Anesthesiol Clin 2009;27:465–84.
7. Foster RT Sr, Borawski KM, South MM, Weidner AC, Webster GD, Amundsen CL. A randomized, controlled trial evaluating 2 techniques of postoperative bladder testing after transvaginal surgery. Am J Obstet Gynecol 2007;197:627.e1–4.
8. Bossuyt PM, Reitsma JB, Bruns DE, Gatsonis CA, Glasziou PP, Irwig LM, et al.. Towards complete and accurate reporting of studies of diagnostic accuracy: the STARD Initiative. Standards for Reporting of Diagnostic Accuracy. Clin Chem 2003;49:1–6.
9. Haylen BT, de Ridder D, Freeman RM, Swift SE, Berghmanns B, Lee J, et al.. An International Urogynecological Association (IUGA)/International Continence Society (ICS) joint report on the terminology for female pelvic floor dysfunction. Neurourol Urodyn 2010;29:4–20.
10. Grimes DA, Peipert JF. Electronic fetal monitoring as a public health screening program: the arithmetic of failure. Obstet Gynecol 2010;116:1397–400.
11. Chaudhri S, Maruthachalam K, Kaiser A, Robson W, Pickard RS, Horgan AF. Successful voiding after trial without catheter is not synonymous with recovery of bladder function after colorectal surgery. Dis Colon Rectum 2006;49:1066–70.
12. Pulvino JQ, Duecy EE, Buchsbaum GM, Flynn MK. Comparison of 2 techniques to predict voiding efficiency after inpatient urogynecologic surgery. J Urol 2010;184:1408–12.
13. Oliphant S, Amburgey OA, Connolly A. Predictors of successful voiding before hospital discharge after urinary stress incontinence surgery. Int Urogynecol J Pelvic Floor Dysfunct 2008;19:805–10.
© 2011 by The American College of Obstetricians and Gynecologists.