Pelvic floor disorders are a common problem affecting women, with a 20% lifetime risk of undergoing surgery for pelvic organ prolapse (POP) or stress urinary incontinence.1 These surgeries generally improve function of the lower urinary tract, but some women develop new voiding dysfunction postoperatively. De novo incomplete bladder emptying (IE) can be a particularly vexing problem for patients and physicians. This problem is generally thought to result from a bladder outlet obstruction, disruption of nerves during surgical dissection, edema, pain, and/or perioperative medications. Suburethral slings pose the highest risk for IE due to bladder outlet obstruction, but women undergoing POP repairs may also be at risk. The rate of catheter use upon hospital discharge varies by the definition used and has been reported up to 32.4% after posterior colporrhaphy,2 43% after combined midurethral sling and prolapse repair,3 and 25% after isolated midurethral sling4 in the Trial of Mid-Urethral Slings (TOMUS) trial. Incomplete bladder emptying generally resolves within 1 week after surgery, but 4.1% to 5.8%5,6 require prolonged bladder drainage for 6 weeks. Persistence of IE beyond 6 weeks typically occurs only after midurethral sling placement and usually resolves with sling release. Sling release or revision for IE ranges from 1.3% to 3.7% after sling placement only,7–9 and from 0.6% to 4.9% with concomitant prolapse repair.6,10
The main risks of unrecognized IE after incontinence and prolapse surgery are urinary tract infection and complete urinary retention. Acute urinary retention of large volumes is painful and may damage the detrusor muscle, further impairing voiding function and prolonging the need for catheterization. Most pelvic floor surgeons screen for IE with a void trial before hospital discharge; those who fail require catheterization until there is adequate recovery of bladder function. A commonly used method for the void trial involves retrograde filling of the bladder with 300 mL of sterile fluid and measurement of the voided volume. Voids greater than two thirds of the instilled volume are deemed adequate and do not require catheterization at the time of discharge. This method has an excellent negative predictive value for catheter reinsertion.11 Risk factors for IE other than slings are inconsistently identified in the literature. The objective of this study was to evaluate risk factors for postoperative IE in our practice. We also were interested in examining the incidence of, and risk factors for, a false-negative void trial. Finally, we sought to describe the postoperative course of patients who develop severe acute urinary retention.
MATERIALS AND METHODS
We performed a case-control study of women who underwent prolapse repair and/or sling by 4 fellowship-trained female pelvic medicine and reconstructive surgeons from June 2011 to April 2016. The patient population was identified by current procedural terminology (CPT) codes for sling, anterior repair, posterior repair, extra- and intraperitoneal vaginal colpopexies, laparoscopic and abdominal sacral colpopexies, and colpocleisis (Appendix A). Both autologous and synthetic slings were included. Exclusion criteria included preoperative catheter use, preoperative postvoid residual (PVR) greater than 150 mL, and postoperative catheter use for reasons other than IE (ie, urethral surgery, cystotomy, etc). All patients underwent a standardized postoperative voiding trial as described above before discharge after surgery. All cases were collected, and the pool of potential controls was identified in an initial chart review. Cases were broadly defined as patients with any postoperative catheter use for IE; these included both those who failed the void trial upon hospital discharge and those who passed the hospital void trial and subsequently developed IE requiring catheterization in the 3 months after surgery. In our institution, standard practice is discharge home with an indwelling Foley catheter on failure of hospital void trial, and repeat void trial in the office on postoperative day 2 or 3. Clean intermittent catheterization is performed if office void trial is failed. The potential controls were those who met inclusion criteria and did not require catheterization postoperatively. The final controls were then selected from this pool for full chart review and data collection in a 1:1 ratio using a random number generator.
We collected data including age, race, medical comorbidities, prolapse stage, overactive bladder symptoms, preoperative PVR, urodynamic parameters [uroflow—PVR, maximum flow rate, and pattern; cystometry—leak point pressure testing and detrusor overactivity (DO); pressure flow—PVR, maximum flow rate, and detrusor pressure and maximum flow], operative information (surgeon, procedures, estimated blood loss, and duration of surgery), duration of postoperative IE, and need for sling revision due to IE. We also identified patients who developed symptomatic IE after passing either the initial hospital void trial or a delayed outpatient void trial for those sent home with a catheter. We recorded the PVR on presentation in these patients with recurrent IE to identify severe bladder distension, defined as PVR greater than 1000 mL. To calculate the incidence of IE outcomes, we used the CPT code query and chart review to determine the number of patients who underwent sling only, prolapse repair only, and prolapse repair with concomitant sling in the entire cohort of eligible patients.
Our sample size was one of convenience. Univariate and multivariate analyses were performed to identify risk factors for IE with an alpha level of .05. T-test, Mann-Whitney U, χ2, and Fisher exact tests were used as appropriate. A survival analysis was performed to compare resolution of IE over time between the groups of patients who had sling only, prolapse repair only, and both.
There were 1552 eligible patients among the 1842 charts identified for review, with a total of 475 patients who required postoperative catheterization for IE. Among the 1077 potential controls, 478 were randomly selected for full chart review and data abstraction (Fig. 1). The mean age of patients included in the case control study was 60.4 ± 12.8 years, and the population was racially diverse (Table 1).
On the basis of the CPT code query and chart review of the 1552 patients eligible for the study, 322 underwent sling only, 550 underwent prolapse repair only, and 680 underwent sling and prolapse repair. The overall incidence of any catheter use among patients eligible for the study was 475 (30.6%) of 1552; the incidence of any catheter use was significantly higher for patients receiving sling and prolapse repair (41.0%) compared with sling only (24.2%) and prolapse repair only (21.5%) (P < 0.0001) (Table 2). The overall incidence of IE more than 1 week was 77 (5.0%) of 1552; again, the incidence was significantly higher for sling and prolapse repair (7.5%) compared with sling only (3.7%) and prolapse repair only (2.5%) (P = 0.00019). The incidence of sling revision was 15 (1.5%) of 1002 (Table 3). The number of patients with IE fell sharply over first 7 to 10 postoperative days, then more gradually thereafter (Fig. 2), and there was no difference between the curves for different surgeries (P = 0.07).
There were no differences between cases of any catheter use and controls for age and race. Any postoperative catheter use was associated with higher uroflow PVR, lower uroflow maximum flow, less DO, and less urinary frequency (Table 3). Cases had more slings, anterior repairs, colpocleisis, and less total intravenous anesthesia (Table 4). There were no differences in type of sling, prolapse severity, or the proportion undergoing prolapse repair only. The median duration of catheter use was 4 days [interquartile range (IQR), 3–6].
Seventy-seven patients (5%) required catheter use longer than 1 week. Compared with controls, this subset of cases had significantly higher preoperative PVR (office, uroflow, and pressure flow), lower uroflow maximum flow, surgeon A, and sling use (Tables 3 and 4). Sling revision was performed in 15 cases and was associated with lower pressure flow maximum flow, stage 2 POP, hysterectomy, and uterosacral suspensions (Tables 3 and 4). The median time to sling revision was 79 days (IQR, 49–97.5), and 13 of 15 patients had resolution of IE after sling revision. No patients who passed void trial on hospital discharge underwent sling revision (negative predictive value, 100%).
Multivariate analysis was performed to determine odds of short- and longer-term catheter use. We adjusted for age and significant factors identified in the univariate analysis. Office PVR, uroflow maximum flow rate, sling, and anterior repair remained significantly associated with any catheter use. Uroflow maximum flow rate, DO, sling placement, and uterosacral suspension remained associated with IE more than 1 week (Table 5).
Thirteen patients developed IE after passing the hospital voiding trial (false negatives); there were 1077 true negatives, so the negative predictive value for any IE was 98%. In addition, 28 patients who were discharged with a catheter passed a repeat office void trial and then developed symptomatic IE. The median time to representation with IE after passing any void trial was 5 days (IQR, 3–7), 29 (70.7%) of 41 had retention lasting more than 1 week, and the median time to resolution of retention was 12 days (IQR, 7–18). No risk factors were identified on univariate analysis. Thirteen patients had severe urinary retention on representation with IE; the median duration of catheter use was 13 days (IQR, 8–17). Ten of these patients had undergone a sling, and none of them required sling revision.
This case-control study identified risk factors for any catheter use after surgeries for incontinence and prolapse (sling placement, higher office preoperative PVR, lower uroflow maximum flow, anterior repair, and general anesthesia), as well as risk factors for catheter use longer than 1 week (sling placement, higher uroflow PVR, lower uroflow maximum flow, with DO being protective). Case-control studies generally cannot provide estimates of incidence; however, because we reviewed all the charts in the eligible population to identify cases and the entire pool of potential controls, we know the denominator for the overall population. Therefore, we were able to determine that the incidence of any catheter use was 30.6%, incomplete emptying longer than 1 week was 5%, and sling revision was 1.5%. These findings are consistent with other published data.
Suburethral sling is generally considered to be the greatest risk factor for IE after pelvic floor surgery. Pubovaginal slings result in higher rates of voiding dysfunction compared with synthetic midurethral slings; however, given the low number of pubovaginal slings in our sample, we did not analyze them separately. In our study, patients who underwent sling with concomitant prolapse repair had higher incidences of overall catheter use and catheter use more than 1 week (41.0% and 7.5%, respectively) compared with patients who had only prolapse repair (21.5% and 2.5%, respectively) or only sling (24.2% and 3.7%, respectively). These are similar to findings from the outcomes following vaginal prolapse repair and mid urethral sling and TOMUS trials. In outcomes following vaginal prolapse repair and mid urethral sling, 42.6% of patients who had a sling and vaginal prolapse repair were discharged with a catheter for IE, compared with 30.0% of patients who had a prolapse repair without sling; at 2 weeks, the rate of catheter use was 3.7% and 0.6%, respectively.12 In TOMUS, 25% of patients were discharged with a catheter due to IE, and 6% required catheterization at 2 weeks postoperative.4
Association of preoperative urinary symptoms and urodynamic parameters with postoperative IE has not been consistently identified in the literature. Secondary analyses of 2 large randomized trials of incontinence surgeries did not find associations between urodynamic parameters and IE, but found preoperative voiding dysfunction symptoms to be predictive. In TOMUS, endorsement of accommodative voiding maneuvers, slow stream, and retropubic route of sling were predictors of IE after midurethral sling,4 and in SISTEr, hesitating urinary stream was predictive of IE after Burch or fascial sling.5,13 Observational studies have found that various urodynamic parameters, including lower uroflow maximum flow and higher PVR, are associated with postoperative IE,14,15 as well as anterior repair.14,16 The associations we found between preoperative measures of higher PVR and lower uroflow maximum flow rate were weak, but precise. The small effect size is in part due to the small units of measurement (10 mL and 10 mL/s increments, respectively) and likely has limited clinical relevance. A larger effect size would be demonstrated by increasing these increments by an order of magnitude, but we felt that these would not be physiologically relevant.
To our knowledge, DO is a novel association with postoperative IE. On the basis of our clinical experience, we did suspect that patients with characteristics of overactive bladder would have less risk of IE. Indeed, urinary frequency and DO were protective against any catheter use and IE more than 1 week in the univariate analysis, and DO remained a significant protector against IE more than 1 week on multivariate analysis. This aligns with our clinical experience and may be related to increased detrusor activity, resulting in better return of voiding function postoperatively.
Finally, our findings confirm that the postoperative void trial has an excellent negative predictive value for IE and sling revision. Patients who developed IE after passing a void trial had prolonged return to normal voiding function, but presentation with severe bladder distention did not seem to worsen their course or prognosis.
Our case-control design allowed for efficient investigation of the rare outcomes of IE more than 1 week and sling revision. The identification of all cases and eligible controls during the study period improved the quality of our study in several ways. We were able to determine incidence IE and sling revision overall and within different surgical groups, which will allow to provide more accurate counseling to our patients. Identification of the total pool of eligible controls and use of a random number generator to select those included in the case-control analyses should mitigate selection bias. In addition, we used chart review, not coding data, to more accurately identify IE. The main limitations of our study are those inherent to its retrospective design—reliance on data imputed to the electronic medical record, information bias, inability to account for patients lost to follow-up, and potential confounding. Finally, based on similar rates of urodynamics in each group, performance of this test does not appear to be associated with IE, but because not all of the patients in our study underwent urodynamics, we cannot draw any meaningful conclusions about its role in the prevention of prolonged postoperative IE. We also cannot evaluate the role of urodynamic testing in the evaluation of patients planning POP repairs with or without concomitant sling.
In summary, the we found that IE after prolapse and incontinence surgery was directly related to sling placement, lower maximum flow on uroflow, higher preoperative PVR, and anterior repair, and inversely related to DO (protective). Passing the hospital void trial is an excellent predictor of normal postoperative voiding function.
The authors acknowledge Eshetu Tefera, MS.
1. Wu JM, Matthews CA, Conover MM, et al. Lifetime risk of stress urinary incontinence
or pelvic organ prolapse
surgery. Obstet Gynecol
2. Book NM, Novi B, Novi JM, et al. Postoperative voiding dysfunction
following posterior colporrhaphy: female pelvic. Med Reconstr Surg
3. Partoll LM. Efficacy of tension-free vaginal tape with other pelvic reconstructive surgery. Am J Obstet Gynecol
4. Norton PA, Nager CW, Chai TC, et al.; Urinary Incontinence
Treatment Network. Risk factors for incomplete bladder emptying
after midurethral sling. Urology
5. Lemack GE, Krauss S, Litman H, et al. Normal preoperative urodynamic testing does not predict voiding dysfunction
after Burch colposuspension versus pubovaginal sling. J Urol
6. Sokol AI, Jelovsek JE, Walters MD, et al. Incidence and predictors of prolonged urinary retention after TVT with and without concurrent prolapse
surgery. Am J Obstet Gynecol
7. Nguyen JN, Jakus-Waldman SM, Walter AJ, et al. Perioperative complications and reoperations after incontinence
surgeries using prosthetic implants. Obstet Gynecol
8. Klutke C, Siegel S, Carlin B, et al. Urinary retention after tension-free vaginal tape procedure: incidence and treatment. Urology
9. Pham KN, Topp N, Guralnick ML, et al. Preoperative Valsalva voiding increases the risk of urinary retention after midurethral sling placement. Int Urogynecol J
10. Smith AL, Karp DR, Lefevre R, et al. LeFort colpocleisis and stress incontinence
: weighing the risk of voiding dysfunction
with sling placement. Int Urogynecol J
11. Kleeman S, Goldwasser S, Vassallo B, et al. Predicting postoperative voiding efficiency after operation for incontinence
. Am J Obstet Gynecol
12. Wei JT, Nygaard I, Richter HE, et al. A midurethral sling to reduce incontinence
after vaginal prolapse
repair. N Engl J Med
13. Sanses TV, Brubaker L, Xu Y, et al. Preoperative hesitating urinary stream is associated with postoperative voiding dysfunction
and surgical failure following Burch colposuspension or pubovaginal rectus fascial sling surgery. Int Urogynecol J
14. Chung SM, Moon YJ, Jeon MJ, et al. Risk factors associated with voiding dysfunction
surgery. Int Urogynecol J
15. Wang KH, Wang KH, Neimark M, et al. Voiding dysfunction
following TVT procedure. Int Urogynecol J Pelvic Floor Dysfunct
16. Hakvoort RA, Dijkgraaf MG, Burger MP, et al. Predicting short-term urinary retention after vaginal prolapse
surgery. Neurourol Urodyn
CPT codes used for chart search.