Catheter-associated urinary tract infection (UTI) is the most common nosocomial infection in the United States1,2 Each year, UTIs account for more than 8.1 million visits to health care providers3 and cost an estimated $1.6 billion in the United States alone.4 In the field of obstetrics and gynecology, UTI is one of the most common complications of pelvic reconstructive surgery, as is urinary retention.5–8 The risk of UTI in this population ranges from 5% to 35% and is significantly increased with exposure to postoperative catheterization.5,7–15 Urinary retention occurs in up to 50% of women undergoing pelvic floor surgery.13,16–18 Currently, most health care providers use transurethral catheterization to manage acute postoperative urinary retention.18,19 Although the duration of catheterization for postoperative urinary retention is usually relatively short, any exposure to transurethral catheterization can increase the risk of UTI.20 Routine antibiotic prophylaxis may be cost-effective, but only if antibiotic therapy does indeed significantly reduce the risk of UTI.21 Whether antibiotics are effective in patients receiving transurethral catheterization after pelvic reconstructive surgery is not currently known.
Given that pelvic floor disorders affect at least one-third of all women in the United States and the number of pelvic floor surgeries performed annually is projected to increase nearly 50% over the next 40 years,22,23 it is increasingly important to find ways to minimize health care costs and reduce postoperative complications for these women. Therefore, we conducted a randomized controlled trial to determine whether nitrofurantoin antibiotic prophylaxis as compared with placebo reduces the risk of postoperative UTI in patients receiving transurethral catheterization after pelvic reconstructive surgery.
MATERIALS AND METHODS
This was a randomized, double-blind, placebo-controlled trial of patients who received postoperative transurethral catheterization after undergoing surgery for pelvic organ prolapse (POP), urinary incontinence, or both by the Division of Urogynecology at Duke University Medical Center from August 2011 through February 2013. The primary outcome was treatment for clinically suspected or culture-proven UTI within 3 weeks after surgery. Approval for this study was obtained from the Duke University Medical Center institutional review board and written informed consent was obtained from all participants at enrollment. The study was registered at ClinicalTrials.gov (ID: NCT01450800).
Potential participants were identified, recruited, and enrolled after providing written informed consent during routine preoperative evaluation. We approached all patients undergoing surgery for pelvic floor disorders (ie, any POP repair, midurethral sling, or both). Patients were excluded if they were undergoing mesh excision, surgery for urethral diverticulum, fistula repair, or sacral neuromodulation. Patients were also excluded if they were younger than 21 years of age, pregnant, had an allergy to nitrofurantoin, had a creatinine clearance less than 60 mL/min, were dependent on catheterization to accomplish voiding preoperatively, or if they did not speak English. Randomization occurred within a few hours after surgery. Participants were excluded from the study before randomization if they sustained intraoperative injury to the urinary tract requiring prolonged postoperative catheterization or if they passed an immediate void trial on the day of surgery (because these participants did not receive additional postoperative catheterization).
Participants who received additional postoperative catheterization were randomized to placebo or nitrofurantoin. For study purposes, we defined additional postoperative catheterization to include all participants who were hospitalized postoperatively with an indwelling transurethral Foley catheter overnight as well as participants discharged on the day of surgery either with an indwelling transurethral Foley catheter or performing intermittent self-catheterization. The decision to place a catheter overnight was determined by surgical procedure and is generally done at our institution for any patient hospitalized overnight after pelvic reconstructive surgery with or without hysterectomy (ie, robotic sacrocolpopexy, uterosacral ligament suspension, vaginectomy or colpocleisis, or sacrospinous ligament fixation). Void trials were conducted through the backfill-assisted void trial method.16 Per our practice guidelines, the bladder was backfilled with 300 mL normal saline, the catheter was removed, the participant was prompted to void immediately, and a postvoid residual volume was measured after voiding was complete through either catheterization or bladder scan. The participant passed the void trial if the postvoid residual volume was less than 100 mL or if she voided more than half of the total voided volume with a voided volume more than 200 mL. Participants who failed their void trial before discharge were discharged home with either a Foley catheter or performing self-catheterization, depending on their preference. If discharged home performing self-catheterization, participants were instructed to continue to catheterize until residual volume was consistently less than 100 mL. Participants discharged with a Foley catheter were scheduled for void trial in the clinic 24–72 hours after discharge.
All participants received routine perioperative intravenous antibiotic prophylaxis with cefazolin dosed by body weight (or an appropriate alternative if allergic) administered within 1 hour of skin incision and redosed every 4 hours while in surgery. At the start of surgery, an indwelling Foley catheter was routinely placed using a sterile technique after administration of perioperative antibiotics. After surgery, eligible participants receiving additional postoperative catheterization were randomized to either one tablet of masked 100 mg nitrofurantoin or matching placebo by mouth daily starting on the morning of postoperative day 1 and continuing each day they received catheterization for up to 7 days. Those requiring catheterization for greater than 7 days were started on prophylactic antibiotics on postoperative day 7 per our practice protocol. Randomization was controlled by the Duke University Medical Center Investigational Drug Services Pharmacy through random permutated blocks of 10 participants each. The study was double-blinded such that both study participants and health care providers were masked to respective treatment allocation and randomization block size for the duration of the study. At the time of randomization, masked study drug was dispensed by the Duke University Medical Center Investigational Drug Services Pharmacy to the study participant with each participant receiving a vial of seven unmarked study drug capsules. Participants were instructed to complete a 1-week catheterization and antibiotic diary starting on postoperative day 1 to record on a daily basis the type of catheter used and any antibiotic medication (including study drug) taken. Participants were instructed to return all unused study medications at or before their 6-week postoperative visit.
The primary outcome was treatment for UTI within the first 3 weeks after surgery. Treatment for UTI was defined to include any treatment received for clinically suspected or culture-proven UTI within 3 weeks of surgery. Clinically suspected treatment was defined to include treatment given empirically on development of urinary symptoms or prescribed based on urine test results. Culture-proven UTI was defined as a urine culture with greater than 100,000 colony-forming units of a single organism. Participants were instructed to contact our practice for therapy and evaluation if they were concerned they had a UTI. In addition, they were provided forms to submit to outside health care providers if they were evaluated at an outside facility and instructed to have those forms completed and faxed to our office. These forms explained their involvement in a research study and contained information for the health care provider to complete regarding any urine testing results obtained or antibiotic treatment prescribed. Our urogynecology clinic nurses routinely contacted all participants postoperatively to assess their progress and ask about symptoms of UTI per our practice protocol. These telephone encounters are documented in the medical record. In these instances, if symptoms of UTI were detected, empiric treatment for UTI may have been started per clinic protocol. Participants who called the urogynecology triage nurse or after-hours physician on-call may also have received empiric treatment per protocol and these telephone encounters are also documented in the medical record. Treatment for UTI in the postoperative period was queried at the participant's postoperative visit at 4 to 6 weeks after surgery. Thus, we used multiple methods to ascertain whether empiric or test-based treatment for UTI occurred in the first 3 weeks after surgery. Planned secondary outcomes were risk factors for treatment for postoperative UTI and prevalence of bacterial resistance to nitrofurantoin on postoperative urine cultures.
Under the assumption of 80% power and an α error of 0.05, we estimated we would need 156 participants to demonstrate a two-thirds reduction in risk of UTI (ie, from 30% to 10%). We chose this parameter based on the fact that a previous study in a similar patient population at our institution showed that risk of UTI in patients receiving catheterization to be 30%24 and, according to the decision analysis by Sutkin et al,21 prophylactic antibiotics are favored if they significantly reduce risk of UTI21 (ie, from 30% to 10%). Because it is uncertain who would require postoperative catheterization and qualify for randomization, we aimed to enroll a minimum of 350 participants preoperatively with a goal of randomizing 156 participants since the rate of postoperative catheterization in the same previous retrospective study conducted at our institution was approximately 50%.24
Categorical variables were analyzed using χ2 or Fisher's exact test; continuous variables were analyzed using Student's t tests. Duration of catheterization and parity were compared using the Mann-Whitney U test for difference in medians. A P value <.05 defined statistical significance. Logistic regression was used to control for potential confounding variables, which were identified by bivariate analyses to determine variables independently associated with postoperative UTI. Statistical analysis was by intention to treat. All statistical analysis was performed using SPSS 20.0 for Windows. Data are presented according to CONSORT guidelines for randomized trials.25
From August 2011 through February 2013, 508 patients were eligible for participation and 37 (74%) provided informed consent to participate (Fig. 1). We subsequently excluded 212 participants based on the following exclusion criteria: 1) passed void trial on day of surgery; 2) declined participation before or on the day of surgery; or 3) had a creatinine clearance less than 60 mL/min. One hundred sixty-three participants were randomized, 82 to nitrofurantoin and 81 to placebo. Of those randomized, there were no adverse events reported from the nitrofurantoin or placebo arms. Four randomized participants were excluded from the final analysis. One in the placebo group was excluded because she had previously been included in the study during a prior surgery. Three other participants were excluded as a result of protocol deviations (one in the nitrofurantoin group because she was on antibiotic therapy for preoperative UTI and two in the placebo group, one with creatinine clearance less than 60 mL/min and one on chronic nitrofurantoin therapy for recurrent UTIs). Of note, when running analyses for primary outcome, the inclusion and exclusion of these participants did not alter the study results (unpublished data). Similarly, if only the participants discharged to home with catheterization (ie, participants discharged after failed void trial either on day of surgery or after at least 1 day of hospitalization) were included, the results of the primary outcome analyses were again unchanged (unpublished data). The final analysis included a total of 159 participants (81 nitrofurantoin, 78 placebo). There were no significant differences between the two groups in baseline or perioperative characteristics, including age, menopausal status, history of recurrent UTIs, preoperative UTI treatment, diabetes, baseline postvoid residual volume, midurethral sling, intraoperative complications, and overnight stay, except for a lower rate of hysterectomy in the nitrofurantoin group (36% in nitrofurantoin compared with 60% in placebo, P<.01; Table 1). Additionally, there were no differences in duration or type of catheterization (Table 2).
The primary outcome, treatment for UTI within 3 weeks of surgery, was seen in 18% (28/159) of all participants and there was no significant difference between the two study groups. Specifically, there were 18 (22%) participants treated for UTI in the nitrofurantoin group compared with 10 (13%) in the placebo group (relative risk 1.73, 95% confidence interval [CI] 0.85–3.52, P=.12; Table 2). The median time from surgery to treatment was 8 days (interquartile range 5–11). Of those who were treated for UTI, 19 (68%) participants had urine testing (urine culture with or without urinalysis), and nine (32%) participants were treated empirically. Urine culture results were mixed but many had positive results, with eight (42%) more than 100,000/colony-forming unit, two (11%) 1,000–100,000 bacteria/colony-forming unit, five (26%) mixed flora, and four (21%) no growth. Of the eight cultures with more than 100,000 bacteria/colony-forming unit, half (50%) had Klebsiella pneumoniae. The other pathogens identified included Escherichia coli, Proteus mirabilis, Klebsiella oxytoca, and Enterococcus species. On susceptibility testing, only one of nine (11%) urine cultures showed resistance to nitrofurantoin with five (56%) specimens being susceptible and three (33%) being inconclusive. Of participants with no growth or mixed flora on urine culture results, most (78%) had suspicious urinalysis or urine dipstick results and, thus, were treated before urine culture results returned.
Bivariate analyses were performed to evaluate risk factors for UTI (Table 3). Treatment for UTI was higher in premenopausal women (29% UTI in premenopausal compared with 12% in postmenopausal, P=.01), lower in diabetics (0% in diabetics compared with 20% in nondiabetics, P=.04), and was significantly associated with duration of catheterization (median 1 day, interquartile range 1–3 for no UTI compared with 2 days, interquartile range 1–4 for UTI, P=.03). Notably, hormone therapy, smoking, history of UTI, severity of prolapse, preoperative postvoid residual volume, creatinine clearance, operative time, estimated blood loss, procedure, type of catheterization, and overnight stay were not significantly associated with treatment for UTI (P>.05 for all). Medication adherence, defined by number of pills taken divided by number of days catheterized, was equal to or greater than 1.0 (“good” adherence) in 88% (96/109) of participants with complete medication data. In these participants, “good” adherence was not associated with either treatment allocation (P=.78) or postoperative UTI treatment (P=.25). Notably, of the participants in the final analysis, eight (two randomized to nitrofurantoin and six randomized to placebo) received additional postoperative antibiotic therapy during the study period. Two received antibiotics for passage of a stitch through rectal tissue during uterosacral ligament suspension, one for a concomitant orthopedic procedure, one for postoperative pneumonia, one for postoperative fever, one for postoperative sinusitis, one for bacterial vaginosis, and one participant started trimethoprim after completing catheterization. A subanalysis (unpublished data) excluding these patients did not change the results of the primary outcome.
A logistic regression model was constructed to evaluate the relationship between prophylactic antibiotics and UTI while controlling for confounding factors. When adjusting for menopausal status, diabetes, preoperative postvoid residual volume, creatinine clearance, hysterectomy, and duration of catheterization, there was still no difference in risk of UTI in participants receiving nitrofurantoin compared with placebo (adjusted odds ratio 1.27, 95% CI 0.38–4.27, P=.70).
Our study demonstrates that antibiotic prophylaxis with 100 mg nitrofurantoin daily during catheterization does not reduce the risk of postoperative UTI in patients receiving short-term transurethral catheterization after pelvic reconstructive surgery. Prior studies published on this topic have important methodologic differences, which likely account for the variability of findings. In 1992 van der Wall et al26 revealed that in men and women undergoing orthopedic, colorectal, or vaginal surgery with planned postoperative catheterization for 3 or more days, prophylaxis with daily ciprofloxacin significantly reduced the risk of postoperative UTI. In 2004, Rogers et al27 performed a multicenter randomized trial examining antibiotic prophylaxis in women undergoing surgery for POP or stress incontinence who received suprapubic catheterization for a median of 10–11 days and found that daily nitrofurantoin significantly decreased the risk of symptomatic UTI during the 6- to 8-week postoperative period. The differences in duration and type of catheterization in these studies change the baseline risk of UTI. Our prevalence of UTI was likely lower because of our shorter observation period and shorter duration of catheterization. In a setting where the baseline prevalence of UTI is lower, antibiotic prophylaxis may no longer reduce the risk of UTI.
Jackson et al15 demonstrated that, in patients undergoing midurethral sling with and without postoperative catheterization, 100 mg therapeutic nitrofurantoin twice daily or placebo for 3 days postoperatively lowered the risk of UTI during the first 6 weeks after surgery. However, the most significant risk factor for UTI in their study was postoperative catheterization, not antibiotic therapy. The authors did not publish results on the risk of UTI in the subset of patients receiving catheterization so we are not able to determine whether nitrofurantoin treatment differentially decreased the risk of UTI in this subset.
Our observed rate and duration of catheterization are similar to other randomized trials, which report voiding function usually returns within the first few days or less after pelvic reconstructive surgery.12,13,16–18 Our risk factors for UTI were similar to those seen in other studies, with the exception of risk of UTI being lower in diabetics as compared with nondiabetics. One possible explanation for this finding is that diabetic patients experienced fewer symptoms in the short term and, therefore, were less likely to present for evaluation during the observation period.
Strengths of our study include the prospective randomized double-blind, placebo-controlled trial design, large sample size, and the use of a study population undergoing transurethral catheterization. Currently the majority of health care practitioners use transurethral catheterization making these results readily applicable to clinical practice.18,19 By defining UTI as treatment received for clinically suspected or culture-proven infection, we mimic standard practice and improve the generalizability of our results. This same definition has been used successfully in many other trials, including the trial by Jackson et al15 as well as the Trial Of Mid-Urethral Slings and the Outcomes following vaginal prolapse repair and mid Urethral Sling and SISTEr randomized trials.8,11,15,28 Although UTI can be readily diagnosed with symptoms at presentation in normal patient populations,29 in our study, not all participants who received UTI treatment had positive urine culture results. Therefore, in the postoperative setting, patients receiving catheterization may present differently and there is a need to develop more accurate modalities to better predict infection.
Limitations of our study include recall bias, which was minimized by using a medication and catheterization diary and verifying patient-reported information with multiple strategies, and selection bias, which was limited by demographic data and a high enrollment rate. Participants with a creatinine clearance less than 60 mL/min were excluded, because this is a contraindication to nitrofurantoin, which led to the exclusion of many patients older than 75–80 years of age. Thus, our findings have limited applicability to elderly patients undergoing pelvic reconstructive surgery. Additionally, our observed UTI risk of 18% overall and 22% in the nitrofurantoin group is lower than the 30% risk used in our power calculation that was based on our previous retrospective study.24 We would have needed 300 participants per arm to be adequately powered to show a significant reduction in risk of UTI from 22% (observed risk in our treatment group) to 13% (observed risk in our placebo group). Our study population includes patients undergoing robotic surgery who likely carry a different UTI risk profile than the exclusively vaginal surgery population examined in our previous study. Notably, we found the nitrofurantoin group had a higher risk of UTI compared with placebo indicating that nitrofurantoin prophylaxis most likely would still not significantly reduce the risk of UTI as compared with placebo, even if a larger study population were to be examined.
Urinary tract infection is a significant health care issue in patients undergoing reconstructive pelvic surgery, especially in patients who receive postoperative catheterization. In an effort to reduce the risk of UTI, health care practitioners often prescribe antibiotics; however, our results suggest this is not an effective strategy for patients receiving short-term postoperative transurethral catheterization. Alternate methods to effectively reduce the risk of UTI may be to continue prophylaxis for several days after catheterization or to use a different (ie, bactericidal not bacteriostatic) antibiotic for prophylaxis. Further investigation is warranted to reduce exposure to unnecessary antibiotic treatments, help improve patient outcomes, and reduce health care costs.
1. Schumm K, Lam TB. Types of urethral catheters for management of short-term voiding problems in hospitalized adults: a short version Cochrane review. Neurourol Urodyn 2008;27:738–46.
2. Stamm WE. Catheter-associated urinary tract infections: epidemiology, pathogenesis, and prevention. Am J Med 1991;91:65S–71S.
3. Schappert SM, Rechtsteiner EA. Ambulatory medical care utilization estimates for 2006. Natl Health Stat Report 2008:1–29.
4. Foxman B. Epidemiology of urinary tract infections: incidence, morbidity, and economic costs. Am J Med 2002;113(suppl 1A):5S–13S.
5. Chung CP, Kuehl TJ, Harris SK, McBride MM, Larsen WI, Yandell PM, et al.. Incidence and risk factors of postoperative urinary tract infection after uterosacral ligament suspension. Int Urogynecol J 2012;23:947–50.
6. Nygaard I, Barber MD, Burgio KL, Kenton K, Meikle S, Schaffer J, et al.. Prevalence of symptomatic pelvic floor disorders in US women. JAMA 2008;300:1311–6.
7. Sutkin G, Alperin M, Meyn L, Wiesenfeld HC, Ellison R, Zyczynski HM. Symptomatic urinary tract infections after surgery for prolapse and/or incontinence. Int Urogynecol J 2010;21:955–61.
8. Wei JT, Nygaard I, Richter HE, Nager CW, Barber MD, Kenton K, et al.. A midurethral sling to reduce incontinence after vaginal prolapse repair. N Engl J Med 2012;366:2358–67.
9. Fok CS, McKinley K, Mueller ER, Kenton K, Schreckenberger P, Wolfe A, et al.. Day of surgery urine cultures identify urogynecologic patients at increased risk for postoperative urinary tract infection. J Urol 2013;189:1721–4.
10. Ingber MS, Vasavada SP, Firoozi F, Goldman HB. Incidence of perioperative urinary tract infection after single-dose antibiotic therapy for midurethal slings. Urology 2010;76:830–4.
11. Nygaard I, Brubaker L, Chai TC, Markland AD, Menefee SA, Sirls L, et al.. Risk factors for urinary tract infection following incontinence surgery. Int Urogynecol J 2011;22:1255–65.
12. Schiotz HA, Malme PA, Tanbo TG. Urinary tract infections and asymptomatic bacteriuria after vaginal plastic surgery. A comparison of suprapubic and transurethral catheters. Acta Obstet Gynecol Scand 1989;68:453–5.
13. Schiotz HA, Tanbo TG. Postoperative voiding, bacteriuria and urinary tract infection with Foley catheterization after gynecological surgery. Acta Obstet Gynecol Scand 2006;85:476–81.
14. Stekkinger E, van der Linden PJ. A comparison of suprapubic and transurethral catheterization on postoperative urinary retention after vaginal prolapse repair: a randomized controlled trial. Gynecol Obstet Invest 2011;72:109–16.
15. Jackson D, Higgins E, Bracken J, Yandell PM, Shull B, Foster RT, et al.. Antibiotic prophylaxis for urinary tract infection after midurethral sling: a randomized controlled trial. Female Pelvic Med Reconstr Surg 2013;19:137–41.
16. 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.
17. Hakvoort RA, Thijs SD, Bouwmeester FW, Broekman AM, Ruhe IM, Vernooij MM, et al.. Comparing clean intermittent catheterisation and transurethral indwelling catheterisation for incomplete voiding after vaginal prolapse surgery: a multicentre randomised trial. BJOG 2011;118:1055–60.
18. Hakvoort RA, Burger MP, Emanuel MH, Roovers JP. A nationwide survey to measure practice variation of catheterisation management in patients undergoing vaginal prolapse surgery. Int Urogynecol J Pelvic Floor Dysfunct 2009;20:813–8.
19. Phipps S, Lim YN, McClinton S, Barry C, Rane A, N'Dow J. Short term urinary catheter policies following urogenital surgery in adults. The Cochrane Database of Systematic Reviews 2006, Issue 2. Art. No.: CD004374. DOI: 10.1002/14651858.CD004374.pub2.
20. Healy EF, Walsh CA, Cotter AM, Walsh SR. Suprapubic compared with transurethral bladder catheterization for gynecologic surgery: a systematic review and meta-analysis. Obstet Gynecol 2012;120:678–87.
21. Sutkin G, Lowder JL, Smith KJ. Prophylactic antibiotics to prevent urinary tract infection during clean intermittent self-catheterization (CISC) for management of voiding dysfunction after prolapse and incontinence surgery: a decision analysis. Int Urogynecol J Pelvic Floor Dysfunct 2009;20:933–8.
22. Bump RC, Norton PA. Epidemiology and natural history of pelvic floor dysfunction. Obstet Gynecol Clin North Am 1998;25:723–46.
23. Wu JM, Kawasaki A, Hundley AF, Dieter AA, Myers ER, Sung VW. Predicting the number of women who will undergo incontinence and prolapse surgery, 2010 to 2050. Am J Obstet Gynecol 2011;205:230.e1–5.
24. Dieter AA, Amundsen CL, Visco AG, Siddiqui NY. Treatment for urinary tract infection after midurethral sling: a retrospective study comparing patients who receive short-term postoperative catheterization and patients who pass a void trial on the day of surgery. Female Pelvic Med Reconstr Surg 2012;18:175–8.
25. Schulz KF, Altman DG, Moher D. CONSORT 2010 statement: updated guidelines for reporting parallel group randomized trials. Obstet Gynecol 2010;115:1063–70.
26. van der Wall E, Verkooyen RP, Mintjes-de Groot J, Oostinga J, van Dijk A, Hustinx WN, et al.. Prophylactic ciprofloxacin for catheter-associated urinary-tract infection. Lancet 1992;339:946–51.
27. Rogers RG, Kammerer-Doak D, Olsen A, Thompson PK, Walters MD, Lukacz ES, et al.. A randomized, double-blind, placebo-controlled comparison of the effect of nitrofurantoin monohydrate macrocrystals on the development of urinary tract infections after surgery for pelvic organ prolapse and/or stress urinary incontinence with suprapubic catheterization. Am J Obstet Gynecol 2004;191:182–7.
28. Albo ME, Richter HE, Brubaker L, Norton P, Kraus SR, Zimmern PE, et al.. Burch colposuspension versus fascial sling to reduce urinary stress incontinence. N Engl J Med 2007;356:2143–55.
29. Bent S, Nallamothu BK, Simel DL, Fihn SD, Saint S. Does this woman have an acute uncomplicated urinary tract infection? JAMA 2002;287:2701–10.