Longer Preoperative Antibiotic Duration Prior to High-Risk Ureteroscopy Does Not Decrease Infectious Complications

Introduction

As the prevalence of nephrolithiasis continues to rise, ureteroscopy for surgical treatment of these stones is becoming one of the most commonly performed procedures in urology. ^{1 –3} This procedure boasts an excellent safety profile with complication rates reported to be 9%–25%, the majority of which are minor in nature. ⁴ Postoperative infectious complications including urinary tract infection (UTI), systemic inflammatory response syndrome (SIRS), and sepsis are uncommon but potentially devastating complications. ^{5 –8}

The optimal perioperative management of an asymptomatic positive urine culture or those at high risk of urinary tract colonization remains a quandary for practicing urologists. The American Urologic Association (AUA) and the European Association of Urology recommend treatment of symptomatic UTIs; however, they do not specifically discuss asymptomatic cases. ^9,10 As a result, there is significant practice variation in antibiotic prophylaxis, even among fellowship-trained endourologists, with some providers prescribing over 10 days of preoperative antibiotics and others not administering any. ¹¹

This is an important problem that affects a patient population already at risk. Prior literature has shown that patient characteristics associated with bacterial colonization, such as the presence of an indwelling foreign body, paraplegia, and bowel interposition for urinary diversion, are significant risk factors for postoperative infectious complications after ureteroscopy. ^5,6,12 These patients are also generally more frail and multimorbid with a higher risk of multi-drug-resistant bacteria. ^13,14 Determining the optimal preoperative antibiotic course for asymptomatic high-risk patients could not only improve surgical outcomes but also facilitate antimicrobial stewardship in this patient population.

To address this critical gap in knowledge, we designed this retrospective study to evaluate the relationship between preoperative antibiotic duration and postoperative ureteroscopy infectious outcomes in both high- and low-risk patients.

Materials and Methods

Starting in November 2015, the University of California San Francisco Urology Clinic has been enrolling new patients with urinary stone disease into a prospective registry called the Registry for Stones of the Kidney and Ureter (ReSKU). ¹⁵ This registry is seamlessly integrated with the electronic medical record system to collect and store patient data specifically related to kidney stone disease. The data points are derived from the electronic health record notes created by healthcare providers during clinic visits, surgical encounters, and follow-up visits as part of their regular care. These data are automatically integrated and stored in the registry database. The methodology and structure of this registry have been previously described. All individuals participating in ReSKU provide written consent, and the study has received approval from the Committee on Human Research (Protocol 14-14533).

We performed a retrospective review of consecutive patients who underwent ureteroscopy for any reason (i.e., treatment of nephrolithiasis, diagnostic, ablation or biopsy of upper tract urothelial carcinoma, or management of stricture disease). Patient demographic characteristics, past medical history, operative characteristics, and postoperative outcomes were also abstracted. High-risk patients (Table 1) were defined as those who had a history of neurological insult leading to paraplegia/quadriplegia, indwelling foley catheter, suprapubic catheter, clean intermittent catheterization, bowel interposition, ureteral stent, or nephrostomy tube placed prior to definitive surgical management, a positive, treated, symptomatic urine culture or asymptomatic urine culture within 60 days of operation. These factors were previously used to identify stone patients at moderate to high risk of postoperative infectious complications for inclusion in a randomized clinical trial. ¹⁶

Pre- and postoperative prophylactic antibiotic durations were categorized as none, less than 3 days, 3–10 days, and more than 10 days. Preoperative prophylactic antibiotics were administered until the day of operation at which time the patient received preoperative antibiotics within 60 minutes of operation. Patients categorized as having received no preoperative antibiotics all received preoperative antibiotics. Duration of indwelling ureteral stents and nephrostomy tubes was categorized as less than 30 days, 30–90 days, and more than 90 days. Preoperative data including urine culture results and prescription of antibiotics were also obtained. A positive urine culture was considered speciation of >100,000 colony-forming units of a specific bacteria or mixed urogenital flora.

The primary outcome was infectious complications occurring within 30 days of the index surgery defined as the occurrence of symptomatic UTI, pyelonephritis, SIRS, or sepsis from a urinary origin. These postoperative outcomes were captured on the day of operation if postoperative admission was indicated or at the time of postoperative follow-up or in the review of the patients’ charts if there was no follow-up visit. Other complications captured included lower urinary tract symptoms, urinary retention, hematuria, flank pain, or reoperation for stent placement.

Students’ t-tests were used for continuous variables and chi-square analysis for categorical variables. Multivariable logistic regression was used to identify predictors of infectious complications and included in the model were age, sex, race, BMI, history of type two diabetes, presence of high-risk comorbidities, operative time, duration of stent or nephrostomy tube dwell time, duration of antibiotics preoperatively, and duration of antibiotics postoperatively. We performed a power analysis to determine whether our cohort was appropriately sized for this analysis. With a noninferiority limit set to 1.2, a sample size of 56 would be required to have 80% power and alpha at a level of 0.05 to evaluate a 10% difference in infectious outcome incidence between groups. All statistics were performed using SPSS™ v27.

Results

A total of 405 patients met our inclusion criteria. These patients were stratified by preoperative prophylactic antibiotic duration with 68% receiving none, 11% receiving <3 days, 11% receiving 3–10 days, and 10% receiving >10 days (Fig. 1). The presence of a positive urine culture (68% received preoperative prophylaxis) or two or more risk factors (58% received preoperative prophylaxis) was associated with higher utilization of preoperative prophylactic antibiotics, whereas longer stent or nephrostomy tube dwell time was not (21% received preoperative prophylaxis).

OPEN IN VIEWER

FIG. 1.

Duration of preoperative prophylactic antibiotics in all high-risk patients, only those with positive preoperative urine cultures, long tube durations or more than one high-risk factor.

There was a higher proportion of women who received preoperative prophylactic antibiotics (Table 1, p = 0.005). In addition, longer durations of antibiotic treatment were favored in those with a history of urinary diversion, the presence of a ureteral stent or nephrostomy tube, and those with positive urine cultures (all p < 0.001). On review of operative characteristics (Table 2), the mean stone size was highest in those who received <3 days of preoperative prophylactic antibiotics (37 mm ± 42 mm, p = 0.01). Access sheaths were also more commonly utilized in this group (85%, p = 0.003). There were no differences in postoperative stent removal timing or operative time. Postoperative antibiotics were prescribed in 17% of those who had no preoperative antibiotics and in 41% of those who had preoperative antibiotics.

The overall rate of infectious complication was 14%. Postoperative UTI/pyelonephritis was found in 9%, whereas the remaining 5% experienced SIRS/sepsis. We performed a sensitivity analysis by comparing the rates of infectious complications in low-risk patients (those without high-risk features). These 509 patients demonstrated an overall infectious complication rate of 7.3%, which was significantly lower than that of the high-risk group (p = 0.004). The sepsis rate was similarly lower at 1.4% though not statistically significant (p = 0.220). On subanalysis of complication rates by risk factors, there were no significant differences in the occurrence of either UTI/pyelonephritis (Fig. 2A) or SIRS/sepsis (Fig. 2B). Among those who did experience a postoperative complication, there was no difference in the rates of preoperative prophylactic antibiotic duration (Fig. 3).

OPEN IN VIEWER

FIG. 2.

Postoperative rates of (A) urinary tract infection or pyelonephritis and (B) sepsis or systemic inflammatory response syndrome.

OPEN IN VIEWER

FIG. 3.

Postoperative complications and receipt of preoperative prophylactic antibiotics.

Predictors of postoperative UTI/pyelonephritis and postoperative SIRS/sepsis were then modeled in two separate multivariable logistic regressions that included age, sex, race, BMI, presence of type 2 diabetes, presence of other high-risk comorbidities, operative time, use of a ureteral access sheath, urine culture status, duration of stent or nephrostomy tube, duration of preoperative prophylactic antibiotics, and duration of postoperative prophylactic antibiotics (Table 3). Women were associated with a 4.135 (odds ratio (OR): 1.768–9.669, p = 0.001) increased odds of UTI compared with men. In both models, the duration of preoperative prophylactic antibiotics was not predictive of infectious complications.

Discussion

Our retrospective study of 405 high-risk ureteroscopies performed at a tertiary care center showed that a longer duration of preoperative antibiotics did not improve the postoperative infectious complication rate, even in those at high risk of urinary tract colonization or who had asymptomatic bacteriuria. This finding persisted in subanalysis of those with multiple high-risk factors. These findings suggest that patients with urinary diversion, neurological insults, and indwelling foreign bodies are at high risk of urinary tract colonization and that treating them with long courses of preoperative prophylactic antibiotics does not lead to improved postoperative outcomes.

There is no consensus regarding preoperative prophylactic antibiotic administration for patients undergoing stone surgery at high risk for postoperative infectious complications, even among members of the endourological society. For low-risk ureteroscopy (URS) patients, single-dose prophylactic antibiotic regimen has been supported ^17,18 ; however, neither the AUA nor European Association of Urology (EAU) addresses high-risk patients or asymptomatic bacteriuria. ^9,10,19 The Infectious Diseases Society of America also has released guidelines on asymptomatic bacteriuria management before endourological operation ²⁰ and recommends a short (1 or 2 dose) course rather than prolonged therapy within 30–60 minutes of operation. These guideline statements however are based largely on cohorts of patients undergoing lower urinary tract procedures such as transurethral resection of prostate.

As a result, even among members of the Endourologic Society, there is a high degree of variability in preoperative antibiotic prescribing patterns. ¹¹ A survey distributed to members of this society included a clinical vignette of a patient with asymptomatic positive urine culture before URS. Roughly 25% responded with 1–3 days of antibiotics, 45% with 4–7 days, and approximately 15% with >7 days or perioperative antibiotics only. This distribution of antibiotic duration was reflected in our sample as well as at the time of the study, we did not have a standardized antibiotic prescription protocol. This variability in prescribing guidelines also occurs across many urological procedures in clinical practice, and a portion of that variability has been attributed to ambiguity surrounding how to account for various patient risk factors in determining optimal antimicrobial prophylaxis. ²¹

Our definition of a high-risk group was in part based on the Endourological Disease Group for Excellence (EDGE) consortium study on preoperative prophylactic antibiotics before percutaneous nephrolithotomy (PCNL) in high-risk patients. ¹⁶ This group defined high risk as the presence of indwelling foreign bodies or positive preoperative urine culture within 3 months of procedure. ¹⁶ Patients in this high-risk group were found to be likely to have positive urine cultures preoperatively which supports the utility in identifying patients at high risk of urinary tract colonization. For high-risk patients, they found that 2-day preoperative antibiotic prophylaxis was associated with an increased risk of postoperative sepsis relative to 7-day antibiotic prophylaxis while for low-risk patients, there was no benefit in the use of 1-week preoperative antibiotic prophylaxis when compared with single-dose prophylaxis. ²²

The presence of an indwelling foreign body including foley catheters, ureteral stents and nephrostomy tubes, urinary diversion, and paraplegia has been shown to be risk factors not only for urinary tract colonization but also for postoperative infectious complications after ureteroscopy. ^5,6,12 Biofilm formation in the case of indwelling foreign bodies ⁷ may act as a scaffold for microorganism colonization. Indwelling bladder catheterization may lead to colonization as soon as 1–3 days after insertion and by 1 month, almost all patients will have bacteriuria. ²³ Ureteral stents and nephrostomy tubes have been implicated in infections after ureteroscopy. ^12,24 Bowel interposition for urinary diversion or augmentation leads to urinary tract colonization and is associated with not only postoperative infection but also stone growth. Paraplegia, spinal cord injury, and other neurological insults leading to poor bladder emptying have been implicated in postoperative infections after ureteroscopy as well. Our data support both these patients with higher risk of having a positive urine culture and with longer tube durations, there was a significantly higher rate of infectious complications on univariate analysis. In practice, urologists recognize the need to mitigate these patients’ risk of infectious complications however because of the absence of randomized data or guidance from national bodies, there is substantial variance in preoperative antibiotic regimens prescribed to patients across providers.

There remains a critical gap in knowledge regarding duration of preoperative antimicrobial prophylaxis and its implications for URS. To our knowledge, there is only one randomized study on this subject which was published recently. ²⁵ The authors randomized patients into enhanced prophylaxis which included IV ceftriaxone for 48 hours before and 24 hours after operation in those with mixed cultures and 5 days preoperatively and 2 days postoperatively of culture-directed antibiotics in those with one or two isolates. For the comparison group, all received IV fluoroquinolones preoperatively and oral antibiotics for 24 hours postoperatively. Their overall complication rate was 21.9% with 15.6% UTI rate in the standard group vs 6.3% in the enhanced group (p = 0.047) on univariate analysis; however, no statistically significant difference on multivariable analysis when considering other clinical factors. The generalizability of this study may be limited as coordination of IV antibiotics in the outpatient setting may be difficult and this study included postoperative antibiotics in all patients. Moreover, this trial did not include details on clinical characteristics such as the presence of urinary diversion, neurological disorders, or foley catheters which may have affected their outcome. Nonetheless, the authors showed a seemingly decreased UTI rate after relatively aggressive antibiotics regimens pre- and postoperatively; however, this was ultimately not statistically significant on multivariable analysis.

Limiting antibiotic exposure has the potential to combat the growing rates of multidrug resistance urinary tract pathogens, especially in vulnerable populations. Uncontrolled antibiotic use has already led to high rates of MDR bacteria leading to reported prevalence as high as 70% in certain countries. ²⁶ In endourology, MDR organisms have been reported in up to 30%–33% of preoperative urine cultures and are an independent predictor of postoperative infectious complications even after appropriate preoperative treatment. ^13,14 The most important measure to help prevent MDR infections is the creation and implementation of a culture of antimicrobial stewardship; however, as a field, urology has an overall poor knowledge of appropriate antimicrobial stewardship highlighting the critical need for further work in this area. Although the data for limited antibiotic utilization in low-risk patients are compelling, ^17,18 high-quality, randomized data are required to lay the groundwork for standardized protocols and guidelines in high-risk patients.

What then is an alternative hypothesis for these findings? Pyelovenous backflow, because of elevations in intrarenal pressure (IRP), has long been hypothesized to be related to infectious complications since the original investigations were reported by Himan et al. ²⁷ An IRP threshold of roughly 40 cmH₂O/30 mmHg generally become the accepted threshold during endourological procedures. ²⁸ Although IRP is known to vary because of access sheath diameter, irrigation method, and instrument presence ex vivo because of technical limitations of obtaining this in vivo during ureteroscopy, the relevance of this threshold has not been well studied. Compared with PCNL patients studied by the EDGE consortium, IRP may be a more significant and as of yet unmeasured risk factor during ureteroscopy. Future studies will likely explore this possibility. In addition, stones colonized by bacteria may not be sufficiently treated with oral or IV antibiotics. The use of antibiotic irrigation at the time of ureteroscopy may reduce infectious complications in these patients. ²⁹

Our study has several important limitations. First, the retrospective nature of the study leads to confounding and selection bias. As a prospectively collected registry, however, the data collected on our patients were of high quality and systematically recorded. Although it is possible that some data could be missed or misrepresented in our dataset, all charts were manually reviewed to determine complication rates, urine culture results, and antibiotic utilization limiting these issues. It is also possible that these high-risk patients were given antibiotics regimens by local providers or bypassed our system to present to local emergency departments when acutely ill postoperatively leading to missed postoperative encounters. However, our infectious complication rate was similar to that of prior studies ^5,6,25 suggesting that if these occurred, they were rare events. Second, we do not routinely perform a repeat urine culture as a test of cure prior to ureteroscopy in patients with asymptomatic bacteriuria. Our registry also does not collect stone culture data and in our practice, we do not routinely capture stone cultures during ureteroscopy. Inadequate treatment of bacteria may have impacted our rates of sepsis as patients with positive stone cultures, even with negative preoperative cultures, have been reported to have higher rates of sepsis likely because of untreated bacteria. ³⁰ However, we would expect then the risk of sepsis to be higher in the short antibiotic duration group which we did not observe. In addition, we regularly employ the dusting technique, which has become increasingly popular with the proliferation of high-powdered lasers, in this setting, obtaining an adequate stone sample for culture may become more difficult without concerted effort. Third, we did not capture information on indications for preoperative stents which may have resulted in differences in postoperative outcomes. We did, however, control for the presence of a positive preoperative urine culture which would capture those who had stents placed for infection. Finally, we did not control for stone location or the type of ureteroscope used during the procedure as these were not captured in our dataset. We did, however, control for the use of an access sheath and operative time, both of which have been suggested as intraoperative risk factors for infectious complications. ⁵

Conclusions

Preoperative prophylactic antibiotic duration did not affect our postoperative infectious complication rate after ureteroscopy even in patients at high risk of urinary tract colonization or with positive, asymptomatic urine cultures. The presence of neurological insult and foreign bodies predicts urinary tract colonization and suggests that these clinical features may be useful in identifying high-risk patients; however, prolonged preoperative antibiosis may not improve postoperative outcomes. Future studies could assess this clinical scenario in a randomized fashion to establish the foundational data toward the standardization of preoperative care.