Less Urinary Tract Infection by Earlier Removal of Bladder Catheter in Surgical Patients Receiving Thoracic Epidural Analgesia

Background and Objectives: It is common practice to catheterize the bladder in the presence of epidural analgesia and to leave the bladder catheter in situ to avoid postoperative urinary retention. However, bladder catheterization carries the risk for urinary tract infection (UTI). The objective of this randomized control trial was to assess whether the incidence of UTI will differ among patients receiving standard care and patients who have the bladder catheterization discontinued on the morning after surgery with the epidural still functioning.

Methods: Patients at low risk for postoperative urinary retention, scheduled for thoracic and abdominal surgery and receiving continuous thoracic epidural analgesia, were randomized on the morning after surgery to 2 groups: in the early removal group (n = 105), the bladder catheter was removed on the same morning after surgery, whereas in the standard group (SG) (n = 110), the bladder catheter was removed when epidural analgesia was discontinued (3-5 days). Urinary bladder volume was assessed by ultrasound. Primary and secondary outcomes were the incidence of UTI and rate of recatheterization.

Results: Two hundred fifteen patients were randomized. There were 17 UTI cases in total, with 15 (14%) in the SG and 2 (2%) in the early removal group (P = 0.004). The incidence of recatheterizations was not different between the 2 groups (P = 0.09) and did not correlate with the site of epidural insertion. When matched for the types of surgery, the duration of hospital stay was longer in the patients who contracted UTI (P = 0.004). There were more patients older than 65 years in the SG.

Conclusions: Leaving the bladder catheter as long as the epidural analgesia is maintained results in a higher incidence of UTI and prolonged hospital stay. Removal of the bladder catheter on the morning after surgery does not lead to higher rate of catheterizations.

The urinary bladder is routinely catheterized during major surgery to monitor urine output throughout the perioperative period and to avoid the risk for urinary retention and bladder overdistension. The catheter is inserted in the patients' bladder at the beginning of surgery and stays for a few days according to different practices.^1,2 Prolonged catheterization is frequently associated with the risk for urinary tract infection (UTI), which increases up to 40% after 5 days of catheterization.³ Catheter-associated UTI is the most common nosocomial infection, accounting for more than 1 million cases each year in US hospitals. In 2002, 13% of the total deaths linked to health care infections were caused by UTI.⁴

There is strong evidence that a mixture of low-concentration local anesthetics and opioids, injected in the epidural space and used as part of multimodal analgesia for major abdominal surgery, provides optimal pain relief, reduces the opioid-related adverse effects, and facilitates early mobilization and early nutrition.⁵ However, the use of epidural during and after surgery has been reported to be responsible for postoperative urinary retention (POUR), with an average incidence of 26%.⁶

Local anesthetics injected in the epidural space act on the sacral and lumbar nerve fibers in the spinal cord, blocking the transmission of afferent and efferent nervous impulses from and to the bladder. This results in decreased sensation of urgency and impaired bladder detrusor contraction. Intrathecal opioids also decrease the urge sensation and detrusor contraction, increasing the bladder capacity and the residual volume and altering sphincter function, thus resulting in impaired coordination between the detrusor contraction and internal urethral sphincter relaxation.⁶ Based on these animal and human findings, it has been common practice to catheterize the bladder in those surgical patients receiving epidural analgesia and to leave the catheter in situ as long as epidural analgesia is maintained.^6,7

The methods of palpation/percussion to assess urinary retention are unreliable and not sensitive enough to measure the volume of urine in the bladder, whereas the requirement of catheterization⁷ is invasive and associated with a number of complications such as catheter-related infections, urethral trauma, and prostatitis.⁸ Measurement of urine volume by ultrasound has been shown to be accurate and easy to perform and allows the clinician to limit unnecessary catheterizations.^9-11

The present study was set up to determine the incidence of UTI in patients having an indwelling bladder catheter and thoracic epidural analgesia for major abdominal and thoracic surgery. Two groups were compared, an early removal group (ERG), in which the bladder catheter was removed on the morning after surgery while the epidural analgesia was still maintained, and a standard or traditional group, in which the bladder catheter was removed when epidural analgesia was discontinued. It was hypothesized that those patients in whom the bladder catheter was removed on the morning after surgery would have a lower incidence of UTI.

METHODS

Patients

The trial was approved by the ethics board of the McGill University Health Centre, and written informed consent was obtained from all patients. The study was conducted from February 1 to October 31, 2008. Patients scheduled for elective major abdominal and thoracic surgery, and with anticipated thoracic epidural analgesia, were seen in the preoperative clinic to determine whether they were eligible for the study. To select a population at minimal risk for POUR, all patients completed a questionnaire on symptoms of lower urinary tract flow obstruction.¹² If positive, they underwent a uroflowmetry test in the urodynamic laboratory. The results were analyzed according to the International Continence Society¹³ by a senior urologist, and those considered at risk for urinary retention were excluded from the study. Similarly, patients with a history of POUR and with medical conditions and surgical conditions recognized to be at risk for POUR were also excluded (Table 1).

Table 1

Anesthesia and Surgical Care

Upon arrival in the operating room, an intravenous cannula was inserted in the forearm under local anesthesia, and midazolam 0.03 mg/kg was administered intravenously (IV). An epidural catheter was inserted between the third and sixth thoracic dermatomes for thoracic surgery and between the 6th and 10th thoracic dermatomes for abdominal surgery. After a test dose of 3 mL lidocaine 2% with epinephrine, 10 mL of bupivacaine 0.5% was administered in the epidural space to establish a bilateral segmental sensory block covering the surgical incision. Bupivacaine 0.25% was infused continuously at the rate of 5 to 10 mL/hr throughout surgery. General anesthesia was then induced with propofol 2 mg/kg, fentanyl 3 μg/kg, and rocuronium 0.8 mg/kg and maintained with end-tidal concentration of desflurane between 5% and 7%. The lungs were ventilated at normocapnia with a mixture of air and oxygen (50%), and intraoperative normothermia (nasopharyngeal temperature between 36.0°C and 36.8°C) was maintained with forced-air warming blankets positioned over exposed parts of the body. After induction of anesthesia, a catheter was inserted in the bladder, and urine output was recorded throughout the surgery. Pressure stockings were placed at the lower limbs, and 5000 U of unfractionated heparin was injected subcutaneously 1 hr after the placement of the thoracic epidural catheter. Twenty minutes before skin incision, 2 g cefazolin with or without 500 mg metronidazole was administered intravenously. If the surgery lasted for more than 5 hrs, a second dose of cefazolin, 1 g, would be administered. Intravenous crystalloids were infused at a rate of 4 to 8 mL/kg per hr. When the hemoglobin dropped to less than 8 g/dL, 2 U of blood was transfused. Hypotension was treated with phenylephrine 0.1 mg as required.

At the end of surgery, the patients were transferred to the postanesthesia care unit, where they stayed until the cardiorespiratory conditions were stable and pain relief was adequate. A continuous epidural infusion containing a mixture of bupivacaine (0.1%) and fentanyl (3 μg/mL) was started at a rate between 6 and 10 mL/hr and incrementally adjusted up to a maximum of 15 mL/hr if the visual analog scale (VAS) for pain (0-10, where 0 = no pain and 10 = excruciating pain) at rest was more than 3.

All patients were visited on the morning after surgery by the surgical team and the acute pain service (APS). All those with stable cardiorespiratory conditions, urinary output of more than 0.5 mL/kg per hr, and functioning epidural analgesia were then allocated, using a computer-generated block randomization schedule, to 2 groups: the ERG, in which the bladder catheter was removed on the same morning, and standard group (SG), in which the bladder catheter was going to be kept as long as epidural analgesia was in use. During the same morning, if the quality of epidural analgesia was not satisfactory (presence of unilateral block, patchy block, lack of sensory block covering the surgical incision), a member of the APS would resite the epidural blockade, and the subject would be kept in the study. The prescription order to either remove or keep the bladder catheter was written in the nursing chart and communicated to the surgical ward nurse in charge of the patients.

Postoperative Pain Management

The APS assessed daily the sensory block and the intensity of pain at rest and on movements and adjusted the epidural infusion to maintain VAS at rest at less than 3 and VAS on coughing at less than 5. In this institution, patients undergoing abdominal surgery receive the epidural infusion for an average of 3 days, whereas those undergoing thoracic surgery, for an average of 5 days. All patients also received oral naproxen 500 mg twice a day and oral acetaminophen 1 g every 6 hrs for the first 5 postoperative days. When the epidural infusion was discontinued, patients received oral slow-release oxycodone (OxyContin) 10 mg twice a day with breakthrough oxycodone 5 to 10 mg every 3 to 4 hrs, as required.

Postoperative Fluid Management, Nutrition, and Mobilization

After surgery, Ringer's lactate solution was infused intravenously at a rate between 80 and 100 mL/hr for the first 24 hrs and followed by 5% dextrose in 0.45% NaCl at the same rate. If urinary output was less than 0.5 mL/kg per hr, a bolus of 300 mL of Ringer's lactate solution was infused. Patients were given oral fluids 24 hrs after surgery and progressed to full fluids and full diet if tolerated. All patients were sitting on the chair on postoperative day 1 and encouraged to walk.

Postoperative Assessment of Urinary Bladder Volume

Once the bladder catheter was removed in the 2 groups, either on the morning after surgery (ERG) or at the same time when the epidural was discontinued (SG), the volume of urine in the bladder was measured by using an ultrasound scanner (Bladderscan, BVI 3000; Verathon Medical Inc, Bothell, Wash), previously calibrated over a range of volumes (50-500 mL) and found to be accurate within 25 mL. Bladder scanning was conducted by the same person (C.Z.), who received 1-day training by a senior nurse in the urodynamic laboratory, and was performed every 3 hrs after the removal of the bladder catheter. The following algorithm, as presented in Figure. 1, was followed according to published guidelines.¹⁴

Figure 1

Diagnosis of UTI

Patients were diagnosed having in-hospital UTI according to international guidelines¹⁵ based on the following characteristics: pyrexia to a temperature of 38°C, urinary tract symptoms (dysuria, increased frequency of urination, urinary urgency, suprapubic pain, burning on micturition, or onset or aggravation of urinary incontinence), and positive urine culture (>10⁷ bacterial colonies of microorganism-forming units per liter within 2 weeks after the removal of bladder catheter). Once the urine and blood samples were sent for culture and sensitivity, an empirical treatment with broad-spectrum antibiotics based on local susceptibility patterns was started. Afterward, when the urine sample was positive, targeted antibiotic therapy was prescribed according to the urine culture results.

Data Collection and Statistical Analysis

The following data were collected: demographic characteristics, type and duration of surgery, amount of intraoperative IV fluids, duration of bladder catheterization, level of epidural insertion, pain intensity at rest and on coughing, length of hospital stay, number of recatheterizations, and number of subjects who contracted UTI.

The primary outcome of interest was the incidence of UTI. The secondary outcome was the incidence of bladder recatheterization.

Descriptive statistics were calculated for demographics and clinical variables.

To determine the magnitude of difference between the groups, Student t test was used for normally distributed variables, whereas Wilcoxon rank sum test for non-normally distributed variables. A χ² statistic was used to test whether distributions of categorical variables differ from one another. Results of all analyses were considered statistically significant when P ≤ 0.05.

Determination of sample size was based on the incidence of UTI in thoracic and abdominal surgery in our institution (16%). This is in agreement with the literature reporting that, in patients with an indwelling catheter placed for 3 to 4 days, the risk to contract UTI ranges from 15% to 40% (from 5% to 10% per day of indwelling bladder catheter).¹⁶ One hundred subjects in each group were required to detect a 4-fold reduction in the incidence of UTI, with a type I error of 0.05 and a power of 80%.

RESULTS

Patients

Three-hundred twenty-one patients were interviewed in the preoperative clinic, and upon exclusion of 57 patients (17 declined, 25 for medical reasons, and 15 for surgical reasons), 264 went for surgery as scheduled (Fig. 2; Table 1). On the morning after surgery, 49 patients were excluded, of which 22 did not receive epidural analgesia (22 had patient-controlled analgesia), 10 patients had extensive pelvic dissection requiring indwelling bladder catheter for a prolonged period, 10 had persistently low urine output (<0.5 mL/kg per hr), and 7 were admitted to the intensive care unit, of which 6 were for cardiorespiratory failure and 1 for massive blood transfusion, leaving 215 for randomization (Fig. 2). Epidural analgesia was found to be satisfactory in all patients, except for 2 patients who had unilateral sensory block. The epidural catheter was resited successfully during the morning, and the patients were kept in the study.

Figure 2

The demographic characteristics of the patients studied after randomization and the clinical data on American Society of Anesthesiologists health status, type and duration of surgery, intraoperative fluid administration, duration of bladder catheterization, intervertebral level of epidural insertion, VAS at rest and on coughing, and length hospital stay are presented in Table 2. The 2 groups were different for age; however, in the category of those older than 65 years, the number and the average age were similar in both groups (SG: n = 45; mean age, 74 years [SD, 6 years]; ERG: n = 36; mean age, 73 years [SD, 6 years]). The intervertebral thoracic level of epidural catheter insertion was similar in both groups: 5 ± 1 and 8 ± 1, respectively, in the thoracic and in the abdominal subgroups of the SG, and 5 ± 1 and 9 ± 1 in the 2 subgroups of the ERG. Duration of surgery was longer in the SG, and VAS on coughing was higher in the ERG. Five of 105 patients in the ERG and 5 of 110 in SG received blood transfusion. There were no differences in the types of surgery between the 2 groups (Table 3). The average epidural infusion rate of bupivacaine and fentanyl was 9 mL/hr (SD, 2 mL/hr).

Table 2

Table 3

Clinical Outcomes

The incidence of UTI was significantly greater in the SG compared with the ERG, with 15 (14%) of 110 patients versus 2 (2%) of 105, respectively (Table 4). The incidence of in-out and 24-hr bladder catheterization were similar in both ERG and SG: 8 versus 2 and 3 versus 0, respectively.

Table 4

The list of subjects who contracted UTI with positive urine culture, together with data on demographic characteristics and on duration of surgery, length of bladder catheterization, and length of hospital stay, is presented in Table 5. There were 15 UTI cases in the SG and 2 in the ERG, with 11 women and 6 men. Those patients who contracted UTI were significantly older than the overall population studied (Table 6), had longer duration of surgery, and stayed longer in the hospital. Table 7 presents the 2 subgroups of patients in the same SG who did and did not contract UTI. Body mass index (BMI) and length of catheterization were similar in both subgroups. There was no difference in mean age, but significantly more patients older than 65 years contracted UTI (11/15; P = 0.006). The duration of surgery was not different, but the length of hospital stay was greater in the group that contracted UTI.

Table 5

Table 6

Table 7

DISCUSSION

The results of this prospective randomized study show that when the bladder catheter is kept for the entire period of epidural analgesia, the incidence of UTI is significantly higher, and the hospital stay is longer. In contrast, when the bladder catheter was removed on the morning after surgery, only 2 subjects of 105 contracted UTI. In addition, early removal of the catheter was not associated with a higher incidence of recatheterization.

The pathogenesis of POUR depends on many factors, particularly the type of surgery, and anesthetic drugs and techniques contribute most. Hyperactivity of the sympathetic nervous system and damage to afferent and efferent parasympathetic fibers interfere with the detrusor muscle activity, leading to bladder distension and voiding dysfunction. For this reason, it is common practice to insert an indwelling catheter in the bladder at the beginning of many surgical procedures and left in situ for a certain period, which is dependent on surgical preferences and traditions.

The mean age of the 17 subjects who contracted UTI was higher than those (n = 198) who did not contract UTI (69 years [SD, 12 years] vs 60 years [SD, 14 years], Table 6), and this is in agreement with the literature reporting that age is an independent factor for UTI.⁸ This is confirmed in Table 7, which shows that more patients in the SG older than 65 years have UTI compared with those who are younger than 65 years. In this same group, sex, BMI, intraoperative IV fluids, and duration of surgery were not significant factors.

Several studies have shown a positive relationship between concentration of epidural local anesthetic and incidence of urinary retention. Conacher et al¹⁷ and Matthews and Govenden¹⁸ independently administered bupivacaine 0.25% in the thoracic epidural space for postoperative analgesia and reported, respectively, 31% and 66% incidence of POUR. In contrast, when bupivacaine 0.1% with opioids was used, the incidence of POUR ranged from 2% and 13%.^19-21 Even in low doses, local anesthetics block the impulse in small myelinated and nonmyelinated autonomic fibers, resulting in impaired sensation to void. The urodynamic effects of intrathecal opioids are mainly caused by the action on the opioid receptors in the spinal cord and in the cerebral structures. The rostral spread of opioids, more evident with the hydrophilic opioids, through the cerebrospinal fluid has also been hypothesized as a possible mechanism of action of intrathecal opioids, but the rapid onset of the urodynamic effects with the concomitant onset of analgesia after intrathecal opioid injection and the reversal of the effects by intrathecal naloxone suggest a spinal site of action. Administration of fentanyl intrathecally causes urethral sphincter relaxation as a result of the potent inhibitory property of fentanyl on the sympathetic fibers. Administration of fentanyl in the lumbar epidural space is associated with higher rate of urinary retention compared with thoracic injection, hence less rostral spread in the CNS and less influence on the urodynamics.⁶

In the present study, the intervertebral level of insertion of thoracic epidural catheter was similar in both SG and ERG, implying that the higher, although not statistically significant, incidence of recatheterization in the ERG was not related to the site of epidural insertion.

The use of urinary drainage during continuous epidural analgesia was challenged almost a decade ago when Basse et al²² showed a low incidence of POUR and UTI after colonic surgery when the bladder catheter was removed on the morning after surgery. In that prospective uncontrolled study, surgical care was accelerated with planned 48-hr epidural analgesia and 2-day hospital stay. The present study included different types of thoracic and abdominal surgery and nonaccelerated surgical care where the subjects were randomly allocated to have the catheter either removed on the morning after surgery or kept as long as the epidural was maintained. Despite the differences between the study of Basse et al²² and ours, the incidence of UTI was low in both (4% and 2%, respectively). Similarly, the incidence of in-out catheterization was similar (9% and 11%, respectively) and well below the 26% reported in the literature.

The risk for UTI increases with increasing duration of catheterization. A 3% to 6% incidence of bacteriuria has been reported after a single in-out catheterization,²³ whereas with an indwelling catheter for a longer period, the risk for UTI increases to 5% to 10% per day of bladder catheterization.¹⁶ The average duration of catheterization in the 10 thoracic procedures was 114 hrs (SD, 44 hrs) and significantly greater compared with 67 hrs (SD, 20 hrs) in the 5 abdominal procedures. Therefore, as the risk to contract UTI is directly related to the duration of catheterization and not to the number of in-out catheterizations, it makes sense to limit the duration of catheterization to a minimum.

All patients who developed UTI had fever and/or urinary tract symptoms, according to the published international guidelines. Catheter-associated bacteriuria is estimated to cause 900,000 additional hospital days per year¹ and directly cause 13% of the deaths related to nosocomial infections in the United States.⁴

In this study, overall hospital stay was longer by 2 days in the SG compared with the ERG (Table 2) and in those subjects in the SG who contracted UTI (median, 14 days) versus those who did not (median, 9 days; P = 0.004) (Table 7). This implies that UTI contributed significantly to delay the discharge of patients from the hospital.

The financial implications of UTI such as hospitalization expenses, antimicrobial prescription, and visits to the physician can be considerable, and it has been estimated that each episode of catheter-related bacteremia costs nearly $2900 per episode.²⁴

In this study, assessment of bladder volume was carried out with bladder scanner using a strict protocol. The bladder scanner shows excellent correlation between the volume scanned and that measured by catheterization.^9-11 Although this equipment has been used in outpatient surgical departments to determine patients' suitability to be discharged, no data are available on surgical inpatients with reference to type of analgesia and monitoring of diuresis.

There are several limitations to the present study. As we restricted to those patients at low preoperative risk for POUR, the findings might not be generalizable. Nevertheless, from the clinical point of view, we excluded only 18% of the subjects, implying that those at risk were a very small proportion. For purposes of conducting accurate clinical research, we were very zealous in selecting patients at low risk, and for this reason, besides reviewing the subjects' history, we asked them to fill a validated questionnaire. Based on our experience, there is presently a lack of preoperative guidelines for appropriate identification of those patients at risk for POUR.

We limited our sample to patients receiving only thoracic epidural analgesia, as we were aware that site of epidural insertion and level of sensory block could impact on the incidence of POUR. Within the range of intervertebral thoracic levels of epidural insertion, we did not find that low thoracic level was associated with more recatheterization.

Another limitation of our study was the small number of UTI cases; 15 patients in the SG had UTI, and 11 of them were older than 65 years. In contrast, only 2 patients (both >65 years old) in the ERG contracted UTI. There is sufficient evidence in the literature that elderly subjects are more prone to UTI. As the number of UTI was small, great care has to be taken in the interpretation of the findings. The odds ratio of having UTI in the SG was 8.1 times that of the ERG. When adjusted for age, the odds ratio was 6.9, suggesting age as a modifier. In conclusion, leaving the bladder catheter in place as long as the epidural catheter was in situ increased significantly the risk for UTI, and this contributed to a delay in hospital discharge. In contrast, removal of the bladder catheter on the morning after surgery not only reduced significantly the incidence of UTI, but also did not lead to an increased rate of recatheterization.

ACKNOWLEDGMENTS

The authors thank the thoracic surgeons and the general surgeons for allowing their patients to be studied; the nurses on 18th, 11th, and 19th surgical wards for their precious collaboration; Dr. Armen Aprikian and Ms. Beverly Cleland for the use of the urodynamic laboratory; and Dr. Nancy Mayo for statistical advice.