Neuraxial anesthesia can result in significant bladder denervation in the perioperative period and can subsequently precipitate urinary retention.1 The dysfunction associated with this transient effect ranges from mild (with delayed initiation of micturition and incomplete bladder emptying) to severe (with urinary retention and bladder overdistension). When alleviated with catheterization, urinary retention can increase morbidity by introducing infection and increasing the length of hospital stay.2-5

Urinary retention is the inability to initiate micturition or to empty the bladder completely. There are no clear defining characteristics of urinary retention, such as a specific volume of urine or elapsed time postoperatively without micturition; however, in accordance with the consensus view in the contemporary literature, urinary retention would be described as an inability to initiate micturition with a bladder volume exceeding 500 mL.6 Urinary retention can be complete or partial, acute or chronic, painful or silent, obstructive or non-obstructive. “Overflow” incontinence secondary to excess intravesical pressure can occur. De novo incontinence secondary to sphincter damage, detrusor overactivity (urgency), or stress (precipitated by increased intra-abdominal pressure) developing in the perioperative period are uncommon occurrences. The long-term consequences of postoperative urinary retention (POUR) are not always immediately apparent in the perioperative period, although increased hospital length of stay and prolonged detrusor dysfunction have been documented.1,6

Neuraxial local anesthetics block the afferent and efferent limbs of the micturition reflex resulting in detrusor dysfunction and the inability to sense a full bladder, thus impairing micturition. Neuraxial opioids enhance this effect by decreasing the sensation of bladder fullness, thus increasing bladder capacity and weakening detrusor contraction through their actions at the spinal level and in the pontine micturition centre.7-9 Other previously identified perioperative and pre-existing risk factors for urinary retention include age, type of surgery, drug side effects, and benign prostatic hypertrophy, but none usually results in the transient, though dense, dysfunction caused by neuraxial anesthesia.

This study was initiated because a recent review on urinary retention did not focus specifically on urinary retention after neuraxial anesthesia and/or analgesia.1 The extensive work by Baldini et al. is a narrative review wherein they aim to give the reader a broad overview of the clinical problem.1 In the present study, we attempt to go beyond a narrative review and perform a systematic assessment of urinary outcomes, including time to micturition, incidence of catheterization, and subsequent frequency of urinary tract infection after neuraxial intervention. The primary aim of this review is to determine the incidence of urinary retention and any associated morbidity in patients following neuraxial anesthesia or analgesia and to identify risk factors prolonging impaired micturition.

Methods

A systematic search of the PubMed, MEDLINE, and EMBASE databases was performed from January 1980 to January 2011 using the medical subject heading (MeSH) words “neuraxial anesthesia” or “neuraxial analgesia” or “epidural” or “intrathecal” or “spinal”. These were combined with the MeSH terms “urinary retention” or “urinary incontinence” or “urinary catheterization” or “micturition” or “post void residual”. The search was limited to articles published in the English language and human adults. Each abstract was evaluated to identify studies where neuraxial anesthesia was utilized and urinary retention, or time to micturition, or post void residual was reported as an outcome. The references of the retrieved articles were hand searched for any relevant studies not captured in the original search. In addition to randomized controlled trials (RCTs), observational studies were also included because of the limited amount of data present in the literature.

Studies included in the review were categorized according to modality of neuraxial anesthesia – intrathecal or epidural. Studies involving combined spinal-epidural techniques were grouped with those utilizing epidural techniques because the effects of the epidural infusion typically outlast the effects of the intrathecal component. Studies were included even if patients did not undergo surgical procedures but were volunteers for urodynamic studies. Data were abstracted using a template created independently to identify the following information: primary author with year of publication, study design, Jadad score for RCTs, number of patients, surgical class, neuraxial medication employed (local anesthetic only, local anesthetic with short- or long-acting opioid, opioid only, undefined), incidence of urinary retention, average time to first micturition, and post void residual (PVR). If reported in the source study, we also abstracted the number of patients requiring catheterization and the incidence of urinary tract infection. Where possible, we reported the statistical significance for the incidence of urinary retention between comparators.

With each neuraxial technique, studies were further subdivided into the following categories: local anesthetic only, local anesthetic with long-acting opioid, local anesthetic with short-acting opioid, and undefined.

The methodology of each RCT was graded according to the criteria published by Jadad et al.10 All RCTs were included regardless of grade, and observational studies were not graded because it became clear after the initial literature review that the level of methodological rigour and study methods were so variable that meta-analysis would not be feasible or appropriate given that there is no easily communicated standard for grading observational studies. Two of the authors (S.C., P.M.) independently performed the literature search and data extraction. Results were combined and differences were resolved through discussion amongst the three authors (S.C., P.M., I.A.).

Results

Initially, 4,465 references were retrieved, and the search yielded 3,662 abstracts when limited to the English language and human subjects. Each abstract was reviewed, but it was not utilized if it did not state clearly that urinary retention, or PVR, or time to micturition was recorded as an outcome. If the abstract did not state specifically that spinal/epidural anesthesia and/or analgesia were employed, it was not retained.

We identified 94 studies (11,162 patients) where neuraxial anesthesia or analgesia was employed and where urinary retention, time to micturition, or PVR was reported as an outcome (Figure). Meta-analysis was not performed owing to the heterogeneity of the definitions of urinary retention, if at all provided, and the significant variability in the dose, type, and use of opioids in the neuraxial medications utilized.

Figure
figure 1

Flow chart of screened, excluded, and included studies

Full size image

In 16 of the 94 RCTs (1,066 patients), time to return of spontaneous micturition after intrathecal anesthesia was assessed as a primary outcome (Table 2).11-26 In 41 studies (5,548 patients), urinary retention or PVR with intrathecal anesthesia was assessed (Table 3),11,13,16,17,20,22,27-65 and in 39 studies (4,938 patients), urinary retention or PVR with epidural anesthesia and/or analgesia was assessed (Table 4).28,33,35,43,48-50,54,66-100 An additional five studies involved both intrathecal and epidural techniques. There is overlap in the numbers of patients/studies reported for Tables 2, 3, and 4 because multiple outcomes and/or multiple neuraxial procedures were examined in several studies. The characteristics of the included studies are detailed in Table 1. Among the 94 studies, 54 of the included studies were RCTs, 27 were prospective observational studies, and 13 were retrospective reviews. None of the included studies designated urinary retention as a primary outcome measure. Among the 94 studies, two studies investigated spinal anesthesia in volunteers with no surgical procedure performed.15,24

Table 1 Characteristics of the 94 studies retained for analysis
Full size table

Of the 55 studies (including 16 RCTs in Table 2) reporting a urologic outcome after intrathecal anesthesia, 41 specifically assessed the incidence of urinary retention, 27 assessed for rate of catheterization, 16 assessed time to micturition, six reported rates of infection, and one reported PVR. Only 25 studies defined criteria for urinary retention. The criteria ranged from quoting bladder volumes (from 150-600 mL) or time frames (from 30 min to two days) to stating that catheterization was necessary without describing indications or that urinary retention was simply as inability to void without other defining criteria (Table 3).

Table 2 Studies with time to micturition as an outcome after intrathecal local anesthetics
Full size table
Table 3 Intrathecal anesthesia studies
Full size table

Forty-two of the 44 studies reporting a urologic outcome after epidural analgesia specifically assessed the incidence of urinary retention. Similarly, only 26 studies defined criteria for urinary retention, but these criteria were as varied as those in the studies reporting intrathecal anesthesia. Thirty-one studies reported catheterization rates, seven reported infection rates, and only one reported PVR.

Local anesthetic type and dose

The incidence of POUR appears correlated with the specific intrathecal local anesthetic utilized. The studies reporting incidence rates of > 20% were those utilizing either tetracaine or bupivacaine, while those employing procaine or lidocaine reported incidence rates < 20% (Table 3). The studies utilizing epidural analgesia are difficult to assess in this respect because of the highly variable durations of infusion (Table 4).

Table 4 - Epidural anesthesia or analgesia studies
Full size table

The 16 RCTs that specifically examined time to return of spontaneous micturition after intrathecal anesthesia as a primary outcome assessed several local anesthetics in differing concentrations, densities, and doses (Table 2). The time to first micturition varied from 103 min15 (2-chloroprocaine) to 462 min (bupivacaine).18 We did not perform a linear regression analysis of the micturition time based on dose because we considered that the varying densities and concentrations utilized would confound the results. Seeing as widely varying doses, concentrations, and densities were utilized even within groups, we did not combine the results of each local anesthetic.

Time to spontaneous micturition correlates with the potency of the local anesthetic administered intrathecally, and it correlates with dose for each specific local anesthetic. Kamphuis et al. showed this with filling cystometric studies comparing bupivacaine with lidocaine.18 The longer lasting and more potent bupivacaine was associated with longer detrusor dysfunction (462 min) compared with lidocaine (233 min). This difference becomes more apparent when varying doses of the same medication (concentration and density) are compared within studies. Ben-David et al., Urmey et al., Kallio et al., and Casati et al. showed this with bupivacaine, lidocaine, articaine, and 2-chloroprocaine, respectively.11,22,25,26 The longest times to spontaneous micturition after intrathecal anesthesia with each of bupivacaine, lidocaine, articaine, and 2-chloroprocaine were 462 min, 260 min, 279 min, and 271 min, respectively.18,19,21,26

Neuraxial opioids

The effects of intrathecal or epidural opioids on bladder function are similar to those of local anesthetics in that the potency and dose of the opioid appears to predict the duration of bladder dysfunction. Morphine in conjunction with intrathecal anesthesia was utilized in only two studies, both non-randomized, and the reported incidence rates of urinary retention were 36% and 25%, respectively.45,64 In contrast, studies in which either intrathecal fentanyl or sufentanil was utilized reported lower incidence rates ranging from 0 to 25%.12,16,18,46,48,57,61 Kamphuis et al. showed that the addition of sufentanil 20 μg prolonged the detrusor dysfunction associated with intrathecal lidocaine 100 mg from 233 to 332 min.18

A similar pattern occurred with neuraxial opioids and epidural analgesia, though the pattern is less distinct (Table 4). Ten studies utilizing long-acting opioids reported incidence rates of 9.2-79.5% (only three studies showed rates < 40%). In 15 studies on short-acting opioids, the rate of urinary retention ranged from 0% to 40%.

Comparison with other anesthetic modalities

Only five studies made comparisons with other anesthetic modalities. Schmittner et al. compared intrathecal with general anesthesia and found no difference in time to micturition.56 Casati et al. also found no difference between intrathecal, general, or peripheral nerve block anesthesia in terms of time to micturition.25 Sungurtekin et al. found no difference between intrathecal anesthesia and local infiltration, while van Veen et al., Young et al., and Anannamchareon et al. reported significantly higher rates of urinary retention with intrathecal anesthesia compared with local infiltration.27,58,61,63

Urinary tract infection

Thirteen of the 94 studies included in this review reported the incidence of urinary tract infection associated with catheterization.28,30,40,61,63,64,66,69,73,75,76,89,100 Six of these reported no infections while seven studies reported rates of < 10%.

Discussion

Our review of the literature identified several themes with respect to the effects of neuraxial anesthesia and analgesia on POUR and other urinary outcomes. First, the potency/dose of local anesthetic and the presence of opioids affect detrusor dysfunction and the time to return of spontaneous micturition. This time period lasts as long as 462 min (bupivacaine) or is as short as 103 min (2-chloroprocaine). With long-acting epidural opioids, the reported incidence of urinary retention is as high as 79.5%. Second, whether detrusor dysfunction specifically results in POUR is unclear, but the incidence of POUR, at least after single-dose intrathecal anesthesia, is low, and complications, such as urinary tract infections, are even less frequent. Complications (urinary tract infection) associated with POUR after epidural analgesia also surface infrequently. However, there are inherent limitations to our analysis. Few of the included studies are randomized trials that compare general with neuraxial anesthesia and include urinary retention as a primary outcome. Furthermore, we included all RCTs regardless of Jadad score and did not grade observational studies to highlight the inadequacy of the current evidence with methodologically sound studies assessing this clinical issue. Indeed, the heterogeneity of definitions and anesthetic management further hampers any ability to offer more quantitative analysis.

Our conclusions must be viewed cautiously owing to several factors, including the aforementioned varying definitions used by the included studies to define urinary retention and the significant heterogeneity in local anesthetic type/dose and opioid doses. Several studies, though implicating intrathecal anesthesia as a risk factor for urinary retention, neither discuss a comparative anesthetic modality nor provide the dose/type of local anesthetic utilized.28,30,33,36,41,43,52-55

Results of studies assessing the urodynamic effects of both intrathecal local anesthetics and opioids tend to concur with our data. Kamphuis et al. performed filling cystometric studies in 30 male patients to estimate detrusor pressure and flow rates.18 The studies were performed both prior to and following intrathecal anesthesia with hyperbaric lidocaine (100 mg) with or without sufentanil (20 μg) or with hyperbaric bupivacaine (10 mg). Patients’ bladders were filled at a constant rate of 50 mL∙min−1 when supine, and filling was stopped when a strong desire to void was felt (the cystometric capacity ~ 500 mL). The urge to void disappeared within 60 sec of the start of injection of intrathecal anesthetic. The recovery of the ability to void normally (using only detrusor muscle, generating intravesical pressures of 40-50 cm H2O) did not occur until the block regressed to the S3 segment. Importantly, the mean duration of detrusor block was significantly longer than somatic motor blockade; 233 min vs 144 min, respectively, for plain lidocaine; 332 min vs 124 min, respectively, for lidocaine with sufentanil; and 462 min vs 233 min, respectively, for hyperbaric bupivacaine.

Kuipers et al. studied the dose effect behaviour of isolated intrathecal opiates (sufentanil or morphine) on detrusor function in 40 healthy male volunteers randomized to receive sufentanil (10 or 20 μg) or morphine (0.1 or 0.3 mg).9 Urodynamic data, including flow rates, bladder pressures, and post void residuals, were recorded hourly. Typically, administration of opiates resulted in dose-dependent decreases in urinary flow rate, increased voiding time, increased post void residual, and diminished urge sensation. Bladder function reverted to normal within 24 hr in all those receiving sufentanil and morphine 0.1 mg. Two participants receiving the higher morphine dose (0.3 mg) did not have full recovery of bladder function within the 24-hr study period. The lower dose of sufentanil (10 μg) resulted in diminished urge (itself a subjective sensation) in six of ten subjects vs nine of ten when the larger dose was used.

The above urodynamic data show the mechanism of detrusor dysfunction, and our data suggest that it is drug and dose related. It is unclear when detrusor dysfunction becomes POUR, although, regardless of the definitions used by source studies, POUR occurred more frequently with high potency local anesthetics, higher doses, and opioids.

There are other purported risk factors for POUR that were not addressed in this review, including the volume of intraoperative fluid administered, type of surgery, and age. Keita et al. prospectively studied 313 adult patients (mean age 46 yr; range 16-88 yr) scheduled for general (86.5%) or neuraxial (13.5%) anesthesia and showed that the volume of intraoperative fluid administration (≥ 750 mL odds ratio [OR] = 2.3), age (≥ 50 yr OR = 2.4), and bladder volume on entry to the postanesthesia care unit (≥ 270 mL OR = 4.8) were independent risk factors for the development of POUR.39 Joelsson-Alm et al. observed that orthopedic surgical patients were 6.87 times (95% confidence interval [CI] 1.76 to 26.79) more likely to develop bladder distension.101 Other studies have implicated colorectal surgery and obstetric status as risk factors. The odds ratio for POUR after colorectal surgery and epidural analgesia was reported to be as great as 4.3 (95% CI 1.2 to 15.9).5 Weiniger et al. showed that intrapartum women receiving epidural analgesia during labour had greater post void residuals independent of fluid volume administered when compared with controls not receiving epidural anesthesia (median 240 mL vs 45 mL, respectively) with no significant difference on postpartum days one or two.102 In addition, advanced age appears to be a risk factor for urinary retention as males > 70 yr having spinal anesthesia for lower limb joint replacement are at higher risk of developing retention than females or those < 69 yr irrespective of anesthetic modality.55,103

Conclusion

Our review of the literature suggests that the duration of detrusor dysfunction caused by neuraxial anesthesia and analgesia is related to the dose/potency of local anesthetic and the use of long-acting neuraxial opioids. This may influence the incidence of POUR. Unfortunately, the defining criteria in the literature are so heterogeneous when present that an accurate estimate of the incidence and subsequent complications is extremely difficult to determine.

There are several limitations to our review. First, only two of the RCTs included in this review specifically compared general with neuraxial anesthesia and the respective incidence of POUR as a primary outcome, inherently resulting in underpowered results. The observational studies included frequently lacked a comparator group. Second, the definitions of POUR were widely variable such that POUR in one study may not be considered POUR in another. Details regarding individual cases were not provided, preventing any ability to rate outcomes according to standardized criteria.

From the available literature, short-acting neuraxial blockade is safe in both inpatient and ambulatory anesthesia given the short duration of detrusor dysfunction with a minimal incidence of POUR and subsequent catheterization/infection. However, clinicians should select agents carefully to strike a balance between duration of somatic blockade and the risk of POUR. We suggest that clinicians remain cognizant of this issue and scan patients at higher risk of urinary retention with catheterization at a bladder volume of 500 mL (~cystometric capacity) given that there is a lack of knowledge regarding the subclinical implications of an episode of retention that “spontaneously” resolves.

Moving forward, standard definitions for urinary retention and indications for catheterization should be developed in conjunction with urologists so that relevant clinical trials comparing anesthetic modalities and incorporating urological endpoints produce data that can be easily used by all clinicians to assess and manage this clinical problem.