Enhancing recovery after Kasai portoenterostomy with epidural analgesia

Abstract

Background:

Biliary atresia (BA) is a rare obstructive cholangiopathy that presents in early infancy. The Kasai portoenterostomy (PE) improves survival with native liver. Epidural analgesia is an appealing option to control pain in this fragile patient population, yet its safety, efficacy, and potential benefits remain unproven.

Methods:

Patients undergoing PE for BA between 2001 and 2016 at a single institution were identified by ICD codes. Preoperative labs, procedure details, and recovery outcomes were reviewed retrospectively. Outcomes of interest were need for postoperative mechanical ventilation, pain scores, normalized opioid administration, return of bowel function, and length of hospital stay after PE.

Results:

Of 47 infants undergoing PE for BA, 25 received epidural analgesia, and 22 did not. Infants with epidurals received less systemic opioids over the first 96 hours post-operatively than those without (p<0.001). Epidurals were associated with lower pain scores between 6 and 30 hours post-operatively (p=0.01 to 0.04), during which the highest median 6-hour interval pain score was 0.2 (IQR 0–1.3) for patients with epidurals yet 2.1 (IQR 1.2–3.3) for patients without. Patients with epidurals (88%, n=22) were more commonly extubated before leaving the operating room than those without (59%, n=13; p=0.02). No significant difference was observed in time to first bowel movement (p=0.48) or first oral feed (p=0.81). However, infants with epidurals had shorter hospital stays after PE than those without (6 days [IQR 5–7] versus 8 days [IQR 6.3–11], p=0.01). No major complications were associated with epidural catheters.

Conclusions:

Epidural analgesia in patients undergoing PE for BA appears safe and effectively controls pain while minimizing the need for systemic opioids. Reduced need for mechanical ventilation post-operatively and shortened hospital stays serve as further evidence for using epidurals to enhance recovery after PE.

Introduction

Biliary atresia (BA) is a rare obstructive cholangiopathy that presents in young infants. Obliteration of the intra- and extrahepatic bile ducts progresses to biliary cirrhosis and death by age 3 years if left untreated (1). Although 80% of treated BA patients ultimately require orthotopic liver transplantation (OLT) by age 20, the Kasai portoenterostomy (PE) is the only intervention shown to improve survival with native liver. The surgery is most effective when performed within the first 60 days of life (1). Appropriate pain control in infants undergoing major surgery, such as PE, is paramount, not only for humane practice but also because optimized post-operative pain control may enhance surgical outcomes and impact sensorineural development, particularly in neonates and young infants (2, 3). Systemic opioids were traditionally used for post-operative analgesia in infants. However, the considerable side effects associated with systemic opioids administered to young infants, such as respiratory depression, sedation, hypotension, vomiting, and constipation have prompted anesthesiologists to consider alternative methods for pain control in this at-risk population. Epidural analgesia is an appealing alternative to systemic opioid administration, but the safety and efficacy of epidurals for postoperative analgesia in infants undergoing major abdominal surgery, particularly the Kasai PE for BA, has not been adequately explored.

In adult patients having abdominal surgery, epidural analgesia is associated with earlier return of bowel function, less pain, and may decrease length of hospital stay (4). As the technique gains popularity in the pediatric realm, data remain sparse regarding the use of epidurals for pain management in young infants, particularly those less than 60 days of age. A few small, retrospective studies have suggested that epidurals offer good perioperative analgesia and may provide additional benefits such as earlier extubation and return of bowel function in infants undergoing major abdominal surgery (5, 6). However, these studies did not include any infants receiving PE. One report of a single-center experience suggested that epidural analgesia was safely utilized in 21 patients undergoing PE (7). However, neither this latter study nor any other to date has directly compared patients who had an epidural for pain control after PE and those who did not. Thus, it remains unclear how epidural analgesia may benefit or compromise patients undergoing PE for BA.

Due to the paucity of available data on the use of epidural analgesia in infants undergoing PE, we performed the first retrospective analysis to compare recovery outcomes between patients after PE for BA who did or did not receive epidural analgesia. We aimed specifically to (1) assess the efficacy of epidural analgesia for controlling pain and minimizing the need for systemic opioids, (2) compare perioperative outcomes for patients with and without epidurals, and (3) evaluate the safety of epidural analgesia in young patients undergoing PE for BA. We hypothesized that epidural analgesia in patients undergoing PE for BA is not only safe and effective, but also, actually enhances recovery after PE.

Materials and Methods

Study design and patient selection

Infants treated for BA at Vanderbilt University Medical Center between January 1, 2001 and December 31, 2016 were identified by ICD codes. All records containing an ICD-9 diagnosis code of 751.61 or ICD-10 code of Q44.2 (i.e., atresia of bile ducts) at 1 year of age or younger were ascertained. Charts were reviewed manually. Children were excluded if the attending pediatric surgeon’s operative report stated that a diagnosis other than BA was found at time of surgical exploration. Cases were also excluded if the patient did not undergo PE, as determined by the operative report. The Vanderbilt Institutional Review Board approved this study (#161414) and waived the requirement for informed consent.

Demographic data abstracted from the medical record included age and weight at PE, gender, race, and gestational age. Disease-related variables included Ohi classification and pre-operative total bilirubin. Pre-operative coagulation studies of interest were prothrombin time (PT), international normalized ratio (INR), partial thromboplastin time (PTT), and platelet count. Operating time was defined as the time from surgical incision to closure. Pre-procedure time was the time elapsed from the patient’s arrival in the operating room (OR) to the initial surgical incision, and post-procedure time reflects time from incision closure to the patient being transported out of room. Total non-operative time in the OR was the sum of pre-procedure and post-procedure times. The primary outcomes of interest were systemic opioid administration and pain scores post-operatively. Pain scores were assessed according to the Face, Legs, Activity, Cry, Consolability (FLACC) scale, and scores ranged from 0 to 10 (8). Additional recovery outcomes of interest were need for mechanical ventilation after surgery, time to first bowel movement, time to first oral feed, epidural or other complications, and length of stay after PE. Patients with epidurals are monitored continuously by nursing in the post-anesthesia care unit and then at specified intervals (depending on patient and epidural characteristics) once they move to the floor or intensive care unit. The pediatric pain service assesses each patient with an epidural daily and more frequently if needed. When assessing for neurological side effects, the team specifically evaluates vital signs, cry, level of consciousness (RASS score), neurologic baseline, and extremity movement (Bromage scale). Disease outcomes of interest were lowest total bilirubin level within 90 days of PE, need for liver transplantation at any time, and vital status.

Normalization of opioid doses

To normalize opioid dosing for the most accurate comparisons, the following factors were used to convert opioids administered intravenously (IV) and by mouth (PO) to a common unit of milligrams of IV morphine per kilogram of body weight: 1 mg IV morphine = 0.1 (IV fentanyl, mcg) = 6.7 (IV hydromorphone, mg) = 0.5 (PO oxycodone, mg) = 0.05 (PO codeine, mg) = 0.33 (PO hydrocodone, mg) (9).

Statistical analysis

Data were summarized using the median and interquartile range. The Wilcoxon rank sum test was applied to two-group continuous outcomes, and Pearson’s chi-squared test was applied to categorical outcomes. Post-operative pain scores were modeled with non-linear time trend analysis. Since pain scores were recorded at inconsistent time intervals, each individual patient’s pain scores were averaged over 6-hour intervals, and the Wilcoxon rank sum test was used for two-group comparisons of median pain score during each 6-hour period, as recently described by Man et. al. (10). Systemic opioid administration was also modeled using non-linear time trend analysis. Systemic opioid doses for each patient were summed over 6-hour intervals and median doses were compared between groups with the Wilcoxon rank sum test. Analyses were performed in R: A language environment for statistical computing (11). All tests were 2-sided and p-values < 0.05 were considered statistically significant.

Results

Between 2001 and 2016, 47 patients underwent open Kasai PE for BA. The median age at operation was 59.0 days (IQR 42.5–66.0). An epidural was placed for post-operative analgesia in 25 patients (53%) and remained in place for a median of 4 days (IQR 3–5) with a range of 2 to 8 days. Table 1 summarizes demographics, pre-operative laboratory studies, perioperative details, and disease outcomes for patients with and without epidural analgesia. The epidural group consisted of a larger proportion of female patients than the non-epidural group (64% vs. 32%, p=0.028; Table 1). Additionally, the epidural group tended to favor infants with higher gestational age compared to the non-epidural group, though the result was not statistically significant (96% vs. 77% born at term, p=0.062; Table 1). Although BA is primarily a disease of the liver, only one patient in the cohort had coagulopathy (PT 16.4, INR 1.4) that precluded placement of an epidural. No significant difference in pre-operative coagulation studies or total bilirubin level was observed between the epidural and non-epidural groups (Table 1). The decision to place an epidural was not standardized for the study period and largely was based on surgeon and anesthesiologist preference. In the first half of the series (2001–2009), 10 of 21 patients (48%) had an epidural placed, which increased insignificantly to 15 of 26 patients (58%) in the second half (p=0.491). Details of epidural placement and infusion regimens are provided in Table 2.

Table 1.

Comparison of patients with and without epidural analgesia following Kasai portoenterostomy for biliary atresia

	N	No epidural n = 22	Epidural n = 25	Combined N = 47	p-value
COHORT & PREOPERATIVE EVALUATION
Gender
Male	47	68% (15)	36% (9)	51% (24)	0.028
Female		32% (7)	64% (16)	49% (23)
Age at PE (days)	47	63.5 [44.5, 73.8]	56.0 [40.0, 64.0]	59.0 [42.5, 66.0]	0.131
Weight at PE (kg)	47	4.5 [3.7, 5.2]	4.3 [3.6, 4.7]	4.4 [3.6, 5.0]	0.436
Race
White, non-Hispanic	47	50% (11)	72% (18)	62% (29)	0.375
White, Hispanic		5% (1)	4% (1)	4% (2)
Black		5% (1)	(0)	2% (1)
Other		41% (9)	24% (6)	32% (15)
Gestational age (weeks)
Pre-term (< 37)	46	23% (5)	4% (1)	13% (6)	0.062
Term (≥ 37)		77% (17)	96% (23)	87% (40)
Ohi classification
Type I	47	(0)	8% (2)	4% (2)	0.278
Type II		23% (5)	12% (3)	17% (8)
Type III		77% (17)	80% (20)	79% (37)
Preoperative Labs
PT (sec)	47	13.8 [13.2, 14.9]	13.4 [12.7, 14.1]	13.6 [13.1, 14.6]	0.220
INR	47	1.1 [1.0, 1.2]	1 [1, 1.1]	1.1 [1, 1.1]	0.452
PTT (sec)	42	33.1 [31.1, 36.5]	32.2 [31.1, 35.4]	32.6 [31.1, 35.4]	0.562
Platelet count (x103)	47	454 [305, 553]	476 [330, 588]	471 [322, 574]	0.488
Total bilirubin (mg/dL)	47	8.6 [7.1, 10.8]	8.4 [7.2, 10.9]	8.4 [7.1, 10.9]	0.941
PERIOPERATIVE DETAILS
OR time
Operating time (min)	47	253 [198.5, 330]	272 [225, 305]	263 [219.5, 316]	0.578
Time in room pre-procedure (min)	46	45 [36, 52]	70 [53, 84]	54 [41, 74.8]	0.002
Time in room post-procedure (min)	46	19 [11, 24]	15 [9, 23]	16.5 [10.2, 23.8]	0.408
Total non-operative time in room (min)	46	53 [67, 70]	87 [70, 112]	72 [63.5, 98.2]	0.004
Intraoperative cholangiogram
No	47	36% (8)	40% (10)	38% (18)	0.798
Yes		64% (14)	60% (15)	62% (29)
Extubated after surgery
No	47	41% (9)	12% (3)	26% (12)	0.023
Yes		59% (13)	88% (22)	74% (35)
Post-operative steroids
No	46	14% (3)	24% (6)	20% (9)	0.408
Yes		86% (18)	76% (19)	80% (37)
Time to first oral feed (days)	47	3 [2, 3]	3 [2, 4]	3 [2, 4]	0.814
Time to first bowel movement (days)	47	3 [2, 4]	3 [2, 4]	3 [2, 4]	0.481
ICU days	47	1 [0, 4]	0 [0, 1]	0 [0, 2]	0.006
Length of hospital stay (days)	47	8 [6.3, 11]	6 [5, 7]	7 [6, 9]	0.008
DISEASE OUTCOMES
Total bilirubin ≤ 2.0 at 90 days
No	47	73% (16)	60% (15)	66% (31)	0.358
Yes		27% (6)	40% (10)	34% (16)
Vital status with native liver
Alive, native liver	47	50% (11)	64% (16)	57% (27)	0.599
Alive, liver transplant		36% (8)	24% (6)	30% (14)
Deceased		14% (3)	12% (3)	13% (6)

PE, portoenterostomy; PT, prothrombin time; INR, international normalized ratio; PTT, partial thromboplastin time; ICU, intensive care unit

Values given are No. (%) or median [Q1, Q3].

Table 2.

Epidural Regimens

Insertion
T-7/8	1 (4%)
T-8/9	11 (44%)
T-9/10	1 (4%)
T-10/11	1 (4%)
Caudal	9 (36%)
Unknown/Not Recorded	2 (8%)
Local anesthetic
Bupivicaine (1/16% - 1/10%)	14 (56%)
Ropivicaine (1/10%)	11 (44%)
Opioid
Fentanyl (1–3 mcg/ml)	18 (72%)
Hydromorphone (2–10 mcg/ml)	6 (24%)
No opioid	1 (4%)
Clonidine
Yes	14 (56%)
No	11 (44%)
Postoperative Regimen
Infusion (ml/kg/hr)	0.3 [0.3, 0.4]
Demand dose (ml/kg)	0.04 [0.03, 0.05]
Demand interval (min)	30 [20, 30]
Total demand doses delivered	20 [12, 15]

Values given are No. (%) or median [Q1, Q3].

The primary outcomes of interest were post-operative pain scores and systemic opioid administration between study groups. Patients receiving epidurals had significantly lower pain scores between 6 and 30 hours post-operatively (Table 3, Figure 1). A surrogate for the efficacy of epidural analgesia is the negligible systemic opioid administration in infants receiving epidurals (Figure 2A). As expected, infants having epidurals received less systemic opioids for the first 96 hours post-operatively (p<0.001, Table 4). The breakdown of systemic opioids administered over each 6-hour post-operative interval is shown in Table 4. Cumulative doses of systemic opioids were also significantly lower in patients with epidurals over the first 96 hours post-operatively (p<0.001, Figure 2B).

Table 3.

Pain scores for each 6-hour post-operative time interval

Post-operative hours	Pain Score
	No epidural	Epidural	p-value
0 – 6	1.7 [0.2, 3.5]	0.5 [0, 2.0]	0.190
6 – 12	1.5 [0.3, 2.5]	0.2 [0, 1.3]	0.037
12 – 18	2.1 [1.2, 3.3]	0 [0, 2.8]	0.040
18 – 24	1.8 [1, 3.8]	0 [0, 2]	0.018
24 – 30	1.9 [1.3, 3.1]	0 [0, 1.3]	0.011
30 – 36	1.3 [0.1, 2.5]	0 [0, 1.8]	0.167
36 – 42	0 [0, 1.5]	0 [0, 0.4]	0.508
42 – 48	1 [0, 1.6]	0 [0, 0]	0.076
48 – 54	0 [0, 1]	0 [0, 0]	0.186
54 – 60	1.3 [0, 1.9]	0 [0, 2.3]	0.516
60 – 66	0 [0, 1.3]	0 [0, 0]	0.359
66 – 72	0 [0, 0.9]	0 [0, 0]	0.334
72 – 78	0 [0, 1.3]	0.5 [0, 1.9]	0.348
78 – 84	0 [0, 2.3]	0 [0, 1.7]	0.548
84 – 90	0 [0, 0.8]	0 [0, 0]	0.253
90 – 96	0.8 [0, 2.5]	0 [0, 2.5]	0.649

Scores were averaged for each patient over each 6-hour post-operative interval, and median pain scores were compared between groups.

Values given are median [Q1, Q3].

Figure 1. Pain scores over time for patients with and without epidurals.

Time trend analysis showing post-operative pain score trends between patients with and without epidural analgesia. Shading represents 95% confidence interval, with darker grey representing overlap of the confidence intervals.

Figure 2. Systemic opioid administration over time for patients with and without epidurals.

Time trend analysis showing post-operative (A) systemic opioid administration and (B) cumulative systemic opioid administration. Cumulative doses reflect the sum of all opioids administered up to a given time point. Systemic opioid doses reflect opioids administered intravenously or orally. Shading represents 95% confidence interval, with darker grey representing overlap of the confidence intervals. Timing of adverse advents in each group indicated by arrows (blue = apnea and/or bradycardia, orange = transfer to ICU, red = reintubation, green = catheter malfunction).

Table 4.

Systemic opioid administration

Post-operative hours	Systemic Opioid Dose
	No epidural	Epidural	p-value
0 – 6	0.12 [0.06, 0.24]	0 [0, 0]	< 0.001
6 – 12	0.11 [0.05, 0.30]	0 [0, 0]	< 0.001
12 – 18	0.10 [0.06, 0.17]	0 [0, 0]	< 0.001
18 – 24	0.11 [0.05, 0.23]	0 [0, 0]	< 0.001
24 – 30	0.10 [0.05, 0.23]	0 [0, 0]	< 0.001
30 – 36	0.05 [0.05, 0.37]	0 [0, 0]	< 0.001
36 – 42	0.20 [0.10, 0.66]	0 [0, 0]	< 0.001
42 – 48	0.10 [0.05, 0.64]	0 [0, 0]	< 0.001
48 – 54	0.10 [0.05, 0.59]	0 [0, 0]	< 0.001
54 – 60	0.10 [0.05, 0.29]	0 [0, 0]	< 0.001
60 – 66	0.10 [0.06, 0.39]	0 [0, 0]	< 0.001
66 – 72	0.12 [0.10, 0.42]	0 [0, 0]	< 0.001
72 – 78	0.06 [0.05, 0.15]	0 [0, 0]	< 0.001
78 – 84	0.15 [0.10, 0.28]	0 [0, 0]	< 0.001
84 – 90	0.10[0.06, 0.20]	0 [0, 0.02]	< 0.001
90 – 96	0.19 [0.05, 0.30]	0 [0, 0.03]	< 0.001

Opioid administration given in units of (mg IV morphine / kg body weight) during each 6-hour post-operative time interval for patients with and without epidural analgesia.

Values given are median [Q1, Q3].

As expected, time in the OR before operative incision was longer for patients who received an epidural, presumably due to the additional time required to place the catheter. Median pre-operative in-room time was 70 minutes (IQR 53–84) for patients that had an epidural and 45 minutes (IQR 36–52) for patients without an epidural (p=0.002, Table 1). Of note, 20 of 25 patients with an epidural had it placed in the OR prior to operative incision. The remaining 5 patients had the epidural placed in the OR after the operation. Infants who had epidurals required mechanical ventilation after surgery less frequently (n=3, 12%) than those who did not have epidurals (n=9, 41%; p=0.023, Table 1). Patients with epidurals spent fewer days in the ICU post-operatively (p=0.006, Table 1). Additionally, the median hospital stay after PE was shorter for patients with epidurals (6 days [IQR 5–7]) than those without (8 days [IQR 6.3–11]; p=0.008, Table 1). Interestingly, patients who had an epidural were more commonly discharged home with a prescription for opioids (n=8, 32% vs. n=0, p=0.004), potentially because they were discharged sooner than non-epidural patients. We did not find a significant difference in time to first bowel movement or time to first oral feed among patients with and without epidurals (p=0.481 and p=0.814, respectively; Table 1).

The most common complications relating to pain control and occurring in both groups were bradycardia and/or apnea. Five of the epidural patients experienced apnea and/or bradycardia that prompted modifications to the epidural drug doses or changes to the basal infusion rate. Of note, these events did not occur in the first 12 hours post-operatively, but rather were observed later in the post-operative course between days one and three. No patients with an epidural catheter required re-intubation following post-operative extubation or transfer to an intensive care setting. Technical complications associated with the catheters themselves included one epidural failure due to a kink that was salvaged and four catheters that migrated out, all on post-operative day three or four. One epidural was removed early on post-operative day four due to parental request. Six patients without an epidural experienced apnea and/or bradycardia. All six patients received systemic opioid within the 12 hours preceding the documented event; however, the exact temporal relationship of opioid administration to adverse event could not be assessed reliably. Three of the six were transferred to the pediatric critical care unit for closer observation, of which two infants required re-intubation. Of note, one patient without an epidural re-presented to the emergency department less than 24 hours after discharge due to concerns for inadequate pain control. Timing of adverse events is indicated by arrows on Figure 2.

No significant difference in disease outcomes was observed between patients who had epidurals and those who did not. A total bilirubin level less than 2.0 within 90 days of PE predicts higher 2-year transplant free survival (12). We did not observe a significant difference in the proportion of patients who achieved a total bilirubin level ≤ 2.0 within 90 days of PE (p=0.358, Table 1). Also, no significant difference was detected in survival among patients who had epidural analgesia after PE and those who did not (p=0.599, Table 1).

Discussion

Our data lay the groundwork for standardizing the perioperative management and enhancing recovery of young infants after Kasai PE for BA. We examined use of epidural analgesia for the first time in this fragile BA patient population and found that epidural analgesia appears to associate with lower pain scores and less systemic opioid administration. From our study, further benefits of epidural analgesia included a reduced need for post-operative mechanical ventilation, ICU admission, and shorter hospital stay. There was no apparent correlation between use of epidural and time to first bowel movement or first oral feed. The benefits of epidural analgesia were not without cost, however, as demonstrated by the significantly longer OR times observed in patients who underwent placement of an epidural catheter.

The benefit from epidural analgesia to reduce systemic opioid administration lasted for the duration of the 96-hour post-operative course that was evaluated in this study. Epidural analgesia appeared to yield its maximum benefit in terms of pain control between 6 and 30 hours post-operatively. Of note, the average duration of epidural analgesia in this cohort was 4 days, which raises the question of whether earlier termination of epidural catheters, and possibly earlier discharge, could be feasible after PE. Although certain advantages of epidural analgesia – namely, earlier and more frequent post-operative extubation and improved pain control – are magnified in the first 30 hours post-operatively, we also observed that systemic opioid administration in patients without epidurals first reached zero at approximately 85 hours after surgery. Taken together, these similar timelines of approximately 96 hours to optimize pain control after PE suggest that patients might still benefit from the epidural through the fourth post-operative day. At our institution, the timing of epidural removal is driven largely by the infant’s ability to tolerate oral feeds. The feasibility and efficacy of terminating epidural analgesia by 96 hours merit further study and support a standardized approach to enhance recovery after PE.

Providers who continue to favor systemic opioids over epidural analgesia deem the risk-to-benefit ratio too high for epidurals (2). Indeed, placement and management of epidural catheters in infants demands the technical expertise of an experienced pediatric anesthesiologist in conjunction with a pediatric pain service familiar with the intricacies of closely monitoring these young patients post-operatively. When administered by experienced teams, epidural analgesia is quite safe. In a review of 307 neonates with epidural catheters for post-operative analgesia reported in the Pediatric Regional Anesthesia Network database, no patient had persistent neurologic deficits, deep infection, spinal cord injury, or epidural hematoma. The most commonly reported complications were catheter malfunction (e.g., dislodgement, kinking), vascular puncture, and catheter contamination. Importantly, no complication resulted in long term sequelae or prolongation of hospitalization (13). Our smaller series also showed no major complications associated with epidural analgesia after PE.

Additional concerns have been raised over decreased epidural drug clearance in infants due to their lower levels of alpha1 acid glycoprotein (AAG), which binds bupivacaine. The most common manifestation of bupivacaine toxicity in infants is cardiac dysrhythmia or respiratory arrest (14). In one study, the combined effect of mild hepatic dysfunction and cholestasis did not appear to affect bupivacaine clearance in patients with biliary atresia, although the lower AAG levels associated with their young age do place these complex patients at risk for drug toxicity (15). Five of the epidural patients in our study experienced bradycardia and/or apnea; however, it is unclear if the arrhythmia was associated with the epidural drugs or some other factor. Importantly, no infant required intervention beyond changes to the epidural drug or infusion rate. One epidural was removed prematurely on post-operative day three due to persistent apneic and bradycardic episodes.

Reduced need for post-operative ventilation, and thus reduced ICU days, along with overall shorter length of hospital stay has the potential to reduce hospital costs significantly despite slightly longer OR times required to place an epidural. Unfortunately, reliable cost data were not available for this study due to variable cost systems in effect across the 15-year study period. However, the use of epidural analgesia has been shown to reduce hospital costs associated with certain procedures in adults, such as pancreatectomy, abdominal wall reconstruction, and open colorectal surgery (16–18). It seems likely that a similar cost benefit would be observed in our patient population for the reasons above, though further investigation is needed to assess this effect.

As with any single-center retrospective cohort study, interpretation of these results is limited by a small sample size and a non-randomized study design. First, for most patients not having epidural analgesia, our records did not provide insight into why an epidural catheter was or was not placed, so we cannot exclude a selection bias between groups that confounds respective outcomes. Importantly, however, our analysis of baseline patient characteristics and pre-operative evaluation did not reveal any significant differences in the preoperative evaluation between study groups. Of note, patients with epidurals were more often female, and they tended to be term infants compared to those without epidurals. This non-significant difference could suggest healthier baseline status among epidural patients, though the lack of difference in any other pre-operative variable studied suggests that the groups were similar at baseline. A second limitation of the study was that pain scores were not recorded at consistent time intervals across patients. We averaged pain scores for each patient over a 6-hour period and then compared medians of the 6-hour averages between intervention groups. If a patient did not have any pain score recorded within a given 6-hour period, a missing data point occurred for that interval. It is likely possible that patients demonstrating less pain had fewer pain assessments, accounting for the absence of certain scores. If missing pain score data are indeed representative of less pain, our median pain scores are over-estimates. Third, although we observed among epidural patients a lower systemic opioid administration, we did not have a means to measure retrospectively the systemic absorption and circulating serum level of opioids and opiate metabolites that were administered through the epidural. Finally, we were unable to account for changes in practice that may have taken place over the 15-year study period, and it is possible that a greater tendency to use epidurals emerged in more recent years given growth of our pediatric surgery and anesthesia departments. Of note, the distribution of epidural and non-epidural patients was similar over the study period, but all patients undergoing PE in the final two years (2015 – 2016, n=5) had an epidural placed. Thus, it is possible that improved outcomes observed in the epidural group were, in part, attributable to other recent changes in operative technique or perioperative management of infants undergoing PE.

In summary, a multicenter randomized control trial would be the most effective way to prove the benefits of post-operative epidural analgesia after PE. However, since no such study has been reported to date, our study fills the current gap in evidence by providing the first report comparing outcomes between patients with and without epidurals after PE. It is also important to note that the approach to postoperative pain management in pediatric patients is evolving rapidly. A multimodal pain management strategy uses alternative techniques in addition to epidural catheters, including non-steroidal anti-inflammatory drugs (NSAIDs), acetaminophen, gabapentin, peripheral nerve blocks, and/or ketamine with the ultimate goal of reducing systemic opioid administration. Future studies should consider other opioid-sparing techniques in addition to epidural catheters for post-operative pain management in this delicate patient population.

Conclusions

Our data support the safety and efficacy of epidural analgesia to control post-operative pain and minimize systemic opioid administration after Kasai PE for BA. Further, patients with epidurals were more often extubated after surgery, had lower pain scores between 6 and 30 hours post-operatively, and required shorter hospital stays. From our observations, we do believe that epidural analgesia provides an important means to stabilize pain control after PE for BA and can set the foundation to enhance recovery after this complex surgery. As with any opiate administration, patients should be carefully monitored post-operatively, particularly for drug toxicity that may be associated with epidural infusion.