Hepaticoduodenostomy versus hepaticojejunostomy after resection of choledochal cyst: A systematic review and meta-analysis

Journal of Pediatric Surgery (2013) 48, 2336–2342

www.elsevier.com/locate/jpedsurg

Review Articles

Hepaticoduodenostomy versus hepaticojejunostomy after resection of choledochal cyst: A systematic review and meta-analysis Sarath Kumar Narayanan a,⁎, Yong Chen b , Kannan Laksmi Narasimhan b , Ralph Clinton Cohen a a

Department of Pediatric Surgery, The Children's Hospital at Westmead, Sydney, NSW 2145, Australia Department of Pediatric Surgery, KK Women's and Children's Hospital, 229899 Singapore

b

Received 25 June 2013; accepted 31 July 2013

Key words: Choledochal cyst; Hepaticoduodenostomy; Hepaticojejunostomy; Meta-analysis

Abstract Background: Excision has been established as a standard management practice for choledochal cysts in the last few decades. The two most commonly performed methods of reconstruction after excision are hepaticoduodenostomy (HD) and Roux-en-Y hepaticojejunostomy (HJ), of which the HJ is favored by most surgeons. Evidence concerning the optimal method of reconstruction is, however, sparse. Materials and Methods: Studies comparing outcomes from HD and HJ after choledochal cyst excision were identified by searching Medline, Ovid, Search Medica, Elsevier Clinicalkey, Google Scholar and Cochrane library. Suitable studies were chosen and data extracted for meta-analysis. Outcomes evaluated included operative time, hospital stay and incidence of postoperative bile leak, cholangitis, reflux/gastritis, anastomotic stricture, bleeding, intestinal obstruction and re-operative rate. Pooled odds ratios (OR) were calculated for dichotomous variables; pooled mean differences (MD) were measured for continuous variables. Results: Six retrospective studies were included in this meta-analysis, comprising a total of 679 patients, 412 of whom (60.7%) underwent HD, and the remainder, 267(39.3%) underwent HJ. Although, HD group had slightly shorter hospital stay (MD: 0.30; 95% CI: −0.22–0.39; P b 0.00001) it showed a higher incidence of postoperative reflux/gastritis (OR: 0.08; 95% CI: − 0.02–0.39; P = 0.002). However, the other outcomes such as bile leak, cholangitis, anastomotic stricture, bleeding, operative time, reoperation rate and adhesive intestinal obstruction did not differ between HD and HJ groups. Conclusions: HD shows higher postoperative reflux/gastritis than HJ but a shorter hospital stay. There are few good-quality studies that compare the outcomes from HD and HJ, meaning that caution should be exercised in the generalization of the results of this meta-analysis, which suggests HD to be comparable with HJ in terms of other complications, operative benefits and outcomes. © 2013 Elsevier Inc. All rights reserved.

⁎ Corresponding author. Department of Pediatric Surgery, The Children's Hospital at Westmead, Cnr Hawkesbury Road and Hainsworth St., Westmead 2145, Sydney, NSW, Australia. Tel.: + 61 02 4706 91149; fax: +61 02 9845 3180. E-mail addresses: [email protected] (S.K. Narayanan), [email protected] (Y. Chen), [email protected] (K.L. Narasimhan), [email protected] (R.C. Cohen). 0022-3468/$ – see front matter © 2013 Elsevier Inc. All rights reserved. http://dx.doi.org/10.1016/j.jpedsurg.2013.07.020

Hepaticoduodenostomy vs hepaticojejunostomy Choledochal cysts are congenital bile duct anomalies with cystic dilatation of the biliary tree and have potential for various complications. Most of the reported cases in the world come from Asia but it is not uncommon in the western world. Complete resection of the extra hepatic bile duct is the accepted management of choledochal cyst [1]. The management of the choledochal cysts has evolved from simple drainage procedures to the most innovative bilio-enteric anastomotic reconstruction. Biliary reconstruction is now commonly performed by one of several techniques, including Roux-en-Y hepaticojejunostomy (HJ), hepaticoduodenostomy (HD), and jejunal interposition hepaticoduodenostomy [2] and [3]. Debate continues regarding the optimal method of biliary reconstruction. With the increasing trend of laparoscopic surgery for choledochal cysts, HD has become more popular among laparoscopic surgeons. Apart from the technical ease of a single anastomosis, the operation has been cited to have several other advantages such as being more physiologic and allows postoperative endoscopic access to the anastomosis if a stricture or stone occurs. However, conventional hepatobiliary surgeons most often prefer a Roux-en-Y reconstruction as the operation has a very long history of safety and is the gold standard for biliary reconstruction. Early attempts to have jejunum loop interposition hepaticoduodenostomy resulted in an unacceptable level of bile gastritis postoperatively that the Japanese surgeons who pushed for this operation gave up the procedure. The current popularity of HD is not based on any fresh evidence in its favor. To compare the clinical outcomes of HD and HJ, we performed a systematic review and comprehensive metaanalysis from indexed published literature.

2337

1.2. Data extraction Three reviewers (SKN, KLN and CY) independently assessed selected studies, extracted and tabulated data from each article: first author, year of publication, study design, follow up period, mean age and number of subjects operated on with each technique, and end point data (operative times, length of postoperative hospital stay, re-operation rates and postoperative complications including bile leak, adhesive intestinal obstruction, cholangitis, reflux/gastritis and anastomotic stricture). The reviewers reached consensus at each stage of the screening process.

1.3. Inclusion criteria To enter the analysis, studies had to (1) report the reconstructive techniques after excision of choledochal cyst; (2) have comparative data for HD and HJ operations available; (3) report on at least one of the outcome measures mentioned below; and (4) have been published as a full paper in a journal, not as meeting abstract or review.

1.4. Exclusion criteria The following criteria were used to exclude studies: (1) studies in which no clear distinction was made between individual complications of reconstruction (post-operative cholangitis, reflux/gastritis, bile leak, adhesive obstruction and anastomotic stricture) and type of operation performed; (2) studies which reported on interposition type or other modifications of HD or other methods of reconstruction and (3) study overlapped with previous report from same institution. In that case, the most informative and/or recent article was chosen for meta-analysis.

1. Materials and methods 1.5. Statistical analysis 1.1. Study selection A Medline, Ovid, Search Medica, Elsevier ClinicalKey, Google Scholar and the Cochrane library search was performed on all studies published between 1950 and 2012 reporting on comparison of HD and HJ for children undergoing excision of extra-hepatic biliary tree for choledochal cysts. The following keywords were used: “choledochal cyst”, “choledochal cyst AND hepaticoduodenostomy”, “choledochal cyst AND hepaticojejunostomy”, “hepaticoduodenostomy”, “hepaticojejunostomy”, “hepaticodochoduodenostomy”. The “related articles” function was used to widen the search. The reference lists of the full articles were also manually searched to identify additional eligible studies. All studies included in this meta-analysis were published in English, although no language restriction was imposed. The latest search was performed in December 2012.

Meta-analysis was performed in line with recommendations from the Cochrane Collaboration and the Quality of Reporting of Meta-analyses guidelines [4,5] and [6]. Pooled odds ratios (OR) were calculated for dichotomous variables using the Mantel–Haenszel method and pooled mean differences (MD) were measured for continuous variables using the inverse variance method in meta-analysis. The odds ratio represents the odds of an adverse event occurring in the treatment (HJ) group compared with the reference (HD) group. The confidence interval (CI) was established at 95% and P values of less than 0.05 were considered statistically significant and an odds ratio of less than 1 favors the treatment group. Statistical heterogeneity was assessed using I2. A fixed effects model was used if I2 b 50% and a random effects model was used if I2 ≥ 50%. Statistical analysis was performed using Review Manager 5.2 (Cochrane Collaboration).

NA d: days y: years

27 d–25 y

4.25

NA m: months

NA

NA

NA NA

7.1 ± 6 9.1 ± 5.8 NA 235 (open) 307 (open) NA

3 m–6 y 4 m–9 y 7.9 ± 1.5 8.7 ± 5.4 NA 5.5 3.1 7.4 ± 2 4.0 ± 3.9 NA 32 22 NA

5.18

6 m–17y

164 ± 51 220 ± 60 NA 2.1 ± 0.2

Retrospective Todani 1981

HD = hepaticoduodenostomy

Retrospective

Retrospective

Shimotokahara 2005 Takada 2005

Retrospective

Retrospective

Mukhopadhyay 2011 Santore 2011

Number of patients

Retrospective

Four analyzed studies gave data on rates of inflammation of the stomach lining as determined clinically [15,16,18,19]

Table 1

2.4. Reflux/gastritis

Liem 2012

As reported in an analyzable form by 3 studies [17,18,19], anastomotic stricture occurred in 1.21% (4/330) of HD and 1.47% (3/204) of HJ patients. There was no statistical difference between these 2 groups. (OR: 1.45; 95% CI: − 0.36–5.79; P = 0.60) (Fig. 3).

Study type

2.3. Anastomotic stricture

Study (Author, Year)

The incidences of cholangitis were described in an analyzable form between the 2 groups of interest in all series [14–18,19]. In 2 of these individual articles, the rates of cholangitis were statistically significantly higher in the HD group while HJ group showed higher rates of cholangitis in 3 other series. In total, cholangitis was reported in 2.47% (9/364) of the HD patients and 2.42% (6/248) of the HJ patients. Pooled OR did not show statistical difference of cholangitis between these 2 groups. (OR: 1.07; 95% CI: − 0.41–2.81; P = 0.89) (Fig. 2).

Characteristics of 6 studies included in the meta-analysis.

2.2. Cholangitis

Laparoscopic method No. (%)

Cystic diameter (mm)

Three studies [17,18] and [19] reported on postoperative bile leak with an incidence of 2.1% (7/330) in the HD group and 2.94% (6/204) in HJ group. There is no statistically significant difference between HD and HJ groups (OR: 1.50; 95% CI; − 0.51–4.39; P = 0.46) (Fig. 1).

3.96

2.1. Bile leak

47.5 ± 15.4 48.3 ± 16.0 NA

Age at operation (Mean ± SD, year)

Follow up time (Mean ± SD, year)

Operative time (Mean ± SD, min)

The initial search yielded 14 articles meeting the inclusion criteria. Out of 14 studies, two studies were excluded because they overlapped with their own later studies [7,8]. Five studies were excluded as they included comparisons of interposition hepaticoduodenostomy or one of its modifications [2,9–11,12]. One study was excluded as the comparison was unclear [13], leaving 6 studies for meta-analysis [14–18,19]. All six included studies are retrospective series. While five studies were published after 2005 one was published in 1981. They included total of 679 patients. Of these, 412 (60.7%) underwent HD, and 267 (39.3%) had HJ. The characteristics of these studies are demonstrated in Table 1. The age of patients varied from 27 days to 25 years. Follow-up was explicitly stated in 5 studies and ranged from 3 to 204 months. Laparoscopic surgeries were performed in majority of patients (99.5%) in one study and in 15% of HD cases in other studies. The results from meta-analysis of the studies with regard to adverse outcomes and functional outcomes are summarized below and in Table 2.

HD: 238 398 (99.5) HJ: 162 HD: 53 0 (0%) HJ: 22 HD: 39 6 (15) HJ: 20 0 (0) HD: 12 NA HJ: 28 HD: 3 NA HJ: 5 HD: 19 NA HJ: 11 HJ = hepaticojejunostomy

2. Results

6.4 ± 0.3 6.7 ± 0.5 NA

S.K. Narayanan et al. Hospital stay (Mean ± SD, days)

2338

Hepaticoduodenostomy vs hepaticojejunostomy Table 2

2339

Outcomes of hepaticoduodenostomy and hepaticojejunostomy.

Study (Author, Year)

Number Bleeding Bile leak Cholangitis Gastritis/gastric Anastomotic Adhesive Reoperation of patients (%) (%) (%) reflux (%) stricture (%) I/O (%) (%)

Liem 2012

HD: 238 0 (0) 4 (1.7) HJ: 162 1 (0.4) 4 (2.5) Mukhopadhyay HD: 53 NA 1 (1.9) 2011 HJ: 22 0 (0) Santore 2011 HD: 39 NA 2 (5) HJ: 20 2 (10) Shimotokahara HD: 12 NA NA 2005 HJ: 28 Takada 2005 HD: 3 NA NA HJ: 5 Todani 1981 HD: 19 0 (0) NA HJ: 11 1 (9.1) HD = hepaticodudenostomy HJ = hepaticojejunostomy

or that was established on the endoscopic examination (Takada et al.). There were no instances of reflux or gastritis in the HJ group (0/217, 0.0%) while all reported cases of reflux/gastritis came from the HD group (18/306, 5.88%). On meta-analysis, significant differences between the groups were found within the studies overwhelmingly favoring HJ (OR: 0.08; 95% CI: − 0.02–0.39; P = 0.002) (Fig. 4).

2.5. Adhesive intestinal obstruction Two studies [14,15] gave details of incidence of adhesive intestinal obstructions with a total of 5.12% (2/39) in the HJ

Fig. 1

5 1 0 2 0 1 0 1 0 0 4 1

(1.7) (2.5) (0) (9.1) (0) (5) (0) (3.6) (0) (0) (21) (9.1)

8 (3.8) 0 (0) 3 (5.7) 0 (0) 0 (0) 0 (0) 4 (33.3) 0 (0) 3 (100) 0 (0) 20

NA NA

NA NA NA 0 (0) 1 (3.6) NA 0 (0) 1 (9.1)

3 (1.3) 0 (0) 0 (0) 1 (4.5) 1 (2.5) 4 (20) NA NA NA

group compared with 0% (0/31) in the HD group. However, there is no statistically significant difference between these two groups (OR: 2.77; 95% CI: − 0.27–27.92; P = 0.39) (Fig. 5).

2.6. Reoperation rates Three studies [17,18,19] cited about reoperation in their series. The reoperation rate in HD group is 1.21% (4/330) while it was 2.45% (5/204) in HJ group. There is no statistically significant difference (OR: 2.14; 95% CI: − 0.67–6.89; P = 0.20) (Fig. 6).

Forest plot comparison, post-operative bile leak.

Fig. 2

2 (0.8) 0 (0) 1 (1.9) 2 (9.1) 1 (2.6) 1 (5) NA

Forest plot comparison, cholangitis.

2340

S.K. Narayanan et al.

Fig. 3

Fig. 4

Forest plot comparison, anastomotic stricture.

Forest plot comparison, reflux/gastritis.

2.7. Other points of interest 2.7.1. Operative time Two studies reported operative time (Table 1). Both studies showed a shorter operative time in HD group. However, one study reported only mean operative time without standard deviation. Therefore, pooled mean differences were unable to be calculated based on current studies. 2.7.2. Hospital stay Two studies [17,19] reported the times of hospital stay (Table 1), both studies showed a slightly shorter period in the HD group. Pooled mean difference (MD: 0.30; 95% CI: − 0.22–0.39; P b 0.00001) shows that the difference is statistically significant (Fig. 7). Santore et al. noted that although there was a trend toward earlier discharge in the HD patients there was no difference between the 2 groups in duration of epidural pain control. Importantly, the HD group had a trend toward shorter time to resumption of a regular diet than the HJ patients (HD =

Fig. 5

4.8 days vs. HJ = 6.1 days; P = 0.08), but this was not significant and both groups had similar intervals of nasogastric decompression. Mortality was not reported by any authors related to the surgery except one group, Mukhopadhyay et al. Their patient had liver cirrhosis at the time of operation (operated at the age of 4 months) that eventually went into liver failure and died after excision of cyst and HJ.

3. Discussion The choice between HD and HJ is still a matter of debate. Our meta-analysis suggests that the HD is comparable to conventional HJ in most postoperative outcomes except a higher rate of gastric reflux. To our knowledge this study is the first systematic review and meta-analysis comparing HD to HJ. HD is more physiologic, but theoretically, the closeness of hepaticoenterostomy to stomach makes HD to have greater chance of cholangitis and bile gastritis. In our meta-

Forest plot comparison, adhesive obstruction.

Hepaticoduodenostomy vs hepaticojejunostomy

Fig. 6

Fig. 7

2341

Forest plot comparison, re-operation rates.

Forest plot comparison, hospital stay.

analysis, none of HJ patients developed bile gastritis, but close to 6% of HD patients had symptoms of gastritis. The incidence of bile gastritis after HD was even higher when examined endoscopically. Interestingly, there is no difference of cholangitis between these two groups. The bile leak and anastomotic stricture are also comparable between HD and HJ. As a more simple procedure compared to HJ, HD does exhibit some advantages such as shorter operative time and hospital stay. It is also easier to perform them laparoscopically, which in turn provides a superior cosmetic result. The procedure can be performed entirely above the transverse colon mesentery, which reduces the potential risk of postoperative adhesions [7]. Our result showed that HJ has higher chance of postoperative intestinal obstruction (5.12%) compared to the HD group (0%). However, the difference is not statistically significant, which could be due to the small patient number in the studies. Based on the evidence of this study, HD as a method of reconstruction after excision of a choledochal cyst is a suitable alternative to HJ. Though reflux/gastritis is higher in the HD group, it could possibly be avoided with a construction of the HD anastomosis at the junction of the first and second portions of the duodenum after an extensive Kocher maneuver to prevent any tension on the anastomosis. This should be well distal to the pylorus and should not impact pyloric function or gastric emptying significantly [17]. There are few limitations in this study. The lack of a randomized trial comparing the two procedures is a key problem. All six included studies are retrospective observational studies which are prone to selection bias and may result in uneven distribution of confounding factors such as age of patient, duration of follow-up and type of choledochal cyst.

Caution should be applied in the interpretation and generalization of this meta-analysis as the numbers are not high and also the overall quantity of these studies is insufficient. Most of the choledochal cyst operations are done in childhood. The longest follow up in the included studies is only seven years. However, some severe complications such as cholangiocarcinoma only develop more than 10 years later. Therefore, apart from randomized trials, studies with long-term follow-up are also necessary to explore the benefits of HD or HJ in the management of choledochal cysts.

References [1] Davidson PM, Auldist AW. Contemporary surgical treatment for choledochal cysts. Pediatr Surg Int 1987;2:157-60. [2] Cosentino CM, Luck SR, Raffensperger JG, et al. Choledochal duct cyst: resection with physiologic reconstruction. Surgery 1992;112(4): 740-7. [3] Edil BH, Olino K, Cameron JL. The current management of choledochal cysts. Adv Surg 2009;43:221-32. [4] Clarke M, Horton R. Bringing it all together: Lancet–Cochrane collaborate on systematic reviews. Lancet 2001;357:1728. [5] Moher D, Cook DJ, Eastwood S, et al. Improving the quality of reports of meta-analyses of randomized controlled trials: the QUOROM statement. Quality of Reporting of Meta-analyses. Lancet 1999;354: 1896-900. [6] Mantel N, Haenszel W. Statistical aspects of the analysis of data from retrospective studies of disease. J Natl Cancer Inst 1959;22:719-48. [7] Nguyen Thanh L, Hien PD, Dung le A, et al. Laparoscopic repair for choledochal cyst: lessons learned from 190 cases. J Pediatr Surg 2010;45(3):540-4. [8] Todani T, Watanabe Y, Urushihara N, et al. Biliary complications after excisional procedure for choledochal cyst. J Pediatr Surg 1995;30(3): 478-81.

2342 [9] Li MJ, Feng JX, Jin QF. Early complications after excision with hepaticoenterostomy for infants and children with choledochal cysts. Hepatobiliary Pancreat Dis Int 2002 May;1(2):281-4. [10] Hara H, Morita S, Ishibashi T, et al. Surgical treatment for congenital biliary dilatation, with or without intrahepatic bile duct dilatation. Hepatogastroenterology 2001;48(39):638-41. [11] Okada A, Higaki J, Nakamura T, et al. Roux-en-Y versus interposition biliary reconstruction. Surg Gynecol Obstet 1992;174(4):313-6. [12] Okada A, Ohguchi Y, Kamata S, et al. Jejunal interposition hepaticoduodenostomy for congenital dilatation of the bile duct (choledochal cyst). J Pediatr Surg 1983;18(5):588-91. [13] Mishra A, Pant N, Chadha R, et al. Choledochal cysts in infancy and childhood. Indian J Pediatr 2007 Oct;74(10):937-43. [14] Todani T, Watanabe Y, Mizuguchi T, et al. Hepaticoduodenostomy at the hepatic hilum after excision of choledochal cyst. Am J Surg 1981;142(5):584-7.

S.K. Narayanan et al. [15] Shimotakahara A, Yamataka A, Yanai T, et al. Roux-en-Y hepaticojejunostomy or hepaticoduodenostomy for biliary reconstruction during the surgical treatment of choledochal cyst: which is better? Pediatr Surg Int 2005;21(1):5-7. [16] Takada K, Hamada Y, Watanabe K, et al. Duodenogastric reflux following biliary reconstruction after excision of choledochal cyst. Pediatr Surg Int 2005;21(1):1-4. [17] Santore MT, Behar BJ, Blinman TA, et al. Hepaticoduodenostomy vs hepaticojejunostomy for reconstruction after resection of choledochal cyst. J Pediatr Surg 2011;46(1):209-13. [18] Mukhopadhyay B, Shukla RM, Mukhopadhyay M, et al. Choledochal cyst: a review of 79 cases and the role of hepaticodochoduodenostomy. J Indian Assoc Pediatr Surg 2011;16(2):54-7. [19] Liem NT, Pham HD, Dung le A, et al. Early and intermediate outcomes of laparoscopic surgery for choledochal cysts with 400 patients. J Laparoendosc Adv Surg Tech A 2012;22(6):599-603.