Systematic review of cystic duct closure techniques in relation to prevention of bile duct leakage after laparoscopic cholecystectomy.

AIM: To study the effect of different techniques of cystic duct closure on bile leakage after laparoscopic cholecystectomy (LC) for biliary disease.
METHODS: A systematic search of MEDLINE, Cochrane and EMBASE was performed. Rate of cystic duct leakage (CDL) was the primary outcome. Risk of bias was evaluated. Odds ratios were analyzed for comparison of techniques and pooled event rates for non-comparative analyses. Pooled event rates were compared for each of included techniques.
RESULTS: Out of 1491 articles, 38 studies were included. A total of 47491 patients were included, of which 38683 (81.5%) underwent cystic duct closure with non-locking (metal) clips. All studies were of low-moderate methodological quality. Only two studies reported separate data on uncomplicated and complicated gallbladder disease. For overall CDL, an odds ratio of 0.4 (95%CI: 0.06-2.48) was found for harmonic energy vs clip closure and an odds ratio of 0.17 (95%CI: 0.03-0.93) for locking vs non-locking clips. Pooled CDL rate was around 1% for harmonic energy and metal clips, and 0% for locking clips and ligatures.
CONCLUSION: Based on available evidence it is not possible to either recommend or discourage any of the techniques for cystic duct closure during LC with respects to CDL, although data point out a slight preference for locking clips and ligatures vs other techniques. No separate recommendation can be made for complicated gallbladder disease.

Core tip: Cystic duct leakage (CDL) is an unwanted, potential life threatening complication of laparoscopic cholecystectomy (LC). Several techniques have been described to securely close the cystic duct during cholecystectomy. Based on available evidence from this systematic review, it is not possible to either recommend or discourage any of the techniques for cystic duct closure during LC with respect to CDL, although data point out a slight preference for locking clips and ligatures vs other techniques. No separate recommendation can be made for complicated gallbladder disease.

INTRODUCTION

Laparoscopic cholecystectomy (LC) is the standard treatment for gallstone disease and associated with low morbidity and mortality. Cystic duct leakage (CDL) is reported in 0.5%-3% of patients following LC[1,2]. Recent articles show that CDL increases to 4%-7% in patients with complicated gallstone disease, such as cholecystitis, pancreatitis, cholangitis and stones in the common bile duct[3-5].

Adequate closure of the cystic duct is essential to prevent CDL, especially in patients with complicated gallstone disease having a higher risk of bile leakage. Although CDL is classified as a minor injury of the bile ducts, it is associated with significant re-intervention rate, increased morbidity and even mortality[6-8]. The most common closure technique during LC is by simple (non-locking) metallic clips. Alternatives are locking clips (e.g., Hem-o-Lok®, Teleflex, Wayne, United States; Lapro-Clip®, Medtronic, Minneapolis, United States; Click’aV®, Grena, Nottingham, United Kingdom) or ligatures (e.g., ENDOLOOP®, Ethicon, Somerville, United States)[4,5]. Locking clips differ from metallic clips as they are made of polymers, are usually absorbable and are designed to lock in place with comparable locking pressure and therefore are thought to provide a more secure closure.

With the introduction of vessel sealing devices, such as LigaSureTM (Medtronic, Minneapolis, United States) and the Harmonic scalpel® (Ethicon, Somerville, United States), their feasibility and outcome in closing the cystic duct during LC are of interest. The aim of this systematic review was to assess the risk of CDL and the CDL rate for different techniques of cystic duct closure after LC, both in uncomplicated and complicated gallbladder disease.

MATERIALS AND METHODS

This review was performed according to the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA), No review protocol was devised before the commencement of this review[9].

Eligibility criteria

Studies were eligible for inclusion when: (1) patients receiving LC for gallstone disease were included; (2) the technique of closure of the cystic duct was described; (3) leakage of the cystic duct was described as an outcome; (4) the article reported original data and the full-text was available; and (5) the study included a minimum of 50 patients per closure technique given the estimated low event rate of CDL. Only studies published following the introduction of LC (1985 and later) were included. No further limits, such as study design and language were predefined. The original author was contacted when the full text of an article was not available in the clinical library.

S

The databases of PubMed/MEDLINE, the Cochrane Library and EMBASE were systematically searched (in July 2018). The complete search is listed in Appendix 1. In summary, the following search terms and synonyms were used; LC, closure of the cystic duct and CDL. All possible techniques of the closure of the cystic duct were also included separately, such as “clips” and “loops”. Free text words were also searched for every keyword. References of the included articles were searched for possible relevant studies.

Selection of studies

Available articles were independently screened for inclusion by two reviewers (van Dijk AH and van Roessel S) on title and abstract. All duplicates were removed. After selection of eligible articles, full-text was assessed independently by both reviewers according to the eligibility criteria.

Collection of data

The following data was extracted from the articles by both reviewers independently; authors, year and country of publication, study design, number of included patients, number of patients with complicated gallstone disease, type of closure technique, failure of closure technique, operating time, duration of hospital stay, leakage of the cystic duct, biloma, bile duct injuries, and intra-abdominal abscesses.

Complicated gallstone disease was defined as patients with cholecystitis, biliary pancreatitis, cholangitis or patients that underwent ERCP prior to LC with or without choledocholithiasis. The failure of the closure technique was defined as the need to use another technique to safely secure the cystic duct during LC, such as placing extra clips or adding a loop. Bile duct injuries were classified according to the Amsterdam classification[10]. Patients who underwent cystic duct closure with an endoloop, a ligature or intracorporeal knot were analyzed in the “ligature” group.

Both reviewers assessed risk of bias. The MINORS score was used to assess bias in non-randomized studies, of which the maximum score is 16[11]. When a study ranked 0-5 points it was listed as high risk of bias, 6-10 as moderate risk of bias and higher than 10 of low risk of bias. For the randomized studies the risk of bias in accordance with the Cochrane Library guideline[12].

Main outcomes

The main outcome of this study was overall CDL and for the subgroups of uncomplicated and complicated gallbladder disease for each technique used to close the cystic duct during LC. Secondary outcomes were the failure of the technique, defined as the need for another technique to safely close the cystic duct, and the occurrence of short-term complications such as a biloma, bile duct injury or intra-abdominal abscess. Operating time and duration of hospital stay were also collected for comparative studies.

Statistical analysis

A proportion of events and a 95%CI were calculated for the main outcome. Comparative studies were grouped per study design and analyzed with the Peto method using Review Manager by the Cochrane Library (RevMan, version 5.3, the Cochrane Collaboration, 2014). This was displayed in a forest plot. The Peto odds ratio methods is commonly used in rare events and was found to be the best performing method[12]. The non-comparative studies were analyzed using R-biomedical statistics (version 3.1.1) and an estimated pooled event rate was calculated with a random effects model and displayed in a forest plot.

A subgroup analysis was planned for the main outcome in the subgroup of complicated vs non-complicated patients. Data from one single arm of a randomized study was used for meta-analysis, if only one single arm of a study could be included for one of the outcomes. Statistical heterogeneity between studies was assessed by viewing the forest plots and the I2 test and shown per study design. An I2 test with a value of 45% and lower was considered as low heterogeneity, of 46%-75% of moderate and above 75% as high. To assess publication bias funnel plots were created.

RESULTS

A total of 1491 articles were defined by searching the databases of MEDLINE, EMBASE and the Cochrane Library. The flow diagram is depicted in Figure 1. After removing 118 duplicates 1271 references were deemed irrelevant after reading the title and abstract. The remaining 102 full texts were assessed for eligibility according to the eligibility criteria. A total of 38 articles met the inclusion criteria and were included in present systematic review. A total of 64 articles were excluded, as they did not meet the inclusion criteria (e.g., case report or conference abstract, see Appendix 2).

Figure 1

PRISMA flow diagram of the inclusion process (last search July 5, 2018).

Included studies

Of the 38 included studies, 14 were used in the comparative analysis and 24 in the non-comparative analysis. Of the comparative studies, 4 were randomized controlled trials[13-16], 3 were prospective[17-19] and 7 had a retrospective design[20-27] (Tables 1 and 2). Six studies compared the harmonic energy with metal clips[13-15,22,23,26], locking vs non-locking clips were the subject of four studies[20,21,24,25] and the other three covered either ligatures[16,19] or a vessel sealing device[17,18].

Table 1

Characteristics of included comparative studies, descending in year of publication

Study, year, country	Study design	Intervention	Control	Aim study
Jain[27], 2011, India	RCT	Ultrasonic shears	Electrocautery	To test the benefit of ultrasonic shears in LC
Redwan[13], 2010, Egypt	RCT	Harmonic shear (Olympus Keymed Sono surg version G2 220–240V 3A)	Titanium clips	To demonstrate the efficiency and safety of the harmonic scalpel
Kandil[14], 2010, Egypt	RCT	Harmonic scalpel (Harmonic ACE, Ethicon Endo-Surgery)	Metal clips	To compare metal clips vs the harmonic scalpel on safety and efficacy in LC
Bessa[15], 2008, Egypt	RCT	Harmonic scalpel (Harmonic ACE, Ethicon Endo-Surgery)	Clip and cautery	To compare the safety and efficacy of the harmonic scalpel vs clip and cautery in LC
Seenu[16], 2004, India	RCT	Absorbable ligature (Vicryl1, Ethicon)	Titanium clips (Ligaclip, Ethicon)	To compare postoperative outcomes after occlusion of the cystic duct with tied knots vs titanium clips
Singal[19], 2018, India	PS	Non-absorbable ligature (Filasilk, Meril)	Titanium clips (Ligaclip, Ethicon)	To study safety and efficacy of silk ligatures compared to clips of closure of the cystic duct
Schulze[17], 2010, Denmark	PS	LigaSure (ForceTriad system, Valleylab)	Titanium clips	To evaluate the safety of the LigaSure system in cholecystectomy
Hüscher[18], 2003, Italy	PS	Harmonic shears (Ultracision, Ethicon Endo-Surgery)	Harmonic shears and absorbable endo-loop	To verify the advantages if ultrasonic dissection
Yang[21], 2014, China	RS	One absorbable clip	Titanium clips	The effectiveness and safety of electrocoagulation after occlusion of the cystic duct and artery with an absorbable clip
Wills[22], 2013, USA	RS	Harmonic scalpel (Harmonic ACE, Ethicon Endo-Surgery) in pts with a cystic duct > 5 mm	Single surgical clip in pts with cystic duct < 5 mm	The comparison of the Harmonic scalpel vs surgical clips in the occlusion of the cystic duct
Matsui[20], 2012, Japan	RS	Locking absorbable clips (Laproclip 8 mm, 12mm, Tyco Healthcare) Locking non-absorbable clip (Hem-o-lok XL, Teleflex Medical)	Endo-loop (SURGITIE, Tyco Healthcare) Suture Metallic clip (ENDO CLIP III, Tyco Healthcare)	To evaluate the effect of locking clips on the leakage from the cystic duct in cholecystectomy
Wu[26], 2011, China	RS	Ultrasonic shears (Harmonic ACE, Ethicon Endo-Surgery)	Clips	To compare conventional LC to SILC
Gelmini[23], 2010, Italy	RS	Harmonic scalpel (Harmonic- Ethicon Endo Surgery)	Clips	To demonstrate that the harmonic scalpel is safe in LC
Rohatgi[24], 2006, United Kingdom	RS	Absorbable locking clips (Laproclip, USS-DG, Tyco)	Titanium clips (Ligaclip, Ethicon)	To compare the efficacy of the locking absorbable clip vs clips in LC
Yano, 2003, Japan[25]	RS	Locking absorbable clips (Laproclip, Davis and Geck)	Ligaclip (metal clip, Ethicon)	To assess if locking clips are safer and less invasive than metal Ligaclips

RS: Retrospective studydesign; RCT: Randomized controlled trial; PS: Prospective study design; LC: Laparoscopic cholecystectomy.

Table 2

Patient and operative characteristics of included comparative studies

Study, year, country	No. of patients	No. of complicated cases	Failure of technique in intervention group	Operating time	Hospital stay
Harmonic scalpel/ shears vs metal clips
Wills[22], 2013, United States	57 vs 148	NR	3	NR	NR
Wu[26], 2011, China	100 vs 100	0 (exclusion criterium)	NR	Mean in minute (SD) 49.2 (13.8) vs 53.3 (24)	NR
Redwan[13], 2010, Egypt	80 vs 80	NR	NR	Mean in minute (SD) 20 (6.8) vs 45 (6.5)	Mean in days (SD) 1 (0.0) vs 1.5 (0.51)
Kandil[14], 2010, Egypt	70 vs 70	NR	NR	Mean in minute (SD) 33.2 (9.6) vs 51.7 (13.8)	Mean in hours (SD) 23.4 (2.29) vs 267.0 (8.94)
Gelmini[23], 2010, Italy	95 vs 90	28 vs 22	17	Median in minute (range) 60 (20-140) vs 80 (45-130)	Median in days (range) 2 (1-16) vs 2 (1-12)
Bessa[15], 2008, Egypt	60 vs 60	0 (exclusion criterium)	NR	Median in minute (range) 32 (18-75) vs 40 (21-85)	NR
Absorbable vs non-absorbable clips
Yang[21], 2014, China	635 vs 728	545 vs 626	NR	Mean in minute (SD) 41.6 (16.5) vs 58.9 (19.4)	2.6 (0.4) vs 2.7 (0.6)
Locking vs non-locking clips
Matsui[20], 2012, Japan	907 vs 110	85 (unknown in which group)	5	NR	NR per group
Rohatgi[24], 2006, United Kingdom	346 vs 148	NR	2	NR	NR
Yano, 2003, Japan[25]	328 vs 444	8 vs 9	0	Mean in minute (SD) 84.6 (1.6) vs 112.7 (2.3)	Mean in days (SD) 7.9 (0.2) vs 8.0 (0.1)
Other
Singal[19], 2018, India	70 silk ligature vs 70 titanium clips	0 (exclusion criterium)	0	NR	NR in detail (“similar”)
Schulze[17], 2010, Denmark	101 Ligasure vs 113 titanium clips	Only elective surgery	NR	NR	NR
Seenu[16], 2004, India	53 absorbable ligature vs 52 titanium clips	NR	NR	Mean in minute 78 vs 66	NR
Hüscher[18], 2003, Italy	331 harmonic shears vs 130 harmonic shears + endoloop	109 vs 68	NR	Mean in minute 76.8 vs 97.5	Mean in days 4.3 vs 5.1

NR: Not reported.

Non comparative analysis consisted of data from one single arm of a randomized trial (25), sixteen retrospective studies[28-43] and 7 studies with a prospective design[44-50]. A single arm of an RCT and 5 other studies reported on harmonic energy[27-29,44-47]. Closure with ligatures was described in six articles[36-38,48-50] and the use of metal clips in nine[30-33,35,40-43]. Only one article described locking clips as a closure technique[39] and one study reported on the LigaSure technique[34]. Only one study with a sufficient number of patients on the use of a stapler as closure technique was available[28].

Risk of bias

Not one of the studies scored higher than 10 on the Minors score. All studies were of low to moderate methodological quality, as shown in Appendix 2A and B. The 5 randomized studies were of a high risk of bias, as shown in Figure 2. The funnel plots in Appendix 4 showed some asymmetry, indicating that publication bias cannot be excluded.

Figure 2

Risk of bias of included randomized comparative studies.

Outcomes

In this review a total of 47491 patients were included, all of which underwent LC. In 38683 (81.5%) patients the cystic duct was closed with (metal, non-locking) clips, a ligature was used in 3604 (7.6%) patients, 1853 (3.9%) patients received locking clips, in 1692 (3.6%) patients the cystic duct was closed by harmonic energy, 1299 received an absorbable clip (2.7%) and 230 (0.5%) cystic ducts were divided by LigaSure. In 130 patients (0.2%) harmonic energy and a ligature were used simultaneously.

The patient and operative characteristics of the comparative analysis are shown in Table 2. Eight of 14 studies included patients with complicated gallstone disease, but none reported separate data on the effect of the used technique in patients with complicated vs uncomplicated gallstone disease. Also, only five studies reported failure of the used technique.

Table 3 shows the clinical outcome of the 14 comparative studies and shows that the morbidity rate after LC is low. Table 4 shows the patient and operative characteristics and clinical outcomes for the non-comparative studies. Eleven of 24 studies did not report patients with complicated gallstone disease and 6 out of 24 did not report failure of the technique. The overall morbidity rate was low.

Table 3

Clinical outcomes of included comparative studies

Study, year, country	No. of patients	Leakage of the cystic duct	Bile duct injury	Biloma	Intra-abdominal abcess
Harmonic scalpel/shears vs metal clips
Wills[22], 2013, United States	57 vs 148	1 vs 1	1 (D) vs 0	1 vs 0	0 vs 0
Wu[26], 2011, China	100 vs 100	0 vs 1	0 vs 0	0 vs 0	0 vs 0
Redwan[13], 2010, Egypt	80 vs 80	0 vs 1	0 vs 0	0 vs 0	0 vs 0
Kandil[14], 2010, Egypt	70 vs 70	0 vs 1	0 vs 0	NR	NR
Gelmini[23], 2010, Italy	95 vs 90	0 vs 0	0 vs 0	NR	2 vs 0
Bessa[15], 2008, Egypt	60 vs 60	0 vs 0	0 vs 0	NR	NR
Locking absorbable vs locking non-absorbable clips
Yang[21], 2014, China	635 vs 728	0 vs 7	NR	NR	1 vs 2
Locking vs non-locking clips
Matsui[20], 2012, Japan	907 vs 110	0 vs 0	NR	NR	NR
Rohatgi[24], 2006, United Kingdom	344 vs 146	0 vs 3	NR	2 vs 2	NR
Yano[25], 2003, Japan	328 vs 444	1 vs 2	2 vs 4 (severity not reported)	NR	NR
Other
Singal[19], 2018, India	70 silk ligature vs 70 titanium clips	0 vs 0	NR	NR	NR
Schulze[17], 2010, Denmark	101 Ligasure vs 113 titanium clips	0 vs 0	0 vs 0	NR	NR
Seenu[16], 2004, India	53 absorbable ligature vs 52 titanium clips	2 vs 2	NR	NR	NR
Hüscher[18], 2003, Italy	331 harmonic shears vs 130 harmonic shears + endoloop	7 vs 3	1 (D) vs 0	NR	0 vs 1

Amsterdam classification was used to identify the severity of the bile duct injuries: B: Major bile duct leaks with or without concomitant biliary strictures; C: Bile duct strictures without bile leakage; D: Complete transection of the duct with or without excision of some portion of the biliary tree.

Table 4

Outcomes of non-comparative studies

Study, year, country	Study design		Intervention	No. of patients	No. of complicated cases	CDL	Biloma/abcess	BDI	Failure of technique
Harmonic scalpel/ shears /stapler
Jain[27], 2011, India	RCT5		Harmonic shears	100	0 (exclusion criterium)	0	0	NR	NR
Ramos[44], 2015, Brazil	PS		Harmonic shears	125	0 (exclusion criterium)	0	NR	NR	NR
Patel[45], 2010, United Kingdom	PS		Harmonic scalpel (LCS-5, Ethicon)2	100	NR	1	NR	1 (D)	NR
Westervelt[46], 2004, United States	PS		Harmonic scalpel (unknown)	100	NR	0	NR	NR	2
Power[47], 2000, Ireland	PS		Harmonic scalpel (Ultracision, Ethicon Endo-Surgery) and clip	282	101	2	NR	NR	NR
Lee[28], 2011, South Korea	RS		Endo-GIA (US Surgical Corp.)1	921	90	0	NR	4 (D)	NR
Tebala[29], 2006, Italy	RS		Harmonic shears (Ultracision, Ethicon Endo- Surgery)	100	0 (exclusion criterium)	0	1/NR	0	2
Ligature
Shah[48], 2010, Nepal	PS		Intracorporeal single ligation	80	19	0	NR	NR	NR
Carvalh[49], 2009, Brazil	PS		Surgical knots (2-0 polyester)4	1000	NR	0	NR	0	NR
Talebpour, 2007, Iran[50]	PS		Absorbable sutures and intracorporeal knots	200	25	0	NR	1 (B)	18
Suo[36], 2013, China	RS		Absorbable thread (VICRYL_ W9215, Ethicon)	1096	296	0	NR /0	0	NR
Golash[37], 2008, Oman	RS		Intracorporeal ligation (3/0 Vicryl)	1000	NR	0	NR	NR	0
Fullum[38], 2005, United States	RS		Two 2–0 PDS Endoloops (Ethicon Endo-Surgery)	105	22	0	NR	NR	NR
Locking clips
Leung[39], 1996, Hong Kong	RS		PDS-clip (Ethicon Endo-surgery)	272	94	3	4 intra-peritoneal collections	NR	45 (30 in complicated group)
Clips
Sinha[40], 2012, India		RS	Ligaclip (titanium clip, Ethicon)4	756	NR	4	NR	NR	NR
Agresta[41], 2011, Italy		RS	Titanium clips4	932	123	1	NR	NR	NR
Feroci[32], 2011, Italy		RS	Polymeric absorbable clip	664	NR	0	0	NR	NR
Ou[42], 2009, China		RS	Clips	10000	NR	6	NR	0	NR
Ojima[35], 2007, Japan		RS	Clips	1127	NR	3	NR	23	NR
Lee[33], 2004, Taiwan		RS	Clips4	1009	78	5	NR/1	2 (D)	NR
Dolan[43], 1999, Nothern Ireland		RS	Titanium clips	303	18	0	1 / 0	NR	0
Wise Unger[30], 1996, United States		RS	Clips	22165	NR	58	NR	NR	NR
Feussner[31], 1991, Germany		RS	Clips	178	28	1	NR	NR	NR
Other
Lewandowski[34], 2006, Poland	RS		LigaSure (Valleylab)	129	NR	0	NR	2 (1B, 1D)	NR

Amsterdam classification was used to identify the severity of the bile duct injuries: B: Major bile duct leaks with or without concomitant biliary strictures; C: Bile duct strictures without bile leakage; D: Complete transection of the duct with or without excision of some portion of the biliary tree.

Only in patients with an inflamed an dilated cystic duct (> 1.0 cm);

Only in patients with an cystic duct < 5 mm; 3severity unknown;

In patients undergoing MLC (mini LC: three ports) or SILC;

Method of closure not described in one study arm, therefore not used as comparative cohort. CDL: Cystic duct leakage; BDI: Bile duct injury; RS: Retrospective study design; PS: Prospective study design; CD: Cystic duct; LC: Laparoscopic cholecystectomy.

CDL

The overall incidence of CDL ranged from 0% to 4% in all 38 studies. The pooled CDL rate was around 0%-1% for each of the techniques (Figures 3-8). Figure 3 shows the forest plot of the comparison of harmonic energy vs metal clips for closing the cystic duct in six studies[13-15,22,23,26]. The overall pooled Peto odds ratio for CDL was 0.4 (95%CI: 0.06-2.48), with low heterogeneity. The use of locking clips vs non-locking clips was compared in three studies (Figure 4); the pooled Peto odds ratio was 0.17 (95%CI: 0.03-0.93)[20,24,25]. Figure 5 depicts the CDL rate for closure of the cystic duct by metal (non-locking) clips in 20 studies (also including the relevant single group from the comparative studies) including 38573 patients; the overall pooled event rate is 0.01% (95%CI: 0-0.01)[13-17,19,21-26,30,31,33,35,40-43]. Harmonic energy was used to close the cystic duct in 14 non-comparative analysis including 1692 patients (Figure 6); the pooled event rate for CDL was 0.01% (95%CI: 0.01-0.02)[13-15,18,22,26-29,44-47]. The overall pooled CDL event rate after application of a ligature to the cystic duct was 0.00% (95%CI: 0-0.01), calculated from eight studies including 3604 patients (Figure 7)[16,36-38,48-50]. Locking clips were used in four studies (including the relevant group from the comparative studies) (in 1853 patients) to close the cystic duct (Figure 8); the overall pooled event rate of CDL was 0.00% (95%CI: 0-0.01)[20,24,25,39].

Figure 3

Forest plot on the comparison of harmonic scalpel and metal clips on cystic duct leakage.

Figure 8

Forest plot of cystic duct leakage following the application of locking clips. CDL: Cystic duct leakage.

Figure 4

Forest plot on the comparison locking and non-locking clips on cystic duct leakage.

Figure 5

Forest plot of cystic duct leakage following the application of metal clips. CDL: Cystic duct leakage.

Figure 6

Forest plot of cystic duct leakage following the application of harmonic scalpel/shears. CDL: Cystic duct leakage.

Figure 7

Forest plot of cystic duct leakage following the application of ligature. CDL: Cystic duct leakage.

CDL in uncomplicated and complicated disease

Sixteen studies of the 38 studies did not specify whether patients had complicated or uncomplicated disease, and seven studies only included patients with uncomplicated disease. Of the 16 studies that did report on complicated and uncomplicated patients, only 2 specified the outcomes per disease entity. Analysis on effectiveness of various techniques of cystic duct closure in patients with uncomplicated vs in those with complicated gallbladder disease could therefore not be performed.

Operating time and hospital stay

Operating time and hospital stay data are listed in Table 2. Five studies compared harmonic energy with metal clips reported on operating time[13-15,23,26]. In four studies operating time was significantly shorter in the harmonic energy group than in the clip group[13-15,23]. In the remaining study the operating time was non-significantly shorter in the harmonic group[26]. The operating time was significantly shorter in the absorbable clip group compared to non-absorbable clips[21]. In one study locking clips were associated with a significantly shorter duration of surgery compared to non-locking clips in one study[25].

Hospital stay was discussed in six studies, two randomized, one prospective and 3 retrospective studies[13,14,18,21,23,25]. In the three studies[13,14,23] reporting on the comparison of harmonic energy and metal clips, two studies showed a significantly shorter hospital stay in the harmonic group[13,14], but in one study the hospital stay was comparable in the harmonic and clip groups[23].

Failure of technique

Only 4 studies described failure of the harmonic scalpel during surgery, mostly resulting in the need to add either clips or ligatures[22,23,29,46]. The failure of the harmonic scalpel was reported in 24 patients from a total of 352 patients (6.7%). Locking clips failed in 52 patients from a total of 1853 (2.8%) during surgery, reported in 4 studies[20,24,25,39]. In 1270 patients 18 ligatures failed (1.4%), which was reported in three studies[19,37,50]. Failure of closure technique usually meant the need to switch to a different closure technique. None of the studies describing failure of technique noted the consequences of failure, such as an increase in complications.

Short term morbidity

Following closure of the cystic duct with harmonic energy 4 of 863 (0.46%) patients developed a biloma or intra-abdominal abscess. After closure with metallic clips four of 3122 patients (0.13%) were diagnosed with a biloma or abscess, compared to 5 of 907 patients (0.55%) who received a locking clip. After closure with a ligature no biloma or abscess was seen in 1096 patients.

The rate of BDI was reported in 20 studies[13-15,17,18,22,23,25,26,28,29,33-36,42,45,49,50]. Overall, only 20 of 17180 patients were diagnosed with BDI (0.12%). Seven BDI were reported in the harmonic group (7 of 1085 patients, 0.65%), all type D. In the patients who had closure of their cystic duct with metal clips 8 BDI were seen (8 of 13421 patients, 0.06%), of which the severity (type D) was reported in only two patients. Two of 328 patients (0.61%) who received a locking clip were diagnosed with a BDI of unknown severity. Two bile duct injuries were found when LigasureTM was used, one type B and one type D (2 of 230 patients, 0.87%). There was only one bile duct injury, type B, in the patients in whom the cystic duct was closed with a ligature (1 of 2296 patients, 0.04%).

Subgroup analysis

It was not possible to identify the subgroup of patients with the highest incidence of CDL, such as patients with complicated gallstones disease, due to lack of subgroup data. Fifteen studies of the 38 studies did not report whether patients with complicated disease were included, and seven studies only included patients with uncomplicated disease. Of the 16 studies that did report complicated patients, only 2 specified the outcomes per subgroup. Therefore, an analysis of CDL associated with the cystic duct closure technique for high-risk patients was not possible.

DISCUSSION

Based on the available evidence as appraised in this systematic review it is not possible to either recommend or discourage any of the techniques for cystic duct closure during LC with respect to CDL, although the data point out a slight preference for locking clips and ligatures vs harmonic energy or (non-locking) metal clips.

No separate recommendation could be made for complicated gallbladder disease as subgroup analysis was not possible due to a lack of reported data per subgroup of complicated and uncomplicated gallbladder disease. The patient populations of the studies included in this systematic review represented clinical practice and comprised a mix of patients with uncomplicated and complicated biliary disease (such as cholecystitis, pancreatitis, cholangitis and common bile duct stones). Several studies however excluded patients with complicated disease. Also, sample sizes are rarely large enough to allow subgroup analysis. When looking at clinical practice different closure techniques are used for different operative scenarios during LC. For example, it is likely that locking clips were used more often in “difficult cholecystectomies”. Therefore, selection bias most likely affected results per technique. The higher bile duct injury rate in locking clips compared to non-locking metal clips may represent preference of locking clips in difficult cholecystectomies, rather than a technique related effect. Vice versa the low bile duct injury rate when ligatures were used may reflect use of this closure technique in less difficult cholecystectomies, or instead be a true favorable effect of this technique.

It is essential to prevent bile duct injuries, regardless of their nature. Type A bile duct injury (Amsterdam Classification[10]), which includes CDL, is previously classified as “minor” injury. Notwithstanding the classification “minor”, type A injuries can be associated with significant morbidity. A recent article shows that mortality related to a type A leakage is 4.2% and sepsis occurs in 15.7% of patients[7]. Although treatment, mostly endoscopically, is usually successful, it would be more prudent to prevent CDL altogether. Whichever closure technique was used, obtaining a critical view of safety (CVS) remains crucial, as can be seen in the occurrence of type D lesions in either technique.

A previous systematic review on the methods of cystic duct closure has been published in 2010, and included only 3 RCTs[51]. The primary outcomes of their Cochrane review include mortality and short term morbidity, but do not focus on CDL. Since then, two more RCT’s, one prospective study and nine retrospective studies have been published, all of which were included in this systematic review. Present review added important information on CDL after various techniques of cystic duct closure and presented an up-to-date overview of all available evidence on the manner of closure of the cystic duct during LC.

This review is hampered by some important limitations. First, the already mentioned lacks of available studies in literature that supply separate data on complicated and uncomplicated gallbladder disease. Second, different study designs were included, such as RCTs, prospective and retrospective studies, to collect enough evidence in large sample sizes. Third, for the purpose of analysis we lumped some techniques into one group; for example, endoloops and intracorporeal knots were both in the “ligature” group. Fourth, reliable comparison of different techniques of cystic duct closure, with event rates around 1%, needs very large sample sizes per studied technique to show a clinically relevant difference. However, 38683 (81.5%) of the 47491 included patients underwent cystic duct closure with non-locking metal clips, leaving the other techniques of interest possibly underpowered. Also, the Peto odds ratio methods are commonly used in rare events and were used in this review because it was the best performing method for pooling our results. Unfortunately, this method does not allow zero events in a group when using it for a comparative analysis, so some included studies in this review could not be used in the pooled comparative analysis. To review the effect of using Peto odds, the same analysis was performed but with 1 event for every group that originally had no events. This resulted in a slightly higher overall pooled event rate, but the effect was very minimal. So, excluding the studies with zero events from the meta-analysis did not result in a big change in the overall pooled event rate. Finally, most studies did not specify cost per particular technique nor cost-effectiveness. Some techniques were clearly more expensive than others, such as the added costs of clip appliers and disposable scalpels/shears. With the ever increasing expenses of our health care system, the choice for a certain method depends in part on the extra costs incurred vs costs saved.

Our group has advocated to report studies on surgery in gallbladder disease in two different entities, uncomplicated and complicated disease[3]. In patients with complicated disease, LC is anticipated to be more difficult than in uncomplicated patients. Consequently, higher risk of complications during the procedure and postoperatively can be expected. We recently published that the risk of bile leakage, including CDL, in patients with complicated gallstone disease is underestimated in literature[3]. Generally, the rate of CDL is reported between 0.5%-3%, but complicated disease is associated with a CDL rate of 4%-7%[1,2]. Based on this risk difference, patients with a high probability of CDL (difficult cholecystectomy) should be included with sufficient high numbers when investigating which technique is superior in cystic duct closure.

It is remarkable that for such a frequently conducted surgical procedure as LC good quality evidence from high sample size trials is lacking to determine which cystic duct closure technique is superior with respect to prevention of CDL. Studies in large populations of patients undergoing LC with or without a high risk of CDL are needed.

ARTICLE HIGHLIGHTS

Research background

Cystic duct leakage (CDL) is reported in laparoscopic cholecystectomy (LC) in 0.5%-3% of patients, and is even reported to increase to up to 4%-7% in patients with complicated gallstone disease.

Research motivation

Although CDL is classified as a minor injury of the bile ducts, it is associated with significant morbidity and even mortality, so adequate closure of the cystic duct is essential to prevent CDL.

Research objectives

Several techniques are used during cholecystectomy to close the cystic duct, but it is currently unknown which technique has the lowest rate of CDL. The aim of this systematic review was to assess the risk of CDL and the CDL rate for different techniques of cystic duct closure after LC, both in uncomplicated and complicated gallbladder disease.

Research methods

A systematic review and meta-analysis was performed according to PRISMA guidelines. A search of MEDLINE, Cochrane and EMBASE was done. Studies were eligible for conclusion when patients underwent cholecystectomy and methods of closure of the cystic duct were described. The primary outcome was leakage of the cystic duct. The risk of bias was evaluated with the MINORS score for non-randomized studies and the Cochrane Library guide for the randomized studies. Odds ratios were analyzed for comparison of techniques and pooled event rates for non-comparative analyses. Pooled event rates were compared for each of included techniques.

Research results

A total of 1491 articles were found by searching the databases. Out of 1491 articles 102 full texts were screened and 38 articles included. A total of 47491 patients were included, of which 38683 (81.5%) underwent cystic duct closure with non-locking (metal) clips. All studies were of low-moderate methodological quality. Only two studies reported separate data on uncomplicated and complicated gallbladder disease. For overall CDL, an odds ratio of 0.4 (95%CI: 0.06-2.48) was found for harmonic energy vs clip closure and an odds ratio of 0.17 (95%CI: 0.03-0.93) for locking vs non-locking clips. Pooled CDL rate was around 1% for harmonic energy and metal clips, and 0% for locking clips and ligatures.

Research conclusions

Based on the available evidence as appraised in this systematic review it is not possible to either recommend or discourage any of the techniques for cystic duct closure during LC with respect to CDL. The data do point out a slight preference for locking clips and ligatures vs harmonic energy or (non-locking) metal clips. This is the first systematic review on methods of cystic duct closure that focuses on CDL. As CDL is an important and potentially serious complication of cholecystectomy, this subject should warrant further research.

Research perspectives

It is interesting to see that no separate recommendation could be made for complicated gallbladder disease as subgroup analysis was not possible due to a lack of reported data per subgroup of complicated and uncomplicated gallbladder disease. It could be hypothesized that cystic duct closure is especially important in these patients and that data on this subject would be readily available. Future research should therefore focus on good quality evidence from high sample size trials that include patients with both uncomplicated and complicated gallstone disease.