Surgical Innovation http://sri.sagepub.com/
Single-Incision Versus Standard Multiple-Incision Laparoscopic Cholecystectomy: A Meta-analysis of Experimental and Observational Studies Nicolò Tamini, Matteo Rota, Elisa Bolzonaro, Luca Nespoli, Angelo Nespoli, Maria Grazia Valsecchi and Luca Gianotti SURG INNOV 2014 21: 528 originally published online 6 March 2014 DOI: 10.1177/1553350614521017 The online version of this article can be found at: http://sri.sagepub.com/content/21/5/528
Published by: http://www.sagepublications.com
On behalf of:
Institute for Research into Cancer of the Digestive System
Additional services and information for Surgical Innovation can be found at: Email Alerts: http://sri.sagepub.com/cgi/alerts Subscriptions: http://sri.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Sep 11, 2014 OnlineFirst Version of Record - Mar 6, 2014 What is This?
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
521017
research-article2014
SRIXXX10.1177/1553350614521017Surgical InnovationTamini et al
In Context: Review
Single-Incision Versus Standard Multiple-Incision Laparoscopic Cholecystectomy: A Meta-analysis of Experimental and Observational Studies
Surgical Innovation 2014, Vol. 21(5) 528–545 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1553350614521017 sri.sagepub.com
Nicolò Tamini, MD1, Matteo Rota, PhD1, Elisa Bolzonaro, MD1, Luca Nespoli, MD1, Angelo Nespoli, MD1, Maria Grazia Valsecchi, PhD1, and Luca Gianotti, MD, ScD1
Abstract Objective. The advantages of single-incision surgery for the treatment of gallstone disease is debated. Previous meta-analyses comparing single-incision laparoscopic cholecystectomy (SILC) and standard laparoscopic multiport cholecystectomy (SLMC) included few and underpowered trials. To overcome this limitation, we performed a metaanalysis of randomized and nonrandomized studies. Methods. A MEDLINE, EMBASE, and Cochrane Library literature search of studies published in and comparing SILC with SLMC was performed. The primary outcome was safety of SILC as measured by the overall rate of postoperative complications and biliary spillage. Feasibility was another primary outcome as measured by the conversion and operative time. Postoperative pain, length of hospital stay, perioperative blood loss, time to return to normal activity, and cosmetic satisfaction were secondary outcomes. Results. We identified 43 studies of which 30 were observational reports and 13 experimental trials, for a total of 7489 patients (2090 SILC and 5389 SLMC). The overall rate of complications was comparable between groups (relative risk [RR] = 1.08; 95% CI = 0.87-1.35; P = .46), as were the rates of biliary spillage (RR = 1.16; 95% CI = 0.73-1.84; P = .53) and conversion rate (RR = 0.88; 95% CI = 0.53-1.46; P = .62). Operative time was in favor of SLMC (weighted mean difference = 0.73; 95% CI = 0.67-0.79; P < .0001). Secondary outcomes favored SILC, but with marginal advantages. Conclusions. SILC is a feasible technique but without any significant advantage over SLMC for relevant end points. Although secondary outcomes favored SILC, the small magnitude of the advantage and the low quality of assessment methods question the clinical significance of these benefits. Keywords single incision, cholecystectomy, laparoscopy, meta-analysis, outcome
Introduction Multiport laparoscopic cholecystectomy is nowadays the gold-standard surgical treatment for gallstonerelated disease, with reduced postoperative pain, shortened hospital stay, faster recovery after surgery, and earlier return to normal function when compared with open surgery.1,2 As laparoscopic techniques have advanced, cholecystectomy has been performed with smaller incisions and/ or fewer ports or using a needlescopic technique, referred to as minilaparoscopic surgery, but without showing objective advantages.3 More recently, laparoscopic cholecystectomy has been shown to be feasible also using a single incision at the umbilicus (single-incision laparoscopic cholecystectomy [SILC]).4 The rationale for using
smaller instruments and/or fewer incisions is to minimize tissue trauma and improve postoperative pain and cosmesis for patients,5 but these new techniques may have potential disadvantages, including a long learning curve, prolonged operative time, and decreased visualization and related increased risk of bile duct injuries.5 Therefore, in the evaluation of technical innovations in surgery, there are several main factors that should be considered before a new procedure achieves widespread acceptance 1
Milano-Bicocca University, San Gerardo Hospital, Monza, Italy
Corresponding Author: Luca Gianotti, Department of Surgery, Milano-Bicocca University, San Gerardo Hospital (4° piano B), Via Pergolesi 33, Monza 20052, Italy. Email: [email protected]
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
529
Tamini et al and implementation. These comprise safety, feasibility, efficacy, comparable or improved clinical results, and a benefit to the patient or health care system in terms of cost-effectiveness. Several randomized controlled trials (RCTs) have been conducted to compare standard laparoscopic multiport cholecystectomy (SLMC) with SILC, yet sample sizes of those RCTs were not large enough for a comprehensive evaluation of this new surgical intervention. Recently, there have been several meta-analyses that synthesized data obtained from the published RCTs to evaluate the feasibility and safety of SILS for cholecystectomy in comparison to that for the traditional multiport technique.6-8 The majority of these previous meta-analyses included a limited number of RCTs. The most recent and comprehensive meta-analysis by Pisanu et al9 pooled results of 12 RCTs with a total of 892 patients. In addition, previous reports often did not assess and evaluate the quality of the available evidence, which is very important for clinical decision making. High-quality RCTs needed to evaluate a new surgical procedure are complex to carry out because of several hurdles such as learning curve effects, ethical and cultural resistance, absence of blinding, and urgent or unexpected conditions during operation in surgical treatment. For this reason, evaluation of non-RCTs may be an appropriate strategy to extend the source of evidence.10 Therefore, we designed a meta-analysis that included available data from experimental and observational studies to evaluate the safety and efficacy of SILC and to determine whether SILC is an acceptable alternative to SLMC for the treatment of gallstone disease.
Methods Search and Identification of Studies We performed a MEDLINE, EMBASE, and Cochrane Library extended literature search of all observational and experimental studies published as original articles written in English from January 1997 to February 2012, comparing SILC and SLMC. The Boolean string [“cholecystectomy” AND “SILS” OR “SILC “ OR “Single site” OR “Single port” OR “LESS” OR “Single incision” OR “Laparo endoscopic” OR “One trocar” OR “Single trocar”, limited to title and abstract] was used to perform the search, following the meta-analysis of observational studies in epidemiology guidelines11 and the PRISMA statement for reporting systematic reviews and metaanalyses of studies evaluating health care interventions.12 Three of the authors (NT, MR, and EB) also reviewed reference lists from all the relevant studies, systematic reviews, and previous meta-analyses to identify additional publications of interests.
Eligibility Criteria Study Type. All experimental (RCTs) and observational (non-RCTs) full-text studies published in the English language and comparing SILC with SLMC were evaluated. Abstracts, opinion pieces, and editorials were excluded because they could not provide detailed information about the methodology, implementation, and outcome measures to be assessed. Among the observational studies, only those reporting a control group were included (retrospective case-control and cohort reports). We included cholecystectomy performed for both simple gallbladder stone disease and for acute cholecystitis in adult and pediatric populations. All studies reported at least 1 or more of the following defined clinical outcomes: overall postoperative complications, biliary spillage, conversion to open procedure, operation time, intraoperative blood loss, postoperative pain evaluation, duration of postoperative hospital stay, cosmesis satisfaction, and time to return to work. If 2 or more reports were from the same institution or team, we selected the most recent or the most complete one. Intervention Type. For the SILC group, we considered any cholecystectomy performed with any commercially available devices (by different companies) and self-modified devices. For the SLMC group, we included all the patients operated with at least 3 trocars.
Data Collection Relevant data were extracted independently and in duplicate from all studies by 2 authors (NT, EB) and included study features, population characteristics, data needed for quality assessment, and the different outcome measures. Population characteristics included the number and type of procedures performed, age, gender, and body mass index (BMI). Checklists and electronic databases were used for data extraction.
Outcome Measures The primary outcome of this meta-analysis was safety of SILC as measured by the overall rate of postoperative complications that developed within 30 days of the surgical procedure and the rate of biliary spillage defined as damage of the biliary tract or bile lick from the liver. Feasibility was another primary outcome as measured by the conversion rate to open surgery and operative time. Postoperative pain, as measured by a 10-point Visual Analogue Scale 24 hours after the procedure, length of hospital stay (LOS), perioperative blood loss, time to return to normal activity, cosmetic satisfaction, and the rate of incisional hernia were secondary outcomes. The
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
530
Surgical Innovation 21(5)
latter outcome was originally measured by means of different scales (5 or 10 items) within the studies included in the meta-analysis. For this reason, we converted the Likert-based cosmetic outcome to a percentage of cosmetic satisfaction.
Egger’s test for funnel plot asymmetry.20 Analyses were conducted using STATA 11.0 (STATA Corp, College Station, TX).
Quality Assessment
Study Selection
Two independent authors (LG and EB) assessed the quality of the observational studies by using the previously validated Newcastle-Ottawa Scale,13 whereas we used the Jadad Scale14 to score the quality of experimental studies. Because in this type of studies double blinding is impossible to perform, in the Jadad Scale we gave a score of 1 to the trials reporting blinding of the patient and independent outcome assessor. Experimental studies with Jadad scores ≥3 and observational studies with NewcastleOttawa Scale score ≥8 were considered to be of high quality.
The process of article selection is described in Figure 1. We identified 220 relevant studies from the electronic literature search, and after a detailed evaluation, 176 of them were excluded because they did not satisfy the inclusion criteria—that is, case-series studies reporting only results of SILC without a comparison with SLMC or studies investigating procedures performed through natural orifices. Moreover, 2 studies21,22 were not considered because we included the updated articles on the same study.23,24
Results
Study and Population Characteristics Statistical Analysis For continuous outcomes, statistical analyses were performed by using the weighted mean difference (WMD) as the summary statistic. The WMD is a dimensionless effect measure that represents the number of standard deviations by which the intervention changes the outcome. For studies that presented continuous data reported as median and range values, we derived the mean and standard deviation by applying the formula developed by Hozo et al15 Using this formula, we could include within our meta-analysis the studies that did not report the mean and standard deviation of the considered outcome. Pooled odds ratios (ORs) were calculated to summarize the effect of binary outcomes. In this setting, we applied the Yates 0.5 correction factor16 for studies containing zero events in either SILC or SLMC groups in order to guarantee the estimability of the OR and, thus, to include the study within the meta-analysis. All pooled outcome measures were determined using random-effects models.17 We assessed the between-studies heterogeneity using the I2 statistic, which is the proportion of total variation contributed by the between-studies variance.18 In the absence of betweenstudies heterogeneity (P > .05), the random-effects pooled estimates are identical to those obtained from fixed-effects models. We generated forest plots for all the investigated outcomes. We also conducted stratified analyses according to geographic area of the study, study quality index, age and BMI of study participants, and surgeon experience in SILC. For this, we used the median value of age, BMI, and study quality index to perform the stratified analyses. We performed a publication bias analysis through the counter-enhanced funnel plot19 and
The present meta-analysis included 43 studies comparing SILS and SLMC, of which 30 were observational repo rts23-52 with 5367 cases and 13 were experimental trials53-65 with 2131 cases, yielding a total of 7489 patients (Table 1).
Primary Outcome Measures The descriptive characteristics of the primary outcome measures are shown in Table 2. A total of 41 studies reported postoperative morbidity. The analysis was conducted in 7399 patients (2015 SILC and 5348 SLMC). The overall rate of complications was comparable between SLMC and SILC (relative risk [RR] = 1.08; 95% CI = 0.87-1.35; P = .46). No significant differences were also observed by analyzing experimental (RR = 1.09; 95% CI = 0.81-1.47; P = .57) and observational studies separately (RR = 1.08; 95% CI = 0.79-1.47; P = .63; Figure 2). The Egger P value for publication bias analysis was .10 (Figure 3). The analysis of postoperative morbidity was stratified according to different potential risk factors. None of the considered stratification variables were significantly related to the occurrence of complications (Table 3). We could retrieve information on bile spillage in 29 studies for a total of 6648 patients (1719 SILC and 4929 SLMC). The rate was comparable between SLMC and SILC (RR = 1.16; 95% CI = 0.73-1.84; P = .53; Figure 4), with no difference in observational (RR = 1.16; 95% CI = 0.65-2.05; P = .62) and experimental trials (RR = 1.17; 95% CI = 0.54-2.53; P = .40). No evidences of publication bias were found for biliary spillage (Egger test P = .21).
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
531
Tamini et al
Figure 1. Process of study selection according to the PRISMA statement.
Abbreviations: RCT, randomized controlled trial; NOTES, natural orifice transluminal endoscopic surgery.
Figure 5 represents the forest plot of the RR for conversion rate to open surgery. This outcome was reported in 10 experimental trials and 27 observational studies, for a total of 3463 patients. The conversion rate was similar in the 2 groups (RR = 0.88; 95% CI = 0.53-1.46; P = .62). Similar results were obtained in a separate analysis of RCTs (RR = 1.23; 95% CI = 0.49-3.08; P = .66) and nonRCTs (RR = 0.75; 95% CI = 0.40-1.40; P = .37). No evidences of publication bias were found for conversion to laparotomy (Egger test P = .87). The needle gallbladder suspension technique was reported in 26/43 (60.4%) of the trials (53.8% in RCTs and 63.3% in non-RCTs). Additional trocars during SILC were needed in 103 patients out of 1693 (6.1%) within 27 observational studies (3 studies did not report data on additional trocar use) and in 28 patients out of 533 (5.2%) in 13 randomized studies. There was no significant difference between observational and experimental studies (P = .49) in the use of additional ports during SILC. The time needed to complete the surgical procedure was reported in 42 studies (29 observational and 13
experimental), with a total of 7427 patients. This outcome was in favor of the SLMC group (WMD = 0.73; 95% CI = 0.67-0.79; P < .0001). Separate analysis for experimental (WMD = 0.74; 95% CI = 0.61-0.88; P < .0001) and observational studies (WMD = 0.72; 95% CI = 0.65-0.79; P < .0001) showed similar results (P for heterogeneity = .81; Figure 6). The P value for publication bias analysis by the funnel plot was .11 (Figure 7). A stratified analysis for potential factors affecting operative time showed a significant heterogeneity among different strata. Low study quality, age >47 years, BMI >27 kg/m2, ASA score >2, and geographic area of study origin seemed to be associated with longer operative time in SILC with respect to SLMC. Previous surgeon experience with SILC was not significantly related to the time needed to complete the operation (Table 4).
Secondary Outcome Measures The descriptive characteristics of the primary outcome measures are shown in Table 5. All secondary outcomes
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
532
Surgical Innovation 21(5)
Table 1. Main Characteristics and Demographic Data of the Studies Included in the Meta-analysis.
Study: First Author (Year) Observational studies Hernandez (2009)23 Kravetz (2009)25 Philipp (2009)26 Chow (2010)27 Fronza (2010)28 Ito (2010)29 McGregor (2011)30 Navarra (2010)31 Thapa (2010)32 Bagloo (2011)33 Cao (2011)34 Chandler (2011)35 Chang (2011)36 Emami (2011)37 Froghi (2012)38 Gangl (2011)39 Garcia-Henriquez (2011)40 Garg (2012)41 Jacob (2011)42 Joseph (2011)43 Khambaty (2011)44 Kilian (2011)45 Prasad (2011)46 Rupp (2011)47 Sinha (2011)48 Vidal (2011)24 Wren (2011)49 Wu (2011)50 Barband (2012)51 Wong (2012)52 Subtotal Experimental studies Aprea (2011)53 Lee (2010)54 Rasic´ (2010)55 Tsimoyiannis (2010)56 Asakuma (2011)57 Bucher (2011)58 Cao (2011)59 Lai (2011)60 Lirici (2011)61 Ma (2011)62 Phillips (2012)63 Sinan (2012)64 Zheng (2012)65 Subtotal Total
SILC Group
Country
Study Quality Indexa
Spain USA USA UK USA Japan UK Italy Nepal USA China USA Singapore USA UK Austria USA India Germany USA USA Germany India USA India Spain Austria China Iran China
6 7 8 6 8 7 9 7 7 7 7 6 6 7 9 7 6 6 9 8 8 8 6 8 7 7 8 5 6 7
100 20 29 41 25 5 11 20 20 40 36 42 30 40 16 67 27 25 36 108 107 16 100 101 240 120 10 100 25 20 1577
46 ± 16 44 38 (18-85) 43 ± 16 44 ± 15 56 ± 12 45 ± 4 45 (28-65) 45 ± 4 45 (24-80) 47 15 (6-19) 48 (25-69) 15 (8-20) 47 ± 17 53 (15-81) 14 (3-18) 44 ± 14 62 (21-81) 48 ± 17 45 38 (20-69) 38 44 ± 18 38 (20-82) 56 (29-70) 59 ± 16 46 ± 13 41 (21-61) 59 (39-66) -
Italy Taiwan Croatia Greece Japan Switzerland China China Italy USA USA Turkey China
2 4 1 3 2 2 3 2 3 2 2 2 2
25 35 48 20 24 75 57 24 20 21 117 17 30 513 2090
46 ± 9 51 ± 14 44 ± 6 49 ± 17 57 (47-62) 42 (18-81) 62 ± 5 52 ± 13 45 (26-63) 57 ± 16 46 (18-77) 49 ± 9 44 ± 11 — —
n
Age, years
SLMC group n
Age, years
M/F
BMI (kg/m2)
25/75 29 ± 6.5 4/16 30.2 4/25 30 (19-43) 12/29 26.6 ± 5 2/23 25.4 ± 5 2/3 23.4 ± 3 2/9 27 ± 1.4 8/12 27 (23-36) 5/15 NA 13/27 27.8 (21-40) 19/17 NA 6/36 NA 12/18 NA 5/35 28.2 ± 7.8 5/11 28.3 ± 6.3 16/51 26 (18-43) 9/18 21.7 (11-33) 5/20 25.1 ± 4.1 19/17 27.8 (19-40) 22/86 28.3 ± 5.8 NA 29.5 7/9 27 (20-33) 44/56 27.7 17/84 29.7 ± 7.4 24/216 NA 43/77 NA 7/3 27.7 ± 3.3 25/75 NA 4/21 29 (22-34.7) 4/16 22.9 ± 2.8 -
30 20 22 58 25 23 24 74 20 83 24 27 30 68 13 67 27 27 37 177 44 20 100 100 3492 120 10 100 30 20 4912
48 ± 17 46 39 (20-75) 51 ± 17 40 ± 12 51 ± 17 56 ± 3 NA 41 ± 3 57 (19-93) 48 15 (3-19) 48 (24-64) 15 (3-22) 48 ± 10 56 (21-87) 15 (6-23) 44 ± 13 60 (21-94) 50 ± 16 48 56 (26-82) 38 41 ± 20 NA 55 (32-76) 62 ± 16 50 ± 14 43 (23-60) 56 (30-65) -
9/21 6/14 4/18 16/42 4/21 10/13 7/17 25/49 1/19 25/58 14/10 5/22 12/18 12/56 3/10 16/51 10/17 4/23 15/22 56/121 NA 10/10 42/58 18/82 NA 38/82 7/3 47/53 5/25 5/15 -
29 ± 7 28.2 29 (17-52) 28.2 ± 5.8 30.7 ± 7 24 ± 5.4 31.7 ± 1.1 NA NA 28.9 (15.3-67.6) NA NA NA 28.1 ± 7.8 26.3 ± 5.8 27 (20-46) 20 (15-43) 24.4 ± 3.8 27.4 (21-52) 30.3 ± 7.5 30.7 25 (18-31) 27.3 32.5 ± 8.6 NA NA 28.4 ± 6.2 NA 29.9 (23.2-34.8) 23.6 ± 2.3 -
9/16 13/22 22/26 5/15 11/13 NA 23/34 8/16 6/14 2/19 28/89 4/13 13/17 — —
25 35 50 20 25 75 51 27 20 22 80 17 30 477 5389
44 ± 10 53 ± 16 44 ± 6 48 ± 10 66 (55-72) 44 (20-78) 60 ± 4 54 ± 12 50 (24-67) 46 ± 12 44 (19-68) 49 ± 14 47 ± 14 — —
6/19 15/20 18/32 1/19 13/12 NA 22/29 11/16 6/14 2/20 24/56 8/9 16/14 — —
23.7 ± 4.6 25.8 ± 3 27 ± 4 25.3 ± 5.3 24.1 (20.4-25.9) 25 (19-34) 29.1 ± 5.1 24.4 ± 2.8 27 (18-30) 30.7 ± 6.1 31 (18.8-45) 27.2 ± 2.9 25.9 ± 4.1 — —
M/F
BMI (kg/m2)
25.9 ± 5.8 24.2 ± 3.4 27 ± 4 23.7 ± 3.2 24 (21.4-27.1) 26 (22-35) 28.6 ± 4.4 25 ± 3 25 (18-29) 28.2 ± 5.3 28.9 (15.4-44.8) 27.3 ± 3.1 24.7 ± 3.4 — —
Abbreviations: SILC, single-incision laparoscopic cholecystectomy; SLMC, standard laparoscopic multiport cholecystectomy; M, male; F, female; BMI, body mass index; NA, not available. a Study quality was evaluated using the Newcastle-Ottawa Scale for observational studies (0 = lowest quality, 10 = highest quality) and the Jadad Scale (0 = lowest quality, 5 = highest quality) for experimental studies.
were in favor of the SILC group. Results are summarized in Table 6. In particular, postoperative pain at 24 hours showed a WMD of −0.27 (95% CI = −0.37 to −0.17; P < .0001). This significant difference was confirmed in both experimental (WMD = −0.25; 95% CI = −0.40 to −0.11; P < .0001) and observational studies (WMD = −0.27; 95% CI = −0.40 to −0.15; P < .0001).
LOS was significantly shorter in the SILC group (WMD = −0.29; 95% CI = −0.37 to −0.21; P < .0001). Experimental (WMD = −0.43; 95% CI = −0.58 to −0.28; P < .0001) and observational studies (WMD = −0.24; 95% CI = −0.33 to −0.14; P < .0001) had a similar trend. We also analyzed and split LOS by geographic region. Studies from the United States had the shortest stay (0.97
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
533
Tamini et al Table 2. Primary Outcomes Investigated Within the Studies Included in the Meta-analysis.a Safety Outcome
Study: First Author, Year Observational studies Hernandez (2009)23 Kravetz (2009)25 Philipp (2009)26 Chow (2010)27 Fronza (2010)28 Ito (2010)29 McGregor (2011)30 Navarra (2010)31 Thapa (2010)32 Bagloo (2011)33 Cao (2011)34 Chandler (2011)35 Chang (2011)36 Emami (2011)37 Froghi (2012)38 Gangl (2011)39 Garcia-Henriquez (2011)40 Garg (2012)41 Jacob (2011)42 Joseph (2011)43 Khambaty (2011)44 Kilian (2011)45 Prasad (2011)46 Rupp (2011)47 Sinha (2011)48 Vidal (2011)24 Wren (2011)49 Wu (2011)50 Barband (2012)51 Wong (2012)52 Subtotal Experimental studies Aprea (2011)53 Lee (2010)54 Rasic´ (2010)55 Tsimoyiannis (2010)56 Asakuma (2011)57 Bucher (2011)58 Cao (2011)59 Lai (2011)60 Lirici (2011)61 Ma (2011)62 Phillips (2012)63 Sinan (2012)64 Zheng (2012)65 Subtotal Total
Postoperative complications (n/N), SILC; SLMC
Bile spillage (n/N), SILC; SLMC
Feasibility Outcome Conversion to Open Laparotomy (n/N), SILC; SLMC
Operating Time (minutes), SILC; SLMC
5/100; 0/30 0/20; 1/20 7/29; 3/22 1/41; 3/58 3/25; 2/25 0/5; 0/23 7/11; 15/24 0/20; 0/74 2/20; 2/20 1/40; 0/83 0/36; 0/24 5/42; 1/27 NA 0/40; 0/68 1/16; 1/13 1/67; 1/67 2/27; 1/27 NA 5/36; 6/37 4/108; 8/177 3/107; 1/44 0/16; 1/20 0/100; 0/100 3/101; 5/100 2/240; 38/3492 8/120; 7/120 0/10; 1/10 2/100; 0/100 0/25; 0/30 3/20; 4/20 65/1522; 101/4855b
1/100; 0/30 0/20; 0/20 0/29; 1/22 1/41; 1/58 0/25; 1/25 0/5; 0/23 0/11; 0/24 NA NA NA NA 0/42; 0/27 NA 0/40; 0/68 0/16; 0/13 0/67; 0/67 0/27; 0/27 NA 1/36; 3/37 5/108; 2/177 3/107; 1/44 NA 0/100; 0/100 0/101; 0/100 1/240; 21/3492 3/120; 2/120 NA 1/100; 0/100 0/25; 0/30 NA 16/1360; 32/4604b
1/100; 0/30 0/20; 0/20 0/29; 0/22 0/41; 4/58 0/25; 0/25 0/5; 0/23 0/11; 0/24 0/20; 1/74 0/20; 2/20 0/40; 0/83 0/36; 0/24 0/42; 0/27 NA 0/40; 0/68 0/16; 0/13 1/67; 0/67 0/27; 0/27 NA 1/36; 3/37 0/108; 2/177 0/107; 0/44 0/16; 0/20 0/100; 0/100 0/101; 0/100 NA 0/120; 0/120 1/10; 0/10 0/100; 0/100 0/25; 0/30 0/20; 0/20 4/1282; 12/1363b
72 (20.7); 71 (16.3) 77 (33); 65.7 (33) 90 (25); 78 (32.4) 126 (42.9); 95.8 (33.1) 100.8 (22); 64.5 (18) 154 (57); 100 (51) 86.9 (9); 79.1 (4.2) 53 (14.5); 40 (10.2) 75 (5.7); 39 (2.2) 102 (26.8); 106 (71.8) 56 (16.7); 36.7 (6) 69.4 (15.7); 64.5 (12.1) 86 (56.1); 58 (27.4) 82.4 (27.6); 63.1 (28.7) 113.3 (33.7); 89.3 (19) 89.3 (38.1); 78.3 (41.6) 98.8 (18.7); 62 (14.4) NA 88 (41.5); 94 (51.3) 67.5 (25.8); 59.8 (36.1) 81.5 (28); 69.1 (21) 65 (17.3); 70 (30.1) 66.8 (5.8); 28.1 (1.4) 76 (29); 75 (33) 26.7 (16.7); 18.4 (4.6) 45 (22); 40 (12.9) 105 (18.3); 106.1 (23.5) 53.5 (24); 49.2 (13.8) 56.3 (13.8); 41.7 (9) 78.7 (21.1); 74 (24.3) 80.8 (25.6); 66.1 (23.1)
0/25; 0/25 7/35; 5/35 0/48; 0/50 1/20; 2/20 0/24; 0/25 12/75; 10/75 2/57; 1/51 0/24; 1/27 1/20; 3/20 3/21; 4/22 45/117; 27/80 1/17; 1/17 2/30; 1/30 74/513; 55/477b 139/2035; 156/5332b
NA 0/35; 0/35 NA 1/20; 2/20 NA 9/75; 6/75 NA 0/24; 0/27 0/20; 0/20 NA 0/117; 0/80 0/17; 0/17 0/30; 0/30 10/338; 8/304b 26/1698; 40/4908b
NA 2/35; 1/35 0/48; 0/50 NA 1/24; 2/25 NA 1/57; 0/51 0/24; 0/27 0/20; 1/20 0/21; 0/22 0/117; 0/80 NA 2/30; 0/30 6/376; 4/340b 10/1658; 16/1703b
41.3 (12); 35.6 (5.8) 71.7 (11.6); 48.4 (10.5) 46 (3.5); 43 (4) 49.7 (9); 37.3 (9.2) 109.3 (2.8); 99 (3) 68.3 (22.7); 70.8 (23.1) 55.2 (12.4); 46.3 (10.8) 43.5 (15.4); 46.5 (20.1) 76.8 (3.4); 48.3 (6.7) 88.5 (40.4); 44.8 (40.4) 57.2 (20.8); 45.2 (20.8) 124.4 (29.7); 64.1 (26.1) 55.6 (25.7); 42.7 (18.6) 68.3 (25.8); 51.7 (17.2) 76.9 (26.0); 61.6 (22.2)
Abbreviations: SILC, single-incision laparoscopic cholecystectomy; SLMC, standard laparoscopic multiport cholecystectomy; NA, not available. a Categorical variables are expressed as number of events and continuous variables as mean (standard deviation). b Sum does not add up to the total number of individuals considered within SILC or SLMC groups as a result of missing information on the considered outcome.
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
534
Surgical Innovation 21(5)
Figure 2. Forest plot for postoperative complications between SILC and SLMC. RRs and 95% CIs are shown.
Abbreviations: SILC, single-incision laparoscopic cholecystectomy; SLMC, standard laparoscopic multiport cholecystectomy; RR, relative risk; CI, confidence interval.
Figure 3. Funnel plot for postoperative complications. Abbreviations: SE, standard error; RR, relative risk.
days in the SILC group and 1.14 in the SLMC group) with respect to Europe (2.31 days in the SILC group and 2.86 in the SLMC group) and Asia (2.31 days in the SILC group and 2.54 in the SLMC group) but without significant difference (P = .17). SILC was associated with a significant reduction in intraoperative blood loss (WMD = −0.16; 95% CI = −0.27 to −0.04; P = .007). Conflicting results (P for heterogeneity
Single-Incision Versus Standard Multiple-Incision Laparoscopic Cholecystectomy: A Meta-analysis of Experimental and Observational Studies Nicolò Tamini, Matteo Rota, Elisa Bolzonaro, Luca Nespoli, Angelo Nespoli, Maria Grazia Valsecchi and Luca Gianotti SURG INNOV 2014 21: 528 originally published online 6 March 2014 DOI: 10.1177/1553350614521017 The online version of this article can be found at: http://sri.sagepub.com/content/21/5/528
Published by: http://www.sagepublications.com
On behalf of:
Institute for Research into Cancer of the Digestive System
Additional services and information for Surgical Innovation can be found at: Email Alerts: http://sri.sagepub.com/cgi/alerts Subscriptions: http://sri.sagepub.com/subscriptions Reprints: http://www.sagepub.com/journalsReprints.nav Permissions: http://www.sagepub.com/journalsPermissions.nav
>> Version of Record - Sep 11, 2014 OnlineFirst Version of Record - Mar 6, 2014 What is This?
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
521017
research-article2014
SRIXXX10.1177/1553350614521017Surgical InnovationTamini et al
In Context: Review
Single-Incision Versus Standard Multiple-Incision Laparoscopic Cholecystectomy: A Meta-analysis of Experimental and Observational Studies
Surgical Innovation 2014, Vol. 21(5) 528–545 © The Author(s) 2014 Reprints and permissions: sagepub.com/journalsPermissions.nav DOI: 10.1177/1553350614521017 sri.sagepub.com
Nicolò Tamini, MD1, Matteo Rota, PhD1, Elisa Bolzonaro, MD1, Luca Nespoli, MD1, Angelo Nespoli, MD1, Maria Grazia Valsecchi, PhD1, and Luca Gianotti, MD, ScD1
Abstract Objective. The advantages of single-incision surgery for the treatment of gallstone disease is debated. Previous meta-analyses comparing single-incision laparoscopic cholecystectomy (SILC) and standard laparoscopic multiport cholecystectomy (SLMC) included few and underpowered trials. To overcome this limitation, we performed a metaanalysis of randomized and nonrandomized studies. Methods. A MEDLINE, EMBASE, and Cochrane Library literature search of studies published in and comparing SILC with SLMC was performed. The primary outcome was safety of SILC as measured by the overall rate of postoperative complications and biliary spillage. Feasibility was another primary outcome as measured by the conversion and operative time. Postoperative pain, length of hospital stay, perioperative blood loss, time to return to normal activity, and cosmetic satisfaction were secondary outcomes. Results. We identified 43 studies of which 30 were observational reports and 13 experimental trials, for a total of 7489 patients (2090 SILC and 5389 SLMC). The overall rate of complications was comparable between groups (relative risk [RR] = 1.08; 95% CI = 0.87-1.35; P = .46), as were the rates of biliary spillage (RR = 1.16; 95% CI = 0.73-1.84; P = .53) and conversion rate (RR = 0.88; 95% CI = 0.53-1.46; P = .62). Operative time was in favor of SLMC (weighted mean difference = 0.73; 95% CI = 0.67-0.79; P < .0001). Secondary outcomes favored SILC, but with marginal advantages. Conclusions. SILC is a feasible technique but without any significant advantage over SLMC for relevant end points. Although secondary outcomes favored SILC, the small magnitude of the advantage and the low quality of assessment methods question the clinical significance of these benefits. Keywords single incision, cholecystectomy, laparoscopy, meta-analysis, outcome
Introduction Multiport laparoscopic cholecystectomy is nowadays the gold-standard surgical treatment for gallstonerelated disease, with reduced postoperative pain, shortened hospital stay, faster recovery after surgery, and earlier return to normal function when compared with open surgery.1,2 As laparoscopic techniques have advanced, cholecystectomy has been performed with smaller incisions and/ or fewer ports or using a needlescopic technique, referred to as minilaparoscopic surgery, but without showing objective advantages.3 More recently, laparoscopic cholecystectomy has been shown to be feasible also using a single incision at the umbilicus (single-incision laparoscopic cholecystectomy [SILC]).4 The rationale for using
smaller instruments and/or fewer incisions is to minimize tissue trauma and improve postoperative pain and cosmesis for patients,5 but these new techniques may have potential disadvantages, including a long learning curve, prolonged operative time, and decreased visualization and related increased risk of bile duct injuries.5 Therefore, in the evaluation of technical innovations in surgery, there are several main factors that should be considered before a new procedure achieves widespread acceptance 1
Milano-Bicocca University, San Gerardo Hospital, Monza, Italy
Corresponding Author: Luca Gianotti, Department of Surgery, Milano-Bicocca University, San Gerardo Hospital (4° piano B), Via Pergolesi 33, Monza 20052, Italy. Email: [email protected]
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
529
Tamini et al and implementation. These comprise safety, feasibility, efficacy, comparable or improved clinical results, and a benefit to the patient or health care system in terms of cost-effectiveness. Several randomized controlled trials (RCTs) have been conducted to compare standard laparoscopic multiport cholecystectomy (SLMC) with SILC, yet sample sizes of those RCTs were not large enough for a comprehensive evaluation of this new surgical intervention. Recently, there have been several meta-analyses that synthesized data obtained from the published RCTs to evaluate the feasibility and safety of SILS for cholecystectomy in comparison to that for the traditional multiport technique.6-8 The majority of these previous meta-analyses included a limited number of RCTs. The most recent and comprehensive meta-analysis by Pisanu et al9 pooled results of 12 RCTs with a total of 892 patients. In addition, previous reports often did not assess and evaluate the quality of the available evidence, which is very important for clinical decision making. High-quality RCTs needed to evaluate a new surgical procedure are complex to carry out because of several hurdles such as learning curve effects, ethical and cultural resistance, absence of blinding, and urgent or unexpected conditions during operation in surgical treatment. For this reason, evaluation of non-RCTs may be an appropriate strategy to extend the source of evidence.10 Therefore, we designed a meta-analysis that included available data from experimental and observational studies to evaluate the safety and efficacy of SILC and to determine whether SILC is an acceptable alternative to SLMC for the treatment of gallstone disease.
Methods Search and Identification of Studies We performed a MEDLINE, EMBASE, and Cochrane Library extended literature search of all observational and experimental studies published as original articles written in English from January 1997 to February 2012, comparing SILC and SLMC. The Boolean string [“cholecystectomy” AND “SILS” OR “SILC “ OR “Single site” OR “Single port” OR “LESS” OR “Single incision” OR “Laparo endoscopic” OR “One trocar” OR “Single trocar”, limited to title and abstract] was used to perform the search, following the meta-analysis of observational studies in epidemiology guidelines11 and the PRISMA statement for reporting systematic reviews and metaanalyses of studies evaluating health care interventions.12 Three of the authors (NT, MR, and EB) also reviewed reference lists from all the relevant studies, systematic reviews, and previous meta-analyses to identify additional publications of interests.
Eligibility Criteria Study Type. All experimental (RCTs) and observational (non-RCTs) full-text studies published in the English language and comparing SILC with SLMC were evaluated. Abstracts, opinion pieces, and editorials were excluded because they could not provide detailed information about the methodology, implementation, and outcome measures to be assessed. Among the observational studies, only those reporting a control group were included (retrospective case-control and cohort reports). We included cholecystectomy performed for both simple gallbladder stone disease and for acute cholecystitis in adult and pediatric populations. All studies reported at least 1 or more of the following defined clinical outcomes: overall postoperative complications, biliary spillage, conversion to open procedure, operation time, intraoperative blood loss, postoperative pain evaluation, duration of postoperative hospital stay, cosmesis satisfaction, and time to return to work. If 2 or more reports were from the same institution or team, we selected the most recent or the most complete one. Intervention Type. For the SILC group, we considered any cholecystectomy performed with any commercially available devices (by different companies) and self-modified devices. For the SLMC group, we included all the patients operated with at least 3 trocars.
Data Collection Relevant data were extracted independently and in duplicate from all studies by 2 authors (NT, EB) and included study features, population characteristics, data needed for quality assessment, and the different outcome measures. Population characteristics included the number and type of procedures performed, age, gender, and body mass index (BMI). Checklists and electronic databases were used for data extraction.
Outcome Measures The primary outcome of this meta-analysis was safety of SILC as measured by the overall rate of postoperative complications that developed within 30 days of the surgical procedure and the rate of biliary spillage defined as damage of the biliary tract or bile lick from the liver. Feasibility was another primary outcome as measured by the conversion rate to open surgery and operative time. Postoperative pain, as measured by a 10-point Visual Analogue Scale 24 hours after the procedure, length of hospital stay (LOS), perioperative blood loss, time to return to normal activity, cosmetic satisfaction, and the rate of incisional hernia were secondary outcomes. The
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
530
Surgical Innovation 21(5)
latter outcome was originally measured by means of different scales (5 or 10 items) within the studies included in the meta-analysis. For this reason, we converted the Likert-based cosmetic outcome to a percentage of cosmetic satisfaction.
Egger’s test for funnel plot asymmetry.20 Analyses were conducted using STATA 11.0 (STATA Corp, College Station, TX).
Quality Assessment
Study Selection
Two independent authors (LG and EB) assessed the quality of the observational studies by using the previously validated Newcastle-Ottawa Scale,13 whereas we used the Jadad Scale14 to score the quality of experimental studies. Because in this type of studies double blinding is impossible to perform, in the Jadad Scale we gave a score of 1 to the trials reporting blinding of the patient and independent outcome assessor. Experimental studies with Jadad scores ≥3 and observational studies with NewcastleOttawa Scale score ≥8 were considered to be of high quality.
The process of article selection is described in Figure 1. We identified 220 relevant studies from the electronic literature search, and after a detailed evaluation, 176 of them were excluded because they did not satisfy the inclusion criteria—that is, case-series studies reporting only results of SILC without a comparison with SLMC or studies investigating procedures performed through natural orifices. Moreover, 2 studies21,22 were not considered because we included the updated articles on the same study.23,24
Results
Study and Population Characteristics Statistical Analysis For continuous outcomes, statistical analyses were performed by using the weighted mean difference (WMD) as the summary statistic. The WMD is a dimensionless effect measure that represents the number of standard deviations by which the intervention changes the outcome. For studies that presented continuous data reported as median and range values, we derived the mean and standard deviation by applying the formula developed by Hozo et al15 Using this formula, we could include within our meta-analysis the studies that did not report the mean and standard deviation of the considered outcome. Pooled odds ratios (ORs) were calculated to summarize the effect of binary outcomes. In this setting, we applied the Yates 0.5 correction factor16 for studies containing zero events in either SILC or SLMC groups in order to guarantee the estimability of the OR and, thus, to include the study within the meta-analysis. All pooled outcome measures were determined using random-effects models.17 We assessed the between-studies heterogeneity using the I2 statistic, which is the proportion of total variation contributed by the between-studies variance.18 In the absence of betweenstudies heterogeneity (P > .05), the random-effects pooled estimates are identical to those obtained from fixed-effects models. We generated forest plots for all the investigated outcomes. We also conducted stratified analyses according to geographic area of the study, study quality index, age and BMI of study participants, and surgeon experience in SILC. For this, we used the median value of age, BMI, and study quality index to perform the stratified analyses. We performed a publication bias analysis through the counter-enhanced funnel plot19 and
The present meta-analysis included 43 studies comparing SILS and SLMC, of which 30 were observational repo rts23-52 with 5367 cases and 13 were experimental trials53-65 with 2131 cases, yielding a total of 7489 patients (Table 1).
Primary Outcome Measures The descriptive characteristics of the primary outcome measures are shown in Table 2. A total of 41 studies reported postoperative morbidity. The analysis was conducted in 7399 patients (2015 SILC and 5348 SLMC). The overall rate of complications was comparable between SLMC and SILC (relative risk [RR] = 1.08; 95% CI = 0.87-1.35; P = .46). No significant differences were also observed by analyzing experimental (RR = 1.09; 95% CI = 0.81-1.47; P = .57) and observational studies separately (RR = 1.08; 95% CI = 0.79-1.47; P = .63; Figure 2). The Egger P value for publication bias analysis was .10 (Figure 3). The analysis of postoperative morbidity was stratified according to different potential risk factors. None of the considered stratification variables were significantly related to the occurrence of complications (Table 3). We could retrieve information on bile spillage in 29 studies for a total of 6648 patients (1719 SILC and 4929 SLMC). The rate was comparable between SLMC and SILC (RR = 1.16; 95% CI = 0.73-1.84; P = .53; Figure 4), with no difference in observational (RR = 1.16; 95% CI = 0.65-2.05; P = .62) and experimental trials (RR = 1.17; 95% CI = 0.54-2.53; P = .40). No evidences of publication bias were found for biliary spillage (Egger test P = .21).
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
531
Tamini et al
Figure 1. Process of study selection according to the PRISMA statement.
Abbreviations: RCT, randomized controlled trial; NOTES, natural orifice transluminal endoscopic surgery.
Figure 5 represents the forest plot of the RR for conversion rate to open surgery. This outcome was reported in 10 experimental trials and 27 observational studies, for a total of 3463 patients. The conversion rate was similar in the 2 groups (RR = 0.88; 95% CI = 0.53-1.46; P = .62). Similar results were obtained in a separate analysis of RCTs (RR = 1.23; 95% CI = 0.49-3.08; P = .66) and nonRCTs (RR = 0.75; 95% CI = 0.40-1.40; P = .37). No evidences of publication bias were found for conversion to laparotomy (Egger test P = .87). The needle gallbladder suspension technique was reported in 26/43 (60.4%) of the trials (53.8% in RCTs and 63.3% in non-RCTs). Additional trocars during SILC were needed in 103 patients out of 1693 (6.1%) within 27 observational studies (3 studies did not report data on additional trocar use) and in 28 patients out of 533 (5.2%) in 13 randomized studies. There was no significant difference between observational and experimental studies (P = .49) in the use of additional ports during SILC. The time needed to complete the surgical procedure was reported in 42 studies (29 observational and 13
experimental), with a total of 7427 patients. This outcome was in favor of the SLMC group (WMD = 0.73; 95% CI = 0.67-0.79; P < .0001). Separate analysis for experimental (WMD = 0.74; 95% CI = 0.61-0.88; P < .0001) and observational studies (WMD = 0.72; 95% CI = 0.65-0.79; P < .0001) showed similar results (P for heterogeneity = .81; Figure 6). The P value for publication bias analysis by the funnel plot was .11 (Figure 7). A stratified analysis for potential factors affecting operative time showed a significant heterogeneity among different strata. Low study quality, age >47 years, BMI >27 kg/m2, ASA score >2, and geographic area of study origin seemed to be associated with longer operative time in SILC with respect to SLMC. Previous surgeon experience with SILC was not significantly related to the time needed to complete the operation (Table 4).
Secondary Outcome Measures The descriptive characteristics of the primary outcome measures are shown in Table 5. All secondary outcomes
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
532
Surgical Innovation 21(5)
Table 1. Main Characteristics and Demographic Data of the Studies Included in the Meta-analysis.
Study: First Author (Year) Observational studies Hernandez (2009)23 Kravetz (2009)25 Philipp (2009)26 Chow (2010)27 Fronza (2010)28 Ito (2010)29 McGregor (2011)30 Navarra (2010)31 Thapa (2010)32 Bagloo (2011)33 Cao (2011)34 Chandler (2011)35 Chang (2011)36 Emami (2011)37 Froghi (2012)38 Gangl (2011)39 Garcia-Henriquez (2011)40 Garg (2012)41 Jacob (2011)42 Joseph (2011)43 Khambaty (2011)44 Kilian (2011)45 Prasad (2011)46 Rupp (2011)47 Sinha (2011)48 Vidal (2011)24 Wren (2011)49 Wu (2011)50 Barband (2012)51 Wong (2012)52 Subtotal Experimental studies Aprea (2011)53 Lee (2010)54 Rasic´ (2010)55 Tsimoyiannis (2010)56 Asakuma (2011)57 Bucher (2011)58 Cao (2011)59 Lai (2011)60 Lirici (2011)61 Ma (2011)62 Phillips (2012)63 Sinan (2012)64 Zheng (2012)65 Subtotal Total
SILC Group
Country
Study Quality Indexa
Spain USA USA UK USA Japan UK Italy Nepal USA China USA Singapore USA UK Austria USA India Germany USA USA Germany India USA India Spain Austria China Iran China
6 7 8 6 8 7 9 7 7 7 7 6 6 7 9 7 6 6 9 8 8 8 6 8 7 7 8 5 6 7
100 20 29 41 25 5 11 20 20 40 36 42 30 40 16 67 27 25 36 108 107 16 100 101 240 120 10 100 25 20 1577
46 ± 16 44 38 (18-85) 43 ± 16 44 ± 15 56 ± 12 45 ± 4 45 (28-65) 45 ± 4 45 (24-80) 47 15 (6-19) 48 (25-69) 15 (8-20) 47 ± 17 53 (15-81) 14 (3-18) 44 ± 14 62 (21-81) 48 ± 17 45 38 (20-69) 38 44 ± 18 38 (20-82) 56 (29-70) 59 ± 16 46 ± 13 41 (21-61) 59 (39-66) -
Italy Taiwan Croatia Greece Japan Switzerland China China Italy USA USA Turkey China
2 4 1 3 2 2 3 2 3 2 2 2 2
25 35 48 20 24 75 57 24 20 21 117 17 30 513 2090
46 ± 9 51 ± 14 44 ± 6 49 ± 17 57 (47-62) 42 (18-81) 62 ± 5 52 ± 13 45 (26-63) 57 ± 16 46 (18-77) 49 ± 9 44 ± 11 — —
n
Age, years
SLMC group n
Age, years
M/F
BMI (kg/m2)
25/75 29 ± 6.5 4/16 30.2 4/25 30 (19-43) 12/29 26.6 ± 5 2/23 25.4 ± 5 2/3 23.4 ± 3 2/9 27 ± 1.4 8/12 27 (23-36) 5/15 NA 13/27 27.8 (21-40) 19/17 NA 6/36 NA 12/18 NA 5/35 28.2 ± 7.8 5/11 28.3 ± 6.3 16/51 26 (18-43) 9/18 21.7 (11-33) 5/20 25.1 ± 4.1 19/17 27.8 (19-40) 22/86 28.3 ± 5.8 NA 29.5 7/9 27 (20-33) 44/56 27.7 17/84 29.7 ± 7.4 24/216 NA 43/77 NA 7/3 27.7 ± 3.3 25/75 NA 4/21 29 (22-34.7) 4/16 22.9 ± 2.8 -
30 20 22 58 25 23 24 74 20 83 24 27 30 68 13 67 27 27 37 177 44 20 100 100 3492 120 10 100 30 20 4912
48 ± 17 46 39 (20-75) 51 ± 17 40 ± 12 51 ± 17 56 ± 3 NA 41 ± 3 57 (19-93) 48 15 (3-19) 48 (24-64) 15 (3-22) 48 ± 10 56 (21-87) 15 (6-23) 44 ± 13 60 (21-94) 50 ± 16 48 56 (26-82) 38 41 ± 20 NA 55 (32-76) 62 ± 16 50 ± 14 43 (23-60) 56 (30-65) -
9/21 6/14 4/18 16/42 4/21 10/13 7/17 25/49 1/19 25/58 14/10 5/22 12/18 12/56 3/10 16/51 10/17 4/23 15/22 56/121 NA 10/10 42/58 18/82 NA 38/82 7/3 47/53 5/25 5/15 -
29 ± 7 28.2 29 (17-52) 28.2 ± 5.8 30.7 ± 7 24 ± 5.4 31.7 ± 1.1 NA NA 28.9 (15.3-67.6) NA NA NA 28.1 ± 7.8 26.3 ± 5.8 27 (20-46) 20 (15-43) 24.4 ± 3.8 27.4 (21-52) 30.3 ± 7.5 30.7 25 (18-31) 27.3 32.5 ± 8.6 NA NA 28.4 ± 6.2 NA 29.9 (23.2-34.8) 23.6 ± 2.3 -
9/16 13/22 22/26 5/15 11/13 NA 23/34 8/16 6/14 2/19 28/89 4/13 13/17 — —
25 35 50 20 25 75 51 27 20 22 80 17 30 477 5389
44 ± 10 53 ± 16 44 ± 6 48 ± 10 66 (55-72) 44 (20-78) 60 ± 4 54 ± 12 50 (24-67) 46 ± 12 44 (19-68) 49 ± 14 47 ± 14 — —
6/19 15/20 18/32 1/19 13/12 NA 22/29 11/16 6/14 2/20 24/56 8/9 16/14 — —
23.7 ± 4.6 25.8 ± 3 27 ± 4 25.3 ± 5.3 24.1 (20.4-25.9) 25 (19-34) 29.1 ± 5.1 24.4 ± 2.8 27 (18-30) 30.7 ± 6.1 31 (18.8-45) 27.2 ± 2.9 25.9 ± 4.1 — —
M/F
BMI (kg/m2)
25.9 ± 5.8 24.2 ± 3.4 27 ± 4 23.7 ± 3.2 24 (21.4-27.1) 26 (22-35) 28.6 ± 4.4 25 ± 3 25 (18-29) 28.2 ± 5.3 28.9 (15.4-44.8) 27.3 ± 3.1 24.7 ± 3.4 — —
Abbreviations: SILC, single-incision laparoscopic cholecystectomy; SLMC, standard laparoscopic multiport cholecystectomy; M, male; F, female; BMI, body mass index; NA, not available. a Study quality was evaluated using the Newcastle-Ottawa Scale for observational studies (0 = lowest quality, 10 = highest quality) and the Jadad Scale (0 = lowest quality, 5 = highest quality) for experimental studies.
were in favor of the SILC group. Results are summarized in Table 6. In particular, postoperative pain at 24 hours showed a WMD of −0.27 (95% CI = −0.37 to −0.17; P < .0001). This significant difference was confirmed in both experimental (WMD = −0.25; 95% CI = −0.40 to −0.11; P < .0001) and observational studies (WMD = −0.27; 95% CI = −0.40 to −0.15; P < .0001).
LOS was significantly shorter in the SILC group (WMD = −0.29; 95% CI = −0.37 to −0.21; P < .0001). Experimental (WMD = −0.43; 95% CI = −0.58 to −0.28; P < .0001) and observational studies (WMD = −0.24; 95% CI = −0.33 to −0.14; P < .0001) had a similar trend. We also analyzed and split LOS by geographic region. Studies from the United States had the shortest stay (0.97
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
533
Tamini et al Table 2. Primary Outcomes Investigated Within the Studies Included in the Meta-analysis.a Safety Outcome
Study: First Author, Year Observational studies Hernandez (2009)23 Kravetz (2009)25 Philipp (2009)26 Chow (2010)27 Fronza (2010)28 Ito (2010)29 McGregor (2011)30 Navarra (2010)31 Thapa (2010)32 Bagloo (2011)33 Cao (2011)34 Chandler (2011)35 Chang (2011)36 Emami (2011)37 Froghi (2012)38 Gangl (2011)39 Garcia-Henriquez (2011)40 Garg (2012)41 Jacob (2011)42 Joseph (2011)43 Khambaty (2011)44 Kilian (2011)45 Prasad (2011)46 Rupp (2011)47 Sinha (2011)48 Vidal (2011)24 Wren (2011)49 Wu (2011)50 Barband (2012)51 Wong (2012)52 Subtotal Experimental studies Aprea (2011)53 Lee (2010)54 Rasic´ (2010)55 Tsimoyiannis (2010)56 Asakuma (2011)57 Bucher (2011)58 Cao (2011)59 Lai (2011)60 Lirici (2011)61 Ma (2011)62 Phillips (2012)63 Sinan (2012)64 Zheng (2012)65 Subtotal Total
Postoperative complications (n/N), SILC; SLMC
Bile spillage (n/N), SILC; SLMC
Feasibility Outcome Conversion to Open Laparotomy (n/N), SILC; SLMC
Operating Time (minutes), SILC; SLMC
5/100; 0/30 0/20; 1/20 7/29; 3/22 1/41; 3/58 3/25; 2/25 0/5; 0/23 7/11; 15/24 0/20; 0/74 2/20; 2/20 1/40; 0/83 0/36; 0/24 5/42; 1/27 NA 0/40; 0/68 1/16; 1/13 1/67; 1/67 2/27; 1/27 NA 5/36; 6/37 4/108; 8/177 3/107; 1/44 0/16; 1/20 0/100; 0/100 3/101; 5/100 2/240; 38/3492 8/120; 7/120 0/10; 1/10 2/100; 0/100 0/25; 0/30 3/20; 4/20 65/1522; 101/4855b
1/100; 0/30 0/20; 0/20 0/29; 1/22 1/41; 1/58 0/25; 1/25 0/5; 0/23 0/11; 0/24 NA NA NA NA 0/42; 0/27 NA 0/40; 0/68 0/16; 0/13 0/67; 0/67 0/27; 0/27 NA 1/36; 3/37 5/108; 2/177 3/107; 1/44 NA 0/100; 0/100 0/101; 0/100 1/240; 21/3492 3/120; 2/120 NA 1/100; 0/100 0/25; 0/30 NA 16/1360; 32/4604b
1/100; 0/30 0/20; 0/20 0/29; 0/22 0/41; 4/58 0/25; 0/25 0/5; 0/23 0/11; 0/24 0/20; 1/74 0/20; 2/20 0/40; 0/83 0/36; 0/24 0/42; 0/27 NA 0/40; 0/68 0/16; 0/13 1/67; 0/67 0/27; 0/27 NA 1/36; 3/37 0/108; 2/177 0/107; 0/44 0/16; 0/20 0/100; 0/100 0/101; 0/100 NA 0/120; 0/120 1/10; 0/10 0/100; 0/100 0/25; 0/30 0/20; 0/20 4/1282; 12/1363b
72 (20.7); 71 (16.3) 77 (33); 65.7 (33) 90 (25); 78 (32.4) 126 (42.9); 95.8 (33.1) 100.8 (22); 64.5 (18) 154 (57); 100 (51) 86.9 (9); 79.1 (4.2) 53 (14.5); 40 (10.2) 75 (5.7); 39 (2.2) 102 (26.8); 106 (71.8) 56 (16.7); 36.7 (6) 69.4 (15.7); 64.5 (12.1) 86 (56.1); 58 (27.4) 82.4 (27.6); 63.1 (28.7) 113.3 (33.7); 89.3 (19) 89.3 (38.1); 78.3 (41.6) 98.8 (18.7); 62 (14.4) NA 88 (41.5); 94 (51.3) 67.5 (25.8); 59.8 (36.1) 81.5 (28); 69.1 (21) 65 (17.3); 70 (30.1) 66.8 (5.8); 28.1 (1.4) 76 (29); 75 (33) 26.7 (16.7); 18.4 (4.6) 45 (22); 40 (12.9) 105 (18.3); 106.1 (23.5) 53.5 (24); 49.2 (13.8) 56.3 (13.8); 41.7 (9) 78.7 (21.1); 74 (24.3) 80.8 (25.6); 66.1 (23.1)
0/25; 0/25 7/35; 5/35 0/48; 0/50 1/20; 2/20 0/24; 0/25 12/75; 10/75 2/57; 1/51 0/24; 1/27 1/20; 3/20 3/21; 4/22 45/117; 27/80 1/17; 1/17 2/30; 1/30 74/513; 55/477b 139/2035; 156/5332b
NA 0/35; 0/35 NA 1/20; 2/20 NA 9/75; 6/75 NA 0/24; 0/27 0/20; 0/20 NA 0/117; 0/80 0/17; 0/17 0/30; 0/30 10/338; 8/304b 26/1698; 40/4908b
NA 2/35; 1/35 0/48; 0/50 NA 1/24; 2/25 NA 1/57; 0/51 0/24; 0/27 0/20; 1/20 0/21; 0/22 0/117; 0/80 NA 2/30; 0/30 6/376; 4/340b 10/1658; 16/1703b
41.3 (12); 35.6 (5.8) 71.7 (11.6); 48.4 (10.5) 46 (3.5); 43 (4) 49.7 (9); 37.3 (9.2) 109.3 (2.8); 99 (3) 68.3 (22.7); 70.8 (23.1) 55.2 (12.4); 46.3 (10.8) 43.5 (15.4); 46.5 (20.1) 76.8 (3.4); 48.3 (6.7) 88.5 (40.4); 44.8 (40.4) 57.2 (20.8); 45.2 (20.8) 124.4 (29.7); 64.1 (26.1) 55.6 (25.7); 42.7 (18.6) 68.3 (25.8); 51.7 (17.2) 76.9 (26.0); 61.6 (22.2)
Abbreviations: SILC, single-incision laparoscopic cholecystectomy; SLMC, standard laparoscopic multiport cholecystectomy; NA, not available. a Categorical variables are expressed as number of events and continuous variables as mean (standard deviation). b Sum does not add up to the total number of individuals considered within SILC or SLMC groups as a result of missing information on the considered outcome.
Downloaded from sri.sagepub.com at UNIV OF UTAH SALT LAKE CITY on November 25, 2014
534
Surgical Innovation 21(5)
Figure 2. Forest plot for postoperative complications between SILC and SLMC. RRs and 95% CIs are shown.
Abbreviations: SILC, single-incision laparoscopic cholecystectomy; SLMC, standard laparoscopic multiport cholecystectomy; RR, relative risk; CI, confidence interval.
Figure 3. Funnel plot for postoperative complications. Abbreviations: SE, standard error; RR, relative risk.
days in the SILC group and 1.14 in the SLMC group) with respect to Europe (2.31 days in the SILC group and 2.86 in the SLMC group) and Asia (2.31 days in the SILC group and 2.54 in the SLMC group) but without significant difference (P = .17). SILC was associated with a significant reduction in intraoperative blood loss (WMD = −0.16; 95% CI = −0.27 to −0.04; P = .007). Conflicting results (P for heterogeneity
