Early versus delayed-interval laparoscopic cholecystectomy for acute cholecystitis A metaanalysis H. Lau, C. Y. Lo, N. G. Patil, W. K. Yuen Department of Surgery, University of Hong Kong Medical Center, Queen Mary Hospital, and Tung Wah Hospital, Pokfulam Road, Hong Kong SAR, People���s Republic of China Received: 22 February 2005/Accepted: 6 July 2005/Online publication: 24 October 2005 Abstract Background: Early laparoscopic cholecystectomy has been advocated for the management of acute cholecys- titis, but little evidence exists to support the superiority of this approach over delayed-interval operation. The current systematic review was undertaken to compare the outcomes and e���cacy between early and delayed- interval laparoscopic cholecystectomy for acute cholecystitis in an evidence-based approach using metaanalytical techniques. Methods: A search of electronic databases, including MEDLINE and EMBASE, was conducted to identify relevant articles published between January 1988 and June 2004. Only randomized or quasi-randomized pro- spective clinical trials in the English language comparing the outcomes of early and delayed-interval laparoscopic cholecystectomy for acute cholecystitis were recruited. Both qualitative and quantitative statistical analyses were performed. The effect size of outcome parameters was estimated by odds ratio or weighted mean difference where feasible and appropriate. Results: A total of four clinical trials comprising 504 patients met the inclusion criteria. Failure of conserva- tive treatment requiring emergency cholecystectomy occurred for 43 patients (23%) in the delayed group. Metaanalyses demonstrated a significantly shortened total length of hospital stay in the early group (weighted mean difference, )1.12 95% confidence interval [CI], )1.42 to )0.99 p 0.001). Pooled estimates did not show any significant differences between the two ap- proaches in terms of operation time, conversion rate, overall complication rate, incidence of bile leakage, and intraabdominal collection. Conclusions: The safety and e���cacy of early and de- layed-interval laparoscopic cholecystectomy for acute cholecystitis were comparable. Because evidence sug- gested that early laparoscopic cholecystectomy reduced the total length of hospital stay and the riskof read- missions attributable to recurrent acute cholecystitis, it is therefore a more cost-effective approach for the management of acute cholecystitis. Key words: Acute cholecystitis ��� Laparoscopic chole- cystectomy ��� Metaanalysis ��� Systematic review The management of acute cholecystitis has evolved with the increasing experience of laparoscopic chole- cystectomy in the past decade [16, 19, 23]. The explosive growth of interest in laparoscopic chole- cystectomy in the early 1990s has dramatically chan- ged the management strategy for acute cholecystitis. In the era of open surgery, early cholecystectomy proved to be beneficial for the management of acute cholecystitis in terms of reducing the morbidity rate and shortening the hospital stay [14, 17]. However, early reports of laparoscopic cholecystectomy for acute cholecystitis frequently showed a higher com- plication rate, a prolonged operation time, and a higher rate of conversion to open surgery [11, 21]. Acute cholecystitis was therefore once considered a relative contraindication for early laparoscopic chole- cystectomy. Conservative treatment of acute chole- cystitis followed by delayed-interval laparoscopic cholecystectomy became a commonly accepted prac- tice in the early 1990s [11]. With the growing experience and improvement in laparoscopic skills, recent studies have demonstrated that laparoscopic cholecystectomy is safe for acute cho- Presented at the Society of American Gastrointestinal and Endoscopic Surgeons (SAGES) meeting, Fort Lauderdale, FL, USA, 13���16 April 2005 Correspondence to: H. Lau Surg Endosc (2006) 20: 82���87 DOI: 10.1007/s00464-005-0100-2 �� Springer Science+Business Media, Inc. 2005

lecystitis [1, 2, 4, 5]. The pendulum has now swung to- ward early laparoscopic cholecystectomy for the man- agement of acute cholecystitis, as in the era of open cholecystectomy for acute cholecystitis. However, exist- ing evidence regarding the clinical benefits of early lap- aroscopic cholecystectomy for patients presenting with acute cholecystitis remains elusive. The current system- atic review was undertaken to evaluate and compare the e���cacy of early and delayed-interval laparoscopic cho- lecystectomy for acute cholecystitis in an evidence-based approach using metaanalytical techniques. Methods Literature search A search of electronic databases, including MEDLINE and EMBASE, was conducted to identify appropriate articles comparing the outcomes of early and delayed-interval laparoscopic cholecystectomy between January 1988 and June 2004. A manual search for other studies also was conducted by a review of the bibliographic reference lists with the retrieved articles. Inclusion criteria Only prospective randomized or quasi-randomized trials that com- pared the outcomes of early and delayed-interval laparoscopic chole- cystectomy were included. Studies were excluded if they were retrospective, nonrandomized, or published in a language other than English. Studies comparing early and late laparoscopic cholecystec- tomy during the index admission for acute cholecystitis also were ex- cluded. Early laparoscopic cholecystectomy was defined as surgery within 72 h after establishment of a clinical diagnosis of acute cholecystitis. Delayed-interval surgery was defined as initial conservative treatment followed by interval laparoscopic cholecystectomy 6 to 10 weeks later. To avoid duplication of data, studies with a similar patient population and studies from the same center were accepted only once. Data extraction Data extraction was performed using a standardized data abstract form. Study and patient characteristics were documented and pre- sented in a table format. The major outcome measures for descriptive and quantitative analyses included conversion rate, operation time, overall complication rate, bile leakage rate, incidence of intraabdom- inal collection, and total length of hospital stay. Data pooling and statistics An effect size for each studied outcome parameter was calculated by the odds ratio (OR) using the Mantel���Haenszel method or the weighted mean difference (WMD) according to Cohen��s method where feasible and appropriate. The choice of each individual sta- tistical method depended on whether the measured event was dichotomous or continuous, whereas the choice of a random or fixed- effect model for analysis depended on the Q statistics. The effect sizes of all trials were tested for heterogeneity by the Q statistics, which were an adaptation of the chi-square goodness-of-fit test. The OR was the ratio across different groups for the odds that the event would occur. A 95% confidence interval (CI) was constructed around the effect size to establish its significance. If the 95% CI of OR in- cluded 1, the two groups were not considered to be statistically dif- ferent. An OR with a 95% CI less than 1 favored early laparoscopic cholecystectomy, whereas a value of greater than 1 favored delayed- interval surgery. The WMD was calculated as the difference in the means of the two treatment arms divided by the pooled standard deviation. In studies for which the standard deviation of the studied outcome data was not reported, the effect size was estimated by using the re- ported p values. For WMD, if the 95% CI crossed the null point (zero), the difference attributable to chance could not be excluded. When the null point fell outside the 95% CI of the WMD for the studied outcome parameters, the observed difference was considered statistically sig- nificant. A qualitative review of the outcome data was performed in the event that a formal metaanalysis was not feasible. Statistical analysis was performed using the software program Statistical Package for the Social Sciences (SPSS Inc., Chicago, IL, USA) and Comprehensive Meta-analysis (Biostat, Englewood, NJ, USA). A p value less than 0.05 was considered statistically significant. Results Four clinical trials that compared the outcomes of early and delayed laparoscopic cholecystectomy were re- cruited in the current review [18, 22, 23, 29]. The four recruited studies with details of their patient character- istics are summarized in Table 1. The sample size of the studies ranged from 41 to 169 patients, for a total of 504 patients. In the three prospective randomized trials, the randomization methods included consecutively num- bered sealed envelopes [23], a computer-generated ran- domization list [22], and unspecified measure [18]. Patients frequently were excluded from the trials when they presented with acute cholangitis or acute pancrea- titis, exhibited symptoms for more than a week, or had undergone previous upper abdominal surgery. Lo et al. [23] included three patients with common ductal stones, which were removed using preoperative endoscopic retrograde cholangiopancreatography with papillotomy. Lai et al. [22] excluded these patients (n = 8) from the subsequent analyses and violated the ������intention-to- treat������ principle. Table 1. Recruited studies and details of patient characteristics Studies Year Countries Intervention Sample size Male:female Age (years) Serralta et al. [29] 2003 Spain Early 82 NA 62 (mean) Delayed 87 N/A 60 (mean) Johansson et al. [18] 2003 Sweden Early 74 63% female 58 (mean) Delayed 71 57% female 55 (mean) Lo et al. [23] 1998 Hong Kong Early 45 26:19 59 (median) Delayed 41 21:20 61 (median) Lai et al. [22] 1998 Hong Kong Early 53 23:30 56 (mean) Delayed 51 15:36 56 (mean) NA, not available 83