The safety of the early removal of prophylactic drainage after liver resection based solely on predetermined criteria: a propensity score analysis

Transcription

1 HPB ORIGINAL ARTICLE The safety of the early removal of prophylactic drainage after liver resection based solely on predetermined criteria: a propensity score analysis Daisuke Hokuto, Takeo Nomi, Satoshi Yasuda, Chihiro Kawaguchi, Takahiro Yoshikawa, Kohei Ishioka, Shinsaku Obara, Takatsugu Yamada & Hiromichi Kanehiro Department of Surgery, Nara Medical University, 840 Shijo-cho, Kashihara-shi, Nara, Japan Abstract Background: Prophylactic drainage after liver resection remains a common practice amongst hepatic surgeons. However, there is little information about the optimal timing of drain removal. Methods: From April 2008 to December 2012 (conventional group), the drains were removed based on the treating surgeon s view. From January 2013 to April 2016 (ERP group), the drains were removed on POD 3 if the bile concentration of the drain discharge was less than three times the serum bilirubin on POD 3, and the amount of drain discharge was <500 ml on POD 3. The postoperative outcomes of the two groups were compared using one-to-one propensity score-matching analysis. Results: One hundred nine patients were extracted from ERP group (n = 226) and conventional group (n = 246). The time to drain removal was significantly shorter in the ERP group than in the conventional group (3 days vs. 5 days, P < 0.001). The frequency of delayed bile leakage or the appearance of symptomatic abdominal fluid collection after drain removal did not differ between the two groups (3% vs. 4%, P = 0.791). Conclusion: Drain removal on POD 3 based on the volume and bile concentration is safe. Received 14 October 2016; accepted 3 December 2016 Correspondence Takeo Nomi, Department of Surgery, Nara Medical University, 840 Shijo-cho, Kashihara-shi, Nara, Japan. t.nomi45@gmail.com Introduction Traditionally, prophylactic drainage after liver resection has been used to prevent post-operative fluid collection, detect postoperative bile leakage, and/or bleeding. 1 However, due to advances in percutaneous drainage techniques, several authors have suggested that the routine use of prophylactic drainage after liver resection is unnecessary. 2 4 A recent systematic review reported a pooled meta-analysis comparing routine abdominal drainage with no abdominal drainage in elective liver surgery. No difference was observed between the two groups in terms of mortality or the frequencies of intra-abdominal fluid collection that required reoperation or infected intra-abdominal fluid collection. 5 Furthermore, several retrospective cohort studies and randomized controlled trials have suggested that abdominal drainage after liver resection may increase the risk of complications such as wound infection, retrograde abdominal infection, and ascitic fluid leakage. 2,3,6,7 Yet many hepatic surgeons continue to use routine drainage following hepatic resection arguing early detection of bile leakage or hemorrhage and reduced need for re-intervention as reasons to support such an approach. 8,9 However routine drainage has also been associated with asymptomatic drain infection by normal skin flora leading to retrograde infection. 8 The effect of timing of drain removal on the incidence of retrograde infection has not been studied. The aim of this study was to confirm a safety of early drain removal after liver resection using propensity score matching analysis. Methods Study population Data from all consecutive patients who underwent liver resection at Nara Medical University, Nara, Japan from April 2008 to April 2016 were retrieved from a prospective database for this retrospective study. Patients who underwent biliary reconstruction or

2 360 HPB additional resection of other organs were excluded. Prophylactic drainage was performed after liver resection in all patients. Drains were placed near each cut surface of the liver. During the period from April 2008 to December 2012, the drains were removed when the surgeon who performed the liver resection decided to remove them by visual examination of the drain discharge. During the period from January 2013 to April 2016, the drains were removed on POD 3, regardless of the surgeon s view, if the bile concentration of the drain discharge was less than three times the serum bilirubin concentration on POD 3, and the amount of the drain discharge was <500 ml on POD 3. Drains that did not meet these criteria were not removed until the surgeon that performed the liver resection decided to remove them, and drains were used for the treatment of bile leakage or symptomatic abdominal fluid collection if necessary. The perioperative outcomes of the patients who underwent liver resection between January 2013 and April 2016 (ERP group) were compared with those of the patients who underwent the procedure between April 2008 and December 2012 (conventional group). The patients baseline characteristics, the operative procedures and outcomes, postoperative complications, and the length of the postoperative hospital stay were analyzed. Surgical procedures Major hepatectomy was defined as resection of three contiguous segments according to Couinaud s classification. Anatomical resection included segmentectomy, sectionectomy, hemihepatectomy, and tri-sectionectomy. The surgical procedure was performed as described previously. 10 At the end of the resection procedure, white gauze was placed on the cut liver surface for 5 min, and the presence/absence of bile leakage was confirmed. Bile leakage tests involving the injection of dye through the cystic duct were not performed routinely. Fibrin glue was applied to the liver surface after parenchymal transection. A silicon rubber closed drain (SILASCON Duple drain; KANEKA MEDICAL PRODUCTS, Tokyo, Japan) was placed near to each cut liver surface. Intraoperative parameters, including blood loss, blood transfusion use, and the duration of surgery, were recorded. Postoperative outcomes Parameters associated with postoperative liver function (i.e., serum liver transferase and bilirubin levels) were measured on POD 1, 3, 5, and 7. During the period from January 2013 to April 2016, examinations of the bile concentration of the drain discharge were conducted on POD 1 and 3. Abdominal CT was not performed routinely after liver resection. Postoperative complications were stratified according to the Clavien Dindo classification. 11 Major complications were defined as those of grade IIIa or above. Bile leakage was defined according to the definitions of the International Study Group of Liver Surgery (ISGLS). 12 Surgical site infections (SSI) were defined according to the Centers for Disease Control guidelines. 13 Liver failure was diagnosed according to the ISGLS definition. 14 Management of bile leakage and abdominal abscesses Drains that did not meet the abovementioned criteria were not removed. If bile was detected in the drain discharge or a patient developed a fever of >38 C after POD 3, an abdominal CT scan was performed to confirm the existence of intra-abdominal fluid collection due to the insufficient drainage of bile or symptomatic abdominal fluid collection. If fluid collection was observed near the drain, the drain was replaced with a new drain using a guidewire under radiological guidance (the drain salvage technique). If this procedure was considered impossible or the drain had already been removed, percutaneous drainage was carried out under ultrasound guidance. If the amount of bile leakage was >100 ml, endoscopic retrograde biliary drainage or endoscopic retrograde nasobiliary drainage was performed. If these procedures could not be performed or were not effective, surgical drainage was considered. Management of massive ascites In the ERP group, drains through which >500 ml of discharge passed on POD 3 were not removed. Such drains were removed when the amount of discharge reached <500 ml/day or on POD 7. In the conventional group, drains through which >500 ml of discharge were not removed until the surgeon that performed the liver resection decided to remove them. Statistical analyses Continuous data are expressed as medians and ranges. Qualitative variables are expressed as frequencies (percentages). The t- test or the Mann Whitney U test was used for intergroup comparisons of quantitative variables as appropriate, whereas the c 2 test or Fisher s exact test was used to compare categorical data. Two-sided P-values of <0.05 were considered statistically significant. A one-to-one propensity matching by calculating was performed. Propensity scores were calculated for each patient using logistic regression analysis involving the following covariates: the frequency of cirrhosis, the preoperative serum albumin level, the frequency of previous hepatectomy, the frequency of a laparoscopic approach, the duration of the operation, the estimated amount of intraoperative blood loss, and the frequency of blood transfusions. All statistical analyses were performed using SPSS for Windows version 22.0 (SPSS Inc.). Results Preoperative characteristics and surgical outcomes The preoperative characteristics and surgical outcomes of the full analysis set are summarized in Table 1 and the propensity matched set are summarized in Table 2. Using one-to-one propensity matching 109 pairs of patients were matched. There was no significant difference in preoperative characteristics and surgical outcomes between the ERP group and the conventional group in the propensity matched set.

3 HPB 361 Table 1 Baseline patient characteristics ERP group n [ 226 Conventional group n [ 246 P-value Age, median (range) 70 (17 85) 69 (29 84) Gender, male, n (%) 167 (74) 174 (71) Present illness, n (%) Hepatocellular carcinoma 140 (62) 159 (65) Colorectal liver metastases 65 (29) 75 (31) Other 21 (9) 12 (5) Cirrhosis, n (%) 31 (14) 53 (22) Previous chemotherapy, n (%) 41 (18) 45 (18) Preoperative laboratory data, median (range) Total bilirubin, mg/dl 0.7 ( ) 0.7 ( ) Albumin, g/dl 4.3 ( ) 4.2 ( ) ICG-R15, % 12.0 ( ) 11.3 ( ) Tumor size, mm, median (range) 24 (6 300) 27 (7 160) Multiple tumors, n (%) 74 (33) 63 (26) Previous hepatectomy, n (%) 62 (27) 35 (14) <0.001 Major hepatectomy 24 (11) 37 (15) Anatomical resection, n (%) 100 (44) 97 (39) Laparoscopic liver resection 48 (21) 34 (14) Operation time (min) 340 ( ) 239 (74 642) <0.001 Blood loss (g) 386 ( ) 580 ( ) Blood transfusion, n (%) 34 (15) 60 (24) ICG-R15, indocyanine green retention rate at 15 min. Table 2 Baseline patient characteristics after propensity score matching ERP group n [ 109 Conventional group n [ 109 P-value Age, median (range) 68 (17 85) 68 (29 84) Gender, male, n (%) 77 (71) 75 (69) Present illness, n (%) Hepatocellular carcinoma 64 (59) 61 (56) Colorectal liver metastases 32 (29) 43 (39) Other 13 (12) 7 (6) Cirrhosis, n (%) 11 (10) 11 (10) Previous chemotherapy, n (%) 15 (14) 20 (18) Preoperative laboratory data, median (range) Total bilirubin, mg/dl 0.7 ( ) 0.7 ( ) Albumin, g/dl 4.3 ( ) 4.3 ( ) ICG-R15, % 11.6 ( ) 10.7 (2 34.2) Tumor size, mm, median (range) 23 (8 240) 25 (8 160) Multiple tumors, n (%) 30 (28) 27 (25) Previous hepatectomy, n (%) 16 (15) 16 (15) Major hepatectomy 11 (10) 18 (18) Anatomical resection, n (%) 48 (44) 42 (17) Laparoscopic liver resection 21 (19) 21 (19) Operation time (min) 282 ( ) 269 (85 642) Blood loss (g) 404 ( ) 474 ( ) Blood transfusion, n (%) 18 (17) 19 (17) ICG-R15, indocyanine green retention rate at 15 min.

4 362 HPB Postoperative outcomes Post-operative outcomes of the propensity matched patients are shown in Table 3. The time to drain removal was significantly shorter in the ERP group than in the conventional group (3 days vs. 5 days, P < 0.001). The proportions of patients who developed delayed bile leakage or symptomatic abdominal fluid collection after drain removal did not differ between the ERP group and the conventional group (P = 0.791). The postoperative hospital stay of the ERP group was significantly shorter than that of the conventional group (9 days vs. 13 days, P < 0.001). Treatment of bile leakage and symptomatic abdominal fluid collection in the propensity matched analysis are shown in Table 4. The proportion of patients treated with percutaneous drainage did not differ between the two groups. Discussion This study investigated the safety of early drain removal after liver resection. After the introduction of ERP, drains were removed on POD 3, regardless of the surgeon s view, if the bile concentration of the drain discharge was less than three times the serum bilirubin concentration on POD 3, and the amount of drain discharge was <500 ml on POD 3. Before the introduction of ERP, drain removal was mainly decided based on the surgeon s Table 3 Postoperative outcomes ERP group a n [ 109 Conventional group a n [ 109 P-value Time to drain removal, days, median (range) 3 (3 82) 5 (4 81) <0.001 Drain removal on POD 3, n (%) 89 (82) 0 (0) <0.001 Drain removal on POD 4 5, n (%) 4 (4) 72 (66) <0.001 Drain removal on POD 6 7, n (%) 4 (4) 10 (9) Drain removal after POD 8, n (%) 12 (11) 27 (25) Bile leakage (BL) 13 (12) 7 (6) Diagnosed before drain removal 11 (10) 5 (5) Diagnosed after drain removal 2 (2) 2 (2) Symptomatic abdominal fluid collection (SAF) 4 (4) 10 (9) Diagnosed before drain removal 1 (1) 3 (3) Diagnosed after drain removal 3 (2) 7 (6) Mortality, n (%) 0 0 Morbidity, n (%) 33 (30) 35 (32) Major complications (worse than Clavien Dindo grade IIIa) 10 (9) 15 (14) Drain salvage technique for BL or SAF 4 (4) 6(6) Percutaneous drainage for BL or SAF 2 (2) 7 (6) Surgical drainage for BL or SAF 0 0 Other 4 (4) 1 (1) Postoperative bleeding 0 0 Ascites 4 (4) 3 (3) Pleural effusion 2 (2) 3 (2) Pneumonia 2 (2) 1 (1) SSI, n (%) 15 (14) 25 (23) Superficial incisional SSI 3 (3) 9 (8) Deep incisional SSI 0 1 (0) Organ/Space SSI 13 (12) 17 (16) Liver failure, n (%) 8 (7) 10 (9) ISGLS grade A 2 (2) 8 (7) ISGLS grade B 6 (6) 1 (1) ISGLS grade C 0 1 (1) Hospital stay, days, median (range) 9 (5 127) 13 (6 154) <0.001 ISGLS, International Study Group of Liver Surgery; SSI, surgical site infection; POD, postoperative day. a ERP group and conventional group are propensity matched groups.

5 HPB 363 Table 4 Treatment of bile leakage and symptomatic abdominal fluid collection ERP group a Conventional group a P-value Bile leakage, n 13 7 Surgically inserted drains alone, n Drain salvage technique, n Percutaneous drainage, n Reoperation, n 0 0 Symptomatic abdominal fluid, n 4 10 Antibiotics only, n Surgically inserted 0 0 drains alone, n Drain salvage technique, n Percutaneous drainage, n Reoperation, n 0 0 ISGLS, International Study Group of Liver Surgery; SSI, surgical site infection; POD, postoperative day. a ERP group and Conventional group are propensity matched groups. view, and the median time to drain removal was 5 days. To confirm the safety of early drain removal, one-to-one propensity score-matching analysis, in which 109 patients were included in each group, was performed. In the ERP group, the drains were removed on POD 3 in 82% (89/109) of patients, and only 3% (3/ 89) of these patients developed delayed bile leakage or symptomatic abdominal fluid collection, while in the conventional group the drains were removed on POD 4 5 in 66% (72/109) of patients, and 4% of them (3/72) developed delayed bile leakage or symptomatic abdominal fluid collection. Thus, drain removal on POD 3 based on the abovementioned criteria was equally as safe as drain removal based on the surgeon s view. Furthermore, the postoperative hospital stay of the ERP group was significantly shorter than that of the conventional group (9 days vs. 13 days, P < 0.001). These results indicate that early drain removal is safe and feasible in patients that undergo liver resection. There have been a few previous studies about the criteria for drain removal after liver resection. Yamazaki et al. reported a drain fluid bilirubin level of <3 mg/dl on POD 3 as a criterion for drain removal. 15 They found that the drain fluid bilirubin level on POD 3 was the strongest predictor of infection. However, they did not prospectively analyze the validity of this approach. In addition, Tanaka et al. reported the following drain removal criterion: (drainage fluid bilirubin concentration/serum bilirubin concentration) drainage fluid volume (ml) on POD 2 and They prospectively analyzed 50 patients to validate their approach and concluded that it was feasible. These criteria were similar to our use of the drainage fluid bilirubin concentration on POD 3 as a criterion. It was reported that prolonged drain placement after liver resection is associated with drain fluid infections, including retrograde infections. 9,15 In addition, drain insertion cannot be used to treat all patients with bile leakage or symptomatic abdominal fluid collection after liver resection. Indeed, some patients require percutaneous drainage or reoperation, even if prophylactic drains are inserted. This is one of the reasons why prophylactic drains are considered to be unnecessary. 2,3,7 However, in this study 27 of 34 patients with bile leakage or symptomatic abdominal fluid collection were treated without percutaneous or surgical drainage. Thus, prophylactic drainage after liver resection might be useful for patients that exhibit bile leakage or symptomatic abdominal fluid collection. On the other hand, in patients without bile leakage or symptomatic abdominal fluid collection, prolonged prophylactic drainage after liver resection is a risk factor for retrograde infections. So, definitive criteria for the removal of prophylactic drains are needed. The current studies criteria for early drain removal were developed based on two easily assessable factors. In the propensity score-matching analysis set, 20 of 109 (18%) patients did not meet our criteria, and 14 of the 20 patients developed bile leakage or symptomatic abdominal fluid collection, which means that the defined criteria were useful for predicting bile leakage and symptomatic abdominal fluid collection. This study had certain limitations. First, the study was retrospective in nature. Perioperative management of the ERP group and the conventional group were slightly different. For example, mechanical bowel preparation and postoperative insertion of nasogastric tube were omitted in the ERP group. And skin closure was performed by intradermal suture in the ERP group, while by stapler in the conventional group. Nevertheless, one-toone propensity score matching was performed. Second, it was not possible to evaluate the effects of not performing prophylactic drainage after liver resection because drains were routinely inserted after elective liver resection. Despite these limitations, this study provides surgeons with valuable information regarding the optimal timing of drain removal after liver resection. In conclusion, this study demonstrated that it is safe to remove prophylactic drains on POD 3, regardless of the surgeon s view, if the bile concentration of the drain discharge is less than three times the serum bilirubin concentration on POD 3, and the amount of drain discharge is <500 ml on POD 3. Conflicts of interest None declared. References 1. Thompson HH, Tompkins RK, Longmire, WP, Jr.. (1983 Apr) Major hepatic resection. A 25-year experience. Ann Surg 197: PubMed PMID: Pubmed Central PMCID: PMC Epub 1983/04/01. eng. 2. Sun HC, Qin LX, Lu L, Wang L, Ye QH, Ren N et al. (2006 Apr) Randomized clinical trial of the effects of abdominal drainage after elective hepatectomy using the crushing clamp method. Br J Surg 93: PubMed PMID: Epub 2006/02/24. eng. 3. Liu CL, Fan ST, Lo CM, Wong Y, Ng IO, Lam CM et al. (2004 Feb) Abdominal drainage after hepatic resection is contraindicated in

6 364 HPB patients with chronic liver diseases. Ann Surg 239: PubMed PMID: Pubmed Central PMCID: PMC Epub 2004/ 01/28. eng. 4. Burt BM, Brown K, Jarnagin W, DeMatteo R, Blumgart LH, Fong Y. (2002 Nov) An audit of results of a no-drainage practice policy after hepatectomy. Am J Surg 184: PubMed PMID: Epub 2002/11/16. eng. 5. Gurusamy KS, Samraj K, Davidson BR. (2007) Routine abdominal drainage for uncomplicated liver resection. Cochrane Database Syst Rev, CD PubMed PMID: Epub 2007/07/20. eng. 6. Squires, MH, 3rd, Lad NL, Fisher SB, Kooby DA, Weber SM, Brinkman A et al. (2015 Apr) Value of primary operative drain placement after major hepatectomy: a multi-institutional analysis of 1041 patients. J Am Coll Surg 220: PubMed PMID: Epub 2015/03/01. eng. 7. Ishizawa T, Zuker NB, Conrad C, Lei HJ, Ciacio O, Kokudo N et al. (2014 May) Using a no drain policy in 342 laparoscopic hepatectomies: which factors predict failure? HPB 16: PubMed PMID: Pubmed Central PMCID: PMC Epub 2013/09/03. eng. 8. Kyoden Y, Imamura H, Sano K, Beck Y, Sugawara Y, Kokudo N et al. (2010 Mar) Value of prophylactic abdominal drainage in 1269 consecutive cases of elective liver resection. J hepato-biliary-pancreatic Sci 17: PubMed PMID: Epub 2009/09/04. eng. 9. Tanaka K, Kumamoto T, Nojiri K, Takeda K, Endo I. (2013 Apr) The effectiveness and appropriate management of abdominal drains in patients undergoing elective liver resection: a retrospective analysis and prospective case series. Surg Today 43: PubMed PMID: Epub 2012/07/17. eng. 10. Hokuto D, Nomi T, Yamato I, Yasuda S, Obara S, Yoshikawa T et al. (2016 Dec) The prognosis of liver resection for patients with four or more colorectal liver metastases has not improved in the era of modern chemotherapy. J Surg Oncol 114: PubMed PMID: Epub 2016/09/30. Eng. 11. Dindo D, Demartines N, Clavien PA. (2004 Aug) Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg 240: PubMed PMID: Pubmed Central PMCID: PMC Epub 2004/07/27. eng. 12. Koch M, Garden OJ, Padbury R, Rahbari NN, Adam R, Capussotti L et al. (2011 May) Bile leakage after hepatobiliary and pancreatic surgery: adefinition and grading of severity by the International Study Group of Liver Surgery. Surgery 149: PubMed PMID: Epub 2011/02/15. eng. 13. Mangram AJ, Horan TC, Pearson ML, Silver LC, Jarvis WR. (1999 Apr) Guideline for prevention of surgical site infection, Hospital Infection Control Practices Advisory Committee. Infect control Hosp Epidemiol 20: quiz PubMed PMID: Epub 1999/04/29. eng. 14. Rahbari NN, Garden OJ, Padbury R, Brooke-Smith M, Crawford M, Adam R et al. (2011 May) Posthepatectomy liver failure: a definition and grading by the International Study Group of Liver Surgery (ISGLS). Surgery 149: PubMed PMID: Epub 2011/01/18. eng. 15. Yamazaki S, Takayama T, Moriguchi M, Mitsuka Y, Okada S, Midorikawa Y et al. (2012 Nov) Criteria for drain removal following liver resection. Br J Surg 99: PubMed PMID: Epub 2012/10/03. eng.