Updates in Surgery (2020) 72:1041–1051 https://doi.org/10.1007/s13304-020-00854-y ORIGINAL ARTICLE Laparoscopic versus open hemihepatectomy: comprehensive comparison of complications and costs at 90 days using a propensity method Francisco Riquelme1 · César Muñoz1 · Fabio Ausania1 · Amelia J. Hessheimer1 · Ferrán Torres2 · David Calatayud1 · Raffaele Sandomenico1 · Rocío García Pérez1 · Joana Ferrer1 · José Fuster1 · Juan Carlos García‑Valdecasas1 · Constantino Fondevila1,3 Received: 19 May 2020 / Accepted: 21 July 2020 / Published online: 30 July 2020 © The Author(s) 2020 Abstract Laparoscopic hemihepatectomy (LHH) may offer advantages over open hemihepatectomy (OHH) in blood loss, recovery, and hospital stay. The aim of this study is to evaluate our recent experience performing hemihepatectomy and compare complications and costs up to 90 days following laparoscopic versus open procedures. Retrospective evaluation of patients undergoing hemihepatectomy at our center 01/2010–12/2018 was performed. Patient, tumor, and surgical characteristics; 90-day complications; and costs were analyzed. Inverse probability of treatment weighting (IPTW) was used to balance covariates. A total of 141 hemihepatectomies were included: 96 OHH and 45 LHH. While operative times were longer for LHH, blood loss and transfusions were less. At 90 days, there were similar rates of liver-specific and surgical complications but fewer medical complications following LHH. Medical complications that arose with greater frequency following OHH were primarily pulmonary complications and urinary and central venous catheter infections. Complications at 90 days were lower following LHH (Clavien–Dindo grade ≥ III OHH 23%, LHH 11%, p = 0.130; Comprehensive Complication Index OHH 20.0 ± 16.1, LHH 10.9 ± 14.2, p = 0.001). While operating costs were higher, costs for hospital stay and readmissions were lower with LHH. Patients undergoing LHH experience a significant reduction in postoperative medical complications and costs, resulting in 90-day cost equity compared with OHH. Keywords Hemihepatectomy · Laparoscopy · Major liver resection · Comprehensive complication index Introduction The laparoscopic approach to abdominal interventions bet- ter preserves the integrity and function of the abdominal wall and limits exposure and manipulation of abdominal viscera, not only leading to improved cosmesis but also * Constantino Fondevila reduced intraabdominal adhesions and wound and wall cfondevila@ub.edu complications. The loss of manual tissue manipulation and 1 Hepatopancreatobiliary Surgery and Transplantation, control is compensated, at least to some degree, by magni- General and Digestive Surgery Service, Hospital Clínic, fied and direct visualization. In the field of hepatic surgery, IDIBAPS, CIBERehd, University of Barcelona, Barcelona, the results of clinical retrospective studies, meta-analyses, Spain and one randomized trial indicate that the aforementioned 2 Medical Statistics Core Facility, Faculty of Medicine, advantages may translate into less perioperative blood loss IDIBAPS, Hospital Clínic Barcelona, Barcelona, Spain, and improved recovery, with decreased postoperative com- and Biostatistics Unit, Universidad Autónoma de Barcelona, Barcelona, Spain plications and hospital stay [1–4]. For these reasons, the 3 laparoscopic approach is now considered standard-of-care Department of General and Digestive Surgery, Hospital Clínic, University of Barcelona, C/ Villarroel 170, for the resection of lesions in the left lateral and anterior 08036 Barcelona, Spain hepatic segments [5]. 13 Vol.:(0123456789) 1042 Updates in Surgery (2020) 72:1041–1051 In contrast with the aforementioned minor liver resec- multiorgan resection (including bowel resection). Approval tions, laparoscopic hemihepatectomy (LHH) is still con- for the study was obtained from our Institution’s Commit- sidered an advanced technique to only be carried out by tee on Ethics in Medical Research prior to initiating data experienced surgeons. While transection may proceed analysis (study number HCB/2018/1179), which waived the across hepatic anatomical planes, these planes remain need for written consent. intraparenchymal, highly vascularized, and not recognized by standard traction–countertraction maneuvers. As well, Preoperative evaluation vascular and biliary structures to one side of the liver have to be ligated while preserving structures on the contralat- Preoperatively, all patients underwent routine laboratory eral side, and the liver has to be separated to a certain tests and contrast-enhanced multislice abdominal computed extent from the inferior vena cava, intraoperative injury to tomography (CT) and and/or gadolinium-enhanced hepatic which can be devastating. Accordingly, a learning curve magnetic resonance imaging (MRI) and were discussed in of at least 55 cases for right hemihepatectomy after prior multidisciplinary tumor board. For colorectal liver metas- experience with minor laparoscopic liver resections has tases, preoperative chemotherapy was indicated in cases of been suggested [5, 6]. initially unresectable disease until conversion to resectabil- Aside from increased complexity, cost associated with ity, borderline resectable disease, and initially resectable LHH can also be increased, in particular related to the disease with synchronous presentation. Initial resection of use of more expensive surgical instruments and devices the primary tumor was performed in cases with synchronous and longer operative times versus open hemihepatectomy presentation. (OHH). Cost analyses have determined that operative costs may be significantly higher for LHH than for OHH, though the increase in upfront cost may be offset, at least in part, Surgical procedures by improved and less costly postoperative recovery and hospital stay [7–10]. All surgeons performing LHH had previously performed a The laparoscopic approach to hepatic resection was minimum of 30 laparoscopic liver resections, and all sur- introduced at our center in 2005, and we currently per- geons performing both LHH and OHH had extensive experi- form approximately 75% of our hepatectomies laparo- ence performing open hepatic surgery. scopically. Based on the results of previous studies, we hypothesized that fewer complications arise after LHH and this difference may translate into not only a shorter Laparoscopic hemihepatectomy and less costly postoperative stay but also fewer readmis- sions and ongoing complications through the end of the The patient was placed with legs spread and in reverse Tren- third postoperative month, whereby the great majority of delenburg. Five or six ports were used, with the 12-mm complications following major hepatectomy has occurred camera port placed along the anterior axillary line. Pneu- [11]. The aim of the present study is, therefore, to com- moperitoneum was maintained 12–14 mmHg. Following prehensively evaluate the recent years of our experience mobilization of the right or left liver from its ligamentous performing hemihepatectomy and to compare cumulative attachments, intraoperative ultrasound was performed to complications and costs measured up to 90 days following confirm the number and size of the lesion(s). Preparation for laparoscopic versus open procedures. intermittent Pringle maneuver was performed by encircling the hepatic hilum with a tourniquet, which was exteriorized through a thoracostomy tube inserted through the abdominal wall above the pubic symphysis. An extrahepatic intraglis- Patients and methods sonian pedicle approach was used to dissect and encircle the ipsilateral hepatic artery and portal vein. For left LHH, Study design the left lobe was pulled upward, and Arantius’ ligament approach was used to encircle the left hepatic vein [12]. For This study is a retrospective evaluation of patients under- right LHH, retrohepatic veins and the hepatocaval ligament going right or left hemihepatectomy at our center between were divided between titanium or Hem-o-lok® clips, fol- January 2010 and December 2018. Exclusion criteria lowed by dissection and encircling of the right hepatic vein included extended hepatectomy, additional biliary or vas- to perform liver hanging maneuver. Ischemic delimitation of cular resection, previous portal vein embolization, previous the transection plane was achieved after pedicle clamping, hepatectomy (including two-stage hepatectomy), emergency and liver transection was performed using a combination of hepatectomy, living donor hepatectomy, and synchronous LigaSure™ (Medtronic, Minneapolis, Minnesota, USA) and 13 Updates in Surgery (2020) 72:1041–1051 1043 CUSA® Excel (Valleylab, Boulder, Colorado, USA). Vascu- of preoperative chemotherapy. Intraoperative variables lar staplers were used to transect both the hepatic pedicles at included associated procedures (e.g., non-anatomical the level of the hilar plate and hepatic veins and the end of resection or radiofrequency ablation in the contralateral parenchymal transection. The surgical specimen was placed hemiliver), the use and duration of intermittent Pringle in an Endo Catch™ Specimen Retrieval Pouch (Medtronic, maneuver, operative time, blood loss, and reasons for Minneapolis, Minnesota, USA) and removed through a Pfan- conversion to open surgery in the LHH group. Surgical nenstiel incision. The incision was then closed, intraabdomi- pathology was evaluated for tumor size and margins, with nal hemostasis was checked, and an abdominal drain was resection margins distancing ≤ 1 mm from tumor cells placed. classified as “R1”. Postoperative outcomes that were recorded included the need for blood transfusion, post- Hybrid hemihepatectomy operative intensive care stay, postoperative hospital stay, complications, re-interventions, unplanned readmissions, The hybrid technique was defined as a planned procedure and mortality. All complications occurring up to 90 days that started with liver mobilization and isolation of hepatic postoperatively were recorded and graded according to pedicle through laparoscopic approach. Thereafter, paren- the Clavien–Dindo [14] and Comprehensive Complica- chymal transection was performed through a 6–8 cm upper tion Index (CCI) systems, the latter of which computes midline laparotomy, which was also used for specimen all the complications a patient experiences in an index extraction. ranging from 0 to 100 [15]. Postoperative complications were categorized as surgical, medical, or liver specific. Liver-specific complications were classified according to Open hemihepatectomy the International Study Group for Liver Surgery (ISGLS) criteria. Bile leak was defined as a drain bilirubin concen- Open approach was performed using a “J”-shaped or right tration > 3-times higher than the serum bilirubin concen- subcostal incision with subxiphoid extension. After explo- tration at or beyond the third postoperative day and graded ration of the abdominal cavity and hemiliver mobilization, as A—requiring no change in patient’s clinical manage- intraoperative ultrasound was performed. The hepatic hilum ment, B—requiring therapeutic intervention but manage- was encircled with a tourniquet to allow subsequent inter- able without surgical re-intervention, or C—requiring sur- mittent Pringle maneuver. The ipsilateral hepatic artery and gical re-intervention [16]. Post-hepatectomy liver failure portal vein were then dissected and clamped in the hepatic (PHLF) was defined as increased international normalized hilum, delimiting the transection plane. CUSA® was used ratio (INR) and hyperbilirubinemia at or beyond the fifth for parenchymal transection. The hemiliver pedicle was sec- postoperative day and graded as A—analytical deteriora- tioned between sutures and the suprahepatic vein using a tion not requiring any change in management, B—devia- vascular stapling device. An abdominal drain was routinely tion from regular postoperative clinical pathway managed placed. without invasive treatment, or C—PHLF requiring an inva- sive procedure [17]. Postoperative ascites was defined as Definition of variables and outcomes drain fluid output > 10 mL/kg/day at or beyond the third postoperative day [18]. In addition to the hospital elec- The main objective of the study was to assess clinical out- tronic medical record, the regional shared electronic medi- comes and costs associated with LHH and OHH up to cal record was checked for readmissions in other centers. 90 days postoperatively. Cases of LHH that resulted in conversion to open surgery were not censored and were included among the LHH group for all analyses based on Costs intention-to-treat. The following patient-related variables were recorded: age, sex, body mass index (BMI), Ameri- Intraoperative, postoperative, and readmission costs were can Society of Anesthesiologists (ASA) class, and previ- obtained, including variable and fixed costs that were ous abdominal surgery (open or laparoscopic surgery or charged in each case. Intraoperative costs included con- percutaneous radiofrequency ablation). Co-morbid condi- sumables (anesthetic drugs, transfusions, surgical instru- tions were quantified according to the age-adjusted Charl- mentation, and devices) and operating room and staff son comorbidity index, which was calculated excluding costs. Postoperative costs included pharmacy, transfusions, liver tumor variables [13]. Liver-specific variables that ICU and/or floor beds, studies and procedures (radiologi- were recorded included principle diagnosis and history of cal, endoscopic, and/or surgical), and other miscellaneous liver disease. Tumor-specific variables included the num- expenditures (laboratory, nursing, and other professional ber, size, extent, and distribution of lesions and history care). Readmission expenses including re-intervention 13 1044 Updates in Surgery (2020) 72:1041–1051 were grouped under total readmission costs. Taking into by treatment prevalence. The following covariates were account the study inclusion period (2010–2018), costs included in the propensity models: age, sex, ASA classi- were adjusted according to the mean annual inflation fication, BMI, age-adjusted Charlson comorbidity index, rate: 2010 + 1.8%, 2011 + 3.2, 2012 + 2.44, 2013 + 1.42, previous abdominal surgery, previous radiofrequency 2014 − 0.15, 2015 − 0.5, 2016 − 0.2, 2017 + 2.97, ablation, cirrhosis, indication for surgery, number of liver 2018 + 1.67 [19]. lesions, diameter of the largest lesion, number of affected liver segments, unilobar versus bilobar liver disease, Data and statistical analysis preoperative chemotherapy and number of cycles, and planned additional procedure. Categorical variables were Categorical variables are described as frequencies and compared using the Chi-square test and continuous vari- percentages and continuous variables as mean ± standard ables using ANOVA with rank-transformed data for both deviation or median (25–75% interquartile range). The raw and IPTW-adjusted analyses. Covariate balance was propensity score method, which simulates the effects of assessed using the standardized difference, which is the a randomized trial for observational data, was used to difference between groups divided by the pooled standard estimate study outcomes. Inverse probability of treatment deviation. Statistical significance was defined as p < 0.05. weighting (IPTW) of the propensity scores was used to Statistical analyses were performed using SAS 9.4 (SAS create a pseudo-population in which study groups were Institute, Inc., Cary, North Carolina, USA). balanced across covariates using data blinded to outcomes. Weights were derived using logistic regression to estimate average treatment effects in treated patients and stabilized Table 1  Patient, tumor, and surgical characteristics used for the propensity model Raw analysis IPTW analysis OHH (n = 96) LHH (n = 45) p value Standard- OHH LHH p value Standardized ized differ- difference ence Age (years) 65 [57–72] 65 [55–69] 0.476 − 0.120 65 [57–72] 67 [57–69] 0.889 0.028 Sex female 38 (40%) 23 (51%) 0.197 − 0.233 42 (45%) 16 (47%) 0.872 − 0.032 Body mass index 26 [23–28] 25 [23–28] 0.329 − 0.177 26 [23–27] 25 [23–28] 0.461 − 0.150 Age-adjusted Charlson comorbidity 2 [1–4] 2 [1–3] 0.406 − 0.153 2 [2, 3] 2 [1–3] 0.669 − 0.087 index ASA class I 4 (4%) 4 (9%) 0.258 0.192 4 (4%) 2 (5%) 0.872 0.032 II 71 (74%) 34 (76%) 0.839 0.037 70 (76%) 27 (77%) 0.885 0.029 III 21 (22%) 7 (16%) 0.380 − 0.163 19 (20%) 6 (18%) 0.810 − 0.048 Previous abdominal procedure 81 (84%) 38 (84%) 0.991 0.002 80 (86%) 30 (87%) 0.866 0.034 Radiofrequency ablation 3 (3%) 0 0.230 − 0.254 2 (2%) 0 (0%) 0.378 − 0.212 Open surgery 42 (44%) 9 (20%) 0.006 − 0.527 37 (39%) 10 (29%) 0.276 − 0.221 Laparoscopic surgery 39 (41%) 29 (64%) 0.008 0.491 43 (46%) 20 (58%) 0.244 0.233 Cirrhosis 3 (3%) 2 (4%) 0.692 0.069 3 (3%) 1 (3%) 0.973 0.007 Malignant disease 89 (93%) 38 (84%) 0.126 0.262 85 (91%) 30 (87%) 0.455 0.142 Preoperative chemotherapy 44 (46%) 29 (64%) 0.039 0.381 49 (52%) 21 (61%) 0.374 0.178 Multiple lesions 49 (51%) 26 (58%) 0.454 0.136 52 (56%) 19 (55%) 0.890 − 0.027 Number of lesions 2 [1–4] 2 [1–3] 0.903 0.023 2 [1–4] 2 [1–4] 0.593 0.109 Number of affected segments 2 [1–3] 2 [2, 3] 0.587 0.100 2 [2, 3] 3 [2, 3] 0.643 0.095 Maximum diameter of lesions (mm) 33 [22–55] 27 [15–45] 0.114 − 0.283 33 [22–55] 32 [15–45] 0.436 − 0.155 Bilobar disease 26 (27%) 11 (24%) 0.739 − 0.060 23 (25%) 10 (29%) 0.683 0.080 Planned associated procedure 30 (31%) 10 (22%) 0.267 − 0.205 27 (29%) 9 (26%) 0.798 − 0.051 Descriptive statistics are frequencies (%) for categorical variables and median [25–75% interquartile range]. Bold marked figures are for p < 0.05 ASA American Society of Anesthesiologists, IPTW inverse probability of treatment weighting, LHH laparoscopic hemihepatectomy, OHH open hemihepatectomy 13 Updates in Surgery (2020) 72:1041–1051 1045 Results IPTW adjustments, covariates were balanced between the two groups, with all p values < 0.05 and standardized Between 01/2010 and 12/2018, 838 liver resections were differences < 0.25. performed at our center. Based on inclusion–exclusion cri- teria, 141 were included for analysis: 96 OHH and 45 LHH Operative characteristics and intraoperative (70% and 64% right hemihepatectomies, respectively). outcomes Patient demographics and tumor and surgical Table 2 lists variables associated with the surgical inter- characteristics ventions and intraoperative outcomes. Rates of any associ- ated intraoperative procedure were similar between the two Baseline patient demographical information and tumor and groups (29% OHH, 26% LHH, p = 0.798). However, when surgical characteristics are listed in Table 1. The average analyzing associated procedures individually, more patients patient in this study was around 65 years of age and over- in the LHH group underwent concomitant radiofrequency weight (BMI 25–26), with an ASA score of II. Men and ablation (2% OHH, 14% LHH), while more patients in women were equally represented among included patients. the OHH group underwent concomitant non-anatomical The indication for hemihepatectomy was malignant tumor resection in the remnant hemiliver (24% OHH, 12% LHH, in 90% of cases (OHH: 71% metastases, 22% primary liver p = 0.010). As well, more patients in the LHH group were tumors; LHH: 76% metastases, 9% primary liver tumors); subjected to intermittent Pringle maneuver (44% OHH, 84% consequentially, the majority of patients had a history of LHH, p < 0.001). preoperative chemotherapy. In terms of tumor characteris- In the LHH group, approach was purely laparoscopic in tics, the median number was two lesions affecting ≥ 2 liver 35 cases (78%) and hybrid in 10 (22%). All hybrid cases segments, and the median size of the largest lesion was were performed during the first 3 years of the study period. approximately 3 cm. Approximately a quarter of patients Among pure laparoscopic cases, conversion to open surgery presented bilobar disease that required an associated pro- was performed in six (17%): three for right upper quadrant cedure in addition to hemihepatectomy. On raw analysis, a adhesions, two for bleeding, and one for technical failure of greater percentage of patients in the LHH group had previ- the laparoscopic CUSA® device. ously received chemotherapy and undergone laparoscopic Operative times were significantly longer by 1 h in the abdominal procedures as opposed to open ones. Following LHH group (210 [180–255] minutes OHH, 270 [225–335] minutes LHH, p < 0.001), but LHH intraoperative blood Table 2  Operative variables and outcomes Raw analysis IPTW analysis OHH (n = 96) LHH (n = 45) p value OHH LHH p value Right hepatectomy 67 (70%) 29 (64%) 0.526 72% 61% 0.237 Associated procedure (any) 30 (31%) 10 (22%) 0.268 29% 26% 0.798 Non-anatomical resection 24 (25%) 5 (11%) 0.018 24% 12% 0.010 Radiofrequency ablation 2 (2%) 5 (11%) 2% 14% Other 4 (4%) 0 3% 0 Pringle maneuver 42 (46%) 38 (86%) < 0.001 44% 84% < 0.001 Pringle maneuver (min) 18 [15–37] 33 [20–45] 0.028 17 [15–36] 30 [15–40] 0.081 Conversion to open s­ urgerya – 6 (17%) – – – – Bleeding – 2 (6%) – – – – Adhesions – 3 (9%) – – – – Technical ­failureb – 1 (3%) – – – – Operative time (min) 210 [180–255] 270 [225–340] < 0.001 210 [180–255] 270 [225–335] < 0.001 Blood loss (mL) 300 [200–500] 200 [100–300] 0.012 300 [150–500] 200 [50–300] 0.018 R1 resection 13 (14%) 5 (11%) 0.656 13 (14%) 4 (11%) 0.612 Descriptive statistics are frequencies (%) for categorical variables and median [25–75% interquartile range] IPTW inverse probability of treatment weighting, LHH laparoscopic hemihepatectomy, OHH open hemihepatectomy a Calculated as a percentage of pure laparoscopic cases (n = 35) b Failure of the laparoscopic CUSA® 13 1046 Updates in Surgery (2020) 72:1041–1051 Table 3  Postoperative complications and outcomes evaluated up to 90 days postoperatively Raw analysis IPTW analysis OHH (n = 96) LHH (n = 45) p value OHH LHH p value At discharge Overall complications 62 (65%) 19 (42%) 0.012 60% 39% 0.039 Clavien–Dindo I/II 46 (48%) 15 (33%) 0.103 44% 29% 0.127 Clavien–Dindo ≥ III 16 (17%) 4 (9%) 0.217 16% 10% 0.416 CCIa 18.0 ± 16.2 10.7 ± 18.3 0.001 16.5 ± 16.1 9.7 ± 14.2 0.009 Overall postoperative stay (days) 8 [711] 6 [5–7] < 0.001 8 [7–10] 6 [5–7] < 0.001 ICU stay (days) 3 [1–4] 1 [0–3] < 0.001 3 [1–4] 1 [0–2] < 0.001 Blood transfusion 17 (18%) 1 (2%) 0.010 15% 1% 0.024 PRBC units 0.36 ± 0.92 0.04 ± 0.30 0.010 0.32 ± 0.87 0.02 ± 0.18 0.018 At 90 days Liver-specific ­complicationsb 16 (17%) 6 (13%) 0.611 16% 12% 0.600 Bile leak (any) 9 (9%) 4 (9%) 0.926 9% 9% 0.948 Grade A 2 (2%) 1 (2%) 0.396 2% 1% 0.610 Grade B 5 (5%) 0 4% 0 Grade C 2 (2%) 2 (4%) 3% 5% Ascites 5 (5%) 2 (4%) 0.846 5% 3% 0.685 Encephalopathy 5 (5%) 0 0.119 4% 0 0.211 ISGLS PHLF (any) 50 (52%) 13 (30%) 0.010 53% 34% 0.053 Grade A 41 (43%) 12 (27%) 0.062 45% 32% 0.213 Grade B 8 (8%) 1 (2%) 8% 2% Grade C 1 (1%) 0 1% 0 General complications 54 (56%) 14 (31%) 0.005 53% 28% 0.010 Surgical complications 23 (24%) 5 (11%) 0.075 22% 12% 0.199 Deep surgical site infection 11 (12%) 3 (7%) 0.375 10% 6% 0.546 Intestinal obstruction 2 (2%) 1 (2%) 0.958 2% 4% 0.459 Hemorrahge/hematoma 8 (8%) 0 0.046 7% 0 0.105 Superficial surgical site infection 8 (8%) 1 (2%) 0.166 9% 2% 0.148 Medical complications 37 (39%) 9 (20%) 0.029 37% 16% 0.022 Arrhythmia 2 (2%) 0 0.330 2% 0 0.443 Pneumonia 11 (12%) 1 (2%) 0.067 10% 1% 0.084 Pleural effusion 6 (6%) 0 0.087 6% 0 0.134 Paralytic ileus 12 (13%) 7 (16%) 0.620 11% 14% 0.688 Urinary tract infection 6 (6%) 0 0.087 6% 0 0.140 CVC infection 8 (8%) 0 0.046 7% 0 0.103 Pulmonary embolism 2 (2%) 1 (2%) 0.958 3% 1% 0.477 Overall complications 71 (74%) 23 (51%) 0.007 69% 44% 0.010 Clavien–Dindo I/II 48 (50%) 18 (40%) 0.267 46% 33% 0.189 Clavien–Dindo ≥ III 23 (24%) 5 (11%) 0.075 23% 11% 0.130 CCIa 21.6 ± 16.1 12.7 ± 18.3 < 0.001 20.0 ± 16.1 10.9 ± 14.2 0.001 Re-interventions at 90 days 6 (6%) 2 (4%) 0.666 6% 5% 0.758 Readmissions at 90 days 17 (18%) 3 (7%) 0.080 17% 4% 0.055 Mortality at 90 days 0 1 (2%) 0.143 0 1% 0.340 Descriptive statistics are frequencies (%) for categorical variables and mean ± standard deviation or median [25–75% interquartile range] CCI Comprehensive Complication Index, ICU intensive care unit, IPTW inverse probability of treatment weighting, ISGLS International Study Group of Liver Surgery, PHLF post-hepatectomy liver failure, PRBC packed red blood cell, SD standard deviation a CCI for entire group; for only patients that developed complications, CCI at discharge was 27.0 ± 12.2 OHH and 25.2 ± 20.8 LHH and at 90 days 28.4 ± 12.0 OHH and 24.8 ± 18.8 LHH b Does not include ISGLS PHLF Grade A 13 Updates in Surgery (2020) 72:1041–1051 1047 loss was less (300 [150–500] mL OHH, 200 [50–300] mL management, occurred at similar rates following both LHH, p = 0.018). There were no differences in rates of “R1” OHH and LHH. Surgical complications also occurred resection (tumor cells ≤ 1 mm from the transection margin) at relatively similar rates, though there was a tendency detected between the two groups: 14% OHH and 11% LHH, toward more postoperative hemorrhage and superficial sur- p = 0.612. gical site infection among OHH patients. Medical compli- cations, on the other hand, occurred at higher rates in the Postoperative complications and outcomes OHH group when compared with LHH (37% OHH, 16% LHH, p = 0.022). Medical complications were primarily Postoperative complications and events are described in pulmonary complications, including pneumonias and pleu- Table 3. Consequent to less intraoperative blood loss, fewer ral effusions, and complications associated with the use patients undergoing LHH received perioperative blood trans- of indwelling catheters, such as urinary tract infections fusions (15% OHH, 1% LHH, p = 0.024). As well, fewer and bloodstream infections arising from central venous complications occurred in the LHH group prior to hospital catheters. discharge (overall 60% OHH, 39% LHH, p = 0.039), and Rates of surgical re-intervention during the first 90 days CCI measured up to discharge was also lower with LHH did not differ between the OHH and LHH groups (6% and (16.5 ± 16.1 OHH, 9.7 ± 14.2, p = 0.009). These differences 5%, respectively, p = 0.758). Six patients undergoing OHH were reflected in a reduction in ICU and overall postopera- were re-operated: two each for hemorrhage and bile leak tive hospital stays for patients in the LHH vs. OHH groups during the index admission and another two for bowel by a median of 2 days for each (p < 0.001 in both cases). obstruction on a subsequent readmission. Two patients By 90 days after surgery, overall complication rates undergoing LHH were re-operated laparoscopically for were 69% OHH and 44% LHH (p = 0.01). Accordingly, bile leak during the index admission. CCI at 90 days was also lower with LHH (20.0 ± 16.1 There was a greater tendency toward readmissions OHH, 10.9 ± 14.2 LHH, p = 0.001). Upon examination among patients undergoing OHH as opposed to LHH (17% of the specific complications that arose during 90 days, vs. 4%, respectively), though this difference did not reach liver-specific complications, including bile leak, ascites, the level statistical significance (p = 0.055). In terms of encephalopathy, and PHLF requiring a change in patient mortality, there was only one death prior to 90 days in Table 4  Cost analysis evaluated up to 90 days postoperatively Raw analysis IPTW analysis OHH (n = 96) LHH (n = 45) p value OHH LHH % difference p value Intraoperative costs Consumables 1801 [1468–2404] 3374 [2898–3889] < 0.001 1777 [1429–2387] 3253 [2958–3889] + 83 < 0.001 OR time 1408 [1167–1748] 1827 [1452–2224] < 0.001 1386 [1161–1745] 1730 [1393–2225] + 29 < 0.001 OR staff 821 [637–1262] 1290 [953–1551] < 0.001 788 [632–1318] 1264 [902–1548] + 60 < 0.001 Total 4309 [3584–5348] 7239 [5542–8244] < 0.001 4339 [3541–5346] 6917 [5314–8363] + 59 < 0.001 Postoperative costs Pharmacy 184 [50–297] 100 [30–175] 0.067 163 [39–274] 100 [27–157] − 40 0.114 Transfusions 90 ± 162 22 ± 111 < 0.001 80 ± 153 10 ± 67 − 88 0.001 Postoperative stay 3896 [2603–4972] 2070 [1648–3313] < 0.001 3574 [1978–4805] 2046 [1560–3012] − 43 < 0.001 ICU 2379 [1121–3311] 797 [0–2212] < 0.001 2212 [0–3188] 761 [0–1592] − 65 < 0.001 Surgical floor 1492 [1183–1872] 1111 [857–1533] < 0.001 1444 [1182–1850] 1182 [887–1560] − 18 0.002 Techniques 142 [67–339] 87 [32–166] 0.008 124 [53–302] 60 [31–177] − 69 0.086 Miscellaneousa 707 [501–1117] 515 [302–632] < 0.001 651 [455–1075] 515 [294–580] − 21 0.002 Total 4990 [3751–6434] 2924 [2079–4449] < 0.001 4817 [2956–6251] 2875 [2066–4109] − 40 < 0.001 Total surgical 9916 [7766–12290] 10,448 [7703– 0.270 9254 [7134–11908] 10,396 [7540– + 12 0.195 admission costs 13730] 13730] Readmission costs 372 ± 1043 232 ± 1195 0.140 331 ± 944 110 ± 699 − 67 0.067 Total cost at 90 days 10,112 [7849– 10,931 [8374– 0.300 9936 [7134–11923] 10,448 [7703– +5 0.286 12433] 14240] 14240] Values are reported in Euros (€) and presented as mean ± standard deviation or median [25–75% interquartile range] a Laboratory tests and medical care during floor stay 13 1048 Updates in Surgery (2020) 72:1041–1051 the entire cohort: a patient that underwent LHH died of between discharge and 90 days among OHH patients and 5% massive pulmonary embolism during the first postopera- more complications between discharge and 90 days among tive week. those undergoing LHH. Additional postoperative complica- tions are not only relevant to patient satisfaction and quality Costs of life but are also known to be a primary factor that serve to increase healthcare costs related to surgery [25, 26]. Perioperative costs are illustrated in Table 4. All operating The overall rate of complications detected by 90 days room costs, including consumable material and costs associ- in this study was ostensibly high: 69% for OHH and 44% ated with OR use and staffing, were higher for LHH versus for LHH. On the other hand, rates of moderate and severe OHH by a median of approximately 2500€ per procedure complications requiring invasive therapy or management (p < 0.001). In contrast, postoperative costs, in particular in intensive care (i.e., Clavien-Dindo III–V: 23% OHH and those related to ICU and floor stays and medical staffing 11% LHH) were relatively low in comparison with other and care, were lower for LHH versus OHH by a median of recent studies examining outcomes following hemihepatec- roughly 2000€ per patient (p < 0.001). Costs for readmis- tomy [6–10]. Moreover, among patients developing compli- sions in Table 4 are presented as the median for all patients cations, CCI, which provides a comprehensive view of all included in each group; costs for only cases where read- postoperative complications, with each graded according to missions occurred were OHH (17 readmissions) 2049€ its clinical impact (i.e., resources required to treat) remained [1229–2411] and LHH 3 readmissions 1396€ [1229–4489], below benchmark values in both groups, both at discharge p = 1.0. Considering that median readmission costs were (< 27.9) and at 3 months (< 32.6) [11]. also reduced by about 200€ among LHH patients (a differ- Liver-specific complications were similar between the ence that did not reach statistical significance, p = 0.067), two groups in this study, though there was a tendency the overall balance of 90-day costs associated with OHH towards a higher rate of grade A PHLF (ongoing abnor- and LHH was similar: 9936€ [7134–11,908] and 10,448€ mal laboratory values requiring no change in clinical man- [7703–14,240], respectively, p = 0.286. agement) following OHH. Rates of surgical complications (intraabdominal abscess/fluid collection, mechanical intes- tinal obstruction, hemorrhage/hematoma, and superficial surgical site infection) were the same. However, rates of Discussion medical complications were significantly higher for OHH. Medical complications not reaching the point of requir- In this study, we meticulously evaluated complications and ing invasive and/or intensive care treatment may have been costs associated with hemihepatectomy arising between under-detected and/or -reported in previous retrospective index procedures up to 90 days postoperatively. The major studies evaluating outcomes following hemihepatectomy findings were a relatively high rate of largely medical com- [6, 7, 10, 27]. We exhaustively evaluated all postoperative plications following OHH that were concomitantly associ- events, including delay in return of normal bowel function, ated with increased ICU and overall postoperative hospital which was surprisingly not improved among patients under- stays as well as higher postoperative costs relative to LHH. going LHH. Manipulation of the bowel in both LHH and Readmission costs for OHH also tended to be higher. In even OHH remains limited with respect to other intraab- contrast, upfront intraoperative costs associated with OHH dominal procedures, yet hemihepatectomy can significantly were lower, and overall costs for the open and laparoscopic impact bowel function based on the aggressive nature of the approaches to hemihepatectomy were equivalent after 3 procedure itself, splanchnic congestion developing second- months of follow-up. ary to significant reduction in the size of the portal sinusoi- Previous studies following hemihepatectomy patients dal bed, and a certain degree of portal hypertension devel- during their index hospitalization up to the end of the first oping in the postoperative period [28]. These events are the postoperative month have also found relative cost equity same regardless of the surgical approach. As well, due to when comparing open and laparoscopic procedures [8–10]. inability to perform manual compression of the cut surface It has been determined for major surgical interventions, how- during transection, we frequently use the intermittent Pringle ever, that limiting the period of observation to the end of the maneuver during laparoscopic hepatectomy to limit blood first postoperative month may overlook significant further loss during parenchymal transection. For this reason, rates morbidity, mortality, and costs, and a 90-day observation of utilization and overall length of hepatic hilar clamping period appears to more accurately capture all relevant post- were significantly greater for LHH than OHH. This, too, operative events [11, 20–24]. In the present study, additional might have contributed to the 14% rate of postoperative par- complications occurring beyond discharge were not negli- alytic ileus following LHH that we observed in our series. gible, with approximately 10% more complications arising Given that return to normal bowel function is one of the key 13 Updates in Surgery (2020) 72:1041–1051 1049 endpoints in some prospective trials aiming to determine the the raw and unadjusted findings. Nonetheless, even IPTW impact of the laparoscopic approach on “functional recov- adjustments cannot achieve the same balanced distribu- ery” following surgery [29–31], this finding is noteworthy tion of unquantified or unidentified variables as in a ran- and a cause for reflection as to whether such an endpoint domized design [35]. Another limitation is the relatively is the most appropriate to evaluate the benefits and impact small sample size, which is partially a consequence of of laparoscopy in the setting of this particular procedure or increasing use of parenchymal-sparing surgical techniques, procedures of a similar nature. particularly for the resection of metastatic lesions [36]. As Pulmonary complications are common following liver well, there is the fact that ten patients in the LHH group resection, arising in > 20% of cases in most series [32, 33]. were operated using a hybrid approach. These patients A French multicenter study specifically focused on pulmo- belonged to the first 3 years of the inclusion period. It is nary complications following major hepatic resection and important to note that all hybrid cases were planned preop- detected a high rate for open procedures (41%) that was eratively and were not unplanned conversions, which have significantly reduced following laparoscopic procedures been shown to be associated with greater perioperative (13%) [34]. In this study, pulmonary complications (includ- morbidity and mortality [37]. Comparing the 35 pure lapa- ing pneumonia, pleural effusion, and pulmonary embolism) roscopic with the ten hybrid hemihepatectomies, both CCI were detected in 19% of patients following OHH but only at discharge and 90 days and costs at discharge and 90 days 2% of patients following LHH. Actual or perceived pain tended to be lower for the hybrid procedures, by and large and other functional limitations associated with open inci- based on the fact that intraoperative costs for the planned sion and costal retraction in OHH contribute to suboptimal hybrid procedures were lower (7260€ [5723–7988] and pulmonary hygiene in the postoperative period. Following 4937€ [3473–6954], respectively, p = 0.024). laparoscopic and even hybrid hemihepatectomy; however, In summary, while laparoscopic hemihepatectomy early abdominal wall pain and function are considerably remains a complex surgical procedure, it appears to offer improved, leading to better respiratory and earlier physical considerable benefits over the open approach in terms of activity. morbidity and costs arising postoperatively, not only dur- Medbery and colleagues introduced the concept of ing index admission but also in subsequent readmissions. “value” into the armamentarium of the benefits offered by These savings in the postoperative period serve to off- laparoscopic liver surgery—that quality (e.g., outcomes) can set the initial increase in operating room costs associated be improved while cutting costs [8]. Unlike the singular and with laparoscopy. In the future, as experience with and largely nascent endpoints that are used in most surgical tri- results following minimally invasive liver surgery continue als, value takes into account not only perioperative events to improve, it is plausible that the laparoscopic or other but also more remote outcome measures, such as physical minimally invasive approaches may eventually become the recovery, ongoing pain, incisional hernias, disease recur- standard-of-care for hemihepatectomy as they already have rence, and survival. Given that the outcome benefits that for other more minor forms of liver resection. laparoscopy offers may only be compounded over time, LHH appears to offer a value advantage over OHH. Furthermore, the value of LHH is likely to improve further still as sur- Author contributions Conception and design: FR, CM, FA, AJH, and CF. Provision of study materials or patients: DC, RGP, JF, JF, JCG- gical expertise increases and upfront costs associated with V, and CF. Collection and assembly of data: FR, CM, and RS. Data the operative procedure progressively reduce due to shorter analysis and interpretation: FR, FA, AJH, and CF. Manuscript writing: operative times and the ability to perform procedures with all authors. Final approval of manuscript: all authors. a more limited range of surgical instruments and devices. This is especially true for new generations of surgeons, who Funding None. train directly in laparoscopy in other abdominal interven- tions (e.g., cholecystectomies, appendectomies, colectomies) Compliance with ethical standards and will also train directly in laparoscopic hepatectomies Conflict of interest F. Riquelme, C. Muñoz, F. Ausania, F. Torres, D. using standardized techniques honed by their predecessors. Calatayud, R. Sandomenico, R. García Péres, J. Ferrer, J. Fuster, and The present study does have limitations related to its J.C. García-Valdecasas have no conflicts of interest to disclose. A.J. retrospective nature and the non-randomized distribution Hessheimer and C. Fondevila are Principal Investigators on a Fondo de of patients. Inverse probability of treatment weighting is a Investigación Sanitaria Grant (PI1800894) from the Spanish Ministry of Health. propensity method that does not remove patients but gives them distinct weights based on over- or underrepresenta- Ethical approval This is stated in Methods (part on "Study Design"): tion of certain influential characteristics. Following IPTW Approval for the study was obtained from our Institution’s Committee adjustments, the results of this study continue to support on Ethics in Medical Research prior to initiating data analysis (study number HCB/2018/1179), which waived the need for written consent. 13 1050 Updates in Surgery (2020) 72:1041–1051 Informed consent Given the nature of the study, informed consent was left hepatic vein and to the common trunk. J Am Coll Surg waived by the Institutional Review Board. 195(5):737–739 13. Charlson M, Szatrowski TP, Peterson J, Gold J (1994) Vali- dation of a combined comorbidity index. J Clin Epidemiol Open Access This article is licensed under a Creative Commons Attri- 47(11):1245–1251 bution 4.0 International License, which permits use, sharing, adapta- 14. Dindo D, Demartines N, Clavien PA (2004) Classification of sur- tion, distribution and reproduction in any medium or format, as long gical complications: a new proposal with evaluation in a cohort of as you give appropriate credit to the original author(s) and the source, 6336 patients and results of a survey. Ann Surg 240(2):205–213 provide a link to the Creative Commons licence, and indicate if changes 15. Slankamenac K, Graf R, Barkun J, Puhan MA, Clavien PA (2013) were made. The images or other third party material in this article are The comprehensive complication index: a novel continuous scale included in the article’s Creative Commons licence, unless indicated to measure surgical morbidity. Ann Surg 258(1):1–7 otherwise in a credit line to the material. If material is not included in 16. Koch M, Garden OJ, Padbury R, Rahbari NN, Adam R, Capussotti the article’s Creative Commons licence and your intended use is not L et al (2011) Bile leakage after hepatobiliary and pancreatic sur- permitted by statutory regulation or exceeds the permitted use, you will gery: a definition and grading of severity by the International need to obtain permission directly from the copyright holder. To view a Study Group of Liver Surgery. Surgery 149(5):680–688 copy of this licence, visit http://creat​iveco​mmons​.org/licen​ses/by/4.0/. 17. Rahbari NN, Garden OJ, Padbury R, Brooke-Smith M, Crawford M, Adam R et al (2011) Posthepatectomy liver failure: a definition and grading by the International Study Group of Liver Surgery (ISGLS). Surgery 149(5):713–724 18. Ishizawa T, Hasegawa K, Kokudo N, Sano K, Imamura H, Beck References Y et al (2009) Risk factors and management of ascites after liver resection to treat hepatocellular carcinoma. Arch Surg 1. Bhojani FD, Fox A, Pitzul K, Gallinger S, Wei A, Moulton CA 144(1):46–51 et al (2012) Clinical and economic comparison of laparoscopic 19. Instituto Nacional de Estadística (Spain) (2016) Índices de Precios to open liver resections using a 2-to-1 matched pair analysis: an de Consumo. Base. Available from: https​://www.ine.es/jaxiT​3/ institutional experience. J Am Coll Surg 214(2):184–195 Tabla​.htm?t=22350​&L=0. 2. Ciria R, Cherqui D, Geller DA, Briceno J, Wakabayashi G 20. Pezzi CM, Mallin K, Mendez AS, Greer GE, Putnam JB Jr (2014) (2016) Comparative short-term benefits of laparoscopic liver Ninety-day mortality after resection for lung cancer is nearly dou- resection: 9000 cases and climbing. Ann Surg 263(4):761–777 ble 30-day mortality. J Thorac Cardiovasc Surg 148(5):2269–2277 3. Fretland AA, Dagenborg VJ, Bjornelv GMW, Kazaryan AM, 21. Swanson RS, Pezzi CM, Mallin K, Loomis AM, Winchester DP Kristiansen R, Fagerland MW et al (2018) Laparoscopic ver- (2014) The 90-day mortality after pancreatectomy for cancer is sus open resection for colorectal liver metastases: the OSLO- double the 30-day mortality: more than 20,000 resections from COMET randomized controlled trial. Ann Surg 267(2):199–207 the national cancer data base. Ann Surg Oncol 21(13):4059–4067 4. Macacari RL, Coelho FF, Bernardo WM, Kruger JAP, Jeismann 22. Talsma AK, Lingsma HF, Steyerberg EW, Wijnhoven BP, Van VB, Fonseca GM et al (2019) Laparoscopic vs. open left lateral Lanschot JJ (2014) The 30-day versus in-hospital and 90-day mor- sectionectomy: an update meta-analysis of randomized and non- tality after esophagectomy as indicators for quality of care. Ann randomized controlled trials. Int J Surg 61:1–10 Surg 260(2):267–273 5. Abu HM, Aldrighetti L, Dagher I, Edwin B, Troisi RI, Alikh- 23. Mise Y, Vauthey JN, Zimmitti G, Parker NH, Conrad C, Aloia anov R et al (2018) The southampton consensus guidelines for TA et al (2015) Ninety-day postoperative mortality is a legitimate laparoscopic liver surgery: from indication to implementation. measure of hepatopancreatobiliary surgical quality. Ann Surg Ann Surg 268(1):11–18 262(6):1071–1078 6. van der Poel MJ, Besselink MG, Cipriani F, Armstrong T, 24. Schiergens TS, Dorsch M, Mittermeier L, Brand K, Kuchenhoff Takhar AS, Van DS et al (2016) Outcome and learning curve in H, Lee SM et al (2015) Thirty-day mortality leads to underestima- 159 consecutive patients undergoing total laparoscopic hemi- tion of postoperative death after liver resection: a novel method to hepatectomy. JAMA Surg 151(10):923–928 define the acute postoperative period. Surgery 158(6):1530–1537 7. Abu HM, Di FF, Syed S, Wiltshire R, Dimovska E, Turner D 25. Vonlanthen R, Slankamenac K, Breitenstein S, Puhan MA, Mul- et al (2013) Assessment of the financial implications for laparo- ler MK, Hahnloser D et al (2011) The impact of complications scopic liver surgery: a single-centre UK cost analysis for minor on costs of major surgical procedures: a cost analysis of 1200 and major hepatectomy. Surg Endosc 27(7):2542–2550 patients. Ann Surg 254(6):907–913 8. Medbery RL, Chadid TS, Sweeney JF, Knechtle SJ, Kooby DA, 26. Staiger RD, Cimino M, Javed A, Biondo S, Fondevila C, Perinel Maithel SK et al (2014) Laparoscopic vs open right hepatec- J et al (2018) The Comprehensive Complication Index (CCI(R)) tomy: a value-based analysis. J Am Coll Surg 218(5):929–939 is a novel cost assessment tool for surgical procedures. Ann Surg 9. Cipriani F, Ratti F, Cardella A, Catena M, Paganelli M, Aldri- 268(5):784–791 ghetti L (2019) Laparoscopic versus open major hepatectomy: 27. Cipriani F, Alzoubi M, Fuks D, Ratti F, Kawai T, Berardi G et al analysis of clinical outcomes and cost effectiveness in a high- (2020) Pure laparoscopic versus open hemihepatectomy: a critical volume center. J Gastrointest Surg 23(11):2163–2173 assessment and realistic expectations—a propensity score-based 10. Wabitsch S, Kastner A, Haber PK, Feldbrugge L, Winklmann analysis of right and left hemihepatectomies from nine European T, Werner S et al (2019) Laparoscopic versus open hemihe- tertiary referral centers. J Hepatobiliary Pancreat Sci 27(1):3–15 patectomy—a cost analysis after propensity score matching. 28. Hessheimer AJ, de la Martinez ML, Adel Al SF, Espinoza AS, Langenbecks Arch Surg 404(4):469–475 Ausania F, Fondevila C (2019) Somatostatin and the “small-for- 11. Rossler F, Sapisochin G, Song G, Lin YH, Simpson MA, size” liver. Int J Mol Sci 20(10):2512 Hasegawa K et al (2016) Defining benchmarks for major liver 29. Wong-Lun-Hing EM, van Dam RM, van Breukelen GJ, Tanis PJ, surgery: a multicenter analysis of 5202 living liver donors. Ann Ratti F, Van HR et al (2017) Randomized clinical trial of open Surg 264(3):492–500 versus laparoscopic left lateral hepatic sectionectomy within an 12. Majno PE, Mentha G, Morel P, Segalin A, Azoulay D, Ober- enhanced recovery after surgery programme (ORANGE II study). holzer J et al (2002) Arantius’ ligament approach to the Br J Surg 104(5):525–535 13 Updates in Surgery (2020) 72:1041–1051 1051 30. ClinicalTrials.gov. Bethesda (MD): National Library of Medicine 35. Torres F, Rios J, Saez-Penataro J, Pontes C (2017) Is propensity (US). 2011 Sep 28. Identifier NCT01441856, The ORANGE II score analysis a valid surrogate of randomization for the avoidance PLUS - Trial: Open Versus Laparoscopic Hemihepatectomy. of allocation bias? Semin Liver Dis 37(3):275–286 Available from: https:​ //clinic​ altri​ als.gov/ct2/show/NCT014​ 41856​ . 36. Moris D, Dimitroulis D, Vernadakis S, Papalampros A, Spartalis 31. ClinicalTrials.gov. Bethesda (MD): National Library of Medicine E, Petrou A et al (2017) Parenchymal-sparing hepatectomy as (US). 2017 Sep 1. Identifier NCT03270917, ORANGE SEG- the new doctrine in the treatment of liver-metastatic colorectal MENTS: Open Versus Laparoscopic Parenchymal Preserving disease: beyond oncological outcomes. Anticancer Res 37(1):9–14 Postero-Superior Liver Segment Resection. Available from: https​ 37. Halls MC, Cipriani F, Berardi G, Barkhatov L, Lainas P, Alzoubi ://clini​caltr​ials.gov/ct2/show/NCT03​27091​7. M et al (2018) Conversion for unfavorable intraoperative events 32. Yang T, Zhang J, Lu JH, Yang GS, Wu MC, Yu WF (2011) Risk results in significantly worse outcomes during laparoscopic liver factors influencing postoperative outcomes of major hepatic resec- resection: lessons learned from a multicenter review of 2861 tion of hepatocellular carcinoma for patients with underlying liver cases. Ann Surg 268(6):1051–1057 diseases. World J Surg 35(9):2073–2082 33. Nobili C, Marzano E, Oussoultzoglou E, Rosso E, Addeo P, Publisher’s Note Springer Nature remains neutral with regard to Bachellier P et al (2012) Multivariate analysis of risk factors jurisdictional claims in published maps and institutional affiliations. for pulmonary complications after hepatic resection. Ann Surg 255(3):540–550 34. Fuks D, Cauchy F, Fteriche S, Nomi T, Schwarz L, Dokmak S et al (2016) Laparoscopy decreases pulmonary complications in patients undergoing major liver resection: a propensity score analysis. Ann Surg 263(2):353–361 13