Superiority of laparoscopic liver resection to open liver resection in obese individuals with hepatocellular carcinoma: A retrospective study

Abstract Aim This study aimed to elucidate the effects of laparoscopic liver resection (LLR) vs open liver resection (OLR) for major complications (Clavien‐Dindo classification grade ≥ IIIa) in obese individuals with hepatocellular carcinoma (HCC). Methods The clinical records of 339 and 733 patients who underwent LLR and OLR, respectively, for HCC between 2000 and 2019 were retrospectively reviewed. Body mass index (BMI) groups were classified according to the definitions of the World Health Organization: underweight group, BMI ≤ 18.4 kg/m2 (LLR vs OLR: 27 vs 47); normal weight, BMI 18.5‐24.9 kg/m2 (211 vs 483); overweight, BMI 25.0‐29.9 kg/m2 (85 vs 181); and obese, BMI ≥ 30.0 kg/m2 (16 vs 22). The effects of obesity on major complications after LLR and OLR were investigated. Results In total, 18 (5.3%) and 127 (17.3%) patients presented with major complications after LLR and OLR, respectively. There was no significant difference in the incidence of major complications after OLR in the four BMI groups. However, a stepwise decrease in the incidence of major complications after LLR was observed from the underweight to the obese group. In addition, a multivariate analysis revealed that increased BMI was an independent preventive factor for major complications after LLR (P = .026, odds ratio: 0.84). The estimated adjusted risk of major postoperative complications decreased with increased BMI in the LLR group, while the risk did not decrease in the OLR group (P for interaction = .048). Conclusion Laparoscopic liver resection is beneficial for obese patients and is superior to OLR.


| INTRODUC TI ON
The prevalence of obesity and its associated diseases is still increasing worldwide. The prevalence of obesity (body mass index [BMI] of ≥30 kg/m 2 ) is 40% in the United States 1 and approximately 20% in Europe. 2 In Japan, obesity is defined by a BMI of ≥25 kg/ m 2 . 3 As of 2018, 32.2% of men and 21.9% of women of ≥20 years of age were classified as obese. 4 Obesity is correlated with comorbidities and technical difficulties in surgery and is considered a risk factor for postoperative complications in several surgical fields. 5,6 Furthermore, several reports have shown that obese patients are at high risk of developing hepatocellular carcinoma (HCC). 7,8 Thus, a higher prevalence of obesity and expansion of the indications for liver resection could increase the number of liver resection procedures among obese patients with HCC in the future. Obesity is associated with an increased risk of postoperative morbidity in individuals undergoing open liver resection (OLR). 9,10 Recently, laparoscopic liver resection (LLR) has been widely performed and is correlated with low morbidity and mortality. 11,12 However, the superiority of LLR to OLR was not evaluated according to BMI. Thus, previous reports cannot support the efficacy and safety of LLR for obese individuals. 13,14 The current study investigated the effects of obesity on major complications (≥grade IIIa based on the Clavien-Dindo classification system 15 ) after LLR and OLR for HCC based on BMI (from underweight to obese). Moreover, the superiority of LLR to OLR in terms of major postoperative complications based on BMI was evaluated.

| Study design and participants
In total, 1072 consecutive patients with HCC who underwent liver resection in our department between January 2000 and December 2019 were included in this study. Patient's height and weight were assessed preoperatively, and BMI was calculated as weight in kilograms (kg) divided by height in meters squared (m 2 ). The patients were allocated to one of four groups based on BMI, as defined by the World Health Organization 16 : underweight group, BMI of ≤18.4 kg/ m 2 ; normal weight group, 18.5 ≤ BMI ≤ 24.9 kg/m 2 ; overweight group, 25.0 ≤ BMI ≤ 29.9 kg/m 2 ; and obesity group, BMI of ≥30.0 kg/ m 2 . The local institutional review board of our institution approved this study (registration no. 1646).

| Surgical procedure
In total, 923 and 149 patients who underwent their first and second hepatic procedure, respectively, were included in the analysis. LLR was performed on 339 patients (LLR group) and OLR on 733 patients (OLR group). As described in our previous study on OLR, 17 in most patients who underwent segmentectomy or more, after Glissonean sheath transection or clamping, an ultrasonic surgical aspirator was used for hepatic dissection during total or unilateral clamping of the hepatic vascular inflow. In the majority of patients who underwent partial hepatic resection, as resection of less than a segmentectomy, an ultrasonic surgical aspirator and bipolar or monopolar forceps was utilized for hepatic dissection with the Pringle maneuver. The major branches of the Glissonean sheath and the hepatic vein were sutured using non-absorbent sutures. Patients who underwent LLR were placed in supine or left-lateral decubitus position and on an average five trocars according to tumor location were used. Hepatic transection was performed using a laparoscopic ultrasonic surgical aspirator and a vessel sealing system with soft coagulation. 18 In general, the Pringle maneuver was applied. Hand-assisted laparoscopy, or the so-called hybrid procedure or laparoscopy-assisted resection, was performed on patients with tumors that are challenging to evaluate via pure laparoscopy due to limited visualization and heavy bleeding. In this study, LLR was defined as all laparoscopic surgeries. Further, the surgical procedures were classified into partial resection, segmentectomy, sectionectomy, and resection of two or more sections according to the Brisbane 2000 Terminology of Liver Anatomy and Resections. 19

| Indication for laparoscopic liver resection
We performed LLR for ≤5-cm solitary lesions located in the peripheral liver segments 2-6 according to the Louisville consensus. 20 Thereafter, we extended the indication for more difficult procedures including major hepatectomy. However, LLR was selected according to tumor location, types of operative procedures, tumor size, proximity to major vessels, and liver function.

| Clinicopathological characteristics and surgical outcomes
The clinical data of all patients were collected prospectively, as shown in Table 1. The date of follow-up was on March 31, 2020.

| Definitions
Patients were diagnosed with diabetes mellitus (DM), hypertension, and dyslipidemia according to the guidelines of the Japan Diabetes Society, 21 Japanese Society of Hypertension, 22

| Outcomes
The primary outcome was the effect of BMI on major complications after LLR and OLR. Patients who underwent LLR and OLR initially presented with risk factors for major complications. Finally, to evaluate the superiority of LLR to OLR in patients with increased BMI, the difference in major complications between patients who underwent LLR and OLR was analyzed via a multivariate analysis.

| Statistical analyses
Categorical variables were presented as numbers and percentages and were compared between groups using the Fisher's exact test or the χ 2 test, as appropriate. Continuous variables were expressed as median (range) and were compared using the Kruskal-Wallis test. .028 .062 Volume of blood loss, mL; median (range) .001 .002 Non-curative surgery .004 .104 Number

| Characteristics of patients who underwent laparoscopic liver resection and open liver resection
There was no significant difference in terms of median BMI among patients who underwent LLR and OLR ( in the LLR group than in the OLR group. However, the distribution of the Child-Pugh scores among the two groups did not significantly differ. The proportion of patients who underwent partial liver resection was higher in the LLR group than in the OLR group, who had a shorter operative time and lower volume of blood loss. Based on tumor-related factors or pathology, the OLR group had a more advanced HCC than the LLR group. The rate of incidence of overall and major postoperative complications was lower in the LLR group than in the OLR group, who had a shorter hospital stay.

| Clinical characteristics of patients who underwent LLR and OLR according to BMI status
The clinicopathological characteristics of patients (n = 1072) after LLR or OLR according to BMI are shown in .828 Wound infection

TA B L E 4 Univariate and multivariate analyses of risk factors associated with major complications after laparoscopic and open liver resection
ALT. However, the distribution of Child-Pugh scores did not remarkably differ among the four groups. Although there was no difference in the type of liver resection among the four BMI groups, the proportion of patients who experienced bleeding indicated a stepwise increase from the underweight to obesity groups. In each BMI group, the proportion of patients who had intraoperative bleeding was higher in the OLR group than in the LLR group (P = .011 in the underweight group and P < .001 in the normal weight, overweight, and obesity groups).

| Major complications after LLR and OLR according to BMI status
The incidence of major complications after LLR indicated a stepwise decrease from the underweight to obese groups (Cochran-Armitage trend test, P = .037, Figure 1A). A multiple analysis using Holm's test indicated that the incidence in the underweight group was higher than that in the normal weight group; however, there were no significant differences among the normal weight, overweight, and obese groups. The incidence of wound infection and intra-abdominal infection did not differ among the four groups, and none of the patients presented with complex venous thromboembolism (Table 3). There was no significant difference in the incidence of each complication among the four BMI groups with zero in-hospital death. In patients who underwent OLR, there was no difference in the incidence of overall or major complications among the four groups ( Figure 1A).

The incidence of wound infection and abdominal infection did not
differ among the four groups, and only one (0.6%) patient in the overweight group had complex venous thromboembolism. One patient in the underweight group and three in the normal weight group died of liver failure. In the normal weight, overweight, and obesity groups, the incidence of major complications was higher in patients who underwent OLR than in those who underwent LLR (underweight group, P = .871; normal weight group, P < .001; overweight group, P = .006; and obesity group, P < .001).

| Risk factors for major complications after LLR and OLR
In all the cohorts including the LLR and the OLR groups, univariate and multivariate logistic regression analyses showed that age, operative time, volume of blood loss, serum albumin concentration, and LLR were independent predictive factors for major postoperative complications (Table S1). Meanwhile, BMI was not associated with major postoperative complications in all the cohorts including the LLR and OLR groups. However, in the LLR group, an increased BMI was considered an independent favorable factor and long operative time was a risk factor for postoperative complications (

TA B L E 4 (Continued)
After adjusting for age, sex, liver function test, and surgery-and tumor-related factors between the LLR and OLR groups, the risk of major postoperative complications in the LLR group decreased with increased BMI, while that in the OLR group increased with greater BMI ( Figure 1B). In patients with a BMI of 25 and 30, but not 18.5, kg/m 2 , the risk was lower in the LLR group than in the OLR group.
In addition, there was significant effect of interaction between patients who underwent LLR and those who underwent OLR according to increase in BMI value (P for interaction = .048, Figure 1B).  33 revealed that underweight patients who underwent LLR had a higher complication rate than normal body weight patients, which corresponded with our results. These phenomena are referred to as the obesity paradox, and its cause has not been fully elucidated. However, underweight-induced hypoalbuminemia, sarcopenia, frailty, malnutrition, and a poor immune function may be correlated with poor outcomes after surgery. [48][49][50][51] In this study, the albumin level was basically good because the liver function reserve (e.g. the Child-Pugh classification) was assessed before hepatectomy. In fact, there was no difference in albumin level among the BMI groups in patients who underwent LLR. In addition, we did not evaluate sarcopenia or frailty; thus, we could not elucidate the cause of the higher proportion of major complications in the underweight group after LLR in our own study. However, we could suggest that sufficient operative space could be obtained during LLR, even in overweight and obese patients, and that this was associated with surgical safety and a decreased incidence of major complications.

| D ISCUSS I ON
The current study had several limitations. That is, it has a retrospective design, and the number of patients was large. Moreover, the research was conducted at a single institution. A BMI of >40 kg/m 2 was considered a risk factor for in-hospital death. 52  Nevertheless, the results of this study should be validated by performing multicenter and international studies.

| CON CLUS ION
Laparoscopic liver resection is superior to OLR in overweight and obese patients in regard to decrease in incidences of postoperative major complications.