Effect of Previous Abdominal Surgery on Laparoscopic Liver Resection: Analysis of Feasibility and Risk Factors for Conversion

Introduction

Liver resection is frequently performed on patients known for previous abdominal surgery. Patients affected by various gastrointestinal malignancies often develop metastases to the liver, and hepatic resection has gained wide acceptance as a potentially curative treatment after primary surgery.^1–3 Also, repeated hepatectomies are increasingly adopted for the treatment of intrahepatic recurrence of liver tumors to extend long-term survival.^4–6

A history of abdominal operations has traditionally been considered a contraindication to the laparoscopic approach, with later surgical procedures being commonly addressed to open surgery. ⁷ In fact, possible dense intra-abdominal adhesions may be difficult to tackle, and inadvertent intraoperative accidents may result as fearful complications of a difficult adhesiolysis.^8,9 Specifically, severe adhesions at the hepatoduodenal ligament may increase the risk of injury to biliary and vascular structures at the hilar area. Nonetheless, the possibility to accomplish adequate mobilization, extrahepatic inflow control, and ultrasound assessment may be limited, thus impairing the feasibility, safety, and oncologic efficiency of the procedures. These dangerous events can be even further magnified by laparoscopy given the lack of manual palpation, and may ultimately threaten the possibility to complete the procedure by laparoscopy, or—more importantly—become a potential source of morbidity and prolonged hospital stay.

With acquisition of expertise, the indication to laparoscopic liver resections (LLR) has progressively been extended to procedures of growing complexity with reported favorable outcomes. As a result, several factors potentially precluding the laparoscopic approach have found to be negotiable with growing experience.^10–13 Yet, LLR in patients known for previous abdominal surgery have been increasingly reported so far. However, specific evidence about their feasibility and safety still lacks, as well as clear evidence on the role of previous abdominal surgery as a risk factor for conversion.^14–16

Herein, the effect of prior abdominal surgery on the feasibility and safety of later LLR was analyzed in a large single-center cohort, along with risk factors for conversion to the open technique.

Materials and Methods

Results

A total of 773 LLR were attempted during the study period (349 PS group; 424 NPS group). As a result of the matching process, 349 of 424 patients of the NPS group were matched with the 349 patients in PS group. After matching, the two groups resulted balanced and comparable on all the five covariates (Table 1). The majority of LLR were performed for malignant lesions in both groups (P = .766), with high prevalence of colorectal liver metastases. A breakdown of diagnosis and type of liver resection is reported in Table 1.

One-hundred eighty-two (n = 182) previous abdominal operations were performed by laparoscopy (52.1%), and 167 by the traditional open technique (47.8%). Previous upper abdominal operations (UPS subgroup, n = 161) accounted for 46.1% of the procedures and included cholecystectomy (n = 76, 21.8%), hepatic resection (n = 59, 16.9%), gastric resection (n = 12, 3.4%), and pancreaticoduodenectomy (n = 3, 0.8%). Baseline characteristics of UPS patients were similar to NPS patients (Table 1). Other abdominal procedures comprised colonic resection (n = 112, 32.1%), rectal resection (n = 66, 18.9%), nephrectomy (n = 6, 1.7%), and kidney transplantation (n = 1, 0.3%).

Discussion

A history of abdominal surgery has traditionally represented a matter of concern in case of later operations. Indeed, it entails some misgivings about the approach to be used (upfront open or possibly minimally invasive) and the need to deal with intra-abdominal adhesions.^7–9

The present study specifically addressed the feasibility of LLR in this setting, showing significantly higher conversion rates in patients with previous surgery compared with patients with virgin abdomen (13.7% and 5.1%, respectively, P = .021), particularly in case of previous upper abdominal surgery (23.0%, P = .015). Furthermore, at multivariable analysis, previous abdominal surgery emerged as a factor independently associated with the risk of conversion (hazard ratio 1.70 [1.31–1.98], P = .033). Altogether, these data suggest that a history of surgery may effectively influence the feasibility of later LLR, and that previous upper abdominal operations might increase the risk of conversion even further, as expected.

In these settings, the role played by hard-to-manage intra-abdominal adhesions on the possibility of conversion and the increased risk of intraoperative unfavorable events (injury to adjacent organs, hemorrhage, R1 resections) are clearly not negligible. Indeed, at analysis of reasons for conversion, the impossibility of a complete laparoscopic adhesiolysis before liver resection was found as the commonest in both PS and UPS group (5.7% and 11.8%). However, the incidence of conversions related to intraoperative accidents (injury of major vascular and biliary structures, adjacent organs, or laceration of liver parenchyma) was low and comparable between the two groups. Carefully interpreted, these findings suggest that dense intra-abdominal adhesions do not necessarily increase the risk of intraoperative accidents, rather the possibility to perform a successful laparoscopic adhesiolysis before the liver resection. Thus, it is sensible to recommend adoption of the laparoscopic approach for liver resections not to be rejected in view of previous surgery per se. This is even further supported by the evidence that adhesiolysis was not even necessary in a considerable proportion of patients with a history of surgery (32.2%). Moreover, it was possible to complete a significant number of attempted laparoscopic resections (86.3% in PS group, 77.0% in UPS group). A reasonable operative strategy for these patients would be to rely on careful approach to the abdominal cavity and ability to readapt the port location, and early precautionary conversion to be timely practiced when hard-to-manage adhesions are found particularly difficult to tackle so as to prevent inadvertent intraoperative accidents.

The risk of intraoperative hemorrhage is also felt as being potentially influenced by a previous surgery. In fact, the dissection of highly vascularized adhesions may increase the intraoperative blood loss. More importantly, dense adhesions in proximity of the hepatoduodenal ligament may threaten the possibility to prepare for a Pringle maneuver and to perform extrahepatic hilar dissection to control the vascular inflow during the transection, when needed. In our experience, the feasibility of an intermittent inflow control during LLR was not significantly affected by adhesions in proximity to the hepatic hilum. Indeed, data from this study show that intermittent clamping of the liver pedicle was equally performed in patients known for previous surgery and virgin abdomen (PS group: 81.1% versus NPS group: 84.8%; P = .792). Moreover, intraoperative blood loss was relatively low and similar between the two groups (250 mL in PS group versus 170 mL in NPS group; P = .270). Nevertheless, conversions for intraoperative bleeding occurred more frequently in the PS group compared with NPS group (3.4% versus 1.4%, respectively, P = .004). However, in all the converted cases, bleeding originated from a hepatic vein or portal branch encountered during transection, and no hemorrhage occurred as a direct consequence of adhesiolysis of highly vascularized adhesions.

The issue of oncologic adequacy of LLR has already been analyzed in several series documenting the effective possibility to obtain radical resections, as in open approach.^28–30 It is worth mentioning that obtaining a surgical margin free from tumor infiltration should always be pursued as the primary aim when performing hepatectomy for malignancy, as a surrogate outcome of the disease-free survival. In the setting of a surgerized abdomen, it should be presumed that intra-abdominal adhesions may threaten an adequate inspection and ultrasound evaluation for complete tumor assessment. Also, the mobilization before parenchymal transection may be demanding, or even impossible to be completed, as much as needed. However, in this series, the incidence of conversions related with inadequate radicality of the procedure was not different between the two groups (9 patients in PS group = 2.5% versus 7 patients in NPS group = 2.0%, P = .892) and unexpected R1 resection rate at a final histological examination resulted low (6 patients in PS group = 1.7% versus 5 patients in NPS group = 1.4%, P = .130). These results suggest that LLR for this subset of patients can be accomplished with oncological radicality comparable to the standard setting. However, the accuracy of intraoperative ultrasound and the opportunity to perform adequate mobilization should always be preserved. ³¹ In this, surgeon's judgment plays a determinant role, and conversion is warranted for the more demanding cases.

The safety and efficacy of LLR for patients known for previous surgery are confirmed by the perioperative outcomes herein reported. Postoperative complication rates were relatively low and comparable between the two groups (13.5% in PS group versus 13.2% in NPS group; P = .578). More importantly, these results were produced according to the intention-to-treat analysis of data, thus including also patients who experienced conversion. This allowed to keep unaltered the balance between treatment groups achieved after matching, and to show real outcomes achievable in clinical practice. The results herein reported suggest that the safety and efficacy of LLR still live even when laparotomy becomes necessary, and perioperative outcomes seem not to suffer from conversion to open when this helps to accomplish a safe procedure. This is of paramount importance in surgical practice, and suggests that conversion should not be felt as a surgeon's failure potentially affecting the intraoperative decision-making, rather as the inadequacy of the technique itself.

Previous abdominal surgery could even increase concerns about longer operative times secondary to the technical challenges of the procedures. ⁷ The present data showed comparable operative times with no significant increases in PS group (190 minutes in PS group and 205 minutes in NPS group, P = .840). This result may be partially explained considering the potentially favorable impact of the laparoscopic approach on adhesiolysis. Indeed, the abdominal distension sustained by the insufflation may stretch the adhesions, and provide the surgeon a work space for dissection that is sensibly wider than in open technique, ultimately simplifying the cut down of adhesions.

The results reported in this study suggest that a history of surgery does not preclude the laparoscopic approach for liver resections, which should not be seen as an absolute contraindication to minimally invasiveness but as a factor potentially increasing the risk of conversion. Undoubtedly, LLR in these settings are demanding procedures, as the need to deal with intra-abdominal adhesions increases the difficulty of the procedures. However, experience in laparoscopic and hepatobiliary surgery along with surgeon's ability in intraoperative decision-making may reduce the occurrence of inadvertent intraoperative accidents by operating early conversions in case of difficult adhesiolysis.

In conclusion, the minimally invasive approach to liver resection appears as an appealing alternative to the open technique even in case of previous abdominal surgery. It proved to be feasible, safe, and oncologically efficient, allowing for a rapid recovery of patients. Conversion to laparotomy should be considered promptly in case of hazardous adhesiolysis.