Laparoscopic versus open distal pancreatectomy: is a randomized trial necessary?

Laparoscopic approaches exist for most operations traditionally performed using open techniques, and an increasing number of organ resections are being performed laparoscopically. The potential benefits of laparoscopic over the open approach include reduced operative blood loss, fewer surgical complications, shorter hospital stays, and faster recovery times. Cholecystectomy and adrenalectomy are two such examples of organ resections more commonly performed laparoscopically than open in the United States, as the complexity of these operations is relatively low for experienced surgeons in properly selected patients. Laparoscopic colectomy had a slower evolution given the increased complexity of the procedure, but the value of the laparoscopic approach was demonstrated through several multicenter randomized trials 1, 2, facilitated by the relatively large number of colectomies performed annually.

Momentum for minimally invasive pancreatectomy (MIP) has been slower – likely due to the retroperitoneal location of the organ, its intimate relationship with the main mesenteric vessels, the complexity of the surgery, the inherent risk of postoperative pancreatic fistula (POPF) formation, and the relatively lower volume of cases, as compared with cholecystectomy and colectomy 3. Of pancreatic resections, distal pancreatectomy (DP) is the most commonly achieved laparoscopically 4. First reported in 1994, laparoscopic distal pancreatectomy (LDP) is being performed with increasing frequency worldwide despite a lack of randomized data supporting this approach over standard open distal pancreatectomy (ODP). Approximately one quarter of all DPs are performed using minimally invasive approaches 5.

Drs Nakamura and colleagues from 69 medical centers in Japan used a sophisticated statistical method (propensity score matching – PSM) to address the comparison of two surgical techniques (LDP and ODP) for removing the left side of the pancreas 6. The key findings between the two cohorts after PSM are that compared with ODP, LDP is associated with lower blood loss (254 ± 384 vs. 499 ± 740, <0.001), fewer blood transfusion (4% vs. 6.8%, P = 0.02), fewer grade B and C POPF (18.4% vs. 28.2%, P < 0.001), a higher percentage of splenic preservations (29.9% vs. 13.2%, P < 0.001), and shorter hospital stays (18.8 ± 14.7 vs. 23.2 ± 18.8, P = 0.001). In great part, these results support data from prior studies.

In 2008 the Central Pancreas Consortium published the first large comparative trial of LDP and ODP 7. This study of 667 patients who underwent distal pancreatectomy between 2002 and 2006 included 159 laparoscopic resections with mixed pathologies from eight medical centers. Twenty (13%) laparoscopic procedures were converted to open. Cohorts were matched by age, American Society of Anesthesiologists score, tumor size, length of resected pancreas, and pathology for open (n = 200) or laparoscopic (n = 142) resection. Results of this study demonstrated that there were no significant differences in OR time (216 vs. 230 min; P = 0.3), development of major pancreatic fistula (18% vs. 11%; P = 0.1), major complication (17% vs. 10%; P = 0.08), or 30-day mortality (1% vs. 0%; P = 0.040) between the groups (ODP vs. LDP, respectively). As compared with ODP, LDP was associated with lower estimated blood loss (357 mL vs. 588 mL; P < 0.01), a higher rate of splenic preservation (30% vs. 12%; P < 0.001), fewer wound infections (5% vs. 15%; P = 0.004), a reduced need for postoperative percutaneous drain placement (6% vs. 15%; P = 0.02), and shorter hospital stays (5.9 vs. 9.0 days; P < 0.01). From this large, multicenter matched cohort study, it became clear that LDP was not only feasible, but could also offer additional benefits as compared to the open approach. Since 2008, these findings have been corroborated by numerous single center 8-12 and multicenter reports 13, 14, as well as those of several meta-analyses 4, 15.

Although the findings of Nakamura et al. are not entirely novel, they contribute to our knowledge by adding sheer volume, as this is the largest existing report comparing the two techniques, with over 700 patients in each arm of the PSM analysis. Use of PSM attempts to recreate a randomized trial using observational data. In this particular case, it should be noted that the results of the PSM patients are quite similar to the results of the entire cohort. The authors are to be congratulated for tremendous collaborative effort required to complete this work. One other multicenter analysis from the Dutch Pancreatic Cancer Group used PSM to compare LDP and ODP cohorts 16. They found that major complications and hospital length of stay did not differ significantly between the approaches in the matched cohorts. The report only includes 64 LDP cases, of which a third were converted to open, and concludes that surgical education is necessary to grow the experience with this approach in the Netherlands.

In further analyzing the design and results of this latest study comparing LDP and ODP from Japan, a few points should be noted. The authors excluded malignant diagnoses from their analysis. Often, these are the more challenging resections, which push the limits of technical ability and decision-making 14. Another important point is that the average patient body mass index (BMI) in this report was 22.3, while the average BMI in the United States is 28.6, and existing reports show us that BMI is tied to postoperative complications following pancreatectomy 17-19. While this does not affect the outcomes of the current study, it raises a question of how translatable the results are for surgeons and patients outside of Japan. The higher percentage of grade B/C POPF and lower percentage of splenic preservation in the ODP group is noted but cannot be explained by the methods section of the paper and may perhaps be a surgeon-specific or technique-related issues. Debate also continues regarding the true relevance of splenic preservation clinically, and as a quality measure. Finally, the shorter hospital stay in the Nakamura LDP group (18.8 ± 14.7 days) remains double that of open cases in the United States (9 ± 7.0 days), which is a cultural difference making it difficult to compare reports across nations 7. Perhaps such a metric should be reported in relative rather than absolute terms.

In summary, the report from Nakamura et al. furthers our existing knowledge of the comparison of LDP and ODP. PSM is another method of matching to correct for inherent biases in observational data and yet comparative assessments of observation data using PSM still fall short of randomized trials. Despite the current accumulated knowledge on LDP and minimally invasive pancreatectomy, several intriguing questions remain: (i) Do we need to randomize patients to ODP and LDP at this point, or should we expend our energies on improving education and patient selection moving forward? (ii) Does any given surgeon deliver better, worse, or equivalent care when performing distal pancreatectomy laparoscopically vs. open? (iii) Is pancreaticoduodenectomy an operation that should be performed using minimally invasive techniques?

As an international community we need to begin answering these questions. On 20 April 2016, the first international meeting of experts will come together to discuss the current state of understanding of minimally invasive pancreatic resection in Sao Paulo, Brazil. This conference is being sponsored through the joint efforts of the IHPBA and the regional HPB societies (AHPBA, E-AHPBA, A-PHPBA) and will focus on questions such as those highlighted by the work of Nakamura et al. The goal of this conference is to clear a path towards improved patient care and improved outcomes following major pancreatic resection.

Conflict of interest

None declared.