The role of SPN in severe necrotizing pancreatitis lacks broadly-accepted guidelines or clear indications for the various non-operative and operative procedures beyond an initial non-operative, “step-up” approach.(11) This study is the largest retrospective collection of SPN cases, using a cross-sectional database of patients in the US, Canada, the UK, Europe and Australia, with 2457 cases recorded between 2007 and 2019, in which we examined preoperative predictors of CD4 complications and mortality. We found a low and stable mortality rate for SPN, with a decreasing rate of ICU complications. The frequency of SPN in ACS-NSQIP decreased significantly after 2010, while the patients undergoing SPN had decreasing frailty scores throughout the study period. There may be a selection bias that SPNs may be occurring mostly at academic, tertiary referral centers which may have been faster to join ACS-NSQIP than later-joining centers; however, the rate of decrease in performance of SPN was significantly more rapid after 2010 than the decrease of pancreaticoduodenectomies during the study period, indicating that fewer SPNs are being performed in ACS-NSQIP contributor hospitals.
ACS-NSQIP does not allow tracking of preoperative procedures that may have been performed before the SPN, so compliance with the “step-up” approach cannot be exactly determined. Post-SPN, additional endoscopic or interventional radiology procedures for WOPN were uncommon. CPT codes for endoscopic procedures post-SPN were quite rare, possibly indicating that SPN was definitive management. Unfortunately, there is no available CPT code for endoscopic pancreatic necrosectomy. The 2010 PANTER trial randomized patients into either primary SPN or a step-up MIS approach to a possible SPN using percutaneous and/or endoscopic transgastric drainage, followed by a second drainage procedure if the initial one failed after 72 hours, and then VARD if there was still no improvement after a further 72 hours. In the step-up group, 26 of 43 (60%) underwent SPN after percutaneous drainage, 24 receiving VARD and 2 having open necrosectomies. Further procedures were required in 33% of step-up cases. In the initial necrosectomy group 42% required additional laparotomies for additional necrosectomies and 33% required additional percutaneous drainage procedures. While ACS-NSQIP cannot determine exactly who underwent a “step-up” approach, our study indicates that the number of SPNs decreased significantly after 2010, as did the number of SPN patients requiring a return trip to the operating room, which we believe is evidence of adoption of an MIS approach such as used in the PANTER algorithm. The decrease in complications after 2010 may be further evidence for a greater use of MIS approaches such as VARD versus open necrosectomy but may also represent a selection bias of managing frail patients via endoscopic and interventional radiologic procedures without SPN.
There was a significant decrease in Frailty Index scores after 2010, indicating that less frail patients with a lower risk of complications were being selected for SPN, which may also have contributed to fewer complications. The modified Frailty Index became less predictive of CD4 complications and mortality with lowered odds ratios compared to an earlier ACS-NSQIP study for the period of 2007 to 2012. (15) This may also be further evidence that less frail patients were being selected for SPN. Whether more frail patients were managed with interventional radiology placed drains or endoscopic debridement without SPN cannot be determined by this paper. We believe the PANTER trial may have influenced providers not just to consider non-operative procedures before surgical procedures, but also to avoid SPN in frail patients altogether.
The PANTER trial showed that the step-up approach did reduce complications but did not affect mortality. Similarly, our study showed that while complications did decrease during the study period, the death rate remained stable. This may indicate that the mortality of severe infected pancreatic necrosis may not be much affected by current SPN procedures due to causes such as multi-system organ failure, late presentation or medical comorbidities. Frailty and non-independent functional status were important predictors of mortality and should be significant adverse risk factors for surgeons considering offering SPN after failure of non-operative management of severe necrotizing pancreatitis. ACS-NSQIP rules indicate the emergency surgery data element should be assigned as a ‘YES' if the surgeon and/or anesthesiologist reported the case as emergent, “usually performed within a short interval of time between patient diagnosis or the onset of related preoperative symptomatology”, and “the patient’s well-being and outcome is potentially threatened by unnecessary delay and the patient’s status could deteriorate unpredictably or rapidly.” The increased mortality and complications for emergency performance of SPN should also be considered an adverse risk factor by surgeons. If non-operative alternatives to urgent SPN are an option, these may in fact be safer. However, we cannot tell from ACS-NSQIP whether suspicion of complications of severe acute pancreatitis such as perforated viscus, hemorrhage, bowel ischemia or uncontrolled sepsis precipitated a decision to perform urgent SPN.(18)
There are several limitations to our study. We cannot tell the exact SPN procedure that was performed, nor the order of procedures including whether the step-up approach was used for any particular patient. We believe the decrease in SPNs is indirect evidence for the acceptance of the PANTER trial results after 2010, and not due to other possible issues with the ACS NSQIP database in capturing CPTs for SPN; however, there may have been other trends in adopting more MIS and non-operative approaches not related to the PANTER trial. We do not have information on the timing of the procedures in relation to the time of onset of severe necrotizing pancreatitis. We also do not know if there was evidence of infected necrotizing pancreatitis or what the abdominal anatomy of the infected collections might have been. CT scan results are not available within the database. A number of adverse outcomes are not included in ACS-NSQIP such as pancreatic fistulae, delayed abdominal closure, incisional hernias, new onset diabetes, need for enteral and parenteral nutrition, need for enzymatic supplementation and prolonged ICU stay. Although we saw a decrease in necrosectomies, the ACS-NSQIP data is submitted only from hospitals and countries that are participating in the ACS-NSQIP and may not represent a statistically valid, nationally representative sample, although the sample is quite large.
This updated analysis of ACS-NSQIP since the study by Kolbe et al for the 2007–2012 period does indicate that there has been a significant decrease in the performance of SPNs(15). SPN may not be obsolete as there is ability to select patients for a step-up approach based on their frailty, functional status, and laboratory results to achieve reduced complications. Given the relative rarity of SPN procedures, we recommend that SPN be included in prospectively collected registries such as current and planned EGS registries, including imaging, anatomic and procedural variables, to obtain more predictive outcome data.