Sven Van Dijk

Acute pancreatitis is one of the most common gastrointestinal conditions requiring acute hospitalisation and has a rising incidence. In the most severe form of pancreatitis, patients will develop necrosis of the pancreatic and peripancreatic tissue. In a majority of patients where infection of the necrotic collections occurs, intervention is necessary. Also, as a result of necrosis of the pancreatic tissue, the pancreatic duct may lose its integrity, which may lead to leakage of pancreatic fluids, resulting in complications like recurrent fluid collections or pancreatic fistula. The aim of the studies in this thesis is to optimize the treatment in patients with complicated acute pancreatitis.

Over the last decade, treatment of acute pancreatitis has gradually developed towards a tailored, multidisciplinary effort, with distinctive roles for gastroenterologists, radiologists and surgeons. The review in chapter 2 summarizes how to diagnose, classify and manage patients with acute pancreatitis, emphasizing the evidence obtained through randomized controlled trials.

Infected necrotizing pancreatitis is treated with the step-up approach, consisting of a primary catheter drainage, followed by a (minimally invasive) necrosectomy, in whom drainage procedures alone do not suffice. In the Netherland, video-assisted retroperitoneal debridement (VARD) is the preferred minimally invasive surgical treatment if a necrosectomy is required. Chapter 3 presents a cohort of 108 patients that underwent a VARD for infected necrotizing pancreatitis. The in-hospital mortality of these patients was 15.7% (17 patients). When comparing the period before and after 2009, there was a significant increase of the number of pre-necrosectomy drainages procedures (from median 1 [IQR 1-2] to 3 [IQR 1-3], p<0.001) and VARD was increasingly postponed (from 45.9 to 65.3 days after onset of disease, p<0.001). The majority of patients (69%) required one VARD as final treatment. The VARD is an efficient procedure for a minimally invasive necrosectomy, when drainage procedures alone do not suffice. The optimal number of pre-operative drainages and timing of necrosectomy remain unclear. Necrotizing pancreatitis may lead to loss of integrity of the pancreatic duct, resulting in leakage of pancreatic fluid. This so-called pancreatic duct disruption or disconnection is associated with a prolonged disease course and particular complications. Since a standard treatment for this condition is currently lacking, chapter 4 presents a systematic review of the literature to compare outcomes of various treatment strategies. Twenty-one observational cohort studies were included comprising a total of 583 relevant patients. Endoscopic transpapillary drainage, endoscopic transluminal drainage, surgical drainage or resection, or combined procedures were the most frequently used treatment strategies. Pooled analysis showed success rates of 81% (95%-CI: 60-92%) for transpapillary and 92% (95%-CI: 77-98%) for transluminal drainage, 80% (95%-CI: 67-89%) for distal pancreatectomy and 84% (95%-CI: 73-91%) for cyst-jejunostomy. However, the success rates of conservative treatment are unknown. Also the diagnosis of disrupted or disconnected pancreatic duct in acute necrotizing pancreatitis is not standardized. In chapter 5 we provided an overview of the available literature of the accuracy of the different diagnostic modalities on this condition. We included 8 studies, evaluating five different diagnostic modalities: endoscopic ultrasound (EUS), endoscopic retrograde cholangiopancreatography (ERCP) and magnetic resonance cholangiopancreatography (MRCP), with or without secretin. This review suggests that various diagnostic modalities are accurate in diagnosing a disrupted pancreatic duct in patients with acute pancreatitis (sensitivity EUS and ERCP: 100%, MRCP without secretin: 83%. A combined cohort of MCRP with and without secretin: 92%). Amylase-measurement in drain fluid should be standardized, as it is a non-invasive, cheap and accurate measurement to detect pancreatic fluid leakage. Given the invasive nature of other modalities, secretin-magnetic resonance cholangiopancreatography or magnetic resonance cholangiopancreatography would be recommended as first diagnostic modality. Further prospective studies on this subject, however, are needed. Chapter 6 presents the rationale and design of the randomized controlled multicenter POINTER trial. In the standard treatment of infected pancreatic necrosis, the step-up approach, the primary catheter drainage is preferably postponed until the stage of walled-off necrosis, a process which usually will take around 4 weeks. This delay stems from the time when open necrosectomy was the standard. It is unclear whether such delay is needed for catheter drainage or whether earlier intervention could actually be beneficial in the current step-up approach. The POINTER trial investigates if immediate catheter drainage in patients with infected necrotizing pancreatitis is superior to the current practice of postponed intervention. In total, 104 adult patients with (suspected) infected necrotizing pancreatitis will be randomized between immediate (within 24 hours) catheter drainage and current standard care of antibiotic treatment and postponed catheter drainage. If a necrosectomy is required, this is preferably postponed until the stage of walled-off necrosis in both treatment arms. Primary outcome is the Comprehensive Complication Index (chapter 7), which is a composite score between 0-100 combining all postoperative complications during follow-up for individual patients. Secondary outcomes include mortality, complications, number of (re-) interventions, hospital and Intensive Care Unit (ICU) lengths of stay, quality adjusted life years (QALYs) and (in)direct costs. Standard follow-up is at 3 and 6 months after randomization. The abovementioned Comprehensive Complication Index (CCI) has shown to be more sensitive than traditional endpoints which may reduce the required sample size in randomized trials of elective surgery. In Chapter 7 the value of the CCI in a highly complicated disease as infected necrotizing pancreatitis was assessed. The CCI was calculated for all 88 patients included in the PANTER trial, comparing a minimally-invasive step-up approach with primary open necrosectomy in the treatment of infected necrotizing pancreatitis. The differences in the original composite endpoint (major complications and death) between study groups were compared with the difference in CCI. The mean CCI was 63.8 (SD 30.5) in the minimally invasive step-up approach, and 72.6 (SD 26.7) in the primary open necrosectomy group. Mean difference of 8.8 points was statistically not significant (p=0.137), where the original primary composited endpoint showed statistical significant differences (69% vs 40%, p=0.006). There were 19/88 patients that would actually have a CCI score above 100, if the formula would not have a cut-off point. This high mean scores in CCI reflect the severity and extent of complications in infected necrotizing pancreatitis, and may therefore be useful for research in this severely ill patients. However, in diseases with many complications the CCI may not be more discriminating than traditional (composite) endpoints. Since the logarithmic scale of the CCI and due to the high mean scores in this logarithmic scale, new complications will lead to diminishing yield in scores. Topics for clinical research are typically determined by clinicians, researchers or pharmaceutical companies, but seldom by patients. In chapter 8 we aimed to assess patients' dominant clinical symptoms and their opinions regarding the most relevant clinical research topics in pancreatitis. A questionnaire was distributed to 925 members of the Dutch Patient Organization for Pancreatic Diseases. Two-hundred-fifty-two patients with acute or chronic pancreatitis responded. The most reported clinical symptoms were pain by 184 (76%) patients and fatigue by 131(54%) patients. Diet during hospital admission was more often seen as major problem by patients with acute pancreatitis than those with chronic pancreatitis; 43% vs. 23%, p=0.007. Inability to work was more often considered as a major problem by chronic pancreatitis patients; 15% vs. 27%, p=0.007. Sixty-eight(29%) patients participated in clinical research which was valued as a positive experience by 68%(45/68). In total, 67% of all patients were willing to participate in future clinical research. Nutrition was most often suggested (54%) as a subject for future research. Patients can have valuable input in steering the direction of clinical research, and should be involved in making decisions about topics and funding for research projects in the future.