Root-cause Analysis of Mortality After Pancreatic Resection (CARE Study): A Multicenter Cohort Study.

OBJECTIVES: Analyze a multicenter cohort of deceased patients after pancreatectomy in high-volume centers in France by performing a root-cause analysis (RCA) to define the avoidable mortality rate. BACKGROUND: Despite undeniable progress in pancreatic surgery for over a century, postoperative outcome remains particularly worse and could be further improved. METHODS: All patients undergoing pancreatectomy between January 2015 and December 2018 and died postoperatively within 90 days after were included. RCA was performed in 2 stages: the first being the exhaustive collection of data concerning each patient from preoperative to death and the second being blind analysis of files by an independent expert committee. A typical root cause of death was defined with the identification of avoidable death. RESULTS: Among the 3195 patients operated on in 9 participating centers, 140 (4.4%) died within 90 days after surgery. After the exclusion of 39 patients, 101 patients were analyzed. The cause of death was identified in 90% of cases. After RCA, mortality was preventable in 30% of cases, mostly consequently to a preoperative assessment (disease evaluation) or a deficient postoperative management (notably pancreatic fistula and hemorrhage). An inappropriate intraoperative decision was incriminated in 10% of cases. The comparative analysis showed that young age and arterial resection, especially unplanned, were often associated with avoidable mortality. CONCLUSIONS: One-third of postoperative mortality after pancreatectomy seems to be avoidable, even if the surgery is performed in high volume centers. These data suggest that improving postoperative pancreatectomy outcome requires a multidisciplinary, rigorous, and personalized management.

Mesh Repair in Crohn's Disease: A Case-Matched Multicenter Study in 234 Patients.

BACKGROUND: There are no specific guidelines for ventral hernia management in Crohn's disease (CD) patients. We aimed to assess the risk of septic morbidity after mesh repair in CD. METHODS: This was a retrospective multicentre study comparing CD and non-CD patients undergoing mesh repair for ventral hernia (primary or incisional hernia). Controls were matched 1:1 for the presence of a stoma, history of surgical sepsis, hernia size and Ventral Hernia Working Group (VHWG) score. All demographic, pre-, intra- and postoperative data were retrieved, including long-term data. RESULTS: We included 234 patients, with 114 CD patients. Both groups had comparable VHWG scores (p = 0.12), hernia sizes (p = 0.11), ASA scores ≥ 3 (p = 0.70), body mass index values (p = 0.14), presence of stoma (CD 21.9% vs. controls 15%, p = 0.16), history of sepsis (14% vs. 6.7%, p = 0.23), rates of malnutrition (4.4% vs. 1.7%, p = 0.46), rates of incisional hernia (93% vs. 95%, p = 0.68) and concomitant procedures (18.4% vs. 11.7%, p = 0.12). CD patients carried a higher risk of postoperative septic morbidity (18.4% vs. 5%, p = 0.001), entero-prosthetic fistula (7% vs. 0, p < 0.01) and mesh withdrawals (5.3% vs. 0, p = 0.011). Ventral hernia recurrence rates were similar (14% vs. 8.3%, p = 0.15). In the univariate analysis, the risk factors for septic morbidity were CD (p = 0.001), malnutrition (p = 0.004), use of biological mesh (p < 0.0001) and concomitant procedure (p = 0.004). The mesh position, the means used for mesh fixation as well as the presence of a stoma were not identified as risk factors. CONCLUSIONS: CD seems to be a risk factor for septic morbidity after mesh repair.

Saccharomyces cerevisiae prevents postoperative recurrence of Crohn's disease modeled by ileocecal resection in HLA-B27 transgenic rats.

BACKGROUND: Postoperative recurrence (POR) after ileocecal resection (ICR) affects most Crohn's disease patients within 3-5 years after surgery. Adherent-invasive Escherichia coli (AIEC) typified by the LF82 strain are pathobionts that are frequently detected in POR of Crohn's disease and have a potential role in the early stages of the disease pathogenesis. Saccharomyces cerevisiae CNCM I-3856 is a probiotic yeast reported to inhibit AIEC adhesion to intestinal epithelial cells and to favor their elimination from the gut. AIM: To evaluate the efficacy of CNCM I-3856 in preventing POR induced by LF82 in an HLA-B27 transgenic (TgB27) rat model. METHODS: Sixty-four rats [strain F344, 38 TgB27, 26 control non-Tg (nTg)] underwent an ICR at the 12th wk (W12) of life and were sacrificed at the 18th wk (W18) of life. TgB27 rats were challenged daily with oral administration of LF82 (109 colony forming units (CFUs)/day (d), n = 8), PBS (n = 5), CNCM I-3856 (109 CFUs/d, n = 7) or a combination of LF82 and CNCM I-3856 (n = 18). nTg rats receiving LF82 (n = 5), PBS (n = 5), CNCM I-3856 (n = 7) or CNCM I-3856 and LF82 (n = 9) under the same conditions were used as controls. POR was analyzed using macroscopic (from 0 to 4) and histologic (from 0 to 6) scores. Luminal LF82 quantifications were performed weekly for each animal. Adherent LF82 and inflammatory/regulatory cytokines were quantified in biopsies at W12 and W18. Data are expressed as the median with the interquartile range. RESULTS: nTg animals did not develop POR. A total of 7/8 (87%) of the TgB27 rats receiving LF82 alone had POR (macroscopic score ≥ 2), which was significantly prevented by CNCM I-3856 administration [6/18 (33%) TgB27 rats, P = 0.01]. Macroscopic lesions were located 2 cm above the anastomosis in the TgB27 rats receiving LF82 alone and consisted of ulcerations with a score of 3.5 (2 - 4). Seven out of 18 TgB27 rats (39%) receiving CNCM I-3856 and LF82 had no macroscopic lesions. Compared to untreated TgB27 animals receiving LF82 alone, coadministration of CNCM I-3856 and LF82 significantly reduced the macroscopic [3.5 (2 - 4) vs 1 (0 - 3), P = 0.002] and histological lesions by more than 50% [4.5 (3.3 - 5.8) vs 2 (1.3 - 3), P = 0.003]. The levels of adherent LF82 were correlated with anastomotic macroscopic scores in TgB27 rats (r = 0.49, P = 0.006), with a higher risk of POR in animals having high levels of luminal LF82 (71.4% vs 25%, P = 0.02). Administration of CNCM I-3856 significantly reduced the levels of luminal and adherent LF82, increased the production of interleukin (IL)-10 and decreased the production of IL-23 and IL-17 in TgB27 rats. CONCLUSION: In a reliable model of POR induced by LF82 in TgB27 rats, CNCM I-3856 prevents macroscopic POR by decreasing LF82 infection and gut inflammation.

The Expression of the Short Isoform of Thymic Stromal Lymphopoietin in the Colon Is Regulated by the Nuclear Receptor Peroxisome Proliferator Activated Receptor-Gamma and Is Impaired during Ulcerative Colitis.

The etiology of inflammatory bowel diseases remains largely unknown. We previously demonstrated that the expression of the peroxisome proliferator activated receptor-gamma (PPARγ) is downregulated in colonic epithelial cells of patients with ulcerative colitis (UC). PPARγ is a nuclear receptor that modulates inflammation. We hypothesized that its deficiency may play a role in the loss of intestinal homeostasis through the control of immunomodulatory factors. We found that thymic stromal lymphopoietin (TSLP), an epithelial cytokine with pleiotropic functions, is regulated by PPARγ. While this cytokine possesses two isoforms, only the short form (sfTSLP) was regulated by PPARγ. sfTSLP mRNA expression was decreased both in PPARγ knock-down Caco2 cells and cells treated with PPARγ antagonist, whereas PPARγ agonists induced the expression of sfTSLP in Caco2 and T-84 cells. The response element activated by PPARγ was identified in the promoter of the sfTSLP gene by chromatin immunoprecipitation and gene reporter assays. The expression of sfTSLP was significantly decreased in the colonic mucosa of UC patients compared to controls and was correlated with PPARγ expression. Our results identified sfTSLP as a new PPARγ-target gene and support the hypothesis that, in UC, PPARγ deficiency in colonic mucosa could play a role in the loss of intestinal tolerance through an impaired sfTSLP expression.