Abstract
Background
Post-operative pancreatic fistula (POPF) formation occurs frequently after a pancreaticoduodenectomy (PD). Recently, a 10-point Fistula Risk Score (FRS) evaluating the likelihood of clinically relevant POPF (CR-POPF) development has been described and validated. This scheme has yet to be evaluated in PD patients managed without intra-operative drain placement.
Methods
Among patients undergoing PD at an academic centre since 2003, a retrospective analysis calculating FRS and its correlation with CR-POPF development was evaluated by logistic regression. Secondary analysis examined presentation and management of CR-POPF in undrained PD patients.
Results
FRS was calculated for 265 patients; 97.7% were managed without operative drains. The overall incidence of CR-POPF was 7.9%. Logistic regression revealed a 1.6-fold increase in CR-POPF risk per 1-point increase in FRS [95% confidence interval (CI) 1.2–2.0]. The negative predictive value in patients with FRS <3 was 100%, whereas the positive predictive value of FRS >6 was 16.7%. The median time to CR-POPF diagnosis was 18 days [interquartile range (IQR) 13–23]; 70.0% required readmission and 10.0% required a laparotomy.
Conclusions
Among patients without operative drainage, CR-POPF often has delayed presentations but most are managed non-operatively. The predictive value of high-risk FRS appears limited; conversely, a low-risk FRS accurately predicts the absence of CR-POPF and seems an appropriate metric for guiding care.
Introduction
Outcomes after a pancreaticoduodenectomy (PD) have steadily improved over the past several decades, in spite of increasing numbers of resections being performed in populations previously considered high risk, such as the elderly.1–3However, a persistently large fraction of patients experience significant morbidity or the need for readmission after a PD.4Of the PD-specific adverse events, a post-operative pancreatic fistula (POPF) is both common and highly morbid. The incidence of POPF has been variably reported as occurring in 2–22% of PD patients after a resection,5–7 and has been associated with increased rates of mortality, readmission, hospital costs and prolonged lengths of stay.4,8,9
The standardization of POPF diagnostic criteria by the International Study Group of Pancreatic Fistula (ISGPF) in 2005 greatly reduced discrepancies in defining POPF occurrence between institutions.5 Broadly, POPF were qualitatively defined as any measurable volume of intraperitoneal fluid collected on or after post-operative day (POD) three with amylase levels three times greater than the upper limit of the normal serum value. The ISGPF characterized clinically insignificant fistulae incidentally discovered after routine post-operative drainage requiring no deviation from routine post-operative care as ‘Grade A’ fistulae. Conversely, POPF necessitating a change in routine management were termed ‘Grade B’, whereas fistulae with life-threatening sequelae or requiring reoperation were termed ‘Grade C’. Collectively, grade B and C fistulae are referred to as ‘clinically-relevant POPF’ (CR-POPF).10,11
The ISGPF classification of POPF was first validated in 2007.9 Subsequently, a comprehensive analysis of hypothetical risk factors for POPF as defined by the ISGPF found that smaller pancreatic duct size, soft gland texture, increased operative blood loss and resections performed for ampullary, duodenal, cystic or islet cell disease were independent risk factors for CR-POPF development.12 Recently, a 10-point Fistula Risk Score (FRS) based on these risk factors has been described by the same group of authors in an effort to characterize the likelihood that a given patient will develop a CR-POPF after a PD.13 While soft gland texture and smaller duct size had been recognized in non-randomized studies and anecdotally as correlating with CR-POPF formation, this straightforward scoring system objectively characterized and attributed relative risk of these factors in relation to POPF development. In an ensuing validation study utilizing patients from three institutions, Miller et al. 14 confirmed the FRS to be predictive of CR-POPF risk. However, as the authors emphasize, PD is highly variable between surgeons and institutions, both in terms of its technical implementation and post-operative management techniques. Accordingly, external validation of the FRS utilizing a wide variety of PD patients is mandated.
The role of closed suction drains after PD remains controversial; in particular they are of interest in examining POPF development and applicability of scoring systems such as the FRS. In recent years it has been shown that PD can be performed safely without usage of closed suction drains in many cases,15–17 although many surgeons continue to have excellent outcomes while employing one or more drains routinely.18 In the original study defining the FRS, Callery et al. utilized at least one drain adjacent to the operative bed. Similarly, over 95% of the cases in the validation by Miller et al. also utilized operative drains. In our practice in a high-volume academic center, we have been utilizing selective intraperitoneal drainage following PD for several years. As such, this appeared to be an ideal cohort to further evaluate the FRS advocated by Callery et al. 13 The primary aim of this study was to evaluate the validity of the FRS in an independent cohort in which the majority of patients did not receive routine operative drainage and evaluate the strength of each component of the FRS in ascribing risk for CR-POPF development. Finally, in a separate analysis, the clinical presentation and course of patients who develop CR-POPF after a PD when routine operative drains are not utilized was analysed, as this population is growing and has not been widely studied.
Patients and methods
Patient selection and defining adverse events
A prospectively maintained database of consecutive patients who underwent PD for any indication at Yale-New Haven Hospital by a single surgeon (R.R.S.) from January 2006 to June 2013 was retrospectively reviewed. The presence of POPF was assessed and classified as grade A, B, or C as defined by the ISGPF. For each case, a corresponding FRS was calculated on a 10-point scale in accordance with the description of gland texture, duct size, etc by Callery et al. 13 in all patients for which the requisite data were available. Operative blood loss was determined in concert with the anaesthesia and nursing teams. Overall operative morbidity was graded in accordance with the modified Clavien classification after 30 days of follow-up, as were mortality and readmission rates.19,20Clavien grades I and II complications were considered minor, whereas grade III complications and above were considered major. Standardized ISGPS definitions were utilized for diagnosis of delayed gastric emptying (DGE).21Fluid drained from the abdomen with positive microbial cultures and amylase content less than three times normal was considered an intra-abdominal abscess. Gastro-/duodenojejunostomy or hepaticojejunostomy leak was defined as radiological or operative identification of anastomic disruption. Patients who underwent surgery for an intraductal papillary mucinous neoplasm that were found to have invasive adenocarcinoma were classified as having adenocarcinoma for the purposes of analysis.
Operative technique
A pylorus-preserving PD was performed except when precluded by disease-specific considerations; in all other cases, a classic (Kausch-Whipple) PD was performed. A pancreaticojejunostomy was performed in a two-layer end-to-side fashion utilizing duct-to-mucosa reconstruction whereas a hepaticojejunostomy was performed in a single-layer end-to-side fashion. A gastro-/duodenojejunostomy was performed using an end-to-side (Billroth II) technique. Closed suction drainage was rarely performed and usually only for difficult biliary anastomoses.
Post-operative care
Patients were admitted to the surgical intensive care unit post-operatively and usually transferred to the surgical ward on post-operative day (POD) one. After 2008, nasogastric tubes were rarely employed. A liquid diet was typically initiated on POD 2–3 and advanced as tolerated. Metoclopramide and octreotide were not routinely given, but were freely administered when clinically indicated. In patients with operatively placed closed suction drains, drain amylase was routinely evaluated and drains were discontinued when clinically indicated. After discharge, all patients were followed for at least 3 months.
Statistical analysis
Summary statistics were used to describe baseline characteristics. Chi-square tests and Wilcoxon’s rank-sum test were used to analyse categorical and continuous variables, respectively. Fisher’s exact test was used to analyse categorical variables with expected values less than five. Binary logistic regression was used to evaluate the accuracy of the Fistula Risk Score in the prediction of CR-POPF. A receiver-operating characteristic (ROC) curve was plotted and the area under the curve (AUC) concordance statistic calculated for model discrimination. The c-statistic varies from 0.5 to 1.0, and the ability of a model to predict outcomes increases as the c-statistic approaches 1. Positive likelihood ratios (LR+) were calculated for each score. LR+ represents the probability that a patient with POPF receives a particular score, divided by the probability that a patient without POPF receives the same score. A LR+ greater than 1 indicates that POPF is more likely whereas a LR+ less than one indicates that POPF is less likely. Statistical analysis was performed with SPSS version 19 (SPSS, Inc., Chicago, IL, USA). Data collection and analysis were performed under a protocol approved by the Yale University School of Medicine Human Investigations Committee.
Results
Fistula incidence and characteristics
During the study period from 2006 to 2013, 297 patients underwent PD by a single surgeon of which 265 had sufficient data available to calculate a FRS (89.2%). Twenty-one patients (7.9%) had CR-POPF (i.e. met criteria for ISGPF grade B or C fistulae). All were described as having a soft pancreas texture and only two underwent PD for pancreatic ductal adenocarcinoma or chronic pancreatitis. Baseline demographic and operative indication data are shown in Table 1. The overall median length of stay for all patients during their index admission was 6 days [range, 3–81; interquartile range (IQR) 5–7 days]. The most frequent indication for PD was pancreatic ductal adenocarcinoma. Drains were placed in only six patients (2.3%); one developed a CR-POPF. Drain amylase was assessed in the remaining five patients and none were greater than three times serum amylase. Four operative deaths were observed (1.5%); none of the deaths were as a result of POPF-related morbidity (three due to pulmonary complications, one due to a myocardial infarction).
Table 1.
Demographic and operative characteristics of patients undergoing Fistula Risk Score classification
| n (%) | |
|---|---|
| Total | 265 |
| Age (years) | |
| Mean (SEM) | 64.2 (0.7) |
| Gender | |
| Male | 136 (51.3) |
| Female | 129 (48.7) |
| Technique | |
| PD | 125 (47.2) |
| PPPD | 140 (52.8) |
| Intra-operative placement of drains | 6 (2.3) |
| Post-operative nasogastric tube | 79 (29.8) |
| Indication | |
| Benign | 64 (24.2) |
| Malignant | 201 (75.8) |
| Pathology | |
| Pancreatic carcinoma | 103 (38.9) |
| Cholangiocarcinoma | 32 (12.1) |
| Ampullary adenocarcinoma | 30 (11.3) |
| Duodenal carcinoma | 13 (4.9) |
| Pancreatic cystic neoplasm | 42 (15.8) |
| Neuroendocrine tumour | 21 (7.9) |
| Other | 24 (9.1) |
PD, pancreaticoduodenectomy; PPPD, pylorus-preserving PD.
FRS and adverse events
All 265 patients were stratified as having negligible (0 points), low (1–2 points), intermediate (3–6 points) or high risk (7–10 points) by FRS (the same classification utilized by Callery et al.). The mean FRS (regardless of CR-POPF development) was 3.66 whereas the median FRS was 3. Increasing risk score correlated with POPF development (P < 0.05), and all patients that developed fistulae were found to have intermediate or high risk scores (see Table 2). In addition, the FRS was also found to correlate with longer lengths of stay regardless of the presence of POPF (P < 0.01). The overall rate of any post-operative adverse event (Clavien–Dindo score ≠ 0) was 83.2%, with major complications (Clavien–Dindo score >2) occurring in 15.1% of cases. Aggregate operative morbidity as measured by Clavien score did not have any statistically significant association with increased FRS, although a trend towards a higher incidence of major complications was observed in patients in the intermediate or high risk FRS groups. No statistically significant correlation of FRS classification and additional selected PD-related complications, such as delayed gastric emptying, intra-abdominal abscess development and anastomotic leak was observed.
Table 2.
Correlation of post-operative adverse events and Fistula Risk Score
| Negligible risk (0 points) n = 17 | Low risk (1–2 points) n = 72 | Intermediate risk (3–6 points) n = 152 | High risk (7–10 points) n = 24 | P-value | |
|---|---|---|---|---|---|
| Median LOS (IQR) | 5 (4–6) | 5 (4–7) | 6 (5–9) | 7 (6–9) | <0.001 |
| ISGPF classification | 0.012 | ||||
| No fistula | 17 (100) | 72 (100) | 135 (88.8) | 20 (83.3) | |
| Grade B | – | – | 11 (7.2) | 4 (16.7) | |
| Grade C | – | – | 6 (3.9) | – | |
| Overall morbidity | 0.160 | ||||
| No complications | 5 (29.4) | 17 (23.6) | 19 (12.6) | 2 (9.1) | |
| Minor (Clavien ≤2) | 10 (58.8) | 49 (68.1) | 103 (68.2) | 17 (77.3) | |
| Major (Clavien ≥3) | 2 (11.8) | 6 (8.3) | 29 (19.2) | 3 (13.6) | |
| Specific complications | |||||
| DGE | 1 (5.8) | 6 (8.3) | 18 (11.8) | 2 (8.7) | 0.763 |
| Abscess | – | 2 (2.8) | 7 (4.6) | 1 (4.2) | 0.762 |
| Anastomotic leak | – | 2 (2.8) | 2 (1.3) | 1 (4.2) | 0.665 |
| Transfusion | 1 (5.8) | 7 (9.7) | 16 (10.5) | 1 (4.2) | 0.741 |
| Death | – | 1 (1.4) | 3 (2.0) | – | 0.776 |
LOS, length of stay; IQR, interquartile range; DGE, delayed gastric emptying.
Evaluating the validity of the FRS
Analysis of the predictive value in our cohort of each component of the FRS is shown in Table 3. A strong correlation with fistula development was observed in cases of soft pancreas texture, decreasing pancreatic duct size and ampullary/duodenal/cystic/islet cell pathology (P < 0.05 for all). Increased operative blood loss was not found to correlate with increasing fistula risk. Rates of CR-POPF and corresponding positive likelihood ratios at each risk score level are shown in Fig. 1. Binary logistic regression revealed a 1-point increase in the FRS to be associated with a 1.6-fold increase in odds of the development of CR-POPF [95% confidence interval (CI) 1.2–2.0, P = 0.001]. The ROC (receiver-operating characteristic) curve concordance statistic for the entire cohort was found to be 0.763 (95% CI 0.684–0.843). None of the low-risk patients (FRS < 3) developed CR-POPF, with a corresponding negative predictive value of 100%. Conversely, in high-risk patients (FRS >6), the positive predictive value was 16.7%.
Table 3.
Correlation of Fistula Risk Score components and presence or absence of clinically relevant post-operative pancreatic fistula (CR-POPF)
| Parameter | No CR-POPF n = 244 | CR-POPF n = 21 | P-value |
|---|---|---|---|
| Gland texture, n (%) | <0.001 | ||
| Firm | 120 (49.2) | – | |
| Soft | 124 (50.8) | 21 (100) | |
| Pathology, n (%) | 0.013 | ||
| Low risk | 96 (39.3) | 2 (9.5) | |
| High risk | 148 (60.7) | 19 (90.5) | |
| Duct diameter (cm), n (%) | 0.001 | ||
| ≥5 | 92 (37.7) | 2 (9.5) | |
| 4 | 51 (20.9) | 5 (23.8) | |
| 3 | 63 (25.8) | 4 (19.0) | |
| 2 | 38 (15.6) | 10 (47.6) | |
| ≤1 | – | – | |
| Estimated blood loss (ml), n (%) | 0.957 | ||
| ≤400 | 141 (57.8) | 13 (61.9) | |
| 401–700 | 62 (25.4) | 5 (23.8) | |
| 701–1000 | 22 (9.0) | 2 (9.5) | |
| >1000 | 19 (7.8) | 1 (4.8) |
‘Low-risk’ pathology is considered resection for pancreatic ductal adenocarcinoma or chronic pancreatitis only. ‘High-risk’ pathology is resection for any other reason.
Figure 1.
Percentage of patients ascribed to each Fistula Risk Score with corresponding percentage of clinically-relevant post-operative pancreatic fistulae. Likelihood ratio (‘LR+’) for fistula development at each score shown
Presentation of CR-POPF in the absence of routine operative drainage
The clinical presentation and management of all patients during the study period with CR-POPF who did not have intraperitoneal drains maintained post-operatively were critically analysed. Of the 21 patients with CR-POPF, one patient did have an intra-operative drain maintained and was therefore excluded. The results are shown in Table 4. Fever and abdominal pain were the most common presenting symptoms (65.0% of patients with either finding) and leukocytosis was the most common objective finding (85.0%). Two patients (10.0%) were diagnosed after a re-laparotomy with the remainder (n = 18, 90.0%) diagnosed after observation of an intra-abdominal fluid collection found to be amylase-rich on subsequent analysis. All patients who were diagnosed radiologically were treated with percutaneous drainage. Nine presented during their index admission (36.0%) with the remainder requiring readmission for management. Among those readmitted, the median time to re-presentation was 16 days after initial discharge (range, 7–42 days).
Table 4.
Presentation of clinically relevant post-operative pancreatic fistula (POPF) in the absence of post-operative intraperitoneal drainage (n = 20)
| Clinical findings, n (%) | |
|---|---|
| Leukocytosis | 17 (85.0) |
| Fever | 13 (65.0) |
| Abdominal pain | 13 (65.0) |
| Tachycardia | 9 (45.0) |
| Nausea/vomiting | 7 (35.0) |
| Anorexia | 7 (35.0) |
| Respiratory | 5 (25.0) |
| Fatigue | 4 (20.0) |
| Hypotension | 4 (20.0) |
| Diarrhoea | 2 (10.0) |
| Altered mental status | 2 (10.0) |
| Chest pain | 2 (10.0) |
| Index admission | |
| Post-operative day of diagnosis, median (IQR) | 18 (13–23) |
| Length of stay (days), median (IQR) | 7 (6–18.5) |
| Readmission | |
| POPF requiring readmission, n (%) | 14 (70.0) |
| Length of stay for readmission (days), median (IQR) | 10.5 (6–14.0) |
| Diagnosis, n (%) | |
| Abdominal CT | 18 (90.0) |
| Exploratory laparotomy | 2 (10.0) |
| Treatment, n (%) | |
| CT-guided IR drainage | 24 (95.0) |
| Antibiotics | 16 (80.0) |
| Total parenteral nutrition | 5 (20.0) |
| Reoperation | 2 (10.0) |
IQR, interquartile range; CT, computed tomography; IR; interventional radiology.
Discussion
Development of pancreatic fistulae remains one of the most common and highly morbid adverse events after PD and multiple interventions have been evaluated that attempt to prevent or mitigate the effects of POPF. Trials evaluating these techniques have thus far shown only a limited ability to impact POPF incidence, even when broadly applied.22–25The development of the ISGPF classification for POPF has greatly improved POPF research through standardization, although alternative schemes are still occasionally utilized.16 The 10-point FRS developed by Callery et al. 13 based upon risk factors first described by Pratt et al.12 to stratify CR-POPF risk13 was recently validated in a multi-institutional trial.14 As nearly all patients in both the original study describing the FRS and the study validating it had operative drains utilized in their care, this study sought to broaden the applicability of the FRS risk classification by evaluating it in a large cohort of patients undergoing PD with rare usage of drains. In summary, the current study found that the FRS was associated with the development of CR-POPF, and was particularly helpful in ruling out the possibility of CR-POPF in low-risk patients. However, its ability to predict the development of CR-POPF in high-risk patients was limited.
In their original assessment of the FRS, Callery et al. 13 prospectively evaluated the performance of the scheme on a second group of patients at their institution and described a concordance statistic (‘c-statistic’) of 0.942 (better = closer to 1.0). In the present study of undrained patients a c-statistic of 0.763 (95% CI 0.684–0.843) was found. The moderately decreased concordance observed compared with the initial FRS study is likely as a result of the decreased positive predictive value in patients ascribed a high-risk FRS (16.7% versus 89% in the original study). This is somewhat magnified as the mean FRS in the present study is nearly a full point higher than the original study (3.66 versus 2.68), in spite of having a lower overall CR-POPF rate. In their validation of the FRS scheme in patients with routine drainage after PD, Miller et al. 13 described a comparable decrease in CR-POPF incidence in high-risk FRS patients (29%) in spite of a similarly elevated mean FRS (3.54), which resulted in a c-statistic of 0.716 and mirrors the present study closely. The decreased predictive ability of high-risk FRS values is likely multifactorial.
Wide variation currently exists in the practice of operative drainage after pancreatic surgery. This has important effects in determining the true incidence of POPF in a particular cohort, as the majority of patients with grade A fistulae may remain undiagnosed if intraperitoneal drains are not routinely utilized, resulting in a higher POPF incidence (grades A/B/C) in studies where drain usage is routine. While differences in technical aspects of PD do not seem to significantly affect the incidence of post-operative adverse events, such as POPF or DGE development,12,26,27groups that do not routinely utilize operative drains typically report a slightly lower incidence of CR-POPF. The overall incidence of 7.9% observed here is comparable to that reported by other groups where operative drains were not routinely employed.16,17 Callery et al. 13 (in the original FRS study) and Miller et al. 13 (in the FRS validation) report rates of CR-POPF of 13.0% and 11.4%, respectively, among their patients (who were routinely drained). The decreased incidence of CR-POPF observed in the present cohort may have contributed to the lower positive predictive value found at higher FRS numbers, as the positive predictive value of a scoring system depends on the incidence of the event in question.
Operative blood loss is a comparatively large component of the FRS, as up to 3 of a possible 10 points in the model are attributed to volume of blood loss. A key finding of the present study was that a difference in blood loss between patients with and without CR-POPF was not observed. As a result, among patients who were granted points for high operative blood loss, the risk for POPF development appears to be less than that ascribed by the FRS. Consequently, this would falsely increase the number of predicted CR-POPFs and also result in the decreased positive predictive value observed. The prevalence of patients in the present study with an elevated blood loss (>400 ml, 41.9%) is similar to that observed in the original FRS and validation studies (36.6% and 36.4%, respectively). The effects of the operative blood loss on CR-POPF incidence was not reported by Miller et al. 13 in their FRS validation of drained PD patients. However, during subgroup analysis, an increased rate of operative blood loss in the current cohort during cases for PDAC and chronic pancreatitis was observed, possibly owing to the increased technical difficulty in these cases and the higher incidence of concomitant vascular resections during oncological resections. Although this trend did not reach statistical significance, the protective effect of the larger duct and firmer gland usually associated with PDAC/chronic pancreatitis resections may have mitigated the adverse effects on CR-POPF formation of increased blood loss endorsed by Callery et al. 13 In summary, both a lack of correlation of CR-POPF with blood loss and a lower overall incidence of CR-POPF contributed to the decreased positive predictive value of the FRS observed in this study.
Conversely, the negative predictive value in patients with a FRS less than three (‘negligible’ or ‘low’ risk patients) was excellent (100%), as none of these patients went on to develop CR-POPF. The FRS validation of drained PD patients by Miller et al. 13 demonstrates a similarly minimal rate of CR-POPF formation in patients with a FRS < 3. This is a crucial finding of the present study and further highlights the utility of the FRS and demonstrates the potential role for the scheme in clinical practice. Conceivably, patients categorized as negligible or low risk could be managed without adjuncts for POPF treatment with confidence. In practices where selective operative drainage is employed, it could also serve as a metric for determining which patients can forgo drainage. In practices that employ fast-track discharge plans or dietary advancement, confidence in that such steps can be taken in patients with a low FRS would be increased. Conversely, in those patients with a higher FRS, interpreting post-operative adverse events in the context of a higher FRS could direct clinical suspicion and diagnostic interventions, leading to earlier POPF diagnosis and treatment. Even although this study demonstrates that not all patients in intermediate- or high-risk FRS categories develop POPF, increased awareness of the risk of POPF development would be useful in patients that do not experience a routine post-operative course.
A secondary objective of this study was to characterize how patients with CR-POPF present in the absence of routine operative drainage, as an early diagnosis of POPF is considered one rationale for the placement of drains. In patients with operative drains, persistent drain output is commonly considered an indicator of POPF development.28 Among patients without drains, the presentation of POPF has not been systematically reported. This study shows that patients typically present within 4 to 6 weeks of operation, most commonly with leukocytosis, fever and persistent abdominal pain. Encouragingly, no mortalities secondary to CR-POPF were observed in spite of a lack of operative drainage, although two patients became critically ill and required reoperation. All patients in the cohort ultimately required drainage, either via relaparotomy or radiologically-guided percutaneous drains. Also, nearly two-thirds did require readmission, with a median duration of 7 days. Delayed presentation and elevated risk for readmission seem to be a consistent risk after a PD, as Ahmad et al. 13 described in a recent multi-centre study evaluating readmission after PD.4In general, among those managed without operative drainage, the incidence of CR-POPF development appears low and compares favourably with previously published data. Although those patients who do develop CR-POPF frequently require intervention and readmission, the majority of these patients are managed with excellent outcomes.
This study was limited in scope by its retrospective nature, although the data presented here were collected prospectively. In spite of this, a large cohort with an incidence of CR-POPF and other post-operative adverse events comparable to published longitudinal case series was assembled and critically assessed.1,2 With the multi-institution validation of the FRS performed by Miller et al. 13 of a large cohort that underwent routine post-operative drainage combined with the large cohort of undrained patients presented here, the FRS scheme has been robustly evaluated in a sizable group of patients that would be comparable to those at many centres. Overall, the FRS reliably predicts the absence of CR-POPF in negligible- and low-risk patients and provides an objective way to characterize POPF risk, rather than relying on inference or surgeon experience. The prognostic ability among patients with an elevated FRS appears somewhat limited. While further model refinement may yield increased accuracy, the current FRS classification appears to be appropriate for general usage and directing fistula-related management decisions.
Conflicts of interest
None declared.
References
- Balcom JH, 4th, Rattner DW, Warshaw AL, Chang Y, Fernandez-del Castillo C. Ten-year experience with 733 pancreatic resections: changing indications, older patients, and decreasing length of hospitalization. Arch Surg. 2001;136:391–398. doi: 10.1001/archsurg.136.4.391. [DOI] [PubMed] [Google Scholar]
- Cameron JL, Riall TS, Coleman J, Belcher KA. One thousand consecutive pancreaticoduodenectomies. Ann Surg. 2006;244:10–15. doi: 10.1097/01.sla.0000217673.04165.ea. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Suzuki S, Kaji S, Koike N, Harada N, Suzuki M. Pancreaticoduodenectomy can be safely performed in the elderly. Surg Today. 2013;43:620–624. doi: 10.1007/s00595-012-0383-6. [DOI] [PubMed] [Google Scholar]
- Ahmad SA, Edwards MJ, Sutton JM, Grewal SS, Hanseman DJ, Maithel SK, et al. Factors influencing readmission after pancreaticoduodenectomy: a multi-institutional study of 1302 patients. Ann Surg. 2012;256:529–537. doi: 10.1097/SLA.0b013e318265ef0b. [DOI] [PubMed] [Google Scholar]
- Bassi C, Dervenis C, Butturini G, Fingerhut A, Yeo C, Izbicki J, et al. Postoperative pancreatic fistula: an international study group (ISGPF) definition. Surgery. 2005;138:8–13. doi: 10.1016/j.surg.2005.05.001. [DOI] [PubMed] [Google Scholar]
- Buchler MW, Friess H, Wagner M, Kulli C, Wagener V, Z’Graggen K. Pancreatic fistula after pancreatic head resection. Br J Surg. 2000;87:883–889. doi: 10.1046/j.1365-2168.2000.01465.x. [DOI] [PubMed] [Google Scholar]
- Halloran CM, Ghaneh P, Bosonnet L, Hartley MN, Sutton R, Neoptolemos JP. Complications of pancreatic cancer resection. Dig Surg. 2002;19:138–146. doi: 10.1159/000052029. [DOI] [PubMed] [Google Scholar]
- Lin JW, Cameron JL, Yeo CJ, Riall TS, Lillemoe KD. Risk factors and outcomes in postpancreaticoduodenectomy pancreaticocutaneous fistula. J Gastrointest Surg. 2004;8:951–959. doi: 10.1016/j.gassur.2004.09.044. [DOI] [PubMed] [Google Scholar]
- Pratt WB, Maithel SK, Vanounou T, Huang ZS, Callery MP, Vollmer CM., Jr Clinical and economic validation of the International Study Group of Pancreatic Fistula (ISGPF) classification scheme. Ann Surg. 2007;245:443–451. doi: 10.1097/01.sla.0000251708.70219.d2. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Gebauer F, Kloth K, Tachezy M, Vashist YK, Cataldegirmen G, Izbicki JR, et al. Options and limitations in applying the fistula classification by the International Study Group for Pancreatic Fistula. Ann Surg. 2012;256:130–138. doi: 10.1097/SLA.0b013e31824f24e4. [DOI] [PubMed] [Google Scholar]
- Vollmer CM, Jr, Sanchez N, Gondek S, McAuliffe J, Kent TS, Christein JD, et al. A root-cause analysis of mortality following major pancreatectomy. J Gastrointest Surg. 2012;16:89–102. doi: 10.1007/s11605-011-1753-x. discussion 102–103. [DOI] [PubMed] [Google Scholar]
- Pratt WB, Callery MP, Vollmer CM., Jr Risk prediction for development of pancreatic fistula using the ISGPF classification scheme. World J Surg. 2008;32:419–428. doi: 10.1007/s00268-007-9388-5. [DOI] [PubMed] [Google Scholar]
- Callery MP, Pratt WB, Kent TS, Chaikof EL, Vollmer CM., Jr A prospectively validated clinical risk score accurately predicts pancreatic fistula after pancreatoduodenectomy. J Am Coll Surg. 2013;216:1–14. doi: 10.1016/j.jamcollsurg.2012.09.002. [DOI] [PubMed] [Google Scholar]
- Miller BC, Christein JD, Behrman SW, Drebin JA, Pratt WB, Callery MP, et al. A Multi-Institutional External Validation of the Fistula Risk Score for Pancreatoduodenectomy. J Gastrointest Surg. 2014;18:172–179. doi: 10.1007/s11605-013-2337-8. [DOI] [PubMed] [Google Scholar]
- Conlon KC, Labow D, Leung D, Smith A, Jarnagin W, Coit DG, et al. Prospective randomized clinical trial of the value of intraperitoneal drainage after pancreatic resection. Ann Surg. 2001;234:487–493. doi: 10.1097/00000658-200110000-00008. discussion 493–494. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Correa-Gallego C, Brennan MF, D’Angelica M, Fong Y, Dematteo RP, Kingham TP, et al. Operative drainage following pancreatic resection: analysis of 1122 patients resected over 5 years at a single institution. Ann Surg. 2013;258:1051–1058. doi: 10.1097/SLA.0b013e3182813806. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Mehta VV, Fisher SB, Maithel SK, Sarmiento JM, Staley CA, Kooby DA. Is it time to abandon routine operative drain use? A single institution assessment of 709 consecutive pancreaticoduodenectomies. J Am Coll Surg. 2013;216:635–642. doi: 10.1016/j.jamcollsurg.2012.12.040. discussion 642–644. [DOI] [PubMed] [Google Scholar]
- Shrikhande SV, Barreto SG, Shetty G, Suradkar K, Bodhankar YD, Shah SB, et al. Post-operative abdominal drainage following major upper gastrointestinal surgery: single drain versus two drains. J Cancer Res Ther. 2013;9:267–271. doi: 10.4103/0973-1482.113380. [DOI] [PubMed] [Google Scholar]
- Clavien PA, Barkun J, de Oliveira ML, Vauthey JN, Dindo D, Schulick RD, et al. The Clavien-Dindo classification of surgical complications: five-year experience. Ann Surg. 2009;250:187–196. doi: 10.1097/SLA.0b013e3181b13ca2. [DOI] [PubMed] [Google Scholar]
- Dindo D, Demartines N, Clavien PA. Classification of surgical complications: a new proposal with evaluation in a cohort of 6336 patients and results of a survey. Ann Surg. 2004;240:205–213. doi: 10.1097/01.sla.0000133083.54934.ae. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Wente MN, Bassi C, Dervenis C, Fingerhut A, Gouma DJ, Izbicki JR, et al. Delayed gastric emptying (DGE) after pancreatic surgery: a suggested definition by the International Study Group of Pancreatic Surgery (ISGPS) Surgery. 2007;142:761–768. doi: 10.1016/j.surg.2007.05.005. [DOI] [PubMed] [Google Scholar]
- Orci LA, Oldani G, Berney T, Andres A, Mentha G, Morel P, et al. Systematic review and meta-analysis of fibrin sealants for patients undergoing pancreatic resection. HPB (Oxford) 2014;16:3–11. doi: 10.1111/hpb.12064. [DOI] [PMC free article] [PubMed] [Google Scholar]
- Vanounou T, Pratt WB, Callery MP, Vollmer CM., Jr Selective administration of prophylactic octreotide during pancreaticoduodenectomy: a clinical and cost-benefit analysis in low- and high-risk glands. J Am Coll Surg. 2007;205:546–557. doi: 10.1016/j.jamcollsurg.2007.05.011. [DOI] [PubMed] [Google Scholar]
- You D, Jung K, Lee H, Heo J, Choi S, Choi D. Comparison of different pancreatic anastomosis techniques using the definitions of the International Study Group of Pancreatic Surgery: a single surgeon’s experience. Pancreas. 2009;38:896–902. doi: 10.1097/MPA.0b013e3181b365f7. [DOI] [PubMed] [Google Scholar]
- Zhou Y, Zhou Q, Li Z, Chen R. Internal pancreatic duct stent does not decrease pancreatic fistula rate after pancreatic resection: a meta-analysis. Am J Surg. 2013;205:718–725. doi: 10.1016/j.amjsurg.2012.07.039. [DOI] [PubMed] [Google Scholar]
- Kunstman JW, Fonseca AL, Ciarleglio MM, Cong X, Hochberg A, Salem RR. Comprehensive analysis of variables affecting delayed gastric emptying following pancreaticoduodenectomy. J Gastrointest Surg. 2012;16:1354–1361. doi: 10.1007/s11605-012-1873-y. [DOI] [PubMed] [Google Scholar]
- Muscari F, Suc B, Kirzin S, Hay JM, Fourtanier G, Fingerhut A, et al. Risk factors for mortality and intra-abdominal complications after pancreatoduodenectomy: multivariate analysis in 300 patients. Surgery. 2006;139:591–598. doi: 10.1016/j.surg.2005.08.012. [DOI] [PubMed] [Google Scholar]
- Hashimoto N, Yasuda C, Ohyanagi H. Pancreatic fistula after pancreatic head resection; incidence, significance and management. Hepatogastroenterology. 2003;50:1658–1660. [PubMed] [Google Scholar]