Abstract
Objectives: Morbidity after pancreaticoduodenectomy (PD) remains high. Computed tomography (CT) of intra-abdominal tissue has not been thoroughly evaluated to establish associations with the occurrence of complications after PD. The current study sought to determine whether differences in non-enhanced visceral attenuation predicted complications after PD.
Methods: Outcomes in patients undergoing PD were analysed according to the Clavien system for classifying complications and the International Study Group on Pancreatic Fistula system for classifying postoperative pancreatic fistula (POPF). Preoperative non-enhanced CT scans were evaluated by a blinded investigator for attenuation of abdominal viscera and fat thickness. Data on pancreatic firmness and pancreatic duct size were collected. Univariate and multivariate analyses were performed.
Results: A total of 134 patients underwent PD for malignant and benign disease. Rates of morbidity, mortality and POPF at 90 days were 61%, 4% and 23%, respectively. Patients with a body mass index of > 25 kg/m2 had higher rates of POPF (P = 0.05) and complications (P < 0.01). In multivariate analysis, patients were more likely to develop any complication as CT attenuation decreased for paraspinus muscle (P < 0.01), spleen (P < 0.03) and liver (P = 0.01) parenchyma.
Conclusions: Postoperative complications after PD remain prevalent. Decreased CT attenuation of abdominal viscera is an independent predictor of morbidity after PD and suggests a high-risk patient physiology for pancreatic resection.
Introduction
Morbidity and mortality rates following operations are often determined at 30 days and seldom refer to the entire in-hospital course. Furthermore, 90-day rates of morbidity and mortality following pancreaticoduodenectomy (PD) at high-volume centres have only recently been published, at 55% and 4.7%, respectively.1 Risk assessment is paramount to stratify patients who are at high risk for poor outcomes. Several validated models, such as the American Society of Anesthesiologists (ASA) class, POSSUM (physiological and operative severity score for the enumeration of mortality and morbidity), Charlson, SOAR (systolic blood pressure, oxygenation, age and respiratory rate) and the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) models, have emerged to aid in patient selection and ameliorate poor surgical outcomes.2–7 Of these, only NSQIP accounts for body mass index (BMI) and none account for computed tomography (CT) attenuation values.
Up to 30% of patients presenting for PD are obese; over 35% of these patients experience complications after PD and 24% experience a serious complication, compared with 14% of patients of normal weight.8 Obesity is associated with increases in the incidence of wound infections and postoperative pancreatic fistula (POPF), estimated blood loss (EBL) and operative times.3,9
Multiphasic CT is critical to operative planning for PD as it is used to evaluate the resectability of tumours and the spatial relationship of structures adjacent to the pancreas. Although CT is critical in the preoperative evaluation of patients scheduled to undergo PD, no risk assessment tool utilizes CT to predict risk.
Computed tomography provides a measure of both abdominal girth and visceral fat. House et al. showed that retrorenal fat of > 2 cm predicted adverse outcomes following PD.10 Computed tomography also provides a measure of X-ray attenuation and thus tissue density for specific regions of interest (RoI), measured in Hounsfield units (HU). Hashimoto et al. described CT enhancement of the pancreas to be a surrogate for pancreatic fibrosis, which accurately predicted POPF.11
Historically, the main sources of morbidity and mortality following PD have been POPF and its sequelae, occurring at rates as high as 30%.12 A critical review of the literature shows that although the rate of POPF remains high, the majority of fistulae are not significant; therefore most morbidity is caused by non-fistulae sources. The metabolic syndrome of obesity is likely to represent a significant contributor to morbidity following PD and assessing for X-ray attenuation in HU may be used as a surrogate for the adiposity of tissues. The present study therefore sought to measure visceral attenuation in patients undergoing PD and to compare these measurements with 90-day morbidity rates.
Materials and methods
Patients
Data for patients undergoing PD for both benign and malignant disease, carried out at the University of Alabama at Birmingham (UAB) by the senior author during 2005–2010 were collected into a prospective database. The ACS-NSQIP and Charlson prediction models were used to guide the collection of perioperative patient characteristics.5 Medical records were retrospectively reviewed to determine 90-day morbidity and mortality rates based on the Clavien system of classification.13 Postoperative pancreatic fistula and delayed gastric emptying (DGE) were classified as previously described by the International Study Group on Pancreatic Fistula (ISGPF) and the International Study Group of Pancreatic Surgery (ISGPS), respectively.14,15
Operations included standard and pylorus-preserving PD. Intraoperative assessment of pancreatic firmness (on a scale of 1–10) and duct diameter (in mm) were documented for each patient. A 15-French round drain was left near the two-layer, end-to-side, duct-to-mucosa pancreaticojejunostomy in all cases. A free-floating silastic stent was used in all cases. Drain amylase levels were routinely evaluated on postoperative day 4 or 5 to determine POPF based upon ISGPF definitions.14
Computed tomography measurements
Radiographic consultation was used to assess each patient's preoperative imaging for quality. Preoperative imaging was reviewed by one observer blinded to patient data and outcomes (JCM). Only patients for whom preoperative non-enhanced CT scans of the abdomen and pelvis were available were included in the CT assessment of visceral attenuation and intra-abdominal fat. Non-enhanced rather than contrasted studies were used to assess parenchymal density in a manner that was not confounded by vascular inflow/outflow and cardiac output characteristics.
Hip girdle, periumbilical and retrorenal fat thicknesses were measured in millimetres as previously described.10 Attenuation measurements (in HU) were placed for pancreatic, splenic, hepatic and paraspinus muscle attenuation. For pancreatic attenuation, the RoI was approximate to but did not include the superior mesenteric artery and did not include the pancreatic duct. For splenic, hepatic and paraspinus muscle attenuation, representative RoIs were peripheral and distant from significant vasculature.
Statistics
Visceral organ attenuation, body fat thickness, POPF and Clavien categories were compared according to demographics and clinical characteristics by using analysis of variance (anova) and chi-squared statistics for continuous and categorical variables, respectively. Continuous variables were analysed using t-tests. Logistic regression was used to determine associations with independent variables. All analyses were performed using sas Version 9.2 (SAS Institute, Inc., Cary, NC, USA).
Results
Patients and outcomes overview
The present group's experience with PD at UAB included 332 patients during 2005–2010. Table 1 shows the characteristics of the entire patient population presenting for PD during the period under study. The mean age of these patients was 62 years and 55% of patients were aged > 65 years. Patient gender was evenly distributed. Mean BMI was 27 kg/m2 and 24% of patients were obese (BMI > 30 kg/m2). Over a third of patients were smokers at the time of surgery. Most patients (53%) underwent PD for malignant disease; the remainder underwent PD for chronic pancreatitis (21%) or another benign aetiology (26%).
Table 1.
Patient cohort, complications overview and computed tomography subset analysis
| Characteristics | All patients (n = 332) | Preoperative non-enhanced CT |
P-value | |
|---|---|---|---|---|
| Yes (n = 134) | No (n = 198) | |||
| Age, years, mean ± SD | 62.1 ± 11.4 | 62.7 ± 12.3 | 61.7 ± 14.1 | 0.52 |
| Elderly, % | 55% | 56% | 54% | 0.81 |
| Male gender, % | 49% | 51% | 47% | 0.43 |
| Body mass index, kg/m2, mean ± SD | 27.0 ± 10.9 | 26.5 ± 5.7 | 27.5 ± 13.2 | 0.39 |
| Obese, % | 24% | 24% | 24% | 0.86 |
| Smokers, % | 35% | 28% | 39% | 0.04 |
| Histology, % | 0.48 | |||
| Pancreatic adenocarcinoma | 40% | 41% | 41% | |
| Ampullary adenocarcinoma | 6% | 5% | 8% | |
| Cholangiocarcinoma | 5% | 5% | 4% | |
| Duodenal cancer | 2% | 1% | 2% | |
| Chronic pancreatitis | 21% | 23% | 18% | |
| Benign or cystic lesion | 26% | 25% | 27% | |
| Length of stay, days, mean ± SD | 11 ± 8 | 11 ± 9 | 11 ± 7 | 0.52 |
| Estimated blood loss, ml, mean ± SD | 334 ± 245 | 337 ± 291 | 332 ± 210 | 0.88 |
| Transfusion, % | 19% | 17% | 20% | 0.60 |
| Complications, % | 66% | 61% | 69% | 0.16 |
| Clavien class 1 | 12% | 7% | 15% | |
| Clavien class 2 | 36% | 37% | 35% | |
| Clavien class 3 | 8% | 5% | 10% | |
| Clavien class 4 | 8% | 8% | 7% | |
| Clavien class 5 | 3% | 4% | 2% | |
| Postoperative pancreatic fistula, % | 21% | 23% | 20% | 0.48 |
| ISGPF grade A | 16% | 19% | 14% | |
| ISGPF grade B | 4% | 2% | 5% | |
| ISGPF grade C | 2% | 2% | 1% | |
CT, computed tomography; SD, standard deviation; ISGPF, International Study Group on Pancreatic Fistula.
All outcomes are evaluations within 90 days of pancreaticoduodenectomy.
The mean total length of stay (LoS) during the initial 90 days after surgery was 11 days including both preoperative and postoperative days. The EBL during PD was 334 ml and 19% of patients required red blood cell transfusions during this timeframe. Overall 90-day morbidity and mortality rates were 66% and 3%, respectively. Significant complications (Clavien class ≥ 3) occurred in only 19% of patients. The overall rate of POPF was 21%; however, the rate of clinically significant (grade B or C) POPF was only 6%.
Preoperative CT subset analysis
Many of the patients referred to the present department presented with CT imaging obtained at an outside facility. Radiologist consultation was sought to evaluate imaging quality. This consultation sought to determine whether patients with appropriate preoperative imaging were representative of the entire UAB series.
Of the 332 patients, 134 patients had appropriate preoperative non-enhanced CT that included all areas in which measurements of fat thickness or visceral density were required. In the remaining patients, preoperative imaging was not available or was of insufficient quality to allow adequate analysis. A subset analysis was performed to compare the patient cohort with adequate preoperative imaging (n = 134) with that in which appropriate preoperative imaging was unavailable (n = 198). As Table 1 shows, all variables except smoking status (P = 0.04) were equivalent (P > 0.05). This suggests that the 134 patients for whom preoperative imaging was adequate were representative of the UAB experience.
Univariate analysis of complications following PD
Patient characteristics and resulting complications in the 134 patients with adequate preoperative imaging are shown in Table 2. A total of 82 (61%) patients experienced a postoperative complication. The most common complications following PD in this cohort were POPF (23%), intra-abdominal abscess (11%), respiratory distress (11%) including pneumonia, DGE (10%) and wound infection (10%). Patients who experienced complications were more likely to be obese, have a BMI of > 25 kg/m2 (P < 0.01), have a higher EBL (P < 0.05), and receive a red blood cell transfusion during the operation or during their first hospital stay (P < 0.01). The occurrence of a postoperative complication increased the mean LoS from 8.4 days to 13.3 days (P < 0.01).
Table 2.
Univariate analysis of patient characteristics and complications in patients with preoperative non-enhanced computed tomography
| Characteristics | Complication (Clavien class 1–5) |
P-value | |
|---|---|---|---|
| Yes (n = 82) | No (n = 52) | ||
| Body mass index, kg/m2, mean ± SD | 28 ± 26 | 24 ± 23 | < 0.01 |
| Obese, % | 31% | 11% | < 0.01 |
| Length of stay, days, mean ± SD | 13.3 ± 11.3 | 8.4 ± 2.1 | < 0.01 |
| Transfusion, % | 26% | 2% | < 0.01 |
| Estimated blood loss, ml, mean ± SD | 377 ± 329 | 268 ± 196 | < 0.05 |
| Age, years, mean ± SD | 62 ± 13 | 63 ± 11 | 0.68 |
| Elderly, % | 54% | 57% | 0.77 |
| Male gender, % | 61% | 55% | 0.45 |
| Smoker, % | 25% | 33% | 0.30 |
| Histology, % | 62% | 55% | 0.44 |
SD, standard deviation.
All outcomes are evaluations within 90 days of pancreaticoduodenectomy.
Computed tomography measurements and univariate analysis
Preoperative non-enhanced abdominal CT images were evaluated for pancreatic, splenic, hepatic and paraspinus muscle attenuation, as well as for fat thickness as depicted in Fig. 1. Mean values were obtained in the group of patients who experienced any postoperative complication and in those who did not. Table 3 shows that mean pancreatic (P = 0.13), splenic (P = 0.03), hepatic (P < 0.01) and muscle (P < 0.01) attenuation were lower in patients who experienced a complication. Increased periumbilical fat thickness (P < 0.01) and decreased pancreatic firmness (P < 0.01) were both associated with an increased incidence of postoperative complications in univariate analysis. Pancreatic firmness did not correlate with pancreatic parenchymal CT attenuation (P = 0.18). Hip girdle (P = 0.22) and retrorenal fat thickness (P = 0.06) were not significantly associated with postoperative complications. Pancreatic duct diameter was not associated with an elevated complication rate following PD (P = 0.19).
Figure 1.
Examples of preoperative computed tomography (CT) attenuation and fat thickness measurements. Axial non-enhanced preoperative CT is evaluated for pancreaticoduodenectomy for treatment of a duodenal carcinoma. (a) Regions of interest with associated attenuation values in Hounsfield units (HU) of the pancreas, spleen, liver and paraspinus muscle. (b) Measurement of periumbilical fat thickness in mm. (c) Measurement of hip girdle fat thickness in mm. (d) Measurement of retrorenal fat thickness in mm
Table 3.
Univariate analysis of computed tomography measurements, intraoperative pancreatic character and 90-day complications
| Variable | Complication (Clavien class 1–5) |
P-value | |
|---|---|---|---|
| Yes | No | ||
| Mean ± SD | Mean ± SD | ||
| Attenuation | |||
| Paraspinus muscle | 39.2 ± 12.7 | 45.4 ± 10.9 | < 0.01a |
| Spleen | 42.1 ± 9.7 | 45.9 ± 9.9 | 0.03a |
| Liver | 48.8 ± 12.7 | 53.4 ± 9.1 | 0.01a |
| Pancreas | 27.9 ± 13.7 | 31.7 ± 14.3 | 0.13 |
| Fat thickness, mm | |||
| Periumbilical | 27.6 ± 18.1 | 20.9 ± 10.7 | < 0.01 |
| Hip girdle | 49.8 ± 17.0 | 46.0 ± 16.9 | 0.22 |
| Retrorenal | 10.9 ± 8.8 | 8.1 ± 8.2 | 0.06 |
| Pancreatic character | |||
| Firmnessb | 0.45 ± 0.22 | 0.54 ± 0.24 | < 0.01 |
| Duct diameter, mm | 3.72 ± 1.64 | 3.99 ± 1.77 | 0.19 |
Remained significant (P < 0.05) in multivariate analysis controlling for age, length of stay, transfusion, periumbilical fat thickness and pancreatic firmness.
Firmness = x/10 where x is a numerical value on a scale of 1–10, on which 10 = a hard, very firm texture and 1 = a very soft texture.
SD, standard deviation.
Table 4 shows that a BMI of > 25 kg/m2 (P = 0.05), decreased pancreatic firmness (P < 0.01) and decreased pancreatic duct diameter (P < 0.01) were associated with an increased rate of POPF in univariate analysis. Visceral attenuation and fat thickness did not correlate with POPF (P < 0.05) in the present cohort.
Table 4.
Univariate analysis of patient, operative and computed tomography characteristics and postoperative pancreatic fistula (POPF)
| POPF |
P-value | ||
|---|---|---|---|
| Yes | No | ||
| Patient characteristics, % | |||
| Body mass index > 25 kg/m2 | 68% | 55% | 0.05 |
| Age > 65 years | 54% | 56% | 0.75 |
| Female | 59% | 50% | 0.15 |
| Smoking | 33% | 36% | 0.70 |
| Operative characteristics | |||
| OR time, h, mean | 3.5 | 3.6 | 0.42 |
| EBL > 250 ml, % | 26% | 74% | 0.68 |
| Transfusion, % | 15% | 85% | 0.29 |
| Attenuation, mean ± SD | |||
| Pancreas | 30.6 ± 13.3 | 29.0 ± 14.0 | 0.60 |
| Liver | 49.3 ± 11.8 | 51.2 ± 11.3 | 0.43 |
| Spleen | 40.9 ± 8.6 | 44.5 ± 10.0 | 0.07 |
| Paraspinus muscle | 38.0 ± 12.4 | 42.9 ± 12.3 | 0.07 |
| Hip girdle | 52.9 ± 17.8 | 47.4 ± 16.5 | 0.11 |
| Periumbilical | 26.8 ± 11.3 | 24.8 ± 17.2 | 0.46 |
| Retrorenal | 8.9 ± 8.4 | 10.3 ± 8.8 | 0.45 |
| Pancreas characteristics | |||
| Firmnessa, mean ± SD | 0.32 ± 0.28 | 0.52 ± 0.49 | < 0.01 |
| Duct diameter, mm, mean ± SD | 3.1 ± 1.4 | 3.9 ± 1.7 | < 0.01 |
Firmness = x/10 where x is a numerical value on a scale of 1–10, on which 10 = a hard, very firm texture and 1 = a very soft texture.
SD, standard deviation; OR, operating room; EBL, estimated blood loss.
Multivariate analysis
The regression model used in the present study controlled for age, LoS, transfusion, periumbilical fat thickness and pancreatic firmness. Decreased splenic [odds ratio (OR) 0.95, 95% confidence interval (CI) 0.90–1.00], paraspinus muscle (OR 0.95, 95% CI 0.91–0.99) and hepatic (OR 0.95, 95% CI 0.91–0.99) attenuation independently predicted postoperative complications following PD. Pancreatic attenuation did not predict complications in this regression model (OR 0.98, 95% CI 0.95–1.01). When the regression model controlled for BMI, visceral attenuation did not predict postoperative complications.
Next, the study sought to determine whether visceral attenuation was associated with specific complications following PD. Findings showed that decreased attenuation of paraspinus muscle predicted postoperative pneumonia (OR 0.87, 95% CI 0.79–0.98). The mean ± standard deviation (SD) attenuation of paraspinus muscle was 24.8 ± 8.1 in patients who experienced postoperative pneumonia compared with 42.2 ± 12.1 in those who did not (P = 0.005). The study also found that liver attenuation predicted wound infection (OR 0.91, 95% CI 0.86–0.98) and intra-abdominal infection or abscess (OR 1.1, 95% CI 1.0–1.2). The mean ± SD attenuation of the liver was 40.1 ± 12.0 and 51.4 ± 11.2, respectively, in patients with and without a wound infection (P = 0.004). The mean ± SD attenuation of the liver was 62.1 ± 16.3 and 50.3 ± 11.29, respectively, in patients with and without an abdominal infection (P = 0.04).
Discussion
This study reports that in patients presenting with non-enhanced preoperative CT, attenuation of viscera is an independent predictor of complications following PD. The present study is one of the few to assess CT attenuation and fat thickness. Interestingly, it demonstrates that, in the present cohort of patients, decreased attenuation of organs not involved in PD (spleen, liver, paraspinus muscle) are predictive of complications. This finding may suggest that overall decreased body density, or perhaps increased adiposity, aids in the identification of those at high risk for complications following PD.
The question of what the decreased attenuation of ‘normal’ viscera indicates remains. Computed tomography differs from traditional radiography in that it displays various X-ray attenuations through a section of the body tissue on a grey scale. X-ray attenuation is measured in HU. Hounsfield units represent a measure of tissue density on non-enhanced images or a combination of density and vascularity following i.v. contrast enhancement with iodinated contrast material.16 By convention, an HU value of 0 describes the attenuation coefficient of water, a value of − 1000 describes that of air, and a value of + 1000 describes that of bone. Thus the spectrum of HU can be used to evaluate the character of tissues in terms of density. Tissue with an increased HU value consists of more collagen, fibrosis and/or bone. By contrast, tissue with a lower HU value consists of more air, fat and/or water. In the present series, decreased attenuation is likely to relate to an increase in adiposity and a decrease in fibrosis. However, this is speculative because this study did not attempt to correlate pathologic specimens of spleen, hepatic or paraspinus muscle tissue from PD patients; this represents a shortcoming of the study. However, when multivariate analysis was controlled for BMI, the attenuation of tissues did not predict complications. This implies that the pathophysiology of adiposity played an important role in the present cohort. Mechanistic validation by pathologic review or serum reactive protein analysis is required.
Body mass index and obesity have been shown to impact on overall survival and complication rates following both operations and non-surgical medical care.17,18 The reasons for these impacts are not entirely understood, but may relate to an imbalance of proinflammatory and anti-inflammatory cytokines in obese subjects.19 Obese patients have impaired natural killer cell, T cell and neutrophil function, as well as an impaired response to stress as a result of imbalances in tumour necrosis factor-α, interleukin-6 and macrophage function.20,21 Moreover, C-reactive protein, amyloid A, fibrinogen and plasminogen activator inhibitor-1 are elevated in obese patients.22–25 Overall, obese patients appear to suffer from a chronic stress physiology that may contribute to the occurrence of postoperative complications. The present group is in the process of examining preoperative reactive proteins, markers of nutrition and inflammation to better understand the present CT findings and patient outcomes after PD. Likewise, prospective studies with multiple radiologic reviewers will be required to determine whether this group's findings can be generalized and used to aid in preoperative risk stratification and patient education in routine practice.
Previous work from the Mayo Clinic showed that pancreatic enhancement on delayed i.v. contrasted CT related to pancreatic fibrosis (attributable to the ‘wash in’ of the extracellular fluid distribution of iodinated contrast media) and a decreased rate of POPF.11 In this regard, pancreatic fibrosis appears to be better evaluated in contrast-enhanced studies. The protocol for enhanced CT varies among institutions. A fair number of the present patients presented with preoperative imaging sourced elsewhere. Thus, the present authors were unable to control the protocol of these imaging studies. Moreover, enhanced studies are limited by patient characteristics such as cardiac output. In addition, the ‘wash in’ and ‘wash out’ of contrast vary depending on the pathology of tissue and according to local blood flow mechanics. Therefore, the present study did not analyse findings in patients with contrasted studies because this study would have been underpowered to make any conclusions in contexts that included so many potential variables.
Previous reports show that surgeon-evaluated pancreatic firmness is predictive of postoperative complications, specifically POPF.26 The present data corroborate these findings. The most notorious complication following PD is POPF. Investigators suggest that POPF is caused by deceased pancreatic firmness or fibrosis or decreased duct size.27,28 Importantly, the present data show that POPF, occurring in 21% of patients, was significant in only a minority of them. Therefore, of all the complications, only a small percentage of those that were clinically significant were related to POPF. This may explain why pancreatic attenuation did not independently predict complications following PD. Although visceral attenuation did not predict POPF, attenuation did predict other complications such as pneumonia and wound infection.
In conclusion, decreased CT attenuation of viscera is an independent predictor of complications following PD. A combination of preoperative patient factors and CT attenuation characteristics may help to better identify patients at high risk for complications following PD. That said, a prospective data collection combined with tissue and serum validation is required to determine the pathophysiology of decreased attenuation in patients undergoing PD.
Conflicts of interest
None declared.
References
- 1.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. 2011;16:89–102. doi: 10.1007/s11605-011-1753-x. [DOI] [PubMed] [Google Scholar]
- 2.Pratt W, Joseph S, Callery MP, Vollmer CM., Jr POSSUM accurately predicts morbidity for pancreatic resection. Surgery. 2008;143:8–19. doi: 10.1016/j.surg.2007.07.035. [DOI] [PubMed] [Google Scholar]
- 3.Noun R, Riachy E, Ghorra C, Yazbeck T, Tohme C, Abboud B, et al. The impact of obesity on surgical outcome after pancreaticoduodenectomy. JOP. 2008;9:468–476. [PubMed] [Google Scholar]
- 4.Charlson ME, Sax FL, MacKenzie CR, Fields SD, Braham RL, Douglas RG., Jr Resuscitation: how do we decide? A prospective study of physicians' preferences and the clinical course of hospitalized patients. JAMA. 1986;255:1316–1322. doi: 10.1001/jama.255.10.1316. [DOI] [PubMed] [Google Scholar]
- 5.Hutter MM, Rowell KS, Devaney LA, Sokal SM, Warshaw AL, Abbott WM, et al. Identification of surgical complications and deaths: an assessment of the traditional surgical morbidity and mortality conference compared with the American College of Surgeons–National Surgical Quality Improvement Program. J Am Coll Surg. 2006;203:618–624. doi: 10.1016/j.jamcollsurg.2006.07.010. [DOI] [PubMed] [Google Scholar]
- 6.Parikh P, Shiloach M, Cohen ME, Bilimoria KY, Ko CY, Hall BL, et al. Pancreatectomy risk calculator: an ACS-NSQIP resource. HPB. 2010;12:488–497. doi: 10.1111/j.1477-2574.2010.00216.x. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 7.Hill JS, Zhou Z, Simons JP, Ng SC, McDade TP, Whalen GF, et al. A simple risk score to predict in-hospital mortality after pancreatic resection for cancer. Ann Surg Oncol. 2010;17:1802–1807. doi: 10.1245/s10434-010-0947-x. [DOI] [PubMed] [Google Scholar]
- 8.Williams TK, Rosato EL, Kennedy EP, Chojnacki KA, Andrel J, Hyslop T, et al. Impact of obesity on perioperative morbidity and mortality after pancreaticoduodenectomy. J Am Coll Surg. 2009;208:210–217. doi: 10.1016/j.jamcollsurg.2008.10.019. [DOI] [PubMed] [Google Scholar]
- 9.Hirose K, Shore AD, Wick EC, Weiner JP, Makary MA. Pay for obesity? Pay-for-performance metrics neglect increased complication rates and cost for obese patients. J Gastrointest Surg. 2011;15:1128–1135. doi: 10.1007/s11605-011-1529-3. [DOI] [PubMed] [Google Scholar]
- 10.House MG, Fong Y, Arnaoutakis DJ, Sharma R, Winston CB, Protic M, et al. Preoperative predictors for complications after pancreaticoduodenectomy: impact of BMI and body fat distribution. J Gastrointest Surg. 2008;12:270–278. doi: 10.1007/s11605-007-0421-7. [DOI] [PubMed] [Google Scholar]
- 11.Hashimoto Y, Sclabas GM, Takahashi N, Kirihara Y, Smyrk TC, Huebner M, et al. Dual-phase computed tomography for assessment of pancreatic fibrosis and anastomotic failure risk following pancreatoduodenectomy. J Gastrointest Surg. 2011;15:2193–2204. doi: 10.1007/s11605-011-1687-3. [DOI] [PubMed] [Google Scholar]
- 12.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]
- 13.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]
- 14.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]
- 15.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]
- 16.Choi H. Critical issues in response evaluation on computed tomography: lessons from the gastrointestinal stromal tumour model. Curr Oncol Rep. 2005;7:307–311. doi: 10.1007/s11912-005-0055-4. [DOI] [PubMed] [Google Scholar]
- 17.Berrington de Gonzalez A, Hartge P, Cerhan JR, Flint AJ, Hannan L, MacInnis RJ, et al. Body mass index and mortality among 1.46 million white adults. N Engl J Med. 2010;363:2211–2219. doi: 10.1056/NEJMoa1000367. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 18.Turrentine FE, Hanks JB, Schirmer BD, Stukenborg GJ. The relationship between body mass index and 30-day mortality risk, by principal surgical procedure. Arch Surg. 2011;147:236–242. doi: 10.1001/archsurg.2011.310. [DOI] [PubMed] [Google Scholar]
- 19.Karalis KP, Giannogonas P, Kodela E, Koutmani Y, Zoumakis M, Teli T. Mechanisms of obesity and related pathology: linking immune responses to metabolic stress. FEBS J. 276:5747–5754. doi: 10.1111/j.1742-4658.2009.07304.x. [DOI] [PubMed] [Google Scholar]
- 20.Nieman DC, Nehlsen-Cannarella SI, Henson DA, Butterworth DE, Fagoaga OR, Warren BJ, et al. Immune response to obesity and moderate weight loss. Int J Obes Relat Metab Disord. 1996;20:353–360. [PubMed] [Google Scholar]
- 21.Womack J, Tien PC, Feldman J, Shin JH, Fennie K, Anastos K, et al. Obesity and immune cell counts in women. Metabolism. 2007;56:998–1004. doi: 10.1016/j.metabol.2007.03.008. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 22.Hotamisligil GS. Inflammation and metabolic disorders. Nature. 2006;444:860–867. doi: 10.1038/nature05485. [DOI] [PubMed] [Google Scholar]
- 23.Rotter V, Nagaev I, Smith U. Interleukin-6 (IL-6) induces insulin resistance in 3T3-L1 adipocytes and is, like IL-8 and tumour necrosis factor-alpha, overexpressed in human fat cells from insulin-resistant subjects. J Biol Chem. 2003;278:45777–45784. doi: 10.1074/jbc.M301977200. [DOI] [PubMed] [Google Scholar]
- 24.Uysal KT, Wiesbrock SM, Marino MW, Hotamisligil GS. Protection from obesity-induced insulin resistance in mice lacking TNF-alpha function. Nature. 1997;389:610–614. doi: 10.1038/39335. [DOI] [PubMed] [Google Scholar]
- 25.Weisberg SP, Hunter D, Huber R, Lemieux J, Slaymaker S, Vaddi K, et al. CCR2 modulates inflammatory and metabolic effects of high-fat feeding. J Clin Invest. 2006;116:115–124. doi: 10.1172/JCI24335. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 26.Mathur A, Pitt HA, Marine M, Saxena R, Schmidt CM, Howard TJ, et al. Fatty pancreas: a factor in postoperative pancreatic fistula. Ann Surg. 2007;246:1058–1064. doi: 10.1097/SLA.0b013e31814a6906. [DOI] [PubMed] [Google Scholar]
- 27.Shrikhande SV, D'Souza MA. Pancreatic fistula after pancreatectomy: evolving definitions, preventive strategies and modern management. World J Gastroenterol. 2008;14:5789–5796. doi: 10.3748/wjg.14.5789. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 28.Ramacciato G, Mercantini P, Petrucciani N, Nigri GR, Kazemi A, Muroni M, et al. Risk factors of pancreatic fistula after pancreaticoduodenectomy: a collective review. Am Surg. 2011;77:257–269. [PubMed] [Google Scholar]