ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS 229 lecystectomy during index admission (during which pancreatitis was diagnosed, n⫽162). Group B patients underwent cholecystectomy following discharge from index admission (n⫽119). Incidence (and timing) of gallstone-related events including recurrent pancreatitis, total length of hospital stay (index admission ⫹ admissions for recurrences and cholecystectomy), and postoperative morbidity and mortality rates were analyzed. Results: Groups were comparable in demographic variables, comorbidity rates, and disease severity. Thirty-nine (32.8%) group B patients experienced gallstone-related events (including 16 cases of recurrent pancreatitis) following discharge from index admission but prior to cholecystectomy. 12.5% of recurrences occurred within 1 week, 37.5% occurred within 2 weeks and 50% occurred within 4 weeks after discharge. Endoscopic sphincterotomy (ES) protected against preoperative recurrent pancreatitis but was associated with higher rates of other biliary events. Median total length of hospital stay was greater for group B than for group A (7 [range, 2-37] days vs. 5 [1-45] days, p⫽0.00). Postoperative recurrence and reoperation were more frequent for group B than for group A (10.1% vs. 3.1%, p⫽0.02 and 0% vs.3.4%, p⫽0.02, respectively). There were no mortalities in either group. Conclusion: Current guidelines suggesting the appropriateness of waiting up to 2 weeks for cholecystectomy following recovery from biliary pancreatitis may place patients at unacceptably high risk for recurrence. ES does not eliminate the need for cholecystectomy in these patients. 120. ROLE OF INTRAOPERATIVE PARATHORMONE MONITORING DURING PARATHYROIDECTOMY IN PATIENTS WITH DISCORDANT LOCALIZATION STUDIES. John I. Lew, Carmen C. Solorzano, Raquel E. Montano, Denise M. Carneiro-Pla, George L. Irvin III; University of Miami Leonard M. Miller School of Medicine, Miami, FL Introduction: Focused parathyroidectomy with concordant preoperative Tc-99m-sestamibi (MIBI) scans and ultrasound localization without intraoperative parathormone monitoring (IPM) for sporadic primary hyperparathyroidism (SPHPT) can be performed at a high rate of success. A significant number of patients with SPHPT, however, present with discordant MIBI and ultrasound localizing tests prior to parathyroidectomy where operative failure is more likely. The purpose of this study is to examine the utility of IPM during focused parathyroidectomy in patients with discordant as well as concordant localization studies. Methods: A retrospective series of 225 consecutive patients with SPHPT had MIBI scans and surgeon performed ultrasound (SUS) prior to parathyroidectomy with IPM from 1999 to 2006 at a single institution. All operative cases were reviewed and the usefulness of IPM was determined. IPM significantly changed operative management when it allowed for: 1) unilateral neck exploration (UNE) instead of bilateral neck exploration (BNE) in patients with discordant studies; and 2) when multigland disease (MGD) was identified by IPM and missed by concordant or discordant studies. Patients were followed at least 6 months postoperatively. Correct parathyroid gland location, presence of MGD and operative outcome were also compared among patients with concordant and discordant imaging studies. Results: The operative success in the 225 consecutive patients was 97% overall. IPM changed the operative management in 29% (66/225) of patients. In the 85 (38%) patients who underwent focused parathyroidectomy with preoperative discordant localization studies, operative success was 93%. IPM changed the operative management in 74% (63/85) of these patients. IPM allowed for 66% (56/85) of these operations to be performed as UNE and confirmed removal of all hypersecreting glands in 78% (7/9) patients with MGD. In 140 (62%) patients with concordant preoperative localization studies, operative success was 99%. IPM changed operative management in only 2% (3/140) patients with MGD not correctly identified by the concordant studies. Of the entire group of 225 patients, there were 12 (5%) patients with MGD where IPM was helpful in successfully treating 10 of these patients. There were 7 (3%) operative failures in the entire series, 5 due to the inability to
find the gland and 2 from missed MGD. Conclusion: Although of marginal benefit in patients with concordant preoperative imaging localization studies, IPM remains essential for performing successful parathyroidectomy in patients with discordant or incorrect concordant MIBI and SUS localization. IPM allows for limited dissection and identifies MGD in the majority of these patients.
CLINICAL TRIALS/OUTCOMES RESEARCH III: BARIATRICS & PERI-OPERATIVE CARE 121. PROPRANOLOL IMPROVES INSULIN SENSITIVITY POSTBURN. Marc G. Jeschke, Gabriela Kulp, William Norbury, David N. Herndon; University of Texas Medical Branch, Galveston, TX Background: Catecholamine induced hyperglycemia and insulin resistance occur after severe burn injury and are associated with increased catabolism, infections, and mortality. Insulin administration decreases blood glucose levels but causes hypoglycemic episodes in as many as 20% of treated burn patients. We hypothesized that attenuation of catecholamine induced stress response using propranolol, a non-selective 1/2 adrenergic receptor antagonist would improve insulin sensitivity in severely burned pediatric patients. Patients and Methods: A prospective, double-blind, randomized, intent to treat trial was performed: demographics (age, gender, burn size, and mortality), daily average blood glucose levels, daily 6 a.m. blood glucose levels, number of patients requiring insulin, and the amount of daily insulin requirements during acute hospitalization were determined. Propranolol was administered to decrease heart rate by 15-20%, approximately 2 mg/kg/day. Similarly packaged placebo was delivered to the control group. Results: Two hundred and thirty-two patients (154 controls, 78 propranolol) were included into the study. There were no differences between the control and propranolol group for age, gender distribution, burn size, third degree burn and length of stay. Mortality was 8% in the control group and 3% in the propranolol group. Average blood glucose levels during acute hospitalization were 136⫾2mg/dl in the control group and 133⫾2 mg/dl in the propranolol group. While 70 control patients (46%) required insulin administration only 18 patients (23%) required insulin administration in the propranolol group, p⬍0.001. The amount of insulin administered throughout acute hospital stay in the control group was on average 555⫾60 IU per patient while it was on average 400⫾40 IU in the propranolol treated group, p⬍0.01. Endogenous insulin levels were not significantly different between groups. Conclusion: Based on our data we suggest that propranolol treatment markedly improves stress induced insulin resistance in severely burned children. 122. REAL MONEY: COMPLICATIONS AND HOSPITAL COSTS IN TRAUMA PATIENTS. Mark R. Hemmila1, Jill L. Jakubus1, Paul M. Maggio1, Wendy L. Wahl1, Justin B. Dimick1, Darrell A. Campbell, Jr.1, Paul A. Taheri2; 1University of Michigan, Ann Arbor, MI; 2University of Vermont College of Medicine, Burlington, VT Introduction: Major postoperative complications are associated with a substantial increase in hospital costs. Trauma patients are known to have a higher rate of complications than the general surgery population. We sought to determine the hospital costs and reimbursement attributable to complications in trauma patients. Methods: Using Methods and definitions from the National Surgical Quality Improvement Program (NSQIP), data were collected on all adult patients admitted to the trauma service for ⬎ 24 hours, with an injury severity score (ISS) ⬎ or ⫽ 5, during a 1-year period 8/20049/2005. Patients who died prior to hospital bed admission were excluded. Cost data were obtained from the hospital’s internal costaccounting database (TSI; Transitions Systems Inc.). Patients were placed in one of three groups: no complications (none), ⬎ or ⫽ 1
230 ASSOCIATION FOR ACADEMIC SURGERY AND SOCIETY OF UNIVERSITY SURGEONS—ABSTRACTS minor complication (minor, e.g. urinary tract infection), or ⬎ or ⫽ 1 major complication (major, e.g. pneumonia). Total hospital charges, costs, and revenue associated with each complication group were calculated. We compared costs and reimbursement for patients with and without complications. We adjusted our analysis for differences in age, gender, ISS, and Glasgow coma scale score (GCS). Results: A total of 512 patients were studied. 330 (64%) patients had no complications, 53 (10%) patients had ⬎ or ⫽ 1 minor complication, and 129 (25%) patients had ⬎ or ⫽ 1 major complication. Trauma patients who suffered a major or minor complication were older and had a higher injury burden (Table 1). Median hospital charges increased from $33,833 for patients without complications to $81,936 for patients with minor complications and $150,885 for patients with major complications. After adjusting for patient risk, minor complications were associated with a $20,019 increase in costs (p⬍0.001), and major complications were associated with a $43,890 increase in costs (p⬍0.001). Despite higher costs, the patients in the ⬎ or ⫽ 1 major complication group generated a higher mean contribution to margin per day when compared to patients without complications ($2,168, p⬍0.001). Conclusions: Understanding the costs and margins associated with traumatic injury provides a window for assessing the locus of potential cost reductions associated with improved quality care. Since the current reimbursement system provides incremental reimbursement for complications, there is little financial incentive for institutions to invest in proactive quality initiatives. To optimize system benefits, payers and providers should develop integrated reimbursement methodologies which align incentives to provide quality care.
TABLE 1
Parameter
N Mean Age, y Mean Injury Severity Score Mean Glasgow Coma Scale Score Mortality, % Median Length of Stay, d (IQR) Median Hospital Charges, $ (IQR) Mean Total Margin, $ (95% CI) Mean Contribution to Margin per Day, $ (95% CI)
No Complications
330 40 17
13.5
⬎ or ⫽ 1 Minor Complication
53 48 23
12.6
p-value
0.002 ⬍0.001
⬎ or ⫽ 1 Major Complication
129 47 28
p-value
⬍0.001 ⬍0.001
0.1
10.0
⬍0.001
3.6% 5 (3–8)
0% 9 (6–13)
0.2 ⬍0.001
14.7% 17 (9–26)
⬍0.001 ⬍0.001
$33,833 (20,793–60,582)
$81,936 (62,383–126,093)
⬍0.001
$150,885 (85,760–248,567)
⬍0.001
$5,073 (3,331–6,816)
$7,505 (1,015–13,994)
0.3
$32,884 (22,502–43,265)
⬍0.001
$994 (650–1,339)
$1,155 (420–1,890)
0.7
$2,168 (1,541–2,794)
⬍0.001
123. CANCER CARE IN THE PEDIATRIC SURGICAL PATIENT: A PARADIGM TO ABOLISH VOLUME-OUTCOME DISPARITIES IN SURGERY. Juan C. Gutierrez1, Leonidas G. Koniaris1, Anne C. Fischer2, Juan E. Sola1; 1University of Miami Miller School of Medicine, Miami, FL; 2UT Southwestern/Children’s Medical Center, Dallas, TX Objective: Extensive data published to date has demonstrated improved patient outcomes for adult cancer patients treated at high volume facilities. Unlike adult cancer care, however, the pediatric surgical patient is treated by a select group of experts at a limited number of facilities with greater standardization of protocols and centralized oversight. We sought to define the prognostic significance of hospital surgical volume on outcomes for two common pediatric
surgical malignancies, neuroblastoma and Wilms tumor. Methods: The Florida Cancer Data System (FCDS) was examined for all patients less than 18 years of age diagnosed with neuroblastoma or Wilms tumor and treated between 1981 and 2004. Medical facilities were ranked according to their operative case volume for each respective disease - those that cumulatively treated greater than approximately 50% of patients were classified as high volume centers (HVC). Results: Of the 869 cases of neuroblastoma identified, 463 patients were treated at 5 HVC and 406 were treated at 61 low volume centers (LVC). There were no differences in gender, age at diagnosis, race, ethnicity or stage of disease between the two groups. A larger proportion of adrenal tumors were treated at HVC (67.0% vs. 60.4% at LVC, P ⫽ 0.005) while LVC treated a larger fraction of cervical tumors (12.6% vs. 5.5% at HVC, P ⫽ 0.005). While there was no difference in the utilization of radiation therapy between groups, chemotherapy was more frequently administered at HVC than at LVC (74.7% vs. 60.3%, P ⬍ 0.001). Five and ten-year survival rates were identical between treatment groups (70.6% and 67.7% at HVC vs. 69.3% and 65.2% at LVC, P ⫽ 0.243). Multivariate analysis identified age at diagnosis and tumor stage as independent prognostic factors. Of the 790 cases of Wilms tumor identified, 395 patients were treated at 5 HVC and 395 were treated at 50 LVC. There were no differences in gender, age of diagnosis, or stage of disease between the two groups. Patients treated at HVC were more frequently identified as being Caucasian (72.4% vs. 68.9% at LVC, P ⫽ 0.049) or of Hispanic origin (21.6% vs. 10.7%, P ⬍ 0.001). Radiation therapy was more frequently administered at LVC than at HVC (39.5% vs. 32.7%, P ⫽ 0.049), while chemotherapy was more frequently administered at HVC than at LVC (87.8% vs. 74.9%, P ⬍ 0.001). Five and ten-year survival rates were identical between treatment groups (91.3% and 89.9% at HVC vs. 89.7% and 88.5% at LVC, P ⫽ 0.698). Multivariate analysis identified ethnicity, tumor stage and use of chemotherapy as independent prognostic factors. Conclusion: Survival rates for patients with neuroblastoma and Wilms tumor are unrelated to the surgical volume of the treatment hospital or patient race. This result stands in stark contrast to a variety of adult malignancies, including carcinomas of the lung, pancreas, colon and esophagus as well as soft tissue sarcomas, in which significant survival improvements are observed at HVC. Furthermore, significant racial disparities are observed in many adult patients with these malignancies. The mechanisms leading to more equivalent treatment of children are under investigation. We speculate that the lack of volume-outcome disparities in pediatric cancer may be related to greater use of standardized protocols by more specialized health care providers, akin to adult care in cancer centers, while the absence of racial disparities might be due to greater equality of access to care. 124. LONG-TERM RESULTS OF LAPAROSCOPIC GASTRIC BANDING: SIX YEAR EXPERIENCE OF A BARIATRIC CENTER OF EXCELLENCE. Rabih Nemr, Joelle Pierre, Rosemarie Hardin, Kevin Brenowitz, Michelle Avitable, George Ferzli; SUNY Downstate Medical Center, Brooklyn, NY Introduction: Morbid obesity is a disease whose prevalence in the United States is on the rise. It is associated with multiple co morbidities including but not limited to hypertension and diabetes. Currently surgical therapy has proven to be the most effective treatment for morbid obesity. Of the procedures that are available, the laparoscopic gastric band (LAGB) is considered by most to be the least invasive. However it requires intensive follow-up. We propose that the expected success of the LAGB has been over shadowed by the high rate of complications and lower than expected patient satisfaction. Methods: Between 2001 and 2007 307 LAGB procedures were performed. Patient’s charts were reviewed and they were contacted via telephone and surveyed. 189 patients were included in the study. Of the remaining 118, four declined to participate. The remaining 114 were not reached for follow up, some secondary to