47.07 The Economic Burden Of Care For Severe Work Related Injuries In A Level-One Trauma Referral Centre

Posted on December 22, 2014

C. T. Robertson-More1, B. Wells1,2, D. Nickerson3, A. Kirkpatrick1,2, C. Ball1,2  1University Of Calgary,General Surgery,Calgary, AB, Canada 2University Of Calgary,Trauma Surgery,Calgary, AB, Canada 3University Of Calgary,Plastic Surgery,Calgary, AB, Canada

Introduction: Work-related injuries (WRI) are common and represent a significant logistical and economic burden to health care systems. It is also possible that insurers and/or public health care systems do not account for the potentially higher cost of caring for these patients when compared to patients with non-WRI (NWRI). The primary aim of this study was to evaluate the demographics, volume, costs and outcomes associated with WRI at a high volume trauma center.

Methods: The Alberta Trauma Registry and clinical information system were used to perform a retrospective cohort study describing all patients with severe WRI (ISS>12) admitted to a high volume, tertiary care trauma referral center between April, 1995 and March, 2013. Patients who died within the emergency department were excluded. Standard statistical methodology was utilized (p<0.05).

Results: Of 14,964 total trauma admissions, 1,270 (8.5%) were for severe WRI. Overall, the patients’ mean age was 45 years with a male to female ratio of 2.8:1 and mean Injury Severity Score (ISS) of 22.7. Blunt (94%), penetrating (4%), and burn (2%) injury mechanisms were observed. Compared to patients with NWRI, the WRI group was significantly younger (41 vs. 46 years, 95% CI: -5.7 to -3.9yrs), typically male (94% vs. 72%, p<0.05), and had fewer pre-injury comorbidities (p<0.05). Although they displayed statistically equivalent ISS, the WRI group had a greater length of stay in the intensive care unit (2.8 vs. 2.3 days, 95% CI: 0.06 to 0.86 days), length of mechanical ventilation (2.2 vs. 1.8 days, 95% CI: 0.08 to 0.68 days), and mean number of surgical/operative procedures (0.86 vs. 0.67 per patient, 95% CI: 0.11 to 0.27). In contrast, significantly fewer patients with WRI died while in hospital (8% vs. 12%, p<0.05). Consequently, more patients with WRI were discharged home without support services (62% vs. 57%, p<0.05) and significantly fewer were transferred to long-term care facilities (0.5 vs. 1%, p<0.05). The acute care economic burden of patients with WRI was significantly higher (p<0.05). Increased costs were related to the care of these patients in the intensive care unit (p<0.05) and operating theatre (p<0.05), as well as for physician compensation (p<0.05).

Conclusion: Patients with WRI admitted to our trauma center were younger, less comorbid, more likely male and had a significantly higher utilization of acute care resources despite a similar ISS when compared to those with NWRI. These increased costs and economic burden in critical care, operative and physician based services are not recovered from work place insurers in public health care systems.

47.08 Successful Observation of Small Traumatic Pneumothoraces in Patients Requiring Aeromedical Transfer

Posted on December 22, 2014

N. Lu1, C. Ursic1, H. Penney1,2, S. Steinemann1, S. Moran1  1University Of Hawaii,Department Of Surgery,Honolulu, HI, USA 2University Of Hawaii,Department Of Radiology,Honolulu, HI, USA

Introduction:  With the widespread use of computed tomography (CT) imaging, the occult pneumothorax (PTX) has become a common finding. It has been shown that it is safe to monitor occult PTX in stable patients, even if they are on positive pressure ventilation. Observation of occult PTX without chest tube placement has been supported for those seen on CT to be <7mm measured perpendicular from lung to chest wall. However, patients transported by air are not optimally monitored and not in the care of practitioners skilled in thoracostomy tube placement.

Methods:  We undertook a retrospective chart review of patients with traumatic PTX who were transported by air over the course of three years (2010-2012) to a level II trauma center that serves 1.3 million people. Occult PTX was defined as a pneumothorax that was not visible on chest radiograph (CXR), but was visible on CT imaging. Patients who did not have an overt PTX or a clinical reason for immediate chest tube placement were divided into two groups: those with PTX<7mm and those with PTX>7 mm.

Results: From 2010 to 2012, 66 patients were transferred with a total of 83 PTX. Eleven PTX in 8 patients were treated with chest tubes placed for clinical reasons such as CPR or needle decompression in the field. For 11 PTX, we have no information about pre-transport CXR or were unable to measure the PTX on CT. Eleven overt PTX were treated with thoracostomy tubes. Of the 10 large occult (>7mm) PTX, 8 were treated with thoracostomy tubes and two were treated with observation in transport. Of the 39 small (<7mm) PTX, 19 were treated with thoracostomy tubes (15 ventilated, 4 not ventilated); and 20 were observed during transportation (5 ventilated, 15 not ventilated). Of all patients without thoracostomy tubes prior to transport, 3 were placed on arrival. One was placed in a patient whose repeat CXR showed the PTX (no longer occult), though the patient was stable.  One was placed in a patient whose follow up CT showed expansion to 8mm and who was to be intubated for an operation. One was placed in a patient with a pre-transfer PTX>7 mm and with copious subcutaneous emphysema which expanded en route. There were 15 total complications. Thirteen were malpositioned and two were related to empyema requiring thoracoscopic drainage.

Conclusion: Patients with small PTX can safely be transported by air without thoracostomy tubes. Only one of 20 patients sent without a chest tube required immediate chest tube placement and, in retrospect, it would have been recommended that a tube be placed prior to transport due to the size of the PTX and the amount of subcutaneous air. Mechanical ventilation prompted more thoracostomy tube placements.  In addition, observation may reduce complications from chest tube placement (malposition, infection, increased number of CXR, increase in hospital length of stay, and delay in returning home). Further studies with large numbers of patients are warranted.

47.09 Unplanned Intensive Care Unit Admissions Following Trauma

Posted on December 22, 2014

J. A. Rubano1, J. A. Vosswinkel1, J. E. McCormack1, E. C. Huang1, M. Paccione1, R. S. Jawa1  1Stony Brook University Medical Center,Trauma,Stony Brook, NY, USA

Introduction: Unplanned Intensive Care Unit (UP-ICU) admission is a key quality measure of the American College of Surgeons Committe on Trauma. We sought to evaluate frequency, timing, risk factors, and morbidity associated with unplanned ICU admission following acute traumatic injury.

Methods: Retrospective analysis of a state-designated level I trauma center's registry.  All adult trauma admissions from January 2007 through December 2013 were considered.  Burns, isolated hip fractures, field/emergency department intubations and patients takend directly to the operating room were excluded.  Univariate and multivariate statistical analyses were performed; p≤ 0.05 was considered significant.

Results: Of 5465 patients meeting study criteria, 85.2% required no ICU (NO-ICU) stay, 10.9% had planned (PL-ICU) admission, and 3.9% were UP-ICU admissions.  Patient demographics are presented in the table.  UP-ICU admissions more frequently had ≥2 National Trauma Data Standard comorbid conditions (65.1%) than NO-ICU (33.2%) and PL-ICU admissions (47.2%), p<0.05. Median length of stay prior to UP-ICU admission was significantly longer than PL-ICU admission (2 days, IQR 0-4 vs. 0 days, IQR 0-0).  UP-ICU admissions had significantly more frequent strokes (2.4% vs 0.5%), MI (14.2% vs. 4.0%), respiratory failure (10.9% vs. 1.7%), pneumonia (30.2% vs. 9.9%), renal failure (7.6% vs. 2.7%), sepsis (10.9% vs. 2.9%), and DVT/PE (11,8% vs. 5.2%) as compared to PL-ICU admissions.  Rates of these complications in the NO-ICU group were each ≤1.1% and correspondingly significantly less than in UP-ICU group.  Finally, UP-ICU patients had a higher mortality (18.4%) than NO-ICU (0.49%, p<0.001) or PL-ICU admission groups (5.71%, p < 0.001).  In subsequent multivariate logistic regression, risk factors for unplanned ICU admission were respiratory failure (odds ratio 3.74, 95% confidence interval 1.62-8.63), PE/DVT (2.27, 1.23-4.18), MI (1.98, 1.05-3.74), and pneumonia (2.60, 1.66-4.08).  Age, presence of ≥ 2 comorbidities, sepsis, and stroke were not risk factors. ISS was slightly negatively associated with UP-ICU admission (OR 0.97 (95% CI 0.95 – 0.99).

Conclusion: Unplanned ICU admission is an infrequent but morbid event. It is associated with a threefold increase in mortality as compared to planned ICU admission.  A slightly lower ISS in UP-ICU would be expected as these patients were not directly admitted to the ICU.  Earlier identification of risk factors may decrease unplanned ICU admission.
 

47.10 Analysis of the Coagulation System in Burn Patients: Perhaps Not As Simple As INR

Posted on December 22, 2014

S. Tejiram1, K. Brummel-Ziedins3, T. Orfeo3, S. Butenas3, B. Hamilton2, J. Marks2, L. Moffatt2, J. Shupp1,2  1MedStar Washington Hospital Center,The Burn Center, Department Of Surgery,Washington, DC, USA 2MedStar Health Research Institute,Firefighters’ Burn And Surgical Research Laboratory,Washington, DC, USA 3University Of Vermont,Department Of Biochemistry,Colchester, VT, USA

Introduction: While a body of literature exists on coagulopathy in trauma patients, understanding of abnormal coagulation in burn patients is limited. Studies have shown alterations in antithrombin, protein C and S levels after burn, but controversy remains over whether burn injury induces coagulopathy. There is no consensus on whether burn patients with variable injury severity are at risk for hyper- or hypocoagulation. Coagulation is a complex process that is frequently assessed only by laboratory values such as PT, PTT, and INR. These measurements do not account for clotting factor dynamics or clot characteristics. Real time assessment of a patient’s coagulation profile may help clinicians better understand the pathophysiology underlying abnormal coagulation in burn patients. Here, we monitored clotting factor levels in a pilot group of burn patients for 96 hours after admission to study potential perturbations in the coagulation system and help elucidate potentially meaningful dynamics in coagulation after burn injury not indicated by INR alone.

Methods: Nine thermally injured patients with total body surface area injuries of 25% or greater who presented to a verified burn center between 2013 and 2014 were included for analysis. Citrated plasma was collected at admission and at regular intervals over a 96 hour period. Clinical laboratory information, specifically PT, PTT, and INR, collected over the same time was compared to levels of factors II, IIa, V, VII, VIII, IX, IXa, X, XI, XIa, antithrombin, and tissue factor pathway inhibitor measured in plasma.

Results: Of the patients profiled, 4 died and 5 survived. Seven patients had factor VIII levels beyond the upper limit of normal range upon admission. Four of these had factor VIII levels 2-3 fold higher than normal. Over the subsequent 24 hours, all patients experienced an initial decrease of factor VIII levels to normal ranges before increasing again above the normal range. Factor IX was also elevated approximately 1.5 times normal levels upon admission in all patients and remained above normal range for all but 2 patients. Conversely, factor VII levels decreased below normal ranges for 3 patients after 24 hours. Only 4 patients had antithrombin levels in normal range upon admission and all patients had antithrombin levels below normal range shortly thereafter for the subsequent 96 hours. Three patients showed an increase in INR and PTT beyond normal range. Clinical laboratory values of INR and PTT remained within normal limits (INR < 1.3 and PTT 23-45s) for all other patients during this 96 hour period.

Conclusion: Dynamic changes in clotting factor levels follow immediately after thermal injury that may not be detected by monitoring of INR and PTT alone. These changes may be important in early identification of coagulopathy in this patient population, which to date is poorly characterized. Further study is warranted to explore the scope of abnormal coagulation in burn patients.

 

48.01 Blood Transfusion & Adverse Surgical Outcomes – the Good, the Bad, & the Ugly

Posted on December 22, 2014

M. Hochstetler1, S. P. Saha1, J. Martin1, A. Mahan1, V. Ferraris1  1University Of Kentucky Chandler Medical Center,Surgery,Lexington, KY, USA

Introduction:
Every experienced surgeon has a patient whose life was saved by a blood transfusion (the GOOD).  On the other hand, an overwhelming amount of evidence suggests that perioperative blood transfusion translates into adverse surgical outcomes (the BAD).  We wondered what patient characteristics, if any, can explain this clinical dichotomy with certain patients benefiting from transfusion while others are harmed by this intervention. 

Methods:
We queried the NSQIP database containing patient information entered between 2010 and 2012 in order to identify mortality and morbidity differences in patients receiving blood transfusion within 72 hours of their operative procedure compared to those who did not receive any blood.  We calculated the relative risk of developing a serious complication or of having operative mortality in propensity matched patients with equivalent risk of having a blood transfusion. 

Results:
There were 470,407 patients in the study group.  Of these, 32,953 patients (7.0%) received at least a single blood transfusion within 72 hours of operation.  The transfusion rate in patients having operative mortality or serious morbidity was 11.3% and 55.4% compared to the transfusion rate of 1.3% and 0% in survivors of operation without complications (both p < 0.001).  Dividing patients into deciles of increasing operative mortality risk or risk of serious morbidity found that patients at the highest risk for development of death or serious complications had non-significant risk of harm from blood transfusion, while patients in the lowest risk deciles had between 8 to 10 fold increased risk of major adverse events associated with transfusion (Figure). 

Conclusions:
We found that high risk patients do not have significant risk from blood transfusion, but the lowest risk patients have between an 8 and 10 fold excess risk of adverse outcomes when they receive a blood transfusion (the UGLY).  We speculate that careful preoperative assessment of transfusion risk, and intervention based on this assessment, could minimize operative morbidity and mortality, especially since the lowest risk patients are more likely to have elective operations and provide time for therapeutic interventions to improve risk profiles. 
 
 

48.02 Evaluation of the accuracy of endoscopic ultrasound for nodal staging in esophageal cancer

Posted on December 22, 2014

V. Bianco1, K. S. Mehta1, M. Sablowsky1, W. E. Gooding1, J. D. Luketich1, A. Pennathur1  1University Of Pittsburgh Medical Center,Department Of Cardiothoracic Surgery,Pittsburgh, PA, USA

Introduction:
The accurate staging of esophageal cancer is important for both prognostic and therapeutic decisions, as well as evaluation of results of treatment.  Of particular importance is the presence of nodal metastases which has a major impact on prognosis and therapeutic approach. Endoscopic ultrasound (EUS) is an important and increasingly used clinical staging modality for the pretreatment evaluation of nodal status in esophageal cancer.   Recently the new AJCC 7TH edition staging system, which incorporates the number of positive nodes in the nodal (N) staging has been adopted.  While the accuracy of EUS nodal staging has been well studied with the previous AJCC staging version, there are limited studies evaluating the accuracy of EUS in N staging using the recently revised esophageal cancer staging (AJCC 7th edition).  The objective of this study was to analyze the accuracy of EUS, in nodal staging of esophageal cancer using the most recent staging system (AJCC 7th edition).

Methods:
We reviewed the records of 172 patients who had undergone esophagectomy without neoadjuvant treatment and collected data which included both clinical and pathological staging.  The preoperative N stage was acquired from EUS and the pathological stage was assigned based on the esophagectomy specimen.  Staging data was recorded for each patient based on the specific criteria for the 7th edition.  The accuracy of identifying EUS in nodal disease was evaluated.

Results:
A total of 172 patients (mean age 67, 137 male, 35 female) underwent esophagectomy for esophageal cancer.  The median number of nodes resected was 20 per patient. For the AJCC 7th edition only 80 of 172 (46%) patients had the N stage correctly classified with EUS (Table).  We identified 92 of 172 (54%) patients who were misclassified (31% under staged, 23% over staged) for AJCC 7. 

Conclusion:
Our results indicate a substantial reduction in the EUS staging accuracy of nodal disease for the AJCC 7th edition.  These  findings suggest that further advancement is necessary for accurate clinical staging of nodal disease preoperatively.  Potential solutions may include refinement of EUS technology with routine use of EUS-FNA, evaluation of other staging modalities such as laparoscopic staging, and the use of  molecular staging with validated biomarkers.
 

48.03 Decade-long Trends of Survival and Cost for Extracorporeal Life Support: results from a modern series

Posted on December 22, 2014

E. B. Pillado1, R. Kashani1, H. Wu1, S. Grant1, C. Hershey1, R. Shemin1, P. Benharash1  1David Geffen School Of Medicine, University Of California At Los Angeles,Division Of Cardiac Surgery,Los Angeles, CA, USA

Introduction: Extracorporeal membrane oxygenation (ECMO) has been used to support patients with advanced cardiac and/or pulmonary failure. More recently, venoarterial (VA) ECMO has been used as an adjunct to CPR, which has increased the number of patients on extracorporeal support. With an increase in ECMO utilization worldwide and the need for a cost efficient healthcare system, the present study aimed to evaluate patient outcomes and hospital costs at our institution.

Methods: A retrospective review of the UCLA Health Extracorporeal Life Support Organization (ELSO) database was performed to identify adult patients who underwent VA-ECMO between 2004-2014. Our institutional Society of Thoracic Surgeons Database was used to extract the volume and type of adult cardiac surgeries, defined as patients having procedures requiring cardiopulmonary bypass as well as heart transplants during the same period. Publicly available cost data was obtained for our institution for ECMO services exclusive of bed cost. STATA 12.1 (College Station, TX) was used to run regression analysis on groups. 

Results:Out of 263 (33% female) patients who underwent venoarterial ECMO, 117 (44%) were weaned, 55 (21%) bridged to transplantation or mechanical assist device, and 91 (35%) expired while on ECMO. The average time on ECMO was 5.3±0.3 days and the mean age was 50.3 (±1.2). The procedural volume for ECMO showed an annual increase of 27%. Success of wean from ECMO showed a non-significant trend towards improvement over the study period (43% in 2004 to 69% in 2013, p=0.17). The average cost per patient was $36,669(±13,951) in 2004 and $32,776(±15,658) in 2014 (p=0.083). During the same period, there were also significant changes in the volume of cardiac transplants and total number of cases at our institution (total heart transplant, p=0.046, total cardiac surgeries, p=0.001, and VA-ECMO patients, p=0.016) (Figure 1).

Conclusion:We have demonstrated that there was a disproportionate increase in VA-ECMO volume when compared to our institutional volume for cardiac surgical and transplant procedures. With more widespread use of ECMO, the hospital costs have increased over the past decade while the cost per patient has remained relatively constant. This may be explained by having shorter periods on ECMO for each patient. The high institutional burden of ECMO and increasing volumes for the procedure mandate better selection criteria and ECMO protocols in order to maintain a cost-efficient healthcare system

 

48.04 Virtual HLA Crossmatching As A Means To Safely Expedite The Transplantation Of Shipped-in Pancreata

Posted on December 22, 2014

B. C. Eby1, T. M. Ellis2, R. R. Redfield1, G. Leverson1, J. S. Odorico1  1University Of Wisconsin,Division Of Transplantation/Department Of Surgery/School Of Medicine And Public Health,Madison, WI, USA 2University Of Wisconsin,Department Of Pathology And Laboratory Medicine,Madison, WI, USA

Introduction:  Cold ischemia time (CIT) accumulates with shipped-in pancreata and limits utilization and outcomes. Flow cytometric HLA crossmatching (FXM) is used to assess histocompatibility between pancreas and recipient before transplantation. Waiting for a FXM, which is typically done once the organ arrives, prolongs the CIT. A ‘virtual crossmatch’ (VXM) can be performed before the transport of the organ using the results of the single antigen bead (SAB) Luminex assay thereby allowing the assessment of recipient donor specific antibodies. This study investigates whether it is acceptable and safe to proceed with transplantation of shipped-in pancreata based solely on VXM results rather than waiting for FXM results.

Methods:  We retrospectively reviewed outcomes of pancreas transplants (n=153 patients) performed from June 2010 to December 2013. Comparisons were made between three patient groups: 1) shipped-in pancreas, VXM only (n=39), 2) shipped-in pancreas, VXM and FXM (n=12), and 3) local pancreas, VXM and FXM (n=102). Graft-survival, patient survival, and CIT were determined.

Results: There were 51 shipped-in pancreata: 39 transplants were performed based on the results of a VXM, not a FXM and 12 were based on FXM results. Transplants that began based on VXM results did have a FXM performed retrospectively. Donor and recipient demographics, immunosuppression regimens and surgical parameters were comparable between groups with the exception of transplantation type. Shipped-in pancreata were primarily solitary pancreas (SP) transplants whereas the local group had primarily SPK transplants (87% SP in Shipped-in VXM only group vs. 100% SP in Shipped-in VXM + FXM group vs. 13% SP in Local group, p < 0.001). Graft survival, death-censored graft survival, and patient survival did not differ between groups. CIT was shorter in the local group than either of the shipped-in groups (15.9h: Shipped-in VXM only vs. 17.5h: Shipped-in VXM + FXM vs. 13.2h: Local, p < 0.001). CIT was compared among pancreata originating from similar destinations. For pancreata shipped in from UNOS regions 3 and 4, proceeding to surgery without a FXM saved 5.1 hours (95% CI [3.25, 6.98]) (p = 0.0001). In no circumstance was the retrospective FXM positive in the VXM negative cases that proceeded to transplantation. 

Conclusion: VXM enables the transplant center to proceed to surgery without waiting for a FXM. For shipped in pancreata, CIT can then be minimized without adversely affecting graft or patient survival. This policy, if widely adopted, could increase utilization of pancreas grafts from further distances. In addition, such a policy is applicable to other transplanted organs.

 

45.04 Perioperative Sepsis Predicts Deep Venous Thrombosis In Colorectal Surgery Patients: A NSQIP Review

Posted on December 22, 2014

Q. Hatch1, D. Nelson1, J. Maykel3, E. Johnson1, B. Champagne2, N. Hyman4, S. Steele1  1Madigan Army Medical Center,Surgery,Tacoma, WA, USA 2Case Western Reserve University School Of Medicine,Cleveland, OH, USA 3University Of Massachusetts Medical School,Worcester, MA, USA 4University Of Chicago,Chicago, IL, USA

Introduction: Little data exists regarding the impact of perioperative sepsis on deep venous thrombosis (DVT) rates in colorectal surgery patients. We sought to quantify the current rate of 30-day DVT in colorectal surgery patients, and hypothesized perioperative sepsis increased the DVT rate in this patient population.

Methods: CPT codes were used to identify patients who underwent major colon or rectal surgery as reported to the National Surgical Quality Improvement Program (NSQIP) in 2010. Patients were stratified by the presence or absence of DVT in the perioperative period, and cohorts were compared in terms of demographics, comorbidities, and the presence of sepsis. Sepsis was defined by specially-trained NSQIP data collectors using Systemic Inflammatory Response Syndrome criteria with an identified infectious source.

Results: Of the 26,554 patients who underwent a major colorectal operation, 462 (1.7%) developed a deep venous thrombosis. Older patients (66 years vs. 61 years), open (77% vs. 23%) or emergency (31% vs. 15%) surgery, chronic steroid use (11% vs. 7%), and perioperative systemic sepsis (39% vs. 13%) were more likely to develop a DVT (P<0.01).  Current smokers and patients undergoing major rectal surgery were less likely to develop DVT (P≤0.02). On multivariate logistic regression analysis (controlling for ASA score, age, major rectal surgery, open surgery, length of hospital stay, smoking, and emergency surgery), systemic sepsis was independently associated with higher rates of DVT (OR=2.6, 95% CI 2.0-3.3, P<0.001).

Conclusion: Perioperative sepsis is a significant risk factor for post-operative deep venous thrombosis in the colorectal surgery population. Increased vigilance for DVT is warranted in patients who experience infectious complications after colorectal surgery.

 

45.05 Radiation Exposure Trends in Children with Appendicitis: Is Image Gently working?

Posted on December 22, 2014

M. M. McGuire1,2, T. A. Oyetunji3, H. T. Jackson1,4, W. Pastor1, D. I. Bulas1,4, F. G. Qureshi1,4  1Children’s National Medical Center,Surgery,Washington, DC, USA 2All Children’s Hospital,Surgery,St. Petersburg, FL, USA 3Howard University College Of Medicine,Washington, DC, USA 4George Washington University School Of Medicine And Health Sciences,Washington, DC, USA

Introduction:  The Alliance for Radiation Safety in Pediatric Imaging was formed in 2006 with a goal to promote radiation protection for children undergoing imaging.  In 2007 the alliance started the Image Gently campaign which promoted lower dosages and practice changes.  We studied the impact of the campaign on imaging trends in children with appendicitis.

Methods:  The Pediatric Health Information System (PHIS) database was queried for all cases of appendicitis from 2004-2013.  Imaging type used and severity of appendicitis was collected.

Results: 104,006 cases of appendicitis were identified with 41% perforated. There was an increase in total imaging over time for all cases (59.46% in 2004 to 65.53% in 2013). Computed tomography scan (CT) use reduced from 37.97% in 2004 to 25.47% in 2013 for all patients. Concomitantly ultrasound (US) use increased for all cases from 26.88% in 2004 to 52.2% in 2013.  Total imaging in perforated appendicitis trended downward (82.43% in 2004 to 65. 20% in 2013). CT scan use also decreased in perforated appendicitis from 49.60% in 2004 to 19.34% in 2013 while US increased from 38.93% in 2004 to 52.15% in 2013. There was increased CT use in all groups from 2004 to 2007 before plateauing and dropping especially in perforated appendicitis. 

Conclusion: Total Imaging has increased in children with appendicitis, with an increasing use of US and reduced use of CT.  CT scan use first increased and then decreased after 2007.  The reduction in CT use and increase in US use may be related to the Image Gently campaign.

45.06 Value of Routine Post-Op Upper GI Esophagogram After Uncomplicated Nissen Fundoplication

Posted on December 22, 2014

A. Bhama1, V. Wu1, B. Nardy1, H. Chong1  1University Of Iowa Hospitals And Clinics,General Surgery,Iowa City, IA, USA

Introduction: Laparoscopic Nissen fundoplication (LNF) is commonly performed for medically refractory gastroesophageal reflux disease and in conjunction with symptomatic hiatal hernia repairs. Though associated with a low incidence of postoperative complications, it may be still be a common practice to obtain postoperative upper gastrointestinal contrast studies (UGI) in the immediate postoperative setting. Extensive literature search did not reveal any literature regarding the use of routine UGI following LNF. The aim of this study is to evaluate the postoperative utility of UGI following uncomplicated LNF in leak assessment.

Methods: A single institution, retrospective review was performed of adult patients who underwent LNF from 2006 to 2012. Indications for LNF were refractory GERD or hiatal hernia repair. 171 patients were identified. Those undergoing reoperative surgery or Collis gastroplasty with LNF were excluded. Ten were excluded for reoperative surgery, and 19 were excluded for undergoing Collis gastroplasty. The final study comprised of 142 patients. UGI and postoperative outcomes were assessed from this group of patients. All UGI were obtained between postoperative days one through three and interpreted by staff radiologists. The cost of study was obtained from the radiology billing department.

Results: Of the 142 patients reviewed, the mean age of patients undergoing surgery was 56.9 years, with a female predominance 72%.  Indication for operation was refractory GERD in 58% (n=83) of patients, paraesophageal hernia repair in 42 %(n=59) of patients. UGI studies were obtained in 94% (n=134) of patients. Of the 134 UGI studies, no leaks were identified; however, one was a false negative study, and the patient required emergent reoperation for gastric perforation found on CT. There was no clinical suspicion for leaks in those who did not undergo UGI. Negative predictive value of UGI for leak was found to be 99.3%. Sensitivity, specificity and PPV are unable to be calculated given that no tests were positive for leak. Cost of each UGI and CT scan with interpretation was $816 and $4020 respectively.

Conclusion: This study suggests that there is little utility of UGI in the early postoperative period following uncomplicated LNF for refractory GERD and paraesophageal hernia repairs. If clinical suspicion of leak arises, a CT scan with oral contrast should be obtained in lieu of an UGI study.

 

45.07 Predictors of Operative Failure in Parathyroidectomy for Primary Hyperparathyroidism

Posted on December 22, 2014

D. C. Cron1, S. Kapeles1, S. Kwon1, P. Kirk1, E. Andraska1, B. McNeish1, D. T. Hughes1  1University Of Michigan,Surgery,Ann Arbor, MI, USA

Introduction:  Cure rates for hyperparathyroidism with parathyroidectomy are typically 95%; however, studies analyzing risk factors for operative failure and persistent hyperparathyroidism are limited. This study sought to identify predictors of operative failure in parathyroidectomy for primary hyperparathyroidism.

Methods:  This is a retrospective review of 2084 adult patients with primary hyperparathyroidism who underwent initial parathyroidectomy at a single center from 1999-2012.  Parathyroidectomy included both directed-approach (59%) and bilateral explorations (41%). The following data were collected: age, sex, perioperative lab values, imaging results, thyroid pathology, intraoperative data including parathyroid hormone (IOPTH) levels, and number of excised glands. Biochemical cure was defined as both an IOPTH decrease of ≥50% from the baseline and a level in the normal range (<75pg/ml). Operative failure (persistent disease) was defined by two or more postop serum calcium measurements ≥10.2mg/dL or a single calcium ≥11.0mg/dL when no other data available. Univariate logistic regression and backwards-stepwise multivariate logistic regression were used to identify significant predictors of operative failure. 

Results: Operative failure was identified in 65 patients (3%), with 35% having reoperative parathyroidectomy as a result. Significant predictors of failure on univariate analysis included: failure to meet IOPTH criteria, lower excised gland weight (< median of 600mg), multiple glands excised, bilateral exploration, lower preop PTH, higher preop calcium, and higher final IOPTH. A correct pre-op sestamibi was protective (Fig. A). Preop ultrasound localization was not associated with operative success. Predictors retained in the multivariate model included: IOPTH criteria met (OR=0.2, CI 0.1-0.3, P<0.01), low gland weight (OR=4.2, CI 2.0-9.0, P<0.01), high preop calcium (OR=1.9, CI 1.3-2.8), and low preop PTH (OR=0.99 CI 0.99-1.0, P=0.03). The C-statistic for this model was 0.80 (Fig. B). 

Conclusion: Operative failure of parathyroidectomy for primary hyperparathyroidism is rare; however, failure to meet IOPTH criteria, low preoperative PTH levels, high preoperative calcium levels, and low gland weight are risk factors for persistent disease.

 

45.08 Substernal Goiter: When is a Sternotomy Required?

Posted on December 22, 2014

L. Nankee1, H. Chen1, D. Schneider1, R. Sippel1, D. M. Elfenbein1  1University Of Wisconsin,Madison, WI, USA

Introduction: Sternotomy for substernal goiters (SSG) is associated with greater morbidity than a cervical approach.  Accordingly, predicting which patients will require a sternotomy is imperative for pre-operative planning.  In this study, we analyzed the pre-operative and post-operative characteristics of patients with SSG compared to those with large goiters contained entirely within the neck, or a cervical goiter (CG). We sought to identify predictors for sternotomy as a surgical approach for the removal of SSG. 

Methods: A retrospective review of the Endocrine Surgery Database was performed.  Patients were included if they had large (>100g) thyroids or any mention of a substernal component during their pre-operative workup.  Between 1995 and 2013, 220 patients met these criteria.  Comparisons were made between patients who had a SSG and patients who had a CG.  Further comparisons were made between those with a SSG who required sternotomy to excise their thyroid and those who underwent cervical incision only.

Results: Of the 220 patients, 127 (58%) patients had SSG, of whom 7 (5.5%) required sternotomy.  On bivariate analysis, there were no differences in gender, BMI, preoperative symptomatology, postoperative complications or length of stay for patients who had SSG vs CG.  Patients with SSG were older (62 + 15 vs 51 + 17 years, p<0.001), more likely to undergo preoperative CT scanning (69% vs. 31%, p<0.001), and less likely to have preoperative hyperthyroidism (10% vs. 29%, p<0.001).  Patients who underwent a sternotomy showed no difference in terms of gender, age, BMI, preoperative hyperthyroidism, or postoperative complications compared to those with SSG who underwent cervical incision thyroidectomy.  All patients who underwent sternotomy underwent preoperative CT scanning and were more likely to have preoperative symptoms of chest pressure and voice complaints.  Furthermore, all patients who underwent sternotomy had extension of the thyroid gland below the aortic arch.  Sternotomy took an average of 2 hours longer than a cervical incision, was associated with significantly more blood loss (600 + 408 vs. 190 + 118, p=0.04), and a longer length of stay (3.1 + 0.9 vs. 1.8 + 1.6 days, p=0.03) than cervical thyroidectomy. 

Conclusion: Sternotomy for SSG is rare.  All patients necessitating sternotomy had extension below the aortic arch on preoperative CT scanning, and were more likely to present complaining of chest pressure and voice issues.  

 

45.09 Potential Targets to Improve in Disparities in Thyroid Surgery Outcomes

Posted on December 22, 2014

L. E. Kuo1, H. Wachtel1, G. C. Karakousis1, D. L. Fraker1, R. R. Kelz1  1Hospital Of The University Of Pennsylvania,Department Of Surgery,Philadelphia, PA, USA

Introduction:  In an effort to improve surgical outcomes and minimize disparities, a great deal of attention has been focused on referral to high volume surgeons. While this is one potential solution, it is not feasible for all patients. Moreover, within high volume practices, thyroidectomy outcomes have been shown to be inferior amongst non-whites. We sought to examine differences in time to surgical consultation and presentation for treatment of benign thyroid conditions in an attempt to identify alternative targets for improving surgical care.

Methods:  We used our prospective endocrine surgery registry from 2002-2012 to identiry patients who underwent thyroidectomy for benign disease. Patients were classified as white or non-white. Patient demographics (race, age, gender, median income), disease history (diagnosis, time until surgical referral, symptoms at surgical consultation), surgical characteristics (mass, pathology) were examined. The primary outcome variable was time to surgical referral. Specimen characteristics were secondary outcome variables. Univariate analyses were performed using the Wilcoxon ranksum and chi square tests, as appropriate. A multivariable linear regression was performed to examine the influence of median income of race on time to surgical referral with adjustment for potential confounders.

Results: 1144 patients were studied. 81.7% were white and 18.3% were non-white. There was no difference in patient age or gender between the groups. Non-white patients had a lower median income than white patients ($46,622, interquartile range(IQR) $30,912:72,762 vs. $75,784, IQR $58,518:90,788, p<0.001). Non-white patients were more likely to report compressive symptoms than whites (53.6% vs. 32.2%, p<0.001), as well as dysphagia (19.6% vs. 13.2%, p=0.017). Non-white patients had a longer median time to surgical referral (0yrs, IQR 0:5) than whites (0yrs, IQR 0:3, p<0.017), and this difference remained after adjusting for confounders. Non-white patients also had a greater median thyroid mass (23.3gm, IQR 12.0:56.4 vs. 59.6gm, IQR 28.8:40.8, p<0.001). Amongst patients with goiter as the indication for surgery, non-white patients had significantly larger glands than white patients: 105.9gm, IQR 50.0:186.5 vs. 70.9gm, IQR 34.3:129.6, p<0.001.

Conclusion: Non-white patients had a longer time to surgical referral, were more likely to report symptoms of compression at the time of surgical referral, and had larger thyroids at the time of resection than white patients. This is the first time a racial disparity has been reported in presentation to surgical care for benign thyroid disease. Disparities in access to surgical care for benign disease are difficult to quantify due to limitations in the availability of clinical information in claims registries and datasets. Further studies are indicated to examine the benefits of referring patients for surgery while their condition is mild enough to be successfully treated by a surgeon. 

 

45.10 Outcomes Using Double-Stapled Technique for Esophagoenteric Anastomosis in Gastric Cancer

Posted on December 22, 2014

A. Falor1, A. Choi1, S. Merchant1, M. Lew2, B. Lee1, I. B. Paz1, R. Nelson3, J. Kim1  1City Of Hope National Medical Center,Division Of Surgical Oncology,Duarte, CA, USA 2City Of Hope National Medical Center,Department Of Anesthesia,Duarte, CA, USA 3City Of Hope National Medical Center,Department Of Biostatistics,Duarte, CA, USA

Introduction: Several surgical techniques to perform esophagoenteric anastomosis for total/proximal gastrectomy have been described including the double-stapled technique (DST), which involves a circular stapled anastomosis across the stapled end of the esophagus. Since prior reports on DST for gastric cancer are limited, our objective was to examine rates of anastomotic leak and stricture with DST for esophagoenteric anastomosis in patients with gastric cancer.

Methods: A single institution review was performed for patients who underwent total/proximal gastrectomy with DST between 2006 and 2014. All DST were performed using the OrVil™ and an end-to-end anastomosis (EEA) stapler. Patient and treatment-related variables were tabulated. Anastomotic leaks were defined as perianastomotic extravasation of oral contrast on radiographic imaging or anastomotic disruption on endoscopy. Stricture was defined as symptomatic anastomotic narrowing requiring dilation.

Results: Of 55 patients who underwent DST between 2006 and 2014, total gastrectomy was performed in 44/55 (80%), proximal gastrectomy in 6/55 (11%) and completion gastrectomy in 5/55 (9%). Eleven patients (20%) had multi-visceral resection at the time of gastrectomy. Fifty patients (91%) had adenocarcinoma on final pathology, and 22 patients (40%) received neoadjuvant chemotherapy. Six patients (10.9%) had undergone radiation therapy prior to completion gastrectomy for recurrent disease. Operative approach was open (n=26/55; 47.2%), laparoscopic (n=26/55; 47.2%), and robotic (n=3/55; 5.4%). The leak rate was 5/55 (9%) occurring at a median of 14 days (5-20 days). The stricture rate was 12/55 (21.8%) occurring at a median of 86 days (40-405 days). Leak and stricture rates improved with increased experience. During the 2012-2014 period, the rates of anastomotic leak and stricture were 0/19 (0%) and 3/19 (15.7%), respectively. The overall complication rate was 21/55 (38.1%) of which 60% were classified as Clavien-Dindo grade III-V complications. On multivariate analysis, none of the aforementioned variables correlated with risk for leak or stricture.

Conclusion: In the largest Western series of DST esophagoenteric anastomoses for gastric cancer, our experience demonstrates that DST is a safe and effective technique with low rates of anastomotic leak and stricture.

 

46.01 Both Systemic Inflammation and Visceral Obesity Are Associated with Colorectal Cancer Recurrence

Posted on December 22, 2014

C. T. Aquina1, A. S. Rickles1, C. P. Probst1, B. J. Hensley1, A. A. Swanger1, K. Noyes1, J. R. Monson1, F. Fleming1  1University Of Rochester,Surgical Health Outcomes & Research Enterprise (S.H.O.R.E.),Rochester, NY, USA

Introduction:  Much recent attention has been directed towards the detrimental effects of a pro-inflammatory state on tumorigenesis and oncologic outcomes. Our research group has previously shown a relationship between elevated metabolically active visceral fat volume and recurrence-free survival in colorectal cancer. Other studies have demonstrated a similar link to colorectal cancer outcomes with laboratory inflammation markers such as the neutrophil-to-lymphocyte ratio. We sought to investigate whether pre-operative inflammatory indices are associated with long-term colorectal cancer outcomes and whether this relationship is mediated by visceral obesity.

Methods:  A single-center retrospective chart review was performed for patients undergoing surgical resection for colorectal cancer between 2000 and 2009. Pre-operative CT scans were used to calculate visceral fat volume (VFV) based upon a previously validated method. Pre-operative laboratory values within 90 days of the date of surgery were used to calculate the neutrophil-to-lymphocyte ratio (NLR). A pre-operative NLR>3 was used as a cut-off to define high NLR according to previous literature. Visceral obesity was defined as VFV>1620cm3 based upon the results of a receiver operating characteristic curve. Five-year recurrence-free survival was defined as locoregional or distant recurrence within 5 years of surgery. Chi-square, Student’s T-test, Kaplan-Meier, and Cox proportional-hazards analysis were used to compare pre-operative and surgical characteristics with recurrence-free survival.  

Results:  Overall, 141 patients met inclusion criteria with 75 patients having a high NLR (53.2%) and 48 patients having visceral obesity (34%). Patients with a high NLR had significantly higher levels of VFV (mean=1741cm3 vs. 1399cm3, p=0.04). On Kaplan-Meier analysis, both high NLR (p=0.03) and visceral obesity (p=0.005) were independently associated with reduced 5-year recurrence-free survival for stage II colorectal cancer but not stages I or III. Using separate Cox proportional-hazards models due to the association between high NLR and visceral obesity, both high NLR (HR=4.04, p=0.04) and visceral obesity (HR=4.05, p=0.03) were associated with a more than 4-fold risk of cancer recurrence within 5 years for stage II colorectal cancer.

Conclusion:  Both pre-operative systemic inflammation, as captured by an elevated pre-operative neutrophil-to-lymphocyte ratio, and visceral obesity are associated with worse recurrence-free survival for stage II colorectal cancer. Additional study is warranted at exploring the association between these two factors and developing strategies at improving outcomes for this high-risk population.

 

46.02 Minimally invasive distal pancreatectomy for cancer: Short-term oncologic outcomes in 1733 patients

Posted on December 22, 2014

M. Abdelgadir Adam1, K. Choudhury3, M. Dinan2, S. Reed2, R. Scheri1, D. Blazer1, S. Roman1, J. Sosa1,2  1Duke University Medical Center,Surgery,Durham, NC, USA 2Duke Clinical Research Institute,Durham, NC, USA 3Duke University Medical Center,Biostatistics,Durham, NC, USA

Introduction: Emerging data from high-volume institutions suggest that minimally invasive distal pancreatectomy (MIDP) provides favorable perioperative outcomes and adequate oncologic resection.  However, it is unclear if these outcomes are generalizable at a population level. This study examines patterns of use and short-term outcomes from MIDP vs. open distal pancreatectomy for pancreatic cancer.  

Methods: Adult patients undergoing distal pancreatectomy were identified from the National Cancer Database, 2010-2011. Descriptive statistics were used to characterize patterns of laparoscopic and/or robotic MIDP use. Multivariable modeling was applied to determine factors associated with use of MIDP and compare short-term outcomes from MIDP vs. open surgery, while adjusting for patient, clinical, and tumor characteristics.

Results: A total of 1,733 patients underwent distal pancreatectomy for cancer: 535 had MIDP and 1,198 had open surgery. Use of MIDP increased 43% between 2010 to 2011, from 220 to 315 cases. Across both study years, the conversion rate from MIDP to open distal pancreatectomy was 23%. MIDP cases were performed at 215 hospitals, with the overwhelming majority of hospitals (97%) performing <10 cases overall. The majority of MIDP cases (67%) were performed at academic institutions. Patients were more likely to undergo MIDP if they were older [odds ratio (OR) 1.02 (95% confidence interval (CI) 1.01-1.04), p<0.01], privately insured [OR 1.41 (CI 1.04-1.92), p=0.03], diagnosed in 2011 (vs. 2010) [OR 1.48 (CI 1.17-1.86), p<0.01], or had a diagnosis of a neuroendocrine malignancy (vs. adenocarcinoma) [OR 1.82 (CI 1.37-2.40), p<0.01]. After adjustment, compared to the open group, those who underwent MIDP were more likely to have negative surgical margins [OR 1.66 (CI 1.12-2.46), p=0.01] and a shorter length of stay [relative risk (RR) 0.82 (0.76-0.89), p<0.01]; the number of lymph nodes removed [RR 0.94 (0.85-1.04), p=0.24], rates of 30-day readmission [OR 1.15 (0.72-1.83), p=0.57] and 30-day mortality [OR 0.34 (0.06-1.80), p=0.20] were similar between groups.     

Conclusion: Use of MIDP for cancer is increasing, with most centers performing a low volume of these procedures. Use of MIDP for body and tail malignancies of the pancreas appears to have short-term outcomes that are similar to those of open procedures with the benefit of a shorter length of hospital stay. Larger studies with longer follow-up should be undertaken to examine clinical outcomes.

 

46.03 Menopausal Status Does Not Predict Recurrence Score Using Oncotype DX Assay

Posted on December 22, 2014

D. N. Carr3, N. Vera3, J. Mullinax1, D. Korz1, W. Sun1, M. Lee1, S. Hoover1, W. Fulp2, G. Acs4, C. Laronga1  1Moffitt Cancer Center And Research Institute,Breast Program,Tampa, FL, USA 2Moffitt Cancer Center And Research Institute,Biostatistics,Tampa, FL, USA 3University Of South Florida College Of Medicine,Tampa, FL, USA 4Women’s Pathology Consultants, Ruffolo Hooper & Associates,Tampa, FL, USA

Introduction:  Adjuvant treatment planning for early stage, estrogen receptor (ER) positive invasive breast cancer has been historically based on menopausal status. The Recurrence Score (RS) from the 21-gene breast cancer assay (ODX) is predictive of distant recurrence in this population, but is rarely applied to younger, premenopausal patients (pts). To evaluate the validity of this historical bias, we sought to evaluate the relationship between menopausal status and the recurrence score derived from the Oncotype DX assay.

Methods:  An IRB-approved retrospective review was conducted of invasive breast cancer pts with known RS. Eligibility for performance of the ODX was based on NCCN guidelines or physician discretion. Data collected included demographics, clinical-pathologic variables, surgery type, adjuvant treatment and outcomes. Menopausal status was documented at time of ODX. Perimenopausal women were classified with premenopausal for statistical analyses. Comparisons on RS were made by menopausal status (premenopausal vs. postmenopausal) using general linear regression model and the exact Wilcoxon Rank Sum Test. 

Results: 607 pts with invasive breast cancer and a RS were identified. Menopausal status was available for 600 pts (166 premenopausal, 434 postmenopausal) comprising our study population. Median age for the entire population was 58yrs (range: 27-84); 47yrs for premenopausal and 62yrs for postmenopausal. Median invasive tumor size was 1.5 cm for both cohorts. No significant differences were seen between cohorts for overall survival, metastatic disease rate, histologic grade, lymphovascular invasion (LVI), nodal status, stage, adjuvant chemotherapy, or endocrine therapy use. Mastectomy rate was higher in the premenopausal group (53.8%), compared to postmenopausal (41.9%) (p=0.0001), and thus receipt of breast irradiation was lower in premenopausal women. Despite the higher mastectomy rate in premenopausal women, a higher local-regional recurrence rate (3% vs. 0.7%; p=0.0384) was observed. Degree of ER expression was lower in premenopausal women (95%) than postmenopausal (100%) (p<0.0001). Median RS was the same (16) for both premenopausal (range: 0-62) and postmenopausal (range: 0-63) women. Tumor size, nodal status, and stage did not affect RS. Menopausal status as a categorical variable was not predictive of RS (p value = 0.7731). Factors predicting higher RS included higher mitotic rates (p<0.0001), higher nuclear grade (p<0.0001), decreased tubule formation (p=0.0001), presence of LVI (p=0.002), high grade (p<0.0001), and lower expression levels of ER/PR (p<0.0001).  

Conclusion: Menopausal status has limited predictive power for distant breast cancer recurrence. We have shown that RS across the spectrum of menopausal status is well distributed in this cohort of women. Therefore, menopausal status alone should not preclude recommendations for performance of ODX in ER-positive, early stage breast cancer.
 

46.04 Tumor Associated Macrophage Expression of Folate Receptor β in Lung Cancer: Prognostic Significance

Posted on December 22, 2014

A. Bain1, A. Vachani6, P. Low7, S. Singhal4, C. Deshpande5  6Perelman School Of Medicine At The University Of Pennsylvania,Department Of Medicine,Philadelphia, PA, USA 1Perelman School Of Medicine At The University Of Pennsylvania,Philadelphia, PA, USA 4Perelman School Of Medicine At The University Of Pennsylvania,Department Of Surgery,Philadelphia, PA, USA 5Perelman School Of Medicine At The University Of Pennsylvania,Department Of Pathology,Philadelphia, PA, USA 7Purdue University,Department Of Chemistry,West Lafayette, IN, USA

Introduction:  This study was undertaken in order to better elucidate the relationship between Folate Receptor Beta (FRβ) positive macrophages and cancer prognosis. Tumor associated macrophages (TAMs) play a key role in promoting inflammation and regulating the immune response to malignancies.  FRβ is a useful target because its expression is limited to activated macrophages in a subset of disease conditions, including neoplasms.  While previous work has suggested a relationship between FRβ+ macrophages and cancer prognosis, to our knowledge no such studies exist for lung adenocarcinoma.

Methods: 69 patients underwent resection for primary lung adenocarcinoma from 2003-2006.  Under IRB approved protocol, a tissue microarray (TMA) was constructed using formalin-fixed, paraffin-embedded specimens from patient tumors.  A TMA section was stained using FRβ-specific monoclonal antibody m909 in antibody diluent (1:100).  Cytoplasmic staining was measured for FRβ+ macrophage frequency and staining intensity. Samples were scored as 0, no staining; 1+, weak; 2+, moderate; and 3+, strong.  Average scores for patients with ≥2 scored TMA tumor cores (n=50) were included in the analysis.  An unweighted product score was derived from the frequency and intensity scores to yield a total score of 0-9.  Clinical measures including tumor staging and survival status were followed for a minimum of 4 years.  The relationship between FRβ expression and survival time was tested using a student's t-test, and differences of tumor staging and survival times were compared by one-way ANOVA.  Statistical analyses were done with Stata 13 (StataCorp., College Station, TX).

Results: 37 patient TMA cores had FRβ product scores ≥4, and 29 patients had product scores ≥6.  The difference in mean survival time for patients with product scores ≥6 (1750.8 days) and patients with product scores <6 (1417.0 days) was 333.8 days (p-value = 0.0493).  Among deceased patients, mean survival time for the ≥6 group (1121.1 days) was greater than for the <6 group (901.5 days).  The difference in mean survival time between patients with Stage I/II disease and Stage III disease was 257.0 days (p-value = 0.1588).  The relationship between tumor stage and FRβ expression was not statistically significant (Fisher’s exact, p-value = 0.180).

Conclusion: Previous clinical studies have suggested that FRβ expression is limited to M2 macrophages and is associated with poor outcomes.  However, our results indicate that FRβ expression may be associated with longer survival times in patients with lung adenocarcinoma.  FRβ expression was also a better prognostic indicator of survival time than tumor stage alone.

46.05 Accuracy of Surgeons In Predicting the Dose of Levothyroxine After Total Thyroidectomy

Posted on December 22, 2014

H. M. Yong1, T. W. Yen1, K. Doffek1, D. B. Evans1, T. S. Wang1  1Medical College Of Wisconsin,Department Of Surgery / Division Of Surgical Oncology,Milwaukee, WI, USA

Introduction:  Following total thyroidectomy, levothyroxine (LT4) is often prescribed at discharge by the surgical team, with follow-up thyroid function tests (TFTs) and dose adjustments primarily managed by the endocrinologist or primary care physician (PCP). The aim of this study was to investigate the accuracy of the surgical team in determining the initial LT4 dose.

Methods:  A retrospective chart review of a prospectively collected, thyroid database was performed of 420 patients with benign thyroid disease who underwent a completion or total thyroidectomy between 1/2009-10/2014 and were prescribed an initial LT4 dose by the surgeon. Data collected included age, gender, body mass index (BMI: kg/m2), pre- and postoperative final pathology, initial surgeon-prescribed LT4 dose, TFTs, subsequent modifications to LT4 dose by the primary care physician/endocrinologist, and the time to achieve biochemical (TSH 0.45 – 4.5 uIU/mL) euthyroidism. All patients had follow-up with an endocrinologist or PCP within 8 weeks of surgery; data on LT4 doses were collected for six months postoperatively.

Results: The final cohort consisted of 289 patients. Median age was 53 years (range, 18-86) and 248 (86%) were female. Median BMI was 31.1 (range, 16.1-63.7). The median LT4 dose initially prescribed was 137 mcg (1.65mcg/kg; range –75-200); 119 (41%) patients achieved normal serum TSH values at initial follow-up and required no dose adjustments. Of the remaining 170 patients, 52 (31%) had elevated TSH levels at initial follow-up, requiring a median adjustment of 25 mcg (range, 11-50). Of the 118 patients with suppressed TSH levels at initial follow-up, the median adjustment was 21.5 mcg (range, 3-60). At 6 months, 59 (34%) of the 170 patients who required an initial dose adjustment had follow-up data; 57 (97%) were euthyroid, achieved at a median of 4.6 months (range, 2–6) and a median of 2 provider visits (range, 2-5).  

For the overall cohort, there was no difference in the proportion of patients who required dose adjustment by BMI or race. However, there was a significant difference in the proportion of patients requiring dose adjustment by age; older patients were more likely to be on too high a dose of LT4 at initial follow-up (p=0.0117; Table).

Conclusion: The initial dose of LT4 prescribed by the surgical team after total thyroidectomy was accurate in less than 50% of patients, with more patients being prescribed a higher dose of LT4 than ultimately required, particularly in the elderly. This underscores the importance of follow-up with thyroid function testing within the initial postoperative period and suggests that preoperative discussion with their referring provider for initial LT4 doses may be appropriate.
 

« Previous Page
Next Page »