82.03 Epidurals are Associated with Increased Morbidity and Length of Stay in Open Ventral Hernia Repairs

Posted on January 18, 2018

S. L. Zhou1, M. C. Helm1, J. H. Helm1, M. I. Goldblatt1  1Medical College Of Wisconsin,Department Of Surgery,Milwaukee, WI, USA

Introduction:

Open ventral hernia repair is a common surgical procedure with multiple options for post-operative pain control.   Intravenous narcotics or epidural anesthesia are commonly used options. The primary objective of this study was to determine the morbidity and mortality associated with the use of epidurals for post-operative pain control.

Methods:
This study was a retrospective review of patients who underwent open ventral hernia repair. Data was abstracted from the National Surgical Quality Improvement Program 2015 dataset. Outpatient procedures, emergency cases, patients who did not receive mesh, or patients who remained inpatient for less than two days were excluded from analyses to identify only complex hernias. 

Results:
In total, 1943 patients met inclusion criteria of which 1009 (51.93%) received any combination of non-epidural post-operative pain relief and 934 (48.07%) received an epidural. The patients who received an epidural had a higher incidence of pulmonary embolism, urinary tract infection and had a longer operative time and length of stay compared to those patients without an epidural (Table 1).

Conclusion:
The use of epidurals was associated with in an increased incidence of pulmonary embolism and urinary tract infections.  In addition, epidural use was associated with an increased operative time and length of stay.  The use of epidurals was not associated with an increased incidence of surgical site infections suggesting that the complexity of hernias between the groups was similar. 

82.04 Thyroidectomy In Older Adults: An ACS-NSQIP Study Of Outcomes

Posted on January 18, 2018

Z. T. Sahli1, G. Ansari1, J. K. Canner1, D. Segev1, M. A. Zeiger1, A. Mathur1  1Johns Hopkins University School Of Medicine,Department Of Surgery,Baltimore, MD, USA

Introduction:  The rise in the geriatric population in the US along with the increasing prevalence of thyroid nodular disease and cancer will lead to a higher number of thyroidectomies performed in this age group.  The impact of thyroidectomy in older adults is not well defined. The aim of our study was to evaluate surgical outcomes after thyroidectomy in older adults as compared to younger adults.

Methods:  We conducted a retrospective cohort study using the American College of Surgeons National Surgery Quality Improvement Program database from 2012-2015. We included and categorized thyroidectomy patients into three age groups (18-64 years, 65-79 years, and ≥80 years) and analyzed 30-day perioperative outcomes using the bivariate X2 test and multivariate logistic regression to estimate risk of outcomes.

Results: Our study identified 39,859 patients who underwent thyroidectomy. Among our cohort, 31,315 (78.56%) patients were between 18-64 years, 8,544 (21.44%) were between 65-79 years, and 904 (2.27%) were ≥80 years. Compared to younger patients, patients ≥80 years were 2.25 times more likely to develop a complication (95% confidence interval [CI]: 1.58-3.20, p<0.001), 1.52 times more likely to have a longer hospital stay (95% CI: 1.18-1.96, p=0.001) and were associated with higher rates of hematoma (16.67%, p<0.001). Compared to younger patients, patients 65-79 years were 1.36 times more likely to develop a complication (95% CI: 1.12-1.64, p<0.001), 1.31 times more likely to have a reoperation (95% CI: 1.07-1.62, p=0.011), and 0.64 times more likely to have a lower rates of related readmission (95% CI: 0.47-0.87, p=0.004).

Conclusion: Patients ≥80 years have significantly higher rates of complications, longer lengths of stay, and incidence of neck hematomas. Patients 65-79 years have higher rates of complications and reoperation rates and lower rates of related readmission. Further studies are needed to risk stratify individuals within the aging population to counsel patients and potentially mitigate these risks.

 

82.02 Perioperative Factors Associated With Postoperative Pain Following Open Ventral Hernia Repair

Posted on January 18, 2018

W. R. Ueland1, M. Plymale1, D. Davenport1, J. Roth1  1University Of Kentucky,Lexington, KY, USA

Introduction: Effective pain control following open ventral and incisional hernia repair (VHR) impacts all aspects of patient recovery. To reduce opioids and enhance pain control, multimodal pain management including use of epidural analgesia, muscle relaxants, and non-opioid analgesics are thought to be beneficial. The purpose of this study was to identify perioperative characteristics associated with patient-reported pain scores.

Methods: After obtaining IRB approval, surgical databases were searched for open VHR cases performed by one surgeon over three years. Modes of pain management and visual analog scale (VAS) pain scores were recorded in twelve-hour intervals to hospital discharge or up to eight days post-operation. Patient characteristics were determined by medical record review. Forward stepwise multivariable regression (p for entry < .05; exit > .10) was used to assess the independent contribution to VAS scores of the preoperative, operative and postoperative factors.

Results: One hundred and seventy-five patients underwent elective open VHR with mesh implantation and were included in the analyses. Average patient age was 55.1 years (+/- 12.8 years) and slightly over half of the patients were female (50.9%). Just over one in ten patients were morbidly obese (BMI ≥40 kg/m2). No significant (p < .01) associations were found between VAS pain scores at any time point based on gender, ASA class, BMI, smoking status, history of cancer, heart disease or COPD. Patient factors independently associated with increased preoperative VAS scores included: preoperative opioid use, open wound, CDC Wound Class II and prior hernia repair(s). Patients with epidural for postoperative pain had significantly decreased VAS pain scores across the time continuum. Operative factors significantly associated with increased preoperative VAS pain score included: median hernia defect size, concomitantly performed procedure(s), duration of operation and estimated blood loss (EBL). Hospital length of stay and postoperative surgical site occurrence were associated with increased VAS pain scores at the preoperative and 0-11 hour postoperative time points. Greater preoperative VAS pain score predicted increased pain at each postoperative time point (all p < .05). 

Conclusion: Preoperative pain status and opioid use are associated with increased VAS pain scores postoperatively. Epidural analgesia effectively results in decreased patient-reported pain. Increased preoperative VAS pain scores are reflected in increased operative complexity measured by operative duration and EBL. 
 

82.01 Emergency General Surgery Patients With Psychiatric Comorbidities and Increased Resource Utilization

Posted on January 18, 2018

A. Lauria2, V. Haney1, J. S. Kim1, A. Kulaylat1, S. Armen1, M. Boltz1, S. Allen1  1Milton S. Hershey Medical Center,Hershey, PA, USA 2Walter Reed Medical Center,Bethesda, MARYLAND, USA

Introduction: Mental health disorders offer a challenge to the care of patients across medical and surgical specialties. The impact of mental health disorders on the emergency general surgery (EGS) population is largely unstudied. We aimed to identify the prevalence of psychiatric disorders in EGS patients, and hypothesized that those with mental health comorbidities who underwent emergent procedures would have worse postoperative outcomes and require more intensive resource utilization.

Methods: Using standard NSQIP practices, data were collected on adult patients admitted for emergent cholecystectomy or appendectomy at a single academic center between 04/01/07 and 01/01/16. Charts were reviewed for psychiatric comorbidities and psychotropic medications. Logistic regression was used to determine the impact of psychiatric comorbidities on postoperative complications, ED visits, and readmission within 30 days.

Results: Of the 641 patients identified (appendectomy n=491 and cholecystectomy n=150), 115 patients (17.9%) had psychiatric comorbidities. Mood disorders were most common (76.5%), followed by anxiety or adjustment disorders (41.7%). Patients with psychiatric comorbidities experienced longer hospitalizations (median 2 vs. 1 days, p<0.001) and required more subsequent ED visits (18.3% vs. 10.3%, p=0.016) compared to those without a psychiatric diagnosis. On multivariable analysis, the presence of psychiatric comorbidities was associated with nearly twice the odds of ED visits within 30 days (OR 1.92, 95% CI 1.01 to 3.68).

Conclusions: Patients with mental health comorbidities who undergo emergency general surgery burden the healthcare system with longer lengths of stay and more ED visits. Protocols and patient education focused on those with mental health disorders may improve these outcomes.

 

81.15 “Practice Patterns and Outcomes of Splenic Flexure Mobilization During Laparoscopic Left Colectomy.”

Posted on January 18, 2018

B. Resio1, K. Y. Pei1  1Yale University School Of Medicine,Surgery,New Haven, CT, USA

Introduction: Historically, splenic flexure mobilization (SFM) was recommended for left colectomies to ensure adequate length for a tension free anastomosis and thus potentially mitigate the incidence of anastomotic complications. Despite the exponential increase in the adoption of the laparoscopic approach to colectomies, the national practice pattern and outcomes of SFM among this group are unknown. This study investigates the use of and outcomes associated with SFM for laparoscopic, partial colectomies with anastomosis for left colon and rectal cancers.

Methods:  The American College of Surgeons NSQIP database from 2007-2015 was queried for elective, laparoscopic, partial colectomies with anastomosis (CPT 44204, 44207, 44208) with/without SFM (CPT 44213) for colon cancer of the left colon or rectum (ICD9 153.2, 153.3, 154.0, 154.1) or (ICD10 C18.6, C18.7, C19, C20). Only cases labeled as elective were included and cases labeled as emergency or ASA 4/5 (life-threatening/ moribund) were excluded.  Primary outcome measures included all complications, superficial, deep and organ space infections, anastomotic leaks, postoperative ileus, return to operating room, death, hospital length of stay (HLOS) and operative time. Logistic regression models were used to compare outcomes, adjusting for patient and operative characteristics.

Results: 17,319 cases were identified of which 39.2% underwent SFM. Specifically, the proportion of SFM for left colon, sigmoid, recto-sigmoid and rectum were 49.3%, 33.3%, 36.9% and 46.9%, respectively. There was an increase in the overall proportion of cases with SFM during the study period (10.9% increase from 2007 to 2015). Compared to colectomies without SFM, patients undergoing SFM had an increase in: all complications (17.6% vs 15.6%, risk adjusted OR 1.11, 95%CI: 1.02-1.20); organ space SSI (4.6% vs 3.4%, risk adjusted OR 1.22, 95%CI: 1.04-1.43); prolonged ileus (10.7% vs 8.1%, risk adjusted OR 1.23 95%CI: 1.05-1.44) and operative time (mean time 234 vs 197 min, p<0.0001, 95%CI: 231.3-236.3 vs 195.4-198.9). There was no significant difference after risk adjustment for superficial SSI, deep SSI, return to operating room, anastomotic leak, death and HLOS.

Conclusion: Splenic flexure mobilization is performed in less than 50% of elective, laparoscopic, left, partial colectomies for colon and rectal cancer and is associated with increased complications, prolonged ileus, and operative time. Study findings support selective splenic flexure mobilization.

 

81.17 Morbidity and Mortality in Patients Undergoing Fecal Diversion as an Adjunct to Wound Healing

Posted on January 18, 2018

R. J. Kucejko1, M. E. Pontell1, D. Scantling1, M. Weingarten1, D. E. Stein1  1Drexel University College Of Medicine,Surgery,Philadelphia, PA, USA

Introduction:  Stomas are routinely created for fecal diversion in chronic, non-healing wounds of the sacrum, ischium and perineum. Aside from re-routing stool from the wound bed, they also improve quality of life and prepare the patient for future reconstructive surgery. While these procedures are commonplace, little is published about their safety, with only two studies in the last twenty years. This study aims to analyze patients from our institutional database as well as the NSQIP national database who underwent fecal diversion for non-healing wounds to clarify the safety of fecal diversion in this group of patients and to identify factors that contribute to elevations in perioperative risk.

Methods:  A retrospective analysis was performed using data from the American College of Surgeons National Surgical Quality Improvement Project database between 2005 and 2015. Patients were selected based on a postoperative diagnosis of chronic ulcer of the skin. Patients were considered to have undergone diversion if the entry contained the procedure code for ileostomy or colostomy. Propensity score matching was conducted based on the NSQIP morbidity score. An additional retrospective analysis was performed on our institutions patient database spanning from 2000 until 2017. All patients who underwent fecal diversion for chronic, non-healing wounds were included. 

Results: 4,849 patients meeting inclusion criteria were identified in the NSQIP database. 859 underwent diversion compared to 3,990 patients who did not. In unmatched data, comparison of the two groups revealed no significant differences in mortality rate, postoperative stroke, need for cardiopulmonary resuscitation, myocardial infarction, need for blood transfusion, deep venous thrombosis, renal failure, organ space or superficial surgical site infection. In matched data, diverted patients had a significantly lower 30 day mortality. 56 patients were identified at our institution that underwent fecal diversion for non-healing wounds. 50% of patients with a preoperative ejection fraction of less than 30% died within 30 days of surgery (p = 0.045, likelihood ratio 6.58).

Conclusion: Fecal diversion in patients with chronic non-healing sacral wounds does not increase 30 day morbidity and mortality, based on NSQIP data.  While the 30 day morbidity does remain high, the subgroup of patients with severe cardiac dysfunction likely represent the majority of these cases and remain at a disproportionately elevated risk, based on our institutional data. It is reasonable to suggest that patients with cardiac risk factors undergoing fecal diversion for chronic wounds should undergo preoperative echocardiography. We propose that a preoperative ejection fraction less than 30% should be seen as a relative contraindication to immediate diversion without further optimization.

81.11 Virtual Postoperative Visits for New Ostomates

Posted on January 18, 2018

T. L. White1, J. Moss1, P. Watts1, J. Cannon1, D. Chu1, G. Kennedy1, S. Vickers1, M. Morris1  1University Of Alabama At Birmingham,Birmingham, AL, USA

Introduction: Post-operative education, discharge instructions, and follow-up appointments provide a foundation for new ostomates leaving the hospital, but a gap in care remains.  Studies show that having a stoma is an independent predictor of hospital readmission. Patients with new stomas utilize resources, including hospital based acute care, twice as much as colorectal patients without an ostomy. Telehealth has an emerging role post-operatively, allowing visual inspection of the patient while providing verbal support during virtual visits before clinic follow-up.  The purposes of this project are to determine the feasibility of Virtual Postoperative Visits (VPOVs), to define specific issues patients want addressed during VPOVs, and to assess whether patients are satisfied with a virtual format. Our hypothesis is that virtual post-operative visits will be feasible and will address patient centered goals of care following discharge with a new stoma.  

Methods: In this pilot project, we recruited 10 patients who attended 2 VPOVs following hospital discharge in addition to routine post-operative WOCN education and a post-operative clinic appointment. The VPOVs were conducted and recorded using UAB approved, HIPAA compliant video conferencing software. Descriptive statistics were used to analyze data gathered from a survey assessing patient satisfaction.     

Results: The mean age of our 10 patients recruited was 40 and 80% were women. Surgical procedures included robotic, laparoscopic, and open colectomies with 80% resulting in an ileostomy and 20% in a colostomy. Of the patients enrolled, 90% successfully completed two VPOVs. The mean time to the first VPOV was 9 days post-discharge (range 2-7 days) and none of the patients enrolled were readmitted.  Ninety percent of patients felt VPOVs helped manage the ostomy and agreed that VPOVs should be part of the discharge plan. All patients felt comfortable with the virtual format. Common themes addressed during VPOVs included pouching issues and skin irritation.  Barriers to enrolling patients into our VPOV pilot study included lack of access to technology and HIPPA compliant software for smart phones. 

 

Conclusion: The immediate post-operative period is a tenuous time for new ostomates. Overall, VPOVs are feasible and patients are very satisfied with VPOVs in addressing their patient centered goals of care.  Bridging the period between hospital discharge and initial clinic follow-up by using culturally sensitive, educational, and timely interventions should be a priority in this population. Future work will focus on large scale implementation of VPOVs for patients with new stomas.

 

81.12 Robotic vs Laparoscopic Resection for Colorectal Disease

Posted on January 18, 2018

T. K. Kleinschmidt1, M. Ferrara1, J. Rosser1, M. Parker1  1Brookwood Baptist Health System,Department Of General Surgery,Birmingham, AL, USA

Introduction: The aim of this study is to compare the results of robotic versus laparoscopic colon resection for all causes performed by two high-volume private practice colon and rectal surgeons.  Current recommendations are at least 150-250 cases on the da Vinci platform are needed to become adept.  This study bears significance as current research shows inconsistent results in comparing the two modalities which could be in part because many of the current literature are meta-analyses incorporating multiple centers, surgeons and disparate patient populations. 

Methods:  Retrospective analysis was performed for all patients who had either robotic or laparoscopic colon resection over a 10 month period.  Outpatient records were reviewed to ascertain demographic data such as: age, BMI, tobacco use, comorbidities and prior surgeries.  Hospital records were accessed to determine: case length in OR, estimated blood loss (EBL), need for conversion of case, length of postoperative stay and complication rates.  Complication rates were defined as: surgical site infections (SSI), need for reoperation and/or loss in quality of life as recorded in follow up office visits.  Statistical analysis of data was performed using R software.

Results: The study included 166 patients (109 in the Robotic surgery (RS) and 57 patients in the Laparoscopic surgery (LS) groups). For RS patients: mean age was 57.8 years of age, 57% were female, mean BMI was 29.9 and the most common procedure performed was LAR (67%) and most common indication was diverticulitis (36.7%). For LS patients: mean age was 64.0 years of age, 47.4% were female, mean BMI was 27.5 and the most common procedure performed and indications were also LAR (50.9%) and diverticulitis (36.1%). Mean operative time was longer in RS group (138.3 versus 125.4 minutes, respectively [p=0.0380]). Estimated blood loss was less in RS (59.6 versus 106.3cc, respectively [p=0.0282]). Mean postoperative length of stay was shorter for RS than LS (2.85 versus 4.0 days [p=0.0046]). Complication rates were similar in LS (26.3%) and RS (20.2%) [p=0.3681], and the most common LS complication was SSI (32.3%%) and most common complication for RS was SSI (42.1%). Conversion of operative technique was similar in RS and in LS (19.3% versus 13.8%, respectively [p=0.3524]).

Conclusion: Robotic surgery for colon resection had decreased EBL, and significantly shorter postoperative hospital stay.  Laparoscopic surgery had shorter operating time.  Similar results were found between postoperative complications and conversion rates of surgery.  In the hands of surgeons who have performed the reported amount of surgeries needed to become adept with the da Vinci Surgical System, robotic surgery has significant advantages for colon resection.

 

81.08 Laparoscopic Surgery for Rectal Prolapse: Short-Term Outcomes Should Not Dictate the Approach

Posted on January 18, 2018

P. L. Rosen1, D. J. Gross1, H. Talus5, V. Roudnitsky2, M. Muthusamy3, G. Sugiyama4, P. J. Chung3  1State University Of New York Downstate Medical Center,Department Of Surgery,Brooklyn, NY, USA 2Kings County Hospital Center,Division Of Trauma And Acute Care Surgery,Brooklyn, NY, USA 3Coney Island Hospital,Department Of Surgery,Brooklyn, NY, USA 4Hofstra Northwell School Of Medicine,Department Of Surgery,Hempstead, NY, USA 5Kings County Hospital Center,Department Of Surgery,Brooklyn, NY, USA

Introduction:
Full thickness rectal prolapse is a debilitating condition for which multiple surgical approaches have been described. Laparoscopic transabdominal approaches are frequently employed, but there is a paucity of data comparing outcomes between laparoscopic transabdominal rectopexy (LR) and laparoscopic transabdominal rectopexy with sigmoidectomy (LRS). Using the American College of Surgeons National Surgical Quality Improvement Program (ACS NSQIP) database, we compared outcomes between these two commonly employed modalities. 

Methods:
Using ACS NSQIP 2010-2015, we identified cases in which LR (CPT 45400) or LRS (CPT 45402) were performed for a postoperative diagnosis of rectal prolapse (ICD 9 569.1). We excluded cases with missing sex, race, BMI, functional status, and ASA classification data. Outcomes of interest included length of stay (LOS), postoperative major morbidity (wound infections, pulmonary complications, cardiovascular complications, renal complications, sepsis/septic shock, bleeding, return to OR) and mortality.  LR and LRS cases were matched using propensity scores. Matching diagnostics were performed and outcomes were evaluated using conditional logistic regression or the Wilcoxon rank-sum test.

Results:
We identified 1,397 patients of which 841 (60.2%) underwent LR and 556 (39.8%) underwent LRS. Patients undergoing LR tended to be older (mean 61.6 vs 55.8 years, p<0.0001), had lower rates of independent functional status (95.7% vs 98.6%, p=0.0072), had higher proportion of African American race (3.8% vs 2.5%, p<0.0001), diabetes treated with medication (7.3% vs 3.8%, p=0.022),  CHF (1.31% vs 0.0%, p=0.0043), bleeding disorders (2.3% vs 0.72%, p=0.031), and ASA class 3 (39.6% vs 30.4%, p=0.0045). Unadjusted comparison between LRS and LR showed increased LOS (median 4 vs 2 days, p<0.0001), increased rates of superficial surgical site infection (SSI) (2.7% vs 0.6%, p=0.0019), bleeding (3.1% vs 1.3%, p=0.036), and sepsis (1.8% vs 0.5%, p=0.025). Propensity scores were then used to match 509 LRS to 509 LR cases with diagnostics showing that the groups were well-balanced across covariates. Conditional logistic regression demonstrated that LRS compared to LR had no statistically significant increased risk of 30-day postoperative complications or mortality. However LRS was associated with increased LOS compared to LR (median 4 vs 2 days, p<0.0001). 

Conclusion:
In this large observational study utilizing a national clinical database we found no differences in 30-day postoperative outcomes between laparoscopic transabdominal rectopexy without sigmoidectomy versus laparoscopic transabdominal rectopexy with sigmoidectomy after propensity score matching. This suggests that long-term outcomes should dictate the choice between these two procedures.

81.09 Proximal Intestinal Diversion for Colorectal Anastomoses: What Are We Preventing?

Posted on January 18, 2018

A. Sunkerneni1, R. J. Kucejko1, D. E. Stein1, J. L. Poggio1  1Drexel University College Of Medicine,Surgery,Philadelphia, PA, USA

Introduction:  Proximal intestinal diversion is often thought to protect colorectal anastomoses from leak, and reduce leak-related morbidity.  While data for rectal cancer convincingly shows improvement in outcomes, the results for abdominal anastomoses are mixed.  Yet, patients are being diverted, incurring the increased morbidity and mortality surrounding the reversal of stomas without clinical benefit.  Prior studies of abdominal anastomoses have been done on limited patient numbers, and did not control for pre-operative morbidity.  The American College of Surgeons’ (ACS) National Surgical Quality Improvement Program (NSQIP) has collected the largest dataset on anastomotic leak to date.  Our aim is to determine risk-adjusted post-operative outcomes including anastomotic leak on patients undergoing colorectal anastomosis to determine which patients receive the most benefit from diversion.

Methods:  A retrospective analysis of the ACS NSQIP Procedure Targeted Colectomy database from 2012 to 2015 was performed. All patients 18 to 90+ years old in the targeted colectomy database were included.  Patients were excluded if any CPT code corresponded to a surgery that did not have an anastomosis, or a surgery with a pelvic anastomosis, or if their leak status was unknown.  Post-operative outcomes were analyzed using chi-squared and Mann-Whitney U tests.   Propensity score-matched cohorts were developed using the NSQIP morbidity score.  The primary outcomes were 30-day mortality, 30-day reoperation rate, and anastomotic leak.

Results: 61,161 patients underwent abdominal colorectal anastomosis over the 4-year period. 8,352 (13.7%) underwent emergent surgery, and were diverted 30.5% of the time, compared to the non-emergent diversion rate of 11.4%.  Matched patients who were emergently diverted had significant improvements in prolonged intubation, septic shock, 30-day return to OR, mortality rate, and operative leak rate, with no significant difference in other outcome measures.  Matched patients who were non-emergently diverted were only noted to have significant improvements in post-operative pneumonia, rate of reintubation, mortality and operative leak rate.  More importantly, these patients had significantly worse rates of organ space infections, AKI, UTI, DVT, sepsis, return to OR and 30-day readmission. 

Conclusion: Proximal fecal diversion for abdominal colorectal anastomosis is a known trade-off between immediate protection and long-term morbidity.  In emergent cases, significant improvements in mortality and leak rate are seen without significant rises of other complications, suggesting the right patients are being diverted.  But, in non-emergent cases, many operative sequela are made worse by diversion, with only modest improvements in leak rates and 30-day mortality.  Surgeons would benefit from a decision tool to better stratify patients undergoing non-emergent abdominal anastomoses to aid in optimal patient selection.

81.07 Revolving Door: The Impact of Length of Stay on Readmissions After Colon and Rectal Operations.

Posted on January 18, 2018

D. Peterson1,3, F. Guzman1, L. Yu4, W. Cirocco1, A. Harzman1, A. Traugott1, M. Arnold1, S. Husain1  1Ohio State University,Colon And Rectal Surgery,Columbus, OH, USA 3Penn State Hershey Medical Center,Surgery,York, PA, USA 4Ohio State University,Biostatistics,Columbus, OH, USA

Introduction:
With recent emphasis on pay for performance model for surgery, length of stay and readmission rates have come under renewed scrutiny. An inverse relationship between length of stay and readmission rates has been suggested raising concerns that early discharges may in fact lead to higher readmission rates. We sought to evaluate the relationship between length of stay and readmission rates and the impact of surgical approach, patient demographics and postoperative complications.  

Methods:
Retrospective chart review was conducted of all colorectal surgeries from September 1, 2011-August 31, 2016 at a tertiary medical center. Patient demographics, comorbidities, postoperative complications, length of stay and readmission rates were evaluated. Logistic regression used to evaluate continuous predictors and Fisher exact test used to evaluate categorical predictors.

Results:
A total of 1319 patients were included. The average length of stay was 10.3 days (median: 7) and 260 (19.7%). At least one complication was noted in 226 (17.36%) patients. Univariate analysis revealed that longer hospital stays correlated with higher readmission rates (R= 0.015, p= 0.00953). However, this association lost its statistical significance with multivariate analysis (p=0.858). Multivariate analysis also revealed that the both increased length of stay and readmissions were strongly associated with pre-existing patient comorbidities (age, diabetes, BMI, COPD, renal dysfunction) as well as postoperative complications (wound infection, abdominal abscess, SVT, PE, pneumonia, UTI). Furthermore, utilization of laparoscopic surgery had statistically significant association with shorter length of stay and lower readmission rates (p=0.004 and 0.02 respectively).

Conclusion:

While length of stay is associated with readmission rates in univariate analysis, this relationship is lost when factors like patient comorbidities, operative outcomes and surgical approach are taken into consideration. Our results also indicate that pre-existing comorbidities and postoperative complications result in prolonged hospitalization and increased readmission rates. Finally, our study not only confirms the well documented beneficial effect of laparoscopic approach on length of stay, it also indicates that minimally invasive approach results in lower readmission rates. While most pre-existing medical conditions leading to longer hospital stays and readmissions are not modifiable, a concerted effort is necessary to minimize postoperative complications and to promote utilization of minimally invasive platform.
 

81.03 Incidence and Risk Factors of C. difficile Infection in Patients with Ileal Pouch-Anal Anastomosis

Posted on January 18, 2018

P. D. Strassle1,3, J. Samples1, E. E. Sickbert-Bennet2,3, D. J. Weber2,3, T. S. Sadiq1, N. Chaumont1  1University Of North Carolina At Chapel Hill, School Of Medicine,Department Of Surgery,Chapel Hill, NC, USA 2University Of North Carolina At Chapel Hill, School Of Medicine,Department Of Medicine,Chapel Hill, NC, USA 3University Of North Carolina At Chapel Hill, Gillings School Of Global Public Health,Department Of Epidemiology,Chapel Hill, NC, USA

Introduction:  In the last 10 years, recognition of Clostridium difficile infection (CDI) in patients with ileal pouch-anal anastomosis (IPAA) has been increasingly recognized. Despite the growing body of literature, conclusions about the incidence and risk factors of CDI in IPAA patients have been limited by single-institution studies, small sample sizes, and short follow-up. The goal of this study was to estimate the incidence and potential risk factors of CDI in patients with IPAA.

Methods: Patients diagnosed with ulcerative colitis, Crohn’s disease, or familial adenomatous polyposis, and undergoing an ileal pouch procedure between 2004 and 2013 in the Truven Health Analytics MarketScan® database were eligible for inclusion. Patients were required to have health insurance coverage for at least 6 months before and 30 days after surgery.  

Kaplan-Meier survival curves were used to estimate the 2-year risk of infection. CDI was identified using ICD-9 CM code 008.45, which has 78.0% sensitivity and 99.7% specificity. Multivariable Cox proportional hazard regression was used to assess the effect of potential risk factors. Risk factors included patient demographics, Charlson comorbidity score, pre-operative CDI (within 6 months of surgery), recent hospitalization (within 30 days of surgery), and use of corticosteroids, biologics, and immunomodulators (within 30 days of surgery). Inverse-probability of censor weights were used to account for differential follow-up. Age was modeled as a linear variable and centered at 40 years old. 

Results: 2,900 patients were included in the analysis. The median follow-up time was 628 days (IQR 287-730). The 2-year cumulative incidence of C. difficile was 3.3% (n=77). Twelve cases (15.6%) occurred during the surgical hospitalization. Patients with previous CDI (HR 7.33, 95% CI 3.85, 13.94) and patients taking corticosteroids (HR 2.19, 95% CI 1.30, 3.71) or biologics (HR 3.59, 95% CI 1.39, 9.24) prior to surgery were significantly more likely to have a CDI after IPAA. No significant differences in the risk of CDI across gender (p=0.51), age (p=0.70), Charlson score (p=0.99), history of recent hospitalization (p=0.50), or immunomodulator use (p=0.30) were seen.

Conclusion: The 2-year incidence of CDI after IPAA is at least 3%. Patients with a history of pre-operative CDI, and those taking corticosteroids or biologics before surgery are more likely to develop a CDI after surgery. 

81.01 Identifying Factors that Decrease Utilization of Adjuvant Chemotherapy in Stage III Colon Cancer

Posted on January 18, 2018

P. M. Schroder1, M. C. Turner1, B. Ezekian1, Z. Sun1, M. A. Adam1, C. R. Mantyh1, J. Migaly1  1Duke University Medical Center,Department Of Surgery,Durham, NC, USA

Introduction:  Standard of care for stage III colon cancer includes adjuvant chemotherapy, which increases survival by nearly 30%. Despite these results, many patients fail to receive adjuvant chemotherapy. We aim to describe the benefit of adjuvant chemotherapy for stage III colon cancer and to determine factors that influence the likelihood of receiving this treatment. 

Methods:  We queried the National Cancer Data Base 2006-2013 for patients with a single primary stage III colon adenocarcinoma and defined two groups: patients who did and did not receive adjuvant chemotherapy. Subgroup analyses were performed for healthy patients (Charlson-Deyo [CD] score = 0), comorbid patients (CD score ≥ 2), and those with post-operative complications (readmitted within 30 days of surgery). Kaplan-Meier (KM) curves were generated and Cox proportional hazard ratios (HR) were calculated to compare overall survival. Odds ratios (OR) for receiving chemotherapy were calculated to identify factors associated with failure to receive adjuvant chemotherapy. 

Results: Of the 74,588 patients included in this study, 54,235 received adjuvant chemotherapy and 20,353 did not. Overall survival was significantly better in the group that received adjuvant chemotherapy (HR of 0.477, p<0.001). Similar results were obtained in our subgroup analyses (see Figure). Adjuvant chemotherapy conferred a survival advantage for healthy patients (HR 0.485, p<0.001), comorbid patients (HR 0.492, p<0.001), and those with post-operative complications (HR 0.358, p<0.001). Several factors were associated with a reduced likelihood of receiving chemotherapy including older age (OR 0.9, p<0.001), black race (OR 0.728, p<0.001), comorbid patients with CD score ≥2 (OR 0.563, p<0.001), positive surgical margins (OR 0.83, p<0.001), and those with post-operative complications (OR 0.605, p<0.001). Patients with private insurance (OR 1.997, p<0.001) or Medicare (OR 2.184, p<0.001) were comparatively more likely to receive adjuvant chemotherapy.

Conclusion: We demonstrate a consistent survival benefit with adjuvant chemotherapy for patients with stage III colon cancer, even for comorbid patients or those with early post-operative complications. Factors such as older age, black race, more comorbidities, positive margins, post-operative complications, and lack of insurance were associated with a reduced likelihood of receiving adjuvant chemotherapy. These data suggest that adjuvant chemotherapy remains critically important for all patients with stage III colon cancer, but particular attention should be paid to utilizing this therapy in higher risk and underserved patients to avoid undertreating these vulnerable populations.

76.05 Should Sentinel Lymph Node Biopsy Be Recommended to All Intermediate Thickness Melanoma Patients?

Posted on January 18, 2018

A. Hanna1, A. J. Sinnamon1, R. Roses1, R. Kelz1, D. Elder1, X. Xu1, B. Pockaj2, D. Fraker1, G. Karakousis1  1University Of Pennsylvania,Philadelphia, PA, USA 2Mayo Clinic,Phoenix, AZ, USA

Introduction:

Sentinel lymph node (SLN) biopsy is routinely recommended for patients with intermediate (1.01 – 4.00 mm) thickness melanoma. Prior institutional data from our group,however, suggested significant variation in the risk for SLN metastasis for these patients and we therefore sought to identify subgroups within this cohort with low risk for SLN positivity using a large national data set.

Methods:

Patients with intermediate thickness melanomas who underwent SLN biopsy from 2010 to 2013 were identified using the National Cancer Database. Clinical and pathologic variables associated with SLN positivity were analyzed using logistic regression. Classification and Regression Tree (CART) analysis was used to risk-stratify patients for SLN positivity.

Results:

Of the 23,440 study patients with intermediate thickness melanoma, 14.7% (95% CI, 14.2% – 15.1%) were found to have a positive SLN. Most (59.9%) patients were male and the median age was 62 years (IQR, 51 – 72 years old). In multivariate logistic regression, increased age (OR = 0.89/10 years, 95% CI 0.88 – 0.90), female gender (OR = 0.85, 95% CI 0.79 – 0.93), absence of lymphovascular invasion (LVI) (OR = 0.31, 95% CI 0.27 – 0.36), absent mitoses (OR = 0.61, 95% CI 0.54 – 0.70), a H&N, upper extremity, or shoulder primary site (OR = 0.55, 95% CI 0.49 – 0.63), decreased thickness (OR = 1.55/mm, 95% CI 1.48 – 1.63), and absent ulceration (OR = 0.74, 95% CI 0.68 – 0.81) all were significantly associated with having a negative SLN. In CART analysis, absent LVI, thickness < 1.7 mm, age < 56, and primary site were significant branch points (Figure 1). In patients 56 years of age or older with absent LVI and intermediate thickness lesions < 1.7 mm (29% of all patients analyzed), the rate of SLN positivity was < 5%.

Conclusion:

Despite a SLN positivity rate of 14.7% overall, there exists significant heterogeneity in the risk for SLN metastasis in patients with intermediate thickness melanoma. In a sizable group of patients (nearly 30% undergoing the procedure), the risk for SLN metastasis approximates that seen in lower risk thin melanomas, where the procedure is offered selectively. For these patients (56 years or older with lower depth intermediate lesions and absent LVI) careful consideration should be made weighing the risks and benefits of the SLN procedure.

73.09 Resident Involvement and Outcomes after Surgery: A Double Edge Sword

Posted on January 18, 2018

M. Zeeshan1, M. Hamidi1, A. Tang1, E. Zakaria1, N. Kulvatunyou1, A. Jain1, L. Gries1, T. O’Keeffe1, B. Joseph1  1University Of Arizona,Tucson, AZ, USA

Introduction:
Diverticular disease is one of the leading causes for outpatient visits and hospitalizations. Resident participation in surgical procedures is essential for training. However, there is paucity of data regarding the outcomes after resident involvement in surgical procedures for diverticulitis. The aim of our study was to determine if the resident participation in surgery correlates with outcomes for patients undergoing surgical procedures in diverticulitis.

Methods:
We analyzed the National Surgical Quality Improvement Program database (2005-2012). We included all patients who had diagnosis of diverticular disease and underwent surgical management. Patients were stratified into two groups based on presence of resident during surgery: attending alone (No-RES) vs. attending with resident (RES). Groups were matched using propensity score matching for demographics, surgical procedure, morbidity probability and comorbidities. Outcomes of interest were compared for patients with and without resident participation in surgery (RES vs no-RES). We performed a sub-analysis of RES group by dividing it into junior (PGY 1-3), and senior residents (PGY 4-5), and fellows (PGY ≥ 6).

Results:
26,172 patients met the inclusion criteria, of which 6912 (3456: No-RES, 3456: RES) were matched. Mean age was 58.8 ± 14.3 years, and 46.7% were males. There was no difference in mortality in both groups (p=0.58), however, overall 30-d complication rates were higher in RES group (18% vs. 15.1%, p<0.01). Operative time (OR time) was longer in the RES group (175 min vs. 142 min, p<0.01), while there was no difference of hospital length of stay (HLOS) between the two groups (p=0.17). Table 1 shows the sub analysis based on level of residency. Mortality rate was highest in senior residents (p<0.01), while operative time was highest in operation performed by fellows (p<0.01).

Conclusion:
Resident involvement in surgical management of diverticulitis increases the rate of complications without an increase in mortality. Resident involvement is an important component of surgical residency. Identifying the factors and increased supervision by attendings may lead to improved outcomes. 
 

73.10 Management of Acute Cholecystitis with Significant Risk of Common Bile Duct Stone:The ‘SaFE’ Approach

Posted on January 18, 2018

K. O. Memeh1, S. Jhajj1, K. Tran1, R. A. Berger1,2, T. S. Riall1, A. Aldridge1,2  1University Of Arizona,Surgery,Tucson, AZ, USA 2Flagstaff Medical Center,Surgery,Flagstaff, AZ, USA

Introduction:

About 3-8% of acute calculous cholecystitis (ACC) present with common bile duct stone (CBDS). The 2010 American Society of Gastrointestinal Endoscopy (ASGE) and the 2016 World Society of Emergency Surgery (WSES) guideline on the management of gallstone with significant risk(high risk[HR] and intermediate risk[IR]) of CBDS recommend pre-operative imaging and ERCP for patient with IR and HR for CBDS respectively. Our group adopted a different approach; primary laparoscopic cholecystectomy (LC) with intraoperative cholangiogram (IOC) for all patients HR and IR for CBDS, and then proceed with intra-operative ERCP (IOERCP) for patients with positive IOC, with the intention of reducing length of stay (LOS) and hospital cost (HoC) without negatively impacting outcome.We believe that this approach is Safe, Fast and cost Effective ( ‘SaFE’) and we thus review the outcome of the ‘SaFE’ approach and compares it with the traditional (ASGE/WSES guided) approach.

Methods:

We retrospectively reviewed the medical record of consecutive patients, 18 years and older presenting with ACC with significant risk for CBDS who underwent LC + IOC +/- IOERCP between Jan 2015 and Feb 2017 in our institution. Patients with cholangitis and pre-operative imaging suggestive of CBD mass (other than stone) were excluded. Patients were stratified into ASGE Intermediate risk (ASGE-IR) and ASGE High risk (ASGE-HR) for CBDS based on the published ASGE criteria. We reviewed pre-operative liver function test, total bilirubin and imaging.Complications( cystic duct leak, post ERCP pancreatitis) and hospital charges (HoC) were evaluated. The student t-test was utilized to analyse difference in LOS when compared to similar patients managed prior to the implementation of the SaFE approach.

Results:

A total of 568 patients presented with ACC and suspicion for CBDS, hence had LC + IOC. IOERCP was performed for positive IOC in 87(15%) patients. Of the 87 patients, 34(39%) was ASGE-HR for CBDS.Medain pre-op T bil was 4.1 and 0.8 for ASGE HR and IR respectively.2 IR patients had negative IOERCP. Average LOS was 1.8 days for both HR and IR patient groups. There was no cystic duct leak and no conversion to open cholecystectomy in any of the 87 patients. Two (1 patient per group) had mild post ERCP pancreatitis. Mean HoC was $10,099 per patient.Prior to implementing the SaFE approach( i.e using the  ASGE/WSES guideline),similar cohort of patients had an average LOS of 3.4 days( p < 0.000) , and mean HoC of $14,320 a diffence of $2,941 with estimated cost saving of $255,867 in the 2 year period.

Conclusion:

Our findings suggest that ACC patients who are ASGE-HR, WSES- HR, and ASGE-IR for CBDS could be managed similarly using the ‘SaFE’ approach with significant reduction in both LOS and HoC without any increase in procedure-related morbidity. 

 

 

 

 

 

 

73.07 Trends in Mortality and Cardiac Complications in Major Abdominal Surgery by Operative Volume.

Posted on January 18, 2018

Y. Sanaiha1, Y. Juo1, K. Bailey1, E. Aguayo1, A. Iyengar1, V. Dobaria1, Y. Seo1, B. Ziaeian2, P. Benharash1  1David Geffen School Of Medicine, University Of California At Los Angeles,Cardiac Surgery,Los Angeles, CA, USA 2David Geffen School Of Medicine, University Of California At Los Angeles,Cardiology,Los Angeles, CA, USA

Introduction:

Cardiovascular complications are the leading cause of death following noncardiac surgery. Major abdominal operations represent the largest category of procedures considered to have elevated risk of cardiovascular complications. The current aim was to examine trends in the incidence of mortality, postoperative myocardial infarction, and cardiac arrest after major abdominal operations and to determine the presence of potential volume-outcome relationships. 

Methods:
We performed a retrospective analysis of the Nationwide Inpatient Sample (NIS) for patients having elective open gastrectomy, pancreatectomy, nephrectomy, splenectomy, and colectomy (major abdominal surgery: “MAS”) during 2008-2014. Chi-squared analysis was used to compare demographic and hospital characteristics between groups. Logistic regression was performed to determine predictors of in-hospital mortality, postoperative cardiac arrest (POCA) and myocardial infarction (POMI).  

Results:
Of the 1,300,794 patients undergoing MAS, 49,589(3.70%) experienced in-hospital mortality, 16,542 (1.24%) POMI, and 9,496 (0.76%) POCA. The annual all-cause mortality and POMI rates remained stable while the incidence of POCA steadily rose.  Average Elixhauser score also increased from 1.8 to 2.2 during this study period. Odds of mortality were significantly lower for medium and large volume hospitals compared to small volume hospitals after adjustment (Table). Hospital operative volume did not significantly impact the odds of POMI or POCA. In contrast, larger hospital bedsize was associated with higher odds of mortality and POCA. Subgroup analysis demonstrated lower odds of mortality with higher operative volume over 2008-2014 for all operations except for splenectomy. Significant risk factors for POMI/POCA included age > 65, peripheral vascular disease, and congestive heart failure, while female gender and higher income quartile had decreased odds of these complications (P<0.02). 

Conclusion:
The rate of POCA amongst patients having MAS has increased in the US without a concomitant rise in POMI or mortality. Hospital operative volume appears to reduce odds of postoperative mortality over the entire study period. The effect of operative volume on rate of postoperative cardiac complications is not consistent over time as odds of POCA are significantly lower for higher volume hospitals only in 2008-2011 population. Operative volume does not significantly impact risk of POMI or POCA in the 2012-2014 subgroup. Increased odds of mortality and POCA at larger hospitals by bedsize could reflect patient or hospital factors that are not well represented in NIS. Non-ischemic causes of POCA need further investigation to delineate opportunities for quality improvement. 
 

73.04 Perforated Peptic Ulcer Surgery: No Difference in Mortality Between Laparoscopic and Open Repair.

Posted on January 18, 2018

V. Gabriel1, A. Grigorian1, S. Schubl1, M. Pejcinovska1, E. Won1, M. Lekawa1, N. Bernal1, J. Nahmias1  1University Of California – Irvine,Division Of Trauma, Burns & Surgical Critical Care,Orange, CA, USA

Introduction:  The lifetime prevalence of perforated peptic ulcer (PPU) in patients with peptic ulcer disease is estimated at 5%. Reported mortality rates after surgery for PPU have ranged from 1 to 24%. A recent meta-analysis by Tan et al demonstrated equivalent morbidity and mortality when comparing laparoscopic repair (LR) to open repair (OR).  However, LR was shown to have lower operative time, less pain, shorter length of stay (LOS), and a lower rate of surgical site infection. We hypothesized a decrease in morbidity and mortality with LR from 2011-2015 compared to 2005-2010. Additionally, we hypothesized a decrease in morbidity and mortality for LR versus OR for the entire duration of 2005-2015.

Methods:  Patients undergoing operative repair of PPU between 2005- 2015 were identified in the NSQIP database by CPT code. Patients with definitive acid-reducing operations were excluded. A comparison of OR from 2005-2010 versus 2011-2015 was performed. A similar comparison was performed for LR. Additionally, a comparison between LR and OR for the entire duration (2005-2015) was conducted. Primary outcomes were the differences in 30-day mortality and overall morbidity. After controlling for significant covariables such as age, American Society of Anesthesiologists class, functional status, pre-operative albumin and creatinine, steroid use, liver disease, time to surgery, and presence of malignancy, a multivariate regression analysis was performed.

Results: 5,413 patients between 2005-2015 were included in the study. From 2005-2010 there were 86 LR cases and 1,924 OR cases.  Between 2011-2015 there were 221 LR cases and 3,182 OR cases. LR demonstrated no difference in 30-day mortality or overall morbidity between the two time periods (p>0.05). There was no significant difference in 30-day mortality for patients undergoing OR between the two time periods. However, overall morbidity (odds ratio (OR), 1.99; 95% CI, 1.71-2.33, p<0.05), development of sepsis (p<0.05), and septic shock (p<0.05) were all more prevalent in patients undergoing OR from 2011-2015. Comparing LR versus OR from 2005-2015, patients undergoing LR had a shorter length of stay (p<0.05), and were less likely to exhibit failure to wean from the ventilator at 2 days (OR, 0.34; 95% CI, 0.18-0.65, p<0.05). 

Conclusion: While a 2.5% increase LR utilization was seen, there was not a decreased morbidity and mortality associated with more recent LR from 2011-2015. This may be secondary to increasing utilization of LR in more debilitated patients over time. When LR was compared to OR there was a significant decrease in LOS. Future prospective research is needed to confirm this finding and evaluate the safety of more widespread adoption of LR for PPU.

 

73.02 Percutaneous Cholecystostomy in Acute Cholecystitis – Predictors of Recurrence & Cholecystectomy

Posted on January 18, 2018

M. N. Bhatt1, M. Ghio1, L. Sadri1, S. Sarkar1, G. Kasotakis1, C. Nasrsule1, B. Sarkar1  1Boston Medical Center,Department Of Trauma And Acute Care Surgery,Boston, MA, USA

Introduction:  Acute cholecystitis (AC) is a common acute illness, with the preferred treatment being cholecystectomy. However, in high-risk patients, a less invasive option of percutaneous cholecystostomy tube placement (PC) is preferable. Patients can subsequently either undergo interval cholecystectomy (IC) or PC can be utilized as definitive treatment. Currently, there is little evidence to guide patient care after PC. We sought to demonstrate the clinical outcomes of PC and identify the predictors of recurrent disease as well as successful IC.

Methods:  A retrospective chart review of patients undergoing PC for AC between 2008 and 2016 at a single tertiary care center was performed. Basic patient demographics, laboratory & imaging findings, and patient outcomes including mortality, readmissions, hospital length of stay (LOS), Intensive Care Unit (ICU) LOS, recurrence, and IC were collected. Univariate and multivariate analyses were performed using logistic regression, Wilcoxon Rank, and multi-variable logistic regression models.

Results: Of 145 patients, 96 (67%) had calculous and 47 (33%) had acalculous cholecystitis. PCs were performed in these patients due to their high preoperative risks; 72 (49%) had chronic prohibitive risks and 73 (51%) had acute prohibitive risks. There were 55 (38%) peri-procedural complications, 44 of which were PC dislodgment. Mean duration of PC was 93 days. Recurrence rate for AC was 18%; median duration to recurrence was 65 days. Patients with calculous cholecystitis were more likely to have AC recurrence (OR = 3.24, p = 0.018), whereas length of antibiotics course or duration of PC had no significant correlation with AC recurrence. 41 (28%) patients underwent IC. Patients with acute prohibitive risks and shorter antibiotics course (≤ 7 days) were more likely to undergo IC (OR = 6.66 & 2.10, p = <0.001 & 0.048), and most were completed laparoscopically (OR = 6.84, p = <0.0001). There were only two peri-operative complications and no peri-operative mortality. Mean hospital and ICU LOS were longer for patients with acalculous cholecystitis compared to calculous (22 vs. 11 days, p = <0.0001). 30-day readmission rate was 29%. Patients with acalculous cholecystitis had higher 30-day readmission rate (OR = 2.42, p = 0.020). 30-day mortality after PC was 9%. The follow up was for 26(3-53) months and survival analysis revealed that patients receiving IC had greater survival compared to PC as a definitive option.

Conclusion: PCs are a viable option for high-risk patients with AC. Calculous cholecystitis is a strong predictor of AC recurrence after PC. A longer (>7 days) antibiotics course is not associated with lower recurrence and should be avoided. Patients undergoing IC have better overall survival. PCs, although safe, should not be considered as a definitive treatment, especially in patients with acute critical illness where a successful IC can be performed laparoscopically with minimal complications.

 

73.03 Opioid Use after Surgery among Preoperative Intermittent Users

Posted on January 18, 2018

E. Harker1, C. A. Keilin1, R. Ahmed1, C. Katzman1, D. C. Cron1, T. Yao3, H. Hu1, J. S. Lee1, C. M. Brummett2, M. J. Englesbe1, J. F. Waljee1  1University Of Michigan,Department Of Surgery,Ann Arbor, MI, USA 2University Of Michigan,Department Of Anesthesiology,Ann Arbor, MI, USA 3University Of Michigan,School Of Public Health,Ann Arbor, MI, USA

Introduction:  A significant number of surgical patients intermittently take opioids prior to elective surgery. Understanding the clinical trajectory of this large number of patients is critical to optimizing their care. We hypothesize that a longer duration of preoperative opioid use will be correlated with a longer duration of postoperative use.

Methods:  We used a national employer-based insurance claims dataset to identify adults age 18 to 64 who were preoperatively either opioid-naïve or intermittent opioid users and who underwent a general, gynecologic, or urologic surgical procedure between January 2010 and March 2014 (N= 309,096). We defined preoperative intermittent opioid users as patients who filled ≤120 days’ supply of opioids between 365 and 31 days before surgery. Our primary explanatory variable was preoperative opioid exposure, measured as the number of months during which an opioid prescription was filled in the year prior to surgery (opioid-naïve, 1 month, 2-3 months, 4-6 months, 7-9 months, >9 months). Our outcome was time until last postoperative opioid script (considered the date of opioid discontinuation). We used survival analysis techniques, including Kaplan-Meier curves to compute estimated proportion of patients continuing to fill opioids postoperatively.

Results: In this cohort, 27% of patients used opioids intermittently in the year before surgery, and the majority of these patients (62%) filled opioids during 1 month preoperatively. Patients with a longer duration of preoperative opioid exposure continued to fill opioids for longer durations postoperatively (Figure). Most patients discontinued opioids after the initial prescription, but the remaining patients continued filling opioids long after surgery. Compared to patients with 1 month of opioid fills preoperatively, patients with >9 months of preoperative opioid fills had a 4-fold longer adjusted mean time until opioid discontinuation (326 vs. 84 days, P<0.001). The estimated proportion of patients continuing to fill ≥1 opioid script beyond 180 days was 90% among patients with >9 months of preoperative opioid use, 23% among patients with 1 month of preoperative use, and 15% among opioid-naïve.

Conclusion: Patients who intermittently use opioids prior to surgery are particularly vulnerable to prolonged postoperative opioid use. The surgical event should be considered an opportunity to wean opioid users postoperatively. Such strategies may have significant positive impact on the overall health and wellness of these surgical patients.

« Previous Page
Next Page »