62.04 Functional Inclusivity of the Northwest London Trauma Network

Posted on January 18, 2018

J. M. Wohlgemut1, J. Davies2, C. Aylwin2, J. J. Morrison3, E. Cole4, N. Batrick2, S. I. Brundage4, J. O. Jansen5  1Aberdeen Royal Infirmary,Department Of Surgery,Aberdeen, SCOTLAND, United Kingdom 2Imperial College Healthcare NHS Trust,St Mary’s Hospital, North West London Trauma Network,London, ENGLAND, United Kingdom 3University Of Maryland,R Adam Cowley Shock Trauma Center,Baltimore, MD, USA 4Queen Mary University Of London,Centre For Trauma Sciences, Blizard Institute,London, ENGLAND, United Kingdom 5University Of Alabama at Birmingham,Division Of Acute Care Surgery, Department Of Surgery,Birmingham, Alabama, USA

Introduction:
Inclusive trauma systems have been shown to improve survival compared to exclusive systems. Metrics exist to assess and validate trauma system outcomes; however, these are clinically focused and do not evaluate the appropriateness of admission patterns, relative to geography and triage category. This is an important consideration in system performance, as the referral of triage negative, non-severely injured patients to a higher echelon of care can overwhelm capability and dilute clinical experience. We propose to term this phenomenon “functional inclusivity” and to define it as the proportion of triage negative, non-severely injured patients, who were injured in proximity to a Level II/III trauma center but were admitted to a Level I facility. The aim of this study was to evaluate the usefulness of this metric in the context of the Northwest London Trauma Network.

Methods:
Retrospective, geospatial, observational analysis of registry data from the Northwest London Trauma Network. We included all adult (≥16 years) patients who were transported by road directly from the scene to the Level I trauma center between 1/1/13 – 31/12/16. Incident location data were geocoded into longitude/ latitude, and drive-times were calculated from each incident location to each hospital in London’s Trauma System, using Google Maps and R. We determined the number and proportion of non-severely injured patients and triage-negative patients, who were injured in closer proximity to a Level II/III but taken to the Level I. We evaluated changes in these metrics over time.

Results:
Of 2051 patients, 907 (44%) were severely injured (ISS ≥15), and 1144 (56%) non-severely injured (ISS 1-15). 795 of the 1144 non-severely injured patients (69%) were injured in closer proximity to a Level II/III than the Level I, but nevertheless taken to the Level I. Similarly, 488 patients were triage-negative, and 229 (47%) of these were injured in closer proximity to a Level II/III than the Level I, but nevertheless taken to the Level I. Patient volume increased over time, from 23 to 85 severely injured patients, and 25 to 85 non-severely injured patients per quarter. However, although the number of non-severely injured and triage-negative patients also increased over time, the corresponding proportions have remained the same.

Conclusion:
This study has demonstrated the potential value of the concept of functional inclusivity in characterising trauma system performance. Further work is required to establish what constitutes an acceptable level of functional inclusivity, and what the denominator should be. The Northwest London Trauma Network is a maturing system. Though the overall number of trauma patients increased over the study period, system exclusivity is not increasing. Future research should focus on validating and further evaluating the concept of functional inclusivity across the London Trauma System as a whole, and in other settings. 
 

62.03 Trauma Patients have Improved Access to Post-Discharge Resources through The Affordable Care Act

Posted on January 18, 2018

A. Y. Williams1, S. N. Davis1, S. K. Hill1, P. Connor1, Y. Lee1, C. C. Butts1, J. Simmons1, S. Brevard1, L. Ding1  1University Of South Alabama Medical Center,Department Of Trauma, Surgical Critical Care And Burn Surgery,Mobile, AL, USA

Introduction: The state of Alabama is one of 19 states opting out of Medicaid expansion under the Affordable Care Act. Many injured patients require post-hospital discharge specialty services and it has been demonstrated that the uninsured have less access to such services and are more likely to be discharged directly to home. Previous studies in Medicaid-expanded states have demonstrated up to a 20% decrease in the uninsured trauma population. The purpose of this study was to evaluate our trauma population payer mix before and after the implementation of the ACA and its effect on disposition.

Methods: This was a retrospective review of adult trauma patients treated at a level I trauma center with a large rural catchment area between 2011-2012 (pre-ACA) and 2014-2015 (post-ACA). Demographics, payer status, patient outcomes and disposition were recorded. 

Results: We identified 3716 patients in the pre-ACA period and 3657 patients in the post-ACA period. Patients in the post-ACA period were more likely to be insured (59.37% vs 55.14%, p<0.001) and utilize post-discharge services (13.48% vs. 9.28%, p<0.001) when compared to the pre-ACA period.  Medicaid patients are more likely than uninsured patients to receive home health services (OR 6.24, CI 2.74-14.18) or intermediate nursing care (OR 16.88, CI 7.05-40.43).

Conclusion: After the implementation of the ACA, our trauma center experienced a statistically significant decrease in the overall number of uninsured trauma patients with increased utilization of extended services after discharge.  However, in our state without Medicaid expansion, this still leaves 40% of our patient population without health insurance.

 

62.02 Are Trends in Fatal and Non-Fatal Firearm Related Injuries in Miami-Dade County Racially Divergent?

Posted on January 18, 2018

A. B. Padiadpu1, S. A. Eidelson1, J. Stoler2, M. B. Mulder1, R. Rattan1, T. L. Zakrison1  1University Of Miami,Ryder Trauma Center, The DeWitt Daughtry Family Department Of Surgery, Miller School Of Medicine,Miami, FL, USA 2University Of Miami,Department Of Geography And Regional Studies, Department Of Public Health Sciences,Miami, FL, USA

Introduction: Firearm related injuries are a significant public health problem in Florida and a growing cause of preventable morbidity and mortality. However, the associated sociodemographic factors remain poorly examined. This study addressed this gap by assessing the racial, ethnic, gender and age demographics of firearm related injuries in the region of Miami-Dade County over a 20-year period.

 

Methods: We retrospectively reviewed all patients involved in firearm incidents who were treated at the only American College of Surgeons verified Level 1 trauma center in Miami-Dade County during the study period. We analyzed the demographic trends and tested for significance (p<0.05) using chi-square test and ANOVA test. Parametric data was expressed as mean ± SD and nonparametric data as median (interquartile range).

 

Results: During 1992-2012, 11839 patients were transported to our trauma center after a firearm-related injury. The population was 30±13 years, 90% male, 66% Black and 34% White. 19% of the population was Hispanic (91% White).  97% of the incidents were intentional. Patients had an Injury Severity Score (ISS) of 14±15 with hospital length of stay (LOS) 2(0-8) days and 13% mortality. The injury and mortality rates trended downward until 2000 and then upward through 2012, more so among blacks (p≤0.001). Black males comprised 60% of the population with mean age of 28±11yrs, ISS of 14±15, LOS 3(0-8) days and 12% mortality. Among these patients, mortality has trended upwards since 2000 (p≤0.001). For patients ≤18 years of age (14% of the population), the median age was 17(16-18) years, ISS 13±15, LOS 2(0-7) days and 10% mortality rate. In this category, mortality has been on the rise over the last 12 years of the study period (p=0.09), significantly so for black pediatric population (p≤0.001).

 

Conclusion: This analysis reveals a recent and sustained increase in firearm related injuries and associated mortality since 2000, an alarming trend that has been driven mainly by the rise in injuries inflicted on younger Black males. These results beckon a focused evaluation and public health intervention to prevent firearm related violence in this highly vulnerable population.

62.01 High Value Care through Standardization of Continuous Renal Replacement Therapy

Posted on January 18, 2018

J. Tseng1, N. Minissian2, R. J. Halbert2, P. Hain2, T. Griner2, H. Rodriguez2, S. Barathan2, R. F. Alban1  1Cedars-Sinai Medical Center,Department Of Surgery,Los Angeles, CA, USA 2Cedars-Sinai Medical Center,Los Angeles, CA, USA

Introduction:
Continuous Renal Replacement Therapy (CRRT) is widely used to manage renal failure in critically ill patients.  Though CRRT utilization has increased nationwide, it is significantly costlier in comparison to intermittent hemodialysis.  Furthermore, the indications for initiation, maintenance, and cessation lack consensus and are not standardized.  In order to target high value care at our institution, we organized a multidisciplinary task force to evaluate current utilization patterns of CRRT, standardize its usage, and assess the outcomes of our intervention.

Methods:
A multidisciplinary task force consisting of intensivists, nursing staff, and nephrologists was established in October 2015 to assess and promote high value utilization of CRRT.  This team created a set of evidence-based guidelines to standardize the initiation, maintenance, and termination of continuous dialysis.  Other interventions were implemented to improve transparency regarding CRRT, mandate daily communication between medical teams and ancillary staff, encourage goals of care discussion, revise electronic order sets, and curb excess lab orders related to dialysis.  Patients receiving CRRT from fiscal years 2013 to 2017 before and after the intervention were compared.

Results:
The total volume of patients on CRRT increased by 104% (216 to 440) from 2013 to 2016, and decreased to 326 patients in 2017.  Similarly, the total number of CRRT days increased by 120% (1490 to 3285) from 2013 to 2016, and decreased to 1879 days in 2017. Prior to our intervention, the average duration of CRRT peaked at a mean of 7.69 ± 7.46 days, or a median of 8 (IQR 3-10) in 2015.  After our intervention, the average duration of CRRT decreased to a mean of 5.76 ± 4.50, or a median of 4 (IQR 3-8) in 2017 (p=0.018).  Solid organ transplant patients utilized continuous dialysis for longer durations compared to non-transplant patients.  The total direct cost of CRRT per case decreased from $12167.44 in 2013 to $10545.96 in 2017, translating to 13% cost savings.  Upon termination of CRRT, the proportion of patients who expired on CRRT decreased from 26.4% in 2013 to 5.8% in 2016, while the proportion of patients expected to transition to hospice care increased from 21.7% to 53.1%.  An increasing number of patients were enrolled in hospice upon hospital discharge after our intervention, from 0.4% in 2014 to 8.8% in 2017 (p<0.001).

Conclusion:
By establishing a task force to critically review utilization of continuous, standardize its usage, and promote daily communication regarding patient progress and goals of care, we significantly reduced the cost and duration of CRRT.  In addition, our intervention was also associated with fewer patients expiring on continuous dialysis, and more patients transitioned to comfort care measures.  Solid organ transplant patients utilize CRRT at higher rates than non-transplant patients, and may be the focus of further efforts to achieve high value care.
 

6.17 National Practice Trends for the Management of Lung Cancer: A Dartmouth Atlas Study

Posted on January 18, 2018

I. C. Bostock1, F. Sheikh1, T. M. Millington1, D. J. Finley1, J. D. Philips1  1Dartmouth Hitchcock Medical Center,Thoracic Surgery,Lebanon, NH, USA

Introduction:
Anatomic resection is the standard of care for early-stage lung cancer. Video-assisted thoracoscopic surgery (VATS) has been established as a safe and effective alternative to an open approach. The aims of this study were to: 1) Characterize open versus VATS surgical practice trends for the management of lung cancer in the United States, and 2) Describe if particular regions of the country utilize minimally invasive surgery more frequently.

Methods:
Using the Dartmouth Atlas Rate Generator, the population of Medicare beneficiaries from the ages of 65 to 99 years with full Part A and B coverage and no HMO coverage for the years of 2006 and 2014 was selected. The diagnosis of lung cancer (ICD-9 codes: 162.0 162.2 162.3 162.4 162.5 162.8 162.9) with subsequent selection of CPT codes to describe thoracoscopic (32601, 32607, 32655, 32663, 32666, 32667, 32668, 32669, 32670, 32671) and open lung resections (32096, 32097, 32141, 32505, 32506, 32507, 32608, 32440, 32442, 32445, 32480, 32482, 32484, 32486, 32488) were queried. Comparison of data between 2006 and 2014 with descriptive statistics and a univariate analysis were performed using student’s t-test and chi-square, as appropriate. A p-value <0.05 was considered statistically significant. Heat maps were generated based on the distribution of cases by geographic region.

Results:
A total of 24,368,333 and 23,921,059 patients for the years of 2006 and 2014, were analyzed. A diagnosis of lung cancer was made in 167,418 patients (0.7%) and 167,506 patients (0.7%) in 2006 and 2014 (p=0.7), respectively. A surgical intervention was performed in 17,249 patients (10.3%) during 2006 and 18,603 patients (11.1%) in 2014 (p=0.01). A VATS approach was performed in 2,512 patients (15%) during 2006 and 9,578 patients (54%) during 2014 (p=0.01). In 2006, California, New York, and New Jersey performed the most VATS procedures, in comparison to 2014, when New York, Florida, and California performed the most VATS procedures. 

Conclusion:
The incidence of lung cancer in the United States was unchanged from the period of 2006 to 2014. A change in surgical practice patterns was evident, with a significant increase in the use of VATS techniques in more than 50% of cases after this eight-year period. This indicates that VATS has become the preferred technique in the population studied.  
 

6.18 ~~Factors During Training Which Predict Future Use Of Minimally Invasive Thoracic Surgery

Posted on January 18, 2018

P. E. Rothenberg1, B. D. Hughes2, I. C. Okereke1  1University Of Texas Medical Branch,Cardiothoracic Surgery,Galveston, TX, USA 2University Of Texas Medical Branch,Department Of Surgery,Galveston, TX, USA

Introduction:
~~Although the use of minimally invasive thoracic surgery has increased with time, the majority of patients undergoing lung and esophageal resections still receive an open approach.  We performed a national survey to analyze factors associated with a propensity to perform minimally invasive thoracic surgery after completing a cardiothoracic training program.

Methods:
~~All cardiothoracic surgery trainees in standard 2 or 3 year programs from 2012 to 2016 were sent an online survey regarding numbers and types of cases performed during their training, current practice patterns as attending surgeons and comfort level with minimally invasive thoracic surgery.  Responses were recorded and analyzed.

Results:
~~Sixty-one trainees responded.  Trainees performed a mean of 113 lobectomies (30-250) during their training, with a mean minimally invasive rate of 53 percent.  Trainees performed a mean of 42 esophagectomies (10-110) during training, with a mean minimally invasive rate of 29 percent.  A higher percentage of minimally invasive lobectomies, compared to all lobectomies, performed during training was associated with a higher percentage of minimally invasive lobectomies performed as an attending physician (p = 0.04) and a greater comfort level with minimally invasive lobectomy (p = 0.01).  A higher percentage of minimally invasive esophagectomies performed during training was associated with a higher rate of minimally invasive esophagectomies performed as an attending physician (p = 0.01) and a greater comfort level with minimally invasive esophagectomy (p < 0.01).  A trainee’s overall case number did not influence their rate of adoption of minimally invasive surgery as attendings for either lobectomy (p = 0.11) or esophagectomy (p = 0.06).

Conclusion:
~~Recent graduates who performed a greater number of overall cases during their training were not more likely to adopt minimally invasive techniques as attending physicians if those cases during training were not performed minimally invasively.  Identifying other factors during the early years of an attending physician’s career which make adoption of minimally invasive techniques more likely may help to increase further the overall prevalence of minimally invasive thoracic surgery nationwide.
 

6.15 Risk Factors Contributing to Cardiac Events Following Thoracic Endovascular Aneurysm Repair (TEVAR)

Posted on January 18, 2018

D. Acheampong1, P. Paul1, P. Boateng1, I. Leitman1  1Mount Sinai School Of Medicine,New York, NY, USA

Introduction:  Cardiac events (CE) following TEVAR have been associated with morbidity and mortality. A large risk-adjusted database was used to understand contributing factors. 

Methods:  A retrospective analysis was performed for completed procedures done from 2010-2015 using the American College of Surgeons -National Surgical Quality Improvement Program (ACS-NSQIP) participant user file. Adult patients (≥18 years) who underwent TEVAR were identified and 30-day outcomes were examined. Initial univariate analysis was conducted on all pre-operative risk factors. Univariate and multivariate analyses were performed to assess risk factors for CE following TEVAR. A P-value of < 0.05 was considered statistically significant.

Results: The study identified 130 out of 2403 (5.4%) patients who underwent TEVAR that developed cardiac events as defined by ACS-NSQIP. Pre-operative leukocytosis, ASA score ³3 and functional dependence were associated with CE post-TEVAR. Underlying major risk factors for CE included emergency operation (53.43% vs 18.58%, p<0.01), ventilator dependence (15.38% vs 0.17%, p<0.01), currently on dialysis (11.53% vs 3.77%, p<0.01), SIRS (19.23% vs 6.30%, p<0.01), sepsis (3.84% vs 0.87%, p<0.01) and septic shock (2.3% vs 0.01%, p<0.01). Patients with postoperative renal failure (3.84% vs 0.99%, p<0.01), unplanned return to operating room (21.53% vs 9.37%, p<0.01) and operation time >180mins (37.7% vs 26.8%, p<0.01) also had increased associated post-operative cardiac events.  CE greatly increased mortality (60.8% vs 2.2%).

Conclusion: CE following TEVAR is associated with significant mortality. Patients with identified risk factors should carefully observed following intervention.  

 

 

6.16 SURGICAL STAGING SUPERIOR TO PET SCAN FOR ASSESSMENT OF DISEASE RESPONSE FOR MEDIASTINAL LYMPHOMA

Posted on January 18, 2018

L. Kane1, H. Savas1, M. DeCamp1, A. Bharat1  1Feinberg School Of Medicine – Northwestern University,Chicago, IL, USA

Introduction:

Mediastinal lymphoma affects young individuals, typically in the second through fourth decades of life, and constitutes over 7% of all lymphomas. The primary treatment modality remains systemic chemotherapy with or without radiation. Response to therapy is determined using PET scan. Unfortunately, in over 25% of patients, PET remains positive and it is unclear whether persistent PET avidity in the mediastinum represents residual disease or inflammatory changes resulting from therapy. Percutaneous image guided biopsy has typically resulted in poor accuracy due to the heterogeneity of the residual mass as well as the difficult nature of needle access. We hypothesized that minimally invasive thoracoscopic techniques would enable better sampling of the PET avid mediastinal lesion, allowing accurate assessment of residual disease following first-line treatment of mediastinal lymphomas.

Methods:
This is a retrospective analysis of a prospectively maintained database. Between January 2009 and December 2015, all patients (n=77) who underwent initial surgical incisional biopsy for diagnosis were included. The surgical biopsies were performed using minimally invasive techniques (video-thoracoscopy or robotic surgery) and required the surgeon to keep performing the biopsy until frozen section was positive or at least until the mass on the ipsilateral hemi-mediastinum was resected. Statistical analysis was performed by a biostatistician using SPSS software. 

Results:
Of the study cohort, 34 patients underwent surgical restaging for PET avid residual mass while 43 either had a complete response with no PET activity or were lost to follow up. The cohort of 34 patients included 76% Caucasians, 50% females, and had a median age of 28 years. The types of lymphoma were predominantly Hodgkins (32%) and Diffuse Large B cell Lymphoma (38%). In these 34 patients with residual PET activity, surgical biopsy revealed presence of lymphoma in 53% of patients. Patients detected to have persistent lymphoma revealed no significant difference in tumor volume reduction compared to those with no residual disease (51% versus 39%) and no significant difference in reduction in PET SUV (68% versus 60%). In all biopsies, significant adhesions between lung and mediastinum were noted, and the median length of the surgical procedure was 75 minutes. However, there were no surgical complications. The length of stay for all patients was less than 24 hours. All patients detected to have residual lymphoma underwent second-line therapy guided by the pathological analysis. 

Conclusion:
While the current standard for patients with mediastinal lymphoma presenting with residual PET activity after completion of first line therapy is surveillance alone, our data suggests that a large number of these patients have residual lymphoma which can be safely diagnosed using minimally invasive surgery. Detection of residual lymphoma has significant implications in further treatment of these patients. 

6.13 Using Mathematical Modeling To Define The Learning Curve In Robot-assisted Thoracoscopic Lobectomy

Posted on January 18, 2018

B. N. Arnold1, D. C. Thomas1, V. Bhatnagar1, J. D. Blasberg1, Z. Wang2, D. J. Boffa1, F. C. Detterbeck1, A. W. Kim3  1Yale University School Of Medicine,Section Of Thoracic Surgery,New Haven, CT, USA 2Yale University School Of Medicine,School Of Public Health,New Haven, CT, USA 3University Of Southern California,Division Of Thoracic Surgery,Los Angeles, CA, USA

Introduction:  Robot-assisted thoracoscopic (RobAT) lobectomy has been shown to be a safe approach to pulmonary lobectomy. There is a learning curve associated with integrating the robotic platform into thoracic surgery. This study sought to define, mathematically, the learning curve for RobAT lobectomy.

Methods:  All patients undergoing RoBAT lobectomy at a single academic medical center from 2010 through 2016 were considered. Covariates included patient demographics, comorbidities, operating time (ORT), length of hospital stay (LOS), estimated blood loss (EBL), and post-operative complications. A cumulative sum (CUSUM) analysis of ORT was performed to identify three distinct phases of the learning curve. Procedures converted to open were omitted from the analysis, but the number of conversions within each phase of RoBAT lobectomy was tallied.

Results: 101 patients met criteria for inclusion. CUSUM analysis identified two inflection points which stratified the population into three phases: cases 1-22, cases 23-63, and cases 64-101. There was a statistically significant difference in operating time and estimated blood loss between phases 1 and 2 (ORT p<0.05, EBL p=0.016), and between phases 1 and 3 (ORT p<0.05, EBL p=0.006). There was no difference in ORT or EBL between phases 2 and 3. There was no statistically significant difference in comorbidities, chest tube duration, LOS, or post-operative complications across the learning curve. Conversion rates in phase 1, 2, and 3 were 12% (3/25), 9% (4/45), and 3% (1/39), respectively. Of these conversions, 2/8 were emergent and occurred in phases 1 and 2.

Conclusion: Based on ORT, the learning curve for RoBAT lobectomy appears to be 22 cases, with mastery achieved after 63 cases. Differences in length of stay, chest tube duration, conversion rate, or complication rate were not observed during the learning phase. Other factors not measured in this study may play a role in the learning process and warrant further study. 
 

6.14 Minimally Invasive Versus Full Sternotomy AVR In Low-risk Patients — Which Will Stand Against TAVR?

Posted on January 18, 2018

S. A. Hirji1, F. Ramirez Del Val1, A. A. Kolkailah1, J. Lee1, S. F. Aranki1, P. S. Shekar1, T. Kaneko1  1Brigham And Women’s Hospital,Division Of Cardiac Surgery, Department Of Surgery,Boston, MA, USA

Introduction: Compared to aortic valve replacement (AVR) via full sternotomy (fAVR), minimally invasive AVR (mAVR) has been associated with improved results. The likely expansion of Transcatheter AVR (TAVR) to low-risk patients demands contemporary outcomes for fAVR versus mAVR in this population. We compared the postoperative outcomes and mid-term survival of these two approaches in a large cohort of low-risk patients. 

Methods:  Between 2002 and 2015, 2,095 low-risk patients (Society of Thoracic Surgeons Predicted Risk of Mortality score (STS PROM ≤  4)) underwent elective isolated AVR, including 1029 (49%) mAVR and 1066 (51%) fAVR. Median follow-up was 5.3 years. 

Results: Compared to mAVR patients, fAVR patients had a significantly higher burden of comorbidities such as diabetes (23% vs 11%), stroke (4% vs 2%), congestive heart failure (CHF) (41% vs 24%) and STS-PROM (1.91±0.95 vs 1.81±0.99), all p ≤  0.05.  However, both groups were similar in terms of gender, age, and preoperative creatinine, p > 0.05. Notably, operative mortality (1.1% vs.1.3%), stroke (3% vs. 2%), and re-operation rates for bleeding (1% vs. 2%) were similar between fAVR and mAVR, respectively, all p > 0.05. Median intensive care unit (ICU) stay (31 hours (interquartile range (IQR) 23,61) vs 42 hours (IQR 24, 68); p=0.075) and hospital length of stay (LOS) (6 days (IQR 5,7) vs 6 days (IQR 5,8); p ≤ 0.001) were significantly shorter among mAVR patients. Adjusted survival analysis identified age (Hazard Ratio (HR) 1.05), chronic kidney disease (HR 4.96), prior sternotomy (HR 1.56), and CHF (HR 2.00) as significant predictors of decreased survival (all p ≤  0.030), while type of intervention, mAVR vs fAVR, was non-contributory (HR 1.58; p=0.49).

Conclusion: In low-risk patients, mAVR results in shorter ICU and hospital LOS, while maintaining similar rate of mortality, stroke, reoperation for bleeding and mid-term survival, compared to fAVR. Therefore, mAVR should stand as a benchmark against TAVR in the low-risk patients.

 

6.10 Financial Impact of Access Site Pseudoaneurysm after Transcatheter Aortic Valve Replacement

Posted on January 18, 2018

A. Iyengar1, E. Aguayo1, Y. Seo1, Y. Sanaiha3, O. Kwon2, R. Satou2, P. Benharash2  1University Of California – Los Angeles,David Geffen School Of Medicine,Los Angeles, CA, USA 2University Of California – Los Angeles,Cardiac Surgery,Los Angeles, CA, USA 3University Of California – Los Angeles,General Surgery,Los Angeles, CA, USA

Introduction:

Vascular injuries are the most common complication following transcatheter aortic valve replacement (TAVR), and significantly contribute to morbidity and mortality in the perioperative period. While reducing the risk of vascular rupture, percutaneous access and smaller delivery devices may adversely impact the incidence of pseudoaneurysms. Although typically benign, the effect of access site pseudoaneurysms on resource utilization remain poorly defined. The purpose of this study was to characterize the impact of access site pseudoaneurysms on hospital costs and readmission rates. 

Methods:  

Retrospective analysis of the National Readmissions Database was performed between January 2012 & December 2014 using the International Classification of Diseases, Ninth Revision procedural codes for TAVR (35.05 and 35.06) and pseudoaneurysm formation (442.3). Costs were standardized to the 2014 US gross domestic product using US Department of Commerce consumer price indices and adjusted for diagnosis related group–based severity. The Kruskal-Wallis and chi-squared tests were used for comparisons between all cohorts.

Results:

Of the 32,976 patients who underwent TAVR, 542 (1.6%) were identified as having the complication of pseudoaneurysm. Development of a pseudoaneurysm was associated with older age (84 vs. 82 years, p=0.009), higher prevalence of peripheral vascular disease (39% vs. 26%, p<0.001), and a higher Elixhauser Comorbidity Index (7 vs 6, p=0.033). While 295 (0.9%) patients were diagnosed with pseudoaneurysms at the index hospitalization, 246 (0.6%) were discovered during a readmission.

At index hospitalization, pseudoaneurysm formation was associated with significantly increased length of stay (8 vs. 5 days, p<0.001) and increased total costs ($68,379 vs. $58,871, p<0.001). Endovascular intervention was utilized in 13% of pseudoaneurysms, while open surgical intervention was required in 2% of cases. Readmissions for pseudoaneurysms were also associated with significantly increased length of stay (5 vs. 4 days, p=0.012) and hospital costs ($20,464 vs. $14,835). Among readmissions, endovascular intervention was utilized in 4.4% of pseudoaneurysms, while open surgical intervention was required in 0.7% of cases.

Conclusion:

Pseudoaneurysm formation is more prevalent in older patients with pre-existing peripheral vascular disease. During both index hospitalization and readmissions, lengths of stay and hospital costs are significantly increased by presence of pseudoaneurysms despite low rates of endovascular or open surgical intervention. Strategies to reduce the formation of pseudoaneurysms after TAVR may serve as a suitable target for improvement in the delivery of quality care.

6.11 Readmissions After Mitral Valve Repair Vs. Replacement in the United States, 2010-2014

Posted on January 18, 2018

Y. Sanaiha1, A. Mantha1,2, Y. Seo1, L. Mukdad1, V. Dobaria1, Y. Juo1, R. Morchi2, P. Benharash1  1David Geffen School Of Medicine, University Of California At Los Angeles,Cardiac Surgery,Los Angeles, CA, USA 2University Of California – Irvine,Cardiothoracic Surgery,Orange, CA, USA

Introduction:
Background: Mitral valve replacement and repair have been debated as the optimal strategy to treat mitral regurgitation. While several studies have demonstrated the superiority of repair strategies from a clinical perspective, readmissions and resource utilization with each modality remains ill defined. Since readmission rates are considered indicators of quality, the present study aimed to evaluate the overall costs of care and rehospitalization rates of the respective strategies in a large representative national sample.

Methods:
Patients who underwent isolated mitral valve repair or replacement from Jan. through June in the 2010-2014 National Readmission Database (NRD) were analyzed. The NRD is an all-payer inpatient database maintained by the Healthcare Cost and Utilization Project (HCUP) that estimates more than 35 million annual U.S. hospitalizations. The primary outcomes were index mortality, length of stay, 30-day and 6-month readmission and GDP-adjusted costs. We utilized hierarchical linear models adjusting for demographics, cardiovascular risk factors,  and Elixhauser Comorbidity Index.

Results:
Of the 54,858 patients enrolled, 29,845 (54%) received replacement and 25,013 (46%) repair. Patients undergoing replacement were more likely to be female (57 vs 40%, P<0.001), older (66 vs. 63 yr, P<0.001), have Medicare (56 vs 42%, p<0.001) and have lower Elixhauser score (5.0 vs. 4.0, P<0.001). Replacement was associated with higher adjusted in-hospital mortality (5.4% vs. 1.2%, OR= 2.6, P<0.001),  higher adjusted costs ($64,158 vs. $43,643, β=0.16, P<0.001), longer hospitalization (14.4 days vs. 8.8 days, IRR:1.17, P<0.001). All-cause readmission at 30 days (19.6% vs 13.5%, OR=1.21, P<0.001) and cost of care 30-day readmission ($17,391 vs. $12,744, β=0.16, p<0.001) were significantly higher after replacement, most commonly due to 1) atrial fibrillation, 2) heart failure exacerbation, and 3) pleural effusion. Similarly replacement had higher adjusted odds of readmission at 6 months (34% vs. 22%, OR:1.26 P<0.001).

Conclusion:
In this study of  U.S. patients who underwent isolated mitral valve surgery from 2011-2014, readmission rates remain high. After adjustment for demographics, comorbidities, and hospital level variation, replacement was associated with greater length of stay, mortality and 30-day readmission. Repair first approach may be beneficial during the index hospitalization and in follow up. Based on available literature and our findings, strategies to maximize repair warrant implementation at the national level and beyond centers of excellence.
 

6.09 Embedding Real-Time Measure of Surgeons’ Cognitive Load into Cardiac Surgery Process Modeling

Posted on January 18, 2018

R. Dias2,7, M. Zenati5,7, H. Conboy6, J. Gabany5, D. Arney3,4, J. Goldman3,4,7, L. Osterweil6, G. Avrunin6, L. Clarke6, S. Yule1,2,7  1Brigham And Women’s Hospital,Department Of Surgery,Boston, MA, USA 2Brigham And Women’s Hospital,STRATUS Center For Medical Simulation,Boston, MA, USA 3Massachusetts General Hospital,Department Of Anesthesia,Boston, MA, USA 4Massachusetts General Hospital,MD PnP Program,Boston, MA, USA 5VA Boston Healthcare System,Division Of Cardiac Surgery,West Roxbury, MA, USA 6University Of Massachusetts,Amherst, MA, USA 7Harvard Medical School,Boston, MA, USA

Introduction:
Surgeons constantly deal with a high-demand operative environment that requires simultaneously processing a large amount of information. In certain situations, high demands imposed by surgical tasks may exceed surgeons’ cognitive resources, leading to a state of cognitive overload. This state may impact negatively on performance, increasing the risk of patient harm. The aim of this study was to investigate the concurrent validity of heart rate variability (HRV) analysis as a real-time and objective measure of surgeons’ cognitive load during cardiothoracic surgery. We also aimed to develop a behavioral framework that embeds surgeons’ physiological data into surgical process modeling for 14 unique high-level stages of cardiothoracic surgery.

Methods:
A heart rate sensor chest strap was used by a cardiac surgeon during 16 consecutive cardiothoracic procedures. Inter-beat intervals (R-R intervals) were captured via a validated smartphone app using a Bluetooth connection, and HRV parameters were calculated using spectral analysis. At the end of each procedure, a modified version of the SURG-TLX questionnaire, a validated tool assessing self-perceived cognitive load, was completed by the surgeon. Using audio-video recordings from real-life cardiac surgeries, the HRV parameters were embedded into the surgical workflow, enabling synchronized visualization of video, audio and cognitive load metrics during specific contexts and stages of cardiothoracic surgery. 

Results:
The HRV parameters presenting statistically significant correlation with SURG-TLX score were standard deviation of normal to normal R-R intervals (SDNN) (r = -0.61, p < 0.001), HRV triangulation index (r = -0.69, p < 0.001), maximum low frequency (LF)/ high frequency (HF) ratio (r = 0.55, p < 0.027), and LF/HF ratio episodes > 2.0 (r = 0.80, p < 0.001). A total of 14 unique stages of coronary artery bypass graft (CABG) were identified and we built a behavioral analysis system incorporating video and physiological data (Figure 1).

Conclusion:
A statistically significant association between HRV parameters and SURG-TLX was found, validating HRV analysis as an objective method of measuring surgeons’ cognitive load. We also developed a framework that enables the synchronization of physiological-based cognitive metrics into the surgical process analysis.  This behavioral framework can be used to monitor surgeons’ cognition in real-time, enhancing the understanding of how specific mental states can impact surgical performance and patient safety. Once this relationship is established, approaches seeking to mitigate the deleterious effects of cognitive overload can be developed.
 

6.06 Correlation of Anastomotic Leak and Neoadjuvant Chemoradiotherapy in Esophageal Cancer

Posted on January 18, 2018

D. Lee1, C. Takahashi2, R. Shridhar3, J. Huston4, K. Meredith1  1Florida State University College Of Medicine,Gastrointestinal Oncology,Sarasota, FL, USA 2Midwestern University,Phoenix, AZ, USA 3University Of Central Florida,Orlando, FL, USA 4Sarasota Memorial Health Care System,Sarasota, FLORIDA, USA

Introduction:  Anastomotic leaks (AL) causes significant morbidity after esophagectomy. Most patients receive neoadjuvant chemoradiation (NCR) prior to esophagectomy which has been associated with increase perioperative complications and mortality. We report on a comparison of AL rates in upfront surgical (US) and NCR patients. 

Methods:  A prospectively managed esophagectomy database was queried for US and NCR patients treated between 1996-2015. Predictors of AL rate were identified using multivariate (MVA) analysis and propensity score matching (PSM). 

Results: We identified 820 patients (US – 288; NCR – 532). Overall AL rate was 5.4%.  Decreased AL rate was observed in NCR patients on MVA (8% vs 4.1%; p = 0.04) but no difference was seen after PSM (7.7% vs 4.2%; p=0.14). MVA of factors associated with decreased AL in US patients included distal esophageal tumors and body mass index (BMI) >25. Age, gender, year of surgery, histology, anastomotic location, and diabetes were not prognostic.  Before PSM, MVA of NCR patients of factors associated with decreased AL revealed that only thoracic anastomosis was prognostic. However, this was not observed after PSM.  MVA of factors associated with decreased AL in all patients revealed thoracic anastomosis, NCR, and BMI 25-30. After PSM, only distal esophageal tumors and thoracic anastomosis were prognostic for decreased AL. 

Conclusion: There is no difference in the AL rate between US and NCR patients.  Decreased AL rate was observed in patients with distal esophageal tumors and thoracic anastomosis.

 

6.07 Perioperative Outcomes Following Esophagectomy With Gastric vs. Non-Gastric Reconstruction?

Posted on January 18, 2018

M. Varasteh Kia2, J. K. Canner1, R. J. Battafarano1, S. C. Yang1, E. L. Bush1, M. V. Brock1, E. R. Haut1,3, S. R. Broderick1  1Johns Hopkins University School Of Medicine,Baltimore, MD, USA 2Johns Hopkins Bloomberg School Of Public Health,Epidemiology And Biostatistics,Baltimore, MD, USA 3Johns Hopkins Bloomberg School Of Public Health,Health Policy And Management,Baltimore, MD, USA

Introduction

To restore gastrointestinal continuity following esophagectomy, tubularized stomach is the preferred conduit.  In scenarios where the stomach cannot be used non-gastric conduits such as jejunal or colonic interpositions are employed. There are inconsistencies between previous studies examining outcomes associated with the use of non-gastric conduits. Using the American College of Surgeons National Surgical Quality Improvement Program (ACS-NSQIP) database, we examined perioperative outcomes in patients reconstructed with gastric and non-gastric conduits to better characterize the relative risks of morbidity and mortality associated with these procedures. 

 

Methods

2006 – 2015 ACS-NSQIP esophagectomy cases were separated into gastric and non-gastric conduits based on CPT codes. Emergent and non-cancer resections were excluded. We examined perioperative differences between the two groups using chi-square and nonparametric Wilcoxon rank sum tests. Unadjusted and adjusted risk ratios of postoperative complications were estimated using Poisson regression with robust error variance.

 

Results

After exclusions, there were 6,321 and 210 patients in the gastric and non-gastric groups, respectively.  In unadjusted comparisons, significant differences were identified in 30-day mortality rate (3.32% vs 10.48%, p < 0.001), prolonged hospitalizations (5.30% vs 10.26%; p = 0.02), median length of stay (10% vs 13%; p < 0.001), reoperation rate (14.27% vs 30.43%, p < 0.001) and operative time (342 vs 384 minutes; p < 0.001) between the gastric and non-gastric groups.  No significant differences were identified in the occurrence of superficial, deep or organ/space surgical site infection, pneumonia, or readmission. After adjusting for age, gender, ethnicity, history of diabetes, smoking status, history of COPD, weight loss >10% in prior 6 months and BMI, there remained significant differences between groups in 30-day mortality (RR 0.33 [0.22-0.50]) for gastric vs non-gastric conduit), prolonged hospitalization (RR 0.51 [0.30-0.90]), and reoperation (RR 0.46 [0.35-0.61]), respectively (Table 1). 

 

Conclusion

The use of non-gastric conduit interposition following esophagectomy for carcinoma is associated with higher chance of mortality. However, the underlying reasons for this difference could not be identified using ACS-NSQIP data. Limitations of this study include its retrospective nature and the inherent limitations of the ACS-NSQIP dataset.  This analysis may help to inform shared decision making when considering alternate conduits for reconstruction for esophageal cancer patients in whom the gastric conduit is not feasible.

6.08 Innominate versus Axillary Artery Cannulation for Hemiarch Repair

Posted on January 18, 2018

M. Eldeiry1, M. Aftab1, K. Yamanaka1, M. S. Mosca1, C. Ghincea1, J. C. Cleveland1, D. Fullerton1, T. B. Reece1  1University Of Colorado Denver,Cardiothoracic Surgery,Aurora, CO, USA

Introduction:

Innominate artery cannulation has gained some popularity over the last decade as an alternative to axillary artery cannulation for providing antegrade cerebral perfusion (ACP) during repair of the ascending aorta and arch. Innominate artery cannulation provides several advantages including avoidance of an additional incision and use of a larger caliber artery to provide ACP. We hypothesize that these advantages make innominate artery cannulation superior to axillary artery cannulation as it can decrease operative times and potentially decrease blood loss.

Methods:

This was a single center retrospective analysis of 177 patients who underwent hemiarch replacement between 2009 and 2016. All patients qualified including emergent cases. Groups were separated by mode of cannulation: axillary vs innominate. Outcomes evaluated included cardiopulmonary bypass (CPB) time, cross-clamp (XC) time, circulatory arrest (CA) time, post-operative transfusions, intensive care unit length of stay, development of any neurological complications, end organ failure, and mortality. Sub-group analysis was performed for elective and emergent cases.

Results:

Axillary and innominate artery cannulation accounted for 42.4% (n=75) and 57.6% (n=102) of cases, respectively. There was no difference in patient characteristics except for a higher incidence of lung disease in the axillary group (21% vs. 9%, p=0.029). More emergent cases were performed in the axillary group (60% vs. 18%, p<0.001).

Operative times are summarized in Figure 1. Innominate cases had shorter CPB and CA times. In the elective subgroup, CA times were shorter for the innominate cases. However, the emergent subgroup displayed no difference.

Less transfusions were given in the innominate group including RBC (2[0,14] vs. 0[0,8], p<0.001), PLT (2[0,7] vs. 2[0,4], p=0.030) and FFP (6[0,20] vs. 3[0,11], p<0.001). A similar trend was observed for RBC and FFP in the elective subgroup. No difference was observed in the emergent subgroup.

There was no statistical difference in remaining outcomes between cases of axillary and innominate cannulation in the combined, elective, and emergent groups.

Conclusions:

Alternate cannulation strategies for open arch anastomoses are evolving with a trend towards utilizing the innominate artery. These data suggest that innominate cannulation is at least equivalent to, and may be superior to, axillary cannulation. The innominate artery provides a larger conduit vessel for perfusion and this decrease in resistance to flow, allowing for faster cooling and rewarming, maybe why CPB times were lower in this group. Innominate cannulation is a safe and potentially advantageous technique for hemiarch repair.

6.04 Poor Pulmonary Function Tests Potentiate The Impact Of Comorbidities After Lobectomy For Lung Cancer

Posted on January 18, 2018

D. C. Thomas1, B. N. Arnold1, M. DeLuzio1, F. C. Detterbeck1, D. J. Boffa1, J. D. Blasberg1, A. W. Kim2  1Yale School Of Medicine,Section Of Thoracic Surgery,New Haven, CT, USA 2University Of Southern California,Division Of Thoracic Surgery,Los Angeles, CA, USA

Introduction: Patients with poor pulmonary function tests (PFTs) and more comorbid conditions, identified by the Charlson comorbidity index (CCI), have been associated independently with an increased risk of perioperative complications after lung cancer surgery. Many large national databases currently available lack PFT data and consequently often rely on surrogates such as comorbidity indices.  This study sought to evaluate the interaction of PFTs and comorbidities on postoperative complications in a large single institution dataset of patients undergoing lobectomy for lung cancer. 

Methods: Patients undergoing lobectomy for lung cancer at an academic medical center from 2008-14 were examined. Patients were stratified by predicted postoperative FEV1 and DLCO:1) low PFTs (either FEV1 or DLCO ≤40%),2) moderate PFTs (both >40%, but not >80%), and 3) high PFTs (both ≥80%). The primary outcome was incidence of any complication in the postoperative period. Variables were analyzed using the χ2 test and predictors of complications using a multivariate model.

Results: A total of 376 patients were identified as having undergone lobectomy for lung cancer.  Low ppoPFTs comprised 9% (34) of patients, while 76% (286) had moderate ppoPFTs, and 15% (56) had high ppoPFTs.  Forty-one percent (154) of patients had a CCI=0, 32% (121) had CCI=1, and 27% (101) had a CCI≥2.  The overall incidence of complications was 36% (136).  Bivariate analysis demonstrated that among patients with high ppoPFFTs and moderate ppoPFTs, as CCI increased from 0 to 1 to ≥2, the incidence of complications increased then plateaued (Figure 1).  However, in patients with low ppoPFTs, increasing CCI from 0 to 1 to ≥2 continued to increase the incidence of complications.  On multivariable, low and moderate ppoPFTs independently predicted complications compared to patients with high ppoPFTs (OR=5.4, P<0.001 and OR=2.2, P=0.05, respectively), while CCI was not independently predictive.

Conclusion: Poor PFTs remain an independent predictor of complications after lobectomy for lung cancer and this finding alone supports utilization of PFTs in the analysis of outcomes. However, poor PFTs appear to have a greater negative impact when the number of comorbidities increases. This effect is distinctly different and not observed among patients with superior PFTs.  Therefore, while the absence of PFT data may not be as impactful when the comorbidities are minimal, the impact of absent PFT data may be more profound when their values are poor and found in patients with a greater number of comorbidities. 

6.05 Long Term Survival and Echocardiographic Findings After Left Ventricular Aneurysmectomy

Posted on January 18, 2018

A. A. Assi1, S. F. Bolling1, H. J. Patel1, M. Deeb1, M. A. Romano1, J. W. Haft1, R. L. Prager1, F. D. Pagani1, P. C. Tang1  1University Of Michigan,Department Of Cardiac Surgery,Ann Arbor, MI, USA

Introduction:
This study investigates the long term outcomes and predictors of mortality for left ventricular (LV) aneurysmectomy.

Methods:
From 1992 to 2017, there were 109 patients who underwent a LV aneurysmectomy procedure. Long term survival was determined from hospital records and the National Death Index. Preoperative demographics, clinical characteristics and features, operative technique and follow up echocardiography findings were analyzed using Cox regression and log-rank to determine variables influencing survival.

Results:
Median age was 63 (IQR=19) years, with 25 (22.9%) females. There were 101 (93%) true and 8 (7%) pseudo-aneurysms. Location of the aneurysm was antero-apical in 92 (84%) and posterior in 17 (16%).  Average preoperative left ventricular diastolic dimension (LVIDD) was 6.7+2.7cm.  Operative technique included primary closure without a patch in 58 (53%) and closure with patch in 51 (47%) patients.  Concomitant surgeries included mitral valve (MV) repair (n=40, 37%), MV replacement (n=5, 5%), tricuspid valve (TV) repair (n=4, 4%), aortic valve (AV) replacement (n=3, 3%), coronary bypass grafting (n=76, 70%; 1.6+1.3 grafts) and VSD closure (n=5, 5%). Redo-sternotomies were performed in 12 (11%) patients. Median echocardiography follow up was 2.9 yrs (IQR=9.0), and was obtained in 59 (54%) patients. LVEF improved from 28+13% to 33+16% (P=0.011).  There was a higher incidence of moderate to severe right ventricular (RV) function at follow-up (12% preoperatively versus 38% at follow-up; P=0.021) and higher incidence in severe TV regurgitation in patents who did not undergo repair (8.9% versus 22.2% respectively; P=0.004). Median echo follow up of MV repair was 3.6 (IQR=9.5) yrs. MV repair led to sustained improvements in MR (P=0.001) where only 2 (5%) experienced recurrent moderate-severe MR. For patients who did not undergo a MV procedure, there was no difference in preop and follow up MR severity (P=0.586). Median patient follow up was 7.1 yrs (IQR=8.5). Overall 5, 10, and 15 year survival were 71.9%, 48.1% and 26.2% respectively (Fig. 1). A multivariable analysis identified concomitant TV repair (P=0.001), increasing preoperative TV regurgitation (P=0.037), and concomitant AV replacement (P=0.086) as independent predictors of mortality.

Conclusion:
Long term survival following LV aneurysmectomy is adversely influenced by RV function.  While sustained improvement in LVEF and decreased MR following MV repair can be expected, RV function continues to decline accompanied by worsening tricuspid regurgitation. Close surveillance and aggressive medical management of RV failure is warranted in this patient population.
 

6.02 Comparison of Outcomes of Minimally Invasive and Open Pneumonectomy

Posted on January 18, 2018

A. Kumar1, H. Devishetty1, T. Demmy1, S. Yendamuri1  1Roswell Park Cancer Institute,Department Of Thoracic Surgery,Buffalo, NY, USA

Introduction:
Over the last decade minimally invasive surgery has been increasingly liberally utilized for the treatment of non-small cell lung cancer (NSCLC). This approach is increasingly used for the conduct of pneumonectomies. We sought to examine short and long term outcomes of the use of minimally invasive surgery for pneumonectomy in the United States.

Methods:
The National Cancer Database was queried for patients undergoing pneumonectomy between 2004 and 2014 and trends examined. As coding for surgical approach is only available from 2010 onwards, this subset was examined for the impact of surgical approach on short and long term outcomes of minimally invasive pneumonectomy. Univariate and multivariate analyses were performed to examine the impact of surgical approach using SPSS.

Results:
18,926 patients had a pneumonectomy between 2004 and 2014. The proportion of patients having a pneumonectomy steadily declined from 9.1% to 4.9% of all anatomic resections over this time period. From 2010 to 2014, 7407 pneumonectomies were performed. 2.0% and 13.5% of these cases were attempted to be performed robot-assisted and by VATS respectively and 1.4% and 8.5% were completed as intended. Intent to treat analysis did not show a difference between minimally invasive (MIS) and open approaches with respect to 30 day and 90 day mortality (6.7% vs 7.0%; P=0.90 and 11.7 vs 12.4%; P=0.51 respectively). There were no differences in age, gender, race, stage, laterality, length of stay and readmission rates between both approaches. Of note, there was no difference in the number of lymph nodes examined between both approaches (21.4 vs 21.9; P=0.56). However, overall survival was higher in patients undergoing MIS pneumonectomy vs. open pneumonectomy in univariable (57 months vs. 40 months; P=0.013) and multivariable analyses (Figure 1).

Conclusion:
Minimally invasive pneumonectomy appears safe with peri-operative outcomes comparable to open surgery.  MIS may improve long term pneumonectomy outcomes but confirmation by controlled studies is needed to reduce biases like high conversion and declining utilization rates observed in our cohort.
 

6.03 Impact of Depression on Readmission Outcomes after Coronary Artery Bypass Graft Surgery

Posted on January 18, 2018

E. A. Aguayo1, Y. Juo1, Y. Seo1, K. L. Bailey1, V. Dobaria1, Y. Sanaiha1, P. Benharash1  1University Of California – Los Angeles,Department Of Surgery/ Division Of Cardiac Surgery,Los Angeles, CA, USA

Introduction:
Postoperative depression affects between 10-40% of patients after cardiac operations and is associated with significantly worse cardiovascular outcomes. Indeed, depression screening has been recommended in the care of patients undergoing cardiac operations. While its exact causes are unknown, the incidence and impact of new onset depression beyond acute follow up remain ill defined. The present study aimed to evaluate the incidence, risk factors, and prognostic implication of depression during 90-day readmission after coronary artery bypass surgery (CABG). 

Methods:
A retrospective cohort study was performed identifying patients with no prior depression who received CABG using the 2010-2014 National Readmissions Database (NRD). Adult CABG patients who were readmitted within 90 days of discharge were categorized based on the presence of new-onset depression. Association between development of postoperative depression and mortality/complications during rehospitalization was examined using logistic regression models, adjusting for comorbid conditions. In-hospital mortality and GDP-adjusted costs were evaluated using hierarchical linear models and comorbidity via the Elixhauser Index. 

Results:
A total of 423,455 patients without prior diagnosis of depression were identified as undergoing CABG during the study period. Nationally, 18.9% of patients were readmitted within 90 days of index discharge, of which 5.6% developed new-onset depression. Within the readmitted cohort, risk factors for new-onset depression included female gender (OR=1.45, 95% CI: 1.33-1.57, p<0.01), experiencing a surgery-related complication (OR=1.18, 95% CI: 1.03-1.34, p=0.01), prolonged stay during index stay (OR=1.23, 95% CI: 1.12-1.35, p<0.01) and being discharged to locations other than home (OR=1.61, 95% CI: 1.48-1.76, p<0.01). Postoperative depression was independently associated with elevated odds of systemic complication during readmission (OR=1.17, 95% CI: 1.09-1.27, p<0.01), prolonged hospital stay (OR=1.61, 95% CI: 1.42-1.82, p<0.01), discharge to disposition other than home (OR=1.51, 95% CI: 1.35-1.69, p<0.01) and higher hospitalization costs (coefficient $906.1, 95% CI: 130.9-1681.3, p=0.02). The overall mortality during readmission was 2.6% while new-onset depression was associated with lower odds of death after adjusting for baseline differences (OR=0.56, 95% CI: 0.42-0.74, p<0.01).

Conclusions:
New-onset of depression following discharge CABG was associated with significantly worse outcomes during rehospitalization. The results of this national study supports institution of psychiatric evaluation and measures to reduce the development of depression following cardiac operations. Programs to incorporate early detection and timely management of depression may reduce systemic complications during readmission and enhance quality of care and postoperative recovery.

« Previous Page
Next Page »