CARDIOTHORACIC SURGERY

Regression of pressure-induced left ventricularhypertrophy is associated with distinct patterns ofdevelopmental gene expressionWilliam E Stansfield MD,* Mauricio Rojas MD, Ruhang Tang PhD,Craig H Selzman MDUniversity of North Carolina, Chapel Hill, NC

INTRODUCTION: Left ventricular hypertrophy (LVH) is a highlyprevalent and an independently robust risk factor for cardiovascularmorbidity and mortality. Therapeutic interventions have uniformlyfocused on preventing progression rather than promoting regression,partly due to the paucity of data regarding LVH regression mecha-nisms. Using our LVH regression model, we evaluated the time-course of changes in developmental gene expression within the con-text of associated physiologic changes.

METHODS: Minimally invasive transverse arch banding was per-formed in C57BL/6 mice. One month later, the band was removedand animals were followed for an additional month. Hearts wereanalyzed grossly, histologically, functionally using echocardiography,and for gene expression using real-time PCR.

RESULTS: Aortic banding resulted in robust hypertrophic re-sponse, observed by heart weight/body weight ratio, histology, echo-cardiography and fetal gene expression. After debanding, wall thick-ness normalized within one week by echocardiography (p�0.005).Histology and HW/BW reversed over similar time. Differentiationgenes (beta-MHC, BNP) trended down in 3 days (p�0.05), whilecellular growth markers (erg-1) required longer. At one month, therewas persistently elevated expression of genes of differentiation, cell-growth, and from the NF-kappaB family (p65, SRF, myocardin)(p�0.05).

CONCLUSIONS: Using a novel LVH regression model, we demon-strated that histologic and functional regression of LVH is associatedwith early transcriptional reversal of key developmental genes, in-cluding the first reported change in myocardin expression. Impor-tantly, altered patterns of gene expression persisted long after physi-ologic reversal appeared complete. Future studies will focus onindependent mechanisms of LVH regression rather than the existingLVH progression paradigm.

The debate between transhiatal and transthoracicesophagectomy: A five-year prospective cohort of17,395 patientsRafe Connors MD, Brian Reuben MD, Leigh Neumayer MD,David Bull MDUniversity of Utah School of Medicine, Salt Lake City, UT

INTRODUCTION: Debate continues over whether transhiatalesophagectomy offers decreased morbidity and mortality comparedto transthoracic esophagectomy. To definitively answer this question,we used the Nationwide In-patient Sample (NIS) database to com-pare morbidity and mortality following transhiatal and transthoracicesophagectomy.

METHODS: The NIS database prospectively followed 17,395 pa-tients across the United States undergoing transhiatal or transtho-

racic esophagectomy between 1999 and 2003. There were 11,914patients (mean age, 61.89 �/� 13.14 years) who underwent trans-hiatal esophagectomy while 5,481 patients (mean age, 61.91 �/�12.67 years) underwent transthoracic esophagectomy. Multivariateregression analysis was used to compare morbidity, mortality andlength of stay.

RESULTS: Overall morbidity and mortality following esophagec-tomy were 50.7 % and 8.8 %, respectively. In-hospital mortalityfollowing transhiatal esophagectomy was 8.91 % compared to 8.47% following transthoracic esophagectomy (p � 0.642). Multivariateregression analysis (accounting for patient and hospital factors)showed no difference in the incidence of wound, infectious, urinary,pulmonary, gastrointestinal, cardiovascular, systemic, procedure re-lated, or overall complications between the two groups. Hospitallength of stay was likewise equal at 18.38 �/� 16.80 days for trans-hiatal esophagectomy patients versus 18.02 �/� 17.21 days fortransthoracic esophagectomy patients (p � 0.904).

CONCLUSIONS: This large volume, multicenter study constitutesthe largest prospective cohort in the literature to compare outcomesfollowing transhiatal and transthoracic esophagectomy. Transhiatalesophagectomy does not result in decreased morbidity and mortalityor decreased length of stay compared to transthoracic esophagec-tomy.

Comparison of video assisted thoracoscopicsurgery to thoracotomy for resection of clinicalstage I non-small cell lung cancerBryan Whitson MD, Adam Boettcher, Ricardo Bardales,Robert Kratzke MD, Peter Dahlberg MD, PhD, Rafael Andrade MD,Michael Maddaus MDUniversity of Minnesota, Minneapolis, MN

INTRODUCTION: Lobectomy for clinical stage I nonsmall cell lungcancer (NSCLC) can be performed by thoracotomy (OT) or video-assisted thoracoscopic surgery (VATS). We compared the operativecharacteristics and postoperative course of VATS with OT lobectomyin patients with clinical stage I NSCLC.

METHODS: Retrospectively, all patients undergoing lobectomy forclinical stage I NSCLC from January 1, 1998 through June 30, 2005were reviewed.

RESULTS: Categorized by operation completed, 147 lobectomies(88 OT, 59 VATS) for clinical stage I NSCLC were performed.Demographics were similar between groups; however, VATS patientshad more hypertension (p�.0114), renal insufficiency (p�.0479)and prior malignancies (p�.0086). The pathologic stage, tumor size,histology, and positive nodes were not different. OT had more totalnodes identified (p�.0001) and shorter ICU stay (p�.0224). TheVATS group had significantly less occurrence of postoperative pneu-monia (p�.0004). VATS trended toward fewer chest tube days,shorter length of stay, and less reoperation. There were no differencesin operative time, blood loss, atrial fibrillation, or ventilator days.Median survival between the cohorts was similar (2.96 years OT v.2.72 years VATS, log rank 0.8039).

S20© 2006 by the American College of Surgeons ISSN 1072-7515/06/$32.00Published by Elsevier Inc.

CONCLUSIONS: Patients undergoing VATS lobectomy for clinicalstage I NSCLC, despite more comorbidities, had fewer postoperativecomplications. The approaches have equivalent operative time,blood loss, length of stay, and survival. Compared to OT, VATSlobectomy for clinical stage I NSCLC appears to be a less morbidoperation.

Deficiency of TNFR1 protects myocardium throughSTAT3, SOCS3, and IL-6, but not p38 MAPK or IL-1betaMeijing Wang MD, MS, Paul Crisostomo MD, Keith Lillemoe MD,Daniel Meldrum MDIndiana University Medical Center, Indianapolis, IN

INTRODUCTION: Suppressor of cytokine signaling (SOCS) pro-teins are major inhibitors of cytokine signaling via the JAK/STATpathway. TNF plays an important role in the development of heartfailure. There is a direct correlation between myocardial function andmyocardial TNF levels in humans. However, it is unknown whetherTNF mediates myocardial inflammation via STAT3/SOCS3 signal-ing in the heart, and if so, whether this effect is through the type 155-kDa TNF receptor (TNFR1). We hypothesized that TNFR1 de-ficiency protects myocardial function and decreases myocardial IL-6production via the STAT3/SOCS3 pathway in response to TNF.

METHODS: Isolated male mouse hearts (n�4/group) from wildtype (WT), TNFR1 knockout (TNFR1KO) were subjected to directTNF infusion (500 pg/ml/min x 30 min) and LVDP, �dP/dT,�dP/dT were continuously recorded. Heart tissue was performed foractive forms of STAT3, p38, SOCS3 (Western blot) and IL-6, IL-1beta (ELISA). p�0.05�statistically significant.

RESULTS: TNFR1KO had significantly better myocardial func-tion (LVDP: 37.9�/�4.4 vs. WT 20.6�/�3.2mmHg; �dP/dt:1176.4�/�159.1 vs. WT 589.1�/�85.6 mmHg/s; �dP/dt:�877.2�/�115.4 vs. WT �448.6�/�61.8mmHg/s) and greaterexpression of SOCS3 after TNF(% of SOCS3/GAPDH: 45�/�4.5% vs. WT 22�/�6.5%). TNFR1 deficiency decreased STAT3activation (% p-STAT3/STAT3: 29�/�6.4% vs. WT 45�/�8.8%). IL-6 decreased in TNFR1KO (150.2�/�3.65 pg/mg pro-tein) vs. WT (211.4�/�26.08). TNFR1 deficiency did not changeexpression of p38 and IL-1 beta following TNF infusion.

CONCLUSIONS: Deficiency of TNFR1 protects myocardiumthrough STAT3, SOCS3, and IL-6, but not p38 MAPK or IL-1 beta.

Regional systolic and remodeling strain differencesduring cardiac remodelingAhmet Kilic MD, G Kwame Yankey MD, Tim Nolan MS,Tieluo Li MD, Jennifer Nash MS, Guangming Cheng MD,Zhongjun Wu PhD, Bartley Griffith MDUniversity of Maryland, Baltimore, MD

INTRODUCTION: The purpose of this study is to investigate differ-ences in systolic and remodeling strains in the infarct, adjacent andremote regions of the left ventricular (LV) myocardium in an ovinepostinfarction model.

METHODS: Sixteen sonomicrometry transducers were placed intothe LV free wall to assess regional contractile function by systolicstrain and structural alteration by remodeling strain. All animalsunderwent daily hemodynamic and strain measurements.

RESULTS: Systolic function of the remote zone was preserved withstrain of �21.2 �3.8, �22.2 � 1.6, �26.0 � 0.2, �28.6 � 3.2 and�27.6 � 1.2 (%) at pre-, post-, 2 weeks, 6 weeks and 8 weekspostinfarction. The adjacent zone showed progressive dysfunctionwith strain of �18.6 � 6.9, �15.4 � 4.3, �10.3 � 9.6, �7.8 �13.9, �7.7 � 10.2. The infarct zone became non-contractile imme-diately postinfarction with strain of �14.0 � 1.5, �6.6 � 11.8,3.1 � 1.5, 2.1 �1.2, 1.4 � 0.8. Using pre-infarction end diastolicgeometry as reference, the remote zone had a progressive increase inremodeling areal strain of 7.2 � 2.3, 1.5 � 3.4, 7.5 � 3.5 and16.0 � 6.9; the adjacent zone had values of 2.0 � 1.3, 12.3 � 3.9,29.6 � 6.7 and 33.4 � 9.8; and the infarct zone with 6.0 � 4.5,13.0 � 5.6, 21.5 � 6.7, 60.1 � 8.7.

CONCLUSIONS: Regional function strain and geometrical struc-tural analysis can be used to illustrate the dynamics and progressionof cardiac remodeling.

Finite element analysis of transmural stress in leftventricular aneurysmAmod P Tendulkar MD, Joseph Walker PhD, Julius Guccione PhD,Mark Ratcliffe MDUniversity of California, San Francisco, Oakland, CA

INTRODUCTION: Normally, there is a transmural variation in stressthat is related to myofiber orientation. Finite element analysis (FEA)is a computational tool used to calculate stress. In REA, infinitesimaldeformation method (IDM) uses stiff linear material properties anddoes not account for myofiber orientation whereas finite deforma-tion method (FDM) uses non-linear physiologic material propertiesbased on myofiber orientation to compute stress. We hypothesizedthat FDM will demonstrate better transmural variation in stress thanIDM.

METHODS: A FE mesh was created from MRI of sheep hearts(n�3) with left ventricular aneurysm (LVA). Aneurysm, border-zoneand remote regions were identified. Validated material propertieswere used for FDM. Stiff,isotropic material properties were used forIDM. Endocardial surface was pressurized to 78.7mmHg. End sys-tolic stresses (circumferential(CS) and longitudinal(LS)) were calcu-lated for every element and then grouped by region and transmurallocation. ANOVA and post-hoc analysis with p�0.05 was deter-mined significant.

RESULTS: Transmural variation was observed in CS calculated byFDM in the endocardium versus midwall and epicardium(p�0.0001) in aneurysm and remote regions. Only remote regionhad transmural variation in calculated CS for IDM (p�0.01). LS wasdifferent in endocardium versus epicardium in aneurysm and remoteregions and endocardium versus midwall in aneurysm region(p�0.0001) for FDM. IDM demonstrated a difference betweenmidwall and epicardium in remote only (p�0.04). Finally, the twomethods were also different (p�0.03).

S21Vol. 203, No. 3S, September 2006 Cardiothoracic Surgery