Search

Whether pre left ventricular assist device (LVAD) sarcopenia is associated with higher incidence of gastrointestinal bleeding (GIB) on LVAD support remains unknown.To study the association between preoperative sarcopenia and post LVAD GIB events, we performed a retrospective, multi-centered study including patients with chest CTs performed ≤ 3 months prior to LVAD implantation at the University of Minnesota (n = 143) and Houston Methodist Hospital (n = 133). To quantify sarcopenia, unilateral pectoralis muscle mass indexed to body surface area (PMI) and attenuation (approximated by mean Hounsfield units; PHUThe study cohort included 276 LVAD recipients with 43 % designated as bridge to transplant at the time of LVAD implantation. High pectoralis muscle mass and tissue attenuation were both protective against GIB events. Each 5 unit increase in PHUPreoperative sarcopenia, as quantified by pectoralis muscle size and attenuation, was associated with the development of recurrent GI bleeding after LVAD implantation. These CT quantitative measures appear to predict not only early mortality but morbidity on LVAD as well.

A 72-year-old patient presented with severe tricuspid regurgitation and patent foramen ovale (PFO) in the setting of severe mediastinal shifting after remote right pneumonectomy. Surgical approach was challenging given the significant herniation of the heart and left lung into the right hemithorax. This report describes tricuspid valve replacement with PFO closure via a right posterolateral thoracotomy and alternative cardiopulmonary bypass cannulation strategy.

Purpose To provide safety and outcomes data on LVAD recipients referred for cardiac rehabilitation (CR). Methods A single center continuous-flow LVAD cohort (n=505) was searched for patients with information on CR participation (n=292). The analysis was then limited to LVAD recipients who fit into one of three main CR categories: those who graduated CR (n=138), those who would have been able to participate but declined (n=61), and those who were too sick to complete or start CR (n=28). To assess the association between CR categories and mortality and hospitalizations on LVAD support, multivariate cox regression and negative binomial regression analyses were performed, respectively. Results The mean age of the cohort with CR data available was 59 +/- 14 years, 44% (128/292) were bridge to transplant (BTT), and 40% (114/292) received a HeartMate 3 device. Among those who started CR and had the opportunity to finish (enough follow up time, insurance coverage), 79% graduated. Those who graduated CR had a 96% survival at 1 year (95% CI 91-98). In the unadjusted and adjusted analyses, those who graduated CR had the highest survival and the lowest incidence rates of hospitalization among the CR categories (Table). Conclusion This study is the largest to date to report outcomes of LVAD recipients referred for cardiac rehabilitation. The lower readmission rates and high survival in the group that graduated cardiac rehab provides further evidence for the safety of CR in LVAD recipients.

Pump thrombosis is a rare and infrequent complication of HeartMate III left ventricular assist device (LVAD). While there are reports of pump thrombosis in the postoperative period, to our knowledge, there have been no prior reports on pump thrombosis in the intraoperative period. Here we present a case of a 24-year-old female who required HeartMate III LVAD implantation for progressive heart failure and the intraoperative period was complicated with pump thrombosis (PT). Managing PT in the intraoperative period is challenging and time-sensitive because of its rare occurrence and paucity of recommendations in diagnosing the PT.

Correction of valvular disease is often undertaken during left ventricular assist device (LVAD) implantation with uncertain benefit. We analyzed clinical outcomes with HeartMate 3 (HM3; Abbott) LVAD implantation in those with various concurrent valve procedures (HM3+VP) with those with an isolated LVAD implant (HM3 alone).The study included 2200 patients with HM3 implanted within the Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy with HeartMate 3 (MOMENTUM 3) trial portfolio who underwent 820 concurrent procedures among which 466 (21.8%) were HM3+VP. VPs included 101 aortic, 61 mitral, 163 tricuspid; 85 patients had multiple VPs. Perioperative complications, major adverse events, and survival were analyzed.Patients who underwent HM3+VP had higher-acuity Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profiles (1-2: 41% vs 31%) compared with no VPs (P .05). The cardiopulmonary bypass time (124 vs 76 minutes; P.0001) and hospital length of stay (20 vs 18 days; P.0001) were longer in HM3+VP. A higher incidence of stroke (4.9% vs 2.4%), bleeding (33.9% vs 23.8%), and right heart failure (41.5% vs 29.6%) was noted in HM3+VP at 0 to 30 days (P .01), with no difference in 30-day mortality (3.9% vs 3.3%) or 2-year survival (81.7% vs 80.8%). Analysis of individual VP showed no differences in survival compared to HM3 alone. No differences were noted among patients with either significant mitral (moderate or worse) or tricuspid (moderate or worse) regurgitation with or without corrective surgery.Concurrent VPs, commonly performed during LVAD implantation, are associated with increased morbidity during the index hospitalization, with no effect on short- and long-term survival. There is sufficient equipoise to consider a randomized trial on the benefit of commonly performed VPs (such as mitral or tricuspid regurgitation correction), during LVAD implantation.

Carcinoid tumours of the heart occur most commonly as a result of metastatic disease and usually affect the right side of the heart. We report a case of a solitary carcinoid metastasis to the interventricular septum without hepatic involvement in a 74-year-old man.

Early readmission (within 30 days) after left ventricular assist device (LVAD) implantation might be a marker for increased mortality. We retrospectively reviewed the records of 277 adults who underwent continuous-flow LVAD implantation from 2005 through 2015 at our institution. The baseline characteristics of patients who were (versus were not) readmitted within 30 days after LVAD implantation were compared. To assess the impact of 30 day readmission on long-term survival, we used multivariate Cox regression. We also compared the cardiac transplant rate between the two groups. Of the 277 patients, 217 (78.3%) underwent LVAD implantation as a bridge-to-transplant; 76 (27.4%) of the 277 were readmitted within 30 days. The most common reason for readmission was volume overload (23.6%), followed by gastrointestinal bleeding (15.8%). Male gender, previous smoking, a higher baseline creatinine level, higher Model for End Stage Liver Disease Excluding INR (MELD-XI) score, and postoperative gastrointestinal bleeding or stroke were each associated with 30 day readmission. In our final multivariate model, increased mortality was also associated with 30 day readmission (hazard ratio, 1.60; 95% confidence interval, 1.1-2.5). Among the 217 bridge-to-transplant patients, the cardiac transplant rate was similar between the two groups: 18.7 transplants per patient-year among those who were readmitted within 30 days versus 19.7 transplants per patient-year among those who were not (p = 0.26). Among our study patients, 30 day readmission after LVAD implantation was frequent and was associated with increased mortality. It is currently unclear whether the general health of those patients was a factor and whether efforts to reduce 30 day readmission would favorably affect longer-term patient outcomes.

Assess the association between digoxin use and gastrointestinal bleeding (GIB) in a multicenter continuous flow left ventricular assist device (LVAD) cohort.Patients implanted with continuous flow LVADs with data on GIB and digoxin use from two centers were included in the analysis (n = 649). GIB events were captured up to 2 years of follow-up. Digoxin use was defined as digoxin prescribed at discharge or within the first 3 months after LVAD implantation. A negative binomial regression model was performed to determine the association between digoxin use and number of GIB events over the follow-up period.Mean age of the cohort was 57 years (±14) and 45% (293/649) were bridge to transplant (BTT). Digoxin was prescribed in 33% of patients. Digoxin use was associated with an unadjusted 32% reduction in the incidence of rate of all cause GIB (IRR 0.68, 95% CI 0.46-0.99, p = 0.049). After adjusting for age, sex, Interagency Registry for Mechanically Assisted Circulatory Support (INTERMACS) profile, renal function, and implanting center there was still a 34% reduction in the incidence rate (IRR 0.67, 95% CI 0.45-0.99, p = 0.048). When limiting the analysis to those with likely arteriovenous malformation associated GIB, the association strengthened (unadjusted: IRR 0.48, 95 % CI 0.26-0.89, p = 0.02, adjusted: IRR 0.47, 95 % CI 0.25-0.9, p = 0.022).In this multicenter study, inclusive of contemporary devices, digoxin use was associated with reduced GIB events. Prospective data will be required to confirm this association.

Background Hemolysis , assessed by elevated serum lactate dehydrogenase (LDH), is strongly associated with HeartMate II pump thrombosis (PT). However, it is unknown whether early intervention for elevated LDH circumvents the risk of serious PT requiring pump exchange. We sought to evaluate the relationship between elevated LDH and clinical outcomes, the effectiveness of early medical intervention, and risk factors for elevated LDH. Methods We studied 268 patients in the prospective, multicenter PREVENT study who had 2 or more LDH measurements at ≥30 days post-implant. Elevated LDH was defined as LDH ≥2.5× upper limit of normal (ULN) for 2 consecutive measurements. Results Fourteen percent of patients had elevated LDH. Stroke-free survival at 6 months was lower in patients with elevated LDH vs patients with normal LDH (83 ± 6% vs 93 ± 2%, p = 0.035). Elevated LDH resolved without intervention in 19% of patients, with intensified medical therapy in 43% and required surgical intervention in 38%. For patients receiving only medical therapy, survival was 94 ± 6% at 6 months post-treatment. In this subgroup, resolution of symptoms with intensified medical therapy was sustained in 15 of 16 patients, with PT occurring in 1 patient at 171 days after initial treatment for elevated LDH (202 days post-implant). Early medical intervention at moderately elevated LDH (2.5× to 3.2× ULN), as compared with higher levels (>3.2× ULN), led to more sustained resolution of symptoms without subsequent PT or need for surgical intervention (91% vs 26% at 6 months post-treatment, p = 0.002). Conclusions Early medical intervention can successfully resolve moderate LDH elevations (2.5× to 3.2× ULN) with a low incidence of death or PT at 6 months post-treatment.

Purpose To determine the association between ACEI/ARB use and post LVAD mortality in a large, multi-center, contemporary CF dataset inclusive of HeartMate 3 devices. Methods Patients at the two centers with a first time continuous flow (CF) LVAD implant with complete medication and survival outcome data were included in the analysis (total n=648). ACEI/ARB utilization was defined as ACEI or ARB use at discharge or within the first 3 months after LVAD implantation. A mutivariate cox regression was performed to test the association between ACEI/ARB use and post LVAD mortality. Results The mean age of the study cohort was 57 years (+/- 14), 45% were bridge to transplant (BTT), and 13% received HeartMate 3 devices. The median length of follow up for the analysis was 608 days. Sixty six percent of patients were on an ACEI or ARB in the 3 months after LVAD implantation. Use of either an ACEI or ARB was associated with an adjusted 35% reduction in the hazard rate of mortality on LVAD support (adjusted HR 0.65, 95% CI 0.48-0.87, p = 0.004). The final model was adjusted for sex, INTERMACS profile, BTT status, serum creatinine at the time of implant, serum albumin, and BMI. Conclusion ACEI/ARB use was associated with reduced post LVAD mortality in this large, multi-center, contemporary CF dataset inclusive of HeartMate 3 devices. Whether or not tolerating an ACEI or ARB is a marker of a healthier LVAD patient or is the direct cause of improved outcomes remains to be proven.

Purpose Pectoralis muscle mass and tissue attenuation obtained on preoperative CT scans are powerful predictors of mortality after LVAD implantation. Gastrointestinal (GI) bleeding remains a common problem on mechanical support. It is unknown whether these skeletal muscle measures, which correlate with frailty and tissue quality, are associated with the development of recurrent GI bleeding. Methods We performed a retrospective, multi-centered study which included patients with chest CTs performed ≤ 3 months prior to LVAD implantation at the University of Minnesota (n=143) and Houston Methodist Hospital (n=133). Unilateral pectoralis muscle mass indexed to body surface area (PMI) and attenuation (approximated by mean Hounsfield units; PHUm) were measured on preoperative chest CT scans. GIB events were captured for the first two years of LVAD support. Negative binomial regression analyses were performed to determine the association between pectoralis muscle measures and number of GIB events on LVAD support. Results The mean age of the study cohort was 59 + /- 13 years, 43 % were bridge to transplant and 62 % had ischemic cardiomyopathy. Thirty-two patients had at last one GI bleed while on LVAD support over 2 years. Both pectoralis muscle mass and tissue attenuation were associated with the incidence rate of GI bleeding on LVAD support. Each 5 unit increase in PHUm was associated with reduction in the incidence rate of GIB (adjusted 18 % reduction, 95 % CI 6-28 %, p = 0.005). Each unit increase in PMI was associated with a reduction in the incidence of GIB (adjusted 19 % reduction, 95 % CI 3- 34 %, p = 0.029). The final models were adjusted for age, sex, INTERMACS profile, BTT status, creatinine and albumin. Conclusion Preoperative pectoralis muscle size and attenuation were associated with the development of recurrent GI bleeding after LVAD implantation in this multicenter cohort. This one time, internal assessment of patient substrate appears to predict not only mortality after LVAD, but morbidity as well.

Purpose The purpose of this study was to determine how LVAD success is defined by LVAD team members across the United States. Methods An online, anonymous survey was shared with 32 LVAD programs across the country. Six case vignettes were provided: 1) young active drug user 2) demanding patient 3) medically complicated patient 4) socially isolated patient 5) patient with financial barriers, and 6) a non-compliant patient. Individual respondents were asked: 1) Would your program implant a patient like this? 2) Would you vote to implant a patient like this? 3) Do you think this LVAD will be successful? Clinical follow-up was provided and respondents were asked: 4) Was this LVAD successful? Results A total of 88 survey responses were completed including 22 (24%) LVAD advance practice providers, 29 (32%) CHF cardiologists, 28 (31%) LVAD coordinators, 8 (9%) surgeons, and 3 social workers (3%). The respondents had an average of 6.5 years of experience in advanced heart failure. Results are shown in the Table. We found frequent differences between individual opinions and program decisions. Responses were also significantly different when stratified by professional role. In addition, respondents often supported implant, even if they predicted it was unlikely to be successful. Finally, the definition of LVAD success varied by years of experience and professional role. Conclusion This national survey quantifies the lack of consensus around the definition of LVAD success. Society guidelines regarding patient selection will continue to face challenges due to lack of consensus in the community.

Purpose To assess the performance of several post cardiac surgery vasoplegia definitions in a large, single center continuous-flow left ventricular assist device (CF-LVAD) cohort. Methods A single center first-time, CF-LVAD dataset was utilized for this analysis (n=505). Vasoactive drips and hemodynamics were collected at the following time points in all subjects with available data: pre-op, immediate post-op, and on post-operative days 1, 2, and 3 (n=325). For each of the four vasoplegia definitions tested (table), infusion requirements for each definition were converted to an equivalent vasoactive intropes score (VIS). VIS were then cacluated for each patient, at each time point, and utilzied to determine if vasoplegia criteria were met. To test the assoication between vasoplegia and post LVAD mortality, logistic regression (60-day death as outcome) and cox regression analyses were performed. To assess for clinical predictors of post-opertive vasoplegia, a mutivariable logisitic regression was performed. Results Among the vasoplegia definitions tested, the Patarroyo definition had the strongest association with death at 60 days (OR 3.55, 95 % CI 1.48-8.52, p=0.005). This defintion was met by 14% of cohort subjects with available data. Variables associated with the development of vasoplegia using this definition in the mutivariate models inlcuded higher creatinine, longer cardiopulmonary bypass time, prior cardiothoracic sugery and lower INTERMACS profile. Two of the vasoplegia definitions tested (Levin, Unified) had little utility to classify vasoplegia due to the requirement for mean arterial pressure to be less than 50 mm Hg. Conclusion Among the post-cardiothoracic surgery vasoplegia definitions tested, the Patrroyo was the most highly associated with early post LVAD mortality in this dataset. Risk factors for vasoplegia were similar to previous published reports. Use of VIS simplified the process of determining vasoplegia infusion. criteria

Purpose To assess for a post left ventricular assist device (LVAD) hemodynamic profile predictive of early mortality. Methods A single center first-time, continuous-flow LVAD dataset was utilized for this analysis (n=505). Vasoactive drips and hemodynamics were collected at the following time points in all subjects with available data: pre-op, immediate post-op, and on post-operative days 1, 2, and 3 (n=325). Vasoactive intropes scores (VIS) were calculated at each time point. Trends before and after LVAD were computed and plotted using Locally Estimated Scatterplot Smoothing (LOESS), stratfied by death at 60 days post LVAD implantation. Mutivariate logistic regression and reciever operator curve analyses were performed to test the association between candidate variables and death at 60 days. Results The distinctive hemodynamic feature of patients who experienced death by 60 days was a failure to increase mean arterial pressure (MAP) from the time of leaving the operating room to post-operative day two accompanied by persistently elevated VIS (Figure). A model incorporating post-operative day two VIS, mean arterial pressure and right atrial pressure had an AUC of 78% to predict 60-day death. The group experiencing early mortality had higher measured cardiac indices, lower systemic vascular resistance, and higher right atrial pressures than patients who survived past 60 days, suggesting overlap between classic right ventricular failure and vasoplegia physiology. In this dataset, half of patients who met the criteria for vasoplegia also met criteria for right ventricular failure (21/41, 51%). Conclusion The failure to increase mean arterial pressure accompanied by high doses of inotropic and vasopressor support early after LVAD is highly associated with early mortality. This hemodynamic profile may be more unifying than existing right heart failure and vasoplegia definitions, which appear to have significant overlap. These findings require validation in an external dataset.

Purpose To test the hypothesis that advanced heart failure patients lose weight prior to LVAD, gain weight after LVAD, and that weight loss prior to LVAD is associated with lower pectroralis muscle measures (PMM) quantified on pre operative chest CT scans. Methods The larger cohort for the BMI trend analysis included patients with first-time LVAD implantation at a single center (n=505). The following PMMs were quantified in patients with chest CTs performed within 3 months prior to LVAD implantation (n=194): pectoralis muscle mass indexed to height in meters squared (PMI) and mean Hounsfield units (PHUm). BMI trends before and after LVAD were computed and plotted using Locally Estimated Scatterplot Smoothing (LOESS). BMI prior to LVAD was then regressed longitudinally. The association between BMI trend and PMMs were assessed by including fixed effects for time, PMMs, and their interactions. Results BMI plots are displayed (Figure). For the entire cohort, BMI decreased prior to LVAD, continued to decrease further after LVAD, and then steadily increased. Across PMM tertiles, the same pattern was observed. At 100 days prior to LVAD, both PMMs were significantly associated with BMI (p Conclusion The trend in BMI across the larger cohort before and after LVAD suggests that LVAD recipients are largely sarcopenic at the time of LVAD. The correlation between more rapid weight loss and lower PMMs supports this hypothesis. As muscle wasting is progressive, and lower PMMs are assoiciated with increased post-LVAD mortality, there may be a role in utilizing PMMs to determining optimal implantation timing.

Purpose To model renal function in an advanced heart failure population both prior to and after left ventricular assist device (LVAD) and to determine the association between pre-operative renal function trends and post-LVAD mortality. Methods Patients with a first time continuous-flow LVAD implant at a single center with renal function data before and after LVAD were included in the analysis (n=410). GFR at each time point was calculated with the CKD-EPI equation. Trends before and after LVAD were computed and plotted using Locally Estimated Scatterplot Smoothing (LOESS). GFR between 200 to 100 days prior to LVAD was regressed longitudinally to calculate a slope of GFR change prior to LVAD. Both the GFR slope and nadir of GFR prior to LVAD (Figure) were assessed as predictors of post LVAD mortality using multivariable cox regression. Results When assessed as a whole, GFR declined in the entire cohort prior to LVAD, improved with medical optimization, improved further post LVAD, and then declined again to a level that was similar to initially measured values. Each unit decrease in the GFR slope was associated with an adjusted 35% increase in post LVAD mortality (HR 1.36, 95% CI 1.15-1.51, p=0.002). With respect to the GFR nadir, each 10-unit decrease was associated with an adjusted 19% increase in post-LVAD mortality (HR 1.19, 95% CI 1.07- 1.30, p = 0.003). Models were adjusted for age, bridge to transplant status, ischemic cardiomyopathy, and diabetes. Conclusion This analysis demonstrates that renal function in the advanced heart failure population prior to LVAD is complex, with declines in renal function prior to LVAD, and the nadir of GFR prior to LVAD being associated with increased post LVAD mortality. This data suggest earlier LVAD placement prior to measurable renal deterioration may be beneficial.

Purpose To describe adverse events associated with subclavian access intra-aortic balloon pump (IABP) before and after cardiac transplant utilizing high level, single center data. In order to assess morbidity associated with IABP use, a matched United Network for Organ Sharing (UNOS) cohort without IABP support was created as a comparison group. Methods The study included two patient groups: 1) patients bridged to cardiac transplant as Status 2 with subclavian access IABP at single center and 2) a UNOS cohort of transplant recipients (between 2014-2018) matched on hemodynamic Status 2 requirements and wait times. Patient characteristics and post-transplant outcomes were compared. Complications specific to the subclavian IABP using the single center data were collected and reported. Results The single center subclavian access IABP cohort included 28 patients and was compared to the matched UNOS cohort of 202 patients. There was no significant difference in age, ischemic cardiomyopathy diagnosis, or wait times between the two cohorts (Table 1). Post-transplant length of stay and need for post-transplant dialysis was higher in the IABP supported patients. The one-year survival probabilities were similar (subclavian IABP: 96%, matched UNOS cohort: 93%, log rank p = 0.97). Balloon pump repositioning (36%) and replacement (21%) were common. Post-operative infections attributable to the graft site occurred in 17% of patients. Conclusion Patients transplanted on subclavian access IABP had comparable survival at 1 year to a UNOS matched cohort, however IABP re-positioning, replacement and post-operative infections were common. The matched cohort analysis suggests the IABP itself may be associated with longer length of stay and higher rates of post-transplant dialysis, although patient acuity may have accounted for this difference despite hemodynamic matching.

Purpose Measured glomerular filtration rate (GFR) improves initially after left ventricular assist device (LVAD) implantation and then declines. The reasons for this decline are incompletely understood. We hypothesize that the decline in measured GFR after LVAD is associated with the timing of increasing body mass after surgery. Methods Patients with a first time continuous-flow LVAD implant at a single center with renal function data and body mass index (BMI) data before and after LVAD were included in the analysis (n=410). GFR at each time point was calculated with CKD-EPI equation. Trends before and after LVAD were computed and plotted using Locally Estimated Scatterplot Smoothing (LOESS). Linear mixed effects modeling was performed to test the association between GFR and BMI accounting for correlated, repeated measures within persons. Results Plots of GFR and BMI before and after LVAD implantation are displayed (Figure). BMI decreased prior to LVAD, decreased further after LVAD, and then steadily increased. GFR declined in the cohort prior to LVAD, improved with medical optimization, improved further post LVAD, and then declined again. Across the cohort there was a strong association between GFR and BMI (p Conclusion The displayed trends suggest GFR is over-estimated in the late stages of heart failure and in the immediate post LVAD period due to low creatinine production from sarcopenia. Use of creatinine in the advanced heart failure population may lead to inaccurate GFR estimates. This has broad implications for treatment of the heart failure population.

Purpose The impact of pre-operative acute liver injury on post left ventricular assist device (LVAD) mortality is unknown. Methods The study cohort included adults with first time, continuous-flow LVAD implantation at a single center. Patients with complete laboratory data between -5 and -1 days prior to LVAD were included in the analysis. To assess for patients with acute liver injury pattern, AST and ALT were plotted, log transformed, and smoothed using the Locally Estimated Scatterplot Smoothing (LOESS) function. As AST and ALT tracked together, log AST alone was utilized for the remainder of the analysis. The association between peak and mean log AST prior to LVAD were assessed as predictors of post-LVAD mortality using multivariate cox regression. Results The study cohort consisted of 366 patients. The median age was 58 +/- 14 years old, 54.6% were bridge to transplant (BTT), and 53.3% had ischemic cardiomyopathy. The figure visually demonstrates the pattern of log AST prior to LVAD, stratified by patients who experienced death by 60 days post-implant. In the multivariate analysis, for each 1 unit increase in the peak log AST in the pre-LVAD time period, there was an adjusted 22% increase in post-LVAD mortality (HR 1.22, 95% CI 1.03 - 1.45, p=0.023). For each 1 unit increase in the mean log AST in the pre-LVAD time period, there was an adjusted 30% increase in post-LVAD mortality (HR 1.30, 95% CI 1.05 - 1.61, p=0.016). Models were adjusted for age, sex, BTT status, serum albumin, and ischemic cardiomyopathy. Conclusion Peak AST and higher mean AST levels prior to LVAD implantation were associated with increased post-LVAD mortality, despite improvement in liver injury pattern prior to LVAD. Ongoing organ dysfunction may be part of the mechanism of poor outcomes of patients who have experienced shock prior to LVAD placement.

There is insufficient data on patients with small body size to determine if this should be considered a risk factor for continuous-flow left ventricular assist device (CF-LVAD) support. We sought to evaluate survival outcomes, adverse events, and functional status of CF-LVAD patients with body surface area (BSA)

Background High levels of discordance between activated partial thromboplastin time (aPTT) and anti-factor Xa (anti-FXa) have been reported in patients with continuous-flow left ventricular assist devices (CF-LVADs). We sought to determine the biochemical basis for discordance by simultaneous measurement of coagulation factors with aPTT and anti-FXa. Methods Consecutive CF-LVAD patients admitted for intravenous unfractionated heparin (IV-UFH) therapy were studied. IV-UFH was monitored based on an anti-FXa. International normalized ratio (INR), aPTT, lactate dehydrogenase, and coagulation factors II, V, VII, VIII, IX, X, XI, and XII were measured simultaneously with anti-FXa. Results The discordance rate between anti-FXa and aPTT was 74.7% overall (568 of 760 samples from 62 patients). In samples in which coagulation factors were simultaneously measured, the discordance rate was 66% (49 of 74 samples from 52 patients). In patients with INR ≤ 1.0 the discordance level was 31.8% vs 82.7% with INR > 1.8. No difference was observed in the rate of discordance between samples with low vs normal vitamin K–dependent factors (65.9% vs 64.7%, p = 0.99). Low factor XII was found in 15%; of these, 90.9% exhibited anti-FXa/aPTT discordance. Low factor V was found in 8.2%, and all were discordant. Factor VIII was elevated in 97% of samples, and 65.7% were discordant. Conclusions In addition to the effects of warfarin administration on aPTT, the influence of other factors, such as low factor XII or V, may contribute to the discordance between aPTT values and anti-FXa. Coagulation factor levels may assist in individualization of anti-coagulation targets while patients implanted with CF-LVADs are on IV-UFH.

Purpose We previously developed a post LVAD mortality prediction model using multicenter data incorporating pectoralis muscle mass and tissue attenuation obtained on preoperative CT scans. The purpose of this analysis was to evaluate the performance of this model in a prospective cohort including the most contemporary LVADs. Methods Since 2016, all patients who underwent first time LVAD implantation at the University of Minnesota had chest CTs performed prior to surgery (n=82). None of the patients included in the Minnesota Pectoralis Muscle Risk Score (MPRS) development cohort were included in this analysis. Unilateral pectoralis muscle mass indexed to body surface area (PMI) and attenuation (approximated by mean Hounsfield units; PHUm) were measured from pre-operative CT scans. MPRSs were calculated on all patients using the following variables: PHUm, PMI, African American race, creatinine, total bilirubin, body mass index, bridge to transplant status and presence or absence of contrast. Cox regression analysis was performed to test the association between the MPRS and post LVAD mortality. Receiver-operating characteristic (ROC) curves were then generated to test model discrimination at 30 and 90 days post LVAD implantation. Results The prospective cohort included 62% HeartMate 3, 17% HVAD, and 20% HeartMate 2 devices. Each two point increase in MPRS was associated with a 2.6 fold increase in the mortality after LVAD (HR 2.6, 95% CI 1.3-5.4, p=0.01). The area under the curve for the MPRS was 0.73 at 30 days and 0.72 at 90 days. Conclusion The Pectoralis Muscle Risk Score had favorable discrimination in this contemporary prospective cohort, which was largely comprised of patients receiving HeartMate 3 devices. These skeletal muscle measures, which are associated with heart failure stage and frailty, continue to add important information to pre-operative risk assessment.

PURPOSE: Conventional median sternotomy (CMS) is still the standard technique utilized to implant left ventricular assist devices (LVAD). Recent studies suggest that less invasive surgery (LIS) may have advantages; however, robust data on differences in right heart failure (RHF) is lacking. This study aimed to determine the impact of LIS versus CMS on RHF outcomes following LVAD implantation. METHODS: An international multi-centre retrospective cohort study was conducted, including 5 centres. Patients were grouped according to their implantation technique (LIS vs. CMS). Only centrifugal devices were included. RHF was defined according to the INTERMACS definition; sub-analyses for inter-centre differences and propensity-score analyses were also performed. RESULTS: Overall, 405 implants occurred during the study period, with 305 patients implanted via CMS and 100 via LIS. Most pre-implant variables including mean age at implantation, sex, serum creatinine, right atrial pressure, PAPi, cardiac index, and baseline tricuspid regurgitation did not differ between the two groups. RHF grade after LVAD implantation was significantly worse in CMS patients than in LIS patients. Fourty-six, 21%, and 34% of the patients in the CMS group had none, mild-to-moderate, and severe-severe acute RHF compared to 76%, 10%, and 14% in the LIS group respectively, p

Purpose The MOMENTUM 3 trial established superiority of a fully magnetically-levitated HeartMate 3 (HM3) centrifugal pump over the axial flow HeartMate II (HMII) pump on hemocompatibility-related outcomes. Infection-related outcomes were no different between the two devices. We sought to understand the types of infection, temporal trends, clinical correlates and risk markers over 2-years of support. Methods MOMENTUM 3 randomized 1020 patients to either the HM 3 (515) or HM II (505) groups. Morbidity and mortality related outcomes were analyzed between patients with or without major infection complications. Cox regression analyses were utilized to identify characteristics associated with major infectious complications. Results The cumulative number of infections was 1213 (634 HM3 and 579 HMII), and major infection occurred in 58% HM3 and 56% HMII patients (p=0.57). Localized infections unrelated to pump components were most common (n=681/1213; 56%), followed by driveline infection (n=329/1213; 27%), sepsis (n=194/1213; 16%), and other events (n=9/1213; 0.7%). Bacterial pathogens were implicated in 806/1213 events (66%) and significant predictors of infection included gender (men vs. women; HR:0.73, p=0.003), BUN (HR:1.12, p=0.01), preimplant history of cardiac surgery (HR:1.28, p=0.01), and BMI≥30 (HR:1.43, p Conclusion Infection is the commonest adverse effect in patients implanted with contemporary continuous-flow pumps with most such events presenting as events unrelated to the pump components. Whether mechanical circulatory devices confer an immunosuppressive effect leading to propensity for infection warrants closer scrutiny to understand and ameliorate this morbidity. (Funded by Abbott; MOMENTUM 3 NCT02224755)

Purpose The association between pectoralis muscle measurements and objective frailty assessments in the heart failure population has not been previously reported. Methods The study cohort included patients who had first-time CF-LVAD implants at the University of Minnesota between 2017-2018 who had 1) chest CTs and 2) frailty assessments performed in the 3 months prior to LVAD (n=27). Unilateral pectoralis muscle mass indexed to height in meters squared and attenuation (approximated by mean Hounsfield units) were measured on preoperative chest CT scans ​using Slice-O-Matic software. Frailty was assessed with a 6 point Clinical Frailty Scale. Multivariable linear regression was performed to assess the association between pectoralis muscle measures and frailty scores. Results After multivariate adjustment, each frailty score increase (more frail) was associated with an adjusted decrease of 3.9 in the mean pectoralis muscle hounsfield units (95% CI= -7.4 to -0.41, p=0.03). A statistically significant association between pectoralis muscle mass index and frailty scores was not detected (-0.61 95% CI = -1.4 to 0.2; p =0.15). Multivariable models were adjusted for age, sex, and INTERMACS profile. Conclusion In this small sample of patients with simultaneous frailty and pectoralis muscle measurements, a significant association was detected between tissue attenuation and frailty. Further data will be needed to confirm this finding and to determine the degree that frailty can be reversed with LVAD support.

Purpose Employment status has been linked to worse outcomes in heart failure. Employment status at the time of LVAD evaluation has not been assessed as a predictor of post-LVAD patient outcomes. Methods The study cohort consisted of patients who were implanted with continuous flow LVADs at the University of Minnesota between June 2005 and June 2018 who had complete psychosocial evaluation data available. Patients were classified as employed, retired, or unemployed. Cox and negative binomial regression models were used to measure the association between employment status and post LVAD mortality and number of 1 year all cause admission, respectively. Results The study cohort included 428 patients. Patient were almost equally distributed among the three groups with 157 (37%) unemployed, 135 retired (33%), and 136 employed (32%). Patients who were employed leading up to LVAD implantation had higher socioeconomic status, lower BMIs, and lower reported smoking history (table). Being employed (vs. retired/unemployed) at the time of LVAD evaluation was associated with lower HR for mortality (unadjusted HR 0.50, 95% CI 0.34 - 0.73, P Conclusion Patients who were employed at the time of LVAD evaluation had improved survival and lower rates of hospitalization on LVAD support even after adjustment for socioeconomic and clinical variables in this single center analysis.

Purpose Elevated body mass index (BMI) is a well-established risk factor for post-operative adverse events. We hypothesized that patients with higher BMI are more likely to develop end stage renal failure (ESRD) after LVAD implantation, a complication regarded as a strong negative prognostic marker for survival. Methods A single center continuous flow LVAD cohort (n=456) was queried for BMI at the time of LVAD implantation. To determine the relationship between BMI and the probability of ESRD, a multivariable Cox regression analysis was performed. Patients were stratified into 3 groups by BMI ( 30 kg/m2) for analysis. Results The mean age of the cohort was 57±14 years. 80% were male, 54% had ischemic cardiomyopathy, 57% were designated as bridge to transplant and the median pre-implant creatinine was 1.2. The median follow-up time on LVAD support was 1.3 years. 25 of the 456 patients in our cohort (5.5%) developed ESRD. BMI >30 kg/m2 was associated with a significantly increased risk for renal failure (adjusted HR: 6.5, CI: 1.4-29.4; p=0.015). While there was a trend, the risk for the intermediate BMI group did not reach statistical significance (p=0.079) when compared to BMI Conclusion BMI above 30 kg/m2 at the time of LVAD implantation was associated with a significantly increased post-operative risk for renal failure in our cohort. Further studies are warranted to determine the pathophysiology and to develop strategies to mitigate this risk.

Purpose The association between renal function at the time of left ventricular assist device (LVAD) implantation and post LVAD mortality has been previously established. The association between the slope of renal function change in the year prior to LVAD implantation has not been reported. Methods A single center retrospective analysis was performed. Inclusion criteria were patients with chest CT scans performed ≤3 months prior to LVAD implantation with renal function trend available for a year prior to LVAD (n=102). Renal function was smoothed using mixed effects modeling from 365 to 60 days prior to LVAD and a slope was calculated for each patient. The slope of the renal function was assessed as a predictor of post LVAD mortality using multivariable cox regression. Results Patients with the largest deterioration in renal function in the year prior to LVAD were more likely to be INTERMACS 1 and 2 at the time of surgery (Table). A 1 mg/dl increase in creatinine between -365 days and -60 days increased the hazards rate of death after LVAD by a factor of 2.6 (adjusted HR 2.6, 95 % CI 1.03-6.7, p = 0.04). Patients with the largest decline in renal function had lower pectoralis muscle tissue attenuation by CT imaging in the 3 months prior to LVAD (Table). Conclusion Renal function deterioration over the course of the year prior to LVAD was associated with pectoralis muscle sarcopenia and higher post LVAD mortality, even after adjustment for other variables. The change in renal function paired with muscle tissue attenuation by CT scan may further define ideal implantation timing.

Background: Skeletal muscle mass decreases in end-stage heart failure and is predictive of clinical outcomes in several disease states. Skeletal muscle attenuation and quantity as quantified on preoperative chest computed tomographic scans may be predictive of mortality after continuous flow (CF) left ventricular assist device (LVAD) implantation. Methods and Results: A single-center continuous flow-LVAD database (n=354) was used to identify patients with chest computed tomographies performed in the 3 months before LVAD implantation (n=143). Among patients with computed tomography data available, unilateral pectoralis muscle mass indexed to body surface area and attenuation (approximated by mean Hounsfield units [PHU m ]) were measured in each patient with a high intrarater and inter-rater reliability (intraclass correlation coefficients 0.98 and 0.97, respectively). Multivariate Cox regression analyses were performed, censoring at cardiac transplantation, to assess the impact of preoperative pectoralis muscle index and pectoralis muscle mean Hounsfield unit on survival after LVAD implantation. Each unit increase in pectoralis muscle index was associated with a 27% reduction in the hazard of death after LVAD (adjusted hazard ratio, 0.73; 95% confidence interval, 0.58–0.92; P =0.007). Each 5-U increase in pectoralis muscle mean Hounsfield unit was associated with a 22% reduction in the hazard of death after LVAD (adjusted hazard ratio, 0.78; 95% confidence interval, 0.68–0.89; P Conclusions: Pectoralis muscle size and attenuation were powerful predictors of outcomes after LVAD implantation in this data set. This one time, repeatable, internal assessment of patient substrate added valuable prognostic information that was not available on standard preoperative testing.

Background: The HeartMate 3 (HM3) Left Ventricular Assist System (LVAS) (Abbott) is a centrifugal, fully magnetically levitated, continuous-flow blood pump engineered to enhance hemocompatibility and reduce shear stress on blood components. The MOMENTUM 3 trial (Multicenter Study of MagLev Technology in Patients Undergoing Mechanical Circulatory Support Therapy With HeartMate 3) compares the HM3 LVAS with the HeartMate II (HMII) LVAS (Abbott) in advanced heart failure refractory to medical management, irrespective of therapeutic intention (bridge to transplant versus destination therapy). This investigation reported its primary outcome in the short-term cohort (n=294; 6-month follow-up), demonstrating superiority of the HM3 for the trial primary end point (survival free of a disabling stroke or reoperation to replace the pump for malfunction), driven by a reduced need for reoperations. The aim of this analysis was to evaluate the aggregate of hemocompatibility-related clinical adverse events (HRAEs) between the 2 LVAS. Methods: We conducted a secondary end point evaluation of HRAE (survival free of any nonsurgical bleeding, thromboembolic event, pump thrombosis, or neurological event) in the short-term cohort (as-treated cohort n=289) at 6 months. The net burden of HRAE was also assessed by using a previously described hemocompatibility score, which uses 4 escalating tiers of hierarchal severity to derive a total score for events encountered during the entire follow-up experience for each patient. Results: In 289 patients in the as-treated group (151 the HM3 and 138 the HMII), survival free of any HRAE was achieved in 69% of the HM3 group and in 55% of the HMII group (hazard ratio, 0.62; confidence interval, 0.42–0.91; P =0.012). Using the hemocompatibility score, the HM3 group demonstrated less pump thrombosis requiring reoperation (0 versus 36 points, P P =0.02), and fewer nondisabling strokes (6 versus 24 points, P =0.026) than the control HMII LVAS. The net hemocompatibility score in the HM3 in comparison with the HMII patients was 101 (0.67±1.50 points/patient) versus 137 (0.99±1.79 points/patient) (odds ratio, 0.64; confidence interval, 0.39–1.03; P =0.065). Conclusions: In this secondary analysis of the MOMENTUM 3 trial, the HM3 LVAS demonstrated greater freedom from HRAEs in comparison with the HMII LVAS at 6 months. Clinical Trial Registration: URL: http://clinicaltrials.gov . Unique identifier: NCT02224755.

Objectives. To evaluate whether collaboration between existing and new transcatheter aortic valve replacement (TAVR) programs could help reduce the number of cases needed to achieve optimal efficiency. Background. There is a well-documented learning curve for achieving procedural efficiency and safety in TAVR procedures. Methods. A multidisciplinary collaboration was established between the Minneapolis VA Medical Center (new program) and the University of Minnesota (established program since 2012, n=219) 1 year prior to launching the new program. Results. 269 patients treated with TAVR (50 treated in the first year at the new program). Mean age was 76 (±18) years and STS score was 6.8 (±6). Access included transfemoral (n=35, 70%), transapical (n=8, 16%), transaortic (n=2, 4%), and subclavian (n=5, 10%) types. Procedural efficiency (procedural time 158±59 versus 148±62, p=0.27), device success (96% versus 87%, p=0.08), length of stay (5±3 versus 6±7 days, p=0.10), and safety (in hospital mortality 4% versus 6%, p=0.75) were similar between programs. We found no difference in outcome measures between the first and last 25 patients treated during the first year of the new program. Conclusions. Establishing a partnership with an established program can help mitigate the learning curve associated with these complex procedures.

Background Drive-line infections and sepsis lead to significant morbidity and even mortality in patients with ventricular assist devices (VADs). The HeartWare HVAD system is unique compared with other VAD systems in that it has a thin, flexible drive-line and is implanted directly into the pericardial space without the need for a pump pocket. We reviewed 332 patients receiving the HVAD in the pivotal ADVANCE Bridge to Transplant (BTT) trial and Continued Access Protocol (CAP) to determine the incidence of drive-line infections and sepsis. Methods Summary statistics of outcomes and Kaplan–Meier freedom-from-event analyses were performed. The Cochran–Mantel–Haenszel test for homogeneity was used to determine whether there was any site effect on rates of either drive-line exit-site or sepsis infections. Results Drive-line exit-site infections occurred in 16.9% (56 of 332) of patients (0.25 event per patient-year [EPPY]). These infections were primarily treated with antibiotics and most occurred >30 days post-implant (88%, or 66 of 75 drive-line infections). Sepsis occurred in 17.2% (57 of 332) of patients (0.23 EPPY), and sepsis events occurred more frequently >30 days post-implant (86%, or 60 of 70 sepsis events). Of patients with sepsis, 17.5% (10 of 57) died due to sepsis-related neurologic events (7 of 10) and multisystem organ failure (2 of 10) and cardiopulmonary failure (1 of 10), whereas there were no deaths related to drive-line infections. In addition, 17.5% (10 of 57) of patients had a stroke event while under treatment for sepsis (7 hemorrhagic and 3 ischemic strokes), and 3.5% (2 of 57 patients with sepsis) had a device exchange for VAD thrombus (4 days before and 1 day after sepsis diagnosis). Patients with drive-line infections or sepsis had a larger body mass index compared to those without infection (29.4 vs 27.6 kg/m 2 , p = 0.015), and were more frequently diabetic (43.7% vs 31.2%, p = 0.034). Although there was no negative impact on survival in patients with drive-line exit-site infections, there was a trend for reduced survival in patients with sepsis events when compared to those patients without sepsis. Conclusions The incidence of drive-line infection and sepsis was low in patients receiving the HVAD pump. Despite a trend toward reduced survival in patients with sepsis events, the overall survival of patients with either drive-line infections or sepsis was excellent.

Background Preexisting valve pathology is common in patients with end-stage heart failure undergoing left ventricular assist device (LVAD) placement. The indications and subsequent benefits of performing valvular procedures in these patients are unclear. The objective of this study was to determine the impact of performing concurrent surgical valve procedures in a large cohort of patients receiving LVADs. Methods One thousand one hundred six patients received the HeartMate II (HMII) LVAD in the bridge to transplant (n = 470) and destination therapy (n = 636) clinical trials. Of these, 374 patients (34%) had concurrent cardiac surgery procedures as follows: 242 patients (21%) with 281 concurrent valve procedures (VP) (aortic 80, mitral 45, and tricuspid 156), and 641 patients had only HMII LVAD. The focus of this study was to determine the clinical outcomes of patients undergoing HMII + VP compared with those who received HMII alone. Results Patients undergoing HMII + VP were significantly older, had higher blood urea nitrogen levels and central venous pressure, and decreased right ventricular stroke work index; intraoperatively, the median cardiopulmonary bypass times were also longer. The unadjusted 30-day mortality was significantly higher in patients undergoing HMII + VP (10.3% vs 4.8% for LVAD alone, P = .005). Subgroup analysis of individual VPs showed that higher mortality occurred in patients with HMII plus 2 or more VPs (13.5%, P = .04) followed by trends for increased mortality with HMII plus mitral alone (11.5%, P = NS), HMII plus aortic alone (10.9%, P = NS), and HMII plus tricuspid (8.9%, P = NS) procedures. Of these various groups, only patients undergoing HMII + isolated aortic VP ( P = .001) and HMII + multiple VPs ( P = .046) had significantly worse long-term survival compared with patients undergoing HMII alone. Right heart failure and right ventricular assist device use was increased in patients undergoing VPs, but there was no difference in the incidence of bleeding or stroke. Conclusions Patients frequently require concurrent VPs at the time of LVAD placement; these patients are sicker and have higher early mortality. Furthermore, right ventricular dysfunction is increased in these patients. Further studies to develop selection criteria for concurrent valve interventions are important to further improve clinical outcomes.

Background Liver dysfunction in left ventricular assist device (LVAD) recipients is common both before and after implantation. Postoperative liver dysfunction (PLD) develops in some LVAD recipients without preoperative liver dysfunction. The aim of this study was to assess clinical outcomes in such patients. Methods Records of all patients undergoing implantation of a HeartMate II (HM II, St. Jude Medical, Inc, Minneapolis, MN) LVAD at a single center at the University of Minnesota from January 2005 through June 2014 were analyzed. PLD was defined by hypertransaminasemia or hyperbilirubinemia, or both, during the hospitalization for LVAD implantation. Results During the study period, 284 patients underwent HM II implantation. Excluded from analysis were 14 recipients with preoperative liver dysfunction. In the final cohort (n = 270), there were no major difference in preoperative characteristics among those patients with versus without PLD. PLD developed in 129 (47.8%) recipients: 16 (12.4%) had isolated hypertransaminasemia (group I), 76 (58.9%) had isolated hyperbilirubinemia (group II), and 37 (28.7%) had combined hypertransaminasemia and hyperbilirubinemia (group III). Group III LVAD recipients had significantly greater rates of 30-day, 90-day, and 1-year mortality, along with significantly higher transfusion requirements and higher rates of renal replacement therapy, prolonged ventilation, and vasopressor use. Moreover, their mortality risk was significantly higher than that of PLD-free LVAD recipients (hazard ratio, 4.6; 95% confidence interval, 2.1 to 10.1; p Conclusions Isolated hyperbilirubinemia is common after LVAD implantation. In this study, it was not associated with an increase in early or midterm postoperative mortality. However, postoperative combined transaminasemia and hyperbilirubinemia was associated with a significant increase in early and midterm morbidity and mortality. Further research into the pathogenesis of post-LVAD PLD is necessary.

Introduction Post cardiac transplant outcomes of patients supported with the Heartmate 3 left ventricular assist device (LVAD) have not been previously reported. The purpose of this analysis was to report status at the time of transplant and survival by continuous flow (CF) device type. Hypothesis Patients supported with a HeartMate 3 LVAD will have a similar post-transplant survival to other contemporary CF-LVADs. Methods The study cohort included adult patients supported on contemporary, continuous flow (CF) devices in the UNOS registry who underwent cardiac transplant between the years 2014-2018, prior to the 2018 UNOS allocation system change. The analysis was limited to patients who underwent initial LVAD implantation in 2014 and beyond to allow post-transplant comparisons among patients with similar LVAD support times. Multivariable cox regression was performed to assess the association between LVAD type and post-transplant survival. Results The study cohort consisted of 3,117 patients who underwent cardiac transplant while on LVAD support (Heartmate 2: 1,545, Heartmate 3: 179, HeartWare (HVAD): 1,393). No patients supported on Heartmate 3 devices were transplanted for pump thrombosis during this time period. Patients supported on the Heartmate 3 were less likely to be listed 1A for complications at the time of transplant (Heartmate3: 28 % (51/179), Heartmate 2: 44 % (680/1,545), HVAD 37 % (511/1,393, p Conclusions Patients supported with Heartmate 3 LVAD at the time of transplant have similar survival after transplant compared to other contemporary CF devices in this early experience. No patients with a Heartmate 3 were transplanted for device thrombosis.

Purpose The EUROMACS Right-Sided Heart Failure Risk Score was developed to predict right ventricular failure (RVF) after left ventricular assist device (LVAD) placement. The performance of this model outside the derivation cohort has not been tested. Methods EUROMACS Right-Sided Heart Failure Risk Scores were calculated and assessed as predictors of post-operative RVF in a single center continuous flow (CF) LVAD cohort (n=451). Only patients with complete data for risk score calculation and RVF determination were included in the analysis (n = 254). RVF was defined as RVAD within 30 days, continuous inotropic support ≥ 14 days, or inhaled pulmonary vasodilatory use for > 48 hours after LVAD implantation. Results Thirty-nine percent of patients (100/254) had post-operative RVF, however only 9% (23/254) required prolonged inotropic support and 5% (12/254) required RVAD placement. Development of RVF after LVAD was associated with a 45% increase the hazards rate of death on LVAD support (HR 1.45, 95 % CI 0.98-2.2, p =0.066), however this was not statistically significant. The increase in mortality associated with RVF was largely driven by patients who required RVAD support (HR 4.6, 95 % CI 2.3- 9.1, p Conclusion The EUROMACS Right-Sided Heart Failure Risk Score had poor discrimination in this LVAD dataset. These findings further emphasize the difficulty of predicting RVF after LVAD placement. Further work is required to enhance our understanding pre- LVAD right ventricular physiology and to accurately risk stratify patients.

Purpose Right ventricular failure (RVF) after LVAD implantation remains difficult to predict pre-operatively. As right heart failure is associated with muscle wasting, we hypothesized pectoralis muscle tissue attenuation quantified on pre-operative CT scans would enhance the prediction of RVF after LVAD implantation. Methods Patients with a first time continuous flow (CF) LVAD implants from Houston Methodist (n=208) and University of Minnesota (n=224) with complete post-operative RVF data were included in the analysis (total n = 432). As we wanted to create a model that applied to INTERMACS 2-5 patients, we limited the analysis to INTERMACS 2 and higher patients. RVF was defined as RVAD within 30 days of implantation or requiring inotropic support ≥ 14 days. To develop an RVF Risk Model, we implemented backward model selection using logistic regression. Thirty-one candidate hemodynamic and clinical predictors were tested during model development. This method was repeated limiting the analysis to patients with mean pectoralis Hounsfield units (PHUm) available by pre-operative CT scans (total n=188). Receiver-operating characteristic (ROC) curves for the RVF Risk Model and the Pectoralis RVF Risk Model were then generated. Results The RVF Risk Model included the following variables: albumin, log total bilirubin, and INTERMACS profile. The AUC for this model was 69% (95% CI: 62-76%). The Pectoralis RV Risk Model included albumin, creatinine and PHUm. The AUC for this model was 73% (95 % CI: 63-82%). Conclusion Pectoralis muscle tissue attenuation was predictive of RVF after LVAD implantation in this multicenter study. Incorporating this measure into a RVF Risk Model improved model performance. This measure is quantifiable, repeatable, and appears to add information to traditional preoperative assessment.

Purpose We have previously shown that pectoralis muscle mass and tissue attenuation obtained on preoperative CT scans were powerful predictors of mortality after LVAD implantation using single center data. This analysis confirms those findings in a separate LVAD dataset and presents a novel prediction model developed across centers. Methods Patients with chest CTs performed ≤ 3 months before LVAD implantation at University of Minnesota (n=143) and Houston Methodist Hospital (n=133) were identified from larger cohorts (University of Minnesota n=353, Houston Methodist =278). Unilateral pectoralis muscle mass indexed to body surface area (PMI) and attenuation (approximated by mean Hounsfield units; PHUm) were measured on preoperative chest CT scans. To develop the Pectoralis Muscle Risk Score (PMRS), we implemented a cross-validated model selection using Cox's proportional hazards regression model with censoring at transplantation. The final model included PHUm, sex, albumin, year of implant, body mass index, bridge to transplant, INTERMACS profile, and presence or absence of contrast. Receiver-operating characteristic (ROC) curves for 30, 90, and 365 day survival were generated for the model and compared to the HeartMate II Risk Score (HMRS). Results The Pectoralis Risk Score outperformed HMRS at each of the pre-specified time-points. The AUC for PMRS for predicting survival at 30, 90, and 365 days was 78%, 76%, and 76%, respectively. The HMRS had AUC of 63%, 67%, and 63% (Figure). Conclusion The Pectoralis Muscle Risk Score had favorable discrimination in this multicenter dataset. These skeletal muscle measures, which are associated with heart failure stage and frailty, appear to add important information to pre-operative risk assessment.