Accuracy and consistency of online large language model-based artificial intelligence chat platforms in answering patients' questions about heart failure.

BACKGROUND: Heart failure (HF) is a prevalent condition associated with significant morbidity. Patients may have questions that they feel embarrassed to ask or will face delays awaiting responses from their healthcare providers which may impact their health behavior. We aimed to investigate the potential of large language model (LLM) based artificial intelligence (AI) chat platforms in complementing the delivery of patient-centered care. METHODS: Using online patient forums and physician experience, we created 30 questions related to diagnosis, management and prognosis of HF. The questions were posed to two LLM-based AI chat platforms (OpenAI's ChatGPT-3.5 and Google's Bard). Each set of answers was evaluated by two HF experts, independently and blinded to each other, for accuracy (adequacy of content) and consistency of content. RESULTS: ChatGPT provided mostly appropriate answers (27/30, 90%) and showed a high degree of consistency (93%). Bard provided a similar content in its answers and thus was evaluated only for adequacy (23/30, 77%). The two HF experts' grades were concordant in 83% and 67% of the questions for ChatGPT and Bard, respectively. CONCLUSION: LLM-based AI chat platforms demonstrate potential in improving HF education and empowering patients, however, these platforms currently suffer from issues related to factual errors and difficulty with more contemporary recommendations. This inaccurate information may pose serious and life-threatening implications for patients that should be considered and addressed in future research.

Outcomes in patients with aortic stenosis and severely reduced ejection fraction following surgical aortic valve replacement and transcatheter aortic valve replacement.

BACKGROUND: Patients with severe aortic stenosis (AS) and left ventricular (LV) dysfunction demonstrate improvement in left ventricular injection fraction (LVEF) after aortic valve replacement (AVR). The timing and magnitude of recovery in patients with very low LVEF (≤ 25%) in surgical or transcatheter AVR is not well studied. OBJECTIVE: Determine clinical outcomes following transcatheter aortic valve replacement (TAVR) and surgical aortic valve repair (SAVR) in the subset of patients with severely reduced EF ≤ 25%. METHODS: Single-center, retrospective study with primary endpoint of LVEF 1-week following either procedure. Secondary outcomes included 30-day mortality and delayed postprocedural LVEF. T-test was used to compare variables and linear regression was used to adjust differences among baseline variables. RESULTS: 83 patients were enrolled (TAVR = 56 and SAVR = 27). TAVR patients were older at the time of procedure (TAVR 77.29 ± 8.69 vs. SAVR 65.41 ± 10.05, p < 0.001). One week post procedure, all patients had improved LVEF after both procedures (p < 0.001). There was no significant difference in LVEF between either group (TAVR 33.5 ± 11.77 vs. SAVR 35.3 ± 13.57, p = 0.60). Average LVEF continued to rise and increased by 101% at final follow-up (41.26 ± 13.70). 30-day mortality rates in SAVR and TAVR were similar (7.4% vs. 7.1%, p = 0.91). CONCLUSION: Patients with severe AS and LVEF ≤ 25% have a significant recovery in post-procedural EF following AVR regardless of method. LVEF doubled at two years post-procedure. There was no significant difference in 30-day mortality or mean EF recovery between TAVR and SAVR. TRIAL REGISTRATION: Indiana University institutional review board granted approval for above study numbered 15,322.

Body Mass Index and Mortality in Cardiogenic Shock.

We explored the association of body mass index (BMI) with mortality in cardiogenic shock (CS). Using the Cardiogenic Shock Working Group registry, we assessed the impact of BMI on mortality using restricted cubic splines in a multivariable logistic regression model adjusting for age, gender, and race. We also assessed mortality, device use, and complications in BMI categories, defined as underweight (<18.5 kg/m2), normal (18.5-24.9 kg/m2), overweight (25-29.9 kg/m2), obese (30-39.9 kg/m2), and severely obese (>40 kg/m2) using univariable logistic regression models. Our cohort had 3,492 patients with CS (mean age = 62.1 ± 14 years, 69% male), 58.0% HF-related CS (HF-CS), and 27.8% acute myocardial infarction (AMI) related CS. Body mass index was a significant predictor of mortality in multivariable regression using restricted cubic splines (p < 0.0001, p = 0.194 for nonlinearity). When stratified by categories, patients with healthy weight had lower mortality (29.0%) than obese (35.1%, p = 0.003) or severely obese (36.7%, p = 0.01). In HF-CS cohort, the healthy weight patients had the lowest mortality (21.7%), whereas it was higher in the underweight (37.5%, p = 0.012), obese (29.2%, p = 0.003), and severely obese (29.9%, p = 0.019). There was no difference in mortality among BMI categories in AMI-CS.

Durable ventricular assist device in adult patients with single ventricle: a systematic literature review.

The worldwide experience with mechanical circulatory support in adult patients with single ventricle and Fontan palliation is limited. We aimed to see the outcomes in such patients on durable left ventricular assist device. We systematically reviewed the English language literature found on PubMed and Google Scholar, searching the terms "Fontan," "single ventricle," "ventricular assist device," "LVAD," and "mechanical circulatory support," and selected all individual cases with reported outcomes. We identified 4 patients on pulsatile left ventricular assist device, all of whom were successfully bridged to transplant, and 15 patients on continuous flow LVAD, age 25.9 ± 5.8, predominantly males. Mean time on LVAD support was 718.6 ± 616.5 days, and only 2 patients (13.3%) died. Durable left ventricular assist device support in adult patients with Fontan provides excellent results and should be used more frequently.

Predictors of Survival After Heart Transplant in the New Allocation System: A UNOS Database Analysis.

Clinical predictors of posttransplant graft loss since the United Network for Organ Sharing (UNOS) heart allocation system change have not been well characterized. Single organ adult heart transplants from the UNOS database were identified (n = 10,252) and divided into a test cohort (n = 6,869, 67%) and validation cohort (n = 3,383, 33%). A Cox regression analysis was performed on the test cohort to identify recipient and donor risk factors for posttransplant graft loss. Based on the risk factors, a score (max 16) was developed to classify patients in the validation cohort into risk groups of low (≤1), mid (2-3), high (≥4) risk. Recipient factors of advanced age, Black race, recipient blood group O, diabetes, etiology of heart failure, renal dysfunction, elevated bilirubin, redo-transplantation, elevated pulmonary artery pressure, transplant with a durable ventricular assist device, or transplant on extracorporeal membrane oxygenation (ECMO) or ventilator were associated with more posttransplant graft loss. Donor factors of ischemic time and donor age were also associated with outcomes. One year graft survival for the low-, mid-, high-risk groups was 94%, 91%, and 85%, respectively. In conclusion, easily obtainable clinical characteristics at time of heart transplant can predict posttransplant outcomes in the current era.

Outcomes of Patients Transferred to Tertiary Care Centers for Treatment of Cardiogenic Shock: A Cardiogenic Shock Working Group Analysis.

BACKGROUND: Consensus recommendations for cardiogenic shock (CS) advise transfer of patients in need of advanced options beyond the capability of "spoke" centers to tertiary/"hub" centers with higher capabilities. However, outcomes associated with such transfers are largely unknown beyond those reported in individual health networks. OBJECTIVES: To analyze a contemporary, multicenter CS cohort with the aim of comparing characteristics and outcomes of patients between transfer (between spoke and hub centers) and nontransfer cohorts (those primarily admitted to a hub center) for both acute myocardial infarction (AMI-CS) and heart failure-related HF-CS. We also aim to identify clinical characteristics of the transfer cohort that are associated with in-hospital mortality. METHODS: The Cardiogenic Shock Working Group (CSWG) registry is a national, multicenter, prospective registry including high-volume (mostly hub) CS centers. Fifteen U.S. sites contributed data for this analysis from 2016-2020. RESULTS: Of 1890 consecutive CS patients enrolled into the CSWG registry, 1028 (54.4%) patients were transferred. Of these patients, 528 (58.1%) had heart failure-related CS (HF-CS), and 381 (41.9%) had CS related to acute myocardial infarction (AMI-CS). Upon arrival to the CSWG site, transfer patients were more likely to be in SCAI stages C and D, when compared to nontransfer patients. Transfer patients had higher mortality rates (37% vs 29%, < 0.001) than nontransfer patients; the differences were driven primarily by the HF-CS cohort. Logistic regression identified increasing age, mechanical ventilation, renal replacement therapy, and higher number of vasoactive drugs prior to or within 24 hours after CSWG site transfer as independent predictors of mortality among HF-CS patients. Conversely, pulmonary artery catheter use prior to transfer or within 24 hours of arrival was associated with decreased mortality rates. Among transfer AMI-CS patients, BMI > 28 kg/m2, worsening renal failure, lactate > 3 mg/dL, and increasing numbers of vasoactive drugs were associated with increased mortality rates. CONCLUSION: More than half of patients with CS managed at high-volume CS centers were transferred from another hospital. Although transfer patients had higher mortality rates than those who were admitted primarily to hub centers, the outcomes and their predictors varied significantly when classified by HF-CS vs AMI-CS.

Multicenter registry and test bed for extended outpatient hemodynamic monitoring: the hemodynamic frontiers in heart failure (HF2) initiative.

Background: Hemodynamic Frontiers in Heart Failure (HF2) is a multicenter academic research consortium comprised of 14 US institutions with mature remote monitoring programs for ambulatory patients with heart failure (HF). The consortium developed a retrospective and prospective registry of patients implanted with a wireless pulmonary artery pressure (PAP) sensor. Goals/aims: HF2 registry collects demographic, clinical, laboratory, echocardiographic (ECHO), and hemodynamic data from patients with PAP sensors. The aims of HF2 are to advance understanding of HF and to accelerate development of novel diagnostic and therapeutic innovations. Methods: HF2 includes adult patients implanted with a PAP sensor as per FDA indications (New York Heart Association (NYHA) Class III HF functional class with a prior hospitalization, or patients with NYHA Class II or brain natriuretic peptide (BNP) elevation without hospitalization) at a HF2 member site between 1/1/19 to present. HF2 registry is maintained at University of Kansas Medical Center (KUMC). The registry was approved by the institutional review board (IRB) at all participating institutions with required data use agreements. Institutions report data into the electronic registry database using REDCap, housed at KUMC. Results: This initial data set includes 254 patients implanted from the start of 2019 until May 2023. At time of device implant, the cohort average age is 73 years old, 59.8% are male, 72% have NYHA Class III HF, 40% have left ventricular ejection fraction (LVEF) < 40%, 35% have LVEF > 50%, mean BNP is 560 pg/ml, mean N-Terminal pro-BNP (NTproBNP) is 5,490 pg/ml, mean creatinine is 1.65 mg/dl. Average baseline hemodynamics at device implant are right atrial pressure (RAP) of 11 mmHg, pulmonary artery systolic pressure (PASP) of 47 mmHg, pulmonary artery diastolic pressure (PADP) 21 mmHg, mean pulmonary artery pressure (mPAP) of 20 mmHg, pulmonary capillary wedge pressure (PCWP) of 19 mmHg, cardiac output (CO) of 5.3 L/min, and cardiac index (CI) of 2.5 L/min/m2. Conclusion: A real-world registry of patients implanted with a PAP sensor enables long-term evaluation of hemodynamic and clinic outcomes in highly-phenotyped ambulatory HF patients, and creates a unique opportunity to validate and test novel diagnostic and therapeutic approaches to HF.

A case report of constrictive pericarditis following COVID-19 vaccination.

Background: COVID-19 infection and the COVID-19 vaccines have been associated with rare cases of pericarditis. We present a case of constrictive pericarditis (CP) following the vaccine. Case summary: A 19-year-old healthy male started having progressive abdominal pain, emesis, dyspnoea, and pleuritic chest pain 2 weeks after the second dose of Pfizer vaccine. Computed tomography angiography chest revealed bilateral pleural effusions and pericardial thickening with effusion. Cardiac catheterization showed ventricular interdependence. Cardiac magnetic resonance (CMR) showed septal bounce and left ventricular tethering suggestive of CP. A total pericardiectomy was performed with significant symptom improvement. Pathology showed chronic fibrosis without amyloid, iron deposits, or opportunistic infections. Patient had Epstein-Barr Virus (EBV) viraemia 825 IU/mL and histoplasmosis complement-fixation positive with negative serum and urine antigen. Hypercoagulable panel and infectious workup were otherwise negative. The patient had resolution of cardiac symptoms at 3 months of follow-up. Discussion: The patient developed progressive symptoms within 2 weeks of his second Pfizer vaccine. Echocardiogram and CMR had classic signs of CP, and pericardial pathology confirmed fibrotic pericardium. The patient had no prior surgery, thoracic radiation, or bacterial infection. Epstein-Barr Virus viraemia was thought to be reactionary, and histoplasmosis complement likely represented chronic exposure. The timing of symptoms and negative multidisciplinary workup raises the suspicion for COVID vaccine-induced CP. The COVID vaccines benefits far exceed the risks, but complications still can occur. Practitioners should have a high index of suspicion to allow prompt diagnosis of CP.

Intensive Cardiac Rehabilitation Outcomes in Patients with Heart Failure.

INTRODUCTION: Cardiac rehabilitation (CR) has proven to be beneficial for patients with heart failure (HF), potentially reducing morbidity and mortality while improving fitness and psychological outcomes. Intensive cardiac rehabilitation (ICR) represents an emerging form of CR that has demonstrated advantages for patients with various cardiovascular diseases. Nevertheless, the specific outcomes of ICR in patients with HF remain unknown. OBJECTIVES: The purpose of this study is to assess the effectiveness of ICR in patients with HF. METHODS: This retrospective study involved 12,950 patients who participated in ICR at 46 centers from January 2016 to December 2020. Patients were categorized into two groups: the HF group, comprising 1400 patients (11%), and the non-HF group, consisting of 11,550 patients (89%). The primary endpoints included the ICR completion rate, changes in body mass index (BMI), exercise minutes per week (EMW), and depression scores (CESD). A t-test was employed to compare variables between the two groups. RESULTS: The HF group comprises older patients, with 37% being females (compared to 44% in the non-HF group). The ICR completion rate was higher in the non-HF group. After ICR completion, adjusted analyses revealed that patients without HF demonstrated a greater improvement in BMI. There were no differences in fitness, as measured via EMW, or in depression scores, as measured via CESD, between the two groups. CONCLUSIONS: Despite the lower baseline functional status and psychosocial scores of HF patients compared to non-HF patients, patients with HF were able to attain similar or even better functional and psychosocial outcomes after ICR.

Impact of Female Sex on Cardiogenic Shock Outcomes: A Cardiogenic Shock Working Group Report.

BACKGROUND: Studies reporting cardiogenic shock (CS) outcomes in women are scarce. OBJECTIVES: The authors compared survival at discharge among women vs men with CS complicating acute myocardial infarction (AMI-CS) and heart failure (HF-CS). METHODS: The authors analyzed 5,083 CS patients in the Cardiogenic Shock Working Group. Propensity score matching (PSM) was performed with the use of baseline characteristics. Logistic regression was performed for log odds of survival. RESULTS: Among 5,083 patients, 1,522 were women (30%), whose mean age was 61.8 ± 15.8 years. There were 30% women and 29.1% men with AMI-CS (P = 0.03). More women presented with de novo HF-CS compared with men (26.2% vs 19.3%; P < 0.001). Before PSM, differences in baseline characteristics and sex-specific outcomes were seen in the HF-CS cohort, with worse survival at discharge (69.9% vs 74.4%; P = 0.009) and a higher rate of maximum Society for Cardiac Angiography and Interventions stage E (26% vs 21%; P = 0.04) in women than in men. Women were less likely to receive pulmonary artery catheterization (52.9% vs 54.6%; P < 0.001), heart transplantation (6.5% vs 10.3%; P < 0.001), or left ventricular assist device implantation (7.8% vs 10%; P = 0.01). Regardless of CS etiology, women had more vascular complications (8.8% vs 5.7%; P < 0.001), bleeding (7.1% vs 5.2%; P = 0.01), and limb ischemia (6.8% vs 4.5%; P = 0.001). More vascular complications persisted in women after PSM (10.4% women vs 7.4% men; P = 0.06). CONCLUSIONS: Women with HF-CS had worse outcomes and more vascular complications than men with HF-CS. More studies are needed to identify barriers to advanced therapies, decrease complications, and improve outcomes of women with CS.

No Evidence for Oversizing Hearts and Donor Size Impact on 1-Year Survival in Heart Failure Patients With Left Ventricular Assist Device.

The predicted heart mass (PHM) ratio has recently emerged as a better metric for donor-to-recipient size-matching than weight ratios. It is unknown whether this applies to transplant candidates on left ventricular assist device (LVAD) support. Our study examines if PHM ratio is optimal for size-matching specifically in the LVAD patient population. Patients with LVAD who received a heart transplant from January 1997 to December 2020 in the Scientific Registry of Transplant Recipients database were studied. We compared 5 size-matching metrics, including donor-recipient ratios of weight, height, body mass index, body surface area, and PHM. Single and multivariable Cox proportional hazards models for 1-year mortality were calculated. Our sample consisted of 11,891 patients. In our multivariate analysis, we found that patients in the undersized group with PHM ratios <0.83 had a hazard ratio for 1-year mortality of 1.34 (95% confidence interval 1.08 to 1.65, p = 0.007) suggestive of increased mortality with the use of undersized donors. There was no statistical difference in mortality between the matched (PHM ratio 0.83 to 1.2) and oversized group (PHM ratio ≥1.2). In heart transplant recipients on LVAD support, the PHM ratio provides better risk stratification than other metrics. Use of undersized donor hearts with PHM ratio <0.83 confers higher 1-year mortality. Using oversized donor hearts for transplantation in recipients on LVAD support has no benefit.

Risk of Arrhythmic Death in Patients With Nonischemic Cardiomyopathy: JACC Review Topic of the Week.

Nonischemic cardiomyopathy (NICM) is common and patients are at significant risk for early mortality secondary to ventricular arrhythmias. Current guidelines recommend implantable cardioverter-defibrillator (ICD) therapy to decrease sudden cardiac death (SCD) in patients with heart failure and reduced left ventricular ejection fraction. However, in randomized clinical trials comprised solely of patients with NICM, primary prevention ICDs did not confer significant mortality benefit. Moreover, left ventricular ejection fraction has limited sensitivity and specificity for predicting SCD. Therefore, precise risk stratification algorithms are needed to define those at the highest risk of SCD. This review examines mechanisms of sudden arrhythmic death in patients with NICM, discusses the role of ICD therapy and treatment of heart failure for prevention of SCD in patients with NICM, examines the role of cardiac magnetic resonance imaging and computational modeling for SCD risk stratification, and proposes new strategies to guide future clinical trials on SCD risk assessment in patients with NICM.

Translating Pressure Into Practice: Operational Characteristics of Ambulatory Hemodynamic Monitoring Program in the United States.

BACKGROUND: Ambulatory hemodynamic monitoring (AHM) using an implantable pulmonary artery pressure sensor (CardioMEMS) is effective in improving outcomes for patients with heart failure. The operations of AHM programs are crucial to clinical efficacy of AHM yet have not been described. METHODS AND RESULTS: An anonymous, voluntary, web-based survey was developed and emailed to clinicians at AHM centers in the United States. Survey questions were related to program volume, staffing, monitoring practices, and patient selection criteria. Fifty-four respondents (40%) completed the survey. Respondents were 44% (n = 24) advanced HF cardiologists and 30% (n = 16) advanced nurse practitioners. Most respondents practice at a center that implants left ventricular assist devices (70%) or performs heart transplantation (54%). Advanced practice providers provide day-to-day monitoring and management in most programs (78%), and use of protocol-driven care is limited (28%). Perceived patient nonadherence and inadequate insurance coverage are cited as the primary barriers to AHM. CONCLUSIONS: Despite broad US Food and Drug Administration approval for patients with symptoms and at increased risk for worsening heart failure, the adoption of pulmonary artery pressure monitoring is concentrated at advanced heart failure centers, and modest numbers of patients are implanted at most centers. Understanding and addressing the barriers to referral of eligible patients and to broader adoption in community heart failure programs is needed to maximize the clinical benefits of AHM.

Evaluation of adjusted international normalized ratio goal in patients with HeartMate 3 left ventricular assist devices.

BACKGROUND: The HeartMate3 left ventricular assist device (HM3 LVAD) has shown a low incidence of thrombosis, but bleeding risk is as high as 43%. We aim to describe the impact of lower international normalization ratio (INR) goal on clinical outcomes. METHODS: In February 2019, our tertiary care institution lowered INR goal in HM3 patients from manufacturer recommendations to 1.8-2.2 and retrospectively analyzed the data. Two cohorts were compared: patients with lower INR goal upon implant (De novo) and those with subsequently lowered INR goal (Adjusted). The Adjusted group also served as its own historical control. Both groups continued aspirin 81 milligrams daily per manufacturer recommendations. The primary outcomes were incidences of bleed and thrombosis events within 12 months. Secondary outcomes included survival free of disabling stroke or reoperation to remove or replace the device and Rosendaal time in therapeutic range (TTR) over 12 months. RESULTS: Thirty-one patients were evaluated for inclusion with 26 meeting criteria. Within 12 months, incidence of bleeding events was 25% and 28.6% in the De novo and Adjusted groups, respectively. Incidence of thrombotic events within 12 months was 0% in the De novo group and 7.1% in the Adjusted group. Twelve-month survival free of disabling stroke or reoperation to remove or replace the device was higher over 12 months for patients in the De novo group (91.7% vs. 78.6%). Median 12-month TTR was 36%, which was primarily attributable to subtherapeutic deviations. CONCLUSIONS: A lower INR goal may be safe when initiated De novo following implantation of the HM3. This study informs the need for larger prospective studies.

Aortic Thrombosis in Patients on Mechanical Circulatory Support: A Systematic Literature Review.

BACKGROUND: Aortic valve (AV) thrombosis is an uncommon but ominous complication in patients managed with mechanical circulatory support (MCS) devices. In this systematic review, we summarised the data on clinical presentations and outcomes in such patients. METHODS: We searched articles on PubMed and Google Scholar, reporting at least one adult patient with aortic thrombosis on MCS support and where the individual patient data could be extracted. We grouped the patients by the type of MCS (temporary or durable), and the type of the AV (prosthetic, surgically modified, or native) RESULTS: We identified reports on six patients with aortic thrombus on short-term MCS, and on 41 patients on durable left ventricular assist devices (LVADs). On temporary MCS, AV thrombus typically causes no symptoms and is found incidentally pre- or intra-operatively. For those with durable MCS, the occurrence of aortic thrombus forming on prosthetic or surgically modified valves appears to be more related to the intervention on the valve, rather than from the presence of LVAD. The mortality in this group was 18%. In patients with native AV on durable LVAD support, 60% of patients presented with acute myocardial infarction, acute stroke, or acute heart failure, and mortality in this cohort was 45%. In terms of management, heart transplantation was most successful. CONCLUSIONS: While the outcomes of aortic thrombosis were good in patients where temporary MCS was used in the setting of aortic valve surgery, patients with native AV who develop this complication on durable LVAD have high morbidity and mortality. Cardiac transplantation should be strongly considered in eligible candidates because other therapies provide inconsistent results.

PCSK9 Inhibition in Patients After Heart Transplantation: a Retrospective Review and Literature Analysis.

PURPOSE OF REVIEW: Following cardiac transplantation, patients have an increased risk of developing cardiac allograft vasculopathy and atherosclerotic cardiovascular disease. Therefore, aggressive lipid management is indicated. Some patients do not achieve optimal lipid profiles with statin monotherapy, however, or discontinue statins due to intolerance. In this review, we investigated the use of PCSK9 inhibitors as an alternative treatment for hyperlipidemia following cardiac transplantation. RECENT FINDINGS: Nine published articles were identified that included 110 patients treated with alirocumab or evolocumab after cardiac transplantation. PCSK9 inhibitors were tolerated by all patients, and each study demonstrated an effective reduction of low-density lipoprotein ranging from 40 to 87% decrease from baseline. In our study, the 110 patients from literature review were added to a cohort of 7 similar patients from our institution for combined analysis. This report supports that PCSK9 inhibitors should be considered following cardiac transplantation when conventional medial therapy is not tolerated or ineffective.