Artificial intelligence-assisted automated heart failure detection and classification from electronic health records

AIMS

Electronic health records (EHR) linked to Digital Imaging and Communications in Medicine (DICOM), biological specimens, and deep learning (DL) algorithms could potentially improve patient care through automated case detection and surveillance. We hypothesized that by applying keyword searches to routinely stored EHR, in conjunction with AI-powered automated reading of DICOM echocardiography images and analysing biomarkers from routinely stored plasma samples, we were able to identify heart failure (HF) patients.

METHODS AND RESULTS

We used EHR data between 1993 and 2021 from Tayside and Fife (~20% of the Scottish population). We implemented a keyword search strategy complemented by filtering based on International Classification of Diseases (ICD) codes and prescription data to EHR data set. We then applied DL for the automated interpretation of echocardiographic DICOM images. These methods were then integrated with the analysis of routinely stored plasma samples to identify and categorize patients into HF with reduced ejection fraction (HFrEF), HF with preserved ejection fraction (HFpEF), and controls without HF. The final diagnosis was verified through a manual review of medical records, measured natriuretic peptides in stored blood samples, and by comparing clinical outcomes among groups. In our study, we selected the patient cohort through an algorithmic workflow. This process started with 60 850 EHR data and resulted in a final cohort of 578 patients, divided into 186 controls, 236 with HFpEF, and 156 with HFrEF, after excluding individuals with mismatched data or significant valvular heart disease. The analysis of baseline characteristics revealed that compared with controls, patients with HFrEF and HFpEF were generally older, had higher BMI, and showed a greater prevalence of co-morbidities such as diabetes, COPD, and CKD. Echocardiographic analysis, enhanced by DL, provided high coverage, and detailed insights into cardiac function, showing significant differences in parameters such as left ventricular diameter, ejection fraction, and myocardial strain among the groups. Clinical outcomes highlighted a higher risk of hospitalization and mortality for HF patients compared with controls, with particularly elevated risk ratios for both HFrEF and HFpEF groups. The concordance between the algorithmic selection of patients and manual validation demonstrated high accuracy, supporting the effectiveness of our approach in identifying and classifying HF subtypes, which could significantly impact future HF diagnosis and management strategies.

CONCLUSIONS

Our study highlights the feasibility of combining keyword searches in EHR, DL automated echocardiographic interpretation, and biobank resources to identify HF subtypes.

Long term follow up and outcomes of Covid-19 vaccine associated myocarditis in Victoria, Australia: A clinical surveillance study

BACKGROUND

Myocarditis and myopericarditis are well described adverse events of special interest (AESI) following COVID-19 vaccinations. Although reports are reassuring regarding initial clinical outcomes, information about longer term outcomes remains limited. We aimed to further this knowledge and report outcomes to 6 months post diagnosis from a single population cohort.

METHODS

Reports of myocarditis following COVID-19 vaccination were followed up by SAEFVIC (Surveillance of Adverse Events Following Vaccination in the Community), the state-wide vaccine safety service for Victoria, Australia. Confirmed myocarditis cases (Brighton Collaboration Criteria levels 1-3) were followed up via surveys at 1, 3 and 6 months post symptom onset. Responses received between 22 February 2021 and 30 September 2022 were analysed.

RESULTS

87.5 % (N = 182) of eligible participants completed at least 1 survey report. 377 reports were analysed. 76.9 % of completed reports were from male patients. The median age of patients was 21 years [IQR: 16 to 32]. 54.8 % (n = 74) of survey reports at 6 months, reported ongoing symptoms. At all follow-up time points, females were significantly more likely to have ongoing symptoms. At 6 months, 51.9 % of male respondents reported symptom resolution compared to 22.6 % of female patients (p = 0.002). Females were also more likely to continue medication and have ongoing exercise restrictions. However, males were significantly more likely to have higher initial peak troponin results and abnormal initial cardiac imaging investigations.

CONCLUSIONS

There appears to be a significant proportion of patients who experience ongoing symptoms to 6 months post onset amongst patients that experience these AESI. Male patients were more likely to report earlier and more complete symptom recovery, despite significantly higher average initial peak troponin. This difference in phenotypic presentation in females compared to males warrants further investigation and there is a need for longer term follow up data.

Diagnostic Sensitivity of Unenhanced CT for Cerebral Venous Thrombosis: Can Clot Density Measurement Replace CT Venogram?

Objectives Cerebral venous sinus thrombosis is an important cause of stroke in young adults. Noncontrast-enhanced CT head (NECT) is almost always the first investigation.

Our objectives were as follows:

1. How accurately does venous sinus density on NECT predict the presence of clot on CT venogram (CTV)?

2. Whether repeated measurements changed the confidence?

3. How many venous sinus thrombus would be missed if we do not do a CTV?

4. Can clot density measurement replace CTV?

Methods Multicenter case–control study was designed with data from seven hospitals. Inclusion criteria: all CT and magnetic resonance imaging venograms with a prior NECT, performed between 1.1.2018 and 31.12.2018 (12 months), were included. Hounsfield unit (HU) values were calculated at the site of highest density on the NECT. Logistic regression analysis was performed using STATA.

Result Two-hundred seventy-seven cases met the criteria with 33 positive cerebral venous thrombosis (density on NECT 60–92 HU) and 244 negative examinations (density on NECT 31–68 HU). Area under the curve for average clot density on NECT was 0.9984.

Conclusion We found a strong relationship between sinus density on NECT and outcome of CTV. Repeating density measurements did not add any predictive value or changed outcome.

Advances in Knowledge Density 70 HU or higher on NECT always resulted in a positive CTV but would miss a fifth of the positives. Cutoff at 60 HU would not miss any but result in significant false positives. An efficient option could be to limit CTV to sinus densities 60 to 70 HU only. However, a larger study would be required for such change in practice.

UpStreAm doxycycline in ST-eLeVation myocArdial infarction - targetinG infarct hEaling and ModulatIon (SALVAGE-MI trial)

BACKGROUND AND AIMS

Experimental studies demonstrate protective effects of doxycycline on myocardial ischemia-reperfusion injury. The trial investigated whether doxycycline administered prior to reperfusion in patients presenting with ST-elevation myocardial infarction (STEMI) reduces infarct size (IS) and ameliorates adverse left ventricular (LV) remodeling.

METHODS

In this randomized, double-blind, placebo-controlled trial, patients presenting with STEMI undergoing primary percutaneous coronary intervention (PPCI) were randomized to either intravenous doxycycline or placebo prior to reperfusion followed by 7-days of oral doxycycline or placebo. The primary outcome was final IS adjusted for area-at-risk (fIS/AAR) measured on two cardiac magnetic resonance scans ∼6 months apart.

RESULTS

Of 103 participants, 50 were randomized to doxycycline and 53 to placebo and were matched for age (59 ± 12 vs. 60 ± 10 years), male sex (92% vs. 79%), diabetes mellitus (26% vs. 11%) and left anterior descending artery occlusion (50% vs. 49%), all p > 0.05. Patients treated with doxycycline had a trend for larger fIS/AAR (0.79 [0.5-0.9] vs. 0.61 [0.47-0.76], p = 0.06), larger fIS at 6 months (18.8% [12-26] vs. 13.6% [11-21], p = 0.08), but similar acute IS (21.7% [17-34] vs. 19.4% [14-27], p = 0.19) and AAR (26% [20-36] vs. 24.7% [16-31], p = 0.22) compared to placebo. Doxycycline did not ameliorate adverse LV remodeling (%Δend-diastolic volume index, 1.1% [-3.8-8.4] vs. -1.34% [-6.1-5.8], p = 0.42) and was independently associated with larger fIS (regression coefficient = 0.175, p = 0.03).

CONCLUSION

Doxycycline prior to PPCI neither reduced IS acutely or at 6 months nor attenuated adverse LV remodeling. These data raise safety concerns regarding doxycycline use in STEMI for infarct modulation and healing.

Current approaches to the diagnosis and management of amyloidosis

Amyloidosis is a collection of diseases caused by the misfolding of proteins that aggregate into insoluble amyloid fibrils and deposit in tissues. While these fibrils may aggregate to form insignificant localised deposits, they can also accumulate in multiple organs to the extent that amyloidosis can be an immediately life-threatening disease, requiring urgent treatment. Recent advances in diagnostic techniques and therapies are dramatically changing the disease landscape and patient prognosis. Delays in diagnosis and treatment remain the greatest challenge, necessitating physician awareness of the common clinical presentations that suggest amyloidosis. The most common types are transthyretin (ATTR) amyloidosis followed by immunoglobulin light-chain (AL) amyloidosis. While systemic AL amyloidosis was previously considered a death sentence with no effective therapies, significant improvement in patient survival has occurred over the past 2 decades, driven by greater understanding of the disease process, risk-adapted adoption of myeloma therapies such as proteosome inhibitors (bortezomib) and monoclonal antibodies (daratumumab) and improved supportive care. ATTR amyloidosis is an underdiagnosed cause of heart failure. Technetium scintigraphy has made noninvasive diagnosis much easier, and ATTR is now recognised as the most common type of amyloidosis because of the increased identification of age-related ATTR. There are emerging ATTR treatments that slow disease progression, decrease patient hospitalisations and improve patient quality of life and survival. This review aims to update physicians on recent developments in amyloidosis diagnosis and management and to provide a diagnostic and treatment framework to improve the management of patients with all forms of amyloidosis.

Spontaneous Coronary Artery Dissection in an Orthotopic Heart Transplant Recipient

We present the case of acute myocardial infarction secondary to spontaneous coronary artery dissection in a patient 2 weeks post orthotopic heart transplantation. (Level of Difficulty: Advanced.)

Vasoplegia Following Orthotopic Heart Transplantation: Prevalence, Predictors and Clinical Outcomes

BACKGROUND

Patients undergoing heart transplant are at high risk for postoperative vasoplegia. Despite its frequency and association with poor clinical outcomes, there remains no consensus definition for vasoplegia, and the predisposing risk factors for vasoplegia remain unclear. Accordingly, the aim of this study was to evaluate the prevalence, predictors, and clinical outcomes associated with vasoplegia in a contemporary cohort of patients undergoing heart transplantation.

METHODS

This was a retrospective cohort study of patients undergoing heart transplantation from January 2015 to December 2019. A binary definition of vasoplegia of a cardiac index of 2.5 L/min/m² or greater and requirement for norepinephrine (≥5 µg/min), epinephrine (≥4 µg/min), or vasopressin (≥1 unit/h) to maintain a mean arterial blood pressure of 65 mm Hg, for 6 consecutive hours during the first 48 hours postoperatively, was used in determining prevalence. Given the relatively low threshold for the binary definition of vasoplegia, patients were divided into tertiles based on their cumulative vasopressor requirement in the 48 hours following transplant. Outcomes included all-cause mortality, intubation time, intensive care unit length of stay, and length of total hospitalization.

RESULTS

After exclusion of patients with primary cardiogenic shock, major bleeding, or overt sepsis, data were collected on 95 eligible patients. By binary definition, vasoplegia incidence was 66.3%. We separately stratified by actual vasopressor requirement tertile (high, intermediate, low). Stratified by tertile, patients with vasoplegia were older (52.7 ± 10.2 vs 46.8 ± 12.7 vs 44.4 ± 11.3 years, P = .02), with higher rates of chronic kidney disease (18.8% vs 32.3% vs 3.1%, P = .01) and were more likely to have been transplanted from left ventricular assist device support (n = 42) (62.5% vs 32.3% vs 37.5%, P = .03). Cardiopulmonary bypass time was prolonged in those that developed vasoplegia (155 min [interquartile range 135-193] vs 131 min [interquartile range 117-152] vs 116 min [interquartile range 102-155], P = .003). Intubation time and length of intensive care unit and hospital stay were significantly increased in those that developed vasoplegia; however, this difference did not translate to a significant increase in all-cause mortality at 30 days or 1 year.

CONCLUSIONS

Vasoplegia occurs at a high rate after heart transplantation. Older age, chronic kidney disease, mechanical circulatory support, and prolonged bypass time are all associated with vasoplegia; however, this study did not demonstrate an associated increase in all-cause mortality LAY SUMMARY: Patients undergoing heart transplantation are at high risk of vasoplegia, a condition defined by low blood pressure despite normal heart function. We found that vasoplegia was common after heart transplant, occurring in 60%-70% of patients after heart transplant after excluding those with other causes for low blood pressure. Factors implicated included age, poor kidney function, prolonged cardiopulmonary bypass time and preoperative left ventricular assist device support. We found no increased risk of death in patients with vasoplegia despite longer lengths of stay in intensive care and in hospital.

Incidence and predictors of eosinophilic myocardial hypersensitivity in patients receiving home dobutamine

Background

Home inotropes are utilised in those with end-stage heart failure as a bridge to cardiac transplantation. The use of intravenous dobutamine has been linked to cases of eosinophilic myocardial hypersensitivity (EMH), however, little is known about incidence and predictors.

Purpose

We sought to examine the incidence and possible predictors of eosinophilic myocardial hypersensitivity in a cohort of patients on home inotrope therapy at a cardiac transplant centre.

Methods

Patients enrolled in the home inotrope program with progression to heart transplantation or ventricular assist device (VAD) with available myocardial tissue for histopathology, from January 2000 to May 2020 were included. EMH was defined by a pathologist reporting eosinophilic infiltrate with hypersensitivity on myocardial histopathology.

Results

From a cohort of 74 patients, 58% (43) were on dobutamine and 42% (31) were on milrinone. There were zero cases of EMH in those on milrinone. EMH was identified in 14% (6/43) of patients receiving dobutamine. In the dobutamine cohort, the mean age was 52-±12 years, with 22% being female. Non-ischaemic dilated cardiomyopathy encompassed 62%, the remaining 38% were ischaemic cardiomyopathy. Median dobutamine dose (250 [200–282] mcg/min vs. 225 [200–291] mcg/min) and duration of therapy (41 [23–79] days vs. 53 [24–91] days) were similar between those with and without EMH. Rates of known allergy (27% vs. 33%) and asthma (1 patient in each group) were also similar between those with and without EMH. Those with EMH had a median peak eosinophil count of 0.40×109/L (IQR 0.21–0.66×109/L) compared to a peak of only 0.10×109/L (IQR 0.06–0.29×109/L) in the non-EMH cohort. There was a significant difference in the change in absolute eosinophil count between groups; over the duration of dobutamine therapy the median change in eosinophil count was 0.31×109/L (IQR 0.21–0.59×109/L) in the EMH group compared to 0.03×109/L (IQR 0.00–0.14×109/L) in the non-EMH cohort (p=0.02). Peak C-reactive protein was similar between groups (42±46mg/L vs. 44±45mg/L). Mean left ventricular ejection fraction (LVEF) reduced from 19% (±7%) to 17% (±2%) in those with EMH, while LVEF increased from 20% (±7%) to 22% (±9%) in non-EMH patients (Figure 1), p=NS. Re-presentation with heart failure requiring hospitalisation occurred in 83% in the EMH group compared to only 59% in the non-EMH group (p=0.26). The majority of patients with EMH (83%) required VAD as bridge to transplant, compared to only 41% of non-EMH (p=0.05).

Conclusion(s)

EMH occurred in 14% of patients receiving home dobutamine. Patients who developed EMH were more likely to require escalation in treatment to VAD as a bridge to heart transplant. In patients receiving dobutamine a reduction in LVEF, hospitalisation with decompensated heart failure and rising eosinophil count should prompt physicians to consider EMH.

Funding Acknowledgement

Type of funding sources: None. Figure 1

A Single-Center Experience of the Optimal Initial Immunosuppressive Strategy for Preventing Early Acute Cellular Rejection in Orthotopic Heart Transplantation Associated With Renal Dysfunction

BACKGROUND

Renal dysfunction is a common complication following heart transplantation that may be worsened by the early initiation of calcineurin inhibitors. Antithymocyte globulin (ATG) or basiliximab has been used to delay or avoid calcineurin inhibitors. The most effective strategy for preventing early acute cellular rejection in this context is uncertain.

METHODS

We retrospectively reviewed all heart transplant cases between January 2012 and June 2017. The standard therapy consisted of mycophenolate mofetil, prednisolone, and tacrolimus. In patients at high risk of post-transplant renal dysfunction, an early calcineurin inhibitor-free regimen with basiliximab and/or ATG was used. Patients were assigned to cohorts based on the initial immunosuppressive strategy. The primary end point was the freedom rate of acute cellular rejection within 4 weeks post-transplant.

RESULTS

Of 93 cases, 21 patients received standard therapy, 64 patients received an initial calcineurin inhibitor-free regimen with basiliximab, and 8 patients received ATG and basiliximab. Freedom from acute rejection was greater in the ATG plus basiliximab group (all rejection free), compared to 40 (63%) of 64 patients treated with basiliximab and 10 (48%) of 21 patients treated with standard therapy (P < .05, log rank test). In patients treated with basiliximab, early administration (<24 hours) was associated with a higher freedom from acute rejection compared to ≥24 hours, (72% vs 29%, P < .05).

CONCLUSIONS

The combination of ATG and basiliximab was more effective in preventing acute cellular rejection. In those patients treated with basiliximab, rejection rates were no worse than standard therapy; however, it was only effective when administered within 24 hours.

Perspective: APOBEC mutagenesis in drug resistance and immune escape in HIV and cancer evolution

The apolipoprotein B mRNA-editing enzyme, catalytic polypeptide-like (APOBEC) mutational signature has only recently been detected in a multitude of cancers through next-generation sequencing. In contrast, APOBEC has been a focus of virology research for over a decade. Many lessons learnt regarding APOBEC within virology are likely to be applicable to cancer. In this review, we explore the parallels between the role of APOBEC enzymes in HIV and cancer evolution. We discuss data supporting the role of APOBEC mutagenesis in creating HIV genome heterogeneity, drug resistance, and immune escape variants. We hypothesize similar functions of APOBEC will also hold true in cancer.

The bioequivalence of a single intravenous administration of the anesthetic alfaxalone in cyclodextrin versus alfaxalone in cyclodextrin plus preservatives in cats

To demonstrate the bioequivalence of alfaxalone in cyclodextrin (Reference Product) to a formulation of alfaxalone in cyclodextrin also containing the preservatives ethanol, chlorocresol, and benzethonium chloride (Test Product) when administered for the purpose of inducing anesthesia in the cat. Blinded, single-dose, randomized, two-period, two-sequence, cross-over bioequivalence study with a 7-day washout period between treatments. Twenty-four (12 neutered males and 12 intact females), healthy, adult cats weighing 4.1±0.9 kg. Cats were administered 5 mg/kg IV of alfaxalone in the Reference or Test Product using a randomized cross-over design. One-milliliter venous blood samples were collected at predetermined time points to 12 hr after drug administration to determine alfaxalone plasma concentration over time. Alfaxalone concentrations were determined by a validated analytical testing method using HPLC-MS/MS. Plasma profiles of alfaxalone concentration against time were analyzed by noncompartmental analysis. The pivotal variables for bioequivalence were AUC_last and C_max . Equivalence was achieved if the 90% confidence interval for AUC_last and C_max fell into the asymmetric ±20% interval (0.80-1.25). Physiological variables, quality of anesthesia visual analog scale (VAS) scoring and anesthetic event times were recorded. ANOVA or ANCOVA (single time point), RMANOVA or RMANCOVA (multiple time point) was used for normally distributed data. GLIMMIX was used for nonnormally distributed data. VAS scores were analyzed as for blood bioequivalence data. Variables were evaluated for safety and assessed at alpha = 0.10. C_max and AUC_last for Reference and Test Products were statistically bioequivalent. No physiological variables except for a drug by time interaction for respiratory rate differed between treatment groups, and this difference was not clinically relevant. No anesthetic event times or VAS scores for quality of anesthesia were different between treatment groups. Neither formulation caused pain upon injection. The Reference and Test Products are pharmaceutically bioequivalent formulations when administered as a single intravenous administration for the purpose of induction of anesthesia in cats.

Induction of heart failure: haemodynamic comparison of three different canine models

We compared three methods: arteriovenous anastomosis, doxorubicin administration, and combination of anastomosis and doxorubicin, with the intention of designing a simple, stable model of chronic heart failure. Twelve dogs were divided into three groups of four. One group received carotid-jugular anastomosis (Ana series), another group received anastomosis and doxorubicin injection (A/D series), and the last group received only doxorubicin (Dox series). Animals were followed for eight weeks. Fifteen different haemodynamic parameters were tracked and compared to baseline values. After eight weeks, diastolic pressure in the right atrium increased from 3.8±2.0 mmHg at baseline to 5.3±5.9 mmHg in the Ana series, to 6.3±3.3 mmHg in the Dox series and to 8.0±2.0 mmHg in the A/D series ( P<0.05 A/D vs. baseline). Systolic pulmonary wedge pressure increased from 11.6±2.0 mmHg at baseline to 15.5±3.4 mmHg in the Ana series, 14.0±3.7 mmHg in the Dox series and 17.3±4.2 mmHg in the A/D series ( P = NS vs. baseline). Left ventricular ejection fraction decreased from 53.9±10% at baseline to 36.1±5.6% in the Ana series ( P<0.05 vs. baseline), 31.5±5.4% in the Dox series ( P<0.05 vs. baseline) and 25.8±5.8% in the A/D series ( P<0.001 vs. baseline, P<0.05 vs. Ana series and Dox series). In conclusion, eight weeks are not enough to produce stable heart failure using arteriovenous anastomosis alone. Doxorubicin administration alone produces a left ventricular failure. However, a combination of both of these interventions provides a more stable model of right-and left-sided heart failure