Development and external validation of a dynamic risk score for early prediction of cardiogenic shock in cardiac intensive care units using machine learning

BACKGROUND

Myocardial infarction and heart failure are major cardiovascular diseases that affect millions of people in the US with the morbidity and mortality being highest among patients who develop cardiogenic shock. Early recognition of cardiogenic shock allows prompt implementation of treatment measures. Our objective is to develop a new dynamic risk score, called CShock, to improve early detection of cardiogenic shock in cardiac intensive care unit (ICU).

METHODS

We developed and externally validated a deep learning-based risk stratification tool, called CShock, for patients admitted into the cardiac ICU with acute decompensated heart failure and/or myocardial infarction to predict onset of cardiogenic shock. We prepared a cardiac ICU dataset using MIMIC-III database by annotating with physician adjudicated outcomes. This dataset that consisted of 1500 patients with 204 having cardiogenic/mixed shock was then used to train CShock. The features used to train the model for CShock included patient demographics, cardiac ICU admission diagnoses, routinely measured laboratory values and vital signs, and relevant features manually extracted from echocardiogram and left heart catheterization reports. We externally validated the risk model on the New York University (NYU) Langone Health cardiac ICU database that was also annotated with physician adjudicated outcomes. The external validation cohort consisted of 131 patients with 25 patients experiencing cardiogenic/mixed shock.

RESULTS

CShock achieved an area under the receiver operator characteristic curve (AUROC) of 0.821 (95% CI 0.792-0.850). CShock was externally validated in the more contemporary NYU cohort and achieved an AUROC of 0.800 (95% CI 0.717-0.884), demonstrating its generalizability in other cardiac ICUs. Having an elevated heart rate is most predictive of cardiogenic shock development based on Shapley values. The other top ten predictors are having an admission diagnosis of myocardial infarction with ST-segment elevation, having an admission diagnosis of acute decompensated heart failure, Braden Scale, Glasgow Coma Scale, Blood urea nitrogen, Systolic blood pressure, Serum chloride, Serum sodium, and Arterial blood pH.

CONCLUSIONS

The novel CShock score has the potential to provide automated detection and early warning for cardiogenic shock and improve the outcomes for the millions of patients who suffer from myocardial infarction and heart failure.

Existing Nongated CT Coronary Calcium Predicts Operative Risk in Patients Undergoing Noncardiac Surgeries (ENCORES)

BACKGROUND

Preoperative cardiovascular risk stratification before noncardiac surgery is a common clinical challenge. Coronary artery calcium scores from ECG-gated chest computed tomography (CT) imaging are associated with perioperative events. At the time of preoperative evaluation, many patients will not have had ECG-gated CT imaging, but will have had nongated chest CT studies performed for a variety of noncardiac indications. We evaluated relationships between coronary calcium severity estimated from previous nongated chest CT imaging and perioperative major clinical events (MCE) after noncardiac surgery.

METHODS

We retrospectively identified consecutive adults age ≥45 years who underwent in-hospital, major noncardiac surgery from 2016 to 2020 at a large academic health system composed of 4 acute care centers. All patients had nongated (contrast or noncontrast) chest CT imaging performed within 1 year before surgery. Coronary calcium in each vessel was retrospectively graded from absent to severe using a 0 to 3 scale (absent, mild, moderate, severe) by physicians blinded to clinical data. The estimated coronary calcium burden (ECCB) was computed as the sum of scores for each coronary artery (0 to 9 scale). A Revised Cardiac Risk Index was calculated for each patient. Perioperative MCE was defined as all-cause death or myocardial infarction within 30 days of surgery.

RESULTS

A total of 2554 patients (median age, 68 years; 49.7% women; median Revised Cardiac Risk Index, 1) were included. The median time interval from nongated chest CT imaging to noncardiac surgery was 15 days (interquartile range, 3-106 days). The median ECCB was 1 (interquartile range, 0-3). Perioperative MCE occurred in 136 (5.2%) patients. Higher ECCB values were associated with stepwise increases in perioperative MCE (0: 2.9%, 1-2: 3.7%, 3-5: 8.0%; 6-9: 12.6%, P<0.001). Addition of ECCB to a model with the Revised Cardiac Risk Index improved the C-statistic for MCE (from 0.675 to 0.712, P=0.018), with a net reclassification improvement of 0.428 (95% CI, 0.254-0.601, P<0.0001). An ECCB ≥3 was associated with 2-fold higher adjusted odds of MCE versus an ECCB <3 (adjusted odds ratio, 2.11 [95% CI, 1.42-3.12]).

CONCLUSIONS

Prevalence and severity of coronary calcium obtained from existing nongated chest CT imaging improve preoperative clinical risk stratification before noncardiac surgery.

Rapidly Progressive Primary Scleroderma Cardiomyopathy in a Young Adult Characterized by Multimodality Imaging

•Scleroderma-related heart disease is usually secondary to lung disease or PH. •Scleroderma rarely causes systolic HF in young patients or those without CAD. •A multimodality strategy should be used to characterize scleroderma cardiomyopathy.

Cysts Around the Heart: Differential Diagnosis and Multimodality Imaging Strategies for Paracardiac Cysts

•Paracardiac cysts (located adjacent to or inside the heart) are rarely detected. •There is a broad differential diagnosis for cysts encountered with echocardiography. •Multimodality imaging is crucial for comprehensively evaluating paracardiac cysts.

Taking Command of Three-Dimensional Stitching Artifacts: From an Annoyance to an Easy Tool for Navigating Three-Dimensional Transesophageal Echocardiography

Despite many recent advances in three-dimensional (3D) transesophageal echocardiography (TEE) imagining, the process of orienting 3D TEE images is nonintuitive and uses assumptions based on idealized anatomy. Correlating two-dimensional TEE cross-sectional images to 3D reconstructions remains an additional challenge. In this article, we suggest the repurposing of the stitching artifact generated in 2-beat electrocardiogram-gated 3D TEE as a means of exactly orienting 3D images within a patient's unique anatomy. We demonstrate the application of this strategy to assess a normal mitral valve to localize scallops of mitral valve prolapse and to visualize typical left atrial appendage two-dimensional cuts in a 3D space. By taking command of stitching artifacts, cardiac imagers can successfully navigate the complex structures of the heart for optimal, individualized echocardiographic views.

Abstract P012: Agreement In Same Day Research-Quality And Casual Systolic Blood Pressure Measurements: Insights From BETTER-BP

Introduction: Prior studies suggest that BPs measured casually during routine clinic workflow are higher than research-grade measurements. However, BPs in these studies were typically collected on different days. We sought to compare the influence of day of assessment on this difference, within the context of an ongoing randomized trial.

Methods: This is a cross-sectional interim analysis of the BETTER-BP trial, which examines the effect of behavioral economic incentives for medication adherence on BP. Our study focused on concordance between baseline BP measurements and did not analyze trial outcome data. We included the first 171 participants (1/1/20-3/31/22). Research BP was measured in a seated position after a 5 minute rest period using a standard cuff (Omron HEM 907-XL). Casual BP measurements (most recent recording prior to the research measurement) as well as baseline characteristics were abstracted from the EHR. Participants were divided into two groups based on whether their casual BP was measured on the same day (vs different day) as their research BP. The mean systolic BP (SBP) obtained for research versus casual measurements were compared, for the two separate groups, using paired T-tests.

Results: Mean age was 55 years, 49.1% were female, 93.2% were nonwhite, and 48.5% were Spanish-speaking. Common comorbidities were diabetes (48.5%) and obesity (49.1%). Among the 132 patients with same day measurements, mean research SBP was 139.2 ± 20.4 mmHg and casual SBP was 137.9 ± 16.4 mmHg (mean within person difference=1.3 mmHg; P=0.259). Among the 39 patients with different days of measurement (median 14 [IQR 6-47] days), mean research SBP was 136.5 ± 23.3 mmHg and casual SBP was 144.2 ± 18.9 mmHg (mean within person difference= -7.7 mmHg; P=0.007). Casual SBP was higher among patients with measurements on different days compared to patients with same day measurements (P=0.043).

Conclusion: SBP measured casually during clinical practice were concordant with research-quality measurements when obtained on the same day. SBP measurements differed significantly when obtained on different days. These findings suggest that prior reports showing casual BPs to be consistently higher than research BPs may have been influenced by temporal variability.

Pre-transplant immune cell function assay as a predictor of early cardiac allograft rejection

INTRODUCTION

ImmuKnow, an immune cell function assay that quantifies overall immune system activity can assist in post-transplant immunosuppression adjustment. However, the utility of pre-transplant ImmuKnow results representing a patient's baseline immune system activity is unknown. This study sought to assess if pre-transplant ImmuKnow results are predictive of rejection at the time of first biopsy in our cardiac transplant population.

METHODS

This is a single center, retrospective observational study of consecutive patients from January 1, 2018 to October 1, 2020 who underwent orthotopic cardiac transplantation at NYU Langone Health. Patients were excluded if a pre-transplant ImmuKnow assay was not performed. ImmuKnow results were categorized according to clinical interpretation ranges (low, moderate, and high activity), and patients were divided into two groups: a low activity group versus a combined moderate-high activity group. Pre-transplant clinical characteristics, induction immunosuppression use, early postoperative tacrolimus levels, and first endomyocardial biopsy results were collected for all patients. Rates of clinically significant early rejection (defined as rejection ≥ 1R/1B) were compared between pre-transplant ImmuKnow groups.

RESULTS

Of 110 patients who underwent cardiac transplant, 81 had pre-transplant ImmuKnow results. The low ImmuKnow activity group was comprised of 15 patients, and 66 patients were in the combined moderate-high group. Baseline characteristics were similar between groups. Early rejection occurred in 0 (0%) patients with low pre-transplant ImmuKnow levels. Among the moderate- high pre-transplant ImmuKnow group, 16 (24.2%) patients experienced early rejection (P = .033). The mean ImmuKnow level in the non-rejection group was the 364.9 ng/ml of ATP compared to 499.3 ng/ml of ATP for those with rejection (P = .020).

CONCLUSION

Patients with low pre-transplant ImmuKnow levels had lower risk of early rejection when compared with patients with moderate or high levels. Our study suggests a possible utility in performing pre-transplant ImmuKnow to identify patients at-risk for early rejection who may benefit from intensified upfront immunosuppression as well as to recognize those where slower calcineurin inhibitor initiation may be appropriate.

Cor Pulmonale from Concomitant Human Immunodeficiency Virus Infection and Methamphetamine Use

•

HIV and frequent methamphetamine use are two known risk factors for PAH development.

•

HIV and methamphetamine are both associated with higher rates of mortality in PAH.

•

Echocardiography is an effective noninvasive modality for assessing PAH severity.

Postinfarction Ventricular Septal Rupture and Hemopericardium with Tamponade Physiology

•

Ventricular septal rupture and hemopericardium are rare postinfarction complications.

•

Contrast[HYPHEN]enhanced echocardiography can help identify pericardial effusion etiologies.

•

Transcaval percutaneous ventricular assist device implantation is a viable strategy.

Effect of Age on Complication Rates and Outcomes Following First Metatarsophalangeal Arthrodesis for Hallux Rigidus

BACKGROUND

Hallux rigidus is one of the most common toe pathologies in patients greater than 50 years old. Although metatarsophalangeal (MTP) arthrodesis is used to improve pain and function, older patients are often considered to be higher-risk operative candidates. There are minimal data on outcomes of MTP fusion and no studies comparing outcomes between geriatric and younger patients.

METHODS

This is a single-center, prospective study of all patients who underwent MTP fusion between August 1, 2015, and July 1, 2018. Patient-reported outcomes were assessed using the Short Form Health Survey (SF-36), Life-Space Assessment survey (LSA), and visual analog scale (VAS) for pain administered preoperatively as well as 6 months and 1 year postoperatively. Baseline characteristics and operative outcomes were collected from the electronic medical record. Clinical and patient-reported outcomes were compared between patients <65 years old and ≥65 years old. Of 143 included patients, 79 were in the younger group (mean of 56.5 years) and 64 were in the older group (mean of 72.0 years).

RESULTS

Compared with the younger group, the older group was more predominantly female (95.3% older group, 77.2% younger group, P = .002). More patients in the older group had hypertension (73.4% older group, 50.6% younger group, P = .005) and coronary artery disease (9.4% older group, 0.0% younger group, P = .005). Both age groups had similar rates of postoperative wound complications requiring operative irrigation and debridement, deep vein thrombosis, deformity recurrence, and revision surgery. No differences were detected in SF-36, LSA, and VAS surveys administered at baseline, 6 months postoperation, or 1 year postoperation between younger and older patients.

CONCLUSION

Despite more comorbidities, older patients had similar rates of postoperative complications and reported similar outcomes for pain, function, and mobility following MTP fusion. These findings support increased implementation of MTP fusion surgery for older patients with hallux rigidus.

LEVEL OF EVIDENCE

Level II, prospective cohort.

Comorbidities Associated With Poor Outcomes Following Operative Hammertoe Correction in a Geriatric Population

Background:

Although complications following hammertoe correction surgery are rare, older patients with comorbid conditions are often considered poorer operative candidates compared with younger, healthier patients because of a suspected increased risk of adverse outcomes. The aim of this study was to determine if the presence of multiple comorbidities was associated with increased complications or unsuccessful patient-reported outcomes following operative hammertoe correction in geriatric patients.

Methods:

Prospectively collected data was reviewed on 78 patients aged 60 years or older who underwent operative correction of hammertoe deformity. Patient demographics, comorbidities, and postoperative complications were recorded. Patient-reported outcomes were assessed using preoperative and postoperative visual analog scale for pain and Short Form Health Survey Physical and Mental Component Summary with 1 year of follow-up. Patients were divided into 2 groups based on number of comorbidities (0 or 1 vs > 2) and then compared. The average age of patients was 69.4 years and the prevalence of comorbidities in the study population was as follows: 11.5% smokers, 25.6% on blood thinners, 15.4% with rheumatoid arthritis, 7.7% with diabetes mellitus, 2.6% with peripheral arterial disease, 6.4% with chronic obstructive pulmonary disease, 11.5% with coronary artery disease, and 23.1% with osteoporosis.

Results:

Fifty-three patients (67.9%) had 0 or 1 comorbidity and 25 (32.1%) had 2 or more comorbidities. Compared to the 0 or 1 comorbidity group, the presence of multiple comorbidities was associated with an adjusted odds ratio (OR) for superficial wound infection of 4.18 (P = .045) and deformity recurrence requiring surgery OR of 23.15 (P = .032). Patient-reported outcomes were similar between comorbidity groups.

Conclusions:

This study further informs foot and ankle specialists to maintain increased surveillance for postoperative complications and unsuccessful outcomes in patients with multiple comorbidities. Although geriatric patients still report significant improvements in both pain and function, patients with underlying medical conditions should be counseled about their increased risks when pursuing operative hammertoe correction.

Level of Evidence:

Level III, retrospective comparative series.

Effect of Psychotropic Medications on Hammertoe Reconstruction Outcomes

Background:

Hammertoe deformities can seriously affect activity level and footwear. The use of prescription, mood-altering medications is very common, with some estimates as high as 25% of the population. Mood disorders, especially depression, negatively affect the results of medical and operative treatments. This study assessed the relationship of mood-altering medication use with the outcomes and complications of operative reconstruction of hammertoes.

Methods:

Data were prospectively collected from 116 patients who underwent hammertoe reconstruction, including demographic information, medical history, the use of mood-altering psychotropic medications (antidepressants, anxiolytics, hypnotics, and mood stabilizers), and postoperative complications. Preoperative patient-reported outcomes were measured using the visual analog scale (VAS) for pain and Short Form Health Survey (SF-36), which were repeated at 1-year follow-up.

Results:

A total of 36.2% of patients were taking psychotropic medications. Medication and nonmedication groups had similar pain VAS and SF-36 Physical Component Summary (PCS) scores before and after surgery. Compared with nonmedication patients, patients on psychotropic medications had significantly lower SF-36 Mental Component Summary (MCS) scores preoperatively (P = .001) and postoperatively (P = .006), but no significant difference in the change in MCS (ΔMCS) from preoperative to postoperative. Psychotropic medication use was associated with superficial wound infections (P = .048), but not other complications.

Conclusions:

Patients taking psychotropic medications were equally likely to benefit from forefoot reconstruction as nonmedication patients. Preoperative and postoperative PCS and VAS were not significantly different between medication and nonmedication groups. Although the medication group had lower absolute MCS, they reported the same magnitude of improvement in MCS (ΔMCS) as the nonmedication group.

Level of Evidence:

Level II, prospective cohort study.

Outcomes Following Shock Aortic Valve Replacement: Transcatheter Versus Surgical Approaches

OBJECTIVES

To compare transcatheter aortic valve replacement (TAVR) with surgical aortic valve replacement (SAVR) for patients in shock.

BACKGROUND

There are minimal data on the clinical and echocardiographic outcomes for patients in shock that undergo TAVR and no data comparing these outcomes to similar patients undergoing SAVR.

METHODS

This is a single center, retrospective cohort study of patients having Society of Thoracic Surgeons (STS)-defined urgent or emergent AVR for aortic stenosis with clinical signs and symptoms of shock. Inclusion criteria were based on the Society of Cardiovascular Angiography & Interventions (SCAI) shock consensus statement and included: the need for inotropic or vasopressor agents, mechanical ventilation, continuous renal replacement therapy or newly initiated hemodialysis, and/or utilization of mechanical hemodynamic support. Clinical and echocardiographic outcomes for TAVR and SAVR were compared.

RESULTS

Thirty-seven patients met the inclusion criteria for this study (17 TAVR, 20 SAVR). TAVR patients had a higher STS Predicted Risk of Mortality (PROM) score of 22.3% compared to 11.8% for SAVR patients (p = 0.001). No significant differences were found in baseline echocardiographic results. TAVR procedures required less procedure room time (185.9 min TAVR, 348.5 min SAVR, p < 0.001) and fewer intraoperative packed red blood cell (pRBC) transfusions (0.2 units TAVR, 3.4 units SAVR, p < 0.001). TAVR patients also had lower rates of prolonged postoperative ventilation compared to SAVR patients (38.5% TAVR, 75.0% SAVR, p = 0.047). TAVR and SAVR had similar rates of mortality at discharge (2 TAVR, 1 SAVR, p = 0.584), 30-days (2 TAVR, 1 SAVR, p = 0.584), and 1-year (8 TAVR, 5 SAVR, p = 0.149).

CONCLUSIONS

Despite a higher risk TAVR group, patients in shock undergoing either TAVR or SAVR have similar 30-day mortality. At one year, SAVR patients have a numerically better, though not statistically significant, survival. These findings support the use of TAVR for patients in shock with aortic stenosis.