Development and validation of a clinical score to identify echocardiographic left ventricular hypertrophy in patients with cardiovascular disease

Artificial Intelligence-Enabled Model for Early Detection of Left Ventricular Hypertrophy and Mortality Prediction in Young to Middle-Aged Adults

BACKGROUND

Concealed left ventricular hypertrophy (LVH) is a prevalent condition that is correlated with a substantial risk of cardiovascular events and mortality, especially in young to middle-aged adults. Early identification of LVH is warranted. In this work, we aimed to develop an artificial intelligence (AI)-enabled model for early detection and risk stratification of LVH using 12-lead ECGs.

METHODS

By deep learning techniques on the ECG recordings from 28 745 patients (20-60 years old), the AI model was developed to detect verified LVH from transthoracic echocardiography and evaluated on an independent cohort. Two hundred twenty-five patients from Japan were externally validated. Cardiologists' diagnosis of LVH was based on conventional ECG criteria. The area under the curve (AUC), sensitivity, and specificity were applied to evaluate the model performance. A Cox regression model estimated the independent effects of AI-predicted LVH on cardiovascular or all-cause death.

RESULTS

The AUC of the AI model in diagnosing LVH was 0.89 (sensitivity: 90.3%, specificity: 69.3%), which was significantly better than that of the cardiologists' diagnosis (AUC, 0.64). In the second independent cohort, the diagnostic performance of the AI model was consistent (AUC, 0.86; sensitivity: 85.4%, specificity: 67.0%). After a follow-up of 6 years, AI-predicted LVH was independently associated with higher cardiovascular or all-cause mortality (hazard ratio, 1.91 [1.04-3.49] and 1.54 [1.20-1.97], respectively). The predictive power of the AI model for mortality was consistently valid among patients of different ages, sexes, and comorbidities, including hypertension, diabetes, stroke, heart failure, and myocardial infarction. Last, we also validated the model in the international independent cohort from Japan (AUC, 0.83).

CONCLUSIONS

The AI model improved the detection of LVH and mortality prediction in the young to middle-aged population and represented an attractive tool for risk stratification. Early identification by the AI model gives every chance for timely treatment to reverse adverse outcomes.

Poor Renal and Cardiovascular Outcomes in Patients with Biopsy-Proven Diabetic Nephropathy

BACKGROUND

Despite the high mortality of cardiovascular disease (CVD) in diabetic patients with renal injury, few studies have compared cardiovascular characteristics and outcomes between patients with diabetic nephropathy (DN) and non-diabetic renal disease (NDRD).

METHODS

A total of 326 type 2 diabetes mellitus patients with renal biopsy were assigned to DN and NDRD groups. Echocardiography and Doppler ultrasound were performed to evaluate left ventricular hypertrophy (LVH) and peripheral atherosclerosis disease (PAD). Renal and cardiovascular survival rates were compared between the DN and NDRD groups by Kaplan-Meier analysis. Risk factors for renal and cardiovascular events in DN patients were identified by a Cox proportional hazards model.

RESULTS

In total, 179 patients entered the DN group (54.9%) and 147 made up the NDRD group (45.1%). The presence of diabetic retinopathy, family history of diabetes, and dependence on insulin therapy were associated with the presence of DN. DN patients had more CVD with more severe LVH and PAD. Poorer renal (log-rank χ2 = 26.534, p < 0.001) and cardiovascular (log-rank χ2 = 16.257, p < 0.001) prognoses were seen in the DN group. DR (HR 1.539, 95% CI 1.332-1.842), eGFR (HR 0.943, 95% CI 0.919-0.961), and 24-h proteinuria (HR 1.211, 95% CI 1.132-1.387) were identified as risk factors for renal endpoints. Age (HR 1.672, 95% CI 1.487-1.821), HbA1C (HR 1.398, 95% CI 1.197-1.876), and 24-h proteinuria (HR 1.453, 95% CI 1.289-1.672) were associated with cardiovascular endpoints.

CONCLUSION

Patients with DN had more severe CVD along with poorer renal and cardiovascular prognoses than those with NDRD.

Predictive model for the development of hypertensive cardiopathy: A prospective cohort study

INTRODUCTION

Predictive models of cardiovascular conditions are useful tools for risk stratification. The high morbidity and mortality resulting from hypertensive cardiopathy creates a need for the use of tools to predict the risk of cardiovascular disease.

OBJECTIVE

To evaluate the capacity of a model based on risk factors to predict the development of hypertensive cardiopathy after ten years in patients with a diagnosis of essential arterial hypertension.

METHODS

A prospective cohort study was carried out in hypertensive patients cared for at the specialized arterial hypertension physician’s office of the Specialty Policlinic attached to “Carlos Manuel de Céspedes” Hospital, Bayamo Municipality, Granma Province, Cuba, from January 1, 2000 to December 31, 2009. A predictive model was constructed and validated through a process that included the random split of the whole sample in two parts: one for development (parameters estimation) and the other for validation.

RESULTS

The binary regression model adjusted by the “step-by-step backward method,” showed that in step six, 13 variables sufficed to estimate the risk of developing hypertensive cardiopathy. In the estimation sample, the area under the receiver operating characteristic curve obtained for the prediction of hypertensive heart disease was 0.985 (confidence interval: 0.980-0.990; p = <0.0005). In the validation sample the area under the receiver operating characteristic curve was 0.963 (confidence interval: 0.953-0, 0.973, p<0.0005). The calibration of the model was also adequate (p = 0.863).

CONCLUSIONS

The model obtained proved is a clinical and epidemiological surveillance instrument, useful to identify subjects with greater likelihood to acquire hypertensive heart disease, and to stratify their risk in the following ten-year period.

Gender-specific association of decreased estimated glomerular filtration rate and left vertical geometry in the general population from rural Northeast China

Background

Left ventricular hypertrophy (LVH) is common and associated with cardiovascular outcomes among patients with known chronic kidney disease (CKD). However, the link between decreased estimated glomerular filtration rate (eGFR) and left ventricular (LV) geometry remains poorly explored in general population. In this study, we examined the gender-specific association between eGFR and LVH in the general population from rural Northeast China.

Methods

This survey was conducted from July 2012 to August 2013. A total of 10907 participants (5,013 men and 5,894 women) from the rural Northeast China were randomly selected and examined. LV mass index (LVMI) was used to define LVH (LVMI > 46.7 g/m^2.7 in women; > 49.2 g/m^2.7 in men). LV geometry was defined as normal, or with concentric remodeling, eccentric or concentric hypertrophy, according to relative wall thickness (RWT) and LVMI. Mildly decreased eGFR was defined as eGFR ≥ 60 and < 90 ml/min/1.73 m², and moderate-severely decreased eGFR was defined as eGFR < 60 ml/min/1.73 m².

Results

As eGFR decreased, LVH showed a gradual increase in the entire study population. Multivariate analysis revealed a gender-specific relationship between eGFR and LV geometry. Only in men, mildly decreased eGFR was associated with concentric remodeling [odds ratio (OR): =1.58; 95% CI: 1.14–2.20; P < 0.01] and concentric LVH OR = 1.63; 95% CI: 1.15–2.31; P < 0.01). And only in men, moderate-severely decreased eGFR was a risk factor for concentric LVH (OR = 4.56; 95% CI: 2.14–9.73; P < 0.001) after adjusting for confounding factors.

Conclusions

These findings suggested that decreased eGFR was a risk factor for LV geometry in men, and a gender-specific difference should be taken into account in clinical practice.

Left ventricular hypertrophy and arrhythmogenesis

Left ventricular hypertrophy (LVH) poses an independent risk of increased morbidity and mortality, including atrial arrhythmias, ventricular arrhythmias, and sudden cardiac death. The most common causes of LVH are hypertension and valvular heart disease. Electrocardiography and echocardiography are the first steps in the diagnosis and evaluation of therapy in patients with LVH. Cardiac MRI is the gold standard in diagnosis and assessment of response to therapy. Management of LVH should be based on etiology, evidence, and guideline adherence. Timely and optimal management of the underlying cause of LVH results in improvement (regression) of LVH and its related complications.

Diagnostic accuracy of left ventricular hypertrophy in patients with myocardial infarction by computer-assisted electrocardiography (ELECTROPRES)

BACKGROUND

Information is limited about the classification accuracy of electrocardiographic (ECG) criteria for left ventricular hypertrophy (LVH) in the presence of myocardial infarction (MI).

METHODS

We evaluated LVH classification accuracy for a set of 16 ECG criteria and some combinations derived from them in 1642 patients (105 with MI) suspected of coronary heart disease with two-dimensional echocardiography evaluation and a standard 12-lead ECG recorded at the same time. Patients with left bundle branch block had previously been excluded. Measures of classification accuracy included sensitivity, specificity, likelihood ratios, and positive and negative predictive values.

RESULTS

Diagnostic accuracy varied widely for different LVH criteria. The criteria with the best overall performance had highest sensitivity in the presence of MI and sensitivities of approximately 30% with relatively low specificities ranging from 72% to 78%. However, the classification accuracy for them was similar to that for patients without MI. The prevalence of LVH in patients with MI was higher (56%) than in those with no MI (31%). Classification accuracy of the best single previously published LVH criteria was comparable to that of the best combinations of any three of them.

CONCLUSIONS

The classification accuracy of LVH criteria in the presence of MI is comparable to that in patients without MI, in part possibly due to the higher LVH prevalence in the MI group. The presence of a well-validated computer database facilitates comparative evaluation of ECG-LVH criteria and derivation of optimal combinations of criteria for any given clinical application.