Article ID: CJ-23-0798
Background: Detection of left ventricular (LV) abnormalities is essential for patients with preclinical heart failure (HF) to delay progression to clinical HF. Global longitudinal strain (GLS) is a sensitive marker for the early occurrence of subtle abnormalities in LV function, but not all echocardiographic instruments can measure GLS.
Methods and Results: We studied 853 preclinical HF patients to devise a scoring system for predicting low GLS (<16%). The associations of medical history and echocardiographic parameters with low GLS were evaluated using Cox proportional hazards analysis. Model 1 of the system consisted of medical history; for Model 2, conventional echocardiographic parameters were added to Model 1. For Model 1, a score ≥5 points meant prediction of low GLS with 90.2% sensitivity and 62.9% specificity (male=1 point, hypertension=4 points, dyslipidemia=1 point, atrial fibrillation=2 points, history of cardiac surgery=2 points). For Model 2, a score ≥4 points denotes prediction of low GLS with 80.3% sensitivity and 76.5% specificity (male=1 point, hypertension=2 points, atrial fibrillation=2 points, LV mass index >116 g/m2 [male] or >96 g/m2 [female]=1 point, LV ejection fraction <59%=2 points, E/e′ >14=1 point).
Conclusions: Our scoring system provides an easy-to-use evaluation of LV longitudinal myocardial dysfunction, and may prove useful for risk stratification of patients with preclinical HF.
Heart failure (HF) is classified as Stages A–D based on structural changes and symptoms, and this HF classification emphasizes the development and progression of the disease and can be used to describe individuals and populations. Because HF is considered a progressive disorder that can be represented as a clinical continuum, individuals at a particular HF stage require specific management with the long-term goal of avoiding HF progression. Stage A HF patients have been characterized as at high risk of HF even though there is no evidence of structural heart disease or symptoms of HF, whereas Stage B HF refers to structural heart disease without signs or symptoms of HF. Stages A and B HF comprise patients without clinical symptoms, such that these 2 stages are both referred to as “preclinical HF”.
The absolute mortality rate for HF remains approximately 50% within 5 years of diagnosis, although survival has improved.1 A population cohort study reported that 5-year survival rates for Stages A and B HF were as high as 97% and 96%, respectively, whereas they were as low as 60–78% for Stage C HF and extremely low, as low as 20%, for Stage D HF.2 Thus, early detection of left ventricular (LV) structural abnormalities is essential for patients with preclinical HF to delay progression to clinical HF, although the assessment of such dysfunction can be challenging. Although echocardiography plays a pivotal role in the quantification and early detection of anomalies in LV structural findings, it has been reported that 2-dimensional speckle-tracking echocardiographic parameters are also useful for the detection of the early occurrence of LV structural abnormalities. Global longitudinal strain (GLS) in particular is reportedly a sensitive marker for early occurrence of subtle abnormalities of LV myocardial performance, which is helpful for the prediction of outcomes for various cardiac diseases, and superior to conventional echocardiographic parameters. It has been reported that GLS <16% is a likely indicator of significant myocardial dysfunction regardless of the vendor of the detection system or clinical covariate lines.3,4 In fact, the 2022 American Heart Association (AHA) / American College of Cardiology (ACC) / Heart Failure Society of America (HFSA) guideline for the management of HF stated that GLS <16% should be considered an LV structural abnormality regardless of whether LV ejection fraction (LVEF) is preserved.5 Moreover, GLS <16% is defined by the Heart Failure Association of the European Society of Cardiology as one of the minor criteria for HF with preserved ejection fraction.6
Although the measurement of GLS has become simpler and easier with advances in instrumentation, not all echocardiographic instruments can measure GLS. Moreover, it is still difficult to predict low GLS for patients with preclinical HF in routine clinical practice, especially in the outpatient clinic. We therefore designed and conducted a single-center retrospective cohort study to create a scoring system for predicting the occurrence of low GLS (<16%) in patients with preclinical HF (i.e., Stage A and B HF) in routine clinical practice.
In all, 1,463 consecutive patients who had undergone echocardiography at the Okamoto Cardiology Clinic between May 2020 and January 2023 were retrospectively enrolled in this study (Supplementary Figure). Of these patients, 540 were excluded because of LVEF <50%, 44 were excluded because of poor echocardiographic images and 22 were excluded because of Stage C or D HF. Therefore, the final study cohort consisted of 853 patients with Stage A or B HF and LVEF ≥50%.
This study was approved by the Institutional Review Board of Kobe University Hospital Clinical & Translational Research Center (Reference no. B230030) and conformed with the Declaration of Helsinki.
Echocardiographic ExaminationEchocardiographic examinations were performed in all patients using a commercially available echocardiography system (EPIQ System; Philips Medical Systems, Andover, MA, USA). Standard echocardiographic measurements were obtained in accordance with the current guidelines of the American Society of Echocardiography.7 Dedicated software (AutoSTRAIN; Philips Medical Systems) was used for 2-dimensional speckle-tracking strain analysis to evaluate each patient’s GLS. Briefly, apical 4-, 2- and long-axis views, obtained as Digital Imaging and Communications in Medicine-formatted file images, were uploaded onto a personal computer for subsequent off-line GLS analysis, which was expressed as an absolute value in accordance with the aforementioned current guideline.7 For patients with atrial fibrillation, GLS values were the mean of measurements made for ≥3 cardiac cycles.
Definition of ComorbiditiesComorbidities in this study were defined as follows. Hypertension was defined as systolic blood pressure (SBP) >140 mmHg or diastolic blood pressure >90 mmHg or being treated with antihypertensive drugs. Diabetes was defined as HbA1c >6.5% or fasting glucose >126 mg/dL, with or without the use of antidiabetic drugs. Atrial fibrillation was defined as an irregular rhythm with a fluctuating baseline and no discernible P waves by means of 12-lead electrocardiogram.
Definition of Preclinical HFDefinition of preclinical HF was based on the 2022 AHA/ACC/HFSA guideline for the management of HF.5 Stage A HF was defined as at risk for HF and includes hypertension, atherosclerotic cardiovascular disease, diabetes, metabolic syndrome and obesity, exposure to cardiotoxic agents, genetic variant for cardiomyopathy, or a positive family history of cardiomyopathy but without symptoms, structural heart disease, or cardiac biomarkers of stretch or injury. Stage B HF was defined as no symptoms or signs of HF and evidence of 1 of the following: LV wall thickness ≥12 mm, relative wall thickness >0.42, LV mass index (LVMI) >116 g/m2 in men and >95 g/m2 in women, GLS <16%, LV wall motion abnormalities, left atrial volume index ≥29 mL/m2, mean early transmitral flow velocity (E)/early diastolic mitral annular velocity (e′) ratio >15, septal e′ <7 cm/s, lateral e′ <10 cm/s, tricuspid regurgitation velocity >2.8 m/s, estimated pulmonary artery systolic pressure >35 mmHg, and more than moderate valvular heart disease.
Statistical AnalysisContinuous variables are expressed as the mean±SD for normally distributed data and as the median with interquartile range (IQR) for non-normally distributed data. Categorical variables are expressed as frequencies and percentages. The parameters of the 2 subgroups were compared using Student’s t-test or the Mann-Whitney U test depending on data distribution. Proportional differences were evaluated using Fisher’s exact test.
Associations of clinical parameters with low GLS were analyzed by means of linear logistic regression models for univariate and multivariate analyses. In the selection of independent variables for entry into the multivariate model, Pearson’s correlation analyses between independent variables were performed in advance to avoid multicollinearity. If ≥2 variables were used to measure a pathophysiological parameter (e.g., LVEF and LV end-systolic volume as markers of LV ejection performance), the more or most clinically relevant parameter was entered into the model. Variables with a univariate value of P<0.05 were incorporated into the multivariate analysis. In addition, a scoring system for predicting low GLS was created and adjustment factor points were assigned in order to correlate the score with the odds ratio determined by the logistic regression models for multivariate analyses. The calibration of the model was evaluated using receiver operating characteristic (ROC) curve analysis. ROC curve analysis was also used to determine the optimal cut-off value for predicting low GLS for SBP, LVEF, and body mass index. For all steps, P<0.05 was considered statistically significant.
All analyses were performed using commercially available software (MedCalc software version 19.0.7; MedCalc Software, Mariakerke, Belgium).
The baseline characteristics of the 853 patients are summarized in Table 1; 61 patients (7.2%) had GLS <16%. Baseline clinical characteristics of patients with GLS ≥16% and <16% are shown separately in Table 1. Compared with patients with GLS ≥16%, those with GLS <16% were more likely to be older (71.9±13.0 vs. 59.7±19.0 years; P<0.01), male (67.2% vs. 45.7%; P<0.01), have higher SBP (137±20 vs. 131±22 mmHg; P=0.02) and diastolic blood pressure (75±14 vs. 69±14 mmHg; P<0.01), with a greater prevalence of hypertension (93.4% vs. 51.2%; P<0.01), diabetes (24.6% vs. 9.7%; P<0.01), dyslipidemia (54.1% vs. 28.7%; P<0.01), atrial fibrillation (31.1% vs. 6.8%; P<0.01), ischemic heart disease (18.0% vs. 5.4%; P<0.01), and a previous history of cardiac surgery (11.5% vs. 1.9%; P<0.01).
Baseline Characteristics of Patients
| All patients (n=853) |
Patients with GLS ≥16% (n=792) |
Patients with GLS <16% (n=61) |
P value (GLS ≥16% vs. <16%) |
|
|---|---|---|---|---|
| Clinical characteristics | ||||
| Age (years) | 60.6±18.9 | 59.7±19.0 | 71.9±13.0 | <0.01 |
| Male sex | 403 (47.2) | 362 (45.7) | 41 (67.2) | <0.01 |
| BMI (kg/m2) | 23.8±9.1 | 23.7±9.3 | 24.6±4.1 | 0.47 |
| Previous history of cardiac surgery | 22 (2.6) | 15 (1.9) | 7 (11.5) | <0.01 |
| Stages of HF | ||||
| A | 162 (19.0) | 162 (20.5) | 0 (0) | <0.01 |
| B | 691 (81.0) | 630 (79.5) | 61 (100) | <0.01 |
| Hemodynamics | ||||
| SBP (mmHg) | 131±21.9 | 131±22 | 137±20 | 0.02 |
| DBP (mmHg) | 70±14 | 69±14 | 75±14 | <0.01 |
| Heart rate (beats/min) | 76.4±12.0 | 76.2±12.1 | 78.5±10.8 | 0.16 |
| Comorbidities | ||||
| Hypertension | 463 (54.3) | 406 (51.2) | 57 (93.4) | <0.01 |
| Diabetes | 92 (10.8) | 77 (9.7) | 15 (24.6) | <0.01 |
| Dyslipidemia | 260 (30.5) | 227 (28.7) | 33 (54.1) | <0.01 |
| Atrial fibrillation | 73 (8.6) | 54 (6.8) | 19 (31.1) | <0.01 |
| Ischemic heart disease | 54 (6.3) | 43 (5.4) | 11 (18.0) | <0.01 |
| Valvular disease | 22 (2.6) | 15 (1.9) | 7 (11.5) | <0.01 |
| Echocardiographic data | ||||
| LV end-diastolic diameter (mm) | 44.1±5.0 | 44.1±4.9 | 45.2±5.5 | 0.10 |
| LV end-systolic diameter (mm) | 28.5±4.8 | 28.3±4.7 | 31.1±4.8 | <0.01 |
| IVST (mm) | 8.2±2.0 | 8.0±2.0 | 10.0±2.0 | <0.01 |
| Posterior wall thickness (mm) | 8.5±1.7 | 8.4±1.7 | 10.0±1.9 | <0.01 |
| Relative wall thickness | 0.39±0.09 | 0.39±0.09 | 0.45±0.12 | <0.01 |
| LVMI (g/m2) | ||||
| Men | 79.5±21.1 | 77.8±20.1 | 93.8±23.7 | <0.01 |
| Women | 67.3±20.9 | 66.1±19.7 | 94.7±27.1 | <0.01 |
| LAVI (mL/m2) | 30.1±11.8 | 29.3±10.5 | 40.7±20.0 | <0.01 |
| LV end-diastolic volume (mL) | 74.9±23.3 | 75.1±22.9 | 71.8±27.9 | 0.29 |
| LV end-systolic volume (mL) | 29.1±10.3 | 29.0±10.0 | 30.9±13.4 | 0.15 |
| LVEF (%) | 61.3±4.9 | 61.6±4.8 | 57.2±4.6 | <0.01 |
| GLS (%) | 20.3±2.9 | 20.7±2.4 | 14.3±1.6 | <0.01 |
| e´ (cm/s) | ||||
| Septal | 6.8±2.6 | 6.9±2.6 | 5.2±1.7 | <0.01 |
| Lateral | 8.8±3.5 | 8.9±3.5 | 7.1±2.5 | <0.01 |
| E/e´ (mean) | 9.8±4.1 | 9.6±3.8 | 12.8±6.5 | <0.01 |
| TRV (m/s) | 1.9±0.5 | 1.9±0.5 | 2.0±0.6 | 0.71 |
| Estimated PA systolic pressure (mmHg) | 18.9±8.2 | 18.9±8.1 | 19.5±8.7 | 0.61 |
| Valvular heart disease | 75 (8.8) | 63 (8.0) | 12 (20.0) | <0.01 |
Unless indicated otherwise, data are given as the mean±SD for normally distributed continuous data, the median [interquartile range] for non-normally distributed continuous data, or as n (%) for categorical data. BMI, body mass index; DBP, diastolic blood pressure; e´, early diastolic mitral annular velocity; E, early transmitral flow velocity; GLS, global longitudinal strain; HF, heart failure; IVST, interventricular septum thickness; LAVI, left atrial volume index; LV, left ventricular; LVEF, left ventricular ejection fraction; LVMI, left ventricular mass index; PA, pulmonary artery; SBP, systolic blood pressure; TRV, tricuspid regurgitation velocity.
In terms of echocardiographic parameters, patients with GLS <16% had a higher interventricular septum thickness (10.0±2.0 vs. 8.0±2.0 mm; P<0.01), higher posterior wall thickness (10.0±1.9 vs. 8.4±1.7 mm; P<0.01), higher relative wall thickness (0.45±0.12 vs. 0.39±0.09; P<0.01), higher LVMI in both men (93.8±23.7 vs. 77.8±20.1 g/m2; P<0.01) and women (94.7±27.1 vs. 66.1±19.7 g/m2; P<0.01), higher left atrial volume index (40.7±20.0 vs. 29.3±10.5 mL/m2; P<0.01), lower LVEF (57.2±4.6% vs. 61.6±4.8%; P<0.01), lower GLS (14.3±1.6% vs. 20.7±2.4%; P<0.01), lower e′ (septal: 5.2±1.7 cm/s vs. 6.9±2.6 cm/s [P<0.01]; lateral: 7.1±2.5 vs. 8.9±3.5 cm/s [P<0.01]), higher mean E/e′ (12.8±6.5 vs. 9.6±3.8; P<0.01), and a higher prevalence of valvular heart disease (20.0% vs. 8.0%; P<0.01) than patients with GLS ≥16%.
Associations of Clinical Parameters With Low GLSWe investigated the associations of clinical parameters with low GLS using linear logistic regression models for univariate and multivariate analyses and created 2 models for predicting low GLS. Model 1 comprised medical history (Table 2), whereas Model 2 included echocardiographic parameters in addition to medical history (Table 3). The validity of these 2 models was determined using ROC curve analysis (Figure 1). The areas under the curve (AUC) of Models 1 and 2 were 0.83 (95% confidential interval [CI] 0.80–0.85) and 0.88 (95% CI 0.86–0.91), respectively. Thus, Model 2 was shown statistically to be significantly more accurate for predicting low GLS than Model 1 (P<0.01).
Univariate and Multivariate Cox Proportional Hazards Analysis for Predicting Low GLS: Model 1 (Medical History)
| Covariate | Univariate | Multivariate | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |
| Age | 1.05 | 1.03–1.07 | <0.01 | |||
| Male sex | 2.44 | 1.40–4.23 | <0.01 | 2.01 | 1.12–3.61 | 0.02 |
| BMI | 1.00 | 0.99–1.02 | 0.50 | |||
| SBP | 1.01 | 1.00–1.03 | 0.02 | |||
| Hypertension | 13.5 | 4.87–37.7 | <0.01 | 8.52 | 2.99–24.3 | <0.01 |
| Diabetes | 3.03 | 1.62–5.68 | <0.01 | |||
| Dyslipidemia | 2.93 | 1.73–4.97 | <0.01 | 1.87 | 1.06–3.29 | 0.03 |
| Atrial fibrillation | 6.18 | 3.37–11.4 | <0.01 | 4.19 | 2.20–8.01 | <0.01 |
| Ischemic heart disease | 3.83 | 1.86–7.88 | <0.01 | |||
| Previous history of cardiac surgery | 6.71 | 2.63–17.2 | <0.01 | 3.84 | 1.38–10.7 | <0.01 |
CI, confidential interval; OR, odds ratio. Other abbreviations as in Table 1.
Univariate and Multivariate Cox Proportional Hazards Analysis for Predicting Low GLS: Model 2 (Medical History + Echocardiographic Parameters)
| Covariate | Univariate | Multivariate | ||||
|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | |
| Age | 1.05 | 1.03–1.07 | <0.01 | |||
| Male sex | 2.44 | 1.40–4.23 | <0.01 | |||
| BMI | 1.00 | 0.99–1.02 | 0.50 | |||
| SBP | 1.01 | 1.00–1.03 | 0.02 | |||
| Hypertension | 13.5 | 4.87–37.7 | <0.01 | |||
| Diabetes | 3.03 | 1.62–5.68 | <0.01 | |||
| Dyslipidemia | 2.93 | 1.73–4.97 | <0.01 | |||
| Atrial fibrillation | 6.18 | 3.37–11.4 | <0.01 | 5.42 | 2.53–11.6 | <0.01 |
| Ischemic heart disease | 3.83 | 1.86–7.88 | <0.01 | |||
| Previous history of cardiac surgery | 6.71 | 2.63–17.2 | <0.01 | 4.42 | 1.31–14.9 | 0.02 |
| Relative wall thickness (per 0.1 increment) | 2.14 | 1.78–2.58 | <0.01 | 2.09 | 1.65–2.64 | <0.01 |
| LVMI | 1.04 | 1.03–1.05 | <0.01 | 1.02 | 1.01–1.04 | <0.01 |
| LAVI | 1.05 | 1.04–1.07 | <0.01 | |||
| LVEF | 0.82 | 0.78–0.87 | <0.01 | 0.87 | 0.81–0.94 | <0.01 |
| e´ (septal) | 0.70 | 0.61–0.82 | <0.01 | |||
| e´ (lateral) | 0.82 | 0.74–0.90 | <0.01 | |||
| E/e´ (mean) | 1.13 | 1.08–1.18 | <0.01 | |||
| Valvular heart disease | 2.84 | 1.43–5.60 | <0.01 | |||
Abbreviations as in Tables 1,2.
Receiver operating characteristic curve analysis of Model 1 (medical history; red line) and Model 2 (medical history + echocardiographic parameters; blue line). AUC, area under the curve; CI, confidence interval.
Scoring for the Prediction of Low GLS
A total of 10 adjustment factor points were assigned for Model 1 and 9 points were assigned for Model 2 for correlation of the score with the odds ratio determined by the logistic regression models for multivariate analyses (Table 4 for Model 1; Table 5 for Model 2). Optimal cut-off values of SBP >130 mmHg, LVEF <59%, and body mass index >25 kg/m2 for predicting low GLS were identified using ROC curve analysis. The AUC of Model 1 was 0.82 (95% CI 0.79–0.84), and the optimal cut-off value for predicting low GLS was ≥5 points, with a sensitivity of 90.2% and specificity of 62.9% (Figure 2A). The AUC of Model 2 was 0.86 (95% CI 0.83–0.88), and the optimal cut-off value for predicting low GLS was ≥4 points, with a sensitivity of 80.3% and specificity of 76.5% (Figure 2B). Figure 3 summarizes the results of Model 1 and Model 2 for an easy-to-use evaluation of low GLS.
Univariate and Multivariate Cox Proportional Hazards Analysis for Predicting Low GLS: Model 1 (Medical History With Categorical Variable)
| Covariate | Univariate | Multivariate | No. adjustment factor points |
||||
|---|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | ||
| Age >70 years | 3.62 | 2.09–6.26 | <0.01 | ||||
| Male sex | 2.44 | 1.40–4.23 | <0.01 | 2.01 | 1.12–3.61 | 0.02 | 1 |
| BMI >25 kg/m2 | 1.58 | 0.93–2.69 | 0.09 | ||||
| SBP >130 mmHg | 2.01 | 1.17–3.45 | 0.01 | ||||
| Hypertension | 13.5 | 4.87–37.7 | <0.01 | 8.52 | 2.99–24.3 | <0.01 | 4 |
| Diabetes | 3.03 | 1.62–5.68 | <0.01 | ||||
| Dyslipidemia | 2.93 | 1.73–4.97 | <0.01 | 1.87 | 1.06–3.29 | 0.03 | 1 |
| Atrial fibrillation | 6.18 | 3.37–11.4 | <0.01 | 4.19 | 2.20–8.01 | <0.01 | 2 |
| Ischemic heart disease | 3.83 | 1.86–7.88 | <0.01 | ||||
| Previous history of cardiac surgery | 6.71 | 2.63–17.2 | <0.01 | 3.84 | 1.38–10.7 | <0.01 | 2 |
| Total score | 10 | ||||||
Abbreviations as in Tables 1,2.
Univariate and Multivariate Cox Proportional Hazards Analysis for Predicting Low GLS: Model 2 (Medical History + Echocardiographic Parameters With Categorical Variable)
| Covariate | Univariate | Multivariate | No. adjustment factor points |
||||
|---|---|---|---|---|---|---|---|
| OR | 95% CI | P value | OR | 95% CI | P value | ||
| Age >70 years | 3.62 | 2.09–6.26 | <0.01 | ||||
| Male sex | 2.44 | 1.40–4.23 | <0.01 | 3.07 | 1.56–6.06 | <0.01 | 1 |
| BMI >25 kg/m2 | 1.58 | 0.93–2.69 | 0.09 | ||||
| SBP >130 mmHg | 2.01 | 1.17–3.45 | 0.01 | ||||
| Hypertension | 13.5 | 4.87–37.7 | <0.01 | 5.70 | 1.95–16.7 | <0.01 | 2 |
| Diabetes | 3.03 | 1.62–5.68 | <0.01 | ||||
| Dyslipidemia | 2.93 | 1.73–4.97 | <0.01 | ||||
| Atrial fibrillation | 6.18 | 3.37–11.4 | <0.01 | 4.32 | 2.17–8.61 | <0.01 | 2 |
| Ischemic heart disease | 3.83 | 1.86–7.88 | <0.01 | ||||
| Previous history of cardiac surgery | 6.71 | 2.63–17.2 | <0.01 | ||||
| Relative wall thickness >0.42 | 2.74 | 1.61–4.68 | <0.01 | ||||
| LVMI >116 g/m2 (men) / 95 g/m2 (women) | 5.06 | 2.68–9.56 | <0.01 | 3.27 | 1.55–6.88 | <0.01 | 1 |
| LAVI >34 mL/m2 | 3.65 | 2.14–6.23 | <0.01 | ||||
| LVEF <59% | 5.57 | 3.19–9.73 | <0.01 | 4.07 | 2.20–7.54 | <0.01 | 2 |
| e´ (septal) <7 cm/s | 3.91 | 1.90–8.07 | <0.01 | ||||
| e´ (lateral) <10 cm/s | 3.34 | 1.49–7.44 | <0.01 | ||||
| E/e´ (average) >14 | 4.10 | 2.27–7.41 | <0.01 | 2.60 | 1.28–5.30 | <0.01 | 1 |
| Valvular heart disease | 2.84 | 1.43–5.60 | <0.01 | ||||
| Total score | 9 | ||||||
Abbreviations as in Tables 1,2.
Receiver operating characteristic curve analysis of (A) Model 1 (medical history with categorical variables) and (B) Model 2 (medical history + echocardiographic parameters with categorical variables). AUC, area under the curve; CI, confidence interval.
Summary of our scoring system for predicting low global longitudinal strain (GLS). AUC, area under the curve; LVEF, left ventricular ejection fraction; LVMI, left ventricular mass index.
In our study, which enrolled preclinical HF patients, we developed an easy-to-use scoring system for predicting LV longitudinal myocardial dysfunction (GLS <16%).
Utility of GLS for Patients With Preclinical HFThe early detection of subtle LV structural abnormality warrants closer attention because of the growing interest in addressing preclinical HF as the best means of preventing eventual progression to clinical HF. Moreover, it was reported that 56% of adults aged ≥45 years were classified as having preclinical HF, which was also identified as an increased risk of progression to clinical HF and mortality.2 Young et al evaluated the rate of progression of preclinical HF in a study comprising 413 patients with Stage A HF and 413 patients with Stage B HF over a follow-up period of 4 years.8 Among patients with Stage A HF, 34% progressed to Stage B HF and 1% progressed to clinical HF. Among patients with Stage B HF, 6% progressed to clinical HF.8
It has been reported that GLS is a more accurate and sensitive parameter than conventional cardiac functional parameters even for all stages of HF, and abnormal GLS has been reported to be associated with the occurrence of cardiovascular events.9–14 However, the characteristics of the GLS of patients with preclinical HF with few abnormalities can best be demonstrated with conventional cardiac functional parameters. GLS <16% is currently considered to be abnormal, indicating significant myocardial dysfunction regardless of vendor or clinical covariate line.3,4 In addition, current guidelines specify that GLS <16% should be considered an LV structural abnormality.5,6,15 The aim of using GLS for patients with preclinical HF is to aid in the prediction of subclinical LV dysfunction, and it can represent an early marker of LV dysfunction, possibly indicating cardiovascular morbidity and mortality, and thus help avoid progression to clinical HF. Therefore, numerous investigators have confirmed the efficacy of GLS for predicting subclinical LV dysfunction and its association with cardiovascular events in patients with preclinical HF. Modin et al showed that patients who had essential hypertension and GLS <17.2% experienced more cardiovascular events than those who did not.13 Holland et al reported that type 2 diabetes patients with preserved LVEF and GLS <18.9% had significantly worse outcomes than those with a higher percentage during a 10-year follow-up and concluded that GLS was independently associated with the primary endpoint.16 Patients without HF symptoms or LV structural abnormalities but with a history of using cardiotoxins, such as doxorubicin and Trastuzumab, are also considered to be Stage A HF. Hatazawa et al reported that during a long-term follow-up, baseline GLS was identified as the only independent predictor of doxorubicin-related cancer therapeutics-related cardiac dysfunction for patients with malignant lymphoma, and that baseline GLS ≤19% was associated with the development of doxorubicin-related cancer therapeutics-related cardiac dysfunction and hospitalization for HF.14 In addition, Wang et al showed that GLS <16% was associated with a higher risk of development of clinical HF for 290 patients with Stage B HF.12 Although several conventional echocardiographic parameters or biomarkers based on the 2022 AHA/ACC/HFSA guideline for the management of HF had suggested thresholds for structural heart disease, no studies have investigated whether these factors can predict GLS <16%.
How to Use Our Scoring System for Predicting Low GLS in Patients With Preclinical HFModel 1 (Medical History) Although echocardiography plays a pivotal role in the quantification of LV morphology and function, the current situation is that echocardiography cannot be used for all patients in actual clinical practice, especially in outpatient clinics, so that this procedure is actually performed for relatively few select patients. However, it is often difficult to identify high-risk HF patients only on the basis of their medical history. However, the results of the present study suggest that low GLS can be predicted for patients with a score of ≥5 points for Model 1 (medical history only) with a sensitivity of 90.2% and specificity of 62.9%. In addition, the presence of hypertension was determined as high as 4 points in Model 1. It is well known that hypertension is the most important risk factor for the development of HF, and LV myocardial damage may be caused by long-term elevated blood pressure. Thus, hypertension is an important surrogate marker for predicting low GLS, and treatment of hypertension is one element of the holistic management of patients with HF. Although Model 1 is less accurate than Model 2, to which parameters of conventional echocardiography have been added, as described below, it may be useful as a screening tool for the stratification of high-risk patients when echocardiography is not available. If the patients score ≥5 points for Model 1, close follow-up and strict lifestyle control should be implemented, with echocardiography performed as soon as possible.
Model 2 (Medical History + Echocardiographic Parameters) In our study, the addition of simple conventional echocardiographic parameters, including LVMI, LVEF and E/e′, to a patient’s medical history increased the accuracy for predicting low GLS. For patients with a score of ≥4 points, low GLS can be predicted with a sensitivity of 80.3% and specificity of 76.5%. If conventional echocardiography is available in clinical practice, this scoring system may be preferable. A patient with a Model 2 score of ≥4 points should be identified as a high-risk HF patient. Finally, measuring GLS at a suitably equipped institution is also recommended.
Study LimitationsThis study is a single-center retrospective study, so that future prospective studies with larger patient populations from several centers will be needed to validate our findings. Moreover, B-type natriuretic peptide is one of the elements to judge preclinical HF according to the 2022 AHA/ACC/HFSA guideline for the management of HF,5 but we did not measure B-type natriuretic peptide in a sufficient number of patients.
Our scoring system provides an easy-to-use evaluation of LV longitudinal myocardial dysfunction, and may prove useful for risk stratification for patients with Stage A and B HF.
H.T. has received remuneration from AstraZeneca plc, Ono Pharmaceutical Co., Ltd., Pfizer Inc., Otsuka Pharmaceutical Co., Ltd., Daiichi Sankyo Co., Ltd., Eli Lilly and Company, Boehringer Ingelheim GmbH, Abbott Japan LLC, and Novartis International AG. K.H. has received research funding from Daiichi Sankyo Co., Ltd., Actelion Pharmaceuticals Japan, Terumo Corporation, Abbott Vascular Japan, Otsuka Pharmaceutical Co., Ltd., Kowa Co., Ltd., Takeda Pharmaceutical Company Limited, Nihon Medi-Physics Co., Ltd., Novartis Pharma Co., Ltd., Bayer Co., Ltd., Biotronic Japan Co., Ltd., FUJIFILM Toyama Chemical Co., Ltd., Medtronic Japan Co., Ltd., and Sysmex Co., Ltd., and is a member of Circulation Journal’s Editorial Team. The remaining authors have no conflicts of interest to declare.
This study was approved by the Institutional Review Board of Kobe University Hospital Clinical & Translational Research Center (Reference no. B230030).
Please find supplementary file(s);
https://doi.org/10.1253/circj.CJ-23-0798
