Clinical utility of the 2016 ASE/EACVI recommendations for the evaluation of left ventricular diastolic function in the stratification of post-discharge prognosis in patients with acute heart failure

Abstract

Aims

Left ventricular diastolic dysfunction (LVDD) has prognostic significance in heart failure (HF). We aimed to assess the impact of LVDD grade stratified by the updated 2016 echocardiographic algorithm (DD2016) on post-discharge outcomes in patients admitted for acute HF and compare with the previous 2009 algorithm (DD2009).

Methods and results

The study included 481 patients hospitalized for acute decompensated HF. Comprehensive echocardiography and LVDD evaluation were performed just before hospital discharge. The primary endpoint was a composite of cardiovascular death and readmission for HF. The concordance between DD2016 and DD2009 was moderate (κ = 0.44, P < 0.001); the reclassification rate was 39%. During the follow-up (median: 15 months), 127 (26%) patients experienced the primary endpoint. In the Kaplan–Meier analysis, Grade III in DD2016 showed a lower event-free survival rate than Grades I and II (log rank, P < 0.001 and P = 0.048, respectively) and was independently associated with a higher incidence of the primary endpoint than Grade I [hazard ratio 1.89; 95% confidence interval (CI) 1.17–3.04; P = 0.009]. Grade II or III in DD2016, reflecting elevation of left ventricular (LV) filling pressure, added an incremental predictive value of the primary endpoint to clinical variables irrespective of LV ejection fraction. DD2016 was comparable to DD2009 in predicting the endpoint (net reclassification improvement = 11%; 95% CI −7% to 30%, P = 0.23).

Conclusion

Despite simplification of the algorithm for LVDD evaluation, the prognostic value of DD2016 for post-discharge cardiovascular events in HF patients was maintained and not compromised in comparison with DD2009.

left ventricular diastolic dysfunction, echocardiography, heart failure, prognosis

Introduction

Owing to global ageing, the absolute number of individuals newly diagnosed with heart failure (HF) has been progressively increasing.¹ To date, survival after the hospitalization of HF remains quite poor and rates of readmissions for HF does not reduce despite evidence-based treatments.² Therefore, risk stratification of cardiovascular events after hospital discharge has clinical significance in the patients admitted for HF.

The presence and severity of left ventricular (LV) diastolic dysfunction (DD) are associated with prognosis of community-based cohorts or patients with HF irrespective of left ventricular ejection fraction (LVEF).^3–5 The American Society of Echocardiography (ASE) and European Association of Echocardiography (EAE) proposed the algorithm to grade DD severity (DD2009) in their 2009 recommendations.⁶ Although the classification of left ventricular diastolic dysfunction (LVDD) was not primarily intended as a prognostic tool, evaluation of LVDD, and estimation of LV filling pressure based on these recommendations had a certain clinical value in the diagnosis and management of HF. However, its complexity was regarded as a problem to use in clinical practice. These recommendations have been updated by ASE and European Association of Cardiovascular Imaging (EACVI), and the algorithm for DD classification has been revised drastically in 2016 (DD2016).⁷ The validation studies of LV filling pressure between echocardiographic estimation based on DD2016 and invasive measurement revealed the limited accuracy of individual measurements and the necessity and utility of the DD2016 to use multiple echocardiographic parameters in combination.^8,9 However, there are few investigations to explore the prognostic impact of DD2016 in patients with HF.

Accordingly, the present study aimed to examine the prognostic value of classification of LVDD by updated 2016 recommendations and compare with the previous one, through an analysis from the multicentre registry of patients admitted for acute decompensated HF.

Methods

Study design and patient population

This is a post hoc analysis of the Ibaraki Cardiovascular Assessment Study-HF (ICAS-HF) registry, which is a prospective, observational, multicentre registry enrolling patients with acute decompensated HF.¹⁰ The exclusion criteria were as follows: age <20 years, lack of informed consent, and limited life expectancy due to malignant disease. Between June 2012 and March 2015, a total of 838 patients diagnosed with acute decompensated HF, based on the Framingham criteria,¹¹ and requiring hospitalization in 11 participating institutions were enrolled in the registry (Figure 1). Of these, we excluded 357 patients consisting of those who died during hospitalization, and those with severe mitral or aortic regurgitation, severe mitral annulus calcification or mitral stenosis, prior mitral valve surgery, sinus tachycardia, history of atrial fibrillation, atrial flutter, or atrial tachycardia, any kind of pacemaker, hypertrophic cardiomyopathy, constrictive pericarditis, and acute coronary syndrome from the analysis. Finally, the present study consisted of the remaining 481 patients. Patients with LVEF ≥50% were defined as having HF with preserved ejection fraction (HFpEF)¹² and remaining patients were defined as having HF with reduced ejection fraction (HFrEF). Data on demographics, medical comorbidities, medication, blood test, and New York Heart Association (NYHA) functional class were collected. This study was approved by the local ethics committee of each participating institution. All patients provided written informed consent prior to participation.

Figure 1

A flowchart illustrating the study population. ADHF, acute decompensated heart failure; DD, diastolic dysfunction; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction.

Echocardiography

Comprehensive transthoracic echocardiography was performed after resolution of symptoms was achieved by optimization of HF medical treatment. Echocardiographic images were recorded and analysed by two experienced cardiologists blinded to clinical data. Each parameter was determined according to the current ASE/EACVI recommendations.¹³ LV volumes and LVEF were calculated using the Simpson biplane methods of discs. Left atrial volume index (LAVI) was measured from the apical view with the biplane method of disks and was indexed by body surface area. The peak velocity of the early (E), late diastolic (A) mitral inflow, deceleration time of the E wave, and A wave duration were recorded from the apical four-chamber view. Tissue Doppler imaging was used to measure the early peak diastolic velocity (e') at the mitral annular septal and lateral corners. The E/e' ratio was calculated as E divided by the average of the two E' velocities. Peak tricuspid regurgitation (TR) jet velocity and pulmonary venous flow including peak systolic (S) velocity, peak diastolic (D) velocity, and atrial reversal (Ar) duration were obtained. Speckle tracking echocardiography analyses were performed using vendor-independent 2D Cardiac Performance Analysis software (TomTec Imaging System, Munich, Germany) to assess LV global longitudinal strain (LV-GLS).¹⁰

Classification of LVDD

LVDD was assessed according to both 2009 ASE/EAE and 2016 ASE/EACVI recommendations (Figure 2).^6,7 With reference to the 2009 recommendations,⁶ we used the algorithm to grade diastolic dysfunction. DD grading starts with e' velocities and LAVI, and then E/A ratio, deceleration time, average E/e’, Ar-A, and change in the E/A ratio with Valsalva manoeuvre were used. At least two positive measures were required to determine a DD grade. When DD grade could not be classified due to discrepancies between Doppler parameters, they were considered as indeterminate.

Figure 2

Algorithms for diastolic dysfunction grading by echocardiography according to 2016 recommendations (A) and 2009 recommendations (B). Av. E/e‘, mitral E velocity divided by the average of septal and lateral mitral annular e’ velocity; DT, deceleration time of E; LA, left atrial; LVEF, left ventricular ejection fraction; TR, tricuspid regurgitation; Val. ΔE/A, a change in mitral E/A ratio with the Valsalva manoeuvre. A and B were reprinted with permission from references 6 and 7, respectively.

With regard to the 2016 recommendations,⁷ we used the algorithm for evaluating LV filling pressures, as all patients had been admitted with acute HF. If mitral E/A is ≤0.8 and the E velocity is >50 cm/s, or if the E/A is >0.8 but <2, LAVI, E/e' ratio, and TR velocity were used to determine DD grade. When two of the three main parameters were available and only one parameter met the cut-off value in patients with reduced LVEF, pulmonary S/D ratio <1 was used as additional supporting data indicating Grade II DD.⁷ In the remaining cases, if there was a discrepancy between the only two available parameters in the three main criteria, DD grade was reported as indeterminate.

Follow-up and endpoints

The patients were followed up at 3-month intervals from the clinical visit after discharge. The attending physicians obtained follow-up information and updated the status of events on the ICAS Web database.¹⁰ The endpoint was a composite of cardiovascular death or unplanned hospitalization due to worsening HF. Cardiovascular death was defined as HF death, fatal myocardial infarction, sudden death, stroke, or cardiovascular haemorrhage.¹⁰

Statistical analysis

Means are expressed with 1 SD for continuous variables, and medians are presented with interquartile ranges for skewed variables. Categorical variables are presented as absolute numbers and percentages. The intergroup comparisons were performed using one-way analysis of variance with the post hoc Tukey test (normally distributed continuous variables), the Kruskal–Wallis test (skewed continuous variables), the χ² test, or Fisher’s exact test (dichotomous data), as appropriate. The concordance between 2016 and 2009 DD grading system was determined using kappa coefficient. The event-free survival probability was estimated by Kaplan–Meier analysis and differences between survival curves were analysed by the log-rank test. Cox proportional hazard regression models were analysed to evaluate the influence of DD grade on clinical outcomes. The incremental value of elevation of LV filling pressure corresponding to Grade II or III DD2016 over clinical data was assessed in three modelling steps. The first step consisted of clinical variables including age, sex, and NYHA III or IV. Serum brain natriuretic peptide (BNP) level was then added in the second step. LV parameters including LVEF, LV mass index, and LV-GLS were fitted on top of the Step 2 variables. In the final step, DD Grade II or III in 2016 recommendations were added. The improvements of global χ² values were calculated at each step. In addition, we also investigated whether the main parameters in DD2016 (E/e' ratio >14, TR velocity >2.8 m/s, and LAVI >34 mL/m²) have an incremental prognostics value over mitral E/A ratio ≥2 by nested Cox regression model. Finally, to compare the predictive ability of DD grade by 2016 and 2009 recommendations, we used the continuous net reclassification index (NRI).¹⁴ Values of P < 0.05 were considered to be statistically significant. All statistical analyses were performed using SPSS software (version 24.0; SPSS Inc., Chicago, IL, USA) and R software (version 3.4.3; R foundation for Statistical Computing, Vienna, Austria).

Results

Study population

Of 481 patients, 67% were men, with an overall mean age of 70 ± 14 years. In all, 328 (68%) were patients with HFrEF and the remaining 153 (32%) were those with HFpEF. The general characteristics of the patients stratified by DD grade are summarized in Table 1. Mean body mass index was 22.5 kg/m² and there were few obese patients. Serum BNP level had a trend towards increase according to the severity of DD grade irrespective of LVEF. In HFrEF, LVEF and LV-GLS had a trend towards decrease according to the severity of DD grade. Evidence-based medications for HF were prescribed to the majority of patients. TR velocity was feasible in 284 patients (87%) with HFrEF and 132 (86%) with HFpEF.

Table 1

Baseline characteristics stratified by 2016 recommendations

	Total (n = 481)	HFrEF (n = 328)a				HFpEF (n = 153)a
		Grade I DD (n = 122)	Grade II DD (n = 109)	Grade III DD (n = 95)	P-value	Grade I DD (n = 62)	Grade II DD (n = 62)	Grade III DD (n = 21)	P-value
Age (years)	70 ± 14	67 ± 14	74 ± 12b	62 ± 15b,c	<0.001	73 ± 12	77 ± 11	73 ± 14	0.30
Female	158 (33)	28 (23)	41 (38)b	9 (9)c	<0.001	19 (31)	43 (69)b	9 (43)c	<0.001
Body mass index (kg/m2)	22.5 ± 4.2	21.8 ± 3.9	21.3 ± 4.4	23.7 ± 4.9b	0.01	22.0 ± 3.6	22.7 ± 4.2	22.7 ± 3.6	0.56
Systolic blood pressure (mmHg)	118 ± 20	117 ± 21	118 ± 19	109 ± 18b,c	0.004	122 ± 17	128 ± 20	116 ± 16c	0.03
Diastolic blood pressure (mmHg)	65 ± 12	67 ± 13	65 ± 12	65 ± 11	0.34	64 ± 10	64 ± 13	63 ± 10	0.90
Heart rate (bpm)	70 ± 14	70 ± 14	68 ± 12	74 ± 14c	0.007	68 ± 14	65 ± 12	63 ± 13	0.32
NYHA Class III or IV	40 (8)	6 (5)	8 (7)	14 (15)b	0.009	4 (6)	6 (10)	2 (10)	0.80
Ischaemic aetiology	186 (39)	53 (43)	45 (41)	37 (39)	0.93	20 (32)	17 (27)	6 (29)	0.80
Hypertension	258 (54)	60 (49)	60 (55)	43 (45)	0.46	34 (55)	43 (69)	11 (52)	0.18
Diabetes	191 (40)	47 (39)	47 (43)	38 (40)	0.43	23 (37)	27 (44)	4 (19)	0.13
Medication
ACE-I or ARB	299 (62)	86 (70)	65 (60)	55 (58)	0.39	35 (56)	42 (68)	8 (38)c	0.04
Aldosterone antagonist	265 (55)	72 (59)	61 (56)	55 (58)	0.33	32 (52)	28 (45)	11 (52)	0.67
Beta-blocker	337 (70)	98 (80)	76 (70)	65 (68)	0.28	38 (61)	37 (60)	16 (76)	0.36
Loop diuretics	339 (70)	84 (69)	85 (78)b	65 (68)	0.04	39 (63)	44 (71)	15 (71)	0.63
Biochemical
Haemoglobin (g/dL)	12.2 ± 2.3	12.4 ± 2.2	12.1 ± 2.2	13.1 ± 2.4c	0.02	11.6 ± 2.2	11.0 ± 1.7	11.5 ± 2.4	0.32
eGFR (mL/min/1.73 m2)	51 ± 23	54 ± 24	45 ± 21b	57 ± 26c	0.001	50 ± 22	40 ± 17b	50 ± 24	0.03
BNP (pg/mL)	243 (130–479)	214 (120–410)	299b (181–436)	336b (193–687)	<0.001	120 (48–274)	137b (100–383)	196b (122–365)	0.03
Echocardiography
LVEDV (mL)	131 ± 60	139 ± 56	146 ± 59	172 ± 60b,c	<0.001	80 ± 34	85 ± 39	92 ± 33	0.41
LVESV (mL)	80 ± 51	88 ± 43	97 ± 51	119 ± 51b,c	<0.001	33 ± 17	33 ± 18	39 ± 19	0.47
LVEF (%)	43 ± 15	38 ± 9	35 ± 10	32 ± 9b	<0.001	59 ± 9	62 ± 8	60 ± 8	0.13
LV-GLS (%)	−10.5 ± 4.5	−9.4 ± 3.5	−8.8 ± 3.1	−7.5 ± 3.0b,c	0.001	−14.2 ± 3.9	−13.5 ± 5.0	−12.7 ± 5.0	0.42
E (cm/s)	78 ± 32	54 ± 19	79 ± 21b	97 ± 23b,c	<0.001	61 ± 17	84 ± 23b	123 ± 38b,c	<0.001
A (cm/s)	70 ± 28	76 ± 21	81 ± 26	34 ± 10b,c	<0.001	80 ± 22	92 ± 25b	46 ± 14b,c	<0.001
E/A ratio	1.4 ± 1.0	0.8 ± 0.4	1.1 ± 0.4	2.8 ± 0.8b,c	<0.001	0.8 ± 0.3	1.0 ± 0.3	2.8 ± 0.7b,c	<0.001
Deceleration time (ms)	196 ± 78	211 ± 97	198 ± 72	140 ± 40b,c	<0.001	233 ± 62	218 ± 57	207 ± 70	0.20
Septal e' (cm/s)	4.7 ± 1.9	4.1 ± 1.5	3.9 ± 1.3	4.5 ± 2.1b,c	<0.001	5.2 ± 1.5	4.3 ± 1.1b	5.4 ± 1.9c	0.001
Lateral e' (cm/s)	6.1 ± 2.6	5.6 ± 2.5	4.9 ± 1.8	6.2 ± 2.6c	0.001	6.9 ± 2.2	6.0 ± 2.3	7.5 ± 2.1c	0.02
Average E/e'	15.6 ± 7.1	11.9 ± 4.5	18.7 ± 5.4	19.0 ± 6.4b	<0.001	10.6 ± 3.4	17.2 ± 5.5b	20.8 ±9.7b,c	<0.001
PV S/D ratio (n = 271)	1.2 ± 0.6	1.5 ± 0.5	1.2 ± 0.5	0.6 ± 0.2b,c	<0.001	1.56 ± 0.47	1.42 ± 0.45	0.55 ± 0.22b,c	<0.001
LAVI (mL/m2)	47 ± 21	36 ± 14	52 ± 18	57 ± 21b	<0.001	36 ± 15	50 ± 13b	56 ± 21b	<0.001
TR velocity (cm/s)  (n = 416)	2.6 ± 0.6	2.2 ± 0.4	2.6 ± 0.5	2.8 ± 0.6b,c	<0.001	2.3 ± 0.6	2.7 ± 0.5b	3.0 ± 0.6b	<0.001

Results are shown as mean ± SD, median (IQR), or as n (%).

ACE-I, angiotensin-converting enzyme inhibitor; ARB, angiotensin II receptor blocker; BNP, brain natriuretic peptide; DD, diastolic dysfunction; eGFR, estimated glomerular filtration rate; HFpEF, heart failure with preserved ejection fraction; HFrEF, heart failure with reduced ejection fraction; LAVI, left atrial volume index; LVEDV, left ventricular end diastolic volume; LVEF, left ventricular ejection fraction; LVESV, left ventricular end systolic volume; LV-GLS, left ventricular global longitudinal strain; NYHA, New York Heart Association; PV, pulmonary vein; TR, tricuspid regurgitation.

a

Two patients (0.7%) in HFrEF and 8 (5.5%) in HFpEF were unable to determine DD grade.

b

P < 0.05 vs. Grade I.

c

P < 0.05 vs. Grade II.

Reclassification of DD grade according to 2016 recommendation

The concordance between the 2016 and 2009 algorithm was moderate (κ = 0.44, P < 0.001), with a reclassification rate of 38.5%, as shown in Figure 3. From the 172 patients determined as Grade I DD using the 2009 recommendations, 56 (33%) patients were reclassified as more advanced DD (Grade II or III) by 2016 recommendation. Also, 56 (45%) patients in Grade II and 57 (34%) in Grade III DD stratified by 2009 algorithm were reclassified as having milder DD by 2016 algorithm. Fourteen (3%) and 10 patients (2%) were indeterminate DD grade in 2009 and 2016 recommendations, respectively, because of discordance of Doppler measurements or lack of TR signal.

Figure 3

Reclassification of diastolic dysfunction grade. Number of each grade of DD and indeterminate cases as classified by the 2009 and 2016 recommendations are illustrated on the left and right, respectively.

Prognostic value of DD grade by 2016 recommendation in patients with HF

During a median follow-up period of 15 months, 127 primary endpoints occurred including 46 cardiovascular deaths and 81 rehospitalizations for HF. Table 2 showed the distribution of the four main parameters of DD adopted in 2016 recommendations. A significantly high prevalence of restrictive pattern of mitral inflow was noted in patients with adverse events irrespective of LVEF. In patients with HFrEF, the rate of adverse events was lower when more than two of three parameters (LAVI, E/e' ratio, and TR velocity) were normal. In patients with HFpEF, there were no significant differences among the prevalence of the three main parameters. In the nested Cox regression model, there was no incremental prognostic value of three parameters over E/A ratio ≥2 in predicting the adverse cardiovascular events (Supplementary data online, Figure S1).

Table 2

Distribution of 2D and Doppler variables

Echocardiographic parameters	HFrEF (n = 328)			HFpEF (n = 153)
	Events (n = 94)	No events (n = 234)	P-value	Events (n = 33)	No events (n = 120)	P-value
Mitral E/A ratio ≤0.8 + E ≤50 cm/s	12 (13)	40 (17)	0.33	8 (24)	31 (26)	0.83
Mitral E/A ratio ≥2	38 (40)	57 (24)	0.004	11 (33)	10 (8)	0.001
E/A is ≤0.8 and E velocity is >50 cm/s, or E/A is >0.8 to <2
≥2 main parameters negative	13 (14)	57 (24)	0.04	3 (9)	20 (17)	0.41
Three main parameters positive	9 (10)	20 (9)	0.87	4 (12)	12 (10)	0.75
Two of three parameters positive
LAVI >34 mL/m2, E/e' >14	21 (22)	48 (21)	0.71	5 (15)	27 (23)	0.35
LAVI >34 mL/m2, TRV >2.8 m/s	0 (0)	5 (2)	0.33	2 (6)	11 (9%)	0.74
E/e' >14, TRV >2.8 m/s	0 (0)	1 (0.4)	1.00	0 (0)	1 (0.8)	1.00
One parameter positive and PV S/D ratio <1 (when two of the three parameters were available in HFrEF)
LAVI >34 mL/m2	0 (0)	2 (0.9)	1.00
E/e' >14	1 (1)	2 (0.9)	1.00
TRV >2.8 m/s	0 (0)	0 (0)	1.00
Indeterminate (one positive and one negative)	0 (0)	2 (0.9)	1.00	0 (0)	8 (7)	0.20

Data are expressed as n (% in each subgroup). The three main parameters in 2016 recommendations are LAVI >34 mL/m², E/e' >14, and TRV >2.8 m/s.

TRV, tricuspid regurgitant velocity.

Figure 4 illustrated the event-free probability regarding primary endpoint for the patients classified into each DD grade according to both of 2016 and 2009 recommendations. Patients with Grade III DD showed significantly worse prognosis than did patients with Grade I or II and these results were consistent in DD2016 and DD2009. In multivariable Cox hazard model, Grade III DD had a higher risk of composite endpoint than Grade I DD even when controlling for age, sex, NYHA III or IV, and serum BNP level in the entire cohort, patients with HFrEF, and HFpEF, whereas Grade II DD did not show a significant increase in the risk compared to Grade I DD (Figure 5).

Figure 4

Kaplan–Meier survival curves of each diastolic dysfunction grade. Kaplan–Meier survival analysis for the primary endpoint in the entire study population (A and B), patients with HFrEF (C and D), and HFpEF (E and F). Each line indicates event-free rates of subgroups stratified according to the DD grading by the 2016 (A, C, and E) and 2009 recommendations (B, D, and F).

Figure 5

Independent impact of diastolic dysfunction grade on the primary endpoint. To test the impact of DD grade on primary endpoint, we used multivariable Cox regression models adjusted for the age, sex, NYHA III or IV, and BNP levels. Values represent adjusted hazard ratio with error bars indicating 95% CI in the entire cohort (A), patients with HFrEF (B), and HFpEF (C).

Nested regression models showed that Grade II or III in DD2016, representing the elevation of LV filling pressure, had significant incremental value in addition to clinically relevant factors (age, sex, and NYHA III or IV), serum BNP level, and LV parameters (LVEF, LV mass index, and LV-GLS) for the prediction of a composite endpoint irrespective of LVEF (Figure 6). In these models, LV parameters did not show significant incremental value over the clinical factors and serum BNP. However, the prognostic power of these models was further improved by the addition of the DD2016.

Figure 6

Incremental prognostic value of elevation of left ventricular filling pressure estimated by 2016 recommendations. Bar graphs show incremental value of elevation of LV filling pressure estimated by 2016 recommendations to clinical variables (age, sex, and NYHA III or IV), BNP levels and LV parameters (LVEF, LV mass index, and LV-GLS) for the prediction of the primary endpoint in the entire cohort (A), patients with HFrEF (B), and HFpEF (C). LV parameters were indicated as LV in the figures. Elevation of LV filling pressure was defined as Grade II or III DD in 2016 recommendations and was represented as DD2016.

Finally, we compared the predictive ability of DD2016 and DD2009. The NRI did not reach significance using DD2016, instead of DD2009, in addition to the basic clinical model in the entire cohort {NRI = 11% [95% confidence interval (CI) −7% to 30%], P = 0.23}, patients with HFrEF [NRI = 14% (95% CI −8% to 36%), P = 0.22], and HFpEF [NRI = −16% (95% CI −55% to 24%), P = 0.22].

Discussion

Major findings

Outcome-based studies to validate the new and simpler DD grading system proposed in 2016 ASE/EACVI recommendations are limited so far. This is one of the first studies to-date to report a prognostic value of the updated algorithm in patients with overt HF. In the present study, we demonstrated that: (i) concordance between 2016 and 2009 recommendation was moderate and substantial portion of patients were reclassified to different DD grade; (ii) feasibility of the 2016 DD grading system was relatively high in patients with overt HF; (iii) patients with Grade III in DD2016 had a higher risk of primary endpoint than those with Grade I, irrespective of LVEF; (iv) elevated LV filling pressure estimated by DD2016 had an incremental value to clinical variables for prediction of adverse events; and (v) when compared with DD2009, simplification of the algorithm in DD2016 did not cause the deterioration of risk prediction for the adverse cardiovascular events.

Reclassification and feasibility of DD grade by 2016 algorithm

As previously reported,^15–17 the concordance between DD2016 and DD2009 was not high in the present study. The important change of DD2016 is the application of TR velocity as one of three criteria to discriminate Grade II from Grade I. This revision might increase the number of those reclassified to different a DD grade or ‘indeterminate’ DD grade, which may make the new algorithm highly specific but poorly sensitive, and lead to diagnose only more advanced HF in comparison with the previous algorithm.^16,18 However, the prevalence of ‘indeterminate’ classified by DD2016 was only 2% in the present study including patients with overt HF. TR is common particularly in patients with advanced HF because elevated LA pressure leads to a rise in pulmonary arterial pressure, resulting in right ventricular and tricuspid annular dilatation, and development of TR.^19,20 Indeed, the feasibility of TR velocity was relatively high in the present study despite the records after the decongestive therapy and lead to high feasibility of DD grading using the 2016 algorithm.

The strengths and limitations of 2016 recommendations

The 2016 recommendations have been significantly simplified to overcome the complexity of the prior version.⁷ The revision has caused concern about the deterioration of clinical correlation. Although evaluation of prognostic value of DD2016 has still been limited,^21,22 our results might provide the additional clinical utility of 2016 recommendations to evaluate post-discharge outcomes in patients with acute decompensated HF. Recent invasive validation studies have demonstrated that DD2016 has good diagnostic accuracy of elevated LV filling pressure.^8,9,21,23 These favourable results support our findings that DD2016 had an incremental prognostic value in patients with HF. Thus, the strengths of the 2016 recommendations are the non-inferiority of the prognostic value compared to the previous version despite its simple and more practical algorithm for DD grading. In addition, we showed that Grade III DD in 2016 recommendations were an independent predictor of adverse events irrespective of LVEF. The current and previous recommendations stated that restrictive filling pattern of mitral E/A ratio in the combination with LA enlargement in patients with normal LVEF is associated with a poor prognosis similar to those with reduced LVEF showing a restrictive pattern.^6,7 Therefore, our results were not surprising; however, that might justify the simplification of DD2016 through the suggestion that Doppler parameters listed in DD2009 other than E/A >2 are not absolutely necessary to discriminate Grade III. The importance of restrictive pattern of E/A ratio might be further supported by the findings with E/e' ratio >14, TR velocity >2.8 m/s, and LAVI >34 mL/m² did not have an incremental prognostics value over mitral E/A ratio ≥2.

Although elevation of LV filling pressure corresponding to Grade II or III DD estimated by 2016 recommendations had an incremental prognostic value in the present study, the distribution of the echocardiographic parameters indicated the majority of patients with adverse events had the finding of mitral E/A ratio ≥2. Concerning LAVI, E/e', and TR velocity, emphasized in DD2016, patients of HFrEF with cardiovascular events had the higher positive prevalence of three parameters than those without events. However, the significant relationships between the prevalence of three main parameters and events were not found in HFpEF. Also, Grade II DD showed no significant increase of the cardiovascular risk when compared with Grade I DD. While each predictive value of LAVI, E/e', and TR velocity has been proven in patients with HFpEF,²⁴ a substantial overlap of these parameters between HFpEF and normal control has also been reported.¹⁸ The clinical significance to discriminate Grade II DD from Grade I DD using a combination of these three parameters might remain undetermined or be limited especially in HFpEF and require further investigations to be verified.

Clinical implication

Once patients are hospitalized with HF, repeated worsening of HF and admission after discharge makes the prognosis worse.²⁵ Echocardiographic estimation of LV filling pressure to detect residual congestion and increased risk for cardiovascular events is clinically important for the management of patients with HF.²⁶ Our results suggest the DD2016 might be useful to detect such vulnerable patients with elevated LV filling pressure and helpful to adjust the medications and the post-discharge follow-up plans. Of course, it remains unclear whether the DD grade-guided treatment for HF (e.g. optimization of diuretics) improves the prognosis compared with routine care. Future randomized trials would be required to clarify this aspect.

Limitations

First, the present study included a limited number of patients. Second, the current ESC guidelines for HF define those with LVEF in the range of 40–49% as HF with mid-range ejection fraction (EF).¹² Since patients with HF with mid-range EF have clinical characteristics that are more similar to those with HFrEF rather than HFpEF,²⁷ we defined HFrEF as patients with LVEF <50% in the present study. Although this cut-off value of LVEF was in line with the previous studies investigating the utility of DD2016, further studies might be required to clarify the clinical benefit of DD2016 in HF with mid-range EF. Finally, ICAS-HF is a registry of acute decompensated HF; therefore, our analysis included patients had more advanced HF with severe underlying disease. Also, the present study included few patients with obesity or overweight in contrast to HF patients in Europe and the USA. In addition, detection rate of TR velocity affecting on the feasibility of DD grading might be partially dependent on the technical skills of sonographers and physicians of each institutions. These might limit the generalizability of our results.

Conclusions

The simplification of algorithm for evaluation of LV diastolic function in 2016 ASE/EACVI recommendations resulted in a substantial reclassification of DD grades. Nevertheless, the feasibility and prognostic value of the current recommendations in cardiovascular risk stratification after discharge in HF patients was maintained and not compromised in comparison with the previous recommendations.

Acknowledgements

The authors thank all co-investigators for their valuable participation in the data collection: Haruhiko Higuchi (Hitachi General Hospital), Tomofumi Nakatsukasa (University of Tsukuba), Rihito Yamada (Mito Medical Center), Masako Baba (Ibaraki Prefectural Central Hospital), Tsuyoshi Enomoto, (Tsukuba Memorial Hospital), Yuichi Noguchi (Tsukuba Medical Center Hospital), Shoji Suzuki (National Hospital Organization Kasumigaura Medical Center), and Hiroshi Maeda (Ibaraki Seinan Medical Center Hospital).

Funding

This work was supported by JSPS KAKENHI (grant number: 16K09416) and the Japan Heart Foundation Research Grant.

Conflict of interest: none declared.