2019 Volume 83 Issue 3 Pages 604-613
Background: Using the normal values for the East Asian population, we evaluated age- and body size-adjusted left ventricular end-diastolic dimension (LVEDD) and its prognostic impact in a hospital-based population in Japan.
Methods and Results: We retrospectively analyzed data obtained from 4,444 consecutive patients who had undergone both transthoracic echocardiography and electrocardiography at Kitano Hospital in 2013. Those who presented with a history of previous episodes of myocardial infarction and severe or moderate valvular disease or with low ejection fraction (<50%) were excluded from the analysis. We calculated LVEDD adjusted by age and body surface area. A total of 3,474 patients were categorized into 3 groups: 401 with large adjusted LVEDD, 2,829 with normal adjusted LVEDD, and 244 with small adjusted LVEDD. Mean patient age in the large, normal, and small adjusted LVEDD groups was 66.6±18.4, 65.6±15.7, and 62.1±15.5 years, respectively (P<0.001). After adjusting for confounding factors, the excess adjusted 3-year risk of primary outcome of large adjusted LVEDD relative to normal LVEDD was significant (HR, 1.40; 95% CI: 1.08–1.78). The risk for primary outcomes of small adjusted LVEDD relative to normal adjusted LVEDD was significantly lower (HR, 0.55; 95% CI: 0.34–0.85).
Conclusions: Adjusted large LVEDD has a deleterious impact on long-term mortality, whereas small LVEDD carried a significantly lower risk.
Cardiac chamber size is altered in several heart diseases. The volume overload produced by mitral regurgitation causes compensatory left ventricular (LV) dilation, whereas pressure overload mediates LV hypertrophy with little or no increase in chamber size. Previous reports have shown LV dilatation to be a powerful predictor of adverse outcomes such as myocardial infarction (MI) or several other heart diseases independently of LV dysfunction.1–4 Similarly, patients with dilated cardiomyopathy have a large LV chamber, which is an adaptation of LV systolic dysfunction. LV dilatation in dilated cardiomyopathy has been linked poor prognosis.5,6 The importance of chamber size without valvular or myocardial disease, however, has not been elucidated.
LV chamber size is defined by body size, race, age, sex, and physique.7–9 The EchoNoRMAL Study published in 2015 reported age- and body size-adjusted normal references in different races including the East Asian population.10 Using the data in the EchoNoRMAL Study as a reference value, we analyzed the factors associated with large or small LV end-diastolic dimension (LVEDD). In addition, there have been no studies on the relationship between cardiac dimension and its prognostic impact in the Japanese population. Therefore, we also evaluated prognostic impact on cardiac events in a Japanese hospital-based population.
We retrospectively analyzed 4,444 patients who had undergone simultaneous transthoracic echocardiography (TTE) and electrocardiography (ECG) at Kitano Hospital during 2013.11 ECG and TTE were ordered at the discretion of the physician. A flowchart of subject selection is shown in Figure 1. A total of 970 patients who had previous MI (n=420) or severe or moderate valvular disease (aortic stenosis, n=133; aortic regurgitation, n=133; mitral stenosis, n=9; and mitral regurgitation, n=169) and low LV ejection fraction (LVEF; <50%, n=407) were excluded due to the diseases’ effects on cardiac dimensions, in addition to patients with no data on body surface area (BSA; n=11). Based on the TTE and ECG data, and data from the catheter suite’s database, we identified the patients who had a previous MI. The final population consisted of 3,474 patients (Figure 1).
Subject selection according to left ventricular end-diastolic dimension (LVEDD). AR, aortic regurgitation; AS, aortic stenosis; BSA, body surface area; ECG, electrocardiography; EF, ejection fraction; LV, left ventricular; MR, mitral regurgitation; MS, mitral stenosis; OMI, old myocardial infarction; TTE, transthoracic echocardiography.
The research protocol was approved by the Institutional Review Board of Kitano Hospital (approval number: P16-02-005). Informed consent was waived because this was a retrospective study. We disclosed the details of the present study to the public as an opt-out method and the notice clearly informed patients of their right to refuse enrollment. The study protocol conformed to the ethics guidelines of the 1975 Declaration of Helsinki, as reflected in a priori approval by the institution’s human research committee. Patient records and information were anonymized and de-identified before analysis.
Data CollectionUsing the TTE database, we extracted data regarding LV wall thickness, LV diastolic dimension (LVDd), LV systolic dimension (LVDs), left atrium diameter, left atrial volume index (LAVI), LVEF, and BSA. From the ECG database, we extracted cardiac rhythm data and recorded it as it was documented. Therefore, we could not determine whether atrial fibrillation (AF) was paroxysmal or persistent. The LV mass index (LVMI) and relative wall thickness (RWT) were calculated using the formula recommended by the American Society of Echocardiography (ASE) as follows: LVMI={0.8×1.04[(LVDd+LVPWTd+IVSTd)3−(LVDd)3]+0.6}/BSA, where LVDd is the LV diastolic diameter, IVSTd is the diastolic interventricular septal wall thickness, and LVPWTd is the diastolic LV posterior wall thickness, and RWT=(2×LVPWTd)/(LVDd).12
Large, normal, and small adjusted LVEDD were defined according to the formula proposed by the EchoNoRMAL Study.10 The following equations were used to define the upper and lower reference values of LV diastolic dimension divided by BSA for men: 2.98+0.0031×(age) and 2.45−0.0019×(age), and for women: 3.25+0.0015×(age) and 2.58−0.00059×(age). All 3,474 patients were categorized into 3 groups as shown in Figure 1. High LVPWTd was defined as >11 mm in men or >10 mm in women.12 Two-dimensional TTE data were analyzed at baseline. LVEF was measured using the Teichholz method or the modified Simpson rule methods. As supplementary analyses, we calculated the adjusted LV end-systolic dimension (LVESD) according to the following formula:10 the upper and lower reference values of LVESD divided by BSA for men: 2.16−0.0033×(age) and 1.55−0.0044×(age), and for women: 2.17−0.00056×(age) and 1.56−0.0018×(age).
We extracted patient information from the electronic medical records at the present institution, including age, sex, and type of disease (i.e., ischemic heart disease, International Statistical Classification of Diseases and Related Health Problems, Tenth Edition [ICD-10] codes I20, I21, I22, I23, I24, and I25; hypertension [HT], ICD-10 codes I10, I11, I12, I13, I14, and I15; dyslipidemia, ICD-10 code E78; diabetes mellitus [DM], ICD-10 codes E10, E11, E12, E13, and E14; and chronic kidney disease [CKD], ICD-10 code N18). The follow-up data from serial clinic visits were also collected retrospectively during June 2017 from the electronic medical records.
Outcome MeasuresThe primary outcome measure was a composite of all-cause death and major adverse cardiac events (MACE) defined as acute heart failure, acute MI, unstable angina pectoris, cerebral infarction, cerebral hemorrhage, aorta and peripheral vascular disease including the treatment of aortic aneurysm. The secondary outcome measure was all-cause death and MACE.
Statistical AnalysisCategorical variables are presented as n (%). They were compared using the chi-squared test or Fisher’s exact test. Continuous variables are expressed as mean±SD or median (IQR). Based on their distributions, the continuous variables were compared using Student’s t-test or Wilcoxon rank-sum test. To determine the differences between 3 groups, we performed the Dunn post-hoc test in each group.
To analyze the factors associated with large and small adjusted LVEDD, we used a multivariable logistic regression model involving the following potentially independent clinically relevant variables: age >80 years, sex, echocardiographic parameters (high LVPWd defined as >11 mm in men and 10 mm in women, and high LAVI defined as 34 mL/m2), and comorbidities (Table 1). We did not include LVMI or RWT because these parameters were derived from the calculation formula including LVDd.
| Total (n=3,474) |
Normal LVEDD (n=2,829) |
Large LVEDD (n=401) |
Small LVEDD (n=244) |
P-value† | P-value Large vs. Normal† |
P-value Small vs. Normal† |
|
|---|---|---|---|---|---|---|---|
| Age (years) | 65.4±16.0 | 65.6±15.7 | 66.6±18.4 | 62.1±15.5 | <0.001 | 0.0083 | <0.001 |
| Age >80 years‡ | 507 (14.6) | 392 (13.9) | 90 (22.4) | 25 (10.3) | <0.001 | <0.001 | 0.12 |
| Male‡ | 1,796 (51.7) | 1,527 (54.0) | 146 (36.4) | 123 (50.4) | <0.001 | <0.001 | 0.29 |
| BMI (kg/m2) | 23.1±4.2 | 23.1±3.9 | 21.1±4.0 | 26.5±5.0 | <0.001 | <0.001 | <0.001 |
| AF‡ | 329 (9.5) | 267 (9.4) | 41 (10.2) | 21 (8.6) | 0.79 | 0.73 | 0.73 |
| Diabetes‡ | 1,007 (29.0) | 798 (28.2) | 115 (28.7) | 94 (38.5) | 0.0030 | 0.86 | 0.0058 |
| HT‡ | 1,863 (53.6) | 1,487 (52.6) | 233 (58.1) | 143 (58.6) | 0.031 | 0.11 | 0.11 |
| Dyslipidemia‡ | 972 (28.0) | 778 (27.5) | 110 (27.4) | 84 (34.4) | 0.067 | 1.0 | 0.077 |
| IHD‡ | 849 (24.4) | 693 (24.5) | 96 (23.9) | 60 (24.6) | 0.97 | 1.0 | 1.0 |
| CKD‡ | 443 (12.8) | 315 (11.1) | 103 (25.7) | 25 (10.3) | <0.001 | <0.001 | 0.75 |
| LVDd (cm) | 4.59±0.51 | 4.57±0.46 | 5.03±0.52 | 4.06±0.48 | <0.001 | <0.001 | <0.001 |
| Adjusted LVEDD (cm/m2) | 2.88±0.35 | 2.85±0.24 | 3.49±0.23 | 2.30±0.16 | <0.001 | <0.001 | <0.001 |
| LVDs (cm) | 3.02±0.37 | 3.01±0.33 | 3.30±0.41 | 2.68±0.32 | <0.001 | <0.001 | <0.001 |
| Adjusted LVESD (cm/m2) | 1.89±0.24 | 1.87±0.18 | 2.29±0.20 | 1.52±0.13 | <0.001 | <0.001 | <0.001 |
| Normal adjusted LVESD | 2,565 (73.8) | 2,310 (81.6) | 36 (9.0) | 219 (89.8) | <0.001 | <0.001 | 0.0011 |
| Large adjusted LVESD | 884 (25.4) | 517 (18.3) | 365 (91.0) | 2 (0.8) | <0.001 | <0.001 | <0.001 |
| Small adjusted LVESD | 25 (0.7) | 2 (0.1) | 0 (0) | 23 (9.4) | <0.001 | 1.0 | <0.001 |
| IVSTd (cm) | 0.82±0.17 | 0.81±0.17 | 0.79±0.16 | 0.88±0.16 | <0.001 | 0.053 | <0.001 |
| LVPWd (cm) | 0.80±0.14 | 0.80±0.14 | 0.78±0.14 | 0.84±0.14 | <0.001 | 0.11 | <0.001 |
| LVPWd High (M >1.1 cm, F >1.0 cm)‡ |
74 (2.1) | 61 (2.2) | 5 (1.3) | 8 (3.3) | 0.22 | 0.34 | 0.34 |
| RWT | 0.35±0.07 | 0.35±0.06 | 0.31±0.05 | 0.42±0.08 | <0.001 | <0.001 | <0.001 |
| LVMI (g/m2) | 75.1±21.4 | 73.7±19.4 | 93.9±26.5 | 60.3±14.1 | <0.001 | <0.001 | <0.001 |
| High LVMI (M >115 g/m2, F >95 g/m2) |
304 (8.8) | 185 (6.5) | 119 (29.7) | 0 (0) | <0.001 | <0.001 | 0.40 |
| LAD (cm) | 3.49±0.65 | 3.48±0.64 | 3.61±0.75 | 3.41±0.62 | 0.0039 | 0.010 | 0.59 |
| LAVI (mL/m2) | 22.7±12.3 | 22.1±11.1 | 28.6±18.4 | 19.0±9.8 | <0.001 | <0.001 | <0.001 |
| High LAVI (LAVI ≥34 mL/m2)‡ | 368 (11.7) | 263 (10.2) | 90 (25.4) | 15 (7.0) | <0.001 | <0.001 | 0.15 |
| EF (%) | 63.3±4.1 | 63.4±4.0 | 62.8±4.8 | 63.6±3.8 | 0.23 | 0.29 | 1.00 |
| HR (beats/min) | 71.1±15.0 | 70.8±14.7 | 69.5±15.7 | 77.1±16.5 | <0.001 | 0.27 | <0.001 |
Data given as n (%) or mean±SD. †Chi-squared or Fisher’s exact test for categorical variables, and Student’s t-test or Wilcoxon rank sum test for continuous variables. ‡Potential risk-adjusting variables selected for Cox proportional hazard models. AF, atrial fibrillation; BMI, body mass index; BSA, body surface area; CKD, chronic kidney disease; EF, ejection fraction; HR, heart rate; HT, hypertension; IHD, ischemic heart disease; IVSTd, diastolic interventricular septal wall thickness; LAVI, left atrial volume index; LVDd, left ventricular diastolic dimension; LVDs, left ventricular systolic dimension; LVEDD, left ventricular end-diastolic dimension; LVMI, left ventricular mass index; LVPWD, diastolic left ventricular posterior wall thickness; RWT, relative wall thickness; TTE, transthoracic echocardiography.
Next, we compared the 3-year clinical outcomes between the large, normal, and small adjusted LVEDD groups. Cumulative incidences of clinical events were estimated using the Kaplan-Meier method, and the intergroup differences were assessed using the log-rank test. Multivariable Cox proportional hazards models were used to estimate the risk of primary and secondary outcomes associated with a large or small adjusted LVEDD relative to a normal adjusted LVEDD. The results are expressed as hazard ratios (HR) and 95% CI. We selected 10 clinically relevant risk-adjusted variables (Table 1) for the primary and secondary outcomes for use in the main analysis. Proportional hazard assumptions for the large, normal, and small adjusted LVEDD groups were assessed using plots of log (time) vs. log [−log (survival)] stratified by variable and were verified as acceptable. We also evaluated the interactions between each subgroup and the clinical effects of a large and small adjusted LVEDD relative to normal adjusted LVEDD for clinical outcomes.
For the supplemental analysis comparing the risk prediction of LVEDD and LVESD, we analyzed the cumulative incidences and HR regarding the adjusted LVESD. We compared the net re-classification improvement (NRI) and integrated discrimination improvement (IDI) between adjusted LVEDD and adjusted LVESD regarding the improvement in prognosis accuracy.13
All statistical analysis was conducted by physicians (Y.S., T.K., Y.M.) using JMP version 13 (SAS Institute, Chicago, IL, USA) and R 3.4.1 (R Foundation for Statistical Computing, Austria). NRI and IDI are 2 new metrics for the formal assessment of new risk factors, to supplement the improvement in the area under the curve (AUC), and were evaluated using the R package of survIDINRI (version 1.1.1). All reported P-values are 2-tailed, and P<0.05 was considered statistically significant.
A total of 401 patients had large, 2,829 patients had normal, and 244 patients had small adjusted LVEDD (Figure 1). The patient distribution according to sex (male, Figure 2A; female, Figure 2B) was determined using the reference values from the EchoNoRMAL Study.10 The baseline characteristics of the whole patient group are listed in Table 1. There were significant differences in sex, history of DM, HT, and CKD, LV dimensions, wall thickness, and LA dimension between the 3 groups (Table 1). Compared with the normal group, the patients with a large LVEDD were more likely to be older and female, and were more likely to have CKD, lower body mass index (BMI), higher LVMI, lower RWT, and higher LAVI. The patients with small LVEDD were younger than the normal group, and were more likely to have higher BMI, DM, higher LVMI, lower RWT, higher LAVI, and a lower heart rate.
Adjusted left ventricular end-diastolic dimension (LVEDD) according to age for (A) men and (B) women, and patient distribution according to the reference values from the EchoNoRMAL Study.10
According to the multivariable logistic regression analysis, age >80 years, CKD, and high LAVI were independently associated with large adjusted LVEDD, while male sex and AF had a negative association (Figure 3A). DM was an independent factor associated with small adjusted LVEDD, while high LAVI had a negative association (Figure 3B).
Multivariable logistic regression analysis. Factors associated with (A) large adjusted left ventricular end-diastolic dimension (LVEDD) and (B) small adjusted LVEDD. AF, atrial fibrillation; CKD, chronic kidney disease; DM, diabetes mellitus; DysL, dyslipidemia; HT, hypertension; IHD, ischemic heart disease; LAVI, left atrial volume index; LVPWTd, diastolic left ventricular posterior wall thickness.
The median follow-up duration after the index echocardiography was 1,274 days (IQR, 410–1,470 days), with a follow-up rate of 80.9% at 1 year, 74.9% at 2 years, and 67.4% at 3 years. The cumulative 3-year incidence of the primary and of the secondary outcome measures was significantly higher in the large adjusted LVEDD group than for the normal group. The cumulative 3-year incidence of the primary and of the secondary outcome measures was significantly lower in the small adjusted LVEDD group than for the normal group (composite of all-cause death and MACE, Figure 4A; all-cause death, Figure 4B; MACE, Figure 4C). After adjustment for confounders, the excess risk of primary outcomes and all-cause death in the large adjusted LVEDD group relative to that in normal adjusted LVEDD group remained significant (Table 2). The excess risk of primary outcome in the small adjusted LVEDD group relative to that in the normal adjusted LVEDD group remained significant (Table 2).
Cumulative incidence of (A) the primary outcome measure (all-cause death or major adverse cardiac events [MACE]) and (B,C) secondary outcome measures (B, all-cause death; C, MACE) for adjusted left ventricular end-diastolic dimension (LVEDD). MACE were defined as acute heart failure, acute myocardial infarction, unstable angina pectoris, cerebral infarction, cerebral hemorrhage, aortic dissection, and treatment of aortic aneurysm.
| Normal adjusted LVEDD |
Large adjusted LVEDD |
Small adjusted LVEDD |
Variables | Unadjusted | Adjusted | |||
|---|---|---|---|---|---|---|---|---|
| No. patients with event/no. patients at risk (cumulative 3-year incidence [%]) |
HR (95% CI) | P-value | HR (95% CI) | P-value | ||||
| Composite of all-cause death and MACE |
453/2,829 (16.5) |
96/401 (25.8) |
25/244 (12.3) |
Normal adjusted LVEDD |
Ref. | Ref. | ||
| Large adjusted LVEDD |
1.58 (1.26–1.96) |
<0.001 | 1.40 (1.08–1.78) |
0.012 | ||||
| Small adjusted LVEDD |
0.62 (0.40–0.91) |
0.012 | 0.55 (0.34–0.85) |
0.0059 | ||||
| All-cause death | 299/2,829 (10.9) |
70/401 (18.1) |
19/244 (8.7) |
Normal adjusted LVEDD |
Ref. | Ref. | ||
| Large adjusted LVEDD |
1.72 (1.32–2.22) |
<0.001 | 1.59 (1.17–2.13) |
0.0033 | ||||
| Small adjusted LVEDD |
0.72 (0.44–1.12) |
0.15 | 0.70 (0.40–1.14) |
0.17 | ||||
| MACE | 213/2,829 (7.8) |
47/401 (14.4) |
11/244 (6.1) |
Normal adjusted LVEDD |
Ref. | Ref. | ||
| Large adjusted LVEDD |
1.65 (1.19–2.24) |
<0.001 | 1.34 (0.93–1.90) |
0.12 | ||||
| Small adjusted LVEDD |
0.58 (0.30–1.01) |
0.054 | 0.43 (0.18–0.84) |
0.011 | ||||
LVEDD, left ventricular end-diastolic dimension; MACE, major adverse cardiac event.
There were no significant interactions between the subgroup factors and the effect of large or small LVEDD relative to normal LVEDD for primary outcomes, except for sex (Table 3). When stratified by sex, the risk for the primary outcome measures was significantly higher in the large adjusted LVEDD group and lower in the small adjusted LVEDD group than for the normal group in men (Table 3). In women, however, the risk for the primary outcome measures in the large and small adjusted LVEDD group relative to that in the normal adjusted LVEDD group was not significant (Table 3).
| Normal adjusted LVEDD |
Large adjusted LVEDD |
Small adjusted LVEDD |
Variables | Unadjusted | Adjusted | P-value for interaction |
|||
|---|---|---|---|---|---|---|---|---|---|
| No. patients with event/no. patients at risk (cumulative 3-year incidence [%]) |
HR (95% CI) | P-value | HR (95% CI) | P-value | |||||
| Age | 0.64 | ||||||||
| >80 years |
122/392 (33.5) |
33/90 (46.1) |
4/25 (20.0) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.32 (0.89–1.92) |
0.17 | 1.27 (0.79–1.97) |
0.31 | |||||
| Small adjusted LVEDD |
0.46 (0.14–1.10) |
0.084 | 0.45 (0.11–1.20) |
0.12 | |||||
| ≤80 years |
331/2,437 (13.8) |
63/311 (20.3) |
21/219 (11.4) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.54 (0.98–1.91) |
0.0030 | 1.46 (1.07–1.96) |
0.017 | |||||
| Small adjusted LVEDD |
0.69 (0.43–1.05) |
0.086 | 0.60 (0.35–0.96) |
0.032 | |||||
| Sex | 0.0097 | ||||||||
| Male | 290/1,527 (19.4) |
52/146 (37.8) |
8/123 (7.2) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
2.18 (1.60–2.90) |
<0.001 | 1.80 (1.41–2.60) |
0.0010 | |||||
| Small adjusted LVEDD |
0.33 (0.15–0.61) |
<0.001 | 0.30 (0.12–0.61) |
<0.001 | |||||
| Female | 163/1,302 (12.9) |
44/255 (18.6) |
17/121 (17.4) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.38 (0.98–1.91) |
0.065 | 1.03 (0.69–1.50) |
0.88 | |||||
| Small adjusted LVEDD |
1.10 (0.64–1.76) |
0.71 | 0.87 (0.47–1.48) |
0.63 | |||||
| CKD | 0.87 | ||||||||
| Yes | 92/315 (24.2) |
39/103 (37.9) |
3/25 (12.8) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.56 (1.06–2.26) |
0.024 | 1.54 (0.98–2.36) |
0.063 | |||||
| Small adjusted LVEDD |
0.37 (0.09–0.97) |
0.043 | 0.54 (0.13–1.48) |
0.26 | |||||
| No | 361/2,514 (15.4) |
57/298 (21.2) |
22/219 (12.2) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.38 (1.03–1.81) |
<0.001 | 1.32 (0.95–1.78) |
0.092 | |||||
| Small adjusted LVEDD |
0.68 (0.43–1.02) |
0.066 | 0.55 (0.32–0.88) |
0.012 | |||||
| AF | 0.90 | ||||||||
| Yes | 71/267 (28.0) |
18/41 (45.5) |
4/21 (21.0) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.74 (1.01–2.86) |
0.047 | 1.34 (0.70–2.41) |
0.36 | |||||
| Small adjusted LVEDD |
0.61 (0.19–1.48) |
0.31 | 0.59 (0.14–1.68) |
0.36 | |||||
| No | 382/2,562 (15.3) |
78/360 (23.4) |
21/223 (11.5) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.54 (1.20–1.95) |
<0.001 | 1.35 (1.02–1.78) |
0.038 | |||||
| Small adjusted LVEDD |
0.62 (0.39–0.94) |
0.022 | 0.55 (0.32–0.87) |
0.0098 | |||||
| HT | 0.42 | ||||||||
| Yes | 334/1,487 (20.6) |
81/233 (34.3) |
18/143 (13.6) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.73 (1.35–2.19) |
<0.001 | 1.46 (1.10–1.93) |
0.010 | |||||
| Small adjusted LVEDD |
0.54 (0.32–0.84) |
0.0054 | 0.52 (0.29–0.86) |
0.0084 | |||||
| No | 119/1,342 (10.7) |
15/168 (10.9) |
7/101 (10.0) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.01 (0.57–1.67) |
0.97 | 1.21 (0.66–2.06) |
0.52 | |||||
| Small adjusted LVEDD |
0.79 (0.33–1.56) |
0.52 | 0.65 (0.23–1.43) |
0.31 | |||||
| LAVI | 0.23 | ||||||||
| High | 72/263 (24.6) |
33/90 (36.5) |
4/15 (30.8) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.54 (1.01–2.31) |
0.045 | 1.33 (0.84–2.07) |
0.22 | |||||
| Small adjusted LVEDD |
1.24 (0.38–2.99) |
0.69 | 1.17 (0.35–2.88) |
0.77 | |||||
| Normal | 334/2,321 (14.8) |
48/265 (19.4) |
15/199 (8.9) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.29 (0.95–1.73) |
0.11 | 1.38 (1.00–1.86) |
0.048 | |||||
| Small adjusted LVEDD |
0.49 (0.28–0.80) |
0.0027 | 0.48 (0.27–0.78) |
0.018 | |||||
| LVMI | 0.60 | ||||||||
| High | 45/185 (22.2) |
32/119 (32.5) |
N/A | Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.22 (0.77–1.91) |
0.39 | 1.65 (0.92–2.99) |
0.093 | |||||
| Small adjusted LVEDD |
N/A | N/A | |||||||
| Normal | 408/2,644 (16.1) |
64/282 (23.1) |
25/244 (12.3) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.53 (1.16–1.97) |
0.0028 | 1.34 (0.99–1.78) |
0.059 | |||||
| Small adjusted LVEDD |
0.64 (0.42–0.94) |
0.022 | 0.57 (0.34–0.88) |
0.0097 | |||||
| Adjusted LVESD |
0.59 | ||||||||
| High | 125/517 (24.3) |
92/365 (26.8) |
1/2 (35.4) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
1.09 (0.83–1.42) |
0.55 | 1.32 (0.95–1.81) |
0.097 | |||||
| Small adjusted LVEDD |
1.90 (0.11–8.51) |
0.56 | 1.21 (0.07–5.67) |
0.85 | |||||
| Normal, Small |
328/2,312 (14.7) |
4/36 (14.1) |
24/242 (11.9) |
Normal adjusted LVEDD |
Ref. | Ref. | |||
| Large adjusted LVEDD |
0.88 (0.27–2.07) |
0.80 | 1.14 (0.35–2.70) |
0.80 | |||||
| Small adjusted LVEDD |
0.68 (0.43–1.00) |
0.051 | 0.57 (0.35–0.92) |
0.021 | |||||
LVESD, left ventricular end-systolic dimension. Other abbreviations as in Tables 1,2.
Baseline clinical and echocardiographic characteristics and the trend of outcomes between the LVESD groups were generally consistent with those in the adjusted LVESD groups (Supplementary Figures 1,2; Supplementary Table 1; composite of all-cause death and MACE, Supplementary Figure 3A; all cause death, Supplementary Figure 3B; MACE, Supplementary Figure 3C). The excess risk of primary and secondary outcomes in the large adjusted LVESD group relative to that in the normal adjusted LVESD group remained significant (Supplementary Table 2). According to NRI and IDI analysis, the improvement in prognosis accuracy did not differ significantly between adjusted LVESD and adjusted LVEDD (Table 4).
| Adjusted LVESD vs. adjusted LVEDD |
|
|---|---|
| IDI (95% CI) | 1.1 (−1.8 to 4.2) |
| P-value | 0.60 |
| NRI (95% CI) | 24.3 (−63.1 to 27.2) |
| P-value | 0.91 |
IDI, integrated discrimination improvement; NRI, net re-classification improvement. Other abbreviations as in Tables 1,3.
The main findings of this study are as follows: (1) higher age, CKD, and high LAVI were independently associated with large adjusted LVEDD, while male sex and AF showed a negative association; DM was an independent factor associated with small adjusted LVEDD, while high LAVI had a negative association; and (2) large adjusted LVEDD had a deleterious impact on outcome, while small LVEDD had a favorable impact.
Large adjusted LVEDD was associated with CKD and high LAVI in this study. CKD patients have been reported to have a large LV volume.14–16 One mechanism of LV dilatation is anemia and chronic fluid overload in CKD.15 Patients with high LAVI also have a large LV volume,17 and the atria will enlarge in response to pressure and volume overload.18 Another consideration is that LV dilatation is a compensatory mechanism for LV systolic dysfunction. In fact, LV dilatation was related to the risk of worse outcome in patients with MI, dilated cardiomyopathy, and valvular disease.1–6,19 In the present study, we excluded these patients with reduced EF (EF <50), old MI, and moderate and severe valvular disease (aortic stenosis, aortic regurgitation, mitral stenosis, and mitral regurgitation). The present study suggests that an adjusted large LVEDD is still an independent factor associated with worse outcomes even after adjusting for confounders and excluding patients with reduced EF and valvular heart disease.
Small adjusted LVEDD was associated with DM and not having a high LAVI in this study. Non-high LAVI is indicative of a non-stiff LV. In addition, younger age tended to be associated with small LVEDD with a favorable outcome. In contrast, DM patients have a lower LV end-diastolic volume (LVEDV) and are more likely to have concentric remodeling20,21 with cardiac steatosis and diastolic stiffness. HT causes pressure overload and subsequent LV hypertrophy22–24 and diastolic dysfunction,25 which is also the suggested reason as to why HT tended to be linked to small LVEDD. Thus, although associated with favorable outcome, the underlying reasons for small LVEDD differed on a patient-by-patient basis.
The standard value for LVEDD differs according to age, sex, and race.7–10 The age coefficient also differed between men and women: it was large for men compared with that in women. The values of the age coefficient were derived from a systematic review;10 thus, the precise mechanism for the observed difference was not provided. The noted difference is likely to be influenced by differences in the ventricular response. Female sex, however, carries an increased risk of myocardial hypertrophy with small chamber size caused by hormone disturbance after menopause, as reflected by the wide distribution in women according to age (Figure 2B).26,27 The wide distribution is a possible reason for the significant sex-related differences in outcome with adjusted LVEDD size. Another possible reason for the sex-related differences in outcomes is that, because of the low incidence of adverse events in women, non-cardiac mortality comprised a substantial proportion of all-cause death; therefore, there was less power to differentiate the prognostic impact of the size of LVEDD in women than in men.
Despite of the importance of chamber size, only one study has reported on the normal Japanese reference values,8 in which the references were given, in 5-year age increments for men and women, derived from 700 Japanese subjects with various comorbidities but without cardiac disease, without adjustment for body size. In addition, there is a paucity of echocardiographic data on long-term prognosis in Japan. Using the large clinical database, this is the first report in Japan to show the impact of adjusted LVEDD on outcome in a hospital-based population with 3-year follow-up. In this study, we selected adjusted LVEDD as an indicator of ventricular dilatation. LVEDD is a simple indicator and is routinely measured on TTE; LVEDV or LV end-systolic volume (LVESV) derived using the modified Simpson’s method are not always measured on screening. We also showed that the large adjusted LVESD had a worse outcome. On Kaplan–Meier curve analysis, adjusted LVEDD had a better prediction ability than adjusted LVESD, although NRI and IDI indicated an equivalent ability in the present study, probably due to the very limited number of small LVESD subjects. Although the LV systolic function and morphology are complex and tightly related to each other, the effects of large adjusted LVEDD have prompted physicians to investigate the underlying cause of the dimensional change as well as to manage the patients to prevent adverse outcomes. In addition, further large population-based studies in Japan are needed to validate the age-, BSA-, and sex-adjusted normal values on echocardiography, because the normal values used in this study were for an East Asian population not including Japanese subjects,10 instead of the normal values in the JAMP study.8
Study LimitationsThis study had several limitations. First, ECG and TTE were ordered at the discretion of the treating physician, with no standardized indications. Second, patient data were extracted from the electronic medical records, which resulted in a low follow-up rate, especially at 3 years. In addition, information regarding the symptoms was not included. Thus, we had no data on the proportion of heart failure with preserved EF. Third, we adopted normal reference values in an East Asian population not including Japanese ethnicity.10 We used this study because it set a usual value and formula. In fact, a similar trend was seen in the JAMP study.8 Fourth, we did not adopt LVEDV and LVESV because this measurement using the modified Simpson’s method was not always performed. Fifth, this was a single-center study performed in Japan; thus, possible selection bias cannot be excluded despite the large sample size. Finally, there remain unmeasured confounders affecting the long-term prognosis. Nevertheless, we conducted extensive statistical adjustment for the measured confounders.
Patients with large adjusted LVEDD are at a higher long-term risk of clinical events, while small LVEDD had a favorable impact.
The authors declare no conflicts of interest.
Y.S. and T.K. conceived the design, performed statistical analysis, and wrote the manuscript. Y.M. carried out statistical analysis. Y.Y., Y.H., T.I., S.M., E.N., H.H., T.H. and M.I. collected the data and made critical revisions. All authors read and approved the final manuscript.
Please find supplementary file(s);
http://dx.doi.org/10.1253/circj.CJ-18-1095
