The presence of left ventricular (LV) diastolic dysfunction (DD) as characterized by Doppler echocardiography is associated with worse overall mortality both in symptomatic and asymptomatic patients. However, available data on this topic come from referral centers and have been obtained by different, validated algorithms for each single study. Thus, we aimed at determining the feasibility of comprehensive evaluation of LVDD in a primary care outpatient setting and at testing the concordance of different methodological approaches in grading diastolic dysfunction. Eight hundred eighty-five consecutive outpatients, in sinus rhythm, prospectively underwent Doppler echocardiography according to a predetermined protocol. Feasibility of each LV diastolic index and concordance between 3 methods to determine the degree of LVDD, namely the American Society of Echocardiography/European Association of Echocardiography (ASE/EAE) recommendations, the Olmstead County, and the Canberra Study protocols, were tested. Feasibility of all diastolic indexes was high, ranging from 93% of Valsalva maneuver to ≥99% for mitral inflow and tissue Doppler parameters. Diastolic function was not classifiable in 6% to 19% of patients. The concordance for LV diastolic dysfunction degree was fair when comparing the classification of the ASE/EAE with those from Olmstead County (κ = 0.25; reclassification rate 51%) and Canberra Study (κ = 0.27; reclassification rate 43.7%), and was good for the comparison between the Olmstead County and Canberra classifications (κ = 0.68, reclassification rate 27%). In conclusion, feasibility of LV diastolic function measurements is very high and grading diastolic dysfunction is possible in most patients in primary care settings. Substantial differences, however, exist when concordance is tested among 3 documented criteria, resulting in poor concordance of data interpretation and hence patient stratification and clinical management.
Comprehensive evaluation of left ventricular (LV) diastolic dysfunction (DD) is a clinically relevant component of echocardiographic examinations because it assists in symptom interpretation and in predicting prognosis in patients with various cardiac syndromes. The overtime use of diastolic parameters identifies patients who develop signs of function instability and who need urgent intervention to avoid potentially preventable death. A typical scenario is heart failure with Doppler signs of raised filling pressures which is known to limit patient’s exercise capacity and provoke LV arrhythmias. Even in the setting of heart failure with preserved ejection fraction, diastolic measurements play a pivotal role in explaining symptoms and predicting prognosis. 7, Available cut-off values for grading DD are based on individual guidelines, which all originate from renowned international centers with years of expertise in the field. However, various documented studies have different grading algorithms which might make their direct application in daily cardiology outpatient services of limited clinical value. To test this hypothesis, we performed a comprehensive evaluation of LV diastolic function in a consecutive group of outpatients referred by general practitioners to primary care echocardiographic laboratories, to explore the feasibility of DD assessment in a noncardiac center setting and to test the concordance of different methodologic approaches in grading DD.
Methods
For the purpose of the GRAding Diastolic dysfunction in Outpatients (GRADO) Study, 885 consecutive outpatients aged >14 years, referred to primary care laboratories for an echocardiographic examination, were prospectively enrolled. Exclusion criteria were sinus tachycardia (>100 beats/min), atrial fibrillation or flutter, either complex or frequent (i.e., 0.10 ectopic beats/h on Holter monitoring) supraventricular or ventricular arrhythmias, mitral stenosis, any other valvular disease of more than mild severity, severe mitral annulus calcification, previous mitral valvuloplasty, valve prosthesis, permanent pacemaker stimulation, complete right or left bundle branch block, and inadequate acoustic windows. All patients underwent a targeted history, standard 12-lead electrocardiogram, and comprehensive Doppler echocardiography. Body surface area, body mass index (BMI), systolic and diastolic blood pressure, and heart rate were calculated using conventional procedures. Informed consent was obtained from each patient, and the study protocol conformed with the ethics guidelines of the 1975 Declaration of Helsinki according to a priori approval by the institution’s human research committee. Studies were prospectively performed by 3 independent, board-certified cardiologist, with >15 years of clinical experience in performing and interpreting echocardiographic examinations, using commercially available ultrasound systems, according to a predetermined protocol. LV size and function were assessed according to the American Society of Echocardiography/European Association of Echocardiography (ASE/EAE) recommendations. All the indexes for assessing diastolic function were measured from the apical views. Mitral inflow peak early (E) and late diastolic (A) velocities, E/A ratio, E wave deceleration time (DT), and A wave duration (A dur ) were measured at baseline and during the strain phase of the Valsalva maneuver, to reduce LV preload and elicit changes in LV inflow to distinguish normal from pseudonormal patterns. A decrease of 20 cm/s in peak E velocity was considered adequate in patients with reversible restrictive filling pattern, and a relative decrease of ≥50% or an absolute decrease of ≥0.5 in the E/A ratio, based on each classification, a highly specific feature of increased LV filling pressures. Pulmonary vein peak systolic (S) and diastolic (D) flow velocities, S/D ratio, and atrial reversal flow (AR) velocity were recorded and the duration of flow at atrial contraction (AR dur ). The difference between AR dur and A dur were calculated, with an AR dur − A dur >30 ms as an early and reliable marker of increased LV end-diastolic pressure, independent from ejection fraction. Mitral annular velocities (septal and lateral) were measured by pulsed-wave tissue Doppler imaging. Peak systolic (s’), early diastolic (e’), and late diastolic (a’) velocities were measured at the 2 annular sites and averaged. The septal, lateral, and average E/e’ ratios were calculated. Mitral annular velocities are used to draw inferences about LV relaxation and E/e’ can provide an estimate of LV filling pressures. Left atrial volume was obtained from dedicated views and indexed to body surface area (LAVi). All examinations were performed after fine adjustment of both ultrasound and pulsed Doppler beam orientation and sample volume positioning, after careful optimization of spectral gain and filter settings, taking the average of 3 consecutive beats for each measurement.
Three different classifications of LV diastolic dysfunction (DD) were considered and compared in this study. The first was that of the joined ASE/EAE recommendations for the evaluation of LV diastolic function by echocardiography. This classification suggests a practical scheme for grading DD starting from both septal and lateral e’ and LAVi. For abnormal values of LAVi and e’, it recommends a comprehensive evaluation based on the use of specific cut-off values for E/A and DT, of the results of the Valsalva maneuver, of average E/e’, and of AR dur − A dur , finally identifying 3 grades of LVDD. The second classification has been validated in residents aged ≥45 years of Olmsted County, Minnesota, and identifies 4 categories of LV diastolic function, namely normal function, mild, moderate, and severe dysfunction, based on the assessment of resting E/A ratio and DT, E/A ratio at peak Valsalva maneuver, septal E/e’ ratio, S/D <1, and AR dur − A dur >30 ms. In this classification, 2 Doppler criteria consistent with moderate or severe DD are required for patients to be so classified. Subjects with only 1 criterion for moderate or severe dysfunction or those whose indexes were borderline or suggestive for but not clearly definitive for DD are classified as having undetermined dysfunction. The third classification was validated in randomly selected residents, aged 60 to 86 years, of Canberra, Australia. This classification is similar to that of Redfield et al but uses a greater cut-off value to identify abnormal DT (>160 vs >140 ms), does not include recording of LV inflow indexes during Valsalva, and uses lateral E/e’. Overall data were anonymized, and grading of LV diastolic function was performed by an independent investigator who did not take part in data acquisition and was blinded to patients’ clinical and echocardiographic characteristics.
Data were expressed as mean ± standard deviation. The concordance between classification methods for the determination of the degree of LV DD was tested by calculating the kappa coefficient and the overall proportion of agreement. On the basis of the kappa value, the concordance was defined as slight (0 to 0.20), fair (0.21 to 0.40), moderate (0.41 to 0.60), good (0.61 to 0.80), and optimal (0.81 to 1). The overall proportion of agreement expressed the proportion of observations identically classified between 2 methods. The difference between 100% and the proportion of agreement corresponds to the reclassification rate. The comparisons between classifiable and unclassifiable subjects were performed using the Student t test for independent groups for continuous variables and the chi-square test, or the Fisher exact test when appropriate, for categorical variables. A p <0.05 was considered for statistical significance, and all tests were 2 tailed. All analyses were performed using SPSS for Windows release 15.0 (Statistical Packages for Social Sciences Inc., Chicago, Illinois). The reproducibility of LV diastolic measurements in our laboratories was previously reported.
Results
Enrolled patients were aged 61 ± 12 years and 48% were women; LV ejection fraction was <55% in 36 patients (4.1%) and wall motion score index >1 in 47 (5.3%). The most common cardiovascular risk factors and reasons for referral were arterial systemic hypertension in 569 (64.3%), dyslipidemia in 377 (42.6%), type 2 diabetes in 125 (14.1%), current smoking in 112 (12.7%), history of coronary artery disease in 66 (7.5%), thyroid dysfunction in 87 (9.8%), peripheral vascular disease in 72 (8.1%), and chronic kidney disease in 31 patients (3.5%). Feasibility was high for all LV diastolic indexes ( Figure 1 ). Distributions of the main parameters of diastolic function, according to different cut-off values proposed by each convention, are reported in Table 1 . In comparison with the ASE/EAE classification, a smaller proportion of subjects were allocated in the lowest E/A category and a larger proportion in the middle one (p = 0.0014) by Olmstead County and Canberra Study classifications. The single DT cutoff of 160 ms suggested by the Canberra study and included in the ASE/EAE classification identified a significantly lower proportion of patients with abnormally prolonged DT compared to the single cutoff of 140 ms suggested by the Olmstead County classification (p <0.0001). Finally, by comparing the proportions of patients with E/e’ suggestive of raised LV filling pressure across the 3 classifications, the prevalence of patients with definitely abnormal E/e’ was higher using the Canberra study convention than the ASE/EAE classification and further increased using the Olmsted County classification (p <0.0001 for all comparisons). Of the 846 patients in whom systolic and diastolic pulmonary vein flow recording was feasible, an S/D <1 was found in 96 (11.3%). Of the 854 patients in whom AR dur − A dur was feasible, 97 (11.4%) had a value >30 ms. The concordance among the 3 methods for determining the degree of LV diastolic dysfunction was fair when comparing the classification of the ASE/EAE with that from the Olmstead County and Canberra Study classifications, with reclassification rates ranging from 27% to 51%, and suboptimal for the comparison between the Olmstead County and Canberra Study classifications ( Figures 2–4 ). Diastolic function was not classifiable in 6% to 19% of patients, with significantly more patients with undetermined DD according to the ASE/EAE classification compared to those by the Canberra Study (p <0.0001; Figure 5 ). The comparison between subjects who were classifiable in 1 of LVDD degrees and those for whom LVDD category was undetermined is reported in Table 2 .
| E/A | DT (ms) | E/e’ | |
|---|---|---|---|
| ASE/EAE | <0.8: n=217(24.5%) 0.8-1.5: n=544 (61.5%) >1.5: n=124 (14.0%) | >200: n=451 (51.2%) 160-200: n=288 (32.7%) <160: n=142 (16.1%) | Average <8: n=523 (59.3%) 9-12: n=309 (35.0%) >13: n=50 (5.7%) |
| Olmstead County | <0.75: n=156 (17.6%) 0.75-1.5: n=605 (68.4%) >1.5: n=124 (14.0%) | >140: n= 830 (94.2%) | Septal ≥10: n=261 (29.6%) |
| Canberra Study | >160: n= 739 (83.9%) | Lateral ≥10: n=127 (14.4%) |
| ASE/EAE | Olmstead County | Canberra Study | |||||||
|---|---|---|---|---|---|---|---|---|---|
| Classified (n=713) | Indeterminate (n=172) | P value | Classified (n=742) | Indeterminate (n=143) | P value | Classified (n=835) | Indeterminate (n=50) | P value | |
| Age (years) | 59.3 ± 12.6 | 67.0 ± 9.3 | <0.0001 | 60.7 ± 12.3 | 64.3 ± 12.3 | 0.50 | 60.9 ± 12.3 | 60.3 ± 14.1 | 0.75 |
| Female gender | 343 (48.1%) | 87(50.6%) | 0.62 | 355 (47.9%) | 75 (52.4%) | 0.36 | 399 (47.8%) | 31 (62.0%) | 0.07 |
| Hypertension | 444 (62.3%) | 125 (72.7%) | 0.014 | 538 (64.2%) | 98 (68.5%) | 0.39 | 539 (64.6%) | 30 (60.0%) | 0.25 |
| Diabetes mellitus | 88 (12.3%) | 37 (21.5%) | 0.0029 | 116 (13.8%) | 20 (14.0%) | 0.71 | 116 (13.9%) | 9 (18.0%) | 0.36 |
| Dyslipidemia | 284 (39.8%) | 93 (54.1%) | 0.0010 | 358 (42.7%) | 63 (44.1%) | 0.41 | 355 (42.5%) | 22 (44.0%) | 0.99 |
| Active smoking | 93 (13.0%) | 19 (11.0%) | 0.56 | 102 (12.2%) | 21 (21.3%) | 0.86 | 103 (12.3%) | 9 (18.0%) | 0.34 |
| Coronary artery disease | 45 (6.3%) | 21 (12.2%) | 0.013 | 60 (7.1%) | 11 (12.8%) | 0.99 | 60 (7.2%) | 6 (12.0%) | 0.26 |
| Renal disease | 15 (2.1%) | 10 (5.8%) | 0.013 | 25 (3.0%) | 1 (0.7%) | 0.10 | 25 (3.0%) | – | 0.39 |
| Body mass index (kg/m 2 ) | 27.3 ± 4.4 | 28.2 ± 4.1 | 0.015 | 27.4 ± 4.4 | 28.7 ± 4.7 | 0.041 | 27.5 ± 4.4 | 27.1 ± 4.5 | 0.51 |
| Body surface area (m 2 ) | 1.83 ± 0.22 | 1.81 ± 0.21 | 0.17 | 1.83 ± 0.21 | 1.83 ± 0.25 | 0.81 | 1.83 ± 0.22 | 1.77 ± 0.19 | 0.018 |
| Systolic blood pressure (mmHg) | 141.9 ± 19.2 | 143.5 ± 21.2 | 0.37 | 142.4 ± 19.5 | 142.9 ± 17.8 | 0.54 | 142.4 ± 19.7 | 139.2 ± 18.1 | 0.24 |
| Diastolic blood pressure (mmHg) | 81.1 ± 8.9 | 80.9 ± 8.3 | 0.83 | 81.1 ± 8.7 | 81.7 ± 8.2 | 0.67 | 81.1 ± 8.7 | 79.5 ± 8.9 | 0.23 |
| Heart rate (bpm) | 68.2 ± 10.9 | 69.5 ± 10.4 | 0.18 | 68.1 ± 10.5 | 73.8 ± 12.3 | 0.015 | 68.3 ± 10.6 | 71.9 ± 13.8 | 0.07 |
| Indexed LV end-diastolic volume (ml/m 2 ) | 57.4 ± 13.3 | 57.9 ± 13.2 | 0.69 | 57.6 ± 13.2 | 59.4 ± 18.8 | 0.47 | 57.3 ± 13.1 | 59.1 ± 17.1 | 0.47 |
| Indexed LV end-systolic volume (ml/m 2 ) | 20.1 ± 7.8 | 20.7 ± 7.7 | 0.40 | 20.3 ± 7.6 | 22.6 ± 13.4 | 0.89 | 20.1 ± 7.4 | 22.1 ± 12.9 | 0.29 |
| Ejection fraction (%) | 65.3 ± 6.7 | 64.5 ± 6.7 | 0.17 | 65.2 ± 6.6 | 63.4 ± 9.0 | 0.96 | 65.2 ± 6.5 | 64.5 ± 9.2 | 0.59 |
| Indexed LV mass (g/m 2 ) | 91.2 ± 20.8 | 98.9 ± 22.0 | <0.0001 | 92.7 ± 21.4 | 100.6 ± 20.5 | 0.93 | 92.6 ± 21.1 | 94.1 ± 23.4 | 0.67 |
| E/A ratio | 1.26 ± 0.92 | 0.82 ± 0.26 | <0.0001 | 1.19 ± 0.91 | 0.83 ± 0.39 | 0.026 | 1.17 ± 0.86 | 1.23 ± 0.84 | 0.60 |
| Deceleration time (ms) | 202 ± 57 | 240 ± 69 | <0.0001 | 212 ± 58 | 237 ± 64 | 0.65 | 211 ± 58 | 221 ± 63 | 0.31 |
| Indexed LA volume (ml/m 2 ) | 33.5 ± 11.3 | 35.2 ± 10.4 | 0.054 | 33.6 ± 11.2 | 38.6 ± 12.2 | 0.25 | 33.7 ± 11.0 | 36.2 ± 13.1 | 0.18 |
| Average s’ (cm/s) | 9.3 ± 1.9 | 8.6 ± 1.3 | <0.0001 | 9.2 ± 1.8 | 8.8 ± 2.3 | 0.049 | 9.2 ± 1.8 | 8.8 ± 2.0 | 0.17 |
| Average e’ (cm/s) | 10.6 ± 2.8 | 7.7 ± 1.7 | <0.0001 | 10.2 ± 2.8 | 7.7 ± 3.4 | 0.0025 | 10.1 ± 2.8 | 9.1 ± 4.1 | 0.10 |
| Average a’ (cm/s) | 10.9 ± 2.2 | 11.4 ± 2.1 | <0.0001 | 11.0 ± 2.2 | 11.7 ± 2.2 | 0.27 | 11.1 ± 2.2 | 10.7 ± 2.4 | 0.28 |
| Septal E/e’ ratio | 9.0 ± 3.5 | 9.8 ± 3.2 | 0.0039 | 8.9 ± 3.5 | 12.0 ± 3.8 | <0.0001 | 9.1 ± 3.4 | 10.9 ± 4.1 | 0.0038 |
| Lateral E/e’ ratio | 7.2 ± 3.0 | 7.8 ± 3.0 | 0.015 | 7.2 ± 2.9 | 9.8 ± 5.2 | 0.013 | 7.2 ± 2.7 | 9.4 ± 5.2 | 0.0055 |
| Average E/e’ ratio | 7.9 ± 2.9 | 8.6 ± 2.8 | 0.0048 | 7.9 ± 2.9 | 10.6 ± 4.2 | 0.0002 | 7.9 ± 2.8 | 9.9 ± 4.4 | 0.0030 |
| Pulmonary vein flow S (cm/s) | 60.6 ± 14.1 | 57.9 ± 10.9 | 0.0085 | 60.1 ± 13.7 | 57.7 ± 10.0 | 0.91 | 60.4 ± 13.7 | 54.9 ± 11.2 | 0.0016 |
| Pulmonary vein flow D (cm/s) | 47.0 ± 13.1 | 39.8 ± 10.3 | <0.0001 | 46.1 ± 13.2 | 39.2 ± 8.6 | 0.65 | 45.5 ± 12.7 | 47.1 ± 16.4 | 0.50 |
| Pulmonary vein flow S/D ratio | 1.35 ± 0.34 | 1.53 ± 0.40 | <0.0001 | 1.37 ± 0.36 | 1.54 ± 0.44 | 0.025 | 1.38 ± 0.35 | 1.29 ± 0.49 | 0.20 |
| Pulmonary vein flow A rev (cm/s) | 30.3 ± 8.2 | 29.1 ± 5.8 | 0.024 | 30.2 ± 8.1 | 29.1 ± 5.9 | 0.17 | 30.0 ± 7.9 | 31.3 ± 7.4 | 0.25 |
| AR Dur -A Dur (ms) | -3.0 ± 33.0 | -15.0 ± 27.9 | <0.0001 | -4.2 ± 32.9 | -9.3 ± 21.1 | 0.011 | -5.3 ± 32.5 | -5.7 ± 31.3 | 0.93 |
Stay updated, free articles. Join our Telegram channel
Full access? Get Clinical Tree
