Introduction

Diastolic dysfunction (DD) is an alteration of relaxation, filling, and/or distensibility of the left ventricle [1]. DD can lead to diastolic heart failure and increases the risk of readmission rates and in-hospital mortality [2]. The increased prevalence of DD has led to growing interest in early detection in acute care settings [3, 4].

The American Society of Echocardiography (ASE) guidelines outline a detailed algorithm for the diagnosis of DD which includes (1) spectral pulsed wave Doppler of transmitral inflow; (2) pulsed wave Doppler profile of pulmonary venous flow; (3) mitral annulus downward velocity measurements (eʹ) using tissue Doppler imaging (TDI) at the septum (eʹS) and lateral wall (eʹL); and (4) left atrial (LA) volumes [5]. Obtaining these multiple measurements may be time-consuming and difficult for the average EP.

Average peak mitral annulus velocity by TDI (eʹA = [eʹS + eʹL]/2) has been described as an acceptable single-step method for assessing LV relaxation, using eʹA < 9 cm/s as a threshold [6,7,8,9]. TDI measurements can be obtained in 30 s with nearly 100% success rate, even with poor echocardiographic windows [10, 11]. This simplified approach may be more suitable for use by EPs with limited experience in echocardiography.

The purpose of this study was to ascertain inter-rater agreement in DD determination between eʹA < 9 cm/s measured by EPs and cardiologist interpretation of LBEs following the ASE guidelines.

Methods

Study design

This was a secondary data set analysis of LBEs completed as part of a prospective, cross-sectional with longitudinal follow-up study (details provided elsewhere) of patients presenting to the emergency department (ED) with asymptomatic elevated blood pressure [12, 13].

Study protocol and measurements

LBEs were performed based on research staff availability by EPs (two emergency ultrasound fellowship-trained faculties and one emergency ultrasound fellow) who had performed at least 100 LBEs through routine clinical care and who underwent training and demonstrated proficiency in diastology with a board-certified cardiologist. A sonosite M-Turbo ultrasound system equipped with a harmonic 4.0-MHz variable-frequency phased-array transducer was used to obtain images and measurements. Studies were digitally archived for cardiologist review.

EPs utilized electrocardiogram (EKG) rhythm strips to time diastole. EPs determined eʹA by averaging eʹS and eʹL measurements. EPs considered an eʹA < 9 cm/s as evidence of DD without adjustment for age or other risk factors. A board-certified cardiologist with an ASE level III echocardiography certification independently reviewed LBE images while blinded to EP interpretation. The cardiologist rated the images in accordance to the 2009 ASE guidelines [8] and upon reviewing digital recordings of the following: parasternal long view for determination of LV wall thickness, apical four-chamber view for estimation of LA size, E and A measurements, eʹS and eʹL, E/eʹ ratios to assess LA pressure, estimation of LA size, and the EKG rhythm strip (see Table 1 for comparison of data interpretation).

Table 1 Comparison of data utilized by emergency physician vs. cardiologist for determination of diastolic dysfunction
Full size table

Data analysis

EPs and cardiologist indicated DD present, DD absent, or indeterminate for each LBE study. A 3 × 3 contingency table provided a summary of agreement. Inter-rater reliability between EPs and the cardiologist was determined using an unweighted kappa with 95% confidence interval (CI) coefficient using Stata Release 15, StataCorp.

Results

Forty-eight studies were submitted to the cardiologist for review. Cardiologist and EP agreement are summarized in Table 2. Agreement was reached in 41 out of 48 cases (85.4%). The unweighted kappa coefficient was 0.74 (95% CI 0.57–0.92).

Table 2 Agreement between emergency physicians and cardiologist interpretation
Full size table

Discussion

Diastolic dysfunction is prevalent and delays in diagnosis can lead to increased morbidity and mortality. EPs with focused training in diastology may identify diastolic dysfunction with high sensitivity compared to a cardiologist trained in echocardiography. Previous studies have demonstrated that EPs can identify DD with high sensitivity, but either did not include TDI as part of their assessment [14] or reported only moderate agreement with cardiologist interpretation [4]. One study showed that EPs who met minimum requirements for LBEs based on American College of Emergency Physicians guidelines demonstrated high inter-rater agreement in the assessment of DD using primarily TDI, but failed to compare EP to a cardiologist interpretation [15]. Our study addresses the limitations of previous evidence by demonstrating that by following a more simplified approach using eʹA alone, EPs can identify DD with high level of agreement compared to a cardiologist following the ASE guidelines.

Limitations

Our sample size and convenience sampling may have introduced selection bias thus preventing a definitive correlation between eʹA and DD. EPs did not screen for regional wall motion abnormalities. Because wall motion abnormalities of the left ventricular basal segments can influence mitral annulus TDI diastolic velocities, this may have led to an overestimation of DD prevalence. Moreover, comparison was limited to cardiologist interpretation of LBE images, which may not be representative of typical exams obtained by a technician or specialist. A larger, multi-center study comparing EP assessment of eʹA against performance of a comprehensive echocardiogram can help establish external validity.

Conclusions

This study highlights a promising simplified approach for identifying DD by EPs. Relying on eʹA alone achieved good agreement for determination of DD compared to LBE interpretation by cardiologist. Future studies should further investigate this simplified approach as a one-step method of screening for LV DD in the emergency department.