The International Stress Echo Study in Ischemic and Non-ischemic Heart Disease

In a prospective, multicenter, international, observational study design, > 100 SE labs will be networked with systematic clinical, laboratory and imaging data collection at the time of SE and with structured follow-up information at least at 6 months and 1 year, up to 3 years for specific protocols. For all the groups the primary aim is to evaluate the feasibility of several indices of SE in the evaluation of patients with known or suspected specific disease (1,2,3). The secondary aim is to assess the value of each of these parameters in predicting the functional impairment. The tertiary aim is to assess the prognostic value of SE indices for prognostic stratification of the disease in the medium-long-term.

The study theater is the international network of cardiology SE laboratories, and the study is endorsed and promoted by the Italian Society of Echocardiography. The starting point of the recruitment phase was a recent electronic survey by the Italian Society of Echocardiography, in 2015 censoring 134 laboratories with moderate- (>100/year) to high- (>400/year) volume SE activities, which were precisely interrogated for interest in participation to SE2020. The recruitment plan forecasts 500 patients by the end of 2016, with doubling of the rate of enrollment in subsequent years, in parallel with the increasing number of recruiting labs fulfilling quality control criteria, reaching the target number of 100 at the end of the 5-year schedule (4).

Data collection Stress protocols are harmonized according to recent European and North-American scientific societies' guidelines, with semi-supine exercise recommended and pharmacological stress dosages up to 40 mcg/kg/min for dobutamine, up to 0.84 mg/kg in 6 min for dipyridamole, and up to a 4-min step of 200 microg/kg/min for adenosine. With dobutamine, atropine (up to 1 mg) can be administered in patients with suspected CAD (protocol 9), and it is associated with a higher rate of complications in those with a history of neuropsychiatric symptoms, reduced left ventricular function, or small body habitus. The maximal allowed dobutamine dose is 20 mcg/kg/min in patients with aortic stenosis, in whom higher doses are less safe and probably unnecessary (3). All laboratories will share a standardized case report form coded in a database format to facilitate retrieval and communication. For applications outside CAD and for CAD testing with vasodilator stress, no atropine is given on top of pharmacological stress. Although data collection with a dedicated project-specific case report form is allowed, we encourage implementing a dedicated, free ad-hoc system for data storage and reporting developed at the National Research Council, Institute of Clinical Physiology. The software provides a suitable informatics infrastructure for the SE 2020 Italian multicenter study, with an intuitive graphic interface, eye-catching graphic format and convenient reporting option. It could represent the trade-off between the comprehensive information required by scientific standards and the smooth workflow priority of busy, high-volume, clinically-driven activities. The software was developed and tested in Italian and the translation of the last release in other languages (English, Portuguese and Serbian) is currently in progress.

Data analysis Data will be expressed as mean ± standard deviation (normally distributed data, such as wall motion score index), median and inter-quartile (25th, 75th) range (non-normally distributed data, such as B-lines) or per cent frequency (categorical data, such as presence or absence of severe mitral regurgitation), with absolute numbers. One-sample comparisons will be performed using Wilcoxon test, and the chi-squared test without Fisher's correction for categorical data. Event rates will be estimated with Kaplan-Meier curves and compared by the log-rank test. Univariable analyses by Cox proportional hazards models will be performed to assess the association between each candidate variable and outcome. All variables with P <0.20 by univariable analysis will be considered as candidate variables for the multivariable analyses. Goodness of fit of the models will be based on C-statistics and its variants, adjusting for optimism using bootstrap replications (at least 1000). A receiver operating characteristic analysis will be used to obtain the best prognostic predictor for the individual SE variables. We will also analyze the data according to a clinically guided stepwise procedure, where the variables were included in the model in the same order in which they are actually considered by the cardiologist. Statistical significance will be set at p<0.05.

Quality control