Appropriate use criteria implementation with modified Haller index for predicting stress echocardiographic results and outcome in a population of patients with suspected coronary artery disease

Raw data related to the article

Abstract
The hypothesis that modified Haller index (MHI) integration with the existing appropriate use criteria (AUC) categories may
predict exercise stress echocardiography (ESE) results and outcome of patients with suspected coronary artery disease (CAD)
has never been previously investigated. We retrospectively analyzed 1230 consecutive patients (64.8 ± 13.1 years, 58.9%
men) who underwent ESE for suspected CAD between February 2011 and September 2019 at our institution. MHI (chest
transverse diameter over the distance between sternum and spine) was assessed in all patients. A true positive (TP) ESE was
a positive ESE with obstructive CAD according to subsequent coronary angiography. During follow-up time, we evaluated
the occurrence of any of the following: (1) cardiovascular (CV) hospitalizations; (2) Cardiac death or sudden death. Overall,
734 (59.7%), 357 (29.0%) and 139 (11.3%) indications for ESE were classified as appropriate (Group 1), rarely appropriate
(Group 2) and which may be appropriate (Group 3), respectively. A funnel chest (defined by an MHI > 2.5) was detected
in 30.3%, 82.1% and 49.6% of Groups 1, 2 and 3 subjects, respectively (p < 0.0001). On multivariate logistic regression
analysis, male sex (OR 1.41, 95%CI 1.02–2.03, p = 0.01) and type-2 diabetes (OR 3.63, 95%CI 2.49–5.55, p = 0.001) were
directly correlated to a TP ESE, while “rarely appropriate” indication for ESE with MHI > 2.5 (OR 0.16, 95%CI 0.11–0.22,
p < 0.0001) showed a significant inverse correlation with the outcome. During a mean follow-up of 2.5 ± 1.9 years, 299 CV
events occurred: 76.4%, 3.5% and 20.1% in Groups 1, 2 and 3, respectively. On multivariate Cox regression analysis, smoking
(HR 1.33, 95%CI 1.19–1.48), type 2 diabetes (HR 2.28, 95%CI 1.74–2.97), dyslipidemia (HR 3.51, 95%CI 2.33–5.15), betablockers
(HR 0.55, 95%CI 0.41–0.75), statins (HR 0.60, 95%CI 0.45–0.80), peak exercise average E/e′ ratio (HR 1.08, 95%CI
1.06–1.09), positive ESE (HR 3.12, 95%CI 2.43–4.01) and finally “rarely appropriate” indication for ESE with MHI > 2.5
(HR 0.15, 95%CI 0.08–0.23) were independently associated with CV events. The implementation of AUC categories with
MHI assessment may select a group of patients with extremely low probability of both TP ESE and adverse CV events over
a medium-term follow-up. A simple noninvasive chest shape assessment could reduce unnecessary exams.