Prognostic Value of Modified Haller Index in Patients with Suspected Coronary Artery Disease Referred for Exercise Stress Echocardiography.

BACKGROUND: The influence of chest conformation on outcome of patients with suspected coronary artery disease (CAD) is actually unknown.
MATERIALS AND METHODS: This retrospective study included all consecutive patients who underwent exercise stress echocardiography (ESE) for suspected CAD at our institution between February 2011 and September 2019. Modified Haller index (MHI; chest transverse diameter over the distance between sternum and spine) was assessed in all patients. Obstructive CAD was diagnosed by ≥70% stenosis in any epicardial coronary artery. During the follow-up time, we evaluated the occurrence of any of the following: (1) cardiovascular (CV) hospitalizations and (2) cardiac death or sudden death.
RESULTS: A total of 1091 consecutive patients (62.4 ± 12.6 years, 57.2% of men) were included in the study. Patients with normal chest shape (MHI ≤2.5) and those with concave-shaped chest wall (MHI >2.5) were separately analyzed. A positive ESE was diagnosed in 171 patients of which 80.7% had an obstructive CAD (true positive), while 19.3 not (false positive [FP]). Majority of FP ESE (70.9%) derived from concave-shaped chest wall group. During follow-up time (2.5 ± 1.9 years), 9 patients died and 281 were hospitalized because of heart failure (163), acute coronary syndromes (39), and arrhythmias (79). At the multivariate Cox regression analysis, age (heart rate [HR]: 1.02, 95% confidence interval [CI]: 1.01-1.03), MHI >2.5 (HR: 0.39, 95% CI: 0.26-0.56), diabetes mellitus (HR: 4.89, 95% CI: 3.78-6.32), horizontal ST depression ≥1 mm (HR: 2.86, 95% CI: 1.98-4.15), peak exercise average E/e' ratio (HR: 1.08, 95% CI: 1.06-1.10), and peak exercise wall motion score index (HR: 1.79, 95% CI: 1.36-2.35) were independently correlated with outcome.
CONCLUSIONS: Patients with concave-shaped chest wall (MHI >2.5) have a significantly lower probability of CV events than those with normal chest shape (MHI ≤2.5) over a medium-term follow-up. A noninvasive chest shape assessment could identify subjects at lower risk of CV events. Copyright:

INTRODUCTION

Exercise stress echocardiography (ESE) is a well-established and reliable technique for the diagnosis and risk stratification of patients with suspected or known coronary artery disease (CAD).[123]

Normal stress echocardiographic results are associated with an excellent prognosis.[456] On the other hand, the development of new left ventricular wall motion (LVWM) abnormalities during ESE is specific for CAD (specificity 80%–88%)[78] and is associated with an increased risk of adverse cardiac events.[91011]

ESE plays an important role not only in identifying these LVWM abnormalities in the assessment of CAD but also in the evaluation of exercise capacity, heart rate (HR), and blood pressure (BP) response to exercise.[121314]

Moreover, it is useful to quantify exercise-induced changes in left ventricular (LV) diastolic function, mitral regurgitation (MR), mean transaortic pressure gradient, and systolic pulmonary artery pressure (SPAP), providing an accurate prognostic risk stratification.[15161718]

Despite the good specificity of ESE, false-positive (FP) results may occur[1920] and a significant number of patients who undergo ESE for suspected CAD do not have cardiovascular (CV) events during the follow-up.[21]

We have recently reported that chest abnormalities may apparently affect LV functional parameters in subjects with pectus excavatum (PE)[22] and mitral valve prolapse (MVP).[23]

The influence of chest conformation on the outcome of patients with suspected CAD is actually unknown.

Accordingly, in the present study, we aimed at investigating the influence of chest shape, as noninvasively assessed by modified Haller index (MHI),[24] on the occurrence of CV events over a medium-term follow-up in a population of patients with suspected CAD who had undergone ESE.

MATERIALS AND METHODS

Patient selection

A total of 1230 consecutive patients underwent a stress echocardiographic examination for suspected CAD at San Giuseppe MultiMedica Hospital (Milan) between February 2011 and September 2019. This monocentric retrospective observational study was conducted on 1091 consecutive patients (88.7% of the total) who underwent a semisupine ESE for suspected CAD during the same period. The remaining 139 patients who had undergone dipyridamole (4.5%) and dobutamine (6.8%) stress echocardiography were excluded. The main indications for ESE were as follows: chest pain of suspected cardiac origin, resting/stress test electrocardiographic abnormalities, and/or dyspnea of unclear origin. The criteria of exclusion were as follows: history of CAD (prior myocardial infarction and/or percutaneous coronary intervention and/or coronary artery bypass graft), resting severe MR, moderate-to-severe aortic stenosis (AS), hemodynamic instability, relevant comorbidities (chronic obstructive pulmonary disease, decompensated diabetes mellitus, and uncontrolled systemic hypertension), poor echocardiographic acoustic windows, inability to perform physical exercise, and lack of consent.

A preliminary estimation of the pretest probability (PTP) of CAD was performed according to the 2019 ESC guidelines for diagnosis and management of chronic coronary syndromes.[3]

The following parameters were recorded: age, gender, body surface area (BSA), presence of CV risk factors (smoking, hypertension, type 2 diabetes, dyslipidemia, and family history of CAD), and the current medical treatment.

At first, each patient underwent resting BP measurement with a semiautomatic sphygmomanometer (Omron 705IT; Omron, Kyoto, Japan), electrocardiogram (ECG), two-dimensional (2D) transthoracic echocardiography (TTE), and MHI assessment.

ECG data included cardiac rhythm, HR, LV depolarization, and repolarization pattern.

All procedures were in accordance with the ethical standards of our Institutional Research Committee and with the 1964 Helsinki Declaration and its later amendments or comparable ethical standards. The protocol was approved by the local Ethics Committee, and the need for informed consent was not required due to the retrospective nature of this study.

Modified Haller index assessment

MHI was calculated by dividing the maximum latero-lateral external thoracic diameter (using a measuring device graduated in centimeter) by the antero-posterior (A-P) least internal thoracic diameter (measured in centimeter from the focused 2D-TTE) [Figure 1]. This approach yields values of chest shape comparable to those obtainable through chest X-rays, without exposing the subject to radiations,[24] and is routinely used at our center since February 2011.

Figure 1

Modified Haller index assessment in a patient enrolled in the study. (a) The latero-lateral maximum external thoracic diameter, measured with the subject in the standing position and with open arms, by using a rigid ruler in centimeters coupled to a level (the measuring device), placed at the distal third of the sternum, in the point of maximum depression of the sternum. Latero-lateral, latero-lateral. (b) The antero-posterior minor internal thoracic diameter, obtained with the subject in left lateral decubitus position, using the transthoracic echocardiography Philips Sparq ultrasound machine, by placing a 2.5 mHz transducer near the sternum in the left third or fourth intercostal space, to obtain a parasternal long-axis view, and measuring the distance between the true apex of the sector (the point of entry of ultrasound) and the anterior surface of the vertebral body. The vertebral body was identified by using, as reference point, the posterior wall of the descending thoracic aorta, visualized behind the left atrium. antero-posterior, antero-posterior

Exercise stress Doppler echocardiography

Patients underwent incremental exercise test on a semisupine bicycle ergometer (Ergoselect 1200 ELP; Ergoline, Bitz, Germany), with initial workload set at 25 W, increased by 25 W every 2 min.[19]

A 12-lead ECG was continuously monitored. HR, BP, and peripheral arterial oxygen saturation (SaO2) were measured at rest, every 2 min during the test, at peak exercise, and during recovery.

Exercise was interrupted when age-related maximum HR (obtained by subtracting the patient's age from 220) was reached, or in case of symptoms (limiting breathlessness, severe leg muscle fatigue, angina, dizziness, and syncope), fall in systolic BP ≥20 mmHg, complex ventricular arrhythmias, SaO2 decrease to <85%, or significant ECG changes.

Ischemic ST-segment response was defined as occurrence of ≥0.1 mV horizontal or downsloping ST-segment depression below the resting level (measured 60 ms after J-point), ≥0.2 mV upsloping depression (measured 80 ms after J-point), or ≥0.1 mV elevation.[25]

All stress echocardiograms were performed by the same cardiologist (A. S.) with specific training and experience in CV echocardiography, using a commercially available Philips Sparq ultrasound machine (Philips, Andover, Massachusetts, USA) with a 2.5 MHz transducer, on a semirecumbent cycle ergometer.

All conventional echo-Doppler measurements were performed according to the recommendations of the American Society of Echocardiography and the European Association of CV Imaging.[2627]

The following echo-Doppler parameters were collected at rest, during the 2nd min of each stage, and at peak exercise: LV ejection fraction (LVEF) measured by using modified biplane Simpson's rule,[26] LV diastolic function assessed by E/A ratio and average E/e' ratio,[27] degree of concomitant MR or AS,[28] tricuspid annular plane systolic excursion, and finally, SPAP calculated by Bernoulli equation, where SPAP = 4 × tricuspid regurgitation velocity2 + estimated right atrial pressure.[29] For each measurement, at least three cardiac cycles were averaged.

For analysis of LVWM, the left ventricle was divided into 16 segments as recommended by the American Society of Echocardiography and the European Association of CV Imaging.[26] Segments were graded based on systolic thickening and excursion: 1 = normal, 2 = hypokinetic, 3 = akinetic, and 4 = dyskinetic. The wall motion score index (WMSI) was calculated by dividing the sum of the scores of individual segments by the number of segments visualized. Change in WMSI (ΔWMSI) was calculated as the difference between peak and basal score.[19] A hypercontractile response in all wall segments was considered normal (negative ESE), regardless of exercise ECG results. Ischemia was defined as new or worsened wall motion abnormalities during stress as indicated by an increase in wall motion score ≥1 grade in ≥1 segment, or >5% decrease in LVEF at peak stress (positive ESE).[19] Ischemia was not considered present if akinetic segments at rest became dyskinetic during stress. A quad-screen format was used for comparative analysis of echocardiographic images.

Patient evaluation and follow-up

During the follow-up period, each patient underwent a clinical visit (when possible) and/or a detailed telephonic interview to detect the occurrence of adverse CV events. The latter were defisned as any of the following:

CV hospitalizations: Congestive heart failure (CHF), acute coronary syndrome (ACS), and arrhythmias associated with hemodynamic instability

Cardiac death or sudden death.

Statistical analysis

Considering the PE definition based on a Haller index value >2.5,[30] we employed the cutoff of MHI = 2.5, in order to distinguish and analyze two subgroups of subjects: those with normal chest shape (MHI ≤2.5; Group 1) and those with a concave-shaped chest wall (MHI >2.5; Group 2).

For the whole study population and for each group of patients, all continuous data were summarized as mean ± standard deviation (SD), while categorical data were presented as percentage. Continuous variables were compared by the use of an independent two-tailed t-test (for normally distributed variables) or Mann–Whitney test (for nonnormally distributed variables) while categorical variables were compared using the Chi-square test or the Fisher's exact test.

The primary endpoint of the study was to identify the parameters independently associated with the occurrence of adverse CV events during the follow-up period in our study population.

Univariate Cox proportional hazard regression analysis was performed to evaluate the effect of the main demographic, anthropometric, clinical, ECG, and echo-Doppler variables on the occurrence of the abovementioned outcome. For each variable investigated, correspondent hazard ratios with 95% confidence intervals (CIs) were calculated. Variables with P < 0.05 were then entered into a multivariate model.

However, to avoid the issue of multicollinearity, only the following variables were included in the multivariate model: age (as demographic variable), tight chest shape (defined by MHI >2.5) (as anthropometric variable), type 2 diabetes mellitus (as clinical variable), horizontal ST depression ≥1 mm (as electrocardiographic variable), peak exercise average E/e' ratio (as expression of LV diastolic function), and finally, peak exercise WMSI (as expression of LV systolic function). The remaining collinear variables were not included in the multivariate analysis.

The receiver operating characteristic (ROC) curve analysis was performed to establish the sensitivity and the specificity of the main statistically significant continuous variables for predicting CV events. Area under curve (AUC) was estimated.

The event-free survival curves of the variables statistically significant at Cox multivariate analysis were estimated using the Kaplan–Meier method, and the survival curves were compared using the log-rank test.

To evaluate intra- and interobserver variability in the assessment of the main conventional echo-Doppler parameters, the key echocardiographic variables were reassessed in a sized subgroup of 15 patients by the same cardiologist and by a second one (E. R.). The analyses were performed in a blinded manner. Both raters chose the frame on which to perform each measurement. We used the intraclass correlation coefficient (ICC) with its 95% CI as a statistical method for assessing intra- and interobserver measurement variability. An ICC of 0.70 or more was considered to indicate acceptable reliability.

Values of P < 0.05 were considered to indicate statistical significance.

Statistical analysis was performed with SPSS version 25 (SPSS Inc., Chicago, IL, USA).

RESULTS

A total of 1091 patients (mean age: 62.4 ± 12.6 years, 57.2% of men) were retrospectively analyzed.

All basal demographic, anthropometric, and clinical parameters inclusive of CV risk factors, BP values, SaO2 values, ECG data, and the current medical treatment detected in the whole study population and in the two subgroups of patients are summarized in Table 1.

Table 1

Basal demographic, anthropometric, clinical parameters, and main indications for exercise stress echocardiography detected in the whole study population and in the two subgroups of patients

	All patients (n=1091), n (%)	Group 1 patients with MHI ≤2.5 (n=734), n (%)	Group 2 patients with MHI >2.5 (n=357), n (%)	P
Demographics and anthropometrics
Age (years)	62.4±12.6	68.3±11.3	56.6±13.9	<0.0001
Male sex	624 (57.2)	485 (66.1)	139 (38.9)	<0.0001
BSA (m2)	1.88±0.15	1.90±0.15	1.86±0.15	<0.0001
MHI	2.5±0.29	2.34±0.33	2.66±0.26	<0.0001
Cardiovascular risk factors
Smoking	462 (42.3)	376 (51.2)	86 (24.1)	<0.0001
Hypertension	729 (66.8)	615 (83.8)	114 (32.0)	<0.0001
Type 2 diabetes mellitus	205 (18.8)	188 (25.6)	17 (4.8)	<0.0001
Dyslipidemia	476 (43.6)	374 (50.9)	102 (28.6)	<0.0001
Family history of CAD	155 (14.2)	141 (19.2)	14 (3.9)	<0.0001
Resting hemodynamics
SBP (mmHg)	129.6±16.5	132.2±16.7	127.0±16.3	<0.0001
DBP (mmHg)	80.6±9.0	80.3±9.4	80.9±8.6	0.31
SaO2 (%)	97.2±1.6	97.0±1.7	97.4±1.6	0.0002
HR (bpm)	71.1±12.6	69.8±12.8	72.5±12.3	0.001
Resting ECG parameters
Sinus rhythm	1055 (96.7)	698 (95.1)	357 (100)	<0.0001
AF	36 (3.3)	36 (4.9)	0 (0.0)	<0.0001
Normal intraventricular conduction	660 (60.5)	403 (54.9)	257 (72.0)	<0.0001
Left anterior fascicular block	316 (29.0)	231 (31.5)	85 (23.8)	0.01
Right bundle branch block	73 (6.7)	60 (8.2)	13 (3.6)	0.004
Left bundle branch block	27 (2.4)	25 (3.4)	2 (0.6)	<0.0001
Ventricular paced rhythm	15 (1.4)	15 (2.0)	0 (0.0)	<0.0001
Medical treatment
Antiplatelet therapy	376 (34.5)	310 (42.2)	66 (18.5)	<0.0001
Anticoagulant therapy	54 (4.9)	54 (7.3)	0 (0.0)	<0.0001
ACE-i/ARB therapy	500 (45.8)	403 (54.9)	97 (27.2)	<0.0001
Calcium channel blockers	292 (26.8)	234 (31.9)	58 (16.2)	<0.0001
Beta-blocker therapy	433 (39.7)	326 (44.4)	107 (29.9)	<0.0001
Diuretic therapy	177 (16.2)	164 (22.3)	13 (3.6)	<0.0001
Statin therapy	370 (33.9)	278 (37.9)	92 (25.8)	<0.0001
Indications for ESE
Chest pain of suspected cardiac origin	548 (50.2)	396 (53.9)	152 (42.6)	0.0005
Exclusion of CAD in the presence of resting/stress test ECG abnormalities	196 (18.0)	14 (2.0)	182 (51.0)	<0.0001
Dyspnea of unclear origin	347 (31.8)	324 (44.1)	23 (6.4)	<0.0001
PTP (%)	15.8±7.2	25.9±11.6	5.8±2.9	<0.0001

Significant P values are in bold. MHI=Modified Haller index, BSA=Body surface area, CAD=Coronary artery disease, SBP=Systolic blood pressure, DBP=Diastolic blood pressure, SaO2=Oxygen saturation, HR=Heart rate, ECG=Electrocardiogram, AF=Atrial fibrillation, ACE-i=Angiotensin-converting enzyme inhibitor, ARB=Angiotensin II receptor blocker, ESE=Exercise stress echocardiography, PTP=Pretest probability

MHI was adequately measured in all patients enrolled. In comparison to Group 1 subjects, Group 2 subjects (those with concave-shaped chest wall) were predominantly women (P < 0.0001), with significantly greater MHI value (P < 0.0001), significantly smaller BSA (P < 0.0001), and significantly lower prevalence of the common CV risk factors (all P < 0.0001). Moreover, all Group 2 subjects were found with sinus rhythm on resting ECG with a significantly higher prevalence of normal intraventricular conduction (P < 0.0001) and a significantly lower prevalence of left bundle branch block pattern (P < 0.0001) than Group 1 normal chest shape subjects. In addition, Group 2 subjects were less commonly prescribed with antiplatelets and cardioprotective drugs than Group 1 subjects (all P < 0.0001). The main indication for ESE in subjects with MHI >2.5 was the exclusion of CAD in the presence of resting/stress test ECG abnormalities (51% of cases). Finally, the PTP of CAD was significantly lower in Group 2 subjects than Group 1 subjects (5.8% ± 2.9% vs. 25.9% ± 11.6%, P < 0.0001).

Table 2 describes all resting echo-Doppler parameters measured in the study population and in both subgroups of subjects at the basal evaluation. Subjects with MHI >2.5 (Group 2) were diagnosed with significantly lower LV mass index (P < 0.0001) and significantly smaller cardiac chamber dimensions than Group 1 subjects. Resting LVEF of Group 2 subjects was significantly higher than Group 1 patients (59.6% ± 2.5% vs. 55.8% ± 7.9%, P < 0.0001). Moreover, Group 2 subjects were found with significantly lower LV filling pressures, as expressed by the estimated average E/e' ratio (7.0 ± 1.7 vs. 10.7 ± 3.8, P < 0.0001), and significantly lower SPAP (26.1 ± 3.8 mmHg vs. 29.4 ± 6.6 mmHg, P < 0.0001), than normal chest shape subjects (Group 1). Finally, in comparison to Group 1 subjects, Group 2 subjects showed a significantly higher prevalence of mild-to-moderate degenerative MR due to MVP (P < 0.0001) and a significantly lower prevalence of mild-to-moderate degenerative AS (P < 0.0001).

Table 2

Main resting echo-Doppler parameters measured in the whole study population and in the two subgroups of patients

Resting echo-Doppler parameters	All patients (n=1091), n (%)	Group 1 patients with MHI ≤2.5 (n=734), n (%)	Group 2 patients with MHI >2.5 (n=357), n (%)	P
LVEDVi (ml/m2)	46.4±8.5	49.7±8.2	43.2±8.8	<0.0001
RWT	0.40±0.06	0.42±0.06	0.39±0.05	<0.0001
LVMi (g/m2)	98.0±21.5	107.5±24.0	88.6±19.0	<0.0001
Normal LV geometry	484 (44.4)	239 (32.6)	245 (68.6)	<0.0001
Concentric remodeling	239 (21.9)	170 (23.2)	69 (19.3)	0.16
LV concentric hypertrophy	178 (16.3)	161 (21.9)	17 (4.8)	<0.0001
LV eccentric hypertrophy	190 (17.4)	164 (22.3)	26 (7.3)	<0.0001
LVEF (%)	57.7±5.2	55.8±7.9	59.6±2.5	<0.0001
WMSI	1.14±0.24	1.26±0.36	1.03±0.13	<0.0001
E/A ratio*	0.88±0.27	0.86±0.24	0.90±0.30	0.02
Average E/e’ ratio**	8.8±2.7	10.7±3.8	7.0±1.7	<0.0001
LAVi (ml/m2)	32.1±4.8	33.4±4.9	30.9±4.7	<0.0001
RVEDD (mm)	27.0±4.0	28.5±4.5	25.5±3.5	<0.0001
TAPSE (mm)	22.6±2.0	22.7±2.1	22.5±1.9	0.13
Mild-to-moderate degenerative MR	385 (35.3)	215 (29.3)	170 (47.6)	<0.0001
Mild-to-moderate degenerative AS	124 (11.4)	105 (14.3)	19 (5.3)	<0.0001
SPAP (mmHg)	27.7±5.2	29.4±6.6	26.1±3.8	<0.0001

Significant P values are in bold. *Calculated in patients with sinus rhythm, **Measured in all patients. LV=Left ventricular, MHI=Modified Haller index, LVEDVi = left ventricular end-diastolic volume index, RWT=Relative wall thickness, LVMi=Left ventricular mass index, LVEF=Left ventricular ejection fraction, WMSI=Wall motion score index, LAVi=Left atrial volume index, RVEDD=Right ventricular end-diastolic diameter, TAPSE=Tricuspid annular plane systolic excursion, MR=Mitral regurgitation, AS=Aortic stenosis, SPAP=Systolic pulmonary artery pressure

Table 3 reports all hemodynamics, ECG changes, and symptoms recorded at peak exercise in the whole study population and in the two subgroups of patients. BP values showed a physiological response to dynamic exercise. The mean percentage of HR maximum reached (76.7% ± 10.8%) was close to submassimal. Group 2 subjects showed the best exercise tolerance (in terms of Watts and double product reached). Furthermore, we observed a significantly higher prevalence of exercise-induced isolated ventricular premature beats, upsloping ST depression ≥2 mm, and symptoms as atypical chest pain and palpitations in Group 2 subjects than Group 1 subjects (all P < 0.0001).

Table 3

Main hemodynamics, electrocardiogram changes, and symptoms detected at peak exercise in the whole study population and in the two subgroups of patients

	All patients (n=1091), n (%)	Group 1 patients with MHI ≤2.5 (n=734), n (%)	Group 2 patients with MHI >2.5 (n=357), n (%)	P
Hemodynamics at peak exercise
Watts reached	101.1±35.0	92.5±31.5	109.7±38.5	<0.0001
Peak exercise SaO2 (%)	96.7±1.9	96.3±2.2	97.1±1.5	<0.0001
∆SaO2 (%)	−0.5±1.7	−0.7±2.0	−0.3±1.4	0.0007
Peak exercise SBP (mmHg)	172.5±24.3	171.6±23.6	173.4±25.0	0.25
∆SBP (mmHg)	42.9±21.1	39.4±21.2	46.4±21.1	<0.0001
Peak exercise DBP (mmHg)	88.9±12.7	87.8±12.4	90.0±13.1	0.007
∆DBP (mmHg)	8.3±10.5	7.5±10.2	9.1±10.8	0.02
Peak exercise HR (bpm)	121.5±20.0	114.3±19.8	128.8±20.3	<0.0001
Percentage of HR maximum reached (%)	76.7±10.8	75.0±11.6	78.5±10.1	<0.0001
DP (mmHg × bpm)	21088.2±5015.5	19773.26±4972.4	22403.19±5058.7	<0.0001
ECG parameters at peak exercise
No arrhythmias	409 (37.5)	279 (38.0)	130 (36.4)	0.64
Isolated SVPBs	206 (18.8)	170 (23.2)	36 (10.1)	<0.0001
Isolated VPBs	398 (36.4)	218 (29.7)	180 (50.4)	<0.0001
SVT	42 (3.8)	33 (4.5)	9 (2.5)	0.13
AF	27 (2.5)	26 (3.5)	1 (0.3)	0.0005
NSVT	11 (1.0)	10 (1.4)	1 (0.3)	0.11
No ST-segment changes	684 (62.7)	475 (64.7)	209 (58.6)	0.05
Upsloping ST depression ≥2 mm	175 (16.0)	87 (11.8)	88 (24.6)	<0.0001
Downsloping ST depression ≥1 mm	154 (14.1)	110 (15.0)	44 (12.3)	0.26
Horizontal ST depression ≥1 mm	78 (7.2)	62 (8.5)	16 (4.5)	0.02
Exercise-induced symptoms
No symptoms	436 (40.0)	348 (47.4)	88 (24.6)	<0.0001
Dyspnea	292 (26.8)	212 (28.9)	80 (22.5)	0.02
Typical chest pain	121 (11.0)	115 (15.7)	6 (1.7)	<0.0001
Atypical chest pain	70 (6.4)	15 (2.0)	55 (15.4)	<0.0001
Palpitations	163 (15.0)	35 (4.8)	128 (35.8)	<0.0001
Syncope	9 (0.8)	9 (1.2)	0	0.03

Significant P values are in bold. Δ: Absolute difference between peak exercise and rest data. MHI=Modified Haller index, SaO2=Oxygen saturation, SBP=Systolic blood pressure, DBP=Diastolic blood pressure, HR=Heart rate, DP=Double product, ECG=Electrocardiogram, VPBs=Ventricular premature beats, SVPBs=Supra-VPBs, SVT=Supraventricular tachycardia, AF=Atrial fibrillation, NSVT=Nonsustained ventricular tachycardia

The main echo-Doppler data obtained at peak exercise in the study population and in both subgroups of subjects are reported in Table 4. All subgroups showed a physiological hyperkinetic response to exercise. However, Group 2 subjects were found with significantly higher peak exercise LVEF (66.8% ± 5.0% vs. 60.2% ± 11.2%, P < 0.0001) and significantly lower prevalence of exercise-induced myocardial asinergies (P < 0.0001) than Group 1 subjects. Therefore, the prevalence of positive ESE was significantly higher in Group 1 subjects than Group 2 subjects (P < 0.0001). Majority of subgroups of subjects were diagnosed without a significant exercise-induced increase in MR degree. However, Group 1 subjects showed significantly higher peak exercise LVFP assessed by average E/e' ratio (P < 0.0001) and peak exercise SPAP (P < 0.0001) in comparison to Group 2 subjects.

Table 4

Echocardiographic data obtained at peak exercise in the whole study population and in the two subgroups of patients

Echo-Doppler data obtained at peak exercise	All patients (n=1091), n (%)	Group 1 patients with MHI ≤2.5 (n=734), n (%)	Group 2 patients with MHI >2.5 (n=357), n (%)	P
Peak exercise LVEF (%)	63.5±8.1	60.2±11.2	66.8±5.0	<0.0001
∆LVEF (%)	5.8±5.1	4.4±6.4	7.3±3.9	<0.0001
Peak WMSI	1.12±0.11	1.22±0.39	1.03±0.14	<0.0001
∆WMSI	−0.02±0.18	−0.04±0.24	−0.002±0.12	0.005
No exercise-induced myocardial asinergies	802 (73.5)	470 (64.0)	332 (93.0)	<0.0001
Dyssynergy in the LAD territory	129 (11.8)	124 (16.9)	5 (1.4)	<0.0001
Dyssynergy in the RCA territory	88 (8.1)	83 (11.3)	5 (1.4)	<0.0001
Dyssynergy in the LCx territory	47 (4.3)	33 (4.5)	14 (3.9)	0.75
Multivessel distribution	25 (2.3)	24 (3.3)	1 (0.3)	<0.0001
Positive ESE	171 (15.7)	140 (19.1)	31 (8.7)	<0.0001
Peak exercise E/A ratio*	1.05±0.32	1.08±0.36	1.02±0.28	0.006
Normal diastolic function*	276 (26.0)	80 (11.5)	196 (55.0)	<0.0001
Grade 1 diastolic dysfunction*	485 (46.0)	352 (50.4)	133 (37.2)	0.0009
Pseudonormalization of the E/A ratio*	296 (28.0)	268 (38.1)	28 (7.8)	<0.0001
Peak exercise average E/e’ ratio**	11.8±5.5	15.8±7.2	7.8±3.8	<0.0001
Exercise-induced severe MR	84 (7.7)	74 (10.1)	10 (2.8)	<0.0001
Exercise-induced severe AS	33 (3.0)	33 (4.5)	0 (0.0)	<0.0001
Peak exercise TAPSE (mm)	28.6±5.6	28.7±5.4	28.5±5.8	0.57
Peak exercise SPAP (mmHg)	43.2±10.4	46.5±11.9	39.9±8.9	<0.0001

Significant P values are in bold. *Calculated in patients with sinus rhythm, **Measured in all patients. Δ: Absolute difference between peak exercise and rest data. MHI=Modified Haller index, LV: Left ventricular, LVEF=LV ejection fraction, WMSI=Wall motion score index, LAD=Left anterior descending, RCA=Right coronary artery, LCx=Left circumflex, ESE=Exercise stress echocardiography, MR=Mitral regurgitation, AS=Aortic stenosis, TAPSE=Tricuspid annular plane systolic excursion, SPAP=Systolic pulmonary artery pressure

Coronary angiography results

One hundred and seventy-one patients (15.7% of total) were diagnosed with positive ESE. A subsequent coronary angiography revealed that 138 patients (80.7%) had an obstructive CAD, while 33 (19.3%) did not (FP ESE). One hundred and eight patients (78.3%) underwent percutaneous coronary intervention, while the remaining 30 (21.7%) were treated with coronary artery bypass grafting. Majority of patients diagnosed with a FP ESE (70.9%) derived from Group 2 subjects.

Predictors of outcome

The mean follow-up time was 2.5 ± 1.9 years. During follow-up, no patients were lost and 290 adverse CV events were recorded: 9 patients died and 281 patients were hospitalized because of CHF (163 patients), ACS (39 patients), and arrhythmias associated with hemodynamic instability (79 patients).

Table 5 reports uni- and multivariate Cox regression analysis, conducted in order to identify the main demographic, anthropometric, clinical, electrocardiographic, and echo-Doppler variables independently associated with the outcome.

Table 5

Univariate and multivariate Cox regression analysis for detecting the main demographic, anthropometric, hemodynamic, electrocardiographic, and echocardiographic indices independently associated with the occurrence of a cardiovascular event during the follow-up period

Variables	Univariate cox regression analysis			Multivariate cox regression analysis
	HR	95% CI	P	HR	95% CI	P
Age	1.05	1.04-1.06	<0.0001	1.02	1.01-1.03	0.006
Male sex	0.98	0.78-1.23	0.87
Concave-shaped chest wall (MHI >2.5)	0.23	0.16-0.34	<0.0001	0.39	0.26-0.56	<0.0001
Hypertension	2.31	1.74-3.05	<0.0001
Type 2 diabetes mellitus	9.12	7.24-11.5	<0.0001	4.89	3.78-6.32	<0.0001
LVMi	1.02	1.01-1.03	<0.0001
Percentage of HR maximum reached	0.98	0.97-0.99	0.003
Exercise-induced PVBs	1.04	0.81-1.34	0.76
Horizontal ST depression ≥1 mm	5.96	4.47-7.93	<0.0001	2.86	1.98-4.15	<0.0001
Peak exercise average E/e’ ratio	1.13	1.11-1.14	<0.0001	1.08	1.06-1.10	<0.0001
Peak exercise WMSI	5.21	4.19-6.47	<0.0001	1.79	1.36-2.35	<0.0001
Peak exercise SPAP	1.05	1.04-10.6	<0.0001

Significant P values are in bold. HR=Heart rate, MHI=Modified Haller index, LV: Left ventricular, LVMi=LV mass index, PVBs=Premature ventricular beats, WMSI=Wall motion score index, SPAP=Systolic pulmonary artery pressure, CI=Confidence interval

The univariate Cox proportional hazard ratio analysis, performed in the whole population of patients enrolled in the study, showed several clinical, ECG, and echo-Doppler variables that were strongly associated with the occurrence of CV events during the follow-up period. Interestingly, a concave-shaped chest wall (defined by MHI >2.5) showed a statistically significant inverse correlation with the outcome (HR: 0.23, 95% CI: 0.16–0.34, P < 0.0001). At the multivariate Cox regression analysis, age (HR: 1.02, 95% CI: 1.01–1.03, P = 0.006), MHI > 2.5 (HR: 0.39, 95% CI: 0.26–0.56, P < 0.0001), type 2 diabetes mellitus (HR: 4.89, 95% CI: 3.78–6.32, P < 0.0001), horizontal ST depression ≥1 mm (HR: 2.86, 95% CI: 1.98–4.15, P < 0.0001), peak exercise average E/e' ratio (HR: 1.08, 95% CI: 1.06–1.10, P < 0.0001), and peak exercise WMSI (HR: 1.79, 95% CI: 1.36–2.35, P < 0.0001) retained statistical significance.

The ROC curve analysis revealed the following cutoff for age (≥70 years, 70% sensitivity and 69% specificity, AUC = 0.75), peak exercise average E/e' ratio (≥13, 81% sensitivity and 80% specificity, AUC = 0.87), and peak exercise WMSI (≥1.5, 86% sensitivity and 71% specificity, AUC = 0.86), as the cutoff values with maximum sensitivity and specificity for the detection of CV events.

The Kaplan–Meier event-free survival curves, obtained for the six strongest independent predictors of CV events, are depicted in Figure 2.

Figure 2

The Kaplan–Meier event-free survival curves, obtained for the six variables with the strongest independent correlation to CV events. CV=Cardiovascular, MHI=Modified Haller index, WMSI=Wall motion score index

Measurement variability

A detailed intra- and interobserver variability analysis of the main conventional echo-Doppler parameters, measured at rest and at peak exercise in a group of 15 randomly selected subjects, is reported in Supplementary Table 1. Intra- and interobserver agreement between the raters, expressed as ICCs with 95% CIs, ranged from 0.77 to 0.93 and from 0.76 to 0.90, respectively.

Supplementary Table 1

Intra- and interobserver variability analysis of the main conventional echo-Doppler parameters, measured at rest and at peak exercise in a subgroup of 15 randomly selected subjects. CI=Confidence interval, ICC=Intraclass correlation coefficient, LVEF=Left ventricular ejection fraction, SPAP=Systolic pulmonary artery pressure, TAPSE=Tricuspid annular plane systolic excursion

INTRA-AND INTER-OBSERVER VARIABILITY ANALYSIS.
PATIENT LIST	LVEF (%)						Average E/e’ ratio
	REST			PEAK STRESS			REST			PEAK STRESS
	INITIAL MEASUREMENT	REMEASUREMENTS		INITIAL MEASUREMENT	REMEASUREMENTS		INITIAL MEASUREMENT	REMEASUREMENTS		INITIAL MEASUREMENT	REMEASUREMENTS
		Rater 1	Rater 2		Rater 1	Rater 2		Rater 1	Rater 2		Rater 1	Rater 2
1) S. E.	62	60	62	65	66	67	7,5	7,7	7,9	8,2	7,7	8,4
2) R. F.	60	60	59	66	66	68	8,1	7,2	9,2	15,7	15,5	16,1
3) C. F.	63	63	62	64	64	62	7,4	7,8	7,6	16,6	14,5	17,2
4) B. T.	61	61	62	65	65	62	12,3	10,2	14,1	15,8	15,5	21,5
5) P. G.	60	62	61	68	67	66	7,4	7,9	8,5	15,1	16,5	17,5
6) M. E.	62	62	61	67	65	65	14,1	11,3	9,5	16,1	19,2	19,2
7) F. ML	60	60	58	65	65	63	7,3	6,1	7,8	21,9	16,2	16,5
8) B. G.	62	62	60	64	65	61	7,9	7,1	7,7	19,6	16,8	18,5
9) E. B	60	60	60	67	66	66	5,58	7,2	5,49	5,7	6,1	7,8
10) S. M.	62	62	63	66	66	66	6,6	5,7	5,8	7,9	6,7	6,8
11) M. R.	64	64	64	68	68	68	6,35	5,6	6,5	6,57	6,62	6,1
12)J. E.	60	60	60	68	68	68	4,37	6,2	4,8	6,7	7,6	7,2
13) A. P.	61	61	61	67	65	66	5,88	6,7	7,6	6,05	6,3	5,7
14) S. S.	63	63	63	66	66	66	7,22	6,1	7,4	7,72	6,1	8,2
15) G. V.	60	60	60	65	65	62	4,45	5,5	4,8	7,32	6,8	7,8
ICC (95% CI)		0,84 (0,59 -0,94)	0,8 (0,48-0,95)		0,77 (0,45-0,92)	0,76 (0,42-0,91)		0,81 (0,53-0,93)	0,81 (0,53-0,93)		0,93 (0,80-0,97)	0,9 (0,75-0,96)
	TAPSE (mm)						SPAP (mmHg)
	REST			PEAK STRESS			REST			PEAK STRESS
	PATIENT LIS	INITIAL MEASUREMENT	REMEASUREMENTS		INITIAL MEASUREMENT	REMEASUREMENTS		INITIAL MEASUREMENT	REMEASUREMENTS		INITIAL MEASUREMENT	REMEASUREMENTS
	Rater 1	Rater 2	Rater 1	Rater 2	Rater 1	Rater 2	Rater 1	Rater 2
1) S. E.	24	23	25	16	14	17	29	28	28	53	55	52
2) R. F.	16	17	16	15	14	16	32	30	30	60	58	55
3) C. F.	21	20	24	17	16	14	23	26	28	50	54	55
4) B. T.	23	22	26	17	14	15	25	24	30	53	55	50
5) P. G.	22	25	22	28	26	20	26	27	28	49	44	55
6) M. E.	26	30	26	17	15	13	26	27	26	47	44	55
7) F. ML	19	17	19	12	13	10	32	30	33	63	66	69
8) B.G.	21	22	25	20	22	24	29	26	28	41	48	45
9) E. B.	22	24	22	31	24	23	19	23	24	43	37	46
10) S. M.	25	27	25	29	24	23	20	22	25	31	30	37
11) M. R.	31	28	24	32	25	27	18	20	18	32	35	30
12) J. E.	21	24	23	31	30	26	22	24	22	37	39	30
13) A. P.	30	26	33	34	22	28	24	26	24	50	57	53
14) S. S.	23	22	24	33	26	22	33	30	28	38	32	32
15) G. V.	23	20	24	32	27	25	21	20	24	31	28	35
ICC (95% CI)		0,78 (0,47-0,93)	0,77 (0,44-0,91)		0,86 (0,63-0,95)	0,83 (0,57-0,94)		0,85 (0,62-0,94)	0,76 (0,42-0,91)		0,92 (0,78-0,93)	0,89 (0,72-0,96)

DISCUSSION

Noninvasive functional imaging for myocardial ischemia is recommended by the 2019 ESC guidelines for diagnosis and management of chronic coronary syndromes as the initial test to diagnose CAD in symptomatic patients in whom obstructive CAD cannot be excluded by clinical assessment alone (recommendation class I, level of evidence B).[31323334]

The Cardiology Department of San Giuseppe Hospital, in Milan, Italy, is known for its local expertise in stress echocardiography. Hence, the latter is the preferred method for noninvasive functional imaging at this center, and exercise is the preferred stressor used for the evaluation of patients with suspected CAD. The reasons there is a center-wide preference for ESE are because the cardiology team is experienced in performing it; in addition, among the different stress echocardiography modalities, exercise is safer, more physiologic, and less expensive than other noninvasive imaging techniques.[3536] Moreover, ESE has relatively high diagnostic accuracy[7819] and high negative predictive value.[3738] Nevertheless, it is a “green technology,” allowing to avoid exposure to high radiation levels of unnecessary coronary computed tomography scan and/or coronary angiography.[39]

Previous studies demonstrated that several echo-Doppler parameters, assessed by ESE, had a significant prognostic relevance both in ischemic and nonischemic heart disease. The main ESE-derived negative prognostic markers include (1) the site, extent, and severity of new LVWM abnormalities;[40] (2) exercise-induced global dilatation of the LV cavity;[41] (3) low resting LVEF;[42] (4) absence of LV contractile reserve;[43] (5) increasing peak WMSI;[42] (6) time taken for LVWM abnormalities to normalize;[44] (7) achieving submaximal percentage of the predicted maximum HR, even with an otherwise normal ESE;[45] (8) an exercise-induced increase in average E/e' ratio >2 SD above normal;[4647] (9) a maximum exercise-induced LV outflow tract obstruction gradient of ≥50 mmHg in patients with hypertrophic cardiomyopathy;[48] and (10) an exercise-induced pulmonary hypertension (SPAP ≥60 mmHg)[16] and the presence of exercise-induced right ventricular dysfunction[49] in patients with primary MR.

To the best of our knowledge, no previous authors considered the chest wall conformation as a possible prognostic determinant over a medium-term follow-up in patients who had undergone ESE for suspected CAD.

In the present study, the MHI, as anthropometric index, was evaluated together with the main conventional demographic, clinical, ECG, and echo-Doppler parameters, as a possible additive factor independently correlated with the prevalence of adverse CV events during the follow-up period.

We retrospectively analyzed a cohort of 1091 consecutive subjects who underwent a semisupine ESE for excluding ischemic heart disease at our institution between February 2011 and September 2019.

The majority of patients did not show any significant impairment in hemodynamics, any significant arrhythmias, any significant ST-segment changes, and/or limitantsymptoms during exercise. Therefore, ESE was substantially well tolerated in our study population.

Subgroup analysis revealed that subjects with concave-shaped chest wall (MHI >2.5) were predominantly women with small BSA and low prevalence of the common CV risk factors. These patients were usually referred to the echo laboratory for excluding CAD in the presence of resting/stress test ECG abnormalities. They were diagnosed with small cardiac chamber dimensions and/or normal LV geometry and with normal LV systolic and diastolic function. A high prevalence of mild-to-moderate degenerative MR due to MVP was detected in these subjects. Moreover, patients with MHI >2.5 showed an optimal exercise tolerance and a high prevalence of exercise-induced isolated ventricular premature beats, of upsloping ST depression ≥2 mm, and of symptoms such as atypical chest pain and palpitations. In addition, they were found with a significantly lower prevalence of both positive ESE and adverse CV events during the follow-up period in comparison to subjects with normal chest shape. Conversely, the prevalence of true ischemia and CV events was far higher in patients with normal chest conformation (MHI ≤2.5).

Multivariate Cox regression analysis confirmed that traditional demographic (age), clinical (type 2 diabetes mellitus), ECG (horizontal ST depression ≥1 mm), and echo-Doppler (peak exercise average E/e' ratio and peak exercise WMSI) parameters were independently correlated with a higher prevalence of CV events during the follow-up period in patients who had undergone ESE for suspected CAD. Originally, we demonstrated that the patient's chest wall conformation, as assessed by MHI value, might be related to a different prevalence of CV events during the follow-up period. Notably, a concave-shaped chest wall (defined by MHI >2.5) showed a strong negative correlation with the abovementioned outcome. Kaplan–Meier survival analysis revealed that subjects with MHI >2.5 had a significantly lower probability of CV events during the follow-up period in comparison to subjects with normal chest conformation (MHI ≤2.5).

Chest shape assessment by MHI might contribute to clarify the mechanisms responsible for exercise-induced electrocardiographic and echocardiographic changes which often occur in subjects with PE and MVP. In particular, the continuous mechanical stress by the thoracic deformity may have a preponderant role in generating deformities of the mitral annulus, MVP, and MR.[50] Consistent with previous studies,[515253] the present study confirmed the high incidence (47.6%) of concomitant MVP in subjects with concave-shaped chest wall and/or PE. Moreover, a narrow A-P chest diameter might have a prominent or concurrent role in inducing a dynamic LV dyssynchrony, accentuated by exercise, secondary to compressive phenomena rather than intrinsic myocardial dysfunction or true ischemia.[54] The abovementioned LV dyssynchrony may be wrongly interpreted as regional myocardial hypokinesis or dyskinesis due to obstructive CAD also by expert echocardiographers. Therefore, subjects with MHI >2.5 have a significantly increased probability to be diagnosed with a FP ESE. This is the main reason explaining the significantly low probability of CV events over a medium-term follow-up in comparison to subjects with normal chest shape. In other terms, subjects with MHI >2.5, small BSA, and low CV risk profile have an excellent prognosis not only due to the low PTP but also due to the chest shape conformation which facilitates LV dyssynchrony during exercise, easily interpreted as pathological dyssynergy even by experts.

A pretest chest shape assessment may contribute to select and identify patients with a lower probability of exercise-induced inducible myocardial ischemia and with a lower probability of CV events. Notably, both a low PTP and a MHI >2.5 might suggest the clinician to avoid ESE because of the possible influence of a narrow A-P chest diameter on the evaluation of kinetic wall motion during exercise in these subjects and because of the excellent prognosis of these subjects.

Limitations

The main limitation of the present study is its retrospective nature. Second, we analyzed only patients who had undergone ESE, and data from pharmacological stress echocardiography were not collected. However, this decision was in agreement with the current guidelines[3] that recommend ESE as the first choice for patients who are able to perform a physical exercise, because it is safer, more physiologic, and less expensive than other noninvasive imaging techniques.

Moreover, our study contains the limitations inherent to the ESE technique, such as patient motion, hyperventilation, and chest wall movement that can impair the image quality and can complicate the acquisitions of the same images at each stage. In addition, maximum HR reached was close to submaximal, probably because some patients did not adequately discontinue beta-blockers. Finally, we did not use a MET-based assessment of exercise capacity, not available in our echo laboratory.

CONCLUSIONS

Chest shape conformation may influence the prevalence of adverse CV events in patients who undergo ESE for excluding CAD.

The present study demonstrated that subjects with concave-shaped chest wall (MHI >2.5) had a significantly lower prevalence of CV events in comparison to those subjects with normal chest shape (MHI ≤2.5), over a medium-term follow-up.

A noninvasive chest shape assessment could be added to the preliminary evaluation of patients with suspected CAD referred to echo laboratory for performing ESE. This innovative approach could be very helpful to avoid unnecessary examinations.

Ethical clearance

CE-24.2021.

Financial support and sponsorship

Nil.

Conflicts of interest

There are no conflicts of interest.