The International Journal of Cardiovascular Imaging

https://doi.org/10.1007/s10554-020-01913-6

ORIGINAL PAPER

Long‑term prognosis and predictors of outcomes after negative stress

echocardiography
Rayan Jo Rachwan1 · Fakilahyel S. Mshelbwala1 · Rody G. Bou Chaaya1 · Edward A. El‑Am1 · Mohammad Sabra1 ·
Zeina Dardari2 · Ziad A. Jaradat3 · Omar Batal3

Received: 1 May 2020 / Accepted: 4 June 2020

© Springer Nature B.V. 2020

Abstract
Negative stress echocardiography (NSE) is associated with low cardiovascular morbidity and overall mortality. We aimed
to determine the clinical and echocardiographic predictors of overall and cardiovascular outcomes following NSE. Patients
who underwent SE between 2013 and 2017 were reviewed. Patients with a history of solid organ transplant or being evalu-
ated for transplant, history of end-stage renal or liver disease, and positive SE were excluded. NSE results were divided into
negative diagnostic if patient reached target heart rate (THR) and had no wall motion abnormality (WMA) at rest or stress;
negative non-diagnostic if patient had no WMA but did not reach THR or if image quality was non-diagnostic; and abnormal
non-ischemic if patient had a resting WMA not worsened at stress along with a personal history of coronary artery disease
(CAD). New CAD lesion at 1 year was defined as ≥ 50% stenosis on cardiac catheterization. Of 4119 patients with SE, 2575
were included. All-cause mortality rate was 1.1%/year and CAD rate was 3.1%/year. Predictors of all-cause mortality were
age, male gender, history of smoking and being selected for dobutamine SE. Predictors of a new CAD lesion at 1 year were
male gender, diabetes, personal history of CAD and abnormal non-ischemic SE. We identified clinical and echocardiographic
characteristics in a subset of NSE patients who are at higher risk for subsequent adverse events. These characteristics should
be accounted for during the clinical interpretation of SE, and patients found at increased risk for morbidity and mortality
warrant continued follow-up.

Keywords Stress echocardiography · Coronary artery disease · Outcomes

Introduction results [2–5]. Overall mortality, as well as cardiac mortal-

ity in patients with negative SE (NSE), have been reported
Stress echocardiography (SE) is a well-validated tool for to be < 1%/year and 0.5–0.7%/year, respectively [6, 7]. As
the diagnosis of coronary artery disease (CAD). The accu- a result, patients with NSE are usually reassured by their
racy of SE in detecting significant CAD (≥ 50% coronary physicians and require no additional testing. There have been
luminal stenosis) is reported to be 80–90% when performed few large studies that addressed the prognostic role of NSE
by experienced operators [1]. SE has also been shown to over a long-term follow-up period. Furthermore, traditional
have a prognostic role in predicting mortality, with earlier estimates of pretest probability appeared to over-estimate the
studies suggesting favorable outcomes following negative prevalence of CAD in a contemporary cohort and thus pre-
dicting a higher prevalence of negative stress tests [8]. This
study aimed to determine the clinical and echocardiographic
* Omar Batal
[email protected] predictors of new CAD lesion at 1 year following SE and of
all-cause mortality in a large cohort of patients with NSE
1
Department of Medicine, Indiana University School over a period of 3.2 ± 1.2 years.
of Medicine, Indianapolis, IN, USA
2
Ciccarone Center for the Prevention of Heart Disease, Johns
Hopkins School of Medicine, Baltimore, MD, USA
3
Division of Cardiology, Department of Medicine, Indiana
University School of Medicine, 1801 North Senate
Boulevard, Indianapolis, IN, USA

13
Vol.:(0123456789)
 The International Journal of Cardiovascular Imaging

Methods that could interact with the outcomes of the study (such
as beta-blockers) were discontinued prior to the stress
Patient selection and outcomes test unless instructed otherwise by the referring physi-
cian. Patients’ heart rates and rhythms were continuously
The methods and results of our initial study involving the monitored using an electrocardiogram. Blood pressure was
same patient cohort have been described in detail else- obtained at rest and after every stage of stress. SE was
where [9]. We reviewed all SE tests performed in the terminated if the patient achieved target heart rate (THR),
outpatient setting between January 2013 and December experienced severe chest pain or developed ST-segment
2017 at Indiana University Methodist Hospital (N = 4119). elevation or hemodynamic abnormalities; these include
Patients with a history of heart transplant (N = 50), being hypotension (defined as a drop in systolic blood pressure
evaluated for solid organ transplants (N = 1016), and of > 10 mmHg from baseline blood pressure despite an
patients with a history of end-stage renal disease (N = 101) increase in workload), hypertension (defined as a systolic
or end-stage liver disease (N = 94) were excluded. We then blood pressure ≥ 250 mmHg and/or a diastolic blood pres-
excluded those with positive SE results (N = 283, 9.9%). sure ≥ 115 mmHg) and worrisome tachyarrhythmia. SE
In the event a patient had more than one SE during the test was considered diagnostic if patients reached at least
study period, only the first test was selected for analysis. 85% of the age-predicted maximal heart rate (defined as
Data collection included demographics (age, gender, race, 220—age). SE images were reviewed and interpreted by
and body mass index), cardiac risk factors (hypertension, cardiologists who are level II or level III-trained in echo-
diabetes mellitus, hyperlipidemia, previous or current cardiography. Left ventricular ejection fraction was deter-
history of smoking, family history of CAD, and personal mined by the Simpson method or by visual assessment.
history of CAD), SE results and outcomes (overall mortal- Positive SE result was defined as the presence of new
ity, 1-year mortality, and new CAD lesion at 1 year). In or worsening of wall motion abnormality (WMA) post-
patients with a new CAD lesion at 1 year, we determined stress. NSE results were divided into negative diagnostic
if acute coronary syndrome (ACS) was diagnosed. The if patient reached THR (85% of the age-predicted maximal
study was approved by the Indiana University Institutional heart rate) and had no WMA at rest or stress; negative
Review Board. No consent was required given the retro- non-diagnostic if patient had no WMA but did not reach
spective nature of the study. THR or if image quality was reported as non-diagnostic
New CAD lesion at 1 year was defined as the presence per the interpreting physician; and abnormal non-ischemic
of a new obstructive lesion (≥ 50% lesion in ≥ 1 epicardial if patient had a known personal history of CAD and a rest-
coronary arteries or major branches) on cardiac catheteri- ing WMA not worsened post-stress.
zation. The occurrence of ACS among patients with a new
CAD lesion at 1 year was determined by reviewing the
charts of patients with International Classification of Dis- Statistical analysis
ease (ICD)-9 and ICD-10 codes consistent with ACS and/
or with elevated troponin levels (defined as > 0.1 ng/ml) Categorical variables were presented as total number and
and findings were confirmed by a consultant cardiologist percentages, and continuous variables as mean ± stand-
who saw the patient at that time. Mortality was determined ard deviation. Clinical and echocardiographic charac-
by reviewing newspaper obituaries and medical databases. teristics were compared using chi-square analysis for
Only all-cause mortality was surveyed. categorical variables and one-way Anova for normally
distributed variables. In the event a continuous variable
was not normally distributed, tests for comparison were
Echocardiogram image acquisition and analysis done using Kruskal–Wallis. Univariate logistic and Cox
regression analyses were performed to investigate the
Patients underwent either exercise (treadmill or bicycle) association between variables and outcomes. Final mod-
or dobutamine SE. Two-dimensional SE was performed by els were derived in a stepwise fashion using variables
trained sonographers according to standardized protocols with probability to enter of < 0.1 on univariate analy-
[10]. Images were obtained with patients in the left lateral sis. Kaplan–Meier curves were constructed for all-cause
decubitus position and included apical (two-chamber and mortality and were stratified according to the independ-
four-chamber) and parasternal (short-axis and long-axis) ent categorical predictors of all-cause mortality obtained
echocardiographic views according to the recommenda- on multivariate analysis. Statistical significance was set
tions of the American Society of Echocardiography (ASE); at p < 0.05. Statistical testing was performed using Stata
examples of which are shown in Fig. 1 [10]. Medications version 13 (StataCorp, College Station, TX, USA).

13
The International Journal of Cardiovascular Imaging

Fig. 1  Transthoracic echocardiography images were obtained in each chamber views at rest (top row) and stress (bottom row). PLAX par-
patient using a parasternal long-axis and short-axis views at rest (top asternal long-axis, PSAX parasternal short-axis, 2C apical two-cham-
row) and stress (bottom row), and b apical two-chamber and four- ber, 4C apical four-chamber

Results and 6% had a family history of CAD. Common clinical

indications for SE included: the presence of chest pain
Clinical and echocardiographic characteristics (53%), dyspnea (17%), and previous history of CAD (8%).
of the total population The modality of stress testing was exercise SE in 67%
and dobutamine SE in 33%. The mean ejection fraction
A total of 2575 patients met the inclusion criteria and was 60 ± 7%, with 3.4% of the patients having ejection
constituted our study population. Clinical and echo- fractions less than 50%. Seventy-three percent of patients
cardiographic characteristics of the study population reached THR. SE results were normal in 77%, non-diag-
were summarized in Table 1. The cohort mean age was nostic in 17%, and abnormal non-ischemic in 6%. Techni-
59 ± 11 years, body mass index 31 ± 8 kg/m 2, with 54% cal quality was reported on 2258 patients (88%). Of those
females and 70% identified as white. Regarding cardiac stress tests, 63% were of fair/adequate quality; while 18%
risk factors, 35% were diabetic, 67% had hypertension, and 19% were of good and poor quality, respectively. Of
61% had hyperlipidemia, 38% had a prior or current his- poor quality studies, echocardiographic contrast was listed
tory of tobacco use, 38% had a personal history of CAD, to be used in 37% of cases.

13
 The International Journal of Cardiovascular Imaging

Table 1  Descriptive statistics of clinical and stress test variables for selected to undergo dobutamine rather than exercise SE
the total population and patients dead at 1 year (1.6% death at 1 year), the presence of a resting WMA at
Variable Total (N = 2575) Dead at baseline (1.9% death at 1 year), and an ejection fraction
1 year < 50% (2.2% death at 1 year).
(N = 29)

Age (years) 59 ± 11 67 ± 11 Clinical and echocardiographic characteristics

Age ≥ 60 years 1220 (43%) 23 (79%) of NSE patients with new CAD lesion at 1 year
Gender
Male 1180 (46%) 20 (69%) A total of 220 patients with NSE (8.5% of total) underwent
Female 1395 (54%) 9 (31%) cardiac catheterization within 1 year from the index SE, of
Body mass index (kg/m2) 31 ± 8 29 ± 5 whom 80 patients (3.1% of total) had a new CAD lesion.
Race Furthermore, 32 patients of the total population (1.2%) had
White 1812 (70%) 22 (76%) an ACS event at 1 year. Clinical and echocardiographic
Black 677 (26%) 7 (24%) characteristics of the 1-year CAD subset were summarized
Other ethnicities 86 (4%) 0 (0%) in Table 2. When compared to the total population, a new
Diabetes mellitus 913 (35%) 11 (38%) obstructive coronary lesion was significantly (p < 0.05)
Hypertension 1714 (67%) 18 (62%) more likely in patients who were older (4.4% of patients
Hyperlipidemia 1566 (61%) 15 (52%) ≥ 60 years of age), men (4.6% of total), white (3.8% of total),
Prior or current smoking 985 (38%) 15 (52%) and with a higher body mass index. These patients were
Family history of CAD 167 (6%) 0 (0%) also more likely to be diabetic (5.1% of total), hyperten-
Personal history of CAD 678 (26%) 10 (34%) sive (4% of total), hyperlipidemic (4.2% of total), and to
SE type have a personal history of CAD (8.4% of total), a family
Dobutamine 852 (33%) 14 (48%) history of CAD (6% of total) and a prior or current history
Exercise 1723 (67%) 15 (52%) of smoking (4.2% of total). Patients with a non-diagnostic
NSE result or an abnormal non-ischemic SE were more likely to have
Normal 1979 (77%) 22 (76%) a new obstructive coronary lesion at 1 year (4.8% and 11%,
Non-diagnostic 441 (17%) 4 (14%) respectively).
Abnormal non-ischemic 155 (6%) 3 (10%) The clinical and stress test variables associated with new
Rest variables CAD lesion at 1 year on univariate analysis are shown in
EF (%) 60 ± 7 60 ± 13 Table 3. Multivariate analysis showed that male gender
EF < 50% 89 (3.4%) 2 (7%) (p = 0.009), presence of diabetes (p = 0.002), personal his-
IVS diastolic thickness (cm) 1.1 ± 0.2 1.1 ± 0.2 tory of CAD (p < 0.001) and abnormal non-ischemic SE
Stress variables (p = 0.01) were independent predictors of a higher incidence
Peak heart rate (beats/min) 142 ± 19 138 ± 19 of new CAD at 1 year in the study population (Table 4).
Peak systolic blood pressure (mmHg) 175 ± 26 169 ± 28
Peak diastolic blood pressure 82 ± 13 75 ± 17 Clinical and stress test variables associated
(mmHg) with all‑cause mortality
Data are presented as the mean value ± SD or number (%) of patients
A total of 99 patients died during the study follow-up period
CAD Coronary artery disease, SE stress echocardiogram, NSE nega-
tive stress echocardiogram, EF ejesction fraction, IVS interventricular of 3.2 ± 1.2 years. The clinical and stress test variables asso-
septum ciated with all-cause mortality on univariate analysis are
shown in Table 3. Patients with NSE who were older, men,
with a prior or current history of smoking and a personal his-
Clinical and echocardiographic characteristics tory of CAD had higher all-cause mortality during the study
of NSE patients dead at 1 year period. Patients with abnormal non-ischemic results on SE
had higher all-cause mortality (p = 0.044), but there was no
A total of 29 patients with NSE (1.1%) were dead within significant difference between negative and non-diagnostic
1 year from index SE. Clinical and characteristics of this SE. Patients who underwent exercise stress as the SE modal-
subset of patients were summarized in Table 1. Variables ity and those who achieved a lower peak heart rate at stress
associated with > 1% mortality at 1 year were age ≥ 60 years had lower all-cause mortality (p < 0.001 for both).
(2.2% death at 1 year), male gender (1.7% death at 1 year), Cox regression analysis showed that age, male gender,
white race (1.2% death at 1 year), diabetes (1.2% death prior or current history of smoking and being selected to
at 1 year), smoking history (1.5% death at 1 year), being undergo a pharmacologic (i.e. dobutamine) SE as the stress

13
The International Journal of Cardiovascular Imaging

Table 2  Descriptive statistics of clinical and stress test variables for the total population, patients with new CAD at 1 year and patients with no
new CAD at 1 year
Variable Total (N = 2575) New CAD at 1 year (N = 80) No new CAD at 1 year P-value
(N = 2495)

Age (years) 59 ± 11 64 ± 11 59 ± 11 < 0.001

Age ≥ 60 years 1220 (43%) 50 (63%) 1170 (47%)
Gender
Male 1180 (46%) 54 (68%) 126 (5%) < 0.001
Female 1395 (54%) 26 (33%) 1369 (55%)
Body mass index (kg/m2) 31 ± 8 32 ± 5 31 ± 8 0.03
Race
White 1812 (70%) 69 (86%) 1743 (70%) 0.007
Black 677 (26%) 10 (13%) 667 (27%)
Other ethnicities 86 (4%) 1 (1%) 85 (3%)
Diabetes mellitus 913 (35%) 47 (59%) 866 (35%) < 0.001
Hypertension 1714 (67%) 69 (86%) 1645 (66%) < 0.001
Hyperlipidemia 1566 (61%) 66 (80%) 1500 (60%) < 0.001
Prior or current smoking 985 (38%) 40 (50%) 945 (38%) 0.03
Family history of CAD 167 (6%) 10 (13%) 157 (6%) 0.03
Personal history of CAD 678 (26%) 57 (71%) 621 (25%) < 0.001
SE type
Dobutamine 852 (33%) 29 (36%) 823 (33%) 0.541
Exercise 1723 (67%) 51 (64%) 1672 (67%)
NSE result
Normal 1979 (77%) 42 (53%) 1937
Non-diagnostic 441 (17%) 21 (26%, 4.8%) 420 0.002
Abnormal non-ischemic 155 (6%) 17 (21%, 11.0%) 138 < 0.001
Rest variables
EF (%) 60 ± 7 59 ± 10 60 ± 7 0.338
EF < 50% 89 (3.4%) 6 (6.7%) 83
IVS diastolic thickness (cm) 1.1 ± 0.2 1.1 ± 0.2 1.0 ± 0.2 0.002
Stress variables
Peak heart rate (beats/min) 142 ± 19 131 ± 19 143 ± 19 < 0.001
Peak systolic blood pressure (mmHg) 175 ± 26 179 ± 28 176 ± 26 0.108
Peak diastolic blood pressure (mmHg) 82 ± 13 84 ± 14 82 ± 13 0.515

Data are presented as the mean value ± SD or number (%) of patients

CAD coronary artery disease, SE stress echocardiogram, NSE negative stress echocardiogram, EF ejection fraction, IVS Interventricular septum

modality (p = 0.004) were independent predictors of all- [2, 4, 5]. The annual rate of obstructive coronary stenosis
cause mortality in patients with NSE (Table 5 and Fig. 2). following NSE has been scarcely reported and was found
to be 3.1%/year in our study. The outcome of 137 patients
undergoing exercise SE was reported in a previous study
Discussion over a follow-up period of 23 months. The annual rate of
cardiac events, which included myocardial infarction and
To our knowledge, this is the largest contemporary study coronary revascularization, was found to be 1.8%/year [11].
that describes the prognostic significance of NSE. In this In a similar study, 170 patients undergoing exercise stress
study of 2575 patients with NSE, the annual all-cause mor- testing over a period of 28 months were found to have a
tality was found to be 1.1%/year (N = 29) while the study myocardial infarction rate of 0.85%/year and a coronary
mortality rate was found to be 3.8% (N = 99) over a follow- revascularization rate of 1.7%/year [3]. No deaths were
up period of 3.2 ± 1.2 years. These findings were similar recorded in any of the aforementioned studies. In a study by
to those reported in previous long-term follow-up studies Davar et al., 136 women with a high pretest probability of

13
 The International Journal of Cardiovascular Imaging

Table 3  Clinical and stress test Variable New CAD at 1 year All-cause mortality
variables associated with new
CAD at 1 year and all-cause OR (95% CI) P-value HR (95% CI) P-value
mortality on univariate analysis
Normal 1 1
Abnormal non-ischemic SE 5.68 (3.15–10.24) < 0.001 2.39 (1.32–4.33) 0.004
Non-diagnostic SE 2.31 (1.35–3.94) 0.002 1.14 (0.67–1.93) 0.64
Age 1.04 (1.02–1.06) < 0.001 1.06 (1.05–1.09) < 0.001
Body mass index 1.01 (0.99–1.03) 0.343 0.97 (0.95–1.00) 0.081
Male gender 2.52 (1.57–4.06) < 0.001 1.64 (1.09–2.44) 0.016
Black (versus non-black) 0.38 (0.19–0.74) 0.004 0.84 (0.52–1.34) 0.462
Diabetes mellitus 2.68 (1.70–4.21) < 0.001 1.08 (0.71–1.62) 0.726
Hypertension 3.24 (1.71–6.16) < 0.001 1.32 (0.84–2.06) 0.231
Hyperlipidemia 3.12 (1.75–5.60) < 0.001 0.99 (0.66–1.49) 0.981
Prior or current smoking 1.64 (1.05–2.56) 0.03 1.90 (1.28–2.82) 0.002
Personal history of CAD 7.48 (4.57–12.24) < 0.001 2.31 (1.55–3.43) < 0.001
Family history of CAD 2.13 (1.07–4.21) 0.03 1.23 (0.60–2.54) 0.569
Exercise SE 0.86 (0.54–1.37) 0.542 0.40 (0.27–0.60) < 0.001
IVS diastolic thickness 3.46 (1.40–8.57) 0.007 2.05 (0.87–4.86) 0.1
Resting EF 0.97 (0.95–0.99) 0.046 0.99 (0.96–1.02) 0.47
Peak heart rate 0.97 (0.96–0.98) < 0.001 0.98 (0.97–0.99) < 0.001
Peak systolic blood pressure 1.00 (0.99–1.01) 0.07 0.99 (0.98–1.00) 0.075
Peak diastolic blood pressure 1.00 (0.99–1.02) 0.335 0.99 (0.97–1.00) 0.233

CAD coronary artery disease, OR odds ratio, HR hazard ratio, SE stress echocardiography, EF ejection
fraction

Table 4  Clinical and stress test variables associated with new CAD at Table 5  Clinical and stress test variables associated with all-cause
1 year on multivariate analysis death on Cox regression
Variable New CAD at 1 year Variables All-cause death
OR (95% CI) P-value HR (95% CI) P-value

Abnormal non-ischemic SE 2.41 (1.24–4.69) 0.01 Abnormal non-ischemic SE 1.34 (0.72–2.50) 0.356
Age 1.00 (0.98–1.03) 0.605 Age 1.05 (1.02–1.07) < 0.001
Male gender 1.98 (1.19–3.28) 0.009 Male gender 1.59 (1.05–2.40) 0.028
Black race 0.52 (0.26–1.06) 0.073 Body mass index 0.99 (0.96–1.02) 0.546
Diabetes mellitus 2.18 (1.34–3.26) 0.002 Personal history of CAD 1.18 (0.76–1.84) 0.452
Hyperlipidemia 1.14 (0.59–2.23) 0.732 Prior or current smoking 1.86 (1.22–2.83) 0.004
Prior or current smoking 1.01 (0.62–1.66) 0.923 Exercise SE 0.52 (0.34–0.81) 0.004
Personal history of CAD 3.96 (2.19–7.17) < 0.001 Peak systolic blood pressure 0.99 (0.98–1.00) 0.172
Family history of CAD 1.39 (0.67–2.90) 0.351 Peak heart rate 0.99 (0.97–1.00) 0.172
Peak systolic blood pressure 1.01 (1.00–1.02) 0.133
HR hazard ratio, SE stress echocardiography, CAD coronary artery
Peak heart rate 0.99 (0.97–1.00) 0.051 disease
Resting EF 0.99 (0.97–1.03) 0.981

CAD Coronary artery disease, OR odds ratio, SE stress echocardiog-

raphy, EF ejection fraction carry a higher pretest probability of CAD. Besides, and con-
trary to our study, patients with non-diagnostic SE as well as
those with a personal history of CAD and those with resting
CAD were evaluated for CAD with either treadmill or dobu- WMA were excluded from the aforementioned study, which
tamine SE. Ninety-three patients (68.9%) had a NSE and may explain the higher incidence of CAD in our female sub-
were followed up for the occurrence of CAD. The authors population (1.9% at 1 year). In a different study, 164 patients
reported an annual CAD rate of 0.4% in women with NSE who were diagnosed with new-onset significant CAD and
[12]. The higher CAD incidence observed in our study could had a recent NSE (ranging from less than 1 month up to
be partly attributed to the presence of 46% of males who 3 years from diagnosis) were surveyed at a single-center

13
The International Journal of Cardiovascular Imaging

Fig. 2  Kaplan–Meier analysis for all-cause mortality stratified according to age (a), gender (b), history of smoking (c) and stress echocardiogra-
phy modality (d)

emergency department. In this study, 20.7% (34/164) of the older age, male gender, presence of diabetes, personal his-
patients with NSE performed within 3 years of presenta- tory of CAD and abnormal non-ischemic SE results as
tion had significant CAD [13]. While the annual CAD rate independent predictors of new CAD at 1 year. Advanced
was not reported in this study, 27 of the 34 patients newly age is associated with a higher prevalence of cardiovascu-
diagnosed with CAD had their stress tests performed within lar and non-cardiovascular comorbidities thus explaining
1 year from diagnosis. The discrepancy between our results why age was found to be an independent predictor for both
and those of the previous study could be explained by the CAD at 1 year as well as all-cause mortality [5, 14]. Large
fact that patients presenting to the emergency department epidemiological studies have demonstrated that men have
with acute chest pain are more likely to have CAD. In addi- higher all-cause mortality (including cardiovascular causes)
tion, all of the patients included in our study were evaluated than women [15–17]. This disparity in survival between the
in the outpatient setting and only 53% had chest pain as an two genders has been attributed to biological differences
indication for SE. (such as genetic factors, immune response, hormones, and
disease patterns) and behavioral differences (such as com-
Predictors of clinical outcomes pliance with medications and willingness to seek medical
help) [17]. Male gender was independently associated with
While NSE is usually associated with favorable prognosis, a higher rate of new CAD at 1 year and all-cause mortality
we identified a subset of patients that are at a higher risk of in our study. Smoking is a well-established risk factor for
overall as well as cardiovascular adverse outcomes. Patients CAD and all-cause mortality [18, 19]. In our study, smok-
who were older, males, with a prior or current history of ing was found to be an independent predictor of all-cause
smoking and who were selected to undergo dobutamine SE mortality but not of 1-year CAD. It is possible that since
had a higher rate of overall mortality. We also identified patients with 1-year CAD constituted only 3.1% of the total

13
 The International Journal of Cardiovascular Imaging

population, the analysis of this subset of patients was lim- Limitations

ited by confounding interactions. As stated above, patients
who were selected to undergo dobutamine SE had a higher This report is a retrospective study and is subject to the
overall mortality. One plausible explanation could be that limitations of the study design. Since the pretest probabil-
those unable to achieve at least 5 metabolic equivalents are ity is largely determined by the symptoms reported by the
usually referred to dobutamine as opposed to exercise SE. patients, data collection of this variable was not feasible.
Also, an inverse correlation between the ability to achieve Also, the positivity of SE result was determined solely
exercise capacity thresholds and all-cause mortality has been based on echocardiographic and not electrocardiographic
previously established [20, 21]. The presence of diabetes is factors or the presence of chest pain or anginal equiva-
a well-established risk factor for CAD and was found to be lents. Patients who had a negative but non-diagnostic
an independent risk factor for new CAD lesions at 1 year in SE might have had other subsequent workups including
patients with NSE. Similar results were reported by Kama- non-echocardiographic imaging modalities. However, the
lesh et al. who followed 233 patients with NSE and found potential additional data was not collected. While inter-
that those with diabetes were at a higher risk for cardiac observer variability is usually unavoidable, multiple stud-
events [22]. Patients with a personal history of CAD are at ies reported good reproducibility in patients with NSE
a higher risk for future cardiovascular events [23]. While all [36]. Drugs that could affect the results of the test were
patients with abnormal non-ischemic tests included in our generally discontinued according to our center protocol.
study had a known previous history of CAD, the presence However, it is possible that they were erroneously taken
of an abnormal non-ischemic SE result was an independent in certain cases thus preventing corresponding patients to
risk factor for 1-year CAD. This finding was also reported achieve THR and contributing to a false-negative or a non-
in previous studies that showed an increased risk of major diagnostic NSE. In this study, only the all-cause mortality
cardiac events and death in patients with WMA only at rest, was surveyed. Therefore, we were unable to determine the
irrespective of the presence or absence of CAD on subse- cause of death when it occurred (cardiac vs. non-cardiac).
quent investigation [24, 25]. The numbers of deaths or ACS at 1 year were low thus
preventing a multivariate analysis for 1-year mortality and
The role of other imaging modalities in predicting 1-year ACS, respectively. We did not account for ACS
outcomes patients that did not undergo coronary angiography in this
study. Furthermore, the number of CAD or ACS in our
Myocardial infarction and cardiac death following normal study might have been underestimated in the event patients
perfusion studies have been reported in multiple studies to sought medical care at different centers or healthcare sys-
be < 1%/year over a period of 1 to 10 years [26–29]. Otten- tems. The studied population was a referral-based group
hof et al. evaluated 266 patients with known CAD and and might not be accurately representative of the general
normal perfusion imaging over a period of 12 years. The population encountered in the primary care settings. Data
annual mortality rate was found to be 3.1% and was higher were incomplete on the technical quality of SE studies, and
than that reported in our study (1.1%) [30]. This could information about quality post-echocardiographic contrast
be explained by the fact that all the studied patients had use was not available.
a personal history of CAD, compared to only 38% in our
population. Coronary computed tomography angiography
and coronary artery calcium scoring have also been used
to predict mortality and cardiovascular events [31, 32]. In Conclusions
a large study from the CONFIRM registry, patients with no
CAD detected on coronary computed tomography angiog- Patients with NSE have low all-cause mortality (1.1%) and
raphy had an annualized death rate of 0.28% [33]. In one CAD (3.1%) annual rates. However, older patients, male
review article, a coronary artery calcium score of 0 was gender, a prior or current of smoking, and being selected
associated with a 10-year CAD rate of 1.1%. There was an for dobutamine SE were associated with an increased risk
incremental increase in the 10-year CAD rate with increas- of all-cause mortality over the study period. Additionally,
ing calcium, reaching up to 37% with a score > 1000 [34]. male gender, diabetes, personal history of CAD, and the
The role of coronary angiography in predicting outcomes presence of abnormal non-ischemic results on SE predicted
has also been studied. In a previous study of 21,487 patients a higher rate of new CAD lesion at 1 year. These characteris-
who underwent catheterization over 7 years, patients who tics should be accounted for during the clinical interpretation
had normal angiographic findings (N = 4051) had a 7-year of SE and patients found at increased risk for morbidity and
survival rate of 96% as compared to 92% in patients with mortality warrant continued follow-up.
mild CAD (p < 0.001) [35].

13
The International Journal of Cardiovascular Imaging

Funding None. women over two decades: a registry-based study in Ontario, Can-
ada. BMJ Open 6:e012564
16. Crimmins EM, Shim H, Zhang YS, Kim JK (2019) Differences
Compliance with ethical standards between men and women in mortality and the health dimensions
of the morbidity process. Clin Chem 65:135–145
Conflict of interest None declared. All authors take responsibility for 17. Oksuzyan A, Juel K, Vaupel JW, Christensen K (2008) Men: good
all aspects of the reliability and freedom from bias of the data pre- health and high mortality. Sex differences in health and aging.
sented and their discussed interpretation. Aging Clin Exp Res 20:91–102
18. Doyle JT, Dawber TR, Kannel WB, Heslin AS, Kahn HA (1962)
Cigarette smoking and coronary heart disease: combined expe-
rience of the Albany and Framingham studies. N Engl J Med
References 266:796–801
19. Gellert C, Schöttker B, Brenner H (2012) Smoking and all-cause
1. Marwick TH (2018) Stress echocardiography. Springer, Echocar- mortality in older people: systematic review and meta-analysis.
diography, pp 491–519 Arch Intern Med 172:837–844
2. Sicari R, Pasanisi E, Venneri L, Landi P, Cortigiani L, Picano 20. Kokkinos P, Faselis C, Myers J, Sui X, Zhang J, Blair SN (2014)
E (2003) Stress echo results predict mortality: a large-scale Age-specific exercise capacity threshold for mortality risk assess-
multicenter prospective international study. J Am Coll Cardiol ment in male veterans. Circulation 130:653–658
41:589–595 21. Rozanski A, Gransar H, Hayes SW, Friedman JD, Hachamovitch
3. Sawada SG, Ryan T, Conley MJ, Corya BC, Feigenbaum H, Arm- R, Berman DS (2010) Comparison of long-term mortality risk
strong WF (1990) Prognostic value of a normal exercise echocar- following normal exercise vs adenosine myocardial perfusion
diogram. Am Heart J 120:49–55 SPECT. J Nucl Cardiol 17:999–1008
4. Steinberg EH, Madmon L, Patel CP, Sedlis SP, Kronzon I, Cohen 22. Kamalesh M, Matorin R, Sawada S (2002) Prognostic value of a
JL (1997) Long-term prognostic significance of dobutamine echo- negative stress echocardiographic study in diabetic patients. Am
cardiography in patients with suspected coronary artery disease: Heart J 143:163–168
results of a 5-year follow-up study. J Am Coll Cardiol 29:969–973 23. Smith SC, Allen J, Blair SN, Bonow RO, Brass LM, Fonarow GC
5. McCully RB, Roger VL, Mahoney DW, Karon BL, Oh JK, Miller et al (2006) AHA/ACC guidelines for secondary prevention for
FA et al (1998) Outcome after normal exercise echocardiography patients with coronary and other atherosclerotic vascular disease:
and predictors of subsequent cardiac events: follow-up of 1,325 2006 update: endorsed by the National Heart, Lung, and Blood
patients. J Am Coll Cardiol 31:144–149 Institute. J Am Coll Cardiol 47:2130–2139
6. Senior R, Monaghan M, Becher H, Mayet J, Nihoyannopoulos 24. Araujo ACPd, Santos BF, Calasans FR, Pinto IM, Oliveira DP,
P (2005) Stress echocardiography for the diagnosis and risk Melo LD et al (2014) Physical stress echocardiography: predic-
stratification of patients with suspected or known coronary artery tion of mortality and cardiac events in patients with exercise test
disease: a critical appraisal. Supported by the British Society of showing ischemia. Arq Bras Cardiol 103:418–425
Echocardiography. Heart 91:427–436 25. Schlett CL, Banerji D, Siegel E, Bamberg F, Lehman SJ, Ferencik
7. Metz LD, Beattie M, Hom R, Redberg RF, Grady D, Fleischmann M et al (2011) Prognostic value of CT angiography for major
KE (2007) The prognostic value of normal exercise myocardial adverse cardiac events in patients with acute chest pain from the
perfusion imaging and exercise echocardiography: a meta-analy- emergency department: 2-year outcomes of the ROMICAT trial.
sis. J Am Coll Cardiol 49:227–237 JACC Cardiovasc Imaging 4:481–491
8. Batal O, Malhotra S, Harinstein M, Markowitz J, Hickey G, Agar- 26. Frans JT, Russo DJ, Russo D, Clements JP (1985) Prognostic sig-
wal S et al (2019) Performance of traditional pretest probability nificance of normal quantitative planar thallium-201 stress scin-
estimates in stable patients undergoing myocardial perfusion tigraphy in patients with chest pain. J Am Coll Cardiol 6:27–30
imaging. Circulation 12:e008473 27. Pamelia FX, Gibson RS, Watson DD, Craddock GB, Sirowatka
9. Rachwan RJ, Mshelbwala FS, Dardari Z, Batal O (2019) False- J, Beller GA (1985) Prognosis with chest pain and normal thal-
positive stress echocardiograms: predictors and prognostic rel- lium-201 exercise scintigrams. Am J Cardiol 55:920–926
evance. Int J Cardiol 296:157–163 28. Steinberg EH, Koss JH, Lee M, Grunwald AM, Bodenheimer MM
10. Pellikka PA, Nagueh SF, Elhendy AA, Kuehl CA, Sawada SG (1993) Prognostic significance from 10-year follow-up of a quali-
(2007) American Society of Echocardiography recommendations tatively normal planar exercise thallium test in suspected coronary
for performance, interpretation, and application of stress echocar- artery disease. Am J Cardiol 71:1270–1273
diography. J Am Soc Echocardiogr 20:1021–1041 29. Brown K (1991) Prognostic value of thallium-201 myocardial
11. Ismail G, Lo E, Sada M, Conant RD, Shapiro SM, Ginzton LE perfusion imaging. A diagnostic tool comes of age. Circulation
(1995) Long-term prognosis of patients with a normal exercise 83:363–381
echocardiogram and clinical suspicion of myocardial ischemia. 30. Ottenhof MJ, Wai MCTJ, Boiten HJ, Korbee RS, Valkema R, van
Am J Cardiol 14:934–936 Domburg RT et al (2013) 12-Year outcome after normal myo-
12. Davar J, Roberts E, Coghlan J, Evans T, Lipkin D (2001) Prog- cardial perfusion SPECT in patients with known coronary artery
nostic value of stress echocardiography in women with high (⩾ disease. J Nucl Cardiol 20:748–754
80%) probability of coronary artery disease. Postgrad Med J 31. Deseive S, Shaw LJ, Min JK, Achenbach S, Andreini D, Al-
77:573–577 Mallah MH et al (2017) Improved 5-year prediction of all-cause
13. Walker J, Galuska M, Vega D (2010) Coronary disease in emer- mortality by coronary CT angiography applying the CONFIRM
gency department chest pain patients with recent negative stress score. Eur Heart J 18:286–293
testing. West J Emerg Med 11:384 32. o Hartaigh B, Valenti V, Cho I, Schulman-Marcus J, Gransar
14. Diamond GA, Forrester JS (1979) Analysis of probability as an H, Knapper J et al (2016) 15-year prognostic utility of coronary
aid in the clinical diagnosis of coronary-artery disease. N Engl J artery calcium scoring for all-cause mortality in the elderly. Ath-
Med 300:1350–1358 erosclerosis. 246:361–366
15. Rosella LC, Calzavara A, Frank JW, Fitzpatrick T, Donnelly 33. Min JK, Dunning A, Lin FY, Achenbach S, Al-Mallah M, Budoff
PD, Henry D (2016) Narrowing mortality gap between men and MJ et al (2011) Age-and sex-related differences in all-cause

13
 The International Journal of Cardiovascular Imaging

mortality risk based on coronary computed tomography angi- 36. Hoffmann R, Lethen H, Marwick T, Arnese M, Fioretti P, Pingi-
ography findings: results from the International Multicenter tore A et al (1996) Analysis of interinstitutional observer agree-
CONFIRM (Coronary CT Angiography Evaluation for Clini- ment in interpretation of dobutamine stress echocardiograms. J
cal Outcomes: An International Multicenter Registry) of 23,854 Am Coll Cardiol 27:330–336
patients without known coronary artery disease. J Am Coll Car-
diol 58:849–860 Publisher’s Note Springer Nature remains neutral with regard to
34. Hecht HS (2015) Coronary artery calcium scanning: past, present, jurisdictional claims in published maps and institutional affiliations.
and future. JACC Cardiovasc Imaging 8:579–596
35. Kemp HG, Kronmal RA, Vlietstra RE, Frye RL (1986) Seven year
survival of patients with normal or near normal coronary arterio-
grams: a CASS registry study. J Am Coll Cardiol 7:479–483

13