2. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 2
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
R
heumatic heart disease (RHD) remains a major
public health concern, especially in low- and mid-
dle-income countries, where it is the leading cause
of cardiovascular death in children and young adults.1
Late presentation and late diagnosis of patients with
RHD remains a key driver of high morbidity and mor-
tality.2
During the last decade, screening echocardiog-
raphy to identify RHD (termed latent RHD) has emerged
as a powerful tool for active case finding, epidemiology,
and advocacy.3–10
Proponents of screening echocardiog-
raphy have hope that diagnosis of latent RHD and early
initiation of penicillin prophylaxis will prevent advance-
ment to clinically significant disease,11,12
although this
has not yet been proven.
Early in the development of screening echocardi-
ography, it became clear that a standardized set of di-
agnostic criteria were needed to allow replication and
comparison between sites.13,14
An international expert
working group was convened by the World Heart Fed-
eration (WHF), and in 2012, the first evidence-based
criteria were published.15
These criteria take into con-
sideration morphological and functional changes of the
left heart valves and classify cases into the categories of
normal, borderline RHD, or definite RHD. Prospective
population-based screening utilizing the WHF criteria
has shown high sensitivity for RHD, and high specific-
ity has been inferred from studies showing low rates
of RHD when applying the WHF criteria to low-risk
populations.10,16
Now, just half a decade after publication, the WHF
criteria have been widely adopted as the interna-
tional gold standard for latent RHD diagnosis. Active
case finding utilizing these criteria has occurred in 2
dozen populations, and 100 000 patients have been
classified.7–10,17–32
Additionally, several cohorts of chil-
dren initially diagnosed by these criteria have been
followed longitudinally, providing a window into their
outcomes.6,33–40
As a result, the cumulative experience
and evidence around early RHD diagnosis has grown
immensely since publication of the WHF criteria, and a
large amount of data are available from which the per-
formance and weight of each feature might be tested.
Analysis of these data holds promise for development
of less complex criteria, focused specially on screening,
which would allow echocardiography screening to be
more reproducible and be applicable on a larger scale
in endemic populations.
The aim of this study was to utilize 2012 WHF echo-
cardiographic results from several large-scale screening
studies to develop and validate a simplified score for
RHD diagnosis and to determine whether this score was
predictive of RHD progression.
METHODS
Study Design
The data, analytic methods, and study materials will not be
made available to other researchers for purposes of reproduc-
ing the results or replicating the procedure.
This study was designed to develop and validate a score
for diagnosis of definite RHD and further to test whether
the score would be able to predict disease progression. The
study occurred in 3 phases (Figure 1) using data from estab-
lished cohorts of asymptomatic children without history of
acute rheumatic fever or RHD, who had undergone screen-
ing echocardiography. In all phases, expert consensus had
been previously undertaken to determine echocardiographic
classification (normal, borderline, and definite) and features
according to the 2012 WHF criteria.15
Retrospective search of 3 existing deidentified echocardi-
ographic databases was used to gather basic demographic
information (age and sex) and echocardiographic features
at diagnosis. Each component of the 2012 WHF criteria was
captured as present or absent, including each of the 4 in-
dividual criteria necessary for pathological aortic and mitral
regurgitation. Studies were excluded from further analysis if
they contained incomplete reporting of the WHF criteria or
if the final diagnosis was congenital heart disease or other
acquired heart disease. Further, in the derivation stage, echo-
cardiograms classified as borderline RHD were also excluded
to focus on the discriminating features of the extreme phe-
notypes—normal and definite RHD. Outcome was assessed
in a cohort of latent RHD, defined as disease progression by
echocardiogram.
The study was approved by the institutional review com-
mittee of both institutions, and written informed consent was
obtained from all subjects.
Phase 1: Derivation
The derivation cohort included 12 056 echocardiograms of
children screened through the PROVAR study (Programa de
CLINICAL PERSPECTIVE
The World Heart Federation criteria are the cur-
rent gold standard for the diagnosis of latent rheu-
matic heart disease (RHD). As data and experience
using these criteria have grown, concerns related
to the complexity of these criteria for RHD screen-
ing programs have been raised. In this regard, a re-
finement set of echocardiographic criteria for the
diagnosis of RHD, focused specially on screening,
would allow echocardiography screening to be
more reproducible and be applicable on a larger
scale in endemic populations. In the current study,
a simplified score for latent RHD diagnosis, based
on components of the World Heart Federation
criteria, is highly accurate to recognize definite
RHD. Furthermore, the new score provides the first
tool for risk stratification at time of diagnosis of
latent RHD, with potential value for clinical deci-
sion-making regarding secondary prophylaxis.
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
3. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 3
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
Rastreamento da Valvopatia Reumática)41
—a large school-
based screening program in Minas Gerais, Brazil.
In brief, the PROVAR study enrolled asymptomatic chil-
dren aged 5 to 18 years without history of acute rheumatic
fever or RHD.9
Screening echocardiography was per-
formed using standard portable (Vivid-Q; GE Healthcare,
Milwaukee, WI) or handheld (VSCAN, GE Healthcare) echo-
cardiography machines, and images were analyzed via tel-
emedicine by cardiologists in Brazil and the United States.9
For handheld studies in which spectral Doppler is not avail-
able, modified WHF criteria, based on the features of color
regurgitant jet and morphological signs of RHD were used,
as described previously.42,43
Phase 2: Validation
The validation cohort included 7312 echocardiograms of chil-
dren screened in Gulu, Uganda.21,44
In brief, collaboration
between the Uganda Heart Institute (Kampala, Uganda) and
Children’s National Health System undertook several school-
based echocardiographic screening studies.21,44
Asymptomatic children without history of acute rheu-
matic fever/RHD underwent screening echocardiography
using a standard portable echocardiography machine
(Vivid-Q; GE Healthcare), and cardiologists with expertise
in the 2012 WHF criteria analyzed the images. Because the
WHF criteria include spectral Doppler assessment, hand-
held echocardiography equipment was not used in the val-
idation cohort. Image acquisition settings using the highest
fundamental imaging frequency of 3.6 MHz were stan-
dardized for standard portable echocardiography across all
cohorts in this study.
Phase 3: Outcomes Prediction
The ability of this score to predict outcome was tested in an
established and previously reported40
longitudinal echocar-
diographic cohort, derived from 227 children (aged 5–18
years), identified with latent borderline or definite RHD and
enrolled in the Ugandan National RHD Registry. Median
length of echocardiographic follow-up in this cohort was
2.3 years (interquartile range, 2.0–2.9). Retrospective
search of this database captured baseline diagnosis and
presence or absence of score components. Previously de-
termined outcomes were captured, including progression,
stabilization, or regression. Unfavorable outcome or di-
sease progression was defined as worsening in diagnostic
category (borderline to definite), a worsening in the grade
of mitral or aortic valve regurgitation, development or
worsening of grade of mitral stenosis, or death because
of complication of RHD. Regression was defined as an im-
provement in diagnostic category or an improvement in se-
verity of mitral or aortic regurgitation based on the findings
of the last available echocardiogram. Cases that did not
meet these definitions for progression or regression were
considered stable. In case of a nonlinear outcome, such as
initial worsening and then return to baseline, the findings
of the last available echocardiogram were considered as
the final outcome.
Statistical Analysis
Statistical analysis was performed using the Statistical Package
for Social Sciences for Windows, version 22.0 (SPSS, Inc,
Chicago, IL), and R for Statistical Computing, version 2.15.1
(R Foundation, Vienna, Austria). Categorical variables were
compared by χ2
test, and continuous data were compared
using Student unpaired t test when appropriate.
Logistic regression was used to identify the WHF compo-
nent variables associated with definite RHD. The variables
found to be significant at the level P 0.05 in univariate anal-
ysis were initially selected for a multivariable logistic regression
analysis. To investigate the potential effect of multicollinearity
among the WHF variables, changes in the regression coeffi-
cient estimates, in their corresponding standard errors, and in
the variance inflation factor were examined.45
The final model
was then derived by including only relevant variables associ-
ated with definite RHD.
Figure 1. Flowchart of the study popula-
tions.
A, Derivation; B, Validation; and C, Outcomes
prediction. RHD indicates rheumatic heart di-
sease; and WHF, World Heart Federation.
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
4. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 4
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
A point-based scoring system was developed from the
final multivariable logistic regression model. Points were
assigned to each component by rounding the β-coefficient to
the nearest integer,46
and the score corresponded to the sum
of the points, ranging from 0 to 21. Three risk groups (low, in-
termediate, and high) were defined based on predicted prob-
abilities from logistic regression. A low-risk score was set as
a predicted probability of RHD diagnosis of 2%, whereas
high-risk score was set as a predicted risk of 40%, and inter-
mediate risk in the middle range.
To assess the predictive accuracy of the scoring system for
diagnosis of definite RHD, discrimination and calibration were
evaluated using the area under the receiver operating charac-
teristic curves and the Hosmer-Lemeshow test, respectively.47
Additionally, the derivation cohort was randomly split into 2
separate groups as training (80%) and test (20%) samples,
and the discrimination and calibration was evaluated by
applying this fitted model to the test group. Cross-validation
was repeated 1000× to generate an average, overall opti-
mism-adjusted estimate of the C statistic.48–50
The optimism in
the model performance was estimated using bootstrapping
techniques.50
For external validation, the score was applied to
an independent cohort to assess discrimination and calibra-
tion of the score model for definite RHD diagnosis.
The score was also applied to an independent population
of children with latent RHD to assess its discrimination and
calibration in predicting disease progression using logistic re-
gression. The value of the score in predicting unfavorable out-
come was assessed as a time-dependent variable in the Cox
proportional hazards model. RHD progression-free survival
rates of the 3 risk categories were estimated by the Kaplan-
Meier method and compared by the log-rank test.
Additionally, clinical variables that may influence disease
progression, specifically age and penicillin use, were included
in the multivariable Cox regression model.
RESULTS
Derivation Cohort
Of the 12 056 echocardiograms in the derivation cohort,
2028 were excluded for incomplete reporting of the
WHF criteria, 67 were excluded as other structural heart
disease (mainly congenital and nonrheumatic abnor-
malities, especially mitral valve prolapse), and 460 were
excluded because of classification as borderline RHD.
Of the remaining 9501 echocardiographic studies, 58
(0.6%) were classified as definite RHD and 9443 (99.4%)
as normal. Table 1 contains demographics and baseline
echocardiographic features of this cohort (including bord-
erline RHD for comparison of component diagnoses). Of
note, the cohort did not contain any patients with mitral
stenosis or the component aortic valve prolapse.
At univariate analysis, the WHF parameters asso-
ciated with definite RHD were identified (Table 2). All
morphological criteria for mitral valve and 2 morpho-
logical criteria (focal valve thickening and coaptation
defect) for aortic valve involvement were initially select-
ed. Restricted leaflet motion of aortic valve occurred in
only 3 children who had definite RHD. All 4 variables
of pathological regurgitation in both valves were also
associated with RHD diagnosis.
Subsequently, to avoid multicollinearity by including
all statistically significant variables with redundant in-
formation, we selected only the most relevant variable
for pathological regurgitation: length of mitral regurgi-
tation (≥2 cm) and aortic (any) regurgitation. Presence
of regurgitation in multiple views, spectral Doppler
assessment, and regurgitant velocity were considered
redundant and did not enter together in the multivari-
able model. Although leaflet thickening and jet length
were also collected as continuous variables, we in-
cluded in the model as categorical variables, following
the WHF criteria.
For the morphological criteria of the mitral valve, we
included the 2 most frequent changes: anterior leaf-
let thickening (86%) and excessive leaflet tip motion
(52%). For the aortic valve, as coaptation defect causes
valvular regurgitation, only aortic regurgitation entered
into the multivariable model. Focal aortic valve thick-
ening was the morphological variable selected for the
multivariable analysis.
Therefore, the final multivariable logistic regression
model consisted of 5 variables: anterior leaflet thick-
ening, excessive leaflet tip motion, regurgitation jet
length ≥2 cm for mitral valve, and focal thickening and
any regurgitation for aortic valve. These variables were
then included in the development of the scoring sys-
tem, resulting in a total score of 0 to 21 points (Table 3).
Three risk categories were defined for the diagnosis of
definite RHD by estimating predicted probabilities from
logistic regression (Figure 2). According to the predicted
probability, low-risk scores were 0 to 6, intermediate-
risk scores were 7 to 9, and high-risk scores were ≥10.
Echocardiographic images of low, intermediate, and
high risk are shown in Figure 3 (Movies I through III in
the Data Supplement).
Bootstrap resampling showed negligible model opti-
mism. The model validated in the test group showed a C
statistic of 0.9982 (0.9967–0.9998) and well calibrated
with a χ2
of 7.4698 of Hosmer-Lemeshow, P=0.1879.
Cross-validation with replication showed optimal dis-
crimination of the model with an average of C statis-
tic of 0.9984 (95% CI, 0.9958–0.9997). Subsequently,
the model was applied to the subgroup of borderline
to stratify them into risk categories. Of 460 children in
borderline RHD category, 322 were reclassified in low
risk with this score, 20 in high risk, and 118 remained
as intermediate risk.
As the use of handheld echocardiography equipment
could be a confounding factor, the performance of the
score was tested according to the equipment used for
screening in the derivation cohort (VSCAN, 5838 and
Vivid-Q, 4123). There were no differences between the
models related to the equipment used (P=0.222).
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
5. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 5
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
External Validation
Complete WHF data were available for all 7312 echo-
cardiograms included in the validation cohort. Fifty-
two children were classified as definite RHD (0.7%),
254 as borderline (3.5%), and 7006 (95.8%) as nor-
mal. The characteristics of the validation cohort com-
paring with derivation cohort are shown in Table 4.
Children from the validation cohort were younger
with less valve regurgitation but with similar preva-
lence of definite RHD.
The score was then applied to the validation cohort,
and 7167 children were classified in the low-risk, 89
in the intermediate-risk, and 56 in the high-risk group.
The estimated probability of definite RHD in the low-,
intermediate-, and high-risk groups was 1%, 15.7%,
and 57%, respectively. The model showed an opti-
mal calibration and discrimination with a C statistic of
0.9949 (95% CI, 0.9933–0.9965) with a χ2
of 11.652
of Hosmer-Lemeshow; P=0.1132.
Of note, of 7006 children classified as normal by the
standard WHF criteria, 6998 were reclassified into low-
risk, 6 in intermediate-risk, and 2 in high-risk category.
Outcomes Prediction
Clinical and echocardiographic characteristics of the
derivation cohort as compared with the 227 children
in the outcomes cohort are shown in Table 5. At diag-
nosis, 164 (72.2%) were classified as borderline and 63
(27.7%) as definite RHD. The proportion of borderline
RHD at diagnosis was greater in the derivation cohort.
Although focal thickening of aortic valve leaflets was
more frequent in Uganda than in Brazil (6% versus 2%;
P=0.014), the other morphological features of RHD
and valvular regurgitation were similar between the 2
populations.
The original score model was then applied to this ex-
ternal population with latent RHD. The model showed
good discrimination and calibration with a C statistic
of 0.811 for predicting disease progression (95% CI,
0.720–0.902).
A low-risk score had a predicted probability of pro-
gression of ≤5.6%; medium risk ranging from 8.3%
to 17.5%, and high risk of ≥24.6%. The probability
of unfavorable outcome according to the total score
is shown in Figure 4. Based on the score in this latent
Table 1. Demographic and Echocardiographic Characteristics of Participants Who Underwent Rheumatic Heart
Disease Screening Using World Heart Federation Criteria: Derivation Cohort
Variables
Normal
(n=9443)
Borderline
(n=460)
Definite
(n=58)
Overall
Prevalence
Morphological features*
Mitral valve Anterior leaflet thickening† 855 (9) 127 (28) 50 (86) 1032 (10)
Chordal thickening 58 (0.6) 3 (0.7) 6 (10) 67 (0.7)
Restricted leaflet motion 10 (0.1) 7 (1.5) 13 (22) 30 (0.3)
Excessive leaflet tip motion 45 (0.5) 24 (5) 30 (52) 99 (1)
Stenosis with mean gradient ≥4
mm Hg
0 0 0 0
Aortic valve Irregular or focal thickening 54 (0.6) 1 (0.2) 10 (17) 65 (0.7)
Coaptation defect 1 (0.0) 1 (0.2) 4 (7) 6 (0.1)
Restricted leaflet motion 0 0 3 (5) 3 (0.0)
Prolapse 0 0 0 0
Valve regurgitation
Mitral regurgitation Any regurgitation 4563 (48) 430 (93) 56 (98) 5049 (51)
Seen in 2 views 3337 (35) 432 (94) 55 (95) 3824 (38)
Jet length ≥2 cm‡ 204 (2) 356 (77) 50 (86) 609 (6)
Velocity ≥3 m/s for 1 envelope§ 415 (4) 145 (32) 17 (29) 577 (6)
Pansystolic jet in 1 envelope§ 449 (5) 188 (41) 27 (47) 664 (7)
Aortic regurgitation Any regurgitation 163 (2) 81 (18) 21 (36) 265 (3)
Seen in 2 views 74 (0.8) 79 (17) 19 (33) 172 (1.7)
Jet length ≥1 cm‡ 54 (0.6) 75 (17) 21 (36) 150 (1.5)
Velocity ≥3 m/s in early diastole§ 2 (0.0) 19 (0.4) 7 (12) 28 (0.3)
Pandiastolic jet in 1 envelope§ 7 (0.0) 34 (0.7) 14 (24) 55 (0.5)
Data are expressed as absolute number (percentage).
*Congenital mitral valve or aortic valve abnormalities were excluded.
†Abnormal thickening of the anterior mitral valve leaflet ≥3 or 4 mm using harmonic imaging.
‡In at least 1 view.
§These measurements were available in 4123 participants who underwent echocardiographic screening using Vivid-Q.
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
6. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 6
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
RHD population, 140 were in the low-risk, 56 in the
intermediate-risk and 31 in the high-risk group.
Applying the score to predict disease outcomes,
6 of 140 (4.3%) in the low-risk group showed di-
sease progression, whereas 75 (53.6%) had regres-
sion over time. Similarly, of the 31 classified as high
risk, only 3 (9.7%) regressed and 28 (90.3%) either
progressed or remained stable (with an abnormal
echocardiogram) during short-term follow-up. RHD
status during the follow-up in the external valida-
tion cohort according to risk categories is shown in
Figure 5.
Considering 164 children who were classified
as borderline RHD at the time of the first screening
echocardiography, the score reclassified 127 in low-
risk, 27 in intermediate-risk, and 10 in high-risk cat-
egories. Of those children in low risk, 122 (96.1%)
showed regression, whereas 6 of 10 (60%) high-risk
children had disease progression (P0.001). Among
the borderline cases classified in intermediate risk,
81.5% remained stable or showed regression over
time. Figure 6 compares the predicted-to-observed
adverse outcome for each risk score group in the pop-
ulation with latent RHD.
Progression-free survival rates in low-risk children
at 1, 2, and 3 years of follow-up were 100%, 100%,
and 93%, respectively, compared with 98%, 93%, and
70% in the intermediate group and 90%, 60%, and
47% in the high-risk group (Figure 7). The point-based
score in continuous format was associated with disease
progression, with the risk increased for every 1-point
incremental in the score (hazard ratio, 1.270; 95% CI,
1.188–1.358; P0.001). Considering the low risk as a
reference category, the hazard ratio for unfavorable out-
Table 2. Univariate Logistic Regression Analysis to Determine the World Heart Federation Variables
Associated With the Diagnosis of Definite RHD
Variable β-Coefficient SE Z Value P Value
Mitral valve morphology
Anterior leaflet thickening 4.140 0.382 10.823 0.0001
Chordal thickening 2.927 0.451 6.492 0.0001
Restricted leaflet motion 5.608 0.446 12.562 0.0001
Excessive leaflet tip motion 5.411 0.302 17.899 0.0001
Mitral valve regurgitation
Any regurgitation 4.092 1.009 4.055 0.0001
Regurgitation jet length ≥1.5 cm* 4.198 0.432 9.710 0.0001
Regurgitation jet length ≥2 cm 5.651 0.387 14.589 0.0001
Seen in 2 views 3.794 1.009 3.762 0.0001
Velocity ≥3 m/s for 1 envelope 1.873 0.352 5.314 0.0001
Pansystolic jet in 1 envelope 3.050 0.427 7.139 0.0001
Aortic valve morphology
Irregular or focal thickening 3.590 0.373 9.612 0.0001
Coaptation defect 6.550 1.126 5.815 0.0001
Aortic valve regurgitation
Any regurgitation† 3.475 0.284 12.219 0.0001
Jet length ≥1 cm 4.662 0.322 14.456 0.0001
Seen in 2 views 5.181 0.753 6.882 0.0001
Velocity ≥3 m/s in early diastole 6.515 0.887 7.347 0.0001
Pandiastolic jet in 1 envelope 7.895 1.102 7.162 0.0001
RHD indicates rheumatic heart disease.
*A simplified RHD screening protocol used mitral regurgitation ≥1.5 cm or the presence of any aortic
regurgitation to define screen positive for RHD.
†A simplified criterion. Reference approach aortic valve regurgitation, ≥1 cm.
Table 3. Echocardiographic Variables Used for Score Development
Variable β-Coefficient SE Z Value P Value Points
Mitral valve
Anterior leaflet
thickening
2.941 0.597 4.922 0.0001 3
Excessive leaflet
tip motion
3.102 0.543 5.716 0.0001 3
Regurgitation jet
length ≥2 cm
5.601 0.705 7.941 0.0001 6
Aortic valve
Irregular or focal
thickening
4.460 0.970 4.597 0.0001 4
Any regurgitation 4.794 0.718 6.679 0.0001 5
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
7. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 7
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
come in intermediate risk was 5.228 (CI, 1.893–14.436)
and in high risk was 15.475 (CI, 6.132–39.053).
Penicillin was prescribed in 112 children (49.3%)
with the overall adherence of 84.7%. However, pen-
icillin use was not a predictor of unfavorable out-
come (hazard ratio, 1.247; 95% CI, 0.512–3.039;
P=0.627) in a model including the score, sex, and
age at diagnosis. Similarly, the interaction between
Figure 2. Total score points according to the probability of definite rheumatic heart disease (RHD) diagnosis in the derivation set.
The vertical dashed lines represent the cutoff values for low-, intermediate-, and high-risk groups.
Figure 3. Echocardiographic images showing the features of the 3 risk categories.
A, Low-risk group showing normal mitral valve morphology (left) with pathological mitral regurgitation (right); (B) intermediate-risk group defined as isolated
aortic valve focal thickening (left) with pathological aortic regurgitation (right), without any mitral valve changes; (C) high-risk group demonstrating thickened
anterior leaflet of the mitral valve (left), excessive leaflet tip motion (middle), and pathological mitral regurgitation (right). Involvement of both mitral and aortic
valves is also classified as a high risk.
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
8. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 8
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
age and score was not predictor of disease progres-
sion (P=0.869).
DISCUSSION
The publication and widespread adoption of the 2012
WHF criteria15
was a major advancement in the diagno-
sis of latent RHD, providing the best evidence available
at the time on the key features. Now, more than half a
decade later, large datasets utilizing these criteria offer
the opportunity to test the performance and weight of
its individual components, and for the first time, robust
longitudinal data exist to assess outcomes prediction.
In the current study, we have established a simplified
combination of WHF components with nearly perfect
performance to discriminate definite RHD from normal.
Further, we translated this simplified model into an
RHD risk score, which provided an accurate prediction
of RHD outcome in both borderline and definite latent
RHD in an established longitudinal cohort.
In our analysis, we were able to demonstrate near-
perfect discrimination between normal and definite
RHD using only 5 components. This simplified diag-
nostic model addresses one of the major critiques of
the WHF criteria—their complexity.51
The WHF criteria
were intended for RHD diagnosis14,15,52–54
; the ease and
practicality of using them in screening environments,
where most latent RHD is detected, and by nonex-
pert providers, who serve as the primary workforce
in many RHD-endemic areas, has been questioned.13
Table 4. Demographic and Echocardiographic Characteristics of Derivation and Validation Cohorts for Rheumatic Heart
Disease Diagnosis
Variables*
Derivation† (n=9961) Validation (n=7312)
P Value‡Normal (n=9443) Latent (n=518) Normal (n=7006) Latent (n=306)
Age at screening, y 12.9±2.9 13.8±2.7 10.7±2.7 11.0±2.4 0.001
Female sex, n (%) 5159 (54.6) 310 (59.8) 3765 (53.7) 147 (48.0) 0.071
Echocardiographic variables included in the model
Mitral valve
Anterior leaflet thickening 855 (9.1) 177 (34.2) 13 (0.2) 198 (64.7) 0.001
Excessive leaflet tip motion 45 (0.5) 54 (10.4) 1 (0.0) 92 (30.1) 0.085
Regurgitation jet length ≥2 cm 203 (2.1) 406 (78.4) 93 (1.3) 166 (54.2) 0.001
Aortic valve
Irregular or focal thickening 54 (0.6) 11 (2.1) 8 (0.1) 25 (8.2) 0.082
Any regurgitation 163 (1.7) 102 (19.7) 36 (0.5) 39 (12.7) 0.001
*Data are expressed as number (percentage) or as a mean±SD.
†Type of machine: in the derivation cohort, VSCAN was used in 5838 and Vivid-Q in 4123 of the children, whereas in validation
cohort, only Vivid-Q was used.
‡Comparison between derivation and validation cohorts.
Table 5. Demographic and Echocardiographic Characteristics of Children With Latent Rheumatic Heart Disease Comparing
Derivation and Prospective Cohorts
Variables*
Derivation Cohort (n=518) Prospective Cohort (n=227)
P Value†Borderline (n=460) Definite (n=58) Borderline (n=164) Definite (n=63)
Age at screening, y 13.8±2.6 14.5±2.7 11.6±2.6 11.8±2.9 0.001
Female sex, n (%) 273 (59.6) 34 (58.6) 98 (59.8) 39 (61.9) 0.826
Height, cm 159.7±9.9 156.5±4.9 153.3±11.2 154.5±12.6 0.002
Weight, kg 55.8±13.8 51.1±5.9 43.6±10.7 42.9±11.8 0.001
Echocardiographic variables included in the model
Mitral valve
Anterior leaflet thickening 127 (27.6) 50 (86.2) 33 (20.1) 45 (71.4) 0.960
Excessive leaflet tip motion 24 (5.2) 30 (51.7) 6 (3.7) 13 (20.6) 0.385
Regurgitation jet length ≥2 cm 356 (77.4) 50 (86.2) 135 (82.3) 54 (85.7) 0.580
Aortic valve
Irregular or focal thickening 1 (0.2) 10 (17.2) 3 (1.8) 11 (17.5) 0.014
Any regurgitation 81 (17.6) 21 (36.2) 29 (17.7) 16 (25.4) 0.967
*Data are expressed as number (percentage) or as a mean±SD.
†Comparison between derivation and prospective cohorts.
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
9. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 9
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
Simplified screening protocols have been proposed, all
excluding morphological criteria and simplifying the
functional criteria by only measuring regurgitation jet
length.10,13,22,43,44
Although this approach is practical,
it misses isolated morphological abnormalities that
can occur in the absence of pathological regurgitation
early in RHD.13
Additionally, full application of the 2012 WHF crite-
ria requires an echocardiogram with spectral Doppler,
only currently available on fully functional portable
echocardiography machines. The 5 components in-
cluded in our model (mitral regurgitation ≥2 cm,
thickening of the anterior mitral valve leaflet, exces-
sive motion of the anterior mitral leaflet, any aortic
regurgitation, and focal thickening of the aortic valve)
can be assessed by handheld echocardiography equip-
ment, which is substantially less expensive, conse-
quently more available, and has shown to have good
accuracy for latent RHD detection.21,22
Eliminating the
need for spectral Doppler also opens the door for
1-stage screening (without confirmatory follow-up)
for diagnosis, which could reduce costs and improve
the practicality of active RHD case finding.
A refinement set of echocardiographic criteria for
the diagnosis of RHD is fundamental in the context of
global screening efforts. Our findings of simplified crite-
ria may improve the practicality and affordability of ech-
ocardiographic screening programs, particularly using
Figure 4. The probability of unfavorable outcome according to the total score.
The vertical dashed lines represent the cutoff values for low-, intermediate-, and high-risk groups.
Figure 5. Rheumatic heart disease (RHD) status during the follow-up
in the independent cohort with latent RHD according to risk categories
of the score.
During a median follow-up of 2.3 y, 34 had disease progression (3 deaths), 95
remained stable, and 98 demonstrated disease regression.
Figure 6. Predicted vs observed (open bars) unfavorable outcome for
increments in the risk score categories in the external cohort with la-
tent rheumatic heart disease.
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
10. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 10
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
handheld echocardiography devices by nonexperts in
resource-limited environments. Our study overcomes a
major barrier for implementation of echocardiography-
based screening with potential to incorporate this strat-
egy for active surveillance of RHD in endemic areas.
Another major barrier in echocardiographic screen-
ing for RHD remains the lack of data on prognosis af-
ter case detection.51
Follow-up of children diagnosed
with latent RHD has shown that ≈40% of children
with borderline and definite RHD will show disease im-
provement, often reverting to normal, over a period of
only 1 to 2 years irrespective of penicillin prophylaxis.40
Currently, other than severity of valvular dysfunction
at diagnosis,11
there are no reliable markers to predict
individual risk.
This concern is even more pronounced in children
classified as having borderline RHD, as there is almost
certainly some degree of overlap with normal findings
and the reliability of the diagnosis has been debat-
ed.14,17,36,55,56
Although many children do well, longitudi-
nal data have also shown that a diagnosis of borderline
RHD does confer increased risk of acute rheumatic fe-
ver and of adverse cardiac events.36
The present study has several implications, not only
on epidemiological aspects of the disease but also on
the direct approach of the children and their families.
Although there is no randomized trial, the decision to
initiate secondary prophylaxis at diagnosis would be
based on the risk of disease progression, which we be-
lieve is the major contribution of our study to the cur-
rent state of knowledge of RHD.
This RHD risk score developed in this study provides
the first tool for risk stratification at time of diagnosis
of latent RHD. Through a retrospective analysis of large
screening databases, the score was able to stratify risk
in children with both borderline RHD and definite RHD,
with children in the highest risk category showing sig-
nificantly more progression over only 2 to 3 years.
Additionally, the more accurate prediction of disease
progression may also improve the cost-effectiveness of
population screening. Indeed, the prevalence of RHD
relies on echocardiographic criteria because subclinical
disease is 7 to 8× higher than that of clinically mani-
fest disease, which is characterized by the presence of
symptoms related to valvular dysfunction, mainly heart
failure or embolic events or pathological valvular heart
murmur detected during auscultation.57
Limitations
Our study has some limitations. First, this study has an
inherent particularity, which is the lack of gold stand-
ard diagnostic test for RHD. However, a model based
on a simplified set of criteria derived differently from
WHF criteria was our strategy of dealing with the lack
of a gold standard. Additionally, to overcome this lim-
itation, the set of criteria was applied to an independ-
ent longitudinal cohort with latent RHD and showed
an accurate prediction of disease outcome, which is
the key finding of the study.
Second, model and score development was under-
taken retrospectively, utilizing existing cohorts. Features
of latent RHD differ between populations, and our der-
ivation cohort did not have any cases of mitral stenosis
or aortic valve prolapse. However, these more severe
phenotypic presentations do not occur in isolation, typi-
cally being found with other features, restrictive motion
of the mitral and aortic regurgitation, as examples. Our
validation cohort did contain patients with these fea-
tures, and discrimination between definite RHD and
normal remained nearly perfect. Additionally, mitral ste-
nosis is a late finding in RHD, and while undiagnosed
Figure 7. Kaplan-Meier curves for
progression-free survival rate according to
risk categories.
Progression-free survival was significantly differ-
ent among the 3 different risk groups (P 0.001
for each). The 3-y progression-free survivals
were 93%, 70%, and 47% in the low-, inter-
mediate-, and high-risk groups, respectively.
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
11. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 11
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
children with mitral stenosis are detected through
screening, particularly in the lowest resource environ-
ments, the high risk of adverse outcomes in these chil-
dren is well established.35,40
Score performance was
accurate in younger patients earlier in the course of
their disease but needs to be tested in other popula-
tions. However, the use of data from both Brazil and
Uganda increases the chances of reproducibility.
Third, the natural history of RHD varies according
to environmental exposure to streptococcal infections,
genetics, age at diagnosis, and other incompletely un-
derstood factors. Additionally, disease progression was
defined based on longitudinal echocardiograms that
were reinterpreted in series, as it is the standard method
in clinical practice to determine an individual’s out-
come. However, there is a level of uncertainty around
the definition of disease progression or improvement
using echocardiographic definitions, especially related
to measurement variability of morphological rheu-
matic features. Moreover, clinical disease progression,
which includes development of heart failure, the need
for valve intervention, and RHD-related complications,
was not assessed in this study. The definitive validation
of this echocardiographic score will require prospective
longitudinal studies in diverse populations.
Conclusions
We developed an accurate simple score for latent RHD
diagnosis based on the most commonly seen compo-
nent diagnoses in the 2012 WHF criteria. This score is
highly accurate to recognize definite RHD and provides
the first tool for risk stratification, assigning children to
low, intermediate, or high risk based on echocardio-
graphic features at diagnosis. In retrospective applica-
tion, the score performed well for prediction of adverse
outcomes. This new tool needs further validation but
has high potential value for improving the triage of chil-
dren identified with latent RHD, in explaining individual
risk to the child and family, and for influencing the de-
cision to initiate or defer secondary prophylaxis.
ARTICLE INFORMATION
Received April 22, 2018; accepted November 27, 2018.
The Data Supplement is available at https://www.ahajournals.org/doi/
suppl/10.1161/CIRCIMAGING.118.007928.
Correspondence
Maria Carmo P. Nunes, MD, PhD, Department of Internal Medicine, School of
Medicine of the Federal University of Minas Gerais, Av Prof Alfredo Balena,
190 Santa Efigênia, Belo Horizonte–MG 30130 100, Brazil. Email mcarmo@
waymail.com.br
Affiliations
Serviço de Cardiologia e Cirurgia Cardiovascular e Centro de Telessaúde do
Hospital das Clínicas da UFMG, Belo Horizonte, Minas Gerais, Brazil (M.C.P.N.,
B.R.N., A.C.D., K.K.B.O., A.L.P.R.). Department of Internal Medicine, School
of Medicine, Belo Horizonte, Minas Gerais, Brazil (M.C.P.N., B.R.N., A.L.P.R.).
Children’s National Health System, Washington, DC (C.S.). Statistical Depart-
ment, Universidade Federal de Minas Gerais, Belo Horizonte, Minas Gerais,
Brazil (E.M.d.L., E.A.C.). Statistical Department, Universidade Federal do
Paraná, Curitiba, Brazil (J.L.P.d.S.). Uganda Heart Institute, Kampala (E.O.,
T.A., P.L.). The Heart Institute, Cincinnati Children’s Hospital Medical Center,
OH (A.Z.B.).
Sources of Funding
Verizon supported and funded the rheumatic heart disease (RHD) screening
program in Brazil; General Electric Healthcare provided echocardiography e-
quipment; Vitel Net contributed to the development of the RHD cloud plat-
form; and Edwards Lifesciences funded the primary care screening program.
Foundation for Research Support of the State of Minas Gerais and National
Council for Scientific and Technological Development (CNPq) partly supported
the study. Drs Nunes, Colosimo, and Ribeiro were supported, in part, by CNPq.
Disclosures
None.
REFERENCES
1. Watkins DA, Johnson CO, Colquhoun SM, Karthikeyan G, Beaton A,
Bukhman G, Forouzanfar MH, Longenecker CT, Mayosi BM, Mensah GA,
Nascimento BR, Ribeiro ALP, Sable CA, Steer AC, Naghavi M, Mokdad
AH, Murray CJL, Vos T, Carapetis JR, Roth GA. Global, Regional, and
National Burden of rheumatic heart disease, 1990-2015. N Engl J Med.
2017;377:713–722. doi: 10.1056/NEJMoa1603693
2. Zühlke L, Engel ME, Karthikeyan G, Rangarajan S, Mackie P, Cupido B,
Mauff K, Islam S, Joachim A, Daniels R, Francis V, Ogendo S, Gitura B,
Mondo C, Okello E, Lwabi P, Al-Kebsi MM, Hugo-Hamman C, Sheta SS,
Haileamlak A, Daniel W, Goshu DY, Abdissa SG, Desta AG, Shasho BA,
Begna DM, ElSayed A, Ibrahim AS, Musuku J, Bode-Thomas F, Okeahialam
BN, Ige O, Sutton C, Misra R, Abul Fadl A, Kennedy N, Damasceno A, Sani
M, Ogah OS, Olunuga T, Elhassan HH, Mocumbi AO, Adeoye AM, Mntla
P, Ojji D, Mucumbitsi J, Teo K, Yusuf S, Mayosi BM. Characteristics, com-
plications, and gaps in evidence-based interventions in rheumatic heart
disease: the Global Rheumatic Heart Disease Registry (the REMEDY study).
Eur Heart J. 2015;36:1115–1122. doi: 10.1093/eurheartj/ehu449
3. Marijon E, Ou P, Celermajer DS, Ferreira B, Mocumbi AO, Jani D, Paquet C,
Jacob S, Sidi D, Jouven X. Prevalence of rheumatic heart disease detected
by echocardiographic screening. N Engl J Med. 2007;357:470–476. doi:
10.1056/NEJMoa065085
4. Marijon E, Celermajer DS, Tafflet M, El-Haou S, Jani DN, Ferreira B, Mo-
cumbi AO, Paquet C, Sidi D, Jouven X. Rheumatic heart disease screen-
ing by echocardiography: the inadequacy of World Health Organization
criteria for optimizing the diagnosis of subclinical disease. Circulation.
2009;120:663–668. doi: 10.1161/CIRCULATIONAHA.109.849190
5. Beaton A, Okello E, Lwabi P, Mondo C, McCarter R, Sable C. Echocardiog-
raphy screening for rheumatic heart disease in Ugandan schoolchildren.
Circulation. 2012;125:3127–3132. doi: 10.1161/CIRCULATIONAHA.
112.092312
6. Paar JA, Berrios NM, Rose JD, Cáceres M, Peña R, Pérez W, Chen-Mok
M, Jolles E, Dale JB. Prevalence of rheumatic heart disease in children
and young adults in Nicaragua. Am J Cardiol. 2010;105:1809–1814. doi:
10.1016/j.amjcard.2010.01.364
7. Engel ME, Haileamlak A, Zühlke L, Lemmer CE, Nkepu S, van de Wall
M, Daniel W, Shung King M, Mayosi BM. Prevalence of rheumatic heart
disease in 4720 asymptomatic scholars from South Africa and Ethiopia.
Heart. 2015;101:1389–1394. doi: 10.1136/heartjnl-2015-307444
8. Mirabel M, Fauchier T, Bacquelin R, Tafflet M, Germain A, Robil-
lard C, Rouchon B, Marijon E, Jouven X. Echocardiography screen-
ing to detect rheumatic heart disease: a cohort study of schoolchil-
dren in French Pacific Islands. Int J Cardiol. 2015;188:89–95. doi:
10.1016/j.ijcard.2015.04.007
9. Nascimento BR, Beaton AZ, Nunes MC, Diamantino AC, Carmo GA,
Oliveira KK, Oliveira CM, Meira ZM, Castilho SR, Lopes EL, Castro IM,
Rezende VM, Chequer G, Landay T, Tompsett A, Ribeiro AL, Sable C;
PROVAR (Programa de Rastreamento da Valvopatia Reumática) Investiga-
tors. Echocardiographic prevalence of rheumatic heart disease in Brazilian
schoolchildren: data from the PROVAR study. Int J Cardiol. 2016;219:439–
445. doi: 10.1016/j.ijcard.2016.06.088
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
12. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 12
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
10. Roberts K, Maguire G, Brown A, Atkinson D, Reményi B, Wheaton G, Kelly
A, Kumar RK, Su JY, Carapetis JR. Echocardiographic screening for rheu-
matic heart disease in high and low risk Australian children. Circulation.
2014;129:1953–1961. doi: 10.1161/CIRCULATIONAHA.113.003495
11. Cannon J, Roberts K, Milne C, Carapetis JR. Rheumatic heart disease se-
verity, progression and outcomes: a multi-state model. J Am Heart Assoc.
2017;6:e003498. doi: 10.1161/JAHA.116.003498
12. Gewitz MH, Baltimore RS, Tani LY, Sable CA, Shulman ST, Carapetis J,
Remenyi B, Taubert KA, Bolger AF, Beerman L, Mayosi BM, Beaton A,
Pandian NG, Kaplan EL; American Heart Association Committee on Rheu-
matic Fever, Endocarditis, and Kawasaki Disease of the Council on Cardi-
ovascular Disease in the Young. Revision of the Jones Criteria for the di-
agnosis of acute rheumatic fever in the era of Doppler echocardiography:
a scientific statement from the American Heart Association. Circulation.
2015;131:1806–1818. doi: 10.1161/CIR.0000000000000205
13. Dougherty S, Khorsandi M, Herbst P. Rheumatic heart disease screening:
current concepts and challenges. Ann Pediatr Cardiol. 2017;10:39–49.
doi: 10.4103/0974-2069.197051
14. ZühlkeL,MayosiBM.Echocardiographicscreeningforsubclinicalrheumatic
heart disease remains a research tool pending studies of impact on prog-
nosis. Curr Cardiol Rep. 2013;15:343. doi: 10.1007/s11886-012-0343-1
15. Reményi B, Wilson N, Steer A, Ferreira B, Kado J, Kumar K, Lawrenson J,
Maguire G, Marijon E, Mirabel M, Mocumbi AO, Mota C, Paar J, Saxena
A, Scheel J, Stirling J, Viali S, Balekundri VI, Wheaton G, Zühlke L, Cara-
petis J. World Heart Federation criteria for echocardiographic diagnosis of
rheumatic heart disease–an evidence-based guideline. Nat Rev Cardiol.
2012;9:297–309. doi: 10.1038/nrcardio.2012.7
16. Clark BC, Krishnan A, McCarter R, Scheel J, Sable C, Beaton A. Using a
low-risk population to estimate the specificity of the World Heart Fed-
eration criteria for the diagnosis of rheumatic heart disease. J Am Soc
Echocardiogr. 2016;29:253–258. doi: 10.1016/j.echo.2015.11.013
17. Colquhoun SM, Kado JH, Remenyi B, Wilson NJ, Carapetis JR, Steer
AC. Echocardiographic screening in a resource poor setting: border-
line rheumatic heart disease could be a normal variant. Int J Cardiol.
2014;173:284–289. doi: 10.1016/j.ijcard.2014.03.004
18. Rossi E, Felici AR, Banteyrga L. Subclinical rheumatic heart disease in an
Eritrean high-school population, detected by echocardiography. J Heart
Valve Dis. 2014;23:235–239.
19. Ledos PH, Kamblock J, Bourgoin P, Eono P, Carapetis JR. Prevalence of
rheumatic heart disease in young adults from New Caledonia. Arch Car-
diovasc Dis. 2015;108:16–22. doi: 10.1016/j.acvd.2014.07.053
20. Ngaïdé AA, Mbaye A, Kane A, Ndiaye MB, Jobe M, Bodian M, Dioum M,
Sarr SA, Aw F, Mbakop PS, Ba FG, Gaye ND, Tabane A, Bah MB, Coly SM,
Diagne D, Diack B, Diao M, Kane A. Prevalence of rheumatic heart disease
in Senegalese school children: a clinical and echocardiographic screening.
Heart Asia. 2015;7:40–45. doi: 10.1136/heartasia-2015-010664
21. Beaton A, Lu JC, Aliku T, Dean P, Gaur L, Weinberg J, Godown J, Lwabi P,
Mirembe G, Okello E, Reese A, Shrestha-Astudillo A, Bradley-Hewitt T, Scheel
J, Webb C, McCarter R, Ensing G, Sable C. The utility of handheld echo-
cardiography for early rheumatic heart disease diagnosis: a field study. Eur
Heart J Cardiovasc Imaging. 2015;16:475–482. doi: 10.1093/ehjci/jeu296
22. Mirabel M, Bacquelin R, Tafflet M, Robillard C, Huon B, Corsenac P, de
Fremicourt I, Narayanan K, Meunier JM, Noel B, Hagege AA, Rouchon
B, Jouven X, Marijon E. Screening for rheumatic heart disease: evalua-
tion of a focused cardiac ultrasound approach. Circ Cardiovasc Imaging.
2015;8:e002324. doi: 10.1161/CIRCIMAGING.114.002324
23. Webb RH, Gentles TL, Stirling JW, Lee M, O’Donnell C, Wilson NJ. Val-
vular regurgitation using portable echocardiography in a healthy student
population: implications for rheumatic heart disease screening. J Am Soc
Echocardiogr. 2015;28:981–988. doi: 10.1016/j.echo.2015.03.012
24. Spitzer E, Mercado J, Islas F, Rothenbühler M, Kurmann R, Zürcher F,
Krähenmann P, Llerena N, Jüni P, Torres P, Pilgrim T. Screening for rheumatic
heart disease among Peruvian children: a two-stage sampling observational
study. PLoS One. 2015;10:e0133004. doi: 10.1371/journal.pone.0133004
25. Nair B, Viswanathan S, Koshy AG, Gupta PN, Nair N, Thakkar A. Rheu-
matic heart disease in Kerala: a vanishing entity? An echo doppler study
in 5-15-years-old school children. Int J Rheumatol. 2015;2015:930790.
http://dx.doi.org/10.1155/2015/930790
26. Engelman D, Kado JH, Reményi B, Colquhoun SM, Carapetis JR, Wilson
NJ, Donath S, Steer AC. Screening for rheumatic heart disease: quality and
agreement of focused cardiac ultrasound by briefly trained health workers.
BMC Cardiovasc Disord. 2016;16:30. doi: 10.1186/s12872-016-0205-7
27. Sims Sanyahumbi A, Sable CA, Beaton A, Chimalizeni Y, Guffey D, Hos-
seinipour M, Karlsten M, Kazembe PN, Kennedy N, Minard CG, Penny
DJ. School and community screening shows Malawi, Africa, to have a
high prevalence of latent rheumatic heart disease. Congenit Heart Dis.
2016;11:615–621. doi: 10.1111/chd.12353
28. Engelman D, Kado JH, Reményi B, Colquhoun SM, Carapetis JR, Don-
ath S, Wilson NJ, Steer AC. Focused cardiac ultrasound screening for
rheumatic heart disease by briefly trained health workers: a study of
diagnostic accuracy. Lancet Glob Health. 2016;4:e386–e394. doi:
10.1016/S2214-109X(16)30065-1
29. Aliku T, Sable C, Scheel A, Tompsett A, Lwabi P, Okello E, McCarter R, Sum-
mar M, Beaton A. Targeted echocardiographic screening for latent rheu-
matic heart disease in Northern Uganda: evaluating familial risk following
identification of an index case. PLoS Negl Trop Dis. 2016;10:e0004727.
doi: 10.1371/journal.pntd.0004727
30. Yadeta D, Hailu A, Haileamlak A, Gedlu E, Guteta S, Tefera E, Tigabu
Z, Tesfaye H, Daniel W, Mekonnen D, Zelalem M, Tekleab Y, Alemayehu
B, Mekonnen D, Azazh A, Moges T, Hailu T, Abdosh T, Yusuf N, Ayele
D, Wuobshet K, Tadele H, Lemma K, Shiferaw S, Giday A, Mekonnen
D, Alemu G, Zühlke L, Allison TG, Nkomo VT, Engel ME. Prevalence of
rheumatic heart disease among school children in Ethiopia: a multisite
echocardiography-based screening. Int J Cardiol. 2016;221:260–263. doi:
10.1016/j.ijcard.2016.06.232
31. Shrestha NR, Karki P, Mahto R, Gurung K, Pandey N, Agrawal K, Rothen-
bühler M, Urban P, Jüni P, Pilgrim T. Prevalence of subclinical rheumatic
heart disease in Eastern Nepal: a school-based cross-sectional study. JAMA
Cardiol. 2016;1:89–96. doi: 10.1001/jamacardio.2015.0292
32. Saxena A, Desai A, Narvencar K, Ramakrishnan S, Thangjam RS, Kulkarni
S, Jacques’ E Costa AK, Mani K, Dias A, Sukharamwala R. Echocardio-
graphic prevalence of rheumatic heart disease in Indian school children
using World Heart Federation criteria - A multi site extension of RHEU-
MATIC study (the e-RHEUMATIC study). Int J Cardiol. 2017;249:438–442.
doi: 10.1016/j.ijcard.2017.09.184
33. Beaton A, Okello E, Aliku T, Lubega S, Lwabi P, Mondo C, McCarter R,
Sable C. Latent rheumatic heart disease: outcomes 2 years after echo-
cardiographic detection. Pediatr Cardiol. 2014;35:1259–1267. doi:
10.1007/s00246-014-0925-3
34. Bhaya M, Beniwal R, Panwar S, Panwar RB. Two years of follow-up vali-
dates the echocardiographic criteria for the diagnosis and screening of
rheumatic heart disease in asymptomatic populations. Echocardiography.
2011;28:929–933. doi: 10.1111/j.1540-8175.2011.01487.x
35. Engelman D, Wheaton GR, Mataika RL, Kado JH, Colquhoun SM, Remenyi B,
Steer AC. Screening-detected rheumatic heart disease can progress to severe
disease. Heart Asia. 2016;8:67–73. doi: 10.1136/heartasia-2016-010847
36. Rémond M, Atkinson D, White A, Brown A, Carapetis J, Remenyi B, Ro-
berts K, Maguire G. Are minor echocardiographic changes associated with
an increased risk of acute rheumatic fever or progression to rheumatic
heart disease? Int J Cardiol. 2015;198:117–122. doi: 10.1016/j.ijcard.
2015.07.005
37. Saxena A, Ramakrishnan S, Roy A, Seth S, Krishnan A, Misra P, Kalaivani
M, Bhargava B, Flather MD, Poole-Wilson PP. Prevalence and outcome of
subclinical rheumatic heart disease in India: the RHEUMATIC (Rheumatic
Heart Echo Utilisation and Monitoring Actuarial Trends in Indian Children)
study. Heart. 2011;97:2018–2022. doi: 10.1136/heartjnl-2011-300792
38. Zühlke L, Engel ME, Lemmer CE, van de Wall M, Nkepu S, Meiring A,
Bestawros M, Mayosi BM. The natural history of latent rheumatic heart di-
sease in a 5 year follow-up study: a prospective observational study. BMC
Cardiovasc Disord. 2016;16:46. doi: 10.1186/s12872-016-0225-3
39. Bertaina G, Rouchon B, Huon B, Guillot N, Robillard C, Noël B, Nadra
M, Tribouilloy C, Marijon E, Jouven X, Mirabel M. Outcomes of border-
line rheumatic heart disease: a prospective cohort study. Int J Cardiol.
2017;228:661–665. doi: 10.1016/j.ijcard.2016.11.234
40. Beaton A, Aliku T, Dewyer A, Jacobs M, Jiang J, Longenecker CT, Lubega
S, McCarter R, Mirabel M, Mirembe G, Namuyonga J, Okello E, Scheel
A, Tenywa E, Sable C, Lwabi P. Latent rheumatic heart disease: identi-
fying the children at highest risk of unfavorable outcome. Circulation.
2017;136:2233–2244. doi: 10.1161/CIRCULATIONAHA.117.029936
41. Nascimento BR, Beaton AZ, Diamantino AC, Maria do Carmo PN, Reese
AT, Oliveira KK, Carmo GA, Lourenço TV, Ruiz GZ, Rabelo LMM. Com-
parison between different strategies of rheumatic heart disease echocar-
diographic screening in brazil: data from the PROVAR study. J Am Heart
Assoc. 2018;7:e008039. doi: 10.1161/JAHA.117.008039
42. Beaton A, Aliku T, Okello E, Lubega S, McCarter R, Lwabi P, Sable C. The u-
tility of handheld echocardiography for early diagnosis of rheumatic heart
disease. J Am Soc Echocardiogr. 2014;27:42–49. doi: 10.1016/j.echo.
2013.09.013
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
13. Circ Cardiovasc Imaging. 2019;12:e007928. DOI: 10.1161/CIRCIMAGING.118.007928 February 2019 13
Nunes et al; Echocardiography Score for Rheumatic Heart Disease
43. Beaton A, Nascimento BR, Diamantino AC, Pereira GT, Lopes EL, Miri
CO, Bruno KK, Chequer G, Ferreira CG, Lafeta LC, Richards H, Perlman L,
Webb CL, Ribeiro AL, Sable C, Nunes Mdo C. Efficacy of a standardized
computer-based training curriculum to teach echocardiographic iden-
tification of rheumatic heart disease to nonexpert users. Am J Cardiol.
2016;117:1783–1789. doi: 10.1016/j.amjcard.2016.03.006
44. Ploutz M, Lu JC, Scheel J, Webb C, Ensing GJ, Aliku T, Lwabi P, Sa-
ble C, Beaton A. Handheld echocardiographic screening for rheu-
matic heart disease by non-experts. Heart. 2016;102:35–39. doi:
10.1136/heartjnl-2015-308236
45. Vatcheva KP, Lee M, McCormick JB, Rahbar MH. Multicollinearity in re-
gression analyses conducted in epidemiologic studies. Epidemiology (Sun-
nyvale). 2016;6:227. doi: 10.4172/2161-1165.1000227
46. Woodward M, Tunstall-Pedoe H, Peters SA. Graphics and statistics for
cardiology: clinical prediction rules. Heart. 2017;103:538–545. doi:
10.1136/heartjnl-2016-310210
47. Hosmer DW, Hosmer T, Le Cessie S, Lemeshow S. A comparison of
goodness-of-fit tests for the logistic regression model. Stat Med.
1997;16:965–980.
48. Smith GC, Seaman SR, Wood AM, Royston P, White IR. Correcting for op-
timistic prediction in small data sets. Am J Epidemiol. 2014;180:318–324.
doi: 10.1093/aje/kwu140
49. Pavlou M, Ambler G, Seaman SR, Guttmann O, Elliott P, King M, Omar RZ.
How to develop a more accurate risk prediction model when there are few
events. BMJ. 2015;351:h3868. doi: 10.1136/bmj.h3868
50. Steyerberg EW. Clinical Prediction Models: A Practical Approach to Devel-
opment, Validation, and Updating. New York, NY: Springer; 2009.
51. Hunter LD, Monaghan M, Lloyd G, Pecoraro AJK, Doubell AF, Herbst PG.
Screening for rheumatic heart disease: is a paradigm shift required? Echo
Res Pract. 2017;4:R43–R52. doi: 10.1530/ERP-17-0037
52. Saxena A, Zühlke L, Wilson N. Echocardiographic screening for rheu-
matic heart disease: issues for the cardiology community. Glob Heart.
2013;8:197–202. doi: 10.1016/j.gheart.2013.08.004
53. Tani LY. Echocardiographic screening for rheumatic heart disease. Circula-
tion. 2014;129:1912–1913. doi: 10.1161/CIRCULATIONAHA.114.009406
54. Roberts K, Colquhoun S, Steer A, Reményi B, Carapetis J. Screening for
rheumatic heart disease: current approaches and controversies. Nat Rev
Cardiol. 2013;10:49–58. doi: 10.1038/nrcardio.2012.157
55. Bacquelin R, Tafflet M, Rouchon B, Guillot N, Marijon E, Jouven X, Mi-
rabel M. Echocardiography-based screening for rheumatic heart disease:
what does borderline mean? Int J Cardiol. 2016;203:1003–1004. doi:
10.1016/j.ijcard.2015.11.110
56. Herbst P. Screening for asymptomatic rheumatic heart disease: un-
derstanding the mechanisms key to the diagnostic criteria. SA Heart.
2015;12:134–144. doi:10.24170/12-3-1716
57. Rothenbühler M, O’Sullivan CJ, Stortecky S, Stefanini GG, Spitzer E, Estill
J, Shrestha NR, Keiser O, Jüni P, Pilgrim T. Active surveillance for rheumatic
heart disease in endemic regions: a systematic review and meta-analy-
sis of prevalence among children and adolescents. Lancet Glob Health.
2014;2:e717–e726. doi: 10.1016/S2214-109X(14)70310-9
Downloadedfromhttp://ahajournals.orgbyonFebruary3,2019
