HHS Public Access

Author manuscript
Circulation. Author manuscript; available in PMC 2018 December 05.
Author Manuscript

Published in final edited form as:

Circulation. 2017 December 05; 136(23): 2233–2244. doi:10.1161/CIRCULATIONAHA.117.029936.

Latent Rheumatic Heart Disease: Identifying the Children at

Highest Risk of Unfavorable Outcome
Andrea Beaton, MD1, Twalib Aliku, MMed2, Alyssa Dewyer, BS1, Marni Jacobs, PhD, MPH1,
Jiji Jiang, MD, MS1, Chris T Longenecker, MD3, Sulaiman Lubega, MMed4, Robert McCarter,
ScD1, Mariana Mirabel, MD5, Grace Mirembe, MMed6, Judith Namuyonga, MMed4, Emmy
Okello, MMed,PhD4, Amy Scheel, BS1, Emmanuel Tenywa, MMed4, Craig Sable, MD1, and
Author Manuscript

Peter Lwabi, MMed4

1Children’s
National Health System, Washington DC USA 2Gulu University, Gulu Uganda 3Case
Western Reserve University, Cleveland Ohio USA 4Uganda Heart Institute, Kampala, Uganda
5INSERM U970, Paris Cardiovascular Research Center PARCC, Paris, France 6Joint Clinical

Research Centers, Kampala Uganda

Abstract
Background—Screening echocardiography (echo) has emerged as a potentially powerful tool
for early diagnosis of rheumatic heart disease (RHD). The utility of screening echo hinges on the
rate of RHD progression and the ability of penicillin prophylaxis to improve outcome. We report
the longitudinal outcomes of a cohort of children with latent RHD and identify risk factors for
Author Manuscript

unfavorable outcomes.

Methods—This was a prospective natural history study conducted under the Ugandan RHD
registry. Children with latent RHD and ≥1 year of follow-up were included. All echos were re-
reviewed by experts (2012 WHF criteria) for inclusion and evidence of change. Bi- and multi-
variable logistic regression, Kaplan-Meier analysis, as well as Cox proportional hazard models
were developed to search for risk factors for unfavorable outcome and compare progression-free
survival between those treated and not treated with penicillin. Propensity and other matching
methods with sensitivity analysis were implemented for the evaluation of the penicillin effect.

Results—Blinded review confirmed 227 cases of latent RHD: 164 borderline and 63 definite (42
mild, 21 moderate/severe). Median age at diagnosis was 12 years and median follow-up was 2.3
years (IQR 2.0–2.9). Penicillin prophylaxis was prescribed in 49.3% with overall adherence of
Author Manuscript

84.7%. Of children with moderate/severe definite RHD, 47.6% had echo progression (including 2
deaths), and 9.5% echo regression. Children with mild definite and borderline RHD showed 26%
and 9.8% echo progression and 45.2% and 46.3% echo improvement respectively. Of those with
mild definite RHD or borderline RHD, more advanced disease category, younger age, and
morphological mitral valve features were risk factors for an unfavorable outcome.

Correspondence: Andrea Beaton, Children’s National Medical Center, Division of Cardiology, 111 Michigan Avenue NW, Washington
DC, 20010, Fax: 202-476-5700 Phone: 202-476-4189 abeaton@childrennational.org.
CONFLICTS OF INTEREST DISCLOSURES: None
 Beaton et al. Page 2

Conclusions—Latent RHD is a heterogeneous diagnosis with variable disease outcomes.

Author Manuscript

Children with moderate to severe latent RHD have poor outcomes. Children with both borderline
and mild definite RHD are at substantial risk of progression. While long-term outcome remains
unclear, the initial change in latent RHD may be evident during the first 1–2 years following
diagnosis. Natural history data is inherently limited and a randomized clinical trial is needed to
definitively determine the impact of penicillin prophylaxis on the trajectory of latent RHD.

Keywords
Rheumatic heart disease; screening; pediatric; outcomes reserach

INTRODUCTION
RHD is the most common acquired cardiovascular disease among children and young adults.
Author Manuscript

In 2015, there were an estimated 33.4 million prevalent clinical cases with 319,400 deaths
and 10.5 million disability adjusted life years1. The global distribution of RHD is uneven
and strongly associated with conditions of social deprivation. The highest age-standardized
RHD death and prevalence rates are found in South Asia, Oceania, and Central Sub-Saharan
African populations1 though the disease remains endemic in most low and middle income
countries (LIMC)2.

A recent multinational registry (REMEDY) in LIMC showed most RHD patients have
advanced RHD (63.9%) and complications at time of diagnosis3. However, RHD is a
cumulative process and opportunities exist for early intervention. While the initial episode(s)
of acute rheumatic fever (ARF)/RHD almost exclusively occur in childhood, RHD most
commonly presents in adolescents and young adults4, 5. The latent period between ARF and
clinically apparent RHD presents an opportunity for early intervention.
Author Manuscript

Screening echocardiography (echo) has emerged as a potentially powerful tool for early
RHD detection6–8. RHD diagnosed through screening echo has been termed latent RHD,
and classified as borderline or definite9. In screening over 15,000 school children in Uganda,
our team has detected a 3% prevalence of latent RHD10, 11. These findings have been
replicated in other parts of Africa12, 13, Asia14, 15, South and Central America8, 16 and the
Pacific17–20 – uncovering a large population that may benefit from early detection.

The utility of screening echo in RHD endemic populations hinges on two critical
parameters: (1) the rate of RHD progression and (2) the ability of Benzathine penicillin G
(BPG) prophylaxis to improve outcome, a critical requirement of evaluating the suitability of
any condition for screening21. Policy decisions on the role of screening for RHD have
Author Manuscript

stalled because of the lack of high quality data to guide management of latent RHD.

Natural history studies have the potential to estimate the rate of latent RHD progression. In
addition, in the absence of a randomized controlled trial, natural history studies may provide
information on the ability of penicillin to modify disease course. There have been eight
previous published natural history studies of latent RHD8, 22–28. Three of these were
published prior to the publication of standardized criteria for latent RHD diagnosis8, 23, 26,
making them difficult to compare. The remaining 5 studies22, 24, 25, 27, 28 report on relatively

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 3

small cohorts, ranging from 25–55 children limiting the power to detect features that predict
Author Manuscript

progression or regression of disease.

We have studied the largest cohort of children with latent RHD to date, with the hypothesis
that risk factors for unfavorable outcome can be found through study of natural history data.
The included children are part of the Ugandan National RHD registry (The Registry),
enrolled during previous research studies on RHD prevalence10 and use of handheld
echo11, 29, 30, and during RHD outreach and an epidemiological study (unpublished) in
Uganda.

METHODS
Setting
This study took place in Uganda under the auspices of the Ugandan National RHD Registry
Author Manuscript

(The Registry), which is IRB approved by Makerere University, Kampala Uganda, Case
Western Reserve University, Cleveland Ohio, and Children’s National Health System,
Washington DC. The Registry was established in 2010 to capture the burden of clinical RHD
presenting to the Ugandan Heart Institute, the single tertiary cardiology center in Uganda,
and to support children with latent RHD detected thorough school-based screening. In 2015
The Registry expanded to the North (Gulu) where children with latent RHD detected
through school screening were added. The majority of children in this study come from two
large cohorts – children screened as part of the 2011 epidemiological study in Kampala, the
Ugandan capitol and largest city10 (previous 2-year outcomes of this cohort reported22) and
children screened as part of ongoing collaborative research and clinical outreach in Gulu
(2013-present)11, 30 (Figure 1).
Author Manuscript

Participants
Any child with latent RHD (borderline or definite RHD according to the 2012 WHF
criteria9) enrolled in The Registry and with at least 1 year of follow-up data was eligible for
study participation. In addition, a 3-month effort was made to contact patients who had been
lost to clinical follow-up (no follow-up in last 6 months for definite RHD and no follow-up
in last 12 months for borderline RHD) and bring them back for regular clinical care and
registry enrollment (if not previously enrolled). Written informed consent and/or assent (≥ 8
years) for this project was obtained from all children and families during registry enrollment.

Follow-up Assessment
The Registry captures demographic, socioeconomic, and clinical information, including
cardiovascular symptoms, physical exam, medication prescription and adherence at
Author Manuscript

enrollment. Historical data are captured when available and prospective data are gathered at
follow-up visits. Registry staff routinely contacts patients who have lapsed from clinical
follow-up, and if reached, the reason for absence is noted, including death. For this study,
the following variables were abstracted from the first follow-up after screening: age, sex,
registry site, and length of follow-up (days). For study purposes, prescription of every-28-
day intramuscular Benzathine penicillin G BPG, left at the discretion of the treating provider
(Ugandan cardiologist), was determined at initial follow-up visit. Adherence was based on

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 4

days of coverage (each injection providing 28 days of coverage) when possible and when not
Author Manuscript

possible, as a percentage of prescribed injections received.

Echocardiographic Reporting
A pediatric cardiologist with expertise in RHD (AB, CS) blindly re-reported on the original
first post-screening follow-up echos to determine study eligibility. These studies and all
follow-up studies were performed on a high-end echo machine (General Electric VIVID Q,
Milwaukee, WI or Philips iE33, Best, Holland); results are stored on the Children’s National
Health System core echo laboratory PACS system (Philips Xcelera, Best Holland).
Diagnosis was made strictly according to the 2012 WHF criteria9 and reported as definite
RHD, borderline RHD, congenital or other acquired heart disease, or normal. Definite and
borderline RHD were further classified into WHF sub-categories (A–C borderline, A–D
definite)9. Mitral and aortic regurgitation and mitral stenosis were classified as trivial (only
Author Manuscript

regurgitation), mild, moderate, or severe according to published standards31, 32. Children

with definite RHD were classified by severity of regurgitation at the most affected valve
(mild, moderate, severe), or severe if any mitral stenosis was present at baseline. Children
with non-specific valvular abnormalities not meeting criteria for RHD by the 2012 WHF
criteria9 were excluded, as outcomes for these children have been previously reported as
benign22, 24.

All available longitudinal echos (typically every 6 months in definite RHD and every 12
months in borderline RHD) were re-interpreted in series (AB, CS). Serial reporting was
chosen, as it is the standard method in clinical practice to determine an individual’s
longitudinal outcome. Cases were assessed for echocardiographic progression or regression
and the time to event (days) was noted.
Author Manuscript

Echocardiographic progression was defined as:

• A worsening in diagnostic category (borderline to definite)

• A worsening in the severity of regurgitation at the mitral or aortic valve (none or

physiologic, mild, moderate, severe)

• Development of new mitral stenosis or an increase in grade of mitral stenosis

(mild, moderate, severe)

• Death due to complication of RHD

Echocardiographic regression was defined as:

• An improvement in diagnostic category (definite, borderline, normal)

Author Manuscript

• An improvement in severity of regurgitation (severe, moderate, mild,

physiologic, none)

Reasons for echocardiographic progression and regression were noted and time to first echo
noting change recorded. Cases that did not meet these definitions for progression or
regression were considered stable. If a child showed a non-linear outcome (example: initial
worsening and then return to baseline), the findings of the last available echo were utilized to
determine final outcome.

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 5

Since echocardiograms reviews in series were unblinded, two measures of agreement were
Author Manuscript

undertaken to ensure that this procedure did not bias results. First, a subset of randomly
selected echos was re-reviewed by a blinded outside cardiologist (MM) with expertise in the
2012 WHF criteria and in latent RHD to calculate inter-rater agreement. Second, a subset of
final echos were re-reviewed blindly by the original readers to calculate intra-reviewer
agreement on final diagnosis.

Statistical Analysis
Propensity score and other bias-reducing analyses were implemented in Stata (Version 14);
all other analyses were performed using SAS (Version 9.3). Descriptive statistics were used
to summarize disease progression vs. regression by overall category and sub-category.
Cohen’s κappa statistic was used to evaluate the inter-rater agreement of diagnosis.
Moderate and severe definite RHD, including all patients with mitral stenosis, were
Author Manuscript

considered “missed clinical” RHD cases rather than true latent RHD. These “missed
clinical” cases were excluded from risk factor analysis.

Borderline and mild definite RHD group were combined for risk factor analysis utilizing
Fisher’s exact test for categorical variables and the Wilcoxon-Mann-Whitney test for
continuous, non-normally distributed, variables. The outcome categories stable, progression,
regression were dichotomized in 2 ways in subsequent analyses (Table 1). Outcome 1
compared progression vs. the collapsed favorable outcomes stable or regression. The
definition of outcome 2 varied by category of RHD. For mild definite cases, progression or
stable was compared to regression (with the rationale that stable definite RHD has clinical
significance). For borderline cases, progression was compared to the collapsed favorable
outcome stable or regression (with the rationale that stable borderline RHD is of little
immediate consequence). Bivariate logistic regression was used to identify individual risk
Author Manuscript

factors for progression, and factors that were significant at p<0.10 were included as
covariables in multivariable models. Kaplan Meier curves and log-rank test were utilized to
compare the progression free survival distributions of borderline and mild definite RHD.

Following logistic modeling, multivariable Cox proportional hazards modeling was

conducted for outcome 1, adjusting for covariables, to determine risk factors for progression
over the time period studied. To substantially reduce differences between the groups
receiving and not receiving penicillin prophylaxis at baseline, propensity score matched as
well as other forms of matched analysis were implemented. The other forms included
regression-based adjustment for confounders, inverse probability weighting and nearest-
neighbor matching. Sensitivity analyses were conducted to evaluate the likely impact of
uncontrolled differences between the groups receiving and not receiving penicillin.
Author Manuscript

RESULTS
Two hundred and eighty-two cases of latent RHD (> 1 year of follow-up) were identified in
The Registry, with 1,062 corresponding echos. On blinded re-review of the first follow-up
echos (after screening), 50 were determined to be non-specific valve disease not meeting the
2012 WHF criteria for RHD, 5 as other heart disease (4 mitral valve prolapse, 1 sub-aortic
membrane), and 227 as latent RHD: 63 definite RHD and 164 borderline RHD. Of those

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 6

with definite RHD, 42 were classified as mild and 21 as moderate-to-severe (Figure 2).
Author Manuscript

Details of echo diagnosis at baseline are shown in Table 2. Inter- and intra-reviewer
agreement based on the trinomial classification (normal/other, borderline, definite) was
substantial (Kappa (inter)=0.72, 95%CI 0.52–0.93, Kappa (intra)=0.78, 95% CI 0.67–0.88
95%)33.

The characteristics of the 227 children with confirmed latent RHD are detailed in Table 2.
Median age at diagnosis was 12 years, and 60.4% were female. The median duration of
follow-up was 2.4 years (range 1.1 to 5.9 years), with 26 children having greater than 5
years of follow-up. BPG was prescribed in 108 children (49.3%), including 59 of 63
children with definite RHD (93.6%) and 49 of 164 children with borderline RHD (29.9%).
Of those prescribed penicillin, the overall adherence (≥80% days covered or ≥ 80% injection
received) was quite good at 84.7%, with 89 children (82.4%) individually meeting the ≥
80% benchmark. There were no cases of ARF captured in the registry during the follow-up
Author Manuscript

period.

RHD Status
The outcomes of children with moderate and severe RHD were poor (Figure 3). Out of 21
children, only 2 (9.5%) showed echocardiographic regression during the study period. Of the
other 19 children, 9 (42.9%) remained with moderate/severe RHD, 8 (38.1%) progressed
and 2 (9.5%) died during the follow-up period (Figure 2). No child underwent a valve
intervention during the study period.

Among children with mild definite RHD (n=42), 11 (26.2%) children showed progression,
12 (28.6%) children remained stable, and 19 (45.2%) children showed echocardiographic
regression. Penicillin was prescribed to 39 out of 42 children with mild definite RHD
Author Manuscript

(92.9%) and 27 of 39 (69.2%) had adherence ≥ 80%, reflecting effective prophylaxis in

64.3% of children with mild definite RHD (27/42). The most common reasons for
progression included worsening of mitral regurgitation (54.5%) and worsening of aortic
regurgitation (36.4%). The most common reasons for regression included improvement in
mitral regurgitation (65.0%), and improvement of mitral valve morphology (35.0%). Full
details on progression and regression are provided in Table 3.

Among children with borderline RHD (n= 164), 16 (9.8%) showed progression, 72 (43.9%)
remained stable, and 76 (46.3%) showed echocardiographic regression. Penicillin was
prescribed to 49 children (29.9%) with borderline RHD, and 44 (90%) had adherence ≥
80%, reflecting effective prophylaxis in 26.8% of these children (44/164). The most
common reasons for progression included worsening of mitral regurgitation (56.3%) and
Author Manuscript

worsening of aortic regurgitation (36.4%). The most common reasons for regression
included resolution of abnormal mitral valve morphological findings (35.0%), and
improvement of mitral regurgitation (65.0%). The majority of initial progression occurred
during the first 2 years of follow-up. Children with mild definite RHD were more likely to
show progression than children with borderline RHD (p<0.01,Figure 4). Full details on
progression and regression are provided in Table 3.

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 7

There were 5 children who demonstrated a more complex pattern of change. Of these, 4
Author Manuscript

waivered between normal and borderline RHD based solely on small change in MR jet
length (above and below the 2cm cutoff). In these cases, there was no clinically meaningful
change, but only variation in diagnostic category due to strict application of the 2012 WHF
criteria. In the final case, an initial echo showed definite RHD (thickened anterior mitral
leaflet and excessive anterior mitral leaflet motion as well as mild MR). On second echo, the
excessive motion had resolved (then borderline because <2 morphological criteria were
met), and then on echos 3–5, the posterior mitral leaflet became restricted. It is likely in this
child, that this pattern of change simply represents natural RHD progression.

Risk Factors
As penicillin prophylaxis was prescribed to only one-third of children with borderline RHD,
before undertaking risk factor analysis, we evaluated whether there were any
Author Manuscript

echocardiographic or clinical differences between children with borderline RHD who were
or were not initially prescribed prophylaxis. Length of mitral regurgitation was found to be
associated with penicillin prescription (p=0.03), while year of diagnosis, sub-category of
borderline RHD (A–C), presence of aortic insufficiency, and abnormal mitral valve
morphology were not associated (respectively p =0.19, 0.08, 0.65, 0.72).

Table 4 shows the results of the bivariate analysis for risk factors. In Outcome 1, category of
mild definite RHD, morphologically abnormal mitral or aortic valves, pathological aortic
insufficiency, and penicillin prescription were more likely to progress. In Outcome 2, the
same risk factors emerged, except for the addition of mitral valve excessive anterior leaflet
motion as an additional risk factor (Table 4).

In a multivariable model, children who had pathologic aortic insufficiency were more likely
Author Manuscript

to progress (Outcome 1) than those without (OR 5.37, 95% CI 1.54–18.69). In outcome 2,
pathological aortic regurgitation remained an independent risk factor in addition to the mitral
valve morphological features restricted leaflet motion and excessive anterior leaflet motion
(OR 6.09, 95% CI 1.81–20.44, OR 5.29, 95% CI 1.22–22.92, OR 7.31, 95% CI 1.13–46.94,
respectively). Increasing age at diagnosis decreased the odds of progression in both models
(OR 0.83, 95% CI 0.70–0.99 and 0.84, 95% CI 0.71–0.99) for every year increased age for
Outcome 1 and 2 respectively) (Table 5).

Results of Cox proportional hazards regression for Outcome 1 were similar to those obtained
using logistic regression. Risk of progression over time was increased for children with
pathologic aortic insufficiency (HR = 5.69, 95% CI 2.01–16.14), with hazard ratios for other
risk factors similar to the previously reported ORs (data not shown).
Author Manuscript

After adjusting for differences in sex, age, presence of aortic regurgitation as well as
presence and degree of mitral regurgitation, penicillin prophylaxis in children with
borderline RHD was associated with increased odds of progression (outcome 2; OR=2.57,
95% CI 0.81–8.19). As indicated in Table 6, propensity score matched and other forms of
analyses designed to remove confounding did not alter the above result. All of these analyses
produced results consistent with more than a 2-fold increase in the odds of progression (9–
10% absolute risk difference) in penicillin users. However, none of the results achieved

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 8

better than borderline statistical significance (p=0.086 to p=0.098). In addition, sensitivity

Author Manuscript

analysis strongly indicated that accounting for any uncontrolled confounding would likely
lead to a further increase in the greater risk of progression associated with penicillin
prophylaxis in participants affected by borderline RHD.

DISCUSSION
A significant knowledge gap remains surrounding natural history of latent RHD found by
echo screening. Understanding the natural history of latent, echo-detected, RHD remains a
critical component of evaluating the utility of more widespread implementation of
echocardiographic screening programs. Here, we report the natural history of the largest
longitudinal cohort of children with latent RHD to date looking specifically at risk factors
for unfavorable outcomes. Four major findings emerge from our study.
Author Manuscript

First, not all screen-detected RHD should be considered latent RHD as some children were
found to have moderate-to-severe RHD on screening echo and they demonstrated poor
outcomes. In our cohort, only 2 such children (9.5%) showed improvement over the study
period, and there were 2 deaths (9.5%). Similar findings were recently reported from Fiji,
where more than 80% of children with moderate-to-severe RHD at time of screening showed
persistence or progression24. When screening reveals moderate to severe RHD, children
should be considered “missed clinical” cases, and treated as such, with referral to a specialist
provider, given consideration of BPG prescription, and valvular intervention as appropriate
and available.

Second, children with mild definite RHD at time of screening echo show more progression
than children with borderline RHD at screening. In our cohort, 25% of children with mild
definite RHD showed progression compared to only 10% of those with borderline RHD.
Author Manuscript

Progression is challenging to directly compare across cohorts, even in those using the 2012
WHF criteria, due to inconsistent definitions of progression, use of different binomial
outcomes (stable + progression), and inclusion of children with advanced RHD. However
the ratio of progression among those with mild definite RHD compared to those with
borderline in our cohort (2.5:1) is similar to reports from Fiji (2.5:1)24. The progression seen
among children with borderline RHD in our study (10%) is also similar to smaller recent
cohorts from New Caledonia (5%)28 and Fiji (12%)24, although lower than reported from
South Africa (20%)27 and Australia (24%)25.

However, category of RHD itself may not be as great a risk as specific echocardiographic
features. Our third major finding was the identification of risk factors for unfavorable
outcome based on multiple variable model. While initial RHD category (borderline vs. mild
Author Manuscript

definite RHD) was a highly significant risk factor for unfavorable outcome in our bivariate
model, it did not retain statistical significance in the multiple variable model. For outcome 1
(progression vs. stable + regression), only presence of pathological aortic insufficiency
emerged as a significant risk factor, while older age at diagnosis was protective from
progression. In outcome 2 (mild definite RHD progression + stable vs. regression or
borderline RHD progression vs. stable + regression), additional mitral morphological

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 9

features including restrictive leaflet motion and excessive anterior leaflet motion, were also
Author Manuscript

risk factors.

There are two aspects of progression that are of primary interest in terms of prevention of
sequelae, these are first evidence of (initial) progression and rate of progression over time.
Our Kaplan-Meier analysis provided important evidence that initial echocardiographic
progression may be detected during the first 2 years of clinical follow-up. There is also
evidence that RHD progression may not be uniform. The rate of RHD progression is
affected by the ongoing inflammatory response from the first insult and repeat insults in the
form of repeated GAS exposure and ARF recurrences, providing reasons to suspect non-
uniform progression. Data from clinical ARF and RHD are clear that progression of RHD
continues over time.34, 35 In our study, we are referencing only initial progression. Our data
suggests that initial evidence of progression may emerge early in follow-up, which has
implications for use of secondary prophylaxis and for retention in care. Longer natural
Author Manuscript

history studies are needed to determine if early evidence of progression is, in fact, predictive
of longer-term outcomes including need for future cardiac medications, interventions and/or
premature death and whether these key indicators of progression occur at differential rates
over time.

Length of follow-up did not emerge as a risk factor in our bivariate analysis. It may be that
we were underpowered to find differences, however the previous report had similarly small
numbers (19 children with ≥ 5 years follow-up24) and was underpowered to conduct a
multivariate analysis.. It is also important to remember that in our study, we are referencing
only initial progression. Data from clinical ARF and RHD is clear that progression of RHD
continues over time34, 35. However our data suggests that disease trajectory may emerge
early in follow-up, which has implications for use of secondary prophylaxis and for retention
Author Manuscript

in care. Longer natural history studies are needed to determine if early trajectory is, in fact,
predictive of longer-term outcomes including need for future cardiac medications,
interventions and/or premature death.

Fourth, the impact of penicillin on improving the trajectory of latent RHD remains unclear.
BPG prophylaxis emerged as a risk factor for progression, a finding also reported from
Australia25. This result failed to reach statistical significance but persisted unabated despite
extensive efforts, including propensity score matching with sensitivity analyses, to remove
extraneous differences between penicillin treated and untreated children. This
counterintuitive finding has important implications for prevention of sequelae in persons
with latent disease and thus needs to be further investigated in a randomized clinical trial.

As more data emerges on the natural history of latent RHD, it is important to consider
Author Manuscript

standardized reporting of outcomes36. The 2012 WHF criteria represent the best available
evidence for the echocardiographic diagnosis and classification of latent RHD, and have
been widely adopted facilitating reproducible diagnostic categories9. Defining latent RHD
trajectory has been less uniform both in what constitutes change and in how trinomial
outcomes are grouped into binomial classifications. Here, as reported by several recent
studies24, 25, we use an expanded definition of change including but not limited to change in
WHF category. Expert agreement on these definitions would facilitate better comparison

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 10

across populations, but until this agreement, clear reporting of definitions and provision of
granular details on reasons for change is essential36. Grouping of disease trajectory into
Author Manuscript

clinically meaningful outcomes is also important. In this study we chose to report outcomes
in 2 ways, first, with only progression vs. stable and regression and second, in the way we
found most clinically relevant, grouping progression and stable definite RHD as the worse
outcome, while only considering progression of borderline RHD as the worse of two
outcomes. Interestingly, these groupings did not affect our multivariate analysis greatly, with
only additional morphological features emerging in Outcome 2.

When possible, it is also important to include multiple variable analysis to determine which
risk factors remain important. In our cohort, many seemingly key variables in bivariate
analysis, including disease category and prescription of BPG prophylaxis, were not
independent risk factors in our multivariate analysis. In part, this reflects the number of
correlated confounders likely to surface in natural history studies. As many longitudinal
Author Manuscript

RHD cohorts remain small, combining data for a larger multicenter analysis remains a top
priority.

Limitations
The main limitation in our study was the inability to determine the impact of BPG
prophylaxis on the trajectory of latent RHD. This limitation is inherent to all natural history
cohorts, reinforcing the need for a well-designed randomized controlled trial in this
population. Based on our data, we would advocate the exclusion of children with moderate-
to-severe definite RHD at time of screening but believe that equipoise exists to justify
inclusion of both mild definite and borderline RHD in such a trial. Similarly, we did not
collect data on streptococcal exposure in this cohort and cannot determine contribution of
repeated GAS exposure to latent disease trajectory. A second limitation is the relatively
Author Manuscript

small number of mild definite RHD cases included in our cohort (n=42). While
comparatively large relative to previous studies8, 22–28, this is still a small number of patients
and may be underpowered to detect important risk factors for unfavorable outcome. Thirdly,
there is selection bias of our included cases: school-going children at time of screening and
only those who were enrolled in the national RHD registry. There are no data to suggest
children who do not attend school are at greater risk of latent RHD or of progression,
however this is possible considering the lower overall socioeconomic status in this group, a
known driver of RHD. Fourth, significant changes in systemic blood pressure can change the
appearance and grade of mitral and aortic regurgitation. We did not capture blood pressure at
each echo, however of 223 patients with data on age, height, and BP at registry enrollment,
221 were normotensive, suggesting this is likely not a driver of change in disease
categorization in our population. Additionally, not all children with screen-detected RHD
Author Manuscript

returned for initial follow-up and/or were enrolled in the national registry. We cannot
determine if or how these losses affected our results. Finally, while initial echos were re-
reviewed blindly for disease classification, we intentionally chose to review serial echos in
an un-blinded fashion (able to look forward and back within a case). This study is among the
first to review and report on serial echos during the follow-up period (as compared to only
first and last), which has provided important information on timing of initial change. We
also believe that side-by-side comparison strengthens the ability of a reviewer to see

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 11

meaningful differences over time, as is standard clinical practice. Additionally, we invited an

Author Manuscript

expert cardiologist outside our group to review a randomly selected batch of echos, and
found substantial agreement in category of diagnosis (normal/other, borderline, definite),
indicating no significant bias was introduced through this method.

Conclusions
Latent RHD is a heterogeneous diagnosis with variable disease outcomes. Children with
moderate-to-severe latent RHD found by echo screening should be considered missed
clinical cases and treated as such, as their outcomes are poor. Children with both borderline
and mild definite RHD are at substantial risk of progression, and at a minimum should be
enrolled in close clinical follow-up. While long-term outcomes remain unclear, the
worsening of latent RHD may be evident through echo assessment during the first 1–2 years
following diagnosis. Natural history data is inherently limited by confounding variables and
Author Manuscript

a randomized controlled clinical trial is needed to definitively determine the impact of BPG
prophylaxis on the trajectory of latent RHD.

Acknowledgments
We would like to thank the Ugandan Heart Institute, the children and families who take part in the Ugandan RHD
Registry, and the nurses, data coordinators, and volunteers who support this work.

SOURCES OF FUNDING: This work was supported by the National Institutes of Health National Center for
Advancing Translational Sciences [grant numbers UL1TR000075 and KL2TR000076 to AB]. Its contents are the
responsibility of the authors and do not necessarily represent the views of the National Center for Advancing
Translational Sciences or the National Institutes of Health. Further support came from the Karp Family Foundation,
the Medtronic Global Health Foundation and the National Institutes of Health (K23 HL123341 to CTL), Gift of
Life International, and General Electric.
Author Manuscript

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.
[PubMed: 28834488]
2. Sims Sanyahumbi, A., Colquhoun, S., Wyber, R., Carapetis, JR. Global Disease Burden of Group A
Streptococcus. In: Ferretti, JJ.Stevens, DL., Fischetti, VA., editors. Streptococcus pyogenes : Basic
Biology to Clinical Manifestations. Oklahoma City (OK): 2016.
3. Zuhlke 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,
Author Manuscript

Elhassan HH, Mocumbi AO, Adeoye AM, Mntla P, Ojji D, Mucumbitsi J, Teo K, Yusuf S, Mayosi
BM. Characteristics, complications, and gaps in evidence-based interventions in rheumatic heart
disease: the Global Rheumatic Heart Disease Registry (the REMEDY study). Eur Heart J. 2015;
36:1115–1122a. [PubMed: 25425448]
4. Zhang W, Mondo C, Okello E, Musoke C, Kakande B, Nyakoojo W, Kayima J, Freers J. Presenting
features of newly diagnosed rheumatic heart disease patients in Mulago Hospital: a pilot study.
Cardiovasc J Afr. 2013; 24:28–33. [PubMed: 23612950]

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 12

5. Okello E, Wanzhu Z, Musoke C, Twalib A, Kakande B, Lwabi P, Wilson NB, Mondo CK, Odoi-
Adome R, Freers J. Cardiovascular complications in newly diagnosed rheumatic heart disease
Author Manuscript

patients at Mulago Hospital, Uganda. Cardiovasc J Afr. 2013; 24:80–85. [PubMed: 23736132]
6. 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. [PubMed: 17671255]
7. Carapetis JR, Hardy M, Fakakovikaetau T, Taib R, Wilkinson L, Penny DJ, Steer AC. Evaluation of
a screening protocol using auscultation and portable echocardiography to detect asymptomatic
rheumatic heart disease in Tongan schoolchildren. Nat Clin Pract Cardiovasc Med. 2008; 5:411–
417. [PubMed: 18398402]
8. Paar JA, Berrios NM, Rose JD, Caceres M, Pena R, Perez 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. [PubMed: 20538135]
9. Remenyi 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, Zuhlke L, Carapetis J. World Heart Federation criteria for echocardiographic diagnosis
Author Manuscript

of rheumatic heart disease--an evidence-based guideline. Nat Rev Cardiol. 2012; 9:297–309.
[PubMed: 22371105]
10. Beaton A, Okello E, Lwabi P, Mondo C, McCarter R, Sable C. Echocardiography screening for
rheumatic heart disease in Ugandan schoolchildren. Circulation. 2012; 125:3127–3132. [PubMed:
22626741]
11. 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 echocardiography for early rheumatic heart disease diagnosis: a field
study. Eur Heart J Cardiovasc Imaging. 2015; 16:475–482. [PubMed: 25564396]
12. Sims Sanyahumbi A, Sable CA, Beaton A, Chimalizeni Y, Guffey D, Hosseinipour 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. [PubMed: 27029239]
13. Engel ME, Haileamlak A, Zuhlke 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
Author Manuscript

Africa and Ethiopia. Heart. 2015; 101:1389–1394. [PubMed: 26076935]

14. Shrestha NR, Kalesan B, Karki P, Sherpa K, Basnet A, Urban P, Pilgrim T. Rheumatic heart
disease: pilot study for a population-based evaluation of prevalence and cardiovascular outcomes
among schoolchildren in Nepal. BMJ Open. 2012; 2doi: 10.1136/bmjopen-2012-001616
15. Bhaya M, Panwar S, Beniwal R, Panwar RB. High prevalence of rheumatic heart disease detected
by echocardiography in school children. Echocardiography. 2010; 27:448–453. [PubMed:
20345448]
16. 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. and investigators P. Echocardiographic prevalence of rheumatic heart disease
in Brazilian schoolchildren: Data from the PROVAR study. Int J Cardiol. 2016; 219:439–445.
[PubMed: 27372607]
17. Mirabel M, Fauchier T, Bacquelin R, Tafflet M, Germain A, Robillard C, Rouchon B, Marijon E,
Jouven X. Echocardiography screening to detect rheumatic heart disease: A cohort study of
schoolchildren in French Pacific Islands. Int J Cardiol. 2015; 188:89–95. [PubMed: 25889336]
Author Manuscript

18. Roberts K, Maguire G, Brown A, Atkinson D, Remenyi B, Wheaton G, Kelly A, Kumar RK, Su
JY, Carapetis JR. Echocardiographic screening for rheumatic heart disease in high and low risk
Australian children. Circulation. 2014; 129:1953–1961. [PubMed: 24622384]
19. Remond MG, Wark EK, Maguire GP. Screening for rheumatic heart disease in Aboriginal and
Torres Strait Islander children. J Paediatr Child Health. 2013; 49:526–531. [PubMed: 23638751]
20. Colquhoun SM, Carapetis JR, Kado JH, Reeves BM, Remenyi B, May W, Wilson NJ, Steer AC.
Pilot study of nurse-led rheumatic heart disease echocardiography screening in Fiji--a novel
approach in a resource-poor setting. Cardiol Young. 2013; 23:546–552. [PubMed: 23040535]

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 13

21. Wilson, JMG., Jungner, G. Public Health Paper Number 34. Geneva: WHO; 1968. Principles and
practice of screening for disease. (http://www.who.int/iris/handle/10665/37650)
Author Manuscript

22. Beaton A, Okello E, Aliku T, Lubega S, Lwabi P, Mondo C, McCarter R, Sable C. Latent
rheumatic heart disease: outcomes 2 years after echocardiographic detection. Pediatr Cardiol.
2014; 35:1259–1267. [PubMed: 24827080]
23. Bhaya M, Beniwal R, Panwar S, Panwar RB. Two years of follow-up validates the
echocardiographic criteria for the diagnosis and screening of rheumatic heart disease in
asymptomatic populations. Echocardiography. 2011; 28:929–933. [PubMed: 21854437]
24. 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. [PubMed: 27933106]
25. Remond M, Atkinson D, White A, Brown A, Carapetis J, Remenyi B, Roberts 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. [PubMed: 26163902]
26. 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
Author Manuscript

RHEUMATIC (Rheumatic Heart Echo Utilisation and Monitoring Actuarial Trends in Indian
Children) study. Heart. 2011; 97:2018–2022. [PubMed: 22076022]
27. Zuhlke L, Engel ME, Lemmer CE, van de Wall M, Nkepu S, Meiring A, Bestawros M, Mayosi
BM. The natural history of latent rheumatic heart disease in a 5 year follow-up study: a
prospective observational study. BMC Cardiovasc Disord. 2016; 16:46. [PubMed: 26892774]
28. Bertaina G, Rouchon B, Huon B, Guillot N, Robillard C, Noel B, Nadra M, Tribouilloy C, Marijon
E, Jouven X, Mirabel M. Outcomes of borderline rheumatic heart disease: A prospective cohort
study. Int J Cardiol. 2017; 228:661–665. [PubMed: 27883978]
29. Beaton A, Aliku T, Okello E, Lubega S, McCarter R, Lwabi P, Sable C. The utility of handheld
echocardiography for early diagnosis of rheumatic heart disease. J Am Soc Echocardiogr. 2014;
27:42–49. [PubMed: 24183541]
30. Ploutz M, Lu JC, Scheel J, Webb C, Ensing GJ, Aliku T, Lwabi P, Sable C, Beaton A. Handheld
echocardiographic screening for rheumatic heart disease by non-experts. Heart. 2016; 102:35–39.
[PubMed: 26438784]
31. Zoghbi WA, Adams D, Bonow RO, Enriquez-Sarano M, Foster E, Grayburn PA, Hahn RT, Han Y,
Author Manuscript

Hung J, Lang RM, Little SH, Shah DJ, Shernan S, Thavendiranathan P, Thomas JD, Weissman NJ.
Recommendations for Noninvasive Evaluation of Native Valvular Regurgitation: A Report from
the American Society of Echocardiography Developed in Collaboration with the Society for
Cardiovascular Magnetic Resonance. J Am Soc Echocardiogr. 2017; 30:303–371. [PubMed:
28314623]
32. Nishimura RA, Otto CM, Bonow RO, Carabello BA, Erwin JP 3rd, Fleisher LA, Jneid H, Mack
MJ, McLeod CJ, O'Gara PT, Rigolin VH, Sundt TM 3rd, Thompson A. 2017 AHA/ACC Focused
Update of the 2014 AHA/ACC Guideline for the Management of Patients With Valvular Heart
Disease: A Report of the American College of Cardiology/American Heart Association Task Force
on Clinical Practice Guidelines. J Am Coll Cardiol. 2017; 70:252–289. DOI: 10.1161/CIR.
0000000000000503 [PubMed: 28315732]
33. Landis JR, Koch GG. The measurement of observer agreement for categorical data. Biometrics.
1977; 33:159–174. [PubMed: 843571]
34. Carapetis JR, Beaton A, Cunningham MW, Guilherme L, Karthikeyan G, Mayosi BM, Sable C,
Author Manuscript

Steer A, Wilson N, Wyber R, Zuhlke L. Acute rheumatic fever and rheumatic heart disease. Nat
Rev Dis Primers. 2016; 2:15084. [PubMed: 27188830]
35. Tompkins DG, Boxerbaum B, Liebman J. Long-term prognosis of rheumatic fever patients
receiving regular intramuscular benzathine penicillin. Circulation. 1972; 45:543–551. [PubMed:
5012243]
36. Karthikeyan G. Measuring and reporting disease progression in subclinical rheumatic heart disease.
Heart Asia. 2016; 8:74–75.

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 14

CLINICAL PERSPECTIVE
Author Manuscript

• What is New?

◦ Rheumatic heart disease (RHD) remains the most common

cardiovascular disease among the world’s youth.

◦ Echocardiographic screening provides a promising tool for early

detection but the natural history of screen-detected RHD (latent
RHD) remains unclear.

◦ In this large series of patients, we show that children with borderline

and definite RHD by echo screening are at substantial risk of
progression, 9% and 26% respectively at a median follow-up of 2.6
years.
Author Manuscript

◦ Further, we identify younger age at diagnosis, and presence of

morphological mitral valve features at diagnosis as independent risk
factors for unfavorable outcomes.

• What are the Clinical Implications?

◦ This study adds to the growing knowledge of latent RHD and the
role of echocardiographic screening as a tool to reduce the global
burden of RHD.

◦ Children with screen-detected RHD remain at substantial clinical

risk.

◦ Children with moderate to severe RHD at screening should be

Author Manuscript

considered for treatment as clinically diagnosed RHD.

◦ Children with borderline or mild definite RHD at screening should,

at a minimum, be maintained in close clinical follow-up.

◦ Additionally, our data highlight the need for a randomized clinical

trial in this population to determine the effect of penicillin
prophylaxis on the clinical course of RHD.
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 15
Author Manuscript
Author Manuscript
Author Manuscript

Figure 1.
Map of Uganda showing the two districts of enrollment.
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 16
Author Manuscript
Author Manuscript
Author Manuscript

Figure 2.
Flowchart of study entry and classification according to CONSORT recommendations.
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 17
Author Manuscript
Author Manuscript

Figure 3.
RHD Trajectory by Initial Diagnostic Classification (Favorable and Unfavorable outcomes
are shown above the bars in solid (favorable) and dashed (unfavorable) bars for Outcome 1
(O1) and Outcome 2 (O2).)
Author Manuscript
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 18
Author Manuscript
Author Manuscript
Author Manuscript

Figure 4.
Progression free Survival Stratified by Disease Status
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 19

Table 1

Binomial Classifications by Outcome.

Author Manuscript

Unfavorable Favorable

Outcome 1 Progression Stable or Regression

Outcome 2 Mild Definite: Stable or Progression Mild Definite: Regression

Borderline: Progression Borderline: Stable or Regression
Author Manuscript
Author Manuscript
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 20

TABLE 2

Characteristics of 227 Children with Latent RHD

Author Manuscript

Demographics

Median Age at Screening (IQR), years 12 (10–14)

Gender – n (%) Female 137 (60.4)

Enrollment Site – n (%) UHI 47 (20.7)

Gulu 180 (79.3)

Median Duration of Follow-up (IQR), years 2.3 (2.0–2.9)

Mean Echocardiograms/Patient (SD) 3.8 (+/ 1.5)

Echocardiographic Characteristics

Definite RHD – n (%) 63 (27.8)

WHF Subcategory of Definite RHD – n (%)

A: Pathological MR + morphological features of the MV 47 (74.6)

Author Manuscript

B: Mitral stenosis 3 (4.8)

C: Pathological AR + morphological features of the AV 6 (9.5)

D: Borderline disease of the MV and AV 7 (11.1)

Severity of Definite RHD – n (%) Mild 42 (66.7)

Moderate 15 (23.8)

Severe 6 (9.5)

Borderline RHD – n (%) 164 (72.2)

WHF Subcategory of Borderline RHD – n (%)

A: Morphological features of the mitral valve 9 (5.5)

B: Pathological MR 135 (82.3)

C: Pathological AR 20 (12.2)
Author Manuscript

Severity of Borderline RHD– n (%) Mild 163 (99.4)

Moderate 1 (0.6)

Severe 0

Secondary Penicillin Prophylaxis

Prescribed BPG – n (%) 108 (47.5)

Data Type – n (%) Continuous (days of coverage) 83 (76.9)

Categorical (# injections received) 25 (23.1)

Mean Adherence (SD) 84.7 (+/− 0.15)

Adherence ≥ 80% – n (%) 89 (82.4)

Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 21

Table 3

Breakdown of Echocardiographic Changes in Cases of Progression and Regression (mild definite and
Author Manuscript

borderline RHD)

Reasons for Progression (n=27 children*)

New pathologic or worsening grade of mitral regurgitation 13 (48.1%)

New pathologic or worsening grade of aortic regurgitation 6 (22.2%)

Newly met criteria for abnormal MV morphology† 15 (55.6%)

Newly met criteria for abnormal AV morphology† 2 (7.4%)

Reasons for Regression (n=95 children*)

Improvement in grade or disappearance of pathologic mitral regurgitation 72 (75.0%)

Improvement in grade or disappearance of pathologic aortic regurgitation 16 (16.7%)

Resolution of Improvement in MV morphological findings‡ 7 (7.3%)

Author Manuscript

Resolution of AV morphological findings‡ 5 (5.2%)

*
Some children showed more than one reason for progression or regression;
†
At least 2 morphological features at the mitral valve including thickened anterior mitral leaflet (≥3mm), restricted leaflet motion, chordal
thickening, excessive anterior mitral leaflet motion OR at least 2 morphological features at the aortic valve including thickening, prolapse, non-
coaptation, or restricted motion;
‡
Less than 2 morphological features of the mitral valve or aortic valve.
Author Manuscript
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 22

Table 4

Clinical risk factors for unfavorable outcome, bivariate logistic regression analysis
Author Manuscript

Outcome 1* Outcome 2*

Variable Odds Ratio (95% CI) P-value† Odds Ratio (95% P-value†
CI)
Disease Category (mild definite) 3.28 (1.39–7.76) 0.007 10.17 (4.59–22.54) <0.001
Length of Follow-up (per year) 1.00 (1.00–1.00) 0.648 1.00 (1.00–1.00) 0.353

Age at Diagnosis (per year) 0.88 (0.75–1.03) 0.099 0.88 (0.77–1.01) 0.073

Gender (female) 1.04 (0.45–2.41) 0.925 0.80 (0.39–1.65) 0.551

Mitral Valve Morphological Features

Anterior leaflet ≥ 3mm 2.84 (1.23–6.57) 0.015 3.09 (1.43–6.42) 0.004

Chordal thickening 4.27 (1.16–15.73) 0.029 4.09 (1.18–14.20) 0.027
Restrictive leaflet motion 3.98 (1.58–10.03) 0.003 5.95 (2.55–13.87) <0.001
Author Manuscript

Excessive anterior leaflet motion 1.56 (0.48–5.00) 0.459 2.93 (1/13–7.61) 0.027
Aortic Valve Morphological Features

Any WHF morphological abnormality 2.86 (1.12–7.27) 0.028 3.66 (1.65–8.13) 0.001
Functional Valve Abnormalities

Length of MR (per 1 cm) 0.72 (0.45–1.15) 0.165 1.05 (0.68–1.62) 0.815

Length of MR ‡ 0.502 ------------- ‡

Pathological AR 3.67 (1.54–8.76) 0.003 3.66 (1.65–8.13) 0.001
Benzathine Penicillin G

Penicillin prescription 2.90 (1.24–6.82) 0.014 3.99 (1.85–8.59) <0.001

Adherence (≥ 80%) 0.94 (0.53–1.67) 0.844 0.73 (0.25–2.09) 0.556

CI: Confidence Interval, WHF: World Heart Federation, MR: Mitral regurgitation, AR: Aortic regurgitation.
Author Manuscript

*
Outcome 1: Progression vs. Stable + Regression (ALL); Outcome 2: Mild definite RHD (Progression + Stable vs. Regression) and Borderline
RHD (Progression vs. Stable + Regression);
†
Type 3 p-value;
‡
Odds ratios for this categorical based on <2cm compared to 2cm, 3cm, and 4cm – none were significant.
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 23

Table 5

Clinical risk factors for RHD progression, multivariable logistic regression analysis
Author Manuscript

Outcome 1* Outcome 2*

Variable Odds Ratio (95% P-value† Odds Ratio (95% P-value†

CI) CI)
Disease Category (borderline/mild definite) 0.58 (0.14–2.37) 0.125 2.24 (0.63–7.97) 0.212

Age at Diagnosis (continuous) 0.83 (0.70–0.99) 0.035 0.84 (0.71–0.99) 0.039

Mitral Valve Morphological Features

Anterior leaflet ≥ 3mm 3.03 (0.29–31.84) 0.355 10.57 (0.69–161.33) 0.090

Chordal thickening 2.55 (0.51–12.86) 0.255 1.56 (0.33–7.43) 0.577

Restrictive leaflet motion 3.39 (0.89–12.83) 0.072 5.29 (1.22–22.92) 0.026

Excessive anterior leaflet motion ‡ ‡ 7.31 (1.13–46.94) 0.036
Aortic Valve Morphological Features
Author Manuscript

Any WHF morphological abnormality 2.27 (0.69–7.48) 0.178 1.28 (0.39–4.20) 0.682

Functional Valve Abnormalities

Pathological AR present 5.37 (1.54–18.69) 0.008 6.09 (1.81–20.44) 0.004

Benzathine Penicillin G

Penicillin prescription (yes/no) 2.37 (0.79–7.12) 0.125 1.99 (0.69–5.76) 0.203

CI: Confidence Interval, WHF: World Heart Federation, AR: Aortic regurgitation.
*
Outcome 1: Progression vs. Stable + Regression (ALL); Outcome 2: Mild definite RHD (Progression + Stable vs. Regression) and Borderline
RHD (Progression vs. Stable + Regression);
†
Type 3 p-value.
‡
Not included in this model.
Author Manuscript
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.

 Beaton et al. Page 24

Table 6

Results of Analyses to Remove Extraneous Differences between Penicillin-Treated and Untreated

Author Manuscript

Average %Risk of Progression

Difference
Method to Remove Extraneous Treatment Treated -
Group Differences Untreated Untreated p-value

Regression Modeling 5.2 8.8 0.097

Inverse Prob Weighting 5.2 8.8 0.093
Nearest Neighbor Matching 8.9 0.098
Propensity Score Matching 9.9 0.086
Author Manuscript
Author Manuscript
Author Manuscript

Circulation. Author manuscript; available in PMC 2018 December 05.