The n e w e ng l a n d j o u r na l of m e dic i n e

original article

Prevalence of Rheumatic Heart Disease

Detected by Echocardiographic Screening
Eloi Marijon, M.D., Phalla Ou, M.D., David S. Celermajer, Ph.D., F.R.A.C.P.,
Beatriz Ferreira, M.D., Ph.D., Ana Olga Mocumbi, M.D., Dinesh Jani, M.D.,
Christophe Paquet, M.D., M.P.H., Sophie Jacob, Ph.D., Daniel Sidi, M.D., Ph.D.,
and Xavier Jouven, M.D., Ph.D.

A bs t r ac t

Background
From the Departments of Pediatric Car­ Epidemiologic studies of the prevalence of rheumatic heart disease have used clinical
diology (E.M., D.S.) and Pediatric Ra­ screening with echocardiographic confirmation of suspected cases. We hypothesized
diology (P.O.), Hôpital Necker–Enfants
Malades, Paris; the University Paris Des­ that echocardiographic screening of all surveyed children would show a significantly
cartes, Georges Pompidou European Hos­ higher prevalence of rheumatic heart disease.
pital, Paris (E.M., X.J.); the Maputo Heart
Institute, Maputo, Mozambique (E.M.,
B.F., A.O.M., D.J.); the Phnom Penh Heart
Methods
Center, Phnom Penh, Cambodia (P.O.); Randomly selected schoolchildren from 6 through 17 years of age in Cambodia and
the Department of Medicine, Sydney Uni­ Mozambique were screened for rheumatic heart disease according to standard clin-
versity, Sydney (D.S.C.); the National Insti­
tute for Public Health Surveillance, Saint ical and echocardiographic criteria.
Maurice, France (C.P.); and INSERM, Unité
780 Avenir, Cardiovascular Epidemiology, Results
Villejuif, France (S.J., X.J.). Address reprint
requests to Dr. Marijon at the Hôpital Clinical examination detected rheumatic heart disease that was confirmed by echo-
­Europeén Georges Pompidou, Service de cardiography in 8 of 3677 children in Cambodia and 5 of 2170 children in Mozam-
Cardiologie 2, 20 rue Leblanc, 75908 Paris bique; the corresponding prevalence rates and 95% confidence intervals (CIs) were
CEDEX 15, France, or at eloi_marijon@
yahoo.fr. 2.2 cases per 1000 (95% CI, 0.7 to 3.7) for Cambodia and 2.3 cases per 1000 (95% CI,
0.3 to 4.3) for Mozambique. In contrast, echocardiographic screening detected 79
N Engl J Med 2007;357:470-6. cases of rheumatic heart disease in Cambodia and 66 cases in Mozambique, corre-
Copyright © 2007 Massachusetts Medical Society.
sponding to prevalence rates of 21.5 cases per 1000 (95% CI, 16.8 to 26.2) and 30.4
cases per 1000 (95% CI, 23.2 to 37.6), respectively. The mitral valve was involved in
the great majority of cases (87.3% in Cambodia and 98.4% in Mozambique).

Conclusions
Systematic screening with echocardiography, as compared with clinical screening, re-
veals a much higher prevalence of rheumatic heart disease (approximately 10 times
as great). Since rheumatic heart disease frequently has devastating clinical consequenc-
es and secondary prevention may be effective after accurate identification of early
cases, these results have important public health implications.

470 n engl j med 357;5  www.nejm.org  august 2, 2007

The New England Journal of Medicine

Downloaded from nejm.org on April 10, 2018. For personal use only. No other uses without permission.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
 Prevalence of Rheumatic Heart Disease in Asia and Africa

I
n poor and developing nations, rheu- tive of two regions of the developing world where
matic heart disease remains a major cause of previous surveys have documented an apparently
morbidity and premature death and imposes high prevalence of rheumatic heart disease in
a substantial burden on health care systems with school-age children2 and because local investiga-
limited budgets.1,2 Nevertheless, primary and sec-tors and authorities were willing and able to par-
ondary prevention efforts may be highly effective.1,3
ticipate in the studies. The two studies were carried
Secondary prevention relies on accurate case detec-
out after prospective planning by a central group
tion for the appropriate use of prophylactic antibi-
of investigators, which included establishing pro-
otics and regular medical surveillance. Exact prev-
tocols for prospectively defined clinical and echo-
alence data are also highly desirable to facilitate
cardiographic criteria for the diagnosis of rheumat-
health care planning. ic valvular abnormalities.
Almost all population-based epidemiologic sur- The first study was carried out in Cambodia
veys have relied on careful clinical examination of
in 2001 and 2002 after approval by the Ministry
school-age children, with confirmation of clini- of Health. The investigators approached the direc-
cally suspected cases by echocardiography. Such tors of randomly selected schools in the capital of
surveys show current prevalence rates of rheu- Phnom Penh, and all agreed to participate. All
matic heart disease of approximately 1 to 5 cases children in each school were invited to participate;
per 1000 among school-age children in developing those children whose parents or guardians gave
countries, with the highest rates in sub-Saharan oral informed consent were brought to the Phnom
Africa.2 Penh Heart Center, where they underwent com-
Cardiac ultrasonography is known to be more prehensive clinical and echocardiographic exami-
sensitive than auscultation for the detection of nation (with a Philips Sonos 4500 4–7 MHz trans-
pathologic valve disease,4 and the recent availabil-
ducer) for signs of rheumatic heart disease. Images
ity of high-quality portable ultrasound equipment were recorded on super-VHS videotape for later
makes it possible to screen large numbers of chil-review by an independent physician who was ex-
dren at schools in developing nations. Further- perienced in diagnosing rheumatic heart disease
more, the 2004 World Health Organization Expert and was not aware of the clinical findings. The
Consultation Report states that echocardiograph- results of this study appeared to confirm our hy-
ically diagnosed, clinically silent rheumatic valve
pothesis that a much higher prevalence of rheu-
involvement should be managed as rheumatic matic heart disease would be found with echocar-
heart disease until proved otherwise.3 We there- diographic screening. We then decided to repeat
fore hypothesized that comprehensive screening, the study with more rigorous randomization pro-
including echocardiography in all children, might cedures for subject selection and with confirma-
reveal a higher prevalence of cases than clinical tion of echocardiographic findings by multiple in­
examination with echocardiographic confirmation dependent observers.
of clinically suspected cases only. We conducted a survey in Mozambique from
We performed two large, population-based May through October 2005 after approval of the
studies of school-age children, one in Southeast survey by the ethics committee of the Ministry of
Asia (Cambodia) and one in sub-Saharan Africa Health and the Education Ministry. A total of 2170
(Mozambique), to assess the feasibility of echocar-
pupils from six public primary schools were ran-
diographic screening and to ascertain whether thisdomly selected from among the 1,140,000 chil-
method would yield a more realistic estimate of dren 6 through 17 years of age living in the capi-
the prevalence of rheumatic heart disease than that
tal city of Maputo. The overall sample was equally
obtained by screening with the use of predomi- distributed among seven class levels, with 310
nantly clinical criteria. subjects per class level, in order to ensure an even
age distribution. Two thirds of the children (1440)
Me thods were from suburban schools and one third (730)
were from urban schools. For each class level,
Setting and survey methods classrooms were selected by means of a systematic
Our study was conducted in Cambodia and Mo- randomization procedure with Epi Info software
zambique, because these countries are representa- (version 3.3.2). In each selected classroom, all chil-

n engl j med 357;5  www.nejm.org  august 2, 2007 471

The New England Journal of Medicine
Downloaded from nejm.org on April 10, 2018. For personal use only. No other uses without permission.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

dren for whom written informed consent could be least two of the following three morphologic ab-
obtained from a parent or guardian (91.6% of chil- normalities of the regurgitant valve: restricted
dren) were included in the study. leaflet mobility, focal or generalized valvular thick-
For each child, any history of acute rheumatic ening, and abnormal subvalvular thickening. For
fever was noted. The participating children then a definite diagnosis of rheumatic heart disease,
underwent a detailed clinical and echocardio- these features had to be identified concordantly by
graphic examination (with a SonoSite 4.2 MHz each of the echocardiographers, all of whom were
transducer) at the school. Absentees were noted experienced in the diagnosis and treatment of
and revisits were made to examine all of them. rheumatic heart disease.
All children in whom cardiac abnormalities were
noted with the portable ultrasound system under- Statistical Analysis
went repeated scanning at the Maputo Heart Insti- Each study enrolled more than 2000 children and
tute (with a Philips Sonos 4500 4–7 MHz trans- had more than 90% power at the 0.01 significance
ducer). These images were recorded on super-VHS level to test the hypothesis that echocardiograph-
videotape for later analysis by three independent ic screening would reveal twice as many cases of
physicians experienced in diagnosing rheumatic rheumatic heart disease as would be expected on
heart disease. the basis of published prevalence data from clin-
In both studies, children with a diagnosis of ical screening studies (5 cases per 1000 children).2
acute rheumatic fever according to the revised Our sample-size calculation was based on the re-
Jones criteria were treated with salicylates and an- sults of previous studies of the clinical prevalence
tibiotic agents.5 Each child in whom rheumatic of rheumatic heart disease; since the prevalence ob-
valve lesions were detected by echocardiography served in our studies was higher than expected,
was followed up with medical surveillance every our sample sizes clearly were overestimated.
6 months, including clinical and echocardiograph- All data were analyzed at INSERM, Unit 780
ic review, initiation of antibiotic prophylaxis in Avenir, Cardiovascular Epidemiology, Villejuif,
cases of lesions considered to be significant ac- France, with SAS software (version 8.2). Descrip-
cording to World Health Organization criteria, and tive data are presented as means ±SD; sample
surgical treatment where clinically indicated. means were compared with an independent-sam-
ples Student’s t-test. A two-sided P value of less
Clinical and echocardiographic definitions than 0.05 was considered to indicate statistical
Clinical examination was performed by physicians significance. We calculated the fraction of children
experienced in the diagnosis of rheumatic heart whose findings were normal by clinical criteria
disease. Careful cardiac auscultation was per- and abnormal by echocardiographic criteria. In
formed with the patient in the supine and left lat- addition, the rate of detection by echocardiograph-
eral decubitus positions. Children in whom an or- ic screening was calculated by means of the total
ganic murmur was detected clinically and the probability formula as follows: rate of detection  =
presence of rheumatic heart disease was confirmed p1 + (1 − p1) × p2, where p1 is the probability of
echocardiographically were classified as having an abnormal result by clinical screening and p2 is
clinically detected rheumatic heart disease. the probability of an abnormal result by echocar-
The echocardiographic criteria were agreed on diographic screening after a normal result by
by all observers in both the Cambodian and the clinical screening.
Mozambique studies before the ultrasound scans
were read by echocardiographers who were un- R e sult s
aware of the clinical findings. Only left-sided
valves were examined for features of rheumatic Cambodia
heart disease; mild tricuspid regurgitation and Of the 3677 children examined in Cambodia, 52.0%
pulmonary regurgitation were frequently noted but were male, and the mean age was 11.7±2.5 years
were not regarded as indicating rheumatic heart (Table 1). Clinical evidence of rheumatic heart dis-
disease. Rheumatic heart disease was defined by ease confirmed by echocardiography was found in
the presence of any definite evidence of mitral- or 8 children, corresponding to a prevalence of 2.2
aortic-valve regurgitation seen in two planes by cases per 1000 children (95% confidence interval
Doppler echocardiography, accompanied by at [CI], 0.7 to 3.7). Five of these eight children ful-

472 n engl j med 357;5  www.nejm.org  august 2, 2007

The New England Journal of Medicine

Downloaded from nejm.org on April 10, 2018. For personal use only. No other uses without permission.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
 Prevalence of Rheumatic Heart Disease in Asia and Africa

filled the revised Jones criteria for acute rheumatic

Table 1. Characteristics of the Patients.*
fever and were treated.
Echocardiographic screening showed that 79 Children in Children in
children had Doppler evidence of left-heart valve Cambodia Mozambique
Characteristic (N = 3677) (N = 2170)
regurgitation and morphologic features of rheu-
Age — yr
matic heart disease, corresponding to a prevalence
of 21.5 cases per 1000 children (95% CI, 16.8 to Mean 11.7±2.5 10.6±2.5
26.2) (Fig. 1). Thus, 71 children were found by Range 6–17 6–17
echocardiography to have clinically silent rheu- Median 12 11
matic heart disease. Of the 79 children with a di- Male sex — no. (%) 1912 (52.0) 1031 (47.5)
agnosis of rheumatic heart disease, 61 (77.2%) had Cases of rheumatic heart disease — no.
only mitral-valve disease, 8 had both aortic- and
Clinical detection 8 5
mitral-valve disease, and 10 had only aortic-valve
disease. No cases of valvular stenosis were de- Echocardiographic detection 79 66
tected. Echocardiographic but not clinical 71 61
detection

Mozambique Prevalence of rheumatic heart disease

Of the 2170 children examined in Mozambique, Clinical detection — no. of cases/ 2.2 (0.7–3.7) 2.3 (0.3–4.3)
1000 children (95% CI)
47.5% were male, and the mean age was 10.6±2.5
years. Three children had a history of acute rheu- Echocardiographic detection — 21.5 (16.8–26.2) 30.4 (23.2–37.6)
no. of cases/1000 children
matic fever but had no cardiac murmur on exami- (95% CI)
nation, one child had known rheumatic heart dis-
Echocardiographic but not clinical 19.3 28.1
ease, and one had known congenital heart disease. detection — no. of cases/1000
Clinical evidence of rheumatic heart disease con- children
firmed by echocardiography was found in 5 chil- Ratio of echocardiographic to clinical 9.8 13.2
dren, corresponding to a prevalence of 2.3 cases per detection
1000 (95% CI, 0.3 to 4.3). Two of these five children Echocardiographic detection in girls 1.3 (0.9–2.1)† 1.7 (1.0–2.8)‡
fulfilled the revised Jones criteria for acute rheu- vs. boys — odds ratio (95% CI)
matic fever and were treated. Involvement of mitral valve — % of 87.3 (78.2–93.0) 98.4 (91.9–99.7)
cases (95% CI)
Echocardiographic screening showed that 124
children (5.7%) had Doppler evidence of left-heart * Plus–minus values are means ±SD.
valve regurgitation. Sixty-six of these children were † P = 0.07.
concordantly confirmed by the three independent ‡ P = 0.04.
experts to have valvular rheumatic damage, corre-
sponding to a prevalence of 30.4 cases per 1000 The prevalence of subclinical rheumatic heart dis-
(95% CI, 23.2 to 37.6) (Fig. 1). Of these 66 children, ease increased with age from 2.4% in the lowest
63 (95.5%) had only mitral-valve disease, 2 had tercile (9 years or less) to 4.1% in the highest ter-
both aortic- and mitral-valve disease, and 1 had cile (12 years or more) (P = 0.04).
only aortic-valve disease (Fig. 2). No cases of val- By applying the observed prevalence in our
vular stenosis were detected. A pairwise kappa test sample of 2170 children to the whole population
showed excellent agreement among the three ob- of children 6 through 17 years of age in Maputo
servers. (1,140,000), we estimate that 2622 cases of rheu-
Of the 456 children with a detected cardiac matic heart disease would be detected by clinical
murmur, 91 were suspected to have organic lesions screening as compared with 34,656 cases that
at auscultation, but only 10 of them had cardiac would be detected by echocardiographic screening.
lesions confirmed by echocardiography (5 had
valvular lesions associated with rheumatic heart Dis cus sion
disease, and 5 had congenital heart lesions). The
prevalence of echocardiographically diagnosed Although there is little doubt that rheumatic heart
rheumatic heart disease was higher in girls than disease remains a major cause of morbidity and
in boys (odds ratio, 1.7; 95% CI, 1.0 to 2.8; P = 0.04) mortality in developing nations, recent reports of
and in suburban than in urban children (P = 0.04). approximately 15 million cases in the developing

n engl j med 357;5  www.nejm.org  august 2, 2007 473

The New England Journal of Medicine
Downloaded from nejm.org on April 10, 2018. For personal use only. No other uses without permission.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
 The n e w e ng l a n d j o u r na l of m e dic i n e

manifested disease in young adulthood.1 Prophy-

Cambodia Mozambique lactic antibiotic therapy and regular medical sur-
40 veillance, if adequately administered, may be high-
35 ly effective in preventing disease progression, even
No. of Cases per 1000 Children
in children with subclinical disease.3
30
The field surveys reported here were planned
25 and coordinated centrally from our institutions
20 located in Paris, and the choice of countries for
this project was based in part on our contacts in
15
and knowledge of government and medical ser-
10 vices in Cambodia and Mozambique. The initially
surprising findings in Cambodia prompted us to
5
undertake a similar study in Mozambique, with
0 more rigorous protocols for randomized selection
Clinical Echocardiographic
Screening Screening of schools and children, as well as protocols for
follow-up of subjects, to assess the generalizability
Figure 1. Prevalence of Rheumatic Valvular Abnormalities of our findings. Although environmental factors,
among Schoolchildren in Cambodia and Mozambique
genetic predisposition, specific virulence factors
as Detected by Clinical Screening with Echocardiograph-
ic Confirmation and by Echocardiographic Screening. of group A streptococcus serotypes, management
I bars AUTHOR:
TheICM Marijon
indicate 95% RETAKE
confidence intervals. 1st of antibiotic prophylaxis, and access to preventive
REG F FIGURE: 1 of 2 2nd care may all affect prevalence rates in specific ar-
3rd
CASE Revised eas, there is no reason to believe that our findings,
Line 4-CEMail SIZE which were remarkably similar in these two loca-
ARTIST: ts H/T H/T
1,2
worldEnon may underestimate the actual 16p6
burden of tions, would be valid only in the countries studied.
Combo
this disease, because these
AUTHOR, estimates
PLEASE NOTE: are based on We believe that this pattern is likely to be repli-
prevalence data
Figure has obtained
been redrawnbyand
comprehensive clinical
type has been reset.
Please check carefully.
cated elsewhere, while acknowledging that this
screening rather than echocardiographic screen- question may require considerably more investi-
ing. Our
JOB: findings suggest that echocardiograph-
35620 ISSUE: 05-17-07 gation.
ic screening might detect approximately 10 times Among more than 100 cross-sectional school
as many cases as clinical screening in school-age surveys of rheumatic heart disease reported to
children in Southeast Asia and sub-Saharan Africa. date, we are aware of only one that included echo-
Even echocardiographic screening of school- cardiography of all surveyed children.8 In this
children may underestimate the true prevalence study, which was conducted in Kenya, the preva-
of rheumatic heart disease, since the major risk lence of echocardiographically detected trivial mi-
factors for rheumatic fever include poverty, over- tral-valve regurgitation was 62 cases per 1000 and
crowding, and poor access to medical services, all the prevalence of clinically detected rheumatic
of which are probably associated with a reduced heart disease was 2.2 cases per 1000. Because
likelihood of attending school. Furthermore, the these investigators did not present any data on
high prevalence rates we observed were for two valve morphology or the characteristics of the re-
populations of children with an average age of only gurgitant jet, it was not clear whether the regur-
10 to 12 years, and community data from Aus- gitation was physiologic or was due to rheumatic
tralian aborigines suggest that the prevalence of or congenital disease. Indeed, many authors have
rheumatic heart disease is much higher in children emphasized the need to avoid overestimating the
over 15 years of age.1,6 prevalence of rheumatic heart disease, which can
There are at least two pressing reasons to ob- result from inclusion of all cases of minor valve
tain accurate prevalence data on rheumatic valvu- regurgitation, and to examine potentially affected
lar abnormalities in children: to guide regional valves for other structural or functional evidence
health service planning and to accurately identify of a rheumatic cause.4,9
affected children to be targeted for secondary There has been some debate concerning the
prevention. Although many cases of rheumatic echocardiographic signs of early rheumatic heart
valvular abnormalities in children may resolve disease.4,9-11 Some groups have preferred an echo-
spontaneously,7 many will progress to clinically cardiographic definition of rheumatic heart dis-

474 n engl j med 357;5  www.nejm.org  august 2, 2007

The New England Journal of Medicine

Downloaded from nejm.org on April 10, 2018. For personal use only. No other uses without permission.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
 Prevalence of Rheumatic Heart Disease in Asia and Africa

2370 Children were randomly

selected

200 Did not have consent provided

by a parent or guardian

2170 Underwent echocardiographic

screening

124 Had suspected rheumatic 5 Had congenital heart disease

2041 Had normal result
heart disease

124 Underwent definitive

echocardiography

58 Had normal result 66 Had rheumatic heart disease

5 Had clinical rheumatic 61 Had subclinical rheumatic

heart disease heart disease

Figure 2. Outcome of Echocardiographic Screening for Rheumatic Heart Disease in Mozambique.

Definitive echocardiography was performed to identify rheumatic heart disease, defined by the presence of any defi­
AUTHOR: Marijon RETAKE 1st
nite evidence of mitral- or aortic-valve
ICM regurgitation seen in two planes on Doppler echocardiography, accompanied
2nd
FIGURE: 2 of 2
by at least two of the following three morphologic abnormalities of the regurgitant 3rd
REG F valve: restricted leaflet mobility,
focal or generalized valvular thickening,
CASE and abnormal subvalvular thickening.
Revised
EMail Line 4-C SIZE
ARTIST: ts H/T H/T
Enon 33p9
Combo
ease based on the length, velocity, and persistence ver. Although
AUTHOR, PLEASE NOTE: portable ultrasound technology now
Figure has been redrawn and type has been reset.
of the regurgitant jet.3,12,13 However, these fea- makes it
Please check carefully. possible to perform most cardiac diag-
tures may depend on the gain settings on the nostic assessments in schools, our findings sup-
ultrasound equipment; thus,JOB:diagnosing
35705 subclini- port the need ISSUE: for a definitive cardiac ultrasound
08-02-07
cal rheumatic heart disease when Doppler crite- examination in suspected cases to avoid overdi-
ria for abnormalities are also associated with agnosis.
morphologic valve changes would be expected to Two recent reports have also advanced argu-
improve the specific diagnosis of rheumatic heart ments in favor of echocardiographic screening in
disease. In addition, the use of morphologic cri- rheumatic heart disease prevalence studies. Meira
teria combined with Doppler criteria for diagnos- and colleagues7 reported that in 34% of their pa-
ing rheumatic valve involvement may result in tients with carditis, clinical examination after the
underdiagnosis of subclinical rheumatic heart dis- acute phase showed normal findings, although
ease, particularly in children who are screened progression to chronic subclinical valvular disease
soon after a first episode of acute rheumatic fe- was confirmed by echocardiography in 82% of

n engl j med 357;5  www.nejm.org  august 2, 2007 475

The New England Journal of Medicine
Downloaded from nejm.org on April 10, 2018. For personal use only. No other uses without permission.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.
 Prevalence of Rheumatic Heart Disease in Asia and Africa

such cases. Furthermore, the latest report on rheu- from charitable organizations. Our results might
matic heart disease from the World Health Orga- encourage a strategy of echocardiography-based
nization states that in areas where rheumatic heart screening programs for rheumatic heart disease,
disease is endemic, echocardiography may be used involving education and both primary and second-
to diagnose “silent but significant rheumatic car- ary prophylaxis, and based on accurate assessment
ditis of insidious onset” and recommends that of the prevalence of rheumatic heart disease.14
such cases be managed as rheumatic heart disease In summary, we have documented a much
until proved otherwise.3 Although echocardio- higher prevalence of rheumatic valvular abnor-
graphic criteria are still not included in the Jones malities in school-age children than was previ-
criteria, this recommendation may represent a new ously suspected in both Southeast Asia and
step toward acknowledgment of subclinical le- sub-Saharan Africa. Comprehensive echocardio-
sions. graphic screening identified approximately 10
Even though portable echocardiographic equip- times as many children with rheumatic heart dis-
ment is now highly developed and accurate, echo- ease as were identified by the traditional strategy
cardiographic screening may not be a practicable of clinical screening with echocardiographic con-
local solution for case identification under the cur- firmation. Because rheumatic heart disease re-
rent financial constraints on provision of diagnos- mains a major cause of morbidity and mortality in
tic services in developing nations. Nevertheless, developing nations, these data have potentially
the observation that approximately 90% of cases important implications for case finding, delivery
of rheumatic heart disease in our study were clini- of effective primary and secondary prevention, and
cally silent, occurring in asymptomatic children adequate planning of health services.
without audible murmurs, suggests that echocar- Supported by grants from La Chaîne de l’Espoir, Paris, and
diographic screening would be a desirable goal to Cadeia da Esperança, Coimbra, Portugal.
No potential conflict of interest relevant to this article was
optimize case identification and targeted (second- reported.
ary) prevention measures. Funds for the purchase We thank the participating staff at the Phnom Penh Heart
of the ultrasound equipment — the major expense Center and the Maputo Heart Institute for their contributions to
the study, and P. Ducimetière, J.P. Empana, and A. Kane at
of an ultrasound-based screening program for INSERM, Unité 780 Avenir, Hôpital Paul Brousse, Assistance
rheumatic heart disease — might be obtainable Publique–Hôpitaux de Paris, for their technical assistance.

Appendix
The principal investigators in this study were P. Ou in Cambodia and E. Marijon and B. Ferreira in Mozambique.

References
1. Carapetis JR, McDonald M, Wilson NJ. 6. Carapetis JR, Currie BJ, Mathews JD. 11. Figueroa FE, Fernandez MS, Valdes P,
Acute rheumatic fever. Lancet 2005;366: Cumulative incidence of rheumatic fever et al. Prospective comparison of clinical
155-68. in an endemic region: a guide to the sus- and echocardiographic diagnosis of rheu-
2. Carapetis JR, Steer AC, Mulholland ceptibility of the population? Epidemiol matic carditis: long term follow up of pa-
EK, Weber M. The global burden of group Infect 2000;124:239-44. tients with subclinical disease. Heart 2001;
A streptococcal diseases. Lancet Infect Dis 7. Meira ZM, Goulart EM, Colosimo EA, 85:407-10.
2005;5:685-94. Mota CC. Long term follow up of rheu- 12. Minich LL, Tani LY, Pagotto LT, Shaddy
3. Rheumatic fever and rheumatic heart matic fever and predictors of severe RE, Veasy LG. Doppler echocardiography
disease: report of a WHO Expert Consul- rheumatic valvar disease in Brazilian chil- distinguishes between physiologic and
tation, Geneva, 29 October–1 November dren and adolescents. Heart 2005;91:1019- pathologic “silent” mitral regurgitation in
2001. Geneva: World Health Organiza- 22. patients with rheumatic fever. Clin Cardiol
tion, 2004. 8. Anabwani GM, Bonhoeffer P. Preva- 1997;20:924-6.
4. Veasy LG. Time to take soundings in lence of heart disease in school children 13. Wilson NJ, Neutze JM. Echocardio-
acute rheumatic fever. Lancet 2001;357: in rural Kenya using colour-flow echocar- graphic diagnosis of subclinical carditis in
1994-5. diography. East Afr Med J 1996;73:215-7. acute rheumatic fever. Int J Cardiol 1995;
5. Special Writing Group of the Commit- 9. Narula J, Kaplan EL. Echocardiograph- 50:1-6.
tee on Rheumatic Fever, Endocarditis, and ic diagnosis of rheumatic fever. Lancet 14. McDonald M, Brown A, Noonan S,
Kawasaki Disease of the Council on Car- 2001;358:2000. Carapetis JR. Preventing recurrent rheu-
diovascular Disease in the Young of the 10. Vasan RS, Shrivastava S, Vijayakumar matic fever: the role of register based pro-
American Heart Association. Guidelines M, Narang R, Lister BC, Narula J. Echo- grammes. Heart 2005;91:1131-3.
for the diagnosis of rheumatic fever: Jones cardiographic evaluation of patients with Copyright © 2007 Massachusetts Medical Society.
Criteria, 1992 update. JAMA 1992;268: acute rheumatic fever and rheumatic car-
2069-73. [Erratum, JAMA 1993;269:476.] ditis. Circulation 1996;94:73-82.

476 n engl j med 357;5  www.nejm.org  august 2, 2007

The New England Journal of Medicine

Downloaded from nejm.org on April 10, 2018. For personal use only. No other uses without permission.
Copyright © 2007 Massachusetts Medical Society. All rights reserved.