Articles

Global, regional, and national burden of congenital heart

disease, 1990–2017: a systematic analysis for the
Global Burden of Disease Study 2017
GBD 2017 Congenital Heart Disease Collaborators*

Summary
Background Previous congenital heart disease estimates came from few data sources, were geographically narrow, Lancet Child Adolesc Health
and did not evaluate congenital heart disease throughout the life course. Completed as part of the Global Burden 2020; 4: 185–200

of Diseases, Injuries, and Risk Factors Study 2017, this study aimed to provide comprehensive estimates of Published Online
January 21, 2020
congenital heart disease mortality, prevalence, and disability by age for 195 countries and territories from
https://doi.org/10.1016/
1990 to 2017. S2352-4642(19)30402-X
This online publication has
Methods Mortality estimates were generated for aggregate congenital heart disease and non-fatal estimates for been corrected. The corrected
five subcategories (single ventricle and single ventricle pathway congenital heart anomalies; severe congenital heart version first appeared at
thelancet.com/child-
anomalies excluding single ventricle heart defects; critical malformations of great vessels, congenital valvular heart
adolescent on February 7, 2020
disease, and patent ductus arteriosus; ventricular septal defect and atrial septal defect; and other congenital
See Comment page 168
heart anomalies), for 1990 through to 2017. All available global data were systematically analysed to generate
*Collaborators listed at the end
congenital heart disease mortality estimates (using Cause of Death Ensemble modelling) and prevalence estimates of the Article
(DisMod-MR 2·1). Systematic literature reviews of all types of congenital anomalies to capture information on
Correspondence to:
prevalence, associated mortality, and long-term health outcomes on congenital heart disease informed subsequent Prof Craig A Sable, Department
disability estimates. of Cardiology, Children’s National
Health System, Washington,
DC 20010, USA
Findings Congenital heart disease caused 261 247 deaths (95% uncertainty interval 216 567–308 159) globally in 2017, [email protected]
a 34·5% decline from 1990, with 180 624 deaths (146 825–214 178) being among infants (aged <1 years). Congenital
heart disease mortality rates declined with increasing Socio-demographic Index (SDI); most deaths occurred in
countries in the low and low-middle SDI quintiles. The prevalence rates of congenital heart disease at birth changed
little temporally or by SDI, resulting in 11 998 283 (10 958 658–13 123 888) people living with congenital heart disease
globally, an 18·7% increase from 1990 to 2017, and causing a total of 589 479 (287 200–973 359) years lived with
disability.

Interpretation Congenital heart disease is a large, rapidly emerging global problem in child health. Without the ability
to substantially alter the prevalence of congenital heart disease, interventions and resources must be used to improve
survival and quality of life. Our findings highlight the large global inequities in congenital heart disease and can serve
as a starting point for policy changes to improve screening, treatment, and data collection.

Funding Bill & Melinda Gates Foundation.

Copyright © 2020 The Author(s). Published by Elsevier Ltd. This is an Open Access article under the CC BY 4·0
license.

Introduction of the world, but the same success rates are not yet seen
Diagnostic and treatment capabilities for congenital in developing regions.
heart disease have dramatically improved over the past The Sustainable Development Goals (SDGs) were For the SDGs see http://www.
80 years. In the Metropolitan Atlanta Congenital Defects adopted by the UN in 2016. SDG 3.2 aims to reduce the un.org/sustainabledevelopment/
blog/2015/12/sustainable-
Program, infant survival with critical congenital heart mortality of neonates to less than 12 deaths per 1000 live development-goals-kick-off-
disease improved from 67·4% for the 1979–93 birth births and the mortality of children to less than 25 deaths with-start-of-new-year/
cohort to 82·5% for the 1994–2005 cohort.1 A Finnish per 1000 live births. SDG 3.4 aims to reduce premature
registry study similarly showed a decrease in early and mortality due to non-communicable diseases (NCDs) by
late post-operative congenital heart disease mortality one-third by 2030. Congenital heart disease accounts for
in 1990–2009 compared with 1953–89.2 Reports from nearly one-third of all congenital birth defects5 and,
Belgium and Sweden found that 90–95% of children therefore, the focus on congenital heart disease is integral
with congenital heart disease born between 1972 and the to eliminating preventable child deaths and NCDs in the
early 1990s survived into adulthood.3,4 These studies show SDG era.6 Nine-tenths of the world’s children born with
substantial improvement in survival in developed regions congenital heart disease live in locations with little to no

www.thelancet.com/child-adolescent Vol 4 March 2020 185

 Articles

Research in context
Evidence before this study mortality as the outcome measures. This study incorporates all
Previous work to describe the global burden of congenital available data sources and previous publications making it the
heart disease was generated from a small number of sources, most comprehensive study on congenital heart disease burden
which were geographically restricted and included only broad to date. Estimates of prevalence were more accurate than in
categories of congenital heart disease. Previous studies did not previous studies because remission rates for specific subtypes
comprehensively evaluate congenital heart disease throughout of congenital heart disease were taken into account and
the lifecourse, instead relying on strong assumptions about congenital heart disease was divided into five subcategories
their distribution and natural history. We did a systematic based on anatomic and clinical similarities for improved clarity
literature review for prevalence, excess mortality, with- of disease burden by lesion. This study not only gives a more
condition mortality, congenital birth defects registries, and accurate assessment of the global burden of congenital heart
hospital administrative data. The literature was also reviewed disease, but also highlights the disproportionate disease burden
for outcomes related to congenital heart disease, most notably in low Socio-demographic Index (SDI) regions compared to
intellectual disability. No previous studies incorporated higher SDI regions.
morbidity or provided years lived with disability (YLDs).
Implications of all the available evidence
Both measures are informative to policy makers when
With the use of GBD 2017 data, this research provides a
assessing lifelong implications and the loss of healthy years in
comprehensive overview of the global burden of congenital
comparison with other childhood diseases. We searched
heart disease and shows a substantial burden of disease
PubMed for English-language research articles describing the
throughout the life cycle. Congenital heart disease burden is
global burden of congenital heart disease published between
especially notable in low SDI regions, where a greater variety
Jan 1, 2010, and June 30, 2019, using the terms “pediatric or
of diseases exist and where data is especially needed to better
childhood or child” AND “congenital heart disease or
inform priorities for policymakers. As countries and
congenital heart defect” AND “global or international or
international agencies implement, monitor, and evaluate the
worldwide or world” AND “burden OR metrics OR incidence OR
UN’s Sustainable Development Goals, refined and updated
mortality OR prevalence OR survival” but did not find
estimates of congenital heart disease burden is crucially
additional applicable work.
important to identify areas in need. This research provides
Added value of this study vital information to governments, stakeholders,
To our knowledge, this is the first report of congenital heart policymakers, and the global health community worldwide,
disease burden using Global Burden of Diseases, Injuries, and for the present and the future, as GBD data continues to be
Risk Factors Study 2017 results, with prevalence, YLDs, and refined and improved.

care and where mortality remains high.7 Despite some The Global Burden of Diseases, Injuries, and Risk
improvements in the past two decades, neonates in Factors Study (GBD) is an international collaboration of
low-income and middle-income countries (LMICs) with researchers who annually produce internally consistent
severe forms of congenital heart disease and without estimates of death and disability throughout the world.
access to surgical treatment are more likely to die before The 2017 update of GBD generated the most up-to-date
their fifth birthday than are those in high income countries estimates of congenital heart disease prevalence and
(HICs).8 Additional premature mortality might occur mortality throughout the world. This analysis placed
secondary to cardiac and pulmonary complications.8 particular emphasis on developing data-driven estimates
Optimal allocation of resources for screening, clinical for the entire lifecourse. Since 2016, GBD has generated
care, and development of cost-effective treatment estimates of congenital heart disease burden by five
strategies requires accurate assessment of the absolute different anatomic subtypes, including consideration of
and relative burden of congenital heart disease remission of some subtypes of congenital heart disease. It
across the world. Previous publications have estimated incorporated many different data sources on mortality and
250 000 annual congenital heart disease deaths globally, prevalence and employed a variety of statistical approaches
1·35 million annual births of neonates with congenital to maximise the robustness of the final results. We report
heart disease globally,5 and 2·4 million people living with here on the GBD 2017 approach and results, to describe
congenital heart disease in the USA alone.9 However, the geographical, temporal, and sociodemographic trends
even the global studies were generated from a small of congenital heart disease from 1990 to 2017.
number of geographically limited sources, included only
broad categories of congenital heart disease, and did Methods
not comprehensively evaluate congenital heart disease Overview
throughout the lifecourse,10 instead relying on strong Detailed methodology for congenital heart disease
See Online for appendix assumptions about their distribution and natural history. estimates is provided in the appendix. Detailed methods

186 www.thelancet.com/child-adolescent Vol 4 March 2020

 Articles

for each analytic step in GBD 2017 are described indoor air pollution, and repro­ductive age-standardised
elsewhere.11–14 Reporting is compliant with the Guidelines smoking prevalence. Covariates having a prespecified
for Accurate and Transparent Health Estimates Reporting negative correlation with congenital heart disease
(GATHER).15 All input data are available online via the mortality were measles vaccine coverage, education
GBD Input Data Sources Tool of the Global Health Data (years per capita), and legality of abortion. For the Global Health Data
Exchange. Exchange see http://ghdx.
healthdata.org
Datasets for estimating congenital heart disease Non-fatal disease estimates
epidemiology were informed by three mechanisms: Starting with GBD 2016, in consultation with GBD
systematic literature review for prevalence, excess experts and paediatric cardiology experts, we have
mortality, and with-condition mortality; congenital birth reorganised cause categories for non-fatal estimates to
defects registries; and hospital administrative data. The be more anatomically oriented, clinically relevant, and
volume of data included in the final models is structured in a way to facilitate incorporation of
summarised in the appendix (pp 12–14 and 20). Each literature, registry, and administrative data to inform
model’s data and fit can be examined in the GBD online levels and trends. These five congenital heart disease
Epi Visualization tool. categories are named descriptively as: single ven­tricle For the Epi Visualization tool
and single ventricle pathway congenital heart anom­ see http://vizhub.healthdata.
org/epi
Data availability alies; severe congenital heart anomalies excluding
Availability of data for fatal and non-fatal cases of single ventricle heart defects; critical malformations of
congenital heart disease varied widely across countries great vessels, congenital valvular heart disease, and
and regions (appendix p 13). Temporal coverage of data patent ductus arteriosus; ventricular septal defect and
by GBD region for each of mortality and non-fatal atrial septal defect; and other congenital heart
analyses are shown in the appendix (p 14). Cause- anomalies.
specific mortality data were identified from most ICD-9 and ICD-10 codes mapped to each category are
countries globally, but from few countries in Africa, shown in the appendix (p 11).
southeast Asia, or Oceania. A flowchart illustrating the Five separate DisMod-MR 2.1 models were developed
estimation process is shown in the appendix (p 15). for each of the congenital heart disease categories.
Deaths due to congenital heart disease were estimated DisMod-MR 2.1 is a compartmental model consisting of
using a common cause-specific mortality framework three states—susceptible, diseased, and dead—with
developed for GBD 2017. Each death was assigned to a state transitions determined by the incidence, remission
single underlying cause. Deaths assigned to ill-defined rate, excess mortality, and other-cause mortality.16
or non-specific causes (eg, “heart disease, unspecified”, DisMod-MR 2.1 uses a geographical cascade to borrow
International Classification of Diseases, 10th revision strength across geography and time, balancing a set of
[ICD-10] code I51.9) or intermediate causes (eg, “heart differential equations to find an internally consistent
failure”, ICD-10 code I50) were reassigned to probable numerical solution for each age, sex, location, and year
causes of death, including congenital heart disease, that best reflects the input data and any specified
with the use of statistical redistribution algorithms. constraints on values of the calculated rates. Incidence
Illustrations of redistribution for ICD-9 and ICD-10 in was set to 0 for all congenital models, as congenital
the USA are shown in the appendix (p 15, 16). Non-fatal conditions occur at the time of birth and by definition
data were identified from a few additional countries there are no incident cases after birth. Remission was
with a low Socio-demographic Index (SDI), although the bounded to 0% for all subcause models except
temporal coverage of non-fatal data was categorically ventricular and atrial septal defect, for which it was
lower than was that of mortality data. A flowchart constrained to be between 0% and 20% per year until
illustrating the estimation process is shown in the age 10 years on the basis of the values that corresponded
appendix (p 18). to observed spontaneous closure in longitudinal
studies.17,18 For most of the severe congenital conditions,
Mortality estimates the mortality associated with the condition is highly
Congenital heart disease mortality was estimated using dependent on access to adequate surgical interventions
the Cause of Death Ensemble model (CODEm), a tool and other medical care during the first hours, weeks,
that runs many different models on the same data and and years of life. Therefore, a decreasing slope on excess
chooses an ensemble of models that best reflects all the mortality was applied to capture the highest risk of
available input data. The covariates used in CODEm mortality from congenital conditions in the neonatal
models for total congenital birth defects and congenital age groups and a subsequent decreasing risk of mortal­
heart disease are listed in the appendix (p 17). Covariates ity from congenital conditions later in life. Model
having a prespecified positive correlation with con­ covariates and calculated coefficients for each are shown
genital heart disease mortality were maternal alcohol in the appendix (pp 21, 22).
consumption, proportion of live births in women aged The proportions of each type of congenital heart
35 years or older, age-standardised diabetes prevalence, disease with each type of sequela were derived using

www.thelancet.com/child-adolescent Vol 4 March 2020 187

 Articles

Number of deaths (all ages) Mortality per 100 000 children younger than 1 year Age-standardised mortality per 100 000
individuals
1990 2017 Percentage 1990 2017 Percentage 1990 2017 Percentage
change, change, change,
1990–2017 1990–2017 1990–2017
Global 398 580 261 247 –34·5% 217·1 131·0 –39·7% 6·3 3·9 –39·0%
(268 422 to (216 567 to (–44·6 to (146·8 to (106·5 to (–49·1 to –27·1)* (4·3 to 8·0) (3·2 to 4·6) (–48·2 to
506 469) 308 159) –19·8)* 277·5) 155·3) –26·0)*
Low SDI 67 340 83 873 24·6% 179·5 167·9 –6·5% 5·5 4·9 –10·6%
(25 251 to (56 785 to (–13·1 to 130·6) (71·7 to 287·6) (115·7 to 221·1) (–35·0 to 63·1) (2·1 to 8·9) (3·4 to 6·4) (–37·9 to 61·8)
109 078) 110 218)
Low-middle SDI 103 086 82 667 –19·8% 225·6 147·9 –34·5% 6·8 4·4 –34·9%
(60 630 to (64 667 to (–37·2 to 8·1) (135·8 to 310·2) (110·5 to 189·4) (–48·3 to –14·6)* (4·0 to 9·2) (3·5 to 5·6) (–49·0 to –12·9)*
140 879) 104 233)
Middle SDI 127 734 58 512 –54·2% 249·3 120·0 –51·8% 7·0 3·5 –49·6%
(98 849 to (50 989 to (–64·8 to –41·8)* (190·1 to (102·8 to 132·4) (–63·5 to –38·5)* (5·5 to 8·5) (3·1 to 3·8) (–61·3 to –36·2)*
154 111) 63 052) 300·6)
High-middle SDI 76 987 26 948 –65·0% 262·8 99·9 –62·0% 7·3 3·0 –59·3%
(61 961 to (24 814 to (–71·0 to –54·8)* (206·5 to (89·2 to 109·1) (–69·3 to –49·2)* (5·8 to 8·4) (2·7 to 3·2) (–66·5 to –47·4)*
89 280) 29 373) 308·3)
High SDI 21 607 8500 –60·7% 101·7 36·6 –64·0% 3·1 1·2 –61·7%
(17 568 to (7653 to (–64·1 to –52·7)* (78·2 to 110·8) (32·8 to 41·7) (–68·0 to –54·5)* (2·5 to 3·4) (1·1 to 1·4) (–65·5 to –52·9)*
23 083) 9992)
Central Europe, eastern Europe, and central Asia
Central Asia 3834 3155 –17·7% 150·0 119·1 –20·6% 4·2 3·5 –17·2%
(3183 to 4356) (2491 to 3815) (–34·2 to 0·7) (125·3 to 170·0) (94·2 to 148·1) (–36·2 to –2·7)* (3·5 to 4·8) (2·7 to 4·2) (–34·0 to 1·2)
Central Europe 4969 1111 –77·6% 212·6 59·4 –72·1% 5·7 1·8 –68·9%
(3898 to 5568) (970 to 1247) (–80·8 to –70·1)* (162·0 to 242·9) (51·9 to 68·4) (–77·2 to –60·9)* (4·4 to 6·4) (1·6 to 2·0) (–73·7 to –57·5)*
Eastern Europe 7409 2751 –62·9% 161·8 60·7 –62·5% 4·7 2·0 –56·8%
(6242 to 8297) (2390 to 3530) (–68·3 to –51·8)* (136·7 to 182·1) (52·0 to 77·7) (–68·8 to –50·0)* (3·9 to 5·3) (1·7 to 2·6) (–63·6 to –42·7)*
High income
Australasia 374 197 –47·5% 70·5 26·3 –62·8% 2·3 0·9 –59·8%
(321 to 429) (159 to 236) (–56·1 to –37·2)* (60·3 to 81·6) (19·8 to 33·9) (–72·3 to –51·0)* (2·0 to 2·6) (0·7 to 1·1) (–67·0 to –50·2)*
High-income 3608 897 –75·2% 104·1 29·3 –71·9% 3·3 1·0 –71·2%
Asia Pacific (2624 to 4060) (792 to 1121) (–79·1 to –63·2)* (73·3 to 120·3) (25·4 to 35·1) (–77·0 to –57·4)* (2·4 to 3·8) (0·8 to 1·2) (–76·1 to –57·0)*
High-income 6430 3522 –45·2% 84·1 37·8 –55·0% 2·8 1·4 –51·0%
North America (5446 to 7079) (3137 to 4311) (–49·3 to –36·0)* (70·0 to 91·9) (33·2 to 45·8) (–59·9 to –46·5)* (2·3 to 3·0) (1·2 to 1·7) (–55·0 to –42·4)*
Southern Latin 2129 1259 –40·9% 159·3 85·6 –46·3% 4·3 2·4 –43·8%
America (1766 to 2450) (946 to 1509) (–54·6 to –24·7)* (131·0 to 185·2) (63·1 to 106·0) (–60·7 to –29·0)* (3·6 to 5·0) (1·8 to 2·9) (–57·2 to –28·2)*
Western Europe 7630 2539 –66·7% 99·4 29·2 –70·6% 3·0 1·0 –68·5%
(5736 to 8323) (2138 to 2 848) (–70·4 to –59·1)* (64·5 to 113·0) (22·3 to 33·2) (–74·5 to –61·4)* (2·2 to 3·3) (0·8 to 1·1) (–72·2 to –60·0)*
Latin America and Caribbean
Andean Latin 1891 2026 7·1% 115·0 108·1 –6·0% 3·4 3·1 –6·9%
America (1359 to 2596) (1338 to 2593) (–44·3 to 55·6) (81·2 to 161·8) (69·5 to 143·4) (–53·5 to 41·5) (2·4 to 4·6) (2·1 to 4·0) (–51·2 to 34·3)
Caribbean 2744 2120 –22·7% 218·6 193·4 –11·5% 6·6 5·4 –17·9%
(1692 to 3 463) (1442 to 2825) (–41·9 to –3·0)* (138·6 to (129·5 to 270·3) (–35·5 to 14·4) (4·1 to 8·3) (3·7 to 7·2) (–38·6 to 3·2)
281·3)
Central Latin 10 767 8965 –16·7% 165·6 129·2 –22·0% 4·8 3·7 –21·8%
America (9594 to 14 534) (6154 to 10 275) (–52·9 to 1·9) (146·8 to (85·3 to 150·7) (–58·5 to –1·9)* (4·3 to 6·4) (2·6 to 4·3) (–55·9 to –4·2)*
222·6)
Tropical Latin 19 295 5962 –69·1% 449·3 144·5 –67·8% 11·3 3·7 –67·1%
America (9542 to 26 614) (5095 to 6896) (–78·5 to –38·6)* (209·9 to (118·1 to 165·1) (–78·1 to –32·1)* (5·6 to 15·5) (3·1 to 4·3) (–77·2 to –35·0)*
625·3)
North Africa and Middle East
North Africa and 60 031 35 710 –40·5% 450·8 211·7 –53·0% 11·8 5·8 –51·0%
Middle East (30 316 to (29 606 to (–54·5% to (221·5 to 653·2) (168·5 to (–64·0% to (6·0 to 16·6) (4·8 to 7·2) (–62·6% to
84 703) 44 359) –0·3%)* 269·5) –22·2%)* –18·6%)*
South Asia
South Asia 75 091 66 984 –10·8% 154·8 138·5 –10·6% 4·8 4·0 –17·0%
(38 351 to (50 484 to (–38·2 to 34·5) (76·6 to 223·5) (95·5 to 176·6) (–38·3 to 29·4) (2·5 to 7·0) (3·0 to 4·9) (–42·1 to 21·0)
109 094) 81 590)
(Table 1 continues on next page)

188 www.thelancet.com/child-adolescent Vol 4 March 2020

 Articles

Number of deaths (all ages) Mortality per 100 000 children younger than 1 year Age-standardised mortality per 100 000
individuals
1990 2017 Percentage 1990 2017 Percentage 1990 2017 Percentage
change, change, change,
1990–2017 1990–2017 1990–2017
(Continued from previous page)
Southeast Asia, east Asia, and Oceania
East Asia 93 878 31 885 –66·0% 254·8 112·5 –55·9% 7·4 3·4 –54·5%
(80 522 to (29 266 to (–72·5 to –58·2)* (217·6 to 312·0) (103·0 to 125·7) (–66·0 to –46·2)* (6·3 to 8·9) (3·1 to 3·7) (–63·6 to –44·1)*
113 589) 35 408)
Oceania 775 1 307 68·8% 244·2 226·4 –7·3% 8·3 7·7 –6·7%
(440 to 1071) (790 to 1793) (23·8 to 126·7)* (137·7 to (126·0 to 327·7) (–35·6 to 32·8) (4·9 to 11·2) (4·8 to 10·5) (–30·5 to 23·6)
348·0)
Southeast Asia 42 680 23 976 –43·8% 268·3 153·1 –42·9% 7·3 4·4 –40·4%
(25 044 to (19 721 to 27 605) (–56·5 to –19·3)* (151·0 to 352·8) (120·2 to 181·0) (–55·3 to –17·6)* (4·3 to 9·6) (3·6 to 5·1) (–53·8 to –15·3)*
55 932)
Sub-Saharan Africa
Central sub- 8152 11 257 38·1% 231·0 189·5 –17·9% 7·7 5·9 –23·4%
Saharan Africa (3377 to 14 002) (7101 to 17 318) (–0·3 to 128·0) (112·1 to 388·3) (121·2 to 288·3) (–43·0 to 30·0) (3·2 to 13·1) (3·8 to 9·0) (–44·2 to 26·7)
Eastern sub- 25 818 26 998 4·6% 208·4 141·5 –32·1% 6·9 4·4 –36·9%
Saharan Africa (11 083 to (19 197 to 38 981) (–21·1 to 97·0) (100·9 to (99·7 to 203·7) (–47·8 to 16·3) (3·1 to 11·9) (3·1 to 6·3) (–52·6 to 17·2)
44 514) 344·1)
Southern sub- 1568 1252 –20·1% 62·7 42·1 –32·8% 2·4 1·5 –34·8%
Saharan Africa (1301 to 1894) (1084 to 1444) (–36·4 to 4·2) (51·0 to 78·2) (35·0 to 50·5) (–48·2 to –10·3)* (2·0 to 2·8) (1·3 to 1·8) (–47·8 to –16·2)*
Western sub- 19 507 27 373 40·3% 164·6 121·1 –26·4% 5·5 4·0 –27·5%
Saharan Africa (7079 to 30 605) (15 756 to 39 149) (7·1 to 121·7)* (68·0 to 261·0) (76·8 to 169·4) (–46·0 to 12·2) (2·1 to 8·6) (2·4 to 5·6) (–44·6 to 14·0)
Data in parentheses are 95% uncertainty intervals. SDI=Socio-demographic Index. *Significant increase or decrease.

Table 1: Global Burden of Diseases, Injuries, and Risk Factors Study 2017 congenital heart disease mortality, 1990–2017

available information to calculate years lived with a health state of con­genital heart disease. Evidence for
disability (YLDs). To assess the distribution of health congenital heart disease-related disability universally
outcomes associated with the congenital causes, we was presented separately for intellectual disability and
extracted data from the same studies identified in the heart failure. We therefore assumed independence of
systematic review done for the estimation of the the probability of each in calculating the proportion of
prevalence of fatal and non-fatal congenital heart disease, cases of congenital heart disease in each combination
this time focused on the long-term health outcomes of of health states (eg, mild heart failure and profound
survivors in cohorts born with each type of congenital intellectual disability).
malformation. For conditions requiring surgical inter­ To assess the distribution of health outcomes
vention shortly after birth to ensure survival, the health associated with the most common causes of heart
states included in the disability weight calculations disease that have origins in childhood, we compared
correspond to the post-surgery outcomes reported in YLDs and years of life lost (YLLs) from congenital heart
cohorts of individuals born with these life-threatening disease with YLDs and YLLs from rheumatic heart
congenital conditions. disease.
Heart failure distribution was derived from the
updated GBD 2017 heart failure analysis using the Role of the funding source
same approach of proportional allocation based on The funder of the study played no role in study design,
cause-specific mortality as previously described.19 data collection, data analysis, data interpretation, or
Several scientific literature sources reported on the writing of the report. All authors had full access to all the
prevalence and severity of intel­ lectual disability in data in the study and had final responsibility for the
populations with congenital heart defect;20–22 the decision to submit for publication.
asymptomatic proportion of individuals with ven­
tricular and atrial septal defects (including those who Results
had undergone repair) was derived from literature Congenital heart disease was the underlying cause of
sources on the long-term outcomes of patients an estimated 261 247 (95% uncertainty interval [UI]
diagnosed with and, in some cases, treated for, septal 216 567–308 159) deaths globally in 2017, a 34·5%
defects at birth.23 All except those with asymptomatic (19·8–44·6) decline from 1990, when the number of
ventricular and atrial septal defects were also assigned deaths was 398 580 (268 422–506 469; table 1). Of all

www.thelancet.com/child-adolescent Vol 4 March 2020 189

 Articles

Mortality rate per 100 000 infants

<34·1
34·1 to <68·5
68·5 to <113·6
113·6 to <151·9
≥151·9

ATG VCT Barbados Comoros Marshall Isl Kiribati

West Africa Eastern
Mediterranean
Solomon Isl FSM

Dominica Grenada Maldives Mauritius Malta

Vanuatu Samoa

Caribbean LCA TTO TLS Seychelles Persian Gulf Singapore Balkan Peninsula Fiji Tonga

Proportion of total infant deaths

<4·1%
4·1% to <6·7%
6·7% to <9·0%
9·0% to <13·2%
≥13·2%

ATG VCT Barbados Comoros Marshall Isl Kiribati

West Africa Eastern
Mediterranean
Solomon Isl FSM

Dominica Grenada Maldives Mauritius Malta

Vanuatu Samoa

Caribbean LCA TTO TLS Seychelles Persian Gulf Singapore Balkan Peninsula Fiji Tonga

(Figure 1 continues on next page)

190 www.thelancet.com/child-adolescent Vol 4 March 2020

 Articles

congenital heart disease deaths in 2017, 180 624

C SDI quintiles
(146 825–214 178; 69%) occurred in infants younger 275 Global
than 1 year (appendix pp 30–66). Congenital heart Low SDI
Low-middle SDI
disease deaths and mortality were highest in the low Middle SDI
and low-middle SDI quintiles. In 2017, regions with 220 High-middle SDI
the highest infant mortality due to congenital heart High SDI

Mortality per 100 000

disease were, in descending order, Oceania, North
165
Africa and the Middle East, the Caribbean, central sub-
Saharan Africa, and southeast Asia (table 1). The
regions with the highest mortality rates also had the 110
highest uncertainty intervals (table 1; figure 1A;
appendix pp 13, 14). Rates of improvement tracked with
55
development status, with more locations with higher
SDIs generally progressing faster than those with
lower SDI. The fastest declines in infant mortality due 0
to congenital heart disease occurred in central Europe,
high-income Asia Pacific, western Europe, tropical D
Latin America, Australasia, and eastern Europe, all of 15

which have experienced declines of at least 60% since

1990. Infant mortality rates due to congenital heart 12
disease have declined across all SDI quintiles except
Proportion of total deaths (%)

low SDI (figure 1C). However, improvement in

congenital heart disease mortality has lagged behind 9
that of other causes. Therefore, the proportion of all
infant deaths caused by congenital heart disease
6
increased in all quintiles except the high-SDI quintile
(figures 1B, 1D).
In the high-SDI quintile in 2017, neonatal preterm 3
birth, other neonatal disorders, neonatal encepha­
lopathy, congenital heart disease, and sudden infant
0
death syndrome were the top five causes of infant 1990 1992 1994 1996 1998 2000 2002 2004 2006 2008 2010 2012 2014 2016
mortality (figure 2; appendix p 21). Congenital heart Year
disease rose to be among the top eight causes of infant
mortality by 2017 in all SDI quintiles. Two conditions Figure 1: Mortality of congenital heart disease in children under 1 year of age, 1990–2017
(A) Mortality per 100 000 infants by country in 2017 and (C) by SDI from 1990 to 2017. (B) Proportion of infant
closely related to congenital heart disease were deaths by country in 2017 and (D) by SDI from 1990 to 2017. Subnational data are available for Brazil, China,
also highly ranked: lower respiratory infections (ie, India, and the USA. ATG=Antigua and Barbuda. FSM=Federated States of Micronesia. LCA=Saint Lucia.
pneumonia), which was ranked in the top five in all SDI=Socio-demographic Index. TLS=Timor-Leste. TTO=Trinidad and Tobago. VCT=Saint Vincent and the
Grenadines.
except the high-SDI quintile, and neonatal preterm
birth, which was ranked either first or second in all
strata of SDI. global level. This effect was more pronounced in the
A total of 11 998 283 (95% UI 10 958 658–13 123 888) middle and high-middle SDI quintiles. Estimates of
people were estimated to be living with congenital heart congenital heart disease prevalence (for all congenital
disease globally in 2017, an 18·7% increase from heart disease and congenital heart disease subtypes) and
10 105 235 (9 159 197–11 150 207) prevalent cases in 1990. number of deaths (for all congenital heart disease), by
The increase is related to improved survival and age group and GBD region are shown in the
population growth, partially offset by ageing. There was appendix (pp 30–66).
an estimated prevalence of 466 566 (429 140–505 806) Since 1990, the rate of YLDs has remained around
patients with congenital heart disease in the USA, of 7 YLDs per 100 000 individuals in high-SDI countries
whom 279 320 (266 461–331 437) are younger than and around 14 YLDs per 100 000 individuals in low-SDI
20 years. The birth prevalence of all congenital heart countries. In 2017, the estimated global number of all-
disease was 1787·6 cases (1587·0–1999·2) per ages YLDs due to congenital heart disease was
100 000 babies in 2017, which has increased by 4·2% 589 479 YLDs (95% UI 287 200–973 359); YLD rates were
(2·8–5·6) since 1990 (table 2). The all-ages global highest in infants younger than 1 year. Infant YLD rates
prevalence of most subcategories of congenital heart varied significantly throughout the world and within
disease remained unchanged from 1990 to 2017 (figure 3). the USA, with the highest rates estimated for western,
The exception was the ventricular and atrial septal defect central, and eastern sub-Saharan Africa, central and
subcategory, which decreased from 1990 to 2017 at the southeast Asia, and some internal provinces of China

www.thelancet.com/child-adolescent Vol 4 March 2020 191

 Articles

1990
Global Low SDI Low-middle SDI Middle SDI High-middle SDI High SDI
Lower respiratory infections 1 1 1 1 2 7
Neonatal preterm birth 2 3 2 2 1 1
Diarrhoeal diseases 3 2 3 4 5 18
Neonatal encephalopathy due to birth asphyxia and trauma 4 4 4 3 3 4
Other neonatal disorders 5 5 5 6 6 6
Congenital heart disease 6 13 8 5 4 2
Neonatal sepsis and other neonatal infections 7 9 6 7 9 8
Tetanus 8 6 10 10 22 82
Protein-energy malnutrition 9 8 9 9 12 53
Measles 10 10 7 13 17 62
Malaria 11 7 11 35 86 115
Other congenital anomalies 12 17 13 8 7 5
Haemolytic disease and other neonatal jaundice 13 12 12 16 15 30
Syphilis 14 11 14 12 18 23
Whooping cough 15 15 17 14 14 51
Drug-susceptible tuberculosis 16 14 16 17 21 49
Neural tube defects 17 19 19 15 8 12
Other meningitis 18 16 15 19 23 19
Digestive congenital anomalies 19 23 20 18 10 10
Sudden infant death syndrome 20 20 21 22 13 3
Pulmonary aspiration and foreign body in airway 21 29 22 11 11 9
H influenzae type B meningitis 22 21 18 21 16 22
Paralytic ileus and intestinal obstruction 23 28 23 20 19 24
Other unspecified infectious diseases 24 24 25 26 25 25
Visceral leishmaniasis 25 18 43 89 96 102

2017

Neonatal preterm birth 1 2 1 1 1 1

Lower respiratory infections 2 1 2 2 5 10
Neonatal encephalopathy due to birth asphyxia and trauma 3 3 3 3 3 3
Other neonatal disorders 4 5 5 5 4 2
Diarrhoeal diseases 5 4 4 7 11 15
Neonatal sepsis and other neonatal infections 6 6 6 6 7 8
Congenital heart disease 7 8 7 4 2 4
Malaria 8 7 8 30 80 117
Syphilis 9 9 10 9 10 19
Other congenital anomalies 10 11 9 8 6 6
Protein-energy malnutrition 11 10 12 12 29 60
Haemolytic disease and other neonatal jaundice 12 15 11 16 19 37
Neural tube defects 13 13 16 13 12 12
Digestive congenital anomalies 14 17 15 11 8 11
Other meningitis 15 14 13 20 30 27
Whooping cough 16 12 17 19 33 53
Sudden infant death syndrome 17 16 18 15 13 5
HIV/AIDS resulting in other diseases 18 18 14 17 22 39
Pulmonary aspiration and foreign body in airway 19 27 22 10 9 9
Drug-susceptible tuberculosis 20 19 20 23 44 78
Measles 21 20 21 22 40 86
H influenzae type B meningitis 22 23 19 25 23 30
Paralytic ileus and intestinal obstruction 23 25 23 14 17 21
Other unspecified infectious diseases 24 22 27 29 18 20
Tetanus 25 21 24 36 68 97

Death rate per 100 000 children

0·0 to 8·27 >8·27 to 24·01 >24·01 to 47·23 >47·23 to 88·16 >88·16 to 227·63 >227·63 to 2214·30

Figure 2: Leading causes of death in children younger than 1 year

The heat map depicts rank order of causes of death (by death rate per 100 000) in children younger than 1 year in 1990 and 2017. The rank of congenital heart disease
increased from ninth to seventh globally and 17th to 11th in low SDI countries from 1990 to 2017. SDI=Socio-demographic Index.

192 www.thelancet.com/child-adolescent Vol 4 March 2020

 Articles

Prevalence at birth per 100 000 babies Prevalence per 100 000 children younger Prevalence per 100 000 individuals Age-standardised prevalence rate
than 1 year (all ages) per 100 000
1990 2017 Percent 1990 2017 Percent 1990 2017 Percent 1990 2017 Percent
change, change, change, change,
1990–2017 1990–2017 1990–2017 1990–2017
Global 1716·1 1787·6 4·2% 1192·4 1233·1 3·4% 187·3 157·0 –16·2% 167·5 170·6 1·9%
(1519·6 to (1587·0 to (2·8 to (1051·2 to (1092·3 to (1·9 to (169·8 to (143·4 to (–17·4 to (152·1 to (155·6 to (0·8 to
1925·8) 1999·2) 5·6)* 1353·9) 1392·7) 4·9)* 206·7) 171·8) –14·9)* 184·8) 186·9) 2·9)*
Low SDI 2517·0 2501·1 –0·6% 1613·0 1611·3 –0·1% 276·8 237·5 –14·2% 184·3 185·1 0·4%
(2212·7 to (2200·5 to (–2·6 to 1·2) (1425·4 to (1424·7 to (–2·0 to 1·7) (250·4 to (215·5 to (–15·5 to (167·7 to (168·5 to (–1·0 to
2823·4) 2807·6) 1809·7) 1812·2) 305·1) 260·9) –12·9)* 202·4) 202·4) 1·9)
Low-middle SDI 1897·8 1867·1 –1·6% 1278·5 1273·8 –0·4% 217·3 182·7 –15·9% 164·5 165·4 0·6%
(1669·9 to (1648·8 to (–3·7 to 0·7) (1131·7 to (1126·3 to (–2·5 to 1·9) (196·8 to (166·6 to (–17·2 to (149·2 to (151·0 to (–0·8 to
2128·2) 2089·7) 1442·7) 1438·8) 239·9) 200·1) –14·5)* 180·5) 180·9) 2·0)
Middle SDI 1384·5 1350·7 –2·4% 1027·9 1001·6 –2·5% 172·4 128·7 –25·4% 150·3 152·4 1·4%
(1217·9 to (1191·8 to (–4·3 to (894·1 to (879·9 to (–4·8 to (155·2 to (116·8 to (–26·9 to (135·7 to (137·8 to (–0·1 to
1567·3) 1531·5) –0·8)* 1179·4) 1137·5) –0·8)* 191·5) 141·5) –23·9)* 166·6) 168·1) 2·7)
High-middle SDI 1207·1 1184·2 –1·9% 949·3 937·4 –1·3% 151·8 112·4 –25·9% 151·0 154·9 2·6%
(1052·7 to (1038·4 to (–4·2 to 0·5) (816·3 to (817·3 to (–3·8 to 1·3) (136·1 to (102·1 to (–27·9 to (135·3 to (139·3 to (0·7 to
1383·7) 1342·2) 1102·6) 1076·3) 169·9) 123·6) –24·1)* 169·0) 171·6) 4·2)*
High SDI 1235·3 1255·5 1·6% 995·4 1020·3 2·5% 156·0 135·0 –13·5% 193·3 196·4 1·6%
(1106·3 to (1129·6 to (–0·2 to 3·6) (878·7 to (910·6 to (0·5 to 4·6)* (144·0 to (125·1 to (–14·8 to (177·1 to (180·9 to (0·2 to
1381·0) 1393·2) 1122·7) 1135·9) 170·0) 145·4) –12·0)* 211·7) 212·6) 2·9)*
Central Europe, eastern Europe, and central Asia
Central Asia 1 582·5 1 542·4 –2·5% 1 215·2 1 180·1 –2·9% 226·8 188·7 –16·8% 187·1 187·2 0·1%
(1 387·5 to (1 361·2 to (–7·0 to 2·2) (1 057·5 to (1 026·7 to (–7·6 to 2·1) (202·5 to (168·9 to (–19·8 to (167·5 to (167·6 to (–3·4 to
1 805·6) 1 751·5) 1 392·5) 1 362·9) 252·7) 210·4) –14·0)* 207·1) 208·7) 3·5)
Central Europe 1 221·5 1 216·5 –0·4% 1 007·2 1 014·6 0·7% 148·7 118·1 –20·6% 177·3 184·2 3·9%
(1 053·9 to (1 096·0 to (–6·0 to 5·4) (861·0 to (904·7 to (–5·2 to 7·0) (133·0 to (108·8 to (–24·0 to (157·5 to (168·4 to (–0·4 to
1 404·5) 1 350·8) 1 167·5) 1 130·5) 165·1) 127·7) –16·6)* 198·3) 200·4) 8·6)
Eastern Europe 1 341·2 1 314·2 –2·0% 1 067·5 1 053·7 –1·3% 146·0 122·4 –16·2% 176·6 178·5 1·1%
(1 167·3 to (1 146·0 to (–4·6 to 0·3) (918·4 to (907·5 to (–4·0 to 1·2) (131·2 to (110·8 to (–17·7 to (157·8 to (158·8 to (–0·8 to
1 542·7) 1 513·7) 1 242·3) 1 220·8) 162·8) 136·1) –14·7)* 197·1) 200·0) 2·9)
High income
Australasia 974·9 1023·6 5·0% 784·7 824·8 5·1% 132·5 121·9 –7·9% 152·6 158·2 3·7%
(854·8 to (891·6 to (–3·1 to (680·8 to (715·0 to (–3·7 to (119·5 to (110·6 to (–12·1 to (136·9 to (142·2 to (–1·3 to
1127·8) 1165·2) 14·1) 910·5) 949·8) 14·7) 145·6) 133·8) –3·0)* 168·9) 174·6) 9·6)
High-income 2116·9 2066·2 –2·4% 1700·2 1690·3 –0·6% 224·2 170·3 –24·1% 290·4 294·0 1·2%
Asia Pacific (1853·4 to (1792·0 to (–5·5 to 1·1) (1461·8 to (1440·5 to (–4·0 to 3·1) (202·6 to (154·9 to (–25·7 to (259·3 to (261·0 to (–1·1 to
2425·1) 2378·9) 1966·5) 1954·9) 247·1) 186·5) –22·4)* 323·6) 326·5) 3·8)
High-income 1232·2 1229·0 –0·3% 995·7 995·8 0·0% 163·2 144·6 –11·4% 191·9 191·7 –0·1%
North America (1091·7 to (1093·0 to (–3·5 to 3·3) (872·4 to (877·9 to (–3·3 to 3·7) (148·9 to (133·2 to (–13·5 to (174·3 to (175·7 to (–2·3 to
1389·6) 1382·1) 1132·3) 1128·3) 179·3) 156·8) –9·2)* 211·4) 209·7) 2·4)
Southern Latin 1711·8 1595·1 –6·8% 1226·0 1165·5 –4·9% 203·8 161·6 –20·7% 199·5 187·5 –6·0%
America (1501·9 to (1401·4 to (–12·9 to (1080·5 to (1023·4 to (–11·2 to (185·7 to (146·9 to (–24·3 to (181·9 to (169·9 to (–10·4 to
1935·9) 1806·9) 0·0) 1385·3) 1328·5) 2·0) 223·9) 177·9) –17·1)* 219·1) 206·5) –1·6)*
Western Europe 963·9 1139·3 18·2% 773·9 920·9 19·0% 127·6 126·1 –1·1% 163·1 181·8 11·5%
(891·7 to (1055·8 to (15·8 to (713·3 to (853·8 to (16·5 to (119·8 to (118·5 to (–2·8 to 0·6) (153·3 to (170·7 to (9·6 to
1043·1) 1226·5) 20·6)* 841·9) 990·7) 21·6)* 136·3) 134·5) 174·0) 193·6) 13·5)*
Latin America and Caribbean
Andean Latin 1370·9 1326·8 –3·2% 964·0 936·9 –2·8% 187·5 152·0 –19·0% 147·5 146·5 –0·7%
America (1211·6 to (1166·8 to (–8·0 to 1·6) (849·6 to (831·4 to (–7·7 to 2·1) (171·8 to (139·9 to (–21·9 to (135·3 to (135·0 to (–4·2 to
1536·3) 1506·3) 1083·3) 1061·3) 204·7) 165·7) –16·0)* 160·9) 159·6) 2·8)
Caribbean 1445·5 1638·0 13·3% 1039·7 1138·8 9·5% 180·1 150·1 –16·6% 161·9 167·2 3·2%
(1271·8 to (1450·9 to (7·8 to (917·2 to (1007·0 to (4·4 to (163·3 to (137·6 to (–19·1 to (147·3 to (152·8 to (0·3 to
1638·0) 1854·8) 19·2)* 1180·8) 1299·6) 14·9)* 198·5) 164·2) –14·0)* 178·2) 183·1) 6·3)*
Central Latin 913·5 894·1 –2·1% 645·2 628·2 –2·6% 135·6 106·0 –21·9% 109·5 108·6 –0·8%
America (800·1 to (785·7 to (–4·4 to 0·2) (564·1 to (554·4 to (–5·0 to (123·6 to (97·4 to (–23·6 to (100·1 to (99·7 to (–2·5 to
1030·3) 1003·6) 735·9) 709·4) –0·2)* 149·7) 115·3) –20·2)* 119·9) 118·2) 0·8)
Tropical Latin 1077·2 983·0 –8·8% 752·2 688·5 –8·5% 140·7 105·6 –25·0% 123·6 122·8 –0·6%
America (939·7 to (868·7 to (–11·0 to (657·1 to (609·3 to (–10·8 to (127·1 to (96·9 to (–27·1 to (111·9 to (112·4 to (–2·8 to
1222·7) 1 103·0) –6·1)* 860·7) 783·1) –5·8)* 155·7) 114·7) –22·7)* 136·5) 134·0) 1·6)
(Table 2 continues on next page)

www.thelancet.com/child-adolescent Vol 4 March 2020 193

 Articles

Prevalence at birth per 100 000 babies Prevalence per 100 000 children younger Prevalence per 100 000 individuals Age-standardised prevalence rate
than 1 year (all ages) per 100 000
1990 2017 Percent 1990 2017 Percent 1990 2017 Percent 1990 2017 Percent
change, change, change, change,
1990–2017 1990–2017 1990–2017 1990–2017
(Continued from previous page)
North Africa and Middle East
North Africa and 1251·3 1303·1 4·1% 903·2 915·4 1·4% 175·7 141·1 –19·7% 132·1 135·0 2·2%
Middle East (1105·6 to (1159·3 to (1·0 to 7·1)* (794·1 to (805·9 to (–1·8 to 4·4) (157·8 to (127·6 to (–21·7 to (119·4 to (122·2 to (0·1 to
1414·0) 1464·2) 1032·7) 1045·6) 195·4) 155·9) –17·9)* 146·3) 149·2) 4·3)*
South Asia
South Asia 2047·8 1987·3 –3·0% 1299·6 1257·8 –3·2% 215·8 165·4 –23·4% 165·7 163·9 –1·1%
(1805·0 to (1751·0 to (–5·2 to (1152·0 to (1117·7 to (–5·2 to (195·6 to (151·2 to (–24·6 to (150·7 to (149·8 to (–2·5 to
2295·6) 2228·6) –0·6)* 1463·1) 1414·8) –1·1)* 237·5) 181·1) –22·0)* 181·3) 179·3) 0·4)
Southeast Asia, east Asia, and Oceania
East Asia 1266·9 1243·7 –1·8% 979·1 973·3 –0·6% 145·9 100·7 –31·0% 142·1 150·6 6·0%
(1088·2 to (1077·8 to (–5·5 to 1·2) (831·3 to (838·8 to (–4·6 to 2·8) (129·5 to (91·1 to (–33·7 to (125·9 to (134·9 to (2·7 to
1473·5) 1430·2) 1156·5) 1122·7) 166·3) 111·5) –28·7)* 161·8) 168·0) 8·8)*
Oceania 1905·4 2088·4 9·6% 1407·4 1521·0 8·1% 265·7 260·4 –2·0% 195·2 206·3 5·7%
(1664·4 to (1844·1 to (2·3 to (1228·5 to (1331·0 to (1·3 to (237·6 to (234·6 to (–6·6 to 3·2) (175·4 to (186·3 to (1·0 to
2168·9) 2380·3) 16·8)* 1622·1) 1737·7) 15·6)* 297·7) 290·4) 217·3) 229·2) 11·0)*
Southeast Asia 1714·5 1686·7 –1·6% 1296·8 1280·8 –1·2% 224·1 168·8 –24·7% 184·3 186·4 1·1%
(1504·1 to (1477·1 to (–4·3 to 1·3) (1127·2 to (1120·3 to (–4·1 to 1·9) (201·4 to (152·5 to (–26·3 to (165·9 to (168·0 to (–0·8 to
1944·3) 1917·9) 1475·9) 1465·5) 249·0) 186·3) –23·0)* 204·4) 206·4) 3·2)
Sub-Saharan Africa
Central sub- 2512·8 2659·8 5·8% 1708·1 1760·9 3·1% 307·3 284·0 –7·6% 193·9 197·2 1·7%
Saharan Africa (2205·7 to (2299·4 to (–0·0 to (1497·0 to (1529·4 to (–2·7 to 9·9) (276·3 to (255·7 to (–11·7 to (175·4 to (178·2 to (–2·5 to
2833·2) 3034·0) 12·9) 1924·2) 1994·2) 340·3) 314·3) –2·9)* 213·2) 217·8) 6·8)
Eastern sub- 2855·5 2669·5 –6·5% 1883·2 1788·0 –5·0% 338·2 291·7 –13·7% 208·6 202·2 –3·1%
Saharan Africa (2504·8 to (2350·6 to (–8·6 to (1655·3 to (1577·6 to (–7·0 to (304·7 to (263·1 to (–15·2 to (188·8 to (183·0 to (–4·6 to
3218·7) 3010·3) –4·4)* 2127·7) 2016·8) –2·9)* 373·4) 320·9) –12·2)* 229·4) 221·9) –1·5)*
Southern sub- 2005·5 2010·9 0·3% 1426·6 1429·2 0·2% 239·1 199·6 –16·5% 187·1 187·3 0·1%
Saharan Africa (1764·9 to (1764·8 to (–2·8 to 3·2) (1254·7 to (1257·8 to (–2·9 to 3·1) (215·5 to (180·3 to (–18·5 to (169·2 to (169·3 to (–2·1 to 2·3)
2249·6) 2257·8) 1615·7) 1619·2) 264·2) 220·4) –14·7)* 205·8) 206·7)
Western sub- 2405·2 2321·3 –3·5% 1641·9 1595·4 –2·8% 288·8 266·5 –7·7% 187·2 183·6 –1·9%
Saharan Africa (2105·2 to (2033·9 to (–7·1 to 0·4) (1441·1 to (1401·8 to (–6·7 to 1·4) (259·7 to (241·1 to (–10·8 to (169·2 to (166·9 to (–5·0 to
2721·1) 2619·2) 1866·0) 1807·5) 319·8) 294·2) –4·7)* 205·9) 201·5) 1·0)
Data are estimate (95% uncertainty interval). SDI=Socio-demographic Index. *Significant increase or decrease.

Table 2: Global Burden of Diseases, Injuries, and Risk Factors Study 2017 prevalence of congenital heart disease, 1990–2017

and India (appendix p 27). From 1990 to 2017, there common causes, but their epidemiology and aggregate
was no significant increase in infant YLD rates burden are quite different (appendix p 29). The
(appendix p 27). aggregate YLDs caused by rheumatic heart disease
Rates of congenital heart disease at birth varied by a were estimated to be substantially larger in 2017
factor of about two across all countries and territories than the number caused by congenital heart disease
in 2017 (Figure S9), with the highest rates observed in (1 900 974 YLDs [95% UI 1 232 815–2 765 992] for
western, central, and eastern sub-Saharan Africa, central rheumatic heart disease versus 589 479 [287 200–973 359]
and southeast Asia, and some internal provinces of for congenital heart disease) but the total number of
China and India, mirroring the YLD rates. The lowest years of life lost (YLLs) caused by congenital heart
rates for congenital heart disease at birth were disease was more than double that of rheumatic heart
concentrated in the Americas, western Europe, north disease (21 634 418 [17 779 611–25 604 823] for congenital
Africa, Australia, and some northern provinces of China. heart disease versus 7 492 586 [6 926 660–8 046 713] for
Mapping age-standardised prevalence revealed a different rheumatic heart disease). Although rheumatic heart
pattern, with many HICs, along with some countries disease often has its origins in childhood, most of both
within sub-Saharan Africa, having the highest age- the fatal and non-fatal burden of rheumatic heart
standardised prevalence rates. disease occurs in adults older than 20 years. By contrast,
When evaluating YLDs for all heart disease in the fatal and non-fatal burden of congenital heart
individuals younger than 20 years, congenital heart disease is highly concentrated in children, especially
disease and rheumatic heart disease are the most infants younger than 1 year.

194 www.thelancet.com/child-adolescent Vol 4 March 2020

 Articles

A
Single ventricle and single ventricle pathway heart defects Ventricular septal defect and atrial septal defect Other congenital cardiovascular defects
Malformations of great vessels, congenital valvular heart Complex congenital heart defects excluding
disease, and patent ductus arteriosis single ventricle and single ventricle pathway
Global Low SDI Low-middle SDI Middle SDI High-middle SDI High SDI
120
Prevalence per 100 000 individuals

96

72

48

24

B
100
Prevalence per 100 000 individuals

80

60

40

20

0
19 0
20 5
00

2007
10

17

90

20 5
00

20 7
10

17
90

20 5
00

2007
10

17

90

20 5
00

2007
10

17

19 0
20 5
00

2007
10

17

90

20 5
00

2007
10

17
0
9

9
9

9
20

20

20

20

20

20
20

20

20

20

20

20
19

19

19

19
19

19

19

19

19

19

Year Year Year Year Year Year

Figure 3: Trends in prevalence of congenital heart disease from 1990 to 2017

Data are prevalence (95% UI). (A) all-age prevalence and (B) age-standardised prevalence by Socio-demographic Index and congenital heart defect sub-category
with 95% uncertainty intervals. SDI=Socio-demographic Index.

Discussion accounts for the observed increase in the estimated

The global prevalence of congenital heart disease at proportion of all congenital heart disease cases in
birth, in 2017, is estimated to be nearly 1·8 cases per patients aged 15 years and older, which increased from
100 live births, a 4·2% (95% UI 2·8–5·6) increase nearly 23% in 1990, to over 28% in 2017. Nearly
since 1990. In comparison with previously determined 12·0 million people were estimated to be living with
prevalence rates, the birth prevalence is likely to be congenital heart disease in 2017 across all age groups,
higher because of the inclusion of data for low and an 18·7% increase from the estimated 10·1 million
middle SDI regions, which have not been previously in 1990. The total number of YLDs from congenital
accounted for, as well as because of biological and social heart disease in 2017 is estimated to be 589 479
factors that warrant additional investigation.5 There was (287 200–973 359), which is on par with attention
a decreasing trend in prevalence for the ventricular and deficit/hyperactivity disorder, peptic ulcer disease,
atrial septal defect subcategory, probably related to an lower respiratory infections, and haemo­ lytic disease
increased proportion of lesions being repaired or and other neonatal jaundice. Cardiac, pulmon­ ary,
having spontaneously regressed, in conjunction with nutritional, haemato­ logical, and gastro­ intestinal
an ageing population. When the population is complications all contribute to YLD, but there are
comprised of a greater proportion of older generations insufficient data available to quantify the effect of
than was the population in 1990, a larger proportion of specific causes of morbidity.8 There is also suggest­ion
ventricular and atrial septal defect lesions are either of increased developmental problems in some child­ren
repaired or have regressed. Ageing probably also with congenital heart disease—such as sensory

www.thelancet.com/child-adolescent Vol 4 March 2020 195

 Articles

impair­ments, intellectual disability, and learning dis­ surgery actually receiving it.28 Pooling resources to care
abilities—independent of the severity of the congenital for both of these diseases makes a stronger case for
heart disease lesion (appendix pp 23, 24).21 investment in surgical services than for either disease
The annual global mortality of congenital heart alone.
disease in 2017 is estimated to be 261 247 (95% UI The ability to accurately estimate both fatal and non-
216 567–308 159). The rate of decline in congenital heart fatal indicators for congenital heart disease has
disease prevalence markedly differs by SDI quintile in historically been, and remains, complicated for many
the first year of life. In the low SDI quintile, the reasons. Despite GBD 2017 mortality estimates being
prevalence of congenital heart disease mortality for similar to previous publications, designating congenital
infants has declined by only 6% since 1990, compared heart disease as the underlying cause of death risks
with declines of more than 50% in the middle, high- underestimating the full disease burden because of the
middle, and high SDI sub­ groups. Although the association between congenital heart disease and other
prevalence in middle and high-middle SDI quintiles has diseases, including genetic syndromes and pulmonary
decreased gradually over the past three decades, and cardiac compli­cations.8 The remission of specific
mortality is still more than double that of HICs. This types of congenital heart disease such as ventricular and
finding is consistent with published evidence showing atrial septal defects also complicates the estimation of
very low survival of patients with complex and critical congenital heart disease prevalence. As with many
congenital heart disease who are older than 1 year diseases, there are variations in data gathering and
and do not have access to advanced treatment and reporting strategies within regions, countries, and
diagnostics.24 Studies in HICs also show substantially hospitals, which include but are not limited to coding
higher long-term morbidity and mortality in patients systems, reporting of multiple congenital heart defects,
with all forms of congenital heart disease than in the and accounting for termination of pregnancies.
general population.25,26 Strickland and colleagues demonstrated that ICD-9
The relative importance of congenital heart disease as Clinical Modification coding can have relatively low
a cause of child mortality is rapidly increasing, as sensitivity and high false-positive fraction for some
evidenced by the increase in the proportion of deaths congenital heart diseases compared with clinical
due to congenital heart disease from 1990 to 2017, for all nomen­clature.29 There continues to be debate
but the high SDI quintile. Other common diseases, such surrounding the so-called unnatural history of patients
as respiratory disease and malnutrition, still rank as after uncomplicated repair of ventricular and atrial
leading causes of death in low SDI countries, but have septal defects in childhood, and whether or not these
fallen in rank in higher SDI countries. This observation individuals do or do not continue to be living with
is consistent with commonly seen shifts from com­ congenital heart disease.8,22 This question also applies to
municable to non-communicable diseases in countries lesions with no haemodynamic conse­quences, such as
with improving economies and public health systems. patent ductus arteriosus and some pulmonary stenosis,
In middle and high-middle SDI quintiles, mortality because they are less likely to come to attention and
from other common childhood killers has declined to a therefore remission rates of these defects is unknown.
greater degree than that from congenital heart disease, Although many patients with moderate or large
and therefore mortality from congenital heart disease is ventricular and atrial septal defects have substantial
accounting for a larger percentage of infant deaths. morbidity and mortality (if not repaired), most people
Although these trends would suggest a shift in invest­ born with small ventricular septal defects, partially
ments and priority toward congenital heart disease, it is closed atrial septal defects, or patent foramen ovales,
important to recognise that the cost and complexity of have no increased morbidity or mortality from this
treating congenital heart disease is much greater than diagnosis. Ventricular septal defects are the most
for many other childhood illnesses. common form of congenital heart disease and at least
In many countries, prevalence and mortality from 25% of all ventricular septal defects spontaneously
rheumatic heart disease parallel lack of access to cardiac close,30 with some studies reporting up to 20% closure
surgery services for congenital heart disease. Although per year. The actual rate of closure varies between
the age distributions of congenital heart disease and studies, in large part because of different method­
rheumatic heart disease mortality and YLDs differ, ology.17,18,31
many emerging cardiac surgery programmes in LMICs Previous studies have used varying techniques to
treat children with congenital heart disease as well as account for the aforementioned issues. The US Centers
children and young adults with rheumatic heart for Disease Control and Prevention reported higher
disease. A recent report provided country-level data for congenital heart disease prevalence than did GBD 2017
burden of rheumatic heart disease with an estimated for the USA, at 1·4 million adults and 1 million children
annual mortality of over 300 000 deaths.27 Insufficient in 2010,9 but these estimates were based on 2010 data
availability of cardiac surgery services in sub-Saharan from Quebec, Canada, that extrapolated the probability
Africa results in well under 3% of children in need of of survival of all congenital heart disease types to the

196 www.thelancet.com/child-adolescent Vol 4 March 2020

 Articles

entirety of births in the USA.32 This approach, in septal defect would not have been counted. All of this
addition to not using primary data from the USA, came together in making GBD 2017 the only effort to
assumes that every person born with any type of comprehensively estimate congenital heart disease
congenital heart disease, regardless of severity, prevalence and mortality globally across all ages and
continues to have disease throughout their life. over time, accounting for alcohol consumption in
Hoffman and colleagues generated global prevalence women of reproductive age by country, inclusion or
estimates of structural heart disease at birth of exclusions of genetic syndromes and pregnancy
1–1·2 cases per 1000 births, but excluded most terminations in reported data, and quantifying variable
ventricular and atrial septal defects.33 Liu and colleagues excess mortality and remission. This analysis
estimated a global birth prevalence of 941 cases per unsurprisingly generated lower predicted long-term
100 000 births.34 Despite not having any adjustments for survival, higher remission, and lower prevalence than
compositional bias, temporal changes in detection, or global or regional extra­polations based on studies from
definitional differences in input data, Liu and HICs,36,37 but also generated estimates that are empirical
colleagues’ estimates were also close to those of GBD in and poised to be highly actionable.
locations with abundant data (eg, western Europe, east The present analysis also has several limitations.
Asia, and USA). By contrast, large differences in global First, it shares all the limitations of the overall GBD
estimates emerged in Liu and colleagues’ study 2017 study, which includes little or no direct data on
compared with GBD because of the inclusion of two congenital heart disease outcomes from many
studies from sub-Saharan Africa that reported countries.11–14 As with all GBD modelling, the method­
congenital heart disease nearly an order of magnitude ology provided in the appendix (pp 12–14, 17–22)
lower than other studies and the use of a simple meta- addresses gaps in knowledge, highlighting what we
regression, in which all data for children younger than don’t know and providing uncertainty intervals for
6 years was considered as equivalent to birth data and each data point. Second, even after adjustment,
the only predictor was region and World Bank income potential underreporting might persist in admin­
category. The approach used in the Modell Global istrative datasets from locations with low diagnostic
Database on Birth Defects has been to assume a capacity and in older age groups. Third, and relatedly,
uniform baseline prevalence in the absence of although subtyping of congenital heart disease is
prevention and treatment efforts and to then adjust the expanded in our dataset, it has yet to capture the full
baseline on the basis of estimates of intervention spectrum of disease. This limitation is particularly true
coverage for selected groups of congenital malform­ for defects that have a wide variety of presentations,
ations.35 severity, and outcomes, such as ventricular septal
Our analysis has several notable advances compared defects, atrial septal defects, Ebstein’s anomaly, patent
with previous GBD estimates and other estimates of ductus arteriosus, and double outlet right ventricle.
congenital heart disease. First, we evolved the Within each of the five subtypes, there is a mix of
prevalence analysis to focus on specific anatomic diseases, especially in the group called critical
groupings of congenital heart disease rather than malformations of great vessels, congenital valvular
categorically classifying each case of congenital heart heart disease, and patent ductus arteriosus. Some
disease as moderate, severe, or critical, as was done for lesions remit spontaneously or are considered to be
GBD 2010, 2013, and 2015. Second, we incorporated functionally cured after surgical or catheter-based
many new sources of data. Before GBD 2016, survival intervention, whereas others can never remit spontan­
estimates were mechanistically tied to neonatal eously or by intervention. Although these issues could
mortality rates, which is unlikely to be accurate, theoretically lead to underestimates of prevalence after
especially in locations where neonatal survival has the neonatal period, this possibility is tempered by the
improved but no congenital heart disease screening or fact that the model results closely reflect most data
treatment exists. The estimates are now driven by the sources that show rapidly decreasing prevalence rates
combined power of data from screening programmes, with age. Fourth, also as described above, the GBD
congenital registries, admin­ istrative data, and data cause-specific mortality framework assigns each death
sources on mortality and survival. Third, we removed to a single underlying cause, which is unlikely to
the assumption that congenital heart disease never capture the full burden of congenital heart disease,
remits. Failing to account for these lesions in remission especially in older age groups and in high-mortality
modelling could result in substantial overestimation of settings where diagnostics and treatment are most
congenital heart disease prevalence, which probably limited. Fifth, we have not comprehensively quantified
occurred in other analyses. Fourth, we included all individuals with multiple congenital heart disease
individuals with congenital heart disease in GBD 2017 lesions, instead constraining the sum of each subtype
estimates. Previous GBD estimates excluded any to match the total. Sixth, cross-sectional data (at a given
individual with chromosomal or genetic syndromes, time point) cannot be extrapolated to make conclusions
meaning that a child with trisomy 21 and atrioventricular about longitudinal outcomes. For example, comparison

www.thelancet.com/child-adolescent Vol 4 March 2020 197

 Articles

of birth and all-age prevalence of congenital heart Robert C Reiner, Gregory A Roth, Mohamad-Hani Temsah,
disease in 2017 does not allow for direct quantification Marcos Roberto Tovani-Palone, Jeffrey A Towbin, Bach Xuan Tran,
Tung Thanh Tran, Nu Thi Truong, Theo Vos, Kia Vosoughi,
of the magnitude of death or remission of congenital Robert G Weintraub, Kidu Gidey Weldegwergs, Zoubida Zaidi,
heart disease over time. Those in older age groups Bistra Zheleva, Liesl J Zuhlke, Christopher J L Murray,
were born in years with lower prevalence rates of Gerard R Martin, Nicholas J Kassebaum
congenital heart disease at birth and lower total global Affiliations
population. Using cross-sectional data to predict death Division of Pediatric Cardiology, Dartmouth-Hitchcock Medical Center,
and remission over time would result in a substantial Lebanon, NH, USA (M S Zimmerman MD); Milken Institute School of
Public Health, George Washington University, Washington, DC, USA
overestimation in the decline of congenital heart (M S Zimmerman); Department of Cardiology, Children’s National
disease prevalence rates with age. Seventh, little data Health System, Washington, DC, USA (M S Zimmerman,
are available on some aspects of congenital heart Prof C A Sable MD, Prof G R Martin MD); School of Medicine, Emory
disease. The most notable data gaps include: the University, Atlanta, GA, USA (A G C Smith BA); Institute for Health
Metrics and Evaluation (M Marie Echko BS, L B Wilner MPH,
distribution of disabling sequelae (including develop­ Prof S I Hay FMedSci, I A Khalil MD, Prof A H Mokdad PhD,
mental disabilities) for which the only information we Prof M Naghavi MD, M R Nixon PhD, R C Reiner Jr PhD,
have identified is from high SDI settings; fetal, G A Roth MD, Prof T Vos PhD, Prof C J L Murray DPhil,
newborn, and population prevalence from many lower- N J Kassebaum MD), Department of Health Metrics Sciences, School
of Medicine (Prof S I Hay, I A Khalil, Prof A H Mokdad,
resource settings, including nearly all of sub-Saharan Prof M Naghavi, R C Reiner Jr, Prof T Vos, Prof C J L Murray,
Africa, Oceania, Southeast Asia, and parts of North N J Kassebaum), Division of Cardiology (G A Roth), and Department of
Africa and the Middle East; and cause-specific mortality Anesthesiology & Pain Medicine (N J Kassebaum), University of
Washington, Seattle, WA, USA; Department of Global Health, Bill &
data from sub-Saharan Africa, southeast Asia, and
Melinda Gates Foundation, Seattle, WA, USA (H E Olsen MA);
Oceania, ideally in conjunction with programmes to Department of Nursing, Aksum University, Aksum, Tigray, Ethiopia
ascertain prevalence in neonates and young children. (H T Atalay MSc); Indian Institute of Public Health, Gandhinagar,
Finally, this analysis did not include any explicit Gujarat, India (A Awasthi PhD); Public Health Foundation of India,
Gurugram, Haryana, India (A Awasthi, Prof R K Kumar MD); The
consideration of coverage of access to cardiac surgery,
Centre for Global Child Health, Hospital for Sick Children, University
catheter-based interventions, or supportive medical of Toronto, Toronto, Ontario, Canada (Z A Bhutta PhD); Center of
care, which could be helpful determinants of degree of Excellence in Women and Child Health, Aga Khan University, Karachi,
disability and long-term survival. Sindh, Pakistan (Z A Bhutta); Department of Administration,
Children’s HeartLink, Edina, MN, USA (J L Boucher MS); Department
The landscape of health—especially child health—is
of Research and Health Technology Assessment, Gorgas Memorial
changing across the world. Accurate and systematically Institute for Health Studies, Panama City, Panama (F Castro MD);
updated disease burden data are essential to track School of Medicine and Surgery, University of Milan Bicocca, Monza,
progress and encourage sustainable solutions for Italy (P A Cortesi PhD); United Nations World Food Programme, New
Delhi, India (M Dubey PhD); Department of Public Health Medicine,
providing high-quality treatment for all children in Bielefeld University, Bielefeld, Germany (F Fischer PhD); School of
need. GBD estimates of congenital heart disease burden Health and Environmental Studies, Hamdan Bin Mohammed Smart
provide a critically important advance in global University, Dubai, United Arab Emirates (S Hamidi DrPH); Center of
population health, helping provide a clearer roadmap Excellence in Behavioral Medicine, Nguyen Tat Thanh University, Ho
Chi Minh, Vietnam (C L Hoang BMedSc, L H Nguyen PhD,
for resource use and allocation to increase access to S H Nguyen BS, T H Nguyen MSc, N T Truong BHlthSci); Department
cardiovascular and NCD care, especially in LMICs of Pediatrics and Child Health (Prof C T Hugo-Hamman BEP,
where progress on congenital heart disease has been Prof L J Zuhlke PhD) and Department of Medicine (J J Noubiap MD,
largely stagnant. The UN has prioritised reduction of Prof L J Zuhlke), University of Cape Town, Cape Town, South Africa
(Prof C T Hugo-Hamman); Pediatric Cardiology, Windhoek, Namibia
premature deaths from NCDs (including congenital and (Prof C T Hugo-Hamman); Center for Applied Pediatric Quality
acquired heart disease), but to meet the SDG target of Analytics, Boston Children’s Hospital, Boston, MA, USA (Prof
“ending preventable deaths of newborns and children K J Jenkins MD); Department of Pediatrics, Harvard Medical School,
under 5 years of age”, health policy makers will need to Boston, MA, USA (Prof K J Jenkins); School of Health Sciences,
Savitribai Phule Pune University, Pune, India (A Kar PhD); Birth
develop specific accountability measures that address Defects and Childhood Disability Research Centre, Pune, India (A Kar);
barriers and improve access to paediatric cardiac Department of Pediatric Cardiology, Amrita Institute of Medical
diagnosis and treatment. Sciences, Ernakulam, India (Prof R K Kumar); School of Medicine,
Boston University, Boston, MA, USA (G F Kwan MD); Partners In
GBD 2017 Congenital Heart Disease Collaborators
Health, Boston, MA, USA (G F Kwan); School of Medicine
Meghan S Zimmerman*, Alison Grace Carswell Smith*,
(D T Mengistu MSc) and Clinical Pharmacy Unit
Craig A Sable, Michelle Marie Echko, Lauren B Wilner,
(K G Weldegwergs MSc), Mekelle University, Mekelle, Ethiopia;
Helen Elizabeth Olsen, Hagos Tasew Atalay, Ashish Awasthi,
Haramaya University, Harar, Ethiopia (L Negesa MSc); Emergency
Zulfiqar A Bhutta, Jackie LeeAnne Boucher, Franz Castro,
Hospital of Bucharest (I Negoi PhD) and Anatomy and Embryology
Paolo Angelo Cortesi, Manisha Dubey, Florian Fischer, Samer Hamidi,
Department (R I Negoi PhD), Carol Davila University of Medicine and
Simon I Hay, Chi Linh Hoang, Christopher Thomas Hugo-Hamman,
Pharmacy, Bucharest, Romania; Department of Cardiology, Cardio-Aid,
Kathy J Jenkins, Anita Kar, Ibrahim A Khalil, Raman Krishna Kumar,
Bucharest, Bucharest, Romania (R I Negoi); Institute for Global Health
Gene F Kwan, Desalegn Tadese Mengistu, Ali H Mokdad,
Innovations, Duy Tan University, Hanoi, Vietnam (C T Nguyen MPH,
Mohsen Naghavi, Lemma Negesa, Ionut Negoi, Ruxandra Irina Negoi,
H L T Nguyen MPH, T T Tran BMedSc); Department of Medicine,
Cuong Tat Nguyen, Huong Lan Thi Nguyen, Long Hoang Nguyen,
Maimonides Medical Center, Brooklyn, NY, USA (S Patel MD);
Son Hoang Nguyen, Trang Huyen Nguyen, Molly R Nixon,
Department of Social and Behavioral Sciences, New York University,
Jean Jacques Noubiap, Shanti Patel, Emmanuel K Peprah,

198 www.thelancet.com/child-adolescent Vol 4 March 2020

 Articles

New York, NY, USA (E K Peprah PhD); Department of Pediatrics, 11 GBD 2017 Mortality Collaborators. Global, regional, and national
King Saud University, Riyadh, Saudi Arabia (M-H Temsah MRCPCH); age-sex-specific mortality and life expectancy, 1950–2017:
College of Medicine, Alfaisal University, Riyadh, Riyadh, Saudi Arabia a systematic analysis for the Global Burden of Disease Study 2017.
(M-H Temsah); Department of Pathology and Legal Medicine, Lancet 2018; 392: 1684–735.
University of São Paulo, Ribeirão Preto, Brazil 12 GBD 2017 Disease and Injury Incidence and Prevalence
(M R Tovani-Palone MSc); Department of Pediatrics, University of Collaborators. Global, regional, and national incidence, prevalence,
Tennessee, Memphis, Tennessee, USA (Prof J A Towbin MD); and years lived with disability for 354 diseases and injuries for
195 countries and territories, 1990–2017: a systematic analysis for the
Department of Health Economics, Hanoi Medical University, Hanoi,
Global Burden of Disease Study 2017. Lancet 2018; 392: 1789–858.
Vietnam (B X Tran PhD); Department of Gastroenterology and
13 GBD 2017 Population and Fertility Collaborators. Population and
Hepatology, Johns Hopkins University, Baltimore, MD, USA
fertility by age and sex for 195 countries and territories, 1950–2017:
(K Vosoughi MD); Preventive Medicine and Public Health Research a systematic analysis for the Global Burden of Disease Study 2017.
Center, Iran University of Medical Sciences, Tehran, Iran (K Vosoughi); Lancet 2018; 392: 1995–2051.
Cardiology Department, Royal Children’s Hospital, Melbourne, VIC, 14 GBD 2017 Causes of Death Collaborators. Global, regional, and
Australia (R G Weintraub MB); Murdoch Childrens Research Institute, national age-sex-specific mortality for 282 causes of death in
Melbourne, VIC, Australia (R G Weintraub); ACS Medical College and 195 countries and territories, 1980–2017: a systematic analysis for
Hospital, Algiers, Algeria (Z Zaidi DrPH); and Children’s HeartLink, the Global Burden of Disease Study 2017. Lancet 2018; 392: 1736–88.
Edina, MN, USA (B Zheleva MBA). 15 Stevens GA, Alkema L, Black RE, et al. Guidelines for Accurate and
*Joint first authors. Transparent Health Estimates Reporting: the GATHER statement.
Lancet 2016; 388: e19–23.
Contributors
MSZ, AGCS, CAS, BZ, GRM, and NJK developed the concept and 16 Flaxman AD, Vos T, Murray C. Integrated Meta-Regression
Framework for Descriptive Epidemiology. Seattle, WA: University of
drafted the manuscript. MSZ, AS, CAS, MME, LBW, MRN, and NJK
Washington Press; 2012.
prepared the tables and figures and completed the scoping review.
17 Miyake T, Shinohara T, Nakamura Y, et al. Spontaneous closure of
AGCS, MME, LBW, MR, and NJK contributed to data preparation and
ventricular septal defects followed up from <3 months of age.
modelling. All other authors provided data, developed models, reviewed Pediatr Int 2004; 46: 135–40.
results, provided guidance on methodology, or reviewed the manuscript,
18 Zhang J, Ko JM, Guileyardo JM, Roberts WC. A review of
and approved the final version of the manuscript. spontaneous closure of ventricular septal defect.
Declaration of interests Proc Bayl Univ Med Cent 2015; 28: 516–20.
HEO was employed by Institute for Health Metrics and Evaluation and 19 Roth GA, Johnson C, Abajobir A, et al. Global, Regional, and
Bill & Melinda Gates Foundation during the study period. All other National Burden of Cardiovascular Diseases for 10 Causes, 1990 to
authors declare no competing interests. 2015. J Am Coll Cardiol 2017; 70: 1–25.
20 Gaynor JW, Stopp C, Wypij D, et al. Neurodevelopmental outcomes
Acknowledgments after cardiac surgery in infancy. Pediatrics 2015; 135: 816–25.
GFK is funded in part by a National Heart, Lung, and Blood Institute 21 Marino BS, Lipkin PH, Newburger JW, et al. Neurodevelopmental
grant (K23HL140133). Dr. Helen Olsen was employed by IHME and outcomes in children with congenital heart disease: evaluation and
BMGF during the study period. management: a scientific statement from the American Heart
Association. Circulation 2012; 126: 1143–72.
Editorial note: the Lancet Group takes a neutral position with respect to
22 Menting ME, Cuypers JA, Opić P, et al. The unnatural history of the
territorial claims in published maps and institutional affiliations.
ventricular septal defect: outcome up to 40 years after surgical
References closure. J Am Coll Cardiol 2015; 65: 1941–51.
1 Oster ME, Lee KA, Honein MA, Riehle-Colarusso T, Shin M, 23 Gabriel HM, Heger M, Innerhofer P, et al. Long-term outcome of
Correa A. Temporal trends in survival among infants with critical patients with ventricular septal defect considered not to require
congenital heart defects. Pediatrics 2013; 131: e1502–08. surgical closure during childhood. J Am Coll Cardiol 2002;
2 Raissadati A, Nieminen H, Haukka J, Sairanen H, Jokinen E. Late 39: 1066–71.
causes of death after pediatric cardiac surgery: a 60-year 24 Schidlow DN, Jenkins KJ, Gauvreau K, et al. Transposition of the
population-based study. J Am Coll Cardiol 2016; 68: 487–98. great arteries in the developing world: surgery and outcomes.
3 Moons P, Bovijn L, Budts W, Belmans A, Gewillig M. Temporal J Am Coll Cardiol 2017; 69: 43–51.
trends in survival to adulthood among patients born with congenital 25 Videbæk J, Laursen HB, Olsen M, Høfsten DE, Johnsen SP.
heart disease from 1970 to 1992 in Belgium. Circulation 2010; Long-term nationwide follow-up study of simple congenital heart
122: 2264–72. disease diagnosed in otherwise healthy children. Circulation 2016;
4 Mandalenakis Z, Rosengren A, Skoglund K, Lappas G, Eriksson P, 133: 474–83.
Dellborg M. Survivorship in children and young adults with 26 Spector LG, Menk JS, Knight JH, et al. Trends in long-term
congenital heart disease in Sweden. JAMA Intern Med 2017; mortality after congenital heart surgery. J Am Coll Cardiol 2018;
177: 224–30. 71: 2434–46.
5 van der Linde D, Konings EE, Slager MA, et al. Birth prevalence of 27 Watkins DA, Johnson CO, Colquhoun SM, et al. Global, regional,
congenital heart disease worldwide: a systematic review and and national burden of rheumatic heart disease, 1990–2015.
meta-analysis. J Am Coll Cardiol 2011; 58: 2241–47. N Engl J Med 2017; 377: 713–22.
6 Zheleva B, Atwood JB. The invisible child: childhood heart disease 28 Zilla P, Yacoub M, Zühlke L, et al. Global unmet needs in cardiac
in global health. Lancet 2017; 389: 16–18. surgery. Glob Heart 2018; 13: 293–303.
7 Tchervenkov CI, Jacobs JP, Bernier PL, et al. The improvement of 29 Strickland MJ, Riehle-Colarusso TJ, Jacobs JP, et al. The importance
care for paediatric and congenital cardiac disease across the of nomenclature for congenital cardiac disease: implications for
world: a challenge for the World Society for Pediatric and research and evaluation. Cardiol Young 2008; 18 (suppl 2): 92–100.
Congenital Heart Surgery. Cardiol Young 2008; 30 Cresti A, Giordano R, Koestenberger M, et al. Incidence and natural
18 (suppl 2): 63–69. history of neonatal isolated ventricular septal defects: do we know
8 Hoffman J. The natural and unnatural history of congenital heart everything? A 6-year single-center Italian experience follow-up.
disease. Chichester: Wiley-Blackwell, 2009. Congenit Heart Dis 2017.
9 Centers for Disease Control and Prevention. Congenital heart 31 Zhao QM, Niu C, Liu F, Wu L, Ma XJ, Huang GY. Spontaneous
defects. Nov 12, 2019. https://www.cdc.gov/ncbddd/heartdefects/ closure rates of ventricular septal defects (6750 consecutive
data.html (accessed Jan 10, 2020). neonates). Am J Cardiol 2019; 124: 613–17.
10 Wren C. The epidemiology of cardiovascular malformations. 32 Gilboa SM, Devine OJ, Kucik JE, et al. Congenital heart defects in
In: Moller JH, Hoffman JIE, eds. Pediatric Cardiovascular Medicine. the United States: estimating the magnitude of the affected
Chichester: Wiley-Blackwell, 2012. population in 2010. Circulation 2016; 134: 101–09.

www.thelancet.com/child-adolescent Vol 4 March 2020 199

 Articles

33 Hoffman JI. The global burden of congenital heart disease. 36 Marelli AJ, Ionescu-Ittu R, Mackie AS, Guo L, Dendukuri N,
Cardiovasc J Afr 2013; 24: 141–45. Kaouache M. Lifetime prevalence of congenital heart disease in the
34 Liu Y, Chen S, Zühlke L, et al. Global birth prevalence of congenital general population from 2000 to 2010. Circulation 2014; 130: 749–56.
heart defects 1970–2017: updated systematic review and 37 Marelli AJ, Mackie AS, Ionescu-Ittu R, Rahme E, Pilote L.
meta-analysis of 260 studies. Int J Epidemiol 2019; 48: 455–63. Congenital heart disease in the general population: changing
35 Modell B. DMW, Moorthie S, Blencowe H, Petrou M, Lawn J. prevalence and age distribution. Circulation 2007; 115: 163–72.
Epidemiological Methods in Community Genetics and the Modell
Global Database of Congenital Disorders. 2016. http://discovery.ucl.
ac.uk/1532179/ (accessed Jan 10, 2020).

200 www.thelancet.com/child-adolescent Vol 4 March 2020