Article Text

Download PDFPDF

Predicting survival in adults with congenital heart disease: what are the odds?
  1. Jonathan Buber1,
  2. Anne Marie Valente2,3,4
  1. 1 Division of Cardiology, Department of Medicine, University of Washington School of Medicine, Seattle, Washington, USA
  2. 2 Department of Cardiology, Boston Children’s Hospital, Boston, Massachusetts, USA
  3. 3 Division of Cardiology, Department of Internal Medicine, Brigham & Women’s Hospital, Boston, Massachusetts, USA
  4. 4 Faculty of Internal Medicine and Pediatrics, Harvard Medical School, Boston, Massachusetts, USA
  1. Correspondence to Dr Anne Marie Valente, Department of Cardiology, Boston Children’s Hospital, Boston, MA 02115, USA; anne.valente{at}

Statistics from

Request Permissions

If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.

The constantly growing rates of survival to adulthood among individuals born with congenital heart disease (CHD) is one of modern medicine’s greatest triumphs. Over the past several years, substantially greater survival rates of patients with CHD have been reported by groups from all around the world, and it is currently expected that over 90% of those born with CHD will reach adulthood.

Thus, it appears that the field of CHD gradually joins other fields in Cardiology, such as advanced heart failure and valvular heart disease, in which vast advancement in treatment options and patient survival have been achieved. Similar to those fields, the focus of CHD-related epidemiological research has been progressively turning towards more accurate evaluation of survival trajectories and identification of populations of patients who may be at a greater risk of mortality and morbidity burdens.

However, even the most meticulously performed survival studies, which typically require large cohorts of patients, numerous years of follow-up and rigorous analysis, may bear paramount shortcomings that limit their applicability to a wide spectrum of patients. Designing and executing large survival analyses that focus on patients with CHD can prove to be an even more challenging task, given the heterogeneity of this population. Close scrutiny of the results of several of the larger CHD-based survival analyses published over the past several years, which are also summarised in  table 1  highlight some of the important limitations of these manuscripts and the difficulty of undertaking such investigations.1–7 Some of these limitations are shared with other population-based survival studies, yet most are distinctive for the patient population with CHD.

Table 1

Causes of death

First and foremost, the vast majority of CHD survival analyses include highly heterogeneous groups of patients with different cardiac defects of varying complexity. Although survival data of patients with specific defects, such as tetralogy of Fallot, are now readily available, similar data on other, less well-studied conditions still lack or are limited by small number of patients. Second, given the relative novelty of many of the surgical and catheter-based therapies for various cardiac defects and the difference in the time of its implementation in different parts of the world, treatment for individual conditions is often not standardised in these studies. Furthermore, medical therapy is almost always based on individual physician-based decision-making, given the near complete absence of evidence-based data. Third, the mode of death of patients with CHD may be difficult to determine. Lack of adequate follow-up by a CHD specialist, relatively low rates of cardioverter defibrillators implantations and lost to follow-up for long durations are just several reasons for the inability to delineate the causes of death. Definitions of the modes of death are not well standardised as well and may differ between studies, with investigators classifying the same mode of death into diverse categories (eg, death from aortic dissection was categorised as death due to cardiac cause, vascular cause or sudden cardiac death, in three different studies). Fourth, the number of patients who get lost to follow-up is generally high as noted above and finally, the type of statistical analysis employed and the methods of patient inclusion (ie, historical cohorts, registries from multiple centres and so on) considerably differ between the studies.

Considering these important alterations between the studies and the wide variations in the number of patients and follow-up durations, one cannot be overtly surprised by the wide spectrum of death rates reported in these papers, ranging from 3.3% to 16% (table 1). Understanding the challenges of planning, performing and waiting for the results of large-scale, prospective studies on this topic, it is imperative to continue and report on the results of CHD survival from large, contemporary registries and retrospective cohorts in order to provide further understanding and plan ahead.

Supplemental material

In their Heart paper, Yu et al 8 provide such much-needed information by reporting on the survival rates within a population of 3068 Australian patients with CHD followed between 2000 and 2015. The distribution of the disease complexity in this manuscript is well in accordance with prior similar reports, and the mortality rates reported by the authors are relatively high, yet comparable to prior studies (table 1). The main findings of this manuscript, which are also highlighted by the authors, shed more light on the trends of mortality and survival among patients with CHD cared for in a contemporary era in a tertiary, dedicated CHD clinic. First, the authors report that the leading cause of death in their population was non-cardiac, a finding which was more prominent in patients with simple forms of CHD. Second, the median age at the time of death was still low—34 years—among patients with complex lesions and their mortality rate was over three times higher than among those with simple lesions. Third, in the same complex-lesion patient population, the most common cause of death of those patients who survived beyond 50 years. These results underscore the enormous success maintained over the past several decades of treating patients with simple cardiac lesions such as bicuspid aortic valves or secundum atrial septal defects, while at the same time demonstrate the road that still needs to be travelled in the caring and in improving the therapeutic options for patients with complex lesions.

With the continuous growth of the population of adults with CHD throughout the world, survival reports will likely continue to be published. Given the heterogeneity of these reports thus far, a need of standardisation of the results is of the utmost importance. Such standardisation in the modes of death, for instance, will allow more uniform interpretations of the outcomes. Furthermore, greater granularity regarding the different types of CHD is needed. The Bethesda classification has been used for almost two decades, yet this designation focuses primarily on the primary anatomical diagnosis to determine the severity of the condition and does not take into account the specific interventions or physiological sequelae that may be present. Redefining the manner in which we consider levels of risk in ACHD patients focused on more than their birth congenital diagnosis and considers additional factors, such as genomics, biomarkers and functional status will most likely result in better planning of healthcare system utilisation and targeted innovations in interventional and medical therapies for the populations who remain at the greatest risk of early mortality and morbidity.



  • Contributors JB designed this work and drafted the manuscript. AMV contributed to the design of the manuscript and critically revised it for intellectual content. Both authors have given final approval of the version submitted.

  • Funding The authors have not declared a specific grant for this research from any funding agency in the public, commercial or not-for-profit sectors.

  • Competing interests None declared.

  • Patient consent Not required.

  • Provenance and peer review Commissioned; internally peer reviewed.

Linked Articles