Objective To determine factors associated with coronary artery disease (CAD) and ischaemic stroke in ageing adult congenital heart disease (ACHD) patients.
Methods We performed a multicentre case–control study, using data from the national CONgenital CORvitia (CONCOR) registry to identify ACHD patients within five participating centres. Patients with CAD were matched (1:2 ratio) with ACHD patients without CAD on age, CHD defect group and gender. Patients with ischaemic stroke (or transient ischaemic attack) were matched similarly. Medical charts were reviewed and a standardised questionnaire was used to determine presence of risk factors.
Results Of 6904 ACHD patients, a total of 55 cases with CAD (80% male, mean age 55.1±12.4 years) and 56 cases with stroke (46% male, mean age 46.9±15.2) were included and matched with control patients. In multivariable logistic regression analysis, traditional atherosclerotic risk factors (hypertension (OR 2.45; 95% CI 1.15 to 5.23), hypercholesterolaemia (OR 3.99; 95% CI 1.62 to 9.83) and smoking (OR 2.25; 95% CI 1.09 to 4.66)) were associated with CAD. In contrast, these risk factors were not associated with ischaemic stroke. In multivariable analysis, stroke was associated with previous shunt operations (OR 4.20; 95% CI 1.36 to 12.9), residual/unclosed septal defects (OR 2.38; 95% CI 1.03 to 5.51) and left-sided mechanical valves (OR 2.67; 95% CI 1.09 to 6.50).
Conclusions Traditional atherosclerotic risk factors were associated with CAD in ACHD patients. In contrast, ischaemic stroke was related to factors (previous shunts, septal defects, mechanical valves) suggesting a cardioembolic aetiology. These findings may inform surveillance and prevention strategies.
- Congenital Heart Disease
- Coronary Artery Disease
Statistics from Altmetric.com
The prevalence of congenital heart disease (CHD) in the adult population is approximately 3–4 per thousand and still rising.1 Mortality rates in CHD have shifted from children towards adults, with a steadily increasing life expectancy.2 Due to these improvements, most patients with CHD live long enough to develop acquired cardiovascular disease such as coronary artery disease (CAD) and stroke.
The number of adult congenital heart disease (ACHD) patients with documented atherosclerotic risk factors is substantial.3 4 Specifically, patients with CHD seem to be at risk for hypertension and diabetes.5–7 Myocardial infarction is now the leading cause of death in adults with simple CHD.3 8 Furthermore, the risk of ischaemic stroke or transient ischaemic attack (TIA) is approximately 10-fold higher in young ACHD patientscompared with the general population.9 10
Despite the increasing risks, little is known about the relation of specific risk factors with CAD and stroke.4 5 10 This information could inform surveillance and prevention strategies. Most previous studies were performed using large population-based registries, which lacked specific data on potential risk factors.11 Our objective was to study the relation of both atherosclerotic and CHD-related risk factors with CAD and ischaemic stroke/TIA in ACHD patients.
We performed a multicentre case–control study, using data from the prospective Dutch nationwide CONgenital CORvitia (CONCOR)12 registry to identify ACHD patients included within five participating centres before March 2016. From 2001 onwards, patients were included in the registry by dedicated research nurses. The CONCOR registry was approved by the ethics boards of all five participating centres and all patients provided informed consent.
Gender, date of birth and type of congenital defects were retrieved from the CONCOR registry. All medical charts of patients with a diagnostic European Paediatric Cardiac coding code for CAD after 2001: myocardial infarction (acute coronary syndrome; non-ST elevation myocardial infarction (NSTEMI),13 ST elevation myocardial infarction (STEMI) and (unstable) angina pectoris), percutaneous coronary intervention (PCI) or coronary artery bypass grafting (CABG) were reviewed. Patients with all types of CHD were included and were classified into nine subgroups according to a combination of the complexity (mild, moderate, severe according to Bethesda) and haemodynamic consequences of the defect (table 1). This classification of the patients with CHD was performed to allow more accurate matching of case and control patients. Patients were excluded if the cause of event was iatrogenic or periprocedural (within 2 weeks after procedure). Other exclusion criteria were absence of at least one complete medical letter and exclusion of CAD as cause of myocardial infarction by cardiac catheterisation. Afterwards, for each patient with CHD with verified CAD, two control ACHD without the outcome parameters were selected, matching on date of birth, CHD defect subgroup and gender. Control patients had to be alive at the reference date (event date-matched case) and all patients with previous CAD or ischaemic stroke (before 2001) were excluded. Similarly, patients with ischaemic cerebrovascular accident (iCVA) or TIA and control patients were selected from CONCOR. Patients were also excluded if the cause of CVA was haemorrhagic.
We collected patient characteristics and surgical data from case and control patients at the time of the primary event (event date of matched case) by detailed review of hospital medical charts. Additionally, the presence and number of atherosclerotic-related and CHD-related risk factors at the time of the reference date were retrieved from medical charts. When not specified in the primary event discharge status, risk factors were collected from the last outpatient visit. Finally, a standardised questionnaire was used to collect information on cardiovascular risk factors and lifestyle, as were present prior to the primary event date. In patients who did not respond to the questionnaire, we retrieved information on risk factors solely from the hospital medical charts.
Atherosclerotic risk factors collected included: (1) hypertension, defined as systolic blood pressure >140 mm Hg at outpatient visit closest to time of inclusion or when previously diagnosed/treated and noted in patient history; (2) hypercholesterolaemia, defined as total cholesterol ≥7 mmol/L in laboratory testing closest to outpatient visit before onset of event or when previously diagnosed/treated; (3) diabetes mellitus, insulin dependent or non-dependent, according to definitions in the European guidelines14; (4) smoking, either former smokers or current smokers; (5) family history of premature (<65 years) non-CHD-related cardiovascular disease, in any first-degree relatives, when noted in medical charts or reported by patients. The total number of risk factors was pooled (0–5) as a measure of the overall atherosclerotic risk profile.
CHD-related risk factors collected included: (1) prior palliative shunt procedure (eg, Blalock-Taussig shunt); (2) presence of left-sided mechanical valve (eg, aortic or mitral position); (3) residual septal defect(s) on echocardiography before reference date (eg,Atrial septal defect(ASD), Ventricular septal defect (VSD) or Atrioventricular septal defect (AVSD)).
For this study, we defined two outcome variables: (1) CAD, either acute coronary syndrome (NSTEMI, STEMI and (unstable) angina pectoris), PCI or CABG15; (2) ischaemic cerebrovascular disease, either ischaemic CVA (ischaemic stroke with acute infarction) or TIA (brief episode of neurological dysfunction caused by temporary cerebral ischaemia without infarction).16
We used SPSS V.23.0 (IBM, Armonk, New York, USA) and R 3.3.1 (R core team) for statistical analysis. Continuous data were described as mean with SD or median with IQR, as appropriate, categorical variables as number with percentage. Differences on baseline variables between cases and controls were assessed using independent samples Student’s t-test, χ2 test or Wilcoxon rank-sum test, as appropriate. The association between baseline variables and outcomes was analysed using univariate logistic regression analysis and expressed as OR. Selected variables with p<0.10 in univariate analysis or when considered a specific risk factor were included in multivariable logistic regression analysis. We did not include strongly correlated variables in multivariable analysis to restrict the number of factors in the analysis, given the limited number of cases. In multivariable analysis, we corrected for age, gender and CHD severity (mild, moderate or severe) to adjust for possible confounding by the matching variables.17 A two-sided p value <0.05 was considered statistically significant.
Of 6904 ACHD patients included in CONCOR within five centres, a total of 55 patients with CAD (80% male, mean age 55.1±12.4 years) and 56 patients with ischaemic stroke/TIA (46% male, mean age 46.9±15.2 years) were included in this study and matched with two control patients (figures 1 and 2). The characteristics of the study population at time of the primary event are shown in tables 1 and 2. Simple shunt lesions were most prevalent in both the CAD and stroke groups. One patient had both CAD and an ischaemic CVA and was included in both stroke and CAD groups, matched with the same two control patients. A total of 20 alive patients of the CAD case–control group and 22 patients of the stroke case–control group did not respond to our call to complete the questionnaire (figures 1, 2).
Coronary artery disease
Of 55 patients with CAD, 10 (18%) patients presented with STEMI, 16 (29%) with NSTEMI, 2 (4%) with out-of-hospital cardiac arrest due to myocardial infarction, 9 (16%) with (unstable) angina pectoris and 14 (25%) did not have symptoms but underwent revascularisation due to significant CAD revealed during (preoperative) diagnostic coronary angiography. A total of 7 (13%) patients underwent CABG and 38 (69%) underwent PCI. Factors associated with CAD in univariate analysis are shown in figure 3 and listed in table 2. Among patients with CAD, 53% had hypertension, 35% had hypercholesterolaemia and 58% were former or current smokers, as compared with 28%, 10% and 45% of control patients, respectively. The added total number (0–5) of traditional atherosclerotic risk factors (smoking, diabetes, hypertension, hypercholesterolaemia, family history) was associated with CAD (OR 1.87 per risk factor; 95% CI 1.36 to 2.57, p<0.001) (figure 3). In multivariable analysis, the presence of hypertension (OR 2.45; 95% CI 1.15 to 5.23, p=0.02), current or former smoking (OR 2.25; 95% CI 1.09 to 4.66, p=0.03) and hypercholesterolaemia (OR 3.99; 95% CI 1.62 to 9.83, p=0.003) remained significantly associated with CAD.
Of 56 patients with ischaemic stroke, 26 (47%) patients had iCVA and 30 (53%) patients had TIA. Factors associated with ischaemic stroke in univariable analysis are listed in table 2 and shown in figure 4. None of the traditional atherosclerotic risk factors (smoking, diabetes, hypertension, hypercholesterolaemia, family history) nor the added total number of risk factors (OR 1.07 per risk factor; 95% CI 0.80 to 1.41) was associated with stroke (figure 4). The number of patients with known supraventricular tachycardia prior to the event date was similar in the stroke (23%) group and control group (23%). Three potential specific risk factors (prior palliative procedure, left-sided mechanical valve and residual septal defect) were (not all significantly) associated with ischaemic stroke in univariate analysis (figure 4). A total of 25% of patients with stroke had a residual septal defect on echocardiography and 23% had a left-sided mechanical valve, as compared with 15% and 13% in the control group, respectively. In multivariable analysis: previous shunt procedure (OR 4.20; 95% CI 1.36 to 12.9, p=0.01), left-sided mechanical valve (OR 2.67; 95% CI 1.09 to 6.50, p=0.03) and residual septal defect (OR 2.38; 95% CI 1.03 to 5.51, p=0.04) were significantly associated with ischaemic stroke.
To our knowledge, this was the first large case–control study, specifically designed to identify ACHD-specific risk factors for acquired cardiovascular disease (CAD and stroke). After carefully reviewing medical charts and using standardised questionnaires, we found that traditional atherosclerotic risk factors such as hypertension, hypercholesterolaemia and smoking were associated with CAD. In contrast, ischaemic stroke was related to factors (previous shunts, septal defects, mechanical valves) suggesting a cardioembolic aetiology.
In our study, multivariable OR of smoking, hypertension and dyslipidaemia for CAD were roughly comparable with the general population.18 A previous study also showed that systemic arterial hypertension and hyperlipidaemia were associated with incident CAD found during coronary angiography in ACHD patients without suspected CAD.19 Previous studies suggested ACHD patients have a healthy lifestyle with more sports participation and less smoking compared with the general population.4 20 Nonetheless, other epidemiological studies revealed that hypertension and diabetes were more prevalent in ACHD patients, especially those with greater complexity.4 5 7 This is remarkable as experimental studies found effective compensatory mechanisms for cyanosis in the coronary circulation, including dilated and tortuous coronary arteries with increased flow.21 22 However, increased medial collagen and smooth muscle loss were also found in histology of coronary arteries of young cyanotic patients.22 Future (experimental) studies are needed to determine the late effects of cyanosis on the vascular wall and whether compensatory mechanisms may contribute to increased risks for hypertension, renal dysfunction and diabetes, despite healthier lifestyle.5 7 23 Considering our results, more effective lifestyle counselling and early treatment of hypertension and diabetes might reduce the risk of CAD similar to the general population. However, this could be particularly important in patients at higher risk such as those with (previous) cyanotic CHD or aortic coarctation.24
We identified both general atherosclerotic risk factors and CHD-related factors in our study population as aetiology of stroke may vary in ACHD patients. Remarkably, there was no relation between atherosclerotic risk factors and ischaemic stroke in our study. This is in contrast to the general population and may also explain the relatively young age of ACHD patients with ischaemic stroke compared with CAD. A previous study also showed that the presence of left-sided mechanical prosthetic valves was associated with ischaemic stroke in ACHD patients.25 The association of previous palliative shunt procedures with stroke may be explained by cyanosis, which was related with stroke in previous studies.25 26 After reviewing medical charts for case and control patients, we found similar prevalence (~23%) of supraventricular tachycardia (SVT) in both groups. In another study, absence of sinus rhythm was associated with stroke. However, in this study, detailed review of medical charts was only performed in patients with stroke, which could lead to underestimation in control patients.25 Subclinical SVTs could have played a role in some patients.
Our study identified risk factors by reviewing medical charts, imaging results and using a standardised questionnaire, providing specific information on the atherosclerotic risk profile and CHD-specific risk factors. Previous studies were population-based registries and provided valuable knowledge on prevalence and incidence, but were less specific regarding risk factors.9–11 Population-based registries often rely on diagnostic codes which can be misclassified or underclassified. In such registries, medical charts cannot usually be reviewed to validate and further specify diagnoses and/or risk factors in the entire population. Most importantly, there has been poor documentation of cardiovascular risk factors, even in hospital records.11 Our study design provided more accurate information on atherosclerotic risk profile as standardised questionnaires were also used.
This retrospective case–control study was designed to determine specific risk factors for CAD and ischaemic stroke but not the incidence of these events. Our matching strategy resulted in comparable case and control groups, but we did not aim to determine the relation of age, gender and CHD type with outcomes, as these were matching variables. There was limited statistical power to detect the association of outcomes with low prevalent potential risk factors such as diabetes, endocarditis, heart failure, cyanosis and devices as statistical power is dependent on prevalence of exposure variables in case–control studies.27
About one-quarter of all patients either died or did not respond to the questionnaire. Alternatively, risk factors were retrieved from hospital charts in this subgroup, which could lead to an underestimation in deceased patients. In addition, cholesterol blood measurements were not performed routinely in our population, reflecting general practice by ACHD cardiologists. We included a combination of smoking or former smoking as risk variable in this retrospective study, although smoking cessation years ago may be less relevant. Our study was limited to patients that survived until inclusion in CONCOR. Although CONCOR is designed to include all types of ACHD, a subgroup of presumably relatively healthy patients who were lost to follow-up or were not receiving specialised care were not studied.28
This case–control study of ACHD patients revealed that traditional atherosclerotic risk factors such as hypertension, hypercholesterolaemia and smoking were associated with CAD. In contrast, ischaemic stroke was not associated with atherosclerotic risk factors but with previous shunt procedures, septal defects on echocardiography and left-sided mechanical valves suggesting an cardioembolic aetiology. Whether modification of these various risk factors prevents acquired cardiovascular disease should be the focus of future studies.
What is already known on this subject?
Ageing adults with congenital heart disease (ACHD) are at risk for coronary artery disease and stroke. The presence of atherosclerotic risk factors in ACHD is substantial.
What might this study add?
This study revealed an association of common atherosclerotic risk factors with coronary artery disease, as is seen in the general population. In contrast, ischaemic stroke was related to factors suggesting a cardioembolic aetiology.
How might this impact on clinical practice?
Physicians should start counselling on heart healthy lifestyle early on, considering its association with coronary artery disease. In addition, they should be aware of the presence of risk factors such as hypertension, and treat according to general guidelines. Future research is needed to determine if more stringent risk factor controls are needed for ACHD.
The authors thank Lia Engelfriet and Sylvia Mantels for their dedicated work on the CONCOR registry.
Contributors All authors attributed in both the conception, design, critical revision and final approval of this manuscript. JPB and IZ analysed and interpreted the data and drafted the manuscript under supervision of senior authors BJMM and BJB.
Funding This work was supported by the Netherlands Heart Institute (NL-HI) and Nuts Ohra foundation. The work described in this study was carried out in the context of the Parelsnoer Institute. Parelsnoer Institute is part of and funded by the Dutch Federation of University Medical Centers.
Competing interests None declared.
Ethics approval AMC Medical Ethics Committee.
Provenance and peer review Not commissioned; externally peer reviewed.
Data sharing statement Unpublished data from the study are available on request.
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.