Congenital Heart Disease

DOI: 10.32604/CHD.2021.014384

ARTICLE

Hypertension and Heart Failure as Predictors of Mortality in an Adult Congenital Heart Defect Population

Cheryl Raskind-Hood1,*, Kashaine A. Gray2,3, Jayne Morgan3 and Wendy M. Book4,*

1Department of Epidemiology, Emory School of Public Health, Atlanta, Georgia, USA
2Department of Public Health, Mercer University College of Health Professions, Atlanta, Georgia, USA
3Division of Cardiovascular Research, Piedmont Hospital, Atlanta, Georgia, USA
4Division of Cardiology, Emory University School of Medicine, Atlanta, Georgia, USA
*Corresponding Authors: Cheryl Raskind-Hood. Email: craskin@emory.edu; Wendy M. Book. Email: wbook@emory.edu
Received: 23 September 2020; Accepted: 05 January 2021

1  Introduction

Congenital heart defects (CHDs) account for ~1% of live births [1] ranging in complexity, often requiring interventions in childhood [2]. In children, increasing anatomic complexity is associated with worse survival [3]. As children with CHD age into adulthood, they face comorbid conditions including cardiovascular complications such as hypertension (HTN), heart failure (HF), and myocardial infarction (MI), and have higher mortality with advancing age compared to younger patients with CHD [1,4]. Previous studies of CHD have identified heart failure as an important contributor to mortality and poor long-term outcomes in an adult CHD population [5–9]. HTN is a significant contributor to the development of heart failure [10,11] in the general population, but once heart failure develops, higher blood pressure is associated with better outcomes than low blood pressure [12–15]. The interplay of HTN and HF with race in the setting of pre-existing CHD in an adult population is unknown. Race may be associated with the increased development of HTN and HF in the general population. The Atherosclerotic Risk in Communities (ARIC) study, which examined HF across race and gender in the general population, noted a discrepancy with cardiovascular risk factors disproportionately affecting blacks more than other races [11]. The evaluation of trends in CHD mortality by race and gender has also shown disparities affecting blacks more than other races, as well as males more than females [8,16]. HF is an increasing cause of hospitalization among adults with CHD (ACHD) in the US, and ACHD patients aged 18–65 years account for a disproportionate number of hospitalizations compared to adults without CHD in the same age range [17]. While anatomic complexity of CHD is a significant contributor to mortality in children [3], HTN and HF are common cardiovascular conditions in adults in the US and often also accompany CHD in adults. While development of HF may be associated with anatomic complexity, and either HF or anatomic complexity may be associated with worse outcomes, it is not known if HF predicts mortality regardless of anatomic complexity. For this reason, the evaluation of whether HF with or without HTN are significant predictors of mortality in the context of age, race, gender and CHD anatomic complexity is imperative. Our aim is to evaluate the impact of HTN and HF, and accompanying covariates as predictors of mortality in ACHD in the context of underlying anatomic complexity. Understanding the contribution of these comorbid conditions to mortality in ACHD may help guide long-term management.

2  Materials and Methods

A de-identified analytic file, created from a linked, de-duplicated CHD repository of Georgia residents developed as part of a pilot CHD surveillance project with the Centers for Disease Control and Prevention (CDC) and Emory University, was used. Methods for that project have previously been described [18] and were approved by the Emory University Institutional Review Board (#IRB0000064051). This cross-sectional retrospective study identified 9,394 adult Georgia residents with CHD, ages 18–64 years as of January 1, 2010, who sought medical services at least once from at least one of seven Georgia healthcare facilities (Appendix A) between January 1, 2008 and December 31, 2010. All adults with CHD were diagnosed with at least one of 55 CHD-related ICD-9-CM diagnostic codes (Appendix B) and categorized by CHD anatomy [18] using a five-level classification scheme [19] that was further refined by a multi-site group of congenital heart disease experts, as a component of a larger parent project and adapted for anatomic complexity. The classification scheme included: (1) Severely complex; (2) Shunts; (3) Valves; (4) Shunts plus valves; and (5) Other CHD anomalies (Appendix B) [18]. However, due to lack of specificity and poor diagnostic accuracy determined through chart abstraction and code validation, code 745.5 used for both atrial septal defect (ASD) (a true CHD) and patent foramen ovale (PFO) (a normal condition), 1,491 patients diagnosed with 745.5 in isolation or in combination with either 746.89 or 746.9 (both defined as unspecified congenital anomaly of the heart) were excluded from the analytic cohort, and another 2,506 cases with an ‘Other’ CHD code were also excluded [20] leaving an analytic dataset with N = 5, 397. The main outcome, mortality, confirmed from Georgia Department of Public Health (GA DPH) Death Certificates for years 2008–2010 served as the dependent variable. The primary comorbid conditions, HTN and HF, were derived from the presence of at least one of 40 HTN-related (Appendix C) or one of 16 HF-related ICD-9-CM codes (Appendix D) appearing on at least one healthcare encounter during the study period; number and percentage of ACHD patients with each of these 40 HTN and 16 HF codes are reported as part of Appendices C and D, respectively. Comorbid groupings of HTN/HF included those having at least one HTN ICD-9-CM code without having any HF ICD-9-CM codes, those having at least one HF ICD-9-CM code without any HTN codes, or those having a combination of both HTN and HF ICD-9-CM codes. We assessed the predictors of age, gender and race because they are established risk factors for CHD. Age was calculated by subtracting the patient’s date of birth from 01/01/2008 and was then classified into four age groups approximating age decades as close as possible: 18–29 years, 30–39 years, 40–49 years, and 50–64 years of age. As noted, two of age intervals include patients whose ages fell outside of the age-specific 10 year or decade intervals. As such, we placed the 18 and 19 year olds with the 20–29 year old decade, and patients who were 60–64 years of age were grouped with the 50–59 year old decade. Patients who were 30–39 years of age and 40–49 years of age were retained in age-specific 10 year or decade intervals. Race was classified into the following 4 categories: White, Black, other (i.e., American Indian/Alaskan Native, Asian, Native Hawaiian/other Pacific Islander, and multi-racial), and unknown. Individuals who had a heart transplantation history were not included.

All analyses were conducted using SAS 9.4 (Cary, NC). Frequencies were conducted for all categorical variables and means and standard deviations were computed for each age in its continuous form. For bivariate analysis, crude or unadjusted odds ratios were computed using chi-square tests to analyze the association of HTN/HF comorbid conditions exposures and other comorbidities with mortality. Bivariate logistic regression analyses were performed to assess the association of mortality with the three HTN/HF primary comorbidities separately (HTN without HF, HF without HTN and both HTN and HF) and with each of the other demographic variables including age, gender, race, and insurance. Multiple logistic regression adjusted models were then conducted to determine the likelihood of dying during the three-year surveillance period for CHD patients diagnosed with HTN without HF, HF without HTN, and both HTN and HF with each model separately, controlling for age group, gender, race, insurance, CHD severity, and interaction terms. The backwards elimination (BWE) approach was applied to determine if any interactions should be included in the models and addressed collinearity by removing inter-correlated variables; SAS default criteria for variable entry and exit in the models (0.05 for SLENTRY and SLSTAY). Odds ratio estimates and Wald confidence intervals in addition to p-values are reported for logistic regressions. The Hosmer-Lemeshow (HL) test statistic option assessed model goodness of fit.

3  Results

For the 5,397 patients included in the analysis, mean age was 33.2 years (SD = 13.7), 61.4% were females, 45.1% were White, 24.6% were Black, 1.1% identified themselves as race ‘other’ which included American Indian, Asians, native Hawaiians/Pacific Islanders, and multi-racial, and 29.2% had race unknown. Patients ranged from 18–64 years of age with more than half comprising the youngest age group, 18–29 years (51.2%, n = 2,761, mean = 22.1 years, SD = 3.4), with 19.3% between 30–39 years of age (n = 1,042, mean = 34.3 years, SD = 2.9), 12.2% were between 40–49 years old (n = 661, mean 44.2 years, SD = 2.9), and 17.3% (n = 933, mean = 57.0 years, SD = 4.3) in the 50–64 years age group. Over 18% (18.3%) of patients had a diagnosis of HTN without HF, 4.4% had a diagnosis of HF without HTN, 8.3% had both HTN and HF and 3.0% of the cohort died during the three-year study period. Patients with a severely complex CHD made up 29.0% of the sample. Public health insurance coverage was the most prevalent, 47.9%, followed by 40.2% with private insurance coverage. Only 1.1% of the cohort indicated that they were uninsured or self-pay patients and health insurance coverage was not indicated for 10.8% of the sample (Tab. 1).

Table 1: Demographics of Adult Congenital Heart Defect (ACHD) Patients, 2008–2010

Tab. 2 presents the association of mortality of patients with CHD who were diagnosed with either HTN without HF, HF without HTN, or with both HTN and HF, along with several demographics. Over the three-year study period, 3.9% of those diagnosed with HTN without HF died (ns) as did 11.8% of those diagnosed with HF without HTN (p < 0.0001), and 15.4% of those diagnosed with both HTN and HF (p < 0.0001). An age-specific trend in mortality was seen during the study period with the risk of dying greatest among older age groups: 8.8% for 50–64 year olds, 3.9% for 40–49 year olds, 1.7% among 30–39 year olds and 1.3% for 18–29 year olds (p < 0.0001). Survival favored females (2.6%) compared to males (3.7%) (p < 0.05). Black patients were more likely to die (5.4%) compared to White patients (2.9%) (p < 0.0001), and those with public health insurance were at significantly greater risk of dying (4.7%) compared to those with private insurance coverage (1.4%) (p < 0.0001).

Table 2: Hypertension, heart failure, and other comorbidities by mortality for adult Congenital Heart Defects (ACHD) patients, 2008–2010

Patients with shunt or valve (non-complex) CHD conditions were at significantly greater risk of mortality (3.4%) compared with those with complex CHD (2.0%, p < 0.0001), and these patients with non-complex lesions were also more likely to be diagnosed with HTN without HF (22.3%) or with both conditions (8.6%) compared to their complex CHD counterparts, 8.4% and 7.5%, respectively (p < 0.0001) (data not shown). This supplemental analysis also revealed that those with complex CHD lesions, were more likely to be diagnosed with HF without HTN (8.0%) compared to those with a shunt or valve CHD condition (3.0%) (p < 0.0001), and having a complex CHD was more prevalent among the youngest adults, 18–29 year olds (64.2.3%) compared to the 30–39 year olds (23.3%), the 40–49 year olds (8.9%) and the oldest age group (5.5%) (p < 0.0001).

An anatomic complexity-specific analysis revealed that those with valve lesions were at significantly greater risk of dying (4.6%) compared with those with a severe lesion (2.0%) or a shunt condition (1.7%) (p < 0.0001) (Tab. 3). Patients with a valve diagnosis were also more likely to have HTN without HF (27.3%) or both HTN and HF (11.5%) compared to those with a complex defect (8.4% and 7.5%, respectively) or a shunt (15.3% and 4.6%, respectively) (p < 0.0001). Valve (3.4%) and shunt patients (2.2%) were less likely to develop HF without HTN compared to patients diagnosed with a complex CHD (8.0%) (p < 0.0001). Among the two older age groups, 40–49 and 50–64 years olds, having a valve condition (52.9% and 69.5%), respectively) was more prevalent than having a shunt (25.4% and 21.1%, respectively) or a complex CHD (21.8% and 9.4%, respectively) (p < 0.0001) (row percentages not displayed in table), and although having a valve condition was more prevalent among females (54.1%) compared to males (45.9%) (p < 0.0001), among males, valve lesions were more common (52.1%) compared to a complex CHD (27.9%) or a shunt diagnosis (20.0%) (p < 0.0001) (row percentages not displayed in table). Blacks were more likely to have a valve lesion (44.8%) compared to a shunt (28.5%) or a complex condition (26.7%) (p < 0.001) (row percentages not displayed in table). Lastly, those with valve lesions were more likely to be public insurance beneficiaries (47.9%) than have private insurance (40.6%), be uninsured or a self-payer (1.4%) or have an unknown insurance type (10.1%) (p < 0.05).

Table 3: Hypertension, heart failure, and other comorbidities by CHD anatomic group for Adult Congenital Heart Defects (ACHD) patients, 2008–2010

Crude logistic models of mortality revealed that the likelihood of dying during the study period was almost 5 times greater for those with HF without HTN (4.96, 95%CI 3.23–7.63), and almost double that for those with both HTN and HF (9.40, 95%CI 7.68–13.05) (Tab. 4). The two oldest age groups were more likely to die compared to the 18–29 year olds, [(40–49 year olds: 3.0, 95%CI 1.81–5.02) and (50–64 year olds: 7.10, 95%CI 4.78–10.54). Patients with non-complex CHD (shunts, valves and shunts + valves) were 1.69 times more likely to die during the study period compared to patients with complex CHDs (95%CI 1.15–2.51). When valve patients were assessed separately from those with shunts, the odds of dying increased to 2.32 times greater compared to patients with severe complexity types of CHD (95%CI 1.55–3.47). Blacks compared to whites were almost twice as likely to die (1.88, 95%CI 1.35–2.64) and those covered by public insurance were over three times more likely to die during the study period compared to those with covered by private health insurance (3.42, 95%CI 2.29–5.09) (Tab. 4).

Table 4: Crude odds for the association of hypertension, heart failure and other comorbidities with mortality for Adult Congenital Heart Defect (ACHD) patients, 2008–2010

Tab. 5 shows final adjusted multivariate logistic regression models that assess the association of mortality for those patients with CHD who are also diagnosed with either HF without HTN or both HTN and HF separately controlling for select demographics; odds ratios, Wald 95% likelihood confidence intervals (CI) and p-values are reported. These models employed the backward elimination approach (BWE) and included age group, gender, race, health insurance coverage and CHD anatomic grouping as independent predictors of mortality along with interaction terms for each of the three primary comorbid conditions (HTN without HF, HF without HTN and both) with each demographic predictor. Since the association of HTN without HF failed to predict mortality, this multivariate model and its results are not reported in Tab. 5. Model 1 includes HF without HTN as the primary comorbid predictor of mortality, while Model 2 includes having both HTN and HF.

Table 5: Adjusted odds of heart failure with and without hypertension controlling for other comorbidities and interactions on mortality for Adult CHD (ACHD) patients, 2008–2010

The log odds equation used for the final adjusted models is:

Logit P (Death) = α+β1(primary comorbidity: [HF w/o HTN] or [Both HTN & HF]) + β2(Age Group)

+ β3(Gender) + β4(Race) + β5(Insurance) + β6(CHD Anatomic Group)

+ β7(primary comorbidity*Age Group) + β8(primary comorbidity*Gender)

+ β9(primary comorbidity*Race) + β10(primary comorbidity*Insurance)

+ β11(primary comorbidity*CHD Anatomic Group)

For Model 1, ACHD patients who had a diagnosis of HF without HTN were almost three times more likely to die (2.85, 95%CI 2.09–3.61), and dying was 1.75 times greater for Blacks than Whites (95%CI 1.21–2.55), and almost three times more likely for 40–49 year olds and over six times greater for 50–64 year olds compared to the youngest 18–29 age group (2.92, 95%CI 1.51–5.65 and 6.38, 95%CI 3.69–11.01, respectively). Patients covered by public health insurance were also twice as likely to die than those covered by private health insurance (2.22, CI95% 1.43–3.54). Patients with a valve lesion were at greater risk of dying during the three-year study period compared to those patients with a complex CHD (1.86, CI95% 1.11–3.13). When the combined effects of age with having a HF without HTN diagnosis on mortality were assessed for ACHD patients, 30–39 and 50–64 year olds with HF without HTN were favored to survive during the study period compared to their younger 18–29 year old counterparts (0.09, 95%CI 0.01–0.60) and (0.10, 95%CI 0.02–0.40), respectively. The BWE approach removed all other interaction terms, except for HF without HTN * age group, for failing to meet criterion for retention in the model. The Hosmer and Lemeshow Goodness-of-Fit Test (HL GoF) revealed that the data adequately fit this adjusted model (X2 = 5.69, df = 8, p < 0.68).

In Model 2, ACHD patients diagnosed with both HTN and HF were almost three times more likely to die (2.71, 95%CI 1.56–4.23) and the oldest age group (50–64 year olds) had significantly greater odds of dying compared to the youngest group of patients with CHD (2.67, 95%CI 1.63–4.40). Patients with CHD who also had concomitant HTN and HF diagnoses were twice as likely to die if covered by public health insurance compared to those with private health insurance coverage (2.34, 95%CI 1.38–3.97). In addition, those with valve CHD lesions had 2.33 times greater mortality compared to those with a complex CHD (95%CI 1.24–4.38). Although race was not shown to influence mortality statistically, Blacks with Concomitant HTN and HF diagnoses were 2.29 times more likely to die compared to Whites diagnosed with both HTN and HF (95%CI 1.06–5.05). Lastly, public insurance beneficiaries with both HTN and HF were favored for survival compared to private insurance beneficiaries with both diagnoses, and the same was true for HTN and HF patients with valve lesions in comparison to those with both comorbid conditions and complex CHDs [(038, 95%CI 0.15–0.97) and (0.25, 95%CI 0.09–0.65). This adjusted model also adequately fit the data (HL GoF X2 = 7.21, df = 8, p < 0.51).

4  Discussion

As adults with CHD represent a large proportion of the CHD population [21], it is important to understand contributors to mortality in the ACHD population, which may differ from a pediatric population. Decisions regarding care are often made based on anatomic complexity alone, although more recent ACHD guidelines integrate physiology [22]. While HF is known to account for a significant proportion of deaths in patients with CHD [5,6], the predictive value of HF with and without HTN on mortality in ACHD has not been clearly elucidated in the context of CHD anatomic complexity. This cross-sectional retrospective study revealed that adults with CHD experience significant mortality regardless of their anatomic CHD complexity, with even those considered to have simple defects experiencing increased mortality. CHD patients with any HF codes accounted for 12.7% (=686/5,397) of the cohort, but comprised 59.5% (=97/163) of all those who died. The prevalence of any HTN among adults at least 18 years of age and older is 29.0% as reported by CDC 2015–2016 [23]. In the current study, the overall proportion of those with HTN was 26.6% for ages 18–64 years, which is slightly less than CDC’s reported U.S. population-based prevalence. In mimicking an age-specific analysis using CDC age groupings of 18–39 years, 40–59 years and 60 and older [23] to our current data, a similar age trend for HTN was found: 13.5% for ages 18–39 years old, 53.7% for ages 40–59 year olds, and 74.3% for those 60 and older. This is comparable to U.S. HTN rates which shows increases for HTN with age from 7.5% between ages 18–39 years, to 33.2% between ages 40–59 years and 63.1% for 60 years and older [23]. While US population-based data reveals the highest prevalence of HTN among the black population (40.3%) [23], prevalence of HTN among blacks in the current study was lower at 27.0%. HF has been reported to affect 2.4% of the US population and is also seen to increase with age [24]. After the age of 79 years, 12% of the population has HF [24]. The proportion of individuals with HF in our cohort (12.7%) is higher than the published population-based prevalence. These findings emphasize the importance of ongoing surveillance for cardiac comorbidities in all patients with heart defects, as the development of HF is an important predictor of mortality in those with complex anatomy, shunt and valve lesions, and may occur at younger ages.

Unadjusted prevalence of ACHD patients diagnosed with HF without HTN and those diagnosed with both HF and HTN died approximately 5 times and almost double that rate, respectively. Overall mortality was 3.0% for this 3-year ACHD sample, and their age-related mortality was 1.3%, 1.7%, 3.9% and 8.8% for 18–29, 30–39, 40–49 and 50–64 year olds, respectively. In a supplemental age-specific analysis of mortality, when we constrained the sample only to those with HF, with or without hypertension, overall mortality in this 3-year study period showed an almost five-fold increase to 14.1% (=97/686), with mortality incrementally increasing from 23.7% (=23/97) in the youngest age group (18–29 year olds) to 46.4% (=45/97) in the oldest (50–64 year olds). Unadjusted logistic regression results showed a lower likelihood of mortality for ACHD patients who had a HF without HTN diagnosis (OR = 4.96) compared to ACHD with both comorbidities (OR = 9.40). Prior large-scale studies have noted a J- or U-shaped mortality curve in those hospitalized with HF possibly related to the worse outcomes seen in those who have HF with lower blood pressure [12,13]. The reason for this finding may be related to the proposed protective effect of blood pressure elevation in those with a HF diagnosis [25] or the known poor prognosis of hypotension [14] in association with HF. Although HTN is a significant risk factor for the development of HF, a similar J-shaped survival curve has been noted in individuals with HTN [26]. Previous studies have shown a higher mortality in men with HF compared to women [27], and have shown a possible U-shaped mortality curve based on age in adults with HF, with a higher mortality in those under 25 years of age and greater than 64 years of age, and the lowest mortality in the middle group [27]. While crude analyses favored females for survival, unlike prior studies, adjusted logistic regression did not demonstrate mortality being impacted by gender. Also, unadjusted mortality results revealed racial differences with black patients with CHD dying 1.88 times the rate of whites, and CHD severity differences for those with valve CHD lesions dying 2.32 times the rate of ACHD patients with complex CHDs.

Results from the final multivariate adjusted logistic regression model which included HF without HTN as the primary comorbidity (Model 1) revealed that Black ACHD patients were 1.75 times more likely to die compared to White ACHD patients, and that the two older age groups, 40–49 year olds and 50–64 year olds, were at risk for greater mortality compared to the youngest ACHD individuals (2.92, 95%CI 1.51–5.65 and 6.38, 95%CI 3.69–11.01, respectively). The adjusted multivariate model for ACHD patients with both HTN and HF diagnoses (Model 2) revealed a comparable increased mortality for the oldest patients (50–64 year olds) when compared to 18–29 years olds (2.67, 95%CI 1.63–4.40). Those with complex CHD tend to be younger, which may account for some of these differences in mortality across anatomic groups. Anatomic CHD group was predictive of mortality in patients with HF without HTN, suggesting less complex heart defects such as valvular disease and shunts may have a cumulative detrimental effect on heart function over time, leading to heart failure in these anatomic groups over time. The development of HF in adults with less complex heart defects may also be related to the interplay of their heart defect with development of acquired comorbid conditions. Mortality within this ACHD population may be more complex and influenced by the effects of having a concomitant HF diagnosis with CHD.

In the multivariate logistic models for those with HF codes in Models 1 and 2, older individuals with ACHD (50–64 year olds) had greater odds of dying compared to the younger age group, 18–29 years [(6.38, 95%CI 3.69–11.01) and (2.67, 95%CI 1.63–4.40), respectively]. In this study, complex heart defects accounted for a larger proportion of CHD in younger adults (29.0% = 1,565/5,397). The younger age groups were the least likely to die during the study window despite having more complex CHD and HF compared to their older age counterparts with less severe CHD and HF, suggesting age has a protective effect for those with comorbid HF. When the interactive effects of age and having HF without HTN on mortality were assessed irrespective of CHD severity, results revealed that one of the younger adult groups, 30–39 year olds, and the oldest age group, 50–64 year olds, were both favored for survival. Khairy et al. also reported a bimodal pattern of mortality which initially peaked in childhood and then again during late adulthood, with the earlier peak perhaps relating to a contribution by children with severe unrepaired or palliated CHDs [7]. Additionally, in an aging CHD population, common comorbid conditions such as diabetes, pulmonary hypertension, atrial fibrillation and other comorbidities may influence mortality [28] and may also contribute to the development of HF.

Lastly, for patients with both HF, with or without HTN, having a valve lesion was associated with a higher likelihood of death compared to those in the complex group or with shunt defects [(HF without HTN: 1.86, 95%CI 1.11–3.13) and (HTN and HF: 2.33, 95%CI 1.24–4.38)]. There are several possible explanations for this finding. The most likely explanation is a protective effect of younger age on mortality. In addition, anatomically complex CHD was defined as defects that typically require intervention or surgery in the first year of life [18,19]. In an ACHD population this may introduce survivor bias, as some children with complex CHD may not reach adulthood. Conversely, the effect of reparative surgery in early childhood may mitigate some of the risk of death in adulthood, and ‘simple’ shunt or valve heart defects may not capture the long-term cumulative effects of valve and shunt lesions on ventricular function, or the cumulative effects of multiple surgeries for valvular heart defects, which may lead to the development of ventricular dysfunction and HF in adulthood, particularly in the older age groups. Aortic valve disease, for example, may require repeated interventions over a lifetime [29,30], may also be associated with other left sided obstructive lesions such as coarctation of the aorta, and both aortic stenosis and regurgitation may contribute to the development of left ventricular dysfunction and HF [29,30]. The finding of increased mortality in all CHD groups warrants further study to understand contributors such as multiple surgeries over a life-time, and the potential consequences of unrepaired or late repair of shunt or valve lesions, the consequences of repeated surgeries for valvular heart defects, and reevaluation of the perception that adults with anatomically simple heart defects have a benign course requiring less frequent surveillance.

This study is limited by its cross-sectional retrospective design, use of administrative data, the inability to definitely verify diagnosis and comorbidity codes, define completeness of repairs, or measure blood pressure. Anatomic groupings are limited by use of ICD-9-CM codes, which may include cases that cross anatomic groups. Eisenmenger syndrome does not have a code in ICD-9-CM, and thus, could not be taken into account. Since our data evaluated individuals from the state of Georgia, results may not be broadly applicable to other geographic regions. In addition, our study window was only three years, and a longer study period may enable better detection of contributors to mortality from HTN. More complete race data would also allow for better assessments of comorbidities that historically are known to be more prevalent within certain racial or ethnic groups, and perhaps, shed light on racial disparities when they exist.

In conclusion, we found in a population-based cohort from Georgia, that adults with heart defects who also had HF experienced significant mortality in a 3-year period of time, and this risk for mortality was age-related. Patients with CHD valve lesions who were also diagnosed with HF, with or without HTN, were at increased risk of death. In older individuals, mortality was higher compared to younger individuals, and Black patients with CHD who had HF without HTN had higher mortality than Whites. Lastly, patients with CHD who also were diagnosed with either HF without HTN or both HTN and HF and who had public health insurance coverage had a greater risk of mortality than those covered by private health insurance carriers. Vigilance for signs and symptoms of HF in adults with all heart defects, and treatment to mitigate comorbid HF and HTN may lessen mortality in adults with CHD regardless of the anatomic complexity of the heart defect. Results from this study suggest factors in addition to anatomic complexity may more accurately predict the contributors to mortality in adults with CHD, emphasizing the need for all ACHD patients to stay in specialty care. Future studies should evaluate other comorbidities that are known to influence health outcomes.

Acknowledgement: The findings and conclusions in this report are those of the authors and do not necessarily represent the official position of the Centers for Disease Control and Prevention. The authors thank Trenton Hoffman, Sr. Data Analyst, for his expertise and assistance in preparing the analytic dataset, and Lindsey Ivey for helping confirm the references.

Data Sharing: The individual deidentified participant data cannot be shared.

Funding Statement: Cheryl Raskind-Hood and Wendy M. Book are supported by a Cooperative Agreement from the Centers for Disease Control and Prevention Cooperative Agreement, Public Health Pilot Project Surveillance of Congenital Heart Defects (CHDs) Focusing on Adolescents and Adults; FOA #DD12-1207.

Conflicts of Interest: The authors declare that they have no conflicts of interest to report regarding the present study.

Appendix A: CHD Data Sources

Appendix B: Congenital Heart Defect Severity Classification [18]

Appendix C: ICD-9-CM Codes Determining Hypertension (HTN) (40 codes)

: Number and Percent of Patients with Hypertension (HTN) Who Have These 40 Codes

Appendix D: ICD-9-CM Codes Determining Heart Failure (HF) (16 codes)

: Number and Percent of Patients with Heart Failure (HF) Who Have These 16 Codes

This work is licensed under a Creative Commons Attribution 4.0 International License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.