Introduction
Depression is commonly unrecognized and untreated [1, 2]. In the United States, its lifetime prevalence in the general population is 17%. Among those who have major depression, 9% commit suicide and only 50% seek treatment [3]. The prevalence of depression in patients with chronic illness is even higher, ranging from 20 to 42% [4–6]. In this setting, its insidious impact on health and association with adverse clinical outcomes in patients with chronic illness have been well documented. Depression has been demonstrated to be associated with the development of coronary artery disease, increased somatic symptoms leading to frequent clinic or emergency department visits, increasing frequency and duration of hospitalization, nonadherence to care recommendations, and increased mortality [7–10]. Depression thus produces physical, emotional, and financial burdens on patients, their families, and society. Given the known risk of depression in patients with chronic illness, clinicians and researchers in the adult congenital heart disease (ACHD) community have found it critical to assess psychological health status in their patients. With advances in medicine, more than 95% of infants with congenital heart disease (CHD) currently survive to adulthood, in contrast to only 25% in 1950 [11]. The number of adults with CHD now exceeds the pediatric CHD population in the Western world [12]. As ACHD prevalence increases, there is an ever-increasing emphasis on improving overall health – physical and psychological. As a result, numerous studies have investigated psychological conditions in ACHD patients and their associations with clinical variables, albeit with differing results [13–15]. In this review, we summarize existing data on the prevalence of depression, clinical features associated with depression, its prognostic impact, and psychological interventions to treat depression in ACHD patients.
Data on Depression in ACHD
As our goal was to systematically review all published data on depression in ACHD, we searched PubMed, Google Scholar, and Journal/Author Name Estimator for publications using the search terms “depression,” “ACHD,” “adult,” “congenital heart disease,” “review,” “single ventricle,” “Fontan,” “psychosocial,” “intervention,” “mental health,” and “psychopathology.” References in relevant articles were also reviewed to identify publications. Studies including patients aged 18 years or older or with a mean/median age older than 18 years were included. We targeted studies in which depression was clearly differentiated from anxiety or other psychological distress using an independent depression scale. Publications involving a large ACHD database review or a postoperative psychiatric evaluation of ACHD patients were excluded. Non-English publications were also excluded. Searching for all articles up to October 17, 2017, we identified 20 publications meeting these criteria; summaries of the findings are presented in Table 1 [16–35].
Year | Authors | Design | Country studied | Sample (n) | Age range (years) | CHD lesion complexity | Reference group | Depression assessed by | Results | Psychiatric interventions | |||
---|---|---|---|---|---|---|---|---|---|---|---|---|---|
Prevalence of depression | Depression compared with reference group | Association with variables | |||||||||||
Clinical | Psychosocial | ||||||||||||
2017 | Eaton et al. [16] | Cross-sectional | Australia | 135 | 18–49 | Various | None | HADS | 6% | – | – | Poor quality of life (linear analogue scale) | Not mentioned |
2016 | Westhoff-Bleck et al. [17] | Cross-sectional | Germany | 150 | 18–70 | Various | General population | BDI-II; HADS; clinical interview (DSM-IV criteria) | 36% (MDD or dysthymia) | Higher (vs. 7%) | – | Poor quality of life (WHOQOL-BREF) | Yes (only 26% of patients recommended for therapy received treatment) |
2016 | Amedro et al. [18] | Cross-sectional | France | 208 | 15–85 (mean 42) | Various with PAH | Within CHD lesion subgroups | HADS | 9% | No significant difference | Higher for female patients than male patients; worsening NYHA class; none with lesion type | Poor perceived health status (SF-36); poor quality of life (CAMPHOR) | Not mentioned |
2015 | Kourkoveli et al. [19] | Prospective cohort | Greece | 60 | Not available (mean 28) | Various | None | Zung SDS; BDI | 32% (Zung SDS); 32% (BDI) | – | Shorter event-free survival; higher risk of having adverse cardiac events | – | No (patients with antidepressant therapy were excluded) |
2015 | Bordin et al. [20] | Cross-sectional | Italy | 35 | 18–48 | Single ventricle | None | HADS | 69% | – | No significant difference with cardiac function | Poor perceived health status (SF-36) | Not mentioned |
2015 | White et al. [21] | Cross-sectional | United States | 80 | Not available (≥18) | Various | None | BDI-II | 13% | – | Sex-mediated medical nonadherence | Poor perceived health status (SF-36) | Not mentioned |
2013 | Eslami et al. [22] | Cross-sectional | Iran | 347 | 18–64 | Various | Non-heart-diseased controls (age and sex matched) | HADS | 42% | Similar | Great somatic symptoms | Perceived financial strain; lower social support | Not mentioned |
2013 | Bang et al. [23] | Cross-sectional | Republic of Korea | 85 | 20–52 | Various | Normative data (age and sex matched) | BDI | 9% | Not mentioned | No significant difference in NYHA class | – | Not mentioned |
2012 | Pike et al. [24] | Cross-sectional | United States | 54 | 15–50 (mean 26) | Single ventricle | Healthy controls (age matched) | PHQ-9 | 32% | Higher | – | Poor quality of life (Satisfaction With Life Scale) | Not mentioned |
2012 | Müller et al. [25] | Cross-sectional | Germany | 767 | 14–67 (median >18 in all CHD lesion subgroups) | Various | General population (age and sex adjusted) | CES-D | 9% | Lower | None with cardiac lesion severity; 14% of the cyanotic group with depression | Poor perceived health status (SF-36) | Not mentioned |
2012 | Riley et al. [26] | Cross-sectional | United Kingdom | 99 | 17–67 (mean 37) | Various | None | HADS | 17% | – | – | Poor perceived health status (SF-36) | Not mentioned |
2011 | Overgaard et al. [27] | Cross-sectional | Denmark | 62 | Not available (median 22) | Single ventricle | Healthy controls (age and sex matched) | HADS | Not mentioned | No significant difference | Worsening functional class (assessed by ACHD cardiologist) | – | Not mentioned |
2009 | Loup et al. [28] | Cross-sectional | Switzerland | 153 | Not available (mean 26) | Various | Standard population (age and sex adjusted) | HADS | Not mentioned | Similar | Older age within CHD groups; longer length of time since last operation | None | Not mentioned |
2009 | Kovacs et al. [29] | Cross-sectional | Canada; United States | 280 | Not available (≥18) | Various | None | BDI-II; clinical interview (DSM-IV criteria) | 12% (BDI-II); 33%* (MDD, BD, or dysthymia) | – | – | Poor perceived health status (SF-36); loneliness; fear of negative evaluation | Yes (approximately 60% of patients interviewed with treatment history) |
2003 | Bromberg et al. [30] | Cross-sectional | United States | 22 | 19–60 | Various | None | BSI; clinical interview (DSM-IV criteria) | 27% (with a depressive episode) | – | Greater illness severity (assessed by ACHD cardiologist) | – | No (patients with treatment history were excluded) |
2002 | Cox et al. [31] | Cross-sectional | United Kingdom | 87 | 17–73 (mean 31) | Various | Orthopedic outpatients | GHQ30; HADS | Not mentioned | Lower | – | – | Not mentioned |
2001 | Saliba et al. [32] | Cross-sectional | France | 89 | 17–49 (median 21) | Single ventricle | Healthy peers | Duke health profile | Not mentioned | Similar | None | None | Not mentioned |
2001 | Popelova et al. [33] | Cross-sectional | Czech Republic | 32 | 19–64 | Complex, cyanotic | None | Zung SDS | 34% | – | Older age; increasing NYHA class | Unemployment | Not mentioned |
2000 | Horner et al. [34] | Cross-sectional | United States | 29 | 26–56 | Complex | None | Clinical interview (DSM-III-R criteria); self-report symptom questionnaire | 14% (MDD) | – | – | – | No (none with treatment history) |
1991 | Brandhagen et al. [35] | Cross-sectional | United States | 168 | 24–42 | Various | Normative data | SCL-90-R (depression scale) | Not mentioned | Higher | None with cardiac lesion severity | – | Not mentioned |
ACHD, adult congenital heart disease; BD, bipolar disorder; BDI, Beck Depression Inventory; BDI-II, Beck Depression Inventory, second edition; BSI, Brief Symptom Inventory; CAMPHOR, Cambridge Pulmonary Hypertension Outcome Review; CES-D, Center for Epidemiologic Studies Depression Scale; CHD, congenital heart disease; DSM-III-R, Diagnostic and Statistical Manual of Mental Disorders, revised third edition; DSM-IV, Diagnostic and Statistical Manual of Mental Disorders, fourth edition; GHQ30, General Health Questionnaire 30; HADS, Hospital Anxiety and Depression Scale; MDD, major depressive disorder; NYHA, New York Heart Association; PAH, pulmonary arterial hypertension; PHQ-9, Patient Health Questionnaire depression module; SCL-90–R, Symptom Checklist-90–revised; SDS, self-rating depression scale; WHOQOL-BREF, World Health Organization Quality of Life, short form.
*Of 58 patients randomly selected for interview.
Prevalence
ACHD patients have unique characteristics and concerns that might lead to an increased prevalence of depression. These include parental overprotection, disability at early age limiting social functioning, developmental challenges, care transitioning, and chronic medical conditions requiring specialized care, all of which might be anticipated to increase psychological vulnerability [36, 37]. This predisposition has been suggested to be a consequence of accumulated, undealt difficulties throughout various developmental stages [37]. The reported prevalence of depression in ACHD, however, is highly variable, ranging from 6 to 69% in the studies included here. A significant component of this variability is likely attributable to multiple variables potentially affecting detected versus true prevalence in any given sample. A closer investigation of these variables is informative in defining the nature of depression in the ACHD population.
Some of the variability in reported depression prevalence is undoubtedly due to the setting in which a given study took place. The present review summarizes data from 13 countries on four continents (two countries in North America, two countries in Asia, eight countries in Europe, and Australia) with widely disparate rates of prevalent depression in the general population [17, 24, 25, 38]. Furthermore, the samples were drawn exclusively from ACHD-specialized clinics, the populations of which will more or less accurately reflect the actual total ACHD population depending on the nature of the health care system in the country from which the study originated. In the Netherlands, a nationalized health care system increases the probability of care continuity, and the relatively small land area of the country makes specialized care reasonably geographically accessible for the entire population. As such, studies from ACHD-specialized clinics in the Netherlands have a much greater probability of providing a representative sample of the overall ACHD population than those drawn from ACHD-specialized clinics in the United States, where the great majority of ACHD patients either receive no care or are seen in non-ACHD clinics. It should not be surprising then that the reported prevalence of depression might differ widely between and within continents (e.g., Asia, 9–42%; North America, 13–33%; Europe, 9–69%). Studies from the United States generally find a higher prevalence compared with normative data, while those from European countries demonstrate similar or lower rates [24, 25, 27, 28, 32, 35].
An additional layer of variability is contributed by differences in the methodology and execution of a particular study. These included differing response rates (lowest, 36%; highest, 100%), type of assessment tools (e.g., self-reported surveys versus a structured psychiatric interview), and sample size, which ranged from 22 to 787 participants (five studies with fewer than 50 participants and eight studies with more than 100 participants). Of particular relevance in this arena was the manner in which depression was identified. Although most of the studies we reviewed used self-reported surveys to screen patients for depression probably for practicality (e.g., shorter time required for assessment, mode of administration convenient for participants and investigators, simple and fast scoring, low costs), on review it appears that structured psychiatric interviews yield more reliable results [17, 29, 30, 39]. Psychiatric interview is the gold standard for the diagnosis of depression and is used to validate self-administered depression scales, which themselves appear to perform unreliably in this population as well as in heart failure patients [29, 40, 41]. Across studies, when a structured psychiatric interview was used for assessment, more patients were found to be depressed as compared with when self-reporting questionnaires were used [17, 29, 30]. When Kovacs et al. [29] assessed depression in a population using both psychiatric interview and the Beck Depression Inventory, psychiatric interview found 33% of patients with depression, while the Beck Depression Inventory, second edition, identified only 12% of patients with depression. Further, psychiatric interview has the additional ability to refine the diagnosis of clinical depression (i.e., major depressive disorder, bipolar disorder, dysthymia, and other types) and to thereby suggest specific interventions. Studies using a structured psychiatric interview may therefore more accurately reflect the actual prevalence of clinical depression, including differential diagnoses in ACHD patients.
With these caveats, ACHD patients as a whole likely have an incidence of depression similar to that of the general population. Among the studies reviewed, nine compared the prevalence of depression with that of a reference group and reported conflicting results from one another (three higher, four similar, and two lower prevalence in ACHD). Seven studies used the general population as a reference, while two used a sick patient group as a control for comparison. In these latter two studies, when compared with patients who had non-ACHD-related illness, ACHD patients were similarly or less depressed [22, 31]. The ACHD population is not monolithic however, and there is significant variability in the prevalence of depression identified depending on both functional class and lesion. Studies including patients with a lower functional class or lower disease complexity yielded a much lower prevalence of depression. Among the studies we reviewed, in the two with the lowest reported prevalence of depression (6 and 9%), almost all patients (≥98%) had New York Heart Association functional class I or II [16, 23]. This may not be surprising as the association of worsening functional status with depression has been well recognized in heart failure patients with normal cardiac anatomy in whom increased New York Heart Association functional class was linked with higher prevalence of depression (11 for class I vs. 42% for class IV) [41].
In addition, although depression was not associated with CHD lesion complexity or cardiac function, patients with either chronic cyanosis or a univentricular heart appear to have a higher depression prevalence compared with ACHD patients in other anatomical groups. In three of five studies including exclusively patients with a single ventricle or cyanotic CHD lesions, depression prevalence was consistently higher (32–69%) than in those studies with mixed CHD lesions [20, 24, 33]. Given these limited data, we suggest that depression among patients with a univentricular heart or chronic cyanosis requires unique investigation. Intuitively, these patients may be more vulnerable to psychological problems than those with other CHD lesions. In adulthood, these patients experience frequent complications even after successful palliation, including arrhythmia, thromboembolic events, protein-losing enteropathy, hepatic dysfunction, progressive cyanosis with associated complications of hyperviscosity and gout, and abnormal cardiorespiratory response to exercise [42]. These clinical symptoms would be anticipated to lead to restrictions in physical, emotional, and social functioning that may result in increased risk of depression [13, 29]. Future studies investigating the prevalence of depression specifically in ACHD patients with cyanosis or a single ventricle may therefore be of uniquely high yield.
The studies reviewed revealed several other clinical and psychological variables associated with prevalent depression in ACHD patients. Clinical variables associated with increased depression prevalence included older age, female sex, increased somatic symptoms, nonadherence to care, and longer length of time since the last operation [18, 21, 22, 25, 27, 28, 30, 33]. The psychosocial variables most frequently associated with depression were poor perceived health status and low quality of life (lower social support, loneliness, fear of negative evaluation, and unemployment were also found to be linked.) [16–18, 20–22, 24–26, 29, 33].
Association with Prognosis
Far from being of only esoteric interest, depression in ACHD appears to predict outcomes. Of 20 reviewed studies, only one was a prospective study and followed up patients for about 5 years to investigate the impact of depression on prognosis. In this study, Kourkoveli et al. [19] found that depression was independently associated with adverse clinical outcomes. They demonstrated that patients with depression had a shorter event-free survival and a two-fold higher risk of death or hospitalization due to heart problems than those without depression. These findings are supported by a retrospective study in which depression was associated with a shorter survival and increased risk of death [43]. The possibility of a link between depression and prognosis should not be surprising as it has been consistently demonstrated in other cardiac disease states. An association between depression and increased cardiac events specifically has been demonstrated both among patients with known coronary artery disease and in the general population [44]. Furthermore, in patients with heart failure, elevated rates of death and hospitalization have been linked to depression [41]. Depression was also associated with poor perceived health status and low quality of life in ACHD. Among the data included in this review, multiple cross-sectional studies identified a correlation between poor perceived health status or quality of life and depression [16–18, 20, 21, 24–26, 29]. Furthermore, one longitudinal study suggested that persistent depressive symptoms were responsible for poor quality of life and unfavorable perceived health status in young ACHD patients [45].
The causal relationship between depression and prognosis in ACHD is unclear. As ACHD patients age, chronic disease–related complications become more common, and heart failure specifically is a major driver of morbidity and mortality [46]. Among patients with heart failure with normal cardiac anatomy, self-assessed health measures are excellent predictors of adverse prognosis, reflecting a link between self-reported health status, depression, and prognosis [47]. The data we reviewed suggest a similar association between depression, poor functional class, and poor perceived health status, and a high risk of adverse clinical outcomes is likely to be present as well in ACHD patients. Furthermore, accumulating evidence hints at a possible causal relationship between depression and adverse outcomes. Among depressed male ACHD patients, nonadherence to care as defined by poor follow-up in a specialized ACHD clinic has been reported [21, 43]. It is possible that depression, by producing negative health behavior, may increase the probability of adverse outcomes given that ACHD-specialized care has been demonstrated to be associated with improved clinical outcomes [48]. Whatever the causal relationship is, the adverse association of depression with both a patient’s perceived health status and a patient’s prognosis suggests identifying and treating depression in ACHD is a potentially high-yield area for improving outcomes.
Interventions
Despite the high prevalence of depression and its adverse prognostic impact, studies on psychiatric treatment for depression in ACHD patients are scarce. Ferguson and Kovacs [36] retrospectively examined 100 ACHD patients who had received psychological assessment at their ACHD-specialized psychological service. Their clients comprised less than 5% of patients followed up clinically in a large ACHD program, of which their center is a part. Twenty-nine percent of reviewed patients met Diagnostic and Statistical Manual of Mental Disorders, fourth edition criteria for clinical depression (23% had major depressive disorder and 6% had dysthymia). Treatment was offered to 87 patients with a diagnosis of any psychiatric disorder, and 14% declined it for unknown reasons. Psychotherapy was the only treatment option offered, and cognitive therapy (92%) was used most frequently. These patients attended a median of eight, 1-h therapy sessions. At the end of treatment, 88% of patients had reduced or no psychological distress. However, this study did not report treatment benefit for patients with depression separately. In the only ongoing prospective study in this population to date, Kovacs et al. [49] developed an eight-session, 90-min intervention program, named Adult Congenital Heart Disease – Coping and Resilience (ACHD-CARE), and tested its feasibility with a pilot study. On the basis of patient-reported needs and wants in treatment collected from previous work, they designed a group therapy with four to six patients facilitated by two therapists [50]. Each session presented a different discussion topic with goals to improve psychosocial functioning, quality of life, and resilience through education, cognitive behavior therapy for coping, and social support. Each group consisted of participants of all ages and of various CHD lesion complexities. This study randomized participants evenly into ACHD-CARE and control (i.e., not receiving ACHD-CARE intervention) groups. A 3-month follow-up survey was collected from all participants following completion of the intervention period. This study is ongoing and will determine the feasibility of a full randomized clinical trial using ACHD-CARE in ACHD centers. Further studies on the efficacy of psychiatric interventions in depressed ACHD patients and their effects on prognosis are nevertheless needed.
Even with evidence that treating depression in ACHD is of benefit, substantial changes to care practices will likely be required to enjoy large-scale benefits from this information. Depression itself is quite treatable, with efficacy rates of up to 80%, even among individuals with severe clinical depression [3]. However, a major reason for treatment failure is nonadherence. In one study, when treatment was recommended to 62 ACHD patients, only 26% received specific therapy [17]. It is possible that the remaining 74% sought mental health services elsewhere at a later time, but this information was not available. One potential reason for this nonadherence to recommended depression care may be that patients are not interested in the type of therapy offered. This possibility is suggested by a comparison of two studies from England and Canada. Despite the higher prevalence of depression in ACHD reported in available data, Diller et al. [43] at the Royal Brompton Hospital in London found that only 3% of 6162 ACHD patients followed up in their clinic between 2000 and 2011 were receiving antidepressant drug therapy. This is likely an underestimation of the actual percentage of patients who were receiving any type of therapy for depression though, as Kovacs et al. [51] in Toronto found that ACHD patients preferred psychotherapy over pharmacotherapy (41 vs. 9%). Tailoring therapy to patient preference may thus significantly help in achieving adherence and thereby maximize the benefits of depression therapy.
In addition to nonadherence, the lack of regular screening for and recognition of depression is a huge impediment to successful treatment. The prevalence of unrecognized and untreated depression in ACHD patients is underscored by studies documenting a high prevalence of depression (14, 27, and 32%) among patients without previous or ongoing psychiatric interventions [19, 30, 34]. In one study, among patients investigated by Bromberg et al. [30] and identified as “emotionally well-adjusted who were not in apparent need of mental health services” by ACHD cardiologists, 27% had a diagnosis of clinical depression. Education in the ACHD community on the importance of depression, its adverse prognostic impact, and broad implementation of effective screening techniques is thus an additional and essential factor required for optimizing psychological care quality. In addition to more reliably identifying patients in need of therapy for depression, there is some suggestion that the treatment adherence rate may increase with an active screening and referral conducted by ACHD clinicians within an established ACHD specialized program. Ferguson and Kovacs [36] observed that among 100 ACHD patients who were referred to their ACHD-specialized psychological service by ACHD specialists as part of routine care, 86% of 87 patients to whom psychotherapy was recommended received treatment. This is a very high rate of adherence compared with that reported in the clinical trial setting. One explanation for this finding might be that most of the patients who actually presented to the psychological service were already determined to proceed with treatment, while patients in research do not anticipate discovering that they have a psychiatric disorder requiring therapy. It is nevertheless possible that psychiatric referral within the established ACHD health care framework provided an opportunity for more thorough education of both clinicians and patients on the importance of psychological problems that resulted in increased awareness and improved receptiveness of patients to therapy.
Conclusions
Despite methodological limitations in existing studies, depression appears at least as common among ACHD patients as in the general population, and has a profound impact on clinical and psychological outcomes. Just as ACHD patients benefit from specialized clinical care for physical conditions, their unique psychological construction may require specialized psychological care. Study results on the efficacy of pharmacotherapy or psychotherapy for depression in patients with or without comorbidity have been conflicting, and practical impediments to the identification and treatment of depression in this population remain [2, 52–56]. Given the potential for significantly impacting outcomes, however, investment in further investigation in this arena is warranted.