Access to the PDF text

Free Article !

Archives of cardiovascular diseases
Volume 108, n° 8-9
pages 437-445 (août 2015)
Doi : 10.1016/j.acvd.2015.03.005
Received : 5 November 2014 ;  accepted : 17 Mars 2015
Social support and the consequences of heart failure compared with other cardiac diseases: The contribution of support received within an attachment relationship
Le soutien social et les conséquences de l’insuffisance cardiaque par rapport à d’autres maladies cardiaques : la contribution du soutien reçu dans une relation d’attachement

Robert G. Maunder a, , b , Robert P. Nolan b, c, Jamie S. Park a, Richard James d, Gary Newton b, e
a Department of Psychiatry, Mount Sinai Hospital, Room 915, Mount Sinai Hospital, 600, University Avenue, M5G 1X5 Toronto, Ontario, Canada 
b Faculty of Medicine, University of Toronto, Toronto, Canada 
c Behavioural Cardiology Research Unit, University Health Network, Toronto, Canada 
d Department of Cardiology, North York General Hospital, Toronto, Canada 
e Department of Cardiology, Mount Sinai Hospital, Toronto, Canada 

Corresponding author.

Interpersonal support is protective in heart disease, but sources of support and the quality of support may change over time, especially with aging and disease progression.


To determine if support received within an attachment relationship with a spouse is more protective than other types.


Subjects were sex- and age-matched cardiac outpatients with (n =40) or without (n =43) heart failure; they were studied with an observer-rated measure of attachment and self-report measures of other variables.


Having heart failure was associated with more depressive symptoms and illness intrusiveness. Although perceived social support did not differ in people with or without heart failure, those with heart failure had a spouse as the primary source of attachment functions less frequently than those without heart failure (50% vs 79%; P =0.006). Not having a spouse as the main provider of attachment functions was a partial mediator of the relationship between disease type (heart failure or no heart failure) and depressive symptoms (β =–0.24, t =–2.2; P =0.03) and deficits in non-attachment support made a further independent contribution (β =–0.24, t =–2.4; P =0.02). Neither perceived social support nor having a spouse serving attachment needs made a significant contribution to illness intrusiveness.


Having someone other than a spouse to provide attachment support is more common in cardiac patients who have heart failure and is associated with an increased risk of depressive symptoms.

The full text of this article is available in PDF format.

Le soutien interpersonnel est un facteur de protection dans les maladies cardiaques. Néanmoins, les sources et la qualité du soutien peuvent changer progressivement, particulièrement avec le vieillissement et la progression de la maladie.


Déterminer si le soutien d’un conjoint représente un facteur de protection plus élevé que le soutien d’un non-conjoint.


Les patients cardiaques non hospitalisés furent choisis selon leur sexe et leur âge, ainsi que l’existence (n =40) ou l’absence (n =43) d’insuffisance cardiaque. L’attachement fut évalué par des observateurs indépendants. Les autres variables furent obtenues via auto-évaluation.


La présence d’insuffisance cardiaque était associée à plus de symptômes dépressifs et à une augmentation du caractère intrusif de la maladie. Le soutien social perçu n’a fait aucune distinction vis-à-vis l’insuffisance cardiaque, cependant ceux souffrant d’insuffisance cardiaque avaient moins souvent un conjoint comme la principale source de fonctions d’attachement (50 %) que ceux sans insuffisance cardiaque (79 % ; p =0,006). L’absence d’un conjoint à titre de principal fournisseur de fonctions d’attachement était un facteur médiateur partiel entre le type de maladie (présence ou absence d’insuffisance cardiaque) et les symptômes dépressifs (β =–0,24, t =–2,2 ; p =0,03). Le manquement au soutient non attachement apportait aussi une contribution indépendante (β =–0,24, t =–2,4 ; p =0,02). Ni le soutien social perçu, ni la présence d’un conjoint répondant aux besoins d’attachement n’apportaient une contribution significative au caractère intrusif de la maladie.


Les patients cardiaques ayant une personne autre qu’un conjoint leur fournissant du soutient d’attachement ont plus souvent une insuffisance cardiaque et ont un risque plus accru de symptômes dépressifs.

The full text of this article is available in PDF format.

Keywords : Social support, Attachment relationship, Heart failure, Depressive symptoms, Illness intrusiveness

Mots clés : Soutien social, Relation d’attachement, Insuffisance cardiaque, Symptômes dépressifs, Intrusion de la maladie

Abbreviations : AAP, MOS-SSS


Interpersonal support includes emotional, tangible and informational support as well as social role engagement [1], and is a protective factor in heart disease, associated with lower incidence of coronary artery disease [2], fewer adverse events [3, 4], increased longevity [5, 6] and better psychosocial function [7, 8]. The link between social relationships and illness is likely to be bidirectional; interpersonal interactions are affected by the progression of the disease, but also affect the course of the illness. Furthermore, the type of relationship in which support is received affects its quality. In particular, support that is received within an attachment relationship, such as from a spouse or committed romantic partner (hereafter referred to as ‘spouse’), may have a greater impact on health than support that is received from people in other types of relationships [9]. The availability and quality of support from a spouse may change as couples age or as chronic illnesses in one or both partners progress. However, little is known about the impact of support from a spouse versus other types of support in people with heart disease.

The current study distinguishes general social support from attachment-related support. The latter concerns interpersonal functions that increase one's sense of emotional security by providing an emotional ‘safe haven’ in times of distress and a ‘secure base’ from which to initiate independent activity [10, 11]. These functions are commonly provided by a spouse, although circumstances, including aging and disease progression, may lead a person to depend on others to provide attachment-related support. We also examine attachment type, a trait-like pattern of response to dynamic interactions in attachment relationships often categorized in four types: secure, dismissing, preoccupied and unresolved [12]. Attachment type influences social support [11] and is linked to a range of health outcomes, with secure attachment associated with better outcomes [13, 14, 15].

We hypothesize that the health benefits of support are diminished when a person with heart disease receives attachment-related support from a non-spouse. The consequences of cardiovascular disease that we studied were depressive symptoms, anxiety symptoms and diminished social/occupational function [16, 17]. We also explored if these relationships vary with sex, age and the presence or absence of heart failure (included as a final common pathway of many cardiac conditions and, therefore, a proxy marker for disease that has progressed further, to a more debilitating disease stage).


This study was a cross-sectional cohort comparison of two groups of outpatients treated for heart disease recruited from a larger study of 98 participants: adults with a confirmed diagnosis of heart failure (New York Heart Association class II or III; n =40) recruited from cardiology clinics at a downtown hospital and a suburban teaching hospital; and age- and sex-matched cardiology patients without heart failure recruited from the downtown hospital (n =43). Subjects were included if they had completed each of the measures of global perceived social support, attachment type and role of the person who provides core attachment needs. All subjects provided written informed consent and the project was approved by the research ethics boards of both hospitals.


Demographic and clinical information (including clinical diagnosis and left ventricular ejection fraction by echocardiogram) were obtained from medical records.

Perceived social support

The Medical Outcomes Study Social Support Scale (MOS-SSS) [1] measures social support as the sum of availability (‘how often is each of the following types of support available…’) of informational-emotional support, tangible support, affectionate support and positive social interaction. Nineteen items that measure availability on a five-point Likert scale are summed. Cronbach's alpha in this sample was 0.96.

Receiving attachment-support from a spouse or other person

The role of the person meeting core attachment needs was determined using the WHOTO scale, which was used previously in young adults to describe the transition from parents to peers and partners serving these functions [18]. The WHOTO asks six questions, two for each attachment function, e.g. ‘who is the person you don’t like to be away from?’ (proximity-seeking), ‘who is the person you want to be with when you are feeling upset or down?’ (safe haven), and ‘who is the person you would want to tell first if you achieved something good?’ (secure base). If a spouse was indicated for either WHOTO question, that attachment function was attributed to the spouse. Thus, the WHOTO yielded three variables: the person who is the object of proximity-seeking, the provider of a safe haven and the provider of a secure base (spouse/other). To simplify the analysis we consolidated these three variables into a single measure, the primary person who serves attachment functions. If a spouse served two or more of the three attachment functions, then the primary attachment figure was designated as ‘spouse’. If a spouse provided zero or one attachment function, the primary attachment figure was indicated as ‘other’. The median number of WHOTO questions for which the spouse was nominated was six for people for whom the primary attachment figure was assigned to ‘spouse’ and zero for people for whom the primary attachment figure was classified as ‘other’.

Attachment type

Attachment type was classified into secure, dismissing, preoccupied and unresolved, using the Adult Attachment Projective Picture System (AAP). The AAP asks a person to describe a series of seven line drawings of attachment situations (e.g. a child and woman face each other, sitting on opposite sides of the child's bed). The individual describes what the characters are thinking and feeling, what led up to the scene and what will happen next [19]. Verbatim transcripts are scored by trained coders with respect to both content and narrative characteristics, and are then classified. The AAP was designed to yield the same classifications as the gold standard Adult Attachment Interview [12].

Secure attachment is characterized by evidence of valuing close relationships, actively seeking repair when a relationship is strained or disrupted and associating such relationships with emotional reciprocity and comfort.

In dismissing attachment, close relationships are devalued, and cognitive strategies that contain or distance negative emotions are used to disavow and diminish perceived threats.

In preoccupied attachment, close relationships are approached with fear of separation or rejection, containing or distancing cognitive strategies are lacking, and action to repair strains and disruptions is absent or ineffective.

Unresolved attachment is a pattern in which there is evidence of unresolved memories of interpersonal trauma and loss [12].

Previous studies have found an interjudge reliability for AAP classifications of 85–90% for four-group classification and 90–92% convergent validity with two-group (secure versus all other types) Adult Attachment Interview classifications [19, 20]. In the current study, AAP transcripts from the first 49 cases were classified blindly and independently by the originators of the AAP, Carol George and Malcolm West. Ten disagreements in the classification of the first 29 cases were resolved by consensus, and inter-rater agreement for the next 20 cases was 100% (overall interjudge reliability of 80%). The remaining cases were classified by one of these raters.

Depression and anxiety symptoms

The Hospital Anxiety and Depression Scale is a 14-item questionnaire that measures current anxiety (seven items) and depressive symptoms (seven items) in non-psychiatric hospital patients [21]. Construct validity has been confirmed in patients with myocardial infarction, breast cancer and stroke [21]. In this cohort, Cronbach's alpha was 0.81 for anxiety and 0.76 for depression.

Social/occupational function

The Illness Intrusiveness Rating Scale probes the degree to which a health condition interferes with 13 domains of living: health, diet, work, active recreation, passive recreation, financial situation, relationship with partner, sex life, family relations, other social relations, self-expression, religious expression and community involvement. This scale has been used in several different chronic illnesses and shows adequate reliability and validity. Cronbach's alpha scores of internal consistency reliability are high, ranging from 0.79 to 0.90 [22, 23]. Participants’ ratings of the degree to which their cardiac condition interfered with each domain on a seven-item Likert scale were summed. In this cohort Cronbach's alpha was 0.90.


Perceived social support can be provided by a spouse or by other people, but some of the MOS-SSS items overlap with attachment functions. In order to measure the impact of attachment-related support independent of generic non-attachment support, the perceived social support variable used was the residual variance in total perceived support score after regressing the dichotomous WHOTO variable (primary person who serves attachment functions – spouse or other) on the raw MOS-SSS total score, with the residual variance saved as a new variable, named non-attachment support.

Comparison of variables between age groups and disease condition groups (heart failure present/absent) were tested by analysis of variance or the χ2 test, as appropriate. A mediation analysis, based on the methods of Baron and Kenny [24], was performed for dependent variables (depression, anxiety, illness intrusiveness) and potential mediators (receiving attachment function from a spouse, attachment type) that had a significant bivariate relationship with disease condition. Covariates (age, non-attachment support) were included if their relationship to dependent variables was at least near significance (P <0.1). Analyses were performed with IBM SPSS Statistics, version 22 (IBM, Armonk, NY, USA). Significance was set at P <0.05 (two-sided).


Participant characteristics are described in Table 1. The prototypic participant was a male (72%) in his late sixties (mean age±standard deviation, 68.7±11.2 years) with coronary artery disease (53%). Fourteen subjects with heart failure (35%) had coronary artery disease compared with 30 subjects without heart failure (70%; P <0.001).

Participants with or without heart failure did not differ in total perceived social support or in non-attachment support (Table 2). However, a spouse less commonly served as the primary source of attachment functions for people with heart failure than for those without heart failure (50% vs 79%, respectively: χ 2 [1,n =83]=7.8; P =0.006). Indeed, people with heart failure were less likely to report being married or in a common-law relationship (55% vs 81%: χ 2 [1,n =83]=6.7; P =0.02). As expected, those with heart failure reported more depressive symptoms and greater illness intrusiveness than those without heart failure (Table 2).

There were no significant differences between age cohorts (<60 years, 60–74 years, ≥75 years) in total perceived social support or non-attachment support (Table 3). The trend towards a spouse being more likely to provide attachment functions with increasing age was not significant. Being widowed was uncommon in all age cohorts (7% overall), whereas being separated or divorced was more common in younger participants (24%, 13% and 3% in the three age cohorts, respectively). With respect to attachment type, a significant trend towards greater prevalence of dismissing attachment with increasing age was observed (8%, 26% and 36% in the three age cohorts, respectively: χ2=6.4; P =0.04).

People receiving attachment functions from a spouse or from someone else did not differ in age, education or current employment status. However, people receiving attachment functions from a spouse were more likely to be male and, naturally, to be married or in a common-law relationship and not living alone (Table 4).

Table 5 indicates that depressive symptoms, anxiety symptoms and illness intrusiveness were not associated with attachment type but were significantly greater in people who did not have a spouse as the main provider of attachment functions. In addition, depressive symptoms and anxiety symptoms were negatively correlated with non-attachment social support (r =–0.22, P =0.04 and r =–0.22, P =0.04, respectively).

A mediation analysis was performed to test whether having a spouse as the main provider of attachment functions modified the relationship between disease type (heart failure versus no heart failure) and the psychosocial consequences with which it was associated: depressive symptoms and illness intrusiveness. The results demonstrate that having a non-spouse serving attachment needs is a partial mediator of the relationship between disease type and depressive symptoms (β =–0.24, t =–2.2; P =0.03) (Table 6). Deficits in non-attachment social support make a further independent contribution to depressive symptoms (β =–0.24, t =–2.4; P =0.02). With respect to illness intrusiveness, neither non-attachment support nor having a non-spouse serving attachment needs made a significant contribution to outcome after accounting for the effect of disease type (Table 6).


Two elements of this study allow for specificity in the interpretation of the relationship between interpersonal phenomena and the consequences of cardiac disease. First, we isolated the correlates of attachment-related and non-attachment-related social support. Second, we distinguished between the effects of disease type, as a proxy for disease progression, and the effects of age.

Having heart failure was associated with a greater likelihood that someone other than a spouse was meeting attachment needs, which is a novel finding, and this in turn was related to depressive symptoms, confirming the hypothesis. Receiving attachment needs from a non-spouse was not related to aging and was not because of becoming widowed. Overall, variance in depressive symptoms was explained by a direct contribution from disease type and by independent contributions from not having a spouse serving attachment needs and less generic social support. Thus, support from a spouse may be more effective for protection against depressive symptoms. It appears that a spouse provides a safe haven and a secure base for emotional regulation in a way that is more effective at reducing depressive symptoms than others. Social support, whether within or without an attachment relationship did not explain variance in illness intrusiveness after accounting for disease type.

With respect to attachment type, we found that dismissing attachment is more prevalent with advancing age, which is consistent with previous studies [25, 26]. There is no consensus currently as to whether the excess of dismissing attachment in older cohorts is explained by changes in attachment type within people over time, by cohort effects or possibly by earlier mortality in other attachment types [27]. A change in attachment pattern favouring more dismissing attachment over time could be understood as an adaptive response to adversity (including disease progression) and loss associated with aging. On the other hand, preoccupied attachment and unresolved attachment may be less adaptive, especially in those who are medically ill, and could increase the risk of earlier mortality.

Attachment type did not differ between cardiology patients with or without heart failure. To the extent that this dichotomous comparison serves as a rough proxy for measuring disease progression, this result suggests that attachment type does not change with progression of heart disease. It is an important caveat, however, that the lower prevalence of coronary artery disease among subjects with heart failure suggests that there are other differences in disease experience between the cohorts in this study, beyond stage of disease. Longitudinal research could better address this question.

To date, there has been little research on the distribution of attachment types among adults who are medically ill using observer-rated methods of classifying attachment. A study of 40 younger adult women with lupus found 38% secure, 5% dismissing, 25% preoccupied and 33% unresolved [28]. A study of 20 adults with spasmodic torticollis found 25% secure, 55% dismissing and 20% preoccupied (unresolved not tested) [29]. Finally, among 59 adults with a ‘physical handicap’, 61% were secure, 22% dismissing, 5% preoccupied and 12% unresolved [30]. These studies are so heterogeneous in the populations studied and in their results that no synthetic interpretation is possible. Nonetheless, the distribution of attachment types in the current study is remarkable for the paucity of secure attachment. It is not known if this is primarily attributable to characteristics of the participants or whether it draws into question the suitability of the AAP for older adults with serious medical conditions.

This study is limited by its cross-sectional design, which precludes drawing conclusions about causality, and by sample size. Its strengths include the use of a validated observer-rated method to classify attachment type and its comparison of age- and sex-matched samples of patients with different types of heart disease. The inference that the characteristics of a cohort of people with heart failure represent the features of those who are further along the trajectory of the natural history of heart disease is an assumption in this cross-sectional cohort study that requires testing by longitudinal research.


The failure of previous efforts to mitigate the progression of cardiovascular disease by bolstering social support suggests the need to attend more closely to the various ways in which support is provided and perceived [6]. The findings of this study are consistent with the hypothesis that attachment-related support provided by a spouse has effects on outcomes of cardiac disease that are independent of non-attachment support. This is a distinction that may guide the development of supportive resources in older people with heart disease.

Disclosure of interest

The authors declare that they have no conflicts of interest concerning this article.


The authors thank Sarwat Abbasi, Katie Watson MacDonell and Michelle Khayeri, and acknowledge the support of Dr. Carol George, Dr. Malcolm West and Dr. Susannah Mak.

Sources of funding: this study was supported by an operating grant (NA 6924) from the Heart and Stroke Foundation (Ontario).


Sherbourne C.D., Stewart A.L. The MOS social support survey Soc Sci Med 1991 ;  32 : 705-714 [cross-ref]
Lett H.S., Blumenthal J.A., Babyak M.A., Strauman T.J., Robins C., Sherwood A. Social support and coronary heart disease: epidemiologic evidence and implications for treatment Psychosom Med 2005 ;  67 : 869-878 [cross-ref]
Kawachi I., Colditz G.A., Ascherio A., and al. A prospective study of social networks in relation to total mortality and cardiovascular disease in men in the USA J Epidemiol Community Health 1996 ;  50 : 245-251 [cross-ref]
Orth-Gomer K., Rosengren A., Wilhelmsen L. Lack of social support and incidence of coronary heart disease in middle-aged Swedish men Psychosom Med 1993 ;  55 : 37-43 [cross-ref]
Barth J., Schneider S., von Kanel R. Lack of social support in the etiology and the prognosis of coronary heart disease: a systematic review and meta-analysis Psychosom Med 2010 ;  72 : 229-238 [cross-ref]
Holt-Lunstad J., Smith T.B., Layton J.B. Social relationships and mortality risk: a meta-analytic review PLoS Med 2010 ;  7 : e1000316
Brummett B.H., Babyak M.A., Barefoot J.C., and al. Social support and hostility as predictors of depressive symptoms in cardiac patients one month after hospitalization: a prospective study Psychosom Med 1998 ;  60 : 707-713 [cross-ref]
Frasure-Smith N., Lesperance F., Gravel G., and al. Social support, depression, and mortality during the first year after myocardial infarction Circulation 2000 ;  101 : 1919-1924 [cross-ref]
Stroebe W., Stroebe M., Abakoumkin G., Schut H. The role of loneliness and social support in adjustment to loss: a test of attachment versus stress theory J Pers Soc Psychol 1996 ;  70 : 1241-1249 [cross-ref]
Bowlby J. Attachment and loss volume 1: attachment  New York: Basic Books (1969). 
Mikulincer M., Shaver P.R. Attachment in adulthood: structure, dynamics, and change  New York: Guilford Press (2007). 
Hesse E. The Adult Attachment Interview: protocol, method of analysis, and empirical studies Handbook of attachment: theory, research and clinical applications New York: Guilford Press (2008).  552-598Chapter 25.
Ciechanowski P., Russo J., Katon W., and al. Influence of patient attachment style on self-care and outcomes in diabetes Psychosom Med 2004 ;  66 : 720-728 [cross-ref]
Kirchmann H., Nolte T., Runkewitz K., and al. Associations between adult attachment characteristics, medical burden, and life satisfaction among older primary care patients Psychol Aging 2013 ;  28 : 1108-1114 [cross-ref]
McWilliams L.A., Bailey S.J. Associations between adult attachment ratings and health conditions: evidence from the National Comorbidity Survey Replication Health Psychol 2010 ;  29 : 446-453 [cross-ref]
Lane D., Carroll D., Ring C., Beevers D.G., Lip G.Y. Mortality and quality of life 12 months after myocardial infarction: effects of depression and anxiety Psychosom Med 2001 ;  63 : 221-230 [cross-ref]
Yohannes A.M., Willgoss T.G., Baldwin R.C., Connolly M.J. Depression and anxiety in chronic heart failure and chronic obstructive pulmonary disease: prevalence, relevance, clinical implications and management principles Int J Geriatr Psychiatry 2010 ;  25 : 1209-1221 [cross-ref]
Fraley R.C., Davis K.E. Attachment formation and transfer in young adults’ close friendships and romantic relationships Personal Relationships 1997 ;  4 : 131-144 [cross-ref]
George C., West M.L. The Adult Attachment Projective Picture System  New York: Guilford Press; (2012). 
West M., George C. Attachment and dysthymia: the contributions of preoccupied attachment and agency of self to depression in women Attach Hum Dev 2002 ;  4 : 278-293 [cross-ref]
Zigmond A.S., Snaith R.P. The hospital anxiety and depression scale Acta Psychiatr Scand 1983 ;  67 : 361-370 [cross-ref]
Devins G.M., Dion R., Pelletier L.G., and al. Structure of lifestyle disruptions in chronic disease: a confirmatory factor analysis of the Illness Intrusiveness Ratings Scale Med Care 2001 ;  39 : 1097-1104 [cross-ref]
Devins G.M., Edworthy S.M., Seland T.P., Klein G.M., Paul L.C., Mandin H. Differences in illness intrusiveness across rheumatoid arthritis, end-stage renal disease, and multiple sclerosis J Nerv Ment Dis 1993 ;  181 : 377-381 [cross-ref]
Baron R.M., Kenny D.A. The moderator-mediator variable distinction in social psychological research: conceptual, strategic, and statistical considerations J Pers Soc Psychol 1986 ;  51 : 1173-1182 [cross-ref]
Diehl M., Elnick A.B., Bourbeau L.S., Labouvie-Vief G. Adult attachment styles: their relations to family context and personality J Pers Soc Psychol 1998 ;  74 : 1656-1669 [cross-ref]
Magai C., Hunziker J., Mesias W., Culver L.C. Adult attachment styles and emotional biases Int J Behav Dev 2000 ;  24 : 301-309
Magai C. Attachment in middle and later life Handbook of attachment: theory, research and clinical applications New York: Guilford Press (2008).  532-551Chapter 24.
Barbasio C., Granieri A. Emotion regulation and mental representation of attachment in patients with systemic lupus erythematosus: a study using the Adult Attachment Interview J Nerv Ment Dis 2013 ;  201 : 304-310 [cross-ref]
Scheidt C.E., Waller E., Malchow H., and al. Attachment representation and cortisol response to the adult attachment interview in idiopathic spasmodic torticollis Psychother Psychosom 2000 ;  69 : 155-162 [cross-ref]
Bakermans-Kranenburg M.J., van I.M.H. The first 10,000 Adult Attachment Interviews: distributions of adult attachment representations in clinical and non-clinical groups Attach Hum Dev 2009 ;  11 : 223-263 [cross-ref]

© 2015  Elsevier Masson SAS. All Rights Reserved.
EM-CONSULTE.COM is registrered at the CNIL, déclaration n° 1286925.
As per the Law relating to information storage and personal integrity, you have the right to oppose (art 26 of that law), access (art 34 of that law) and rectify (art 36 of that law) your personal data. You may thus request that your data, should it be inaccurate, incomplete, unclear, outdated, not be used or stored, be corrected, clarified, updated or deleted.
Personal information regarding our website's visitors, including their identity, is confidential.
The owners of this website hereby guarantee to respect the legal confidentiality conditions, applicable in France, and not to disclose this data to third parties.
Article Outline