br A greater adaptive flexibility may contribute to
A greater adaptive flexibility may contribute to the protective eﬀect of attachment security on depression and anxiety that has been de-monstrated empirically in advanced cancer . In that regard, the meta-analysis of Nissen et al.  suggests an inverse relationship be-tween attachment security and depression in cancer, reporting pooled correlations of 0.29 for attachment anxiety, and of 0.20 for attachment avoidance. Another study examining the mechanisms by which at-tachment security may buﬀer psychological distress in cancer found that attachment anxiety moderated the association of physical symptom burden and depression, with highest levels of depression among in-dividuals with high attachment anxiety and greater symptom burden . This eﬀect may be explained by the tendency of those with high attachment anxiety to amplify distress in response to threat and to be limited in their capacity for flexible 14605-22-2 in the context of serious medical illness.
The relationship of attachment security to existential distress con-structs such as hopelessness, sense of dignity, death anxiety, demor-alization or the desire for hastened death has been little studied. Cicero et al.  found a negative association between attachment security and the Mental Adjustment to Cancer helplessness/hopelessness subscale. However, to our knowledge, no previous study has studied the inter-action of physical symptoms and attachment security on demoraliza-tion. Demoralization has been defined as a clinically relevant syndrome of existential distress that is characterized by a perceived inability to cope, a sense of hopelessness and lack of meaning and purpose in life [10,11]. There is growing evidence that demoralization is a frequent and distinct stress response in advanced cancer and makes a unique contribution to health care outcomes [12–14]. Further, cross-sectional and prospective studies have reported a positive association between physical symptom burden and demoralization in cancer patients [12,15].
In the present study, we aimed to investigate the contribution of attachment security to demoralization among patients with advanced cancer. We particularly tested the hypothesis that attachment security is associated with less vulnerability to demoralization in the presence of greater physical symptom burden. We further compared the results for demoralization with those for depression, a widely used approach to assess psychological distress in cancer. Specifically, we aimed to:
a) examine the association between attachment security and demor-alization, after controlling for demographic factors and physical symptom burden;
b) determine whether attachment security moderated the association of physical symptom burden with demoralization; and
c) explore diﬀerences between the contribution of attachment anxiety and attachment avoidance to demoralization and depression.
We hypothesized that:
1) Attachment security is negatively associated with demoralization, after controlling for demographic factors and physical symptoms. 2) Attachment security moderates the association between physical symptom burden and demoralization, such that the relationship between symptom burden and demoralization is weaker when Journal of Psychosomatic Research 116 (2019) 93–99
attachment security is greater.
2.1. Participants and procedures
For beta decay study, cross-sectional data was obtained from patients with advanced cancer attending outpatient oncology clinics at the Princess Margaret Cancer Centre, University Health Network (UHN) in Toronto, Canada. Advanced cancer was defined by diagnosis of a stage IV solid tumor or stage III lung or ovarian tumor, or any stage pancreatic cancer, all consistent with an expected survival of 12–18 months. Patients were invited to participate in a randomized controlled trial (RCT) of a psy-chotherapeutic intervention called CALM (Managing Cancer And Living Meaningfully ). Patients younger than 18 years, with severe cog-nitive or physical impairment, language barrier, and current receipt of other psychotherapeutic interventions from the Centre's Department of Supportive Care were excluded from the study. Patients provided written informed consent and completed baseline measures before randomization. This study received approval from the UHN Research Ethics Board.
Demographic data were collected from patients using a standar-dized self-report questionnaire. Disease-related characteristics were obtained from medical charts.
We assessed attachment security using the self-report Experiences in Close Relationships Scale in its 16-item modified version (ECR-16). The psychometric performance of this shortened version has been tested in patients with advanced cancer and found to be comparable to the ori-ginal version [17,18]. The two dimensions of attachment security are assessed by attachment anxiety and attachment avoidance subscales. Total scores may range from 8 to 56 for each subscale. Higher subscale scores indicate higher attachment anxiety and attachment avoidance, respectively. The total attachment security scores were reversed for analysis purposes, so that higher scores of the total scale would corre-spond to higher attachment security (total scores may range from 16 to 102).