Prevalence of Depression in Breast Cancer Survivors : a Systematic Review of Observational Studies

s =


Introduction
Breast cancer has one of the highest five-year survival rates among female malignancies i.e. between 80% and 95% (Coleman et al., 2011).Diagnosis of breast cancer might no longer be regarded as fatal but has been increasingly acknowledged as treatable.Nevertheless, cancer diagnosis and its treatment are recognized to be stressful times that underscored the need for clinicians to actively identify its psychological sequelae such as depression in the vulnerable patients.Early detection and treatment of depression in breast cancer sufferers not only significantly improved their quality of life but also increased their survival rates (Weinberger et al., 2010).
The period following completion of treatment and thereafter is also the transition time when women were transformed from the role of breast cancer 'patients' to become 'survivors' (Allen et al., 2009).The definition of breast cancer survivors has become broader; it used to refer to those who were cancer-free for a minimum of 5 years

Prevalence of Depression in Breast Cancer Survivors: a Systematic Review of Observational Studies
Nor Zuraida Zainal 1 *, Nik Ruzyanei Nik-Jaafar 2 , Azlin Baharudin 2 , Zuraida Ahmad Sabki 1 , Chong Guan Ng 1   after their diagnoses (Odle, 2011).This has substantially changed to encompass the period immediately after the completion of treatment, including those who were symptoms-free or had recurrence until his or her death (Odle, 2011).This is also the period whereby survivors may be at risk of developing long-term physical and psychological complications of the disease and/ or its array of treatment including mastectomy, radiotherapy and chemotherapy (Howard-Anderson et al., 2012).While data on the survivors' long-term physical health and their quality of life have considerably accumulated, greater attention to examine the psychological states of the survivors are needed to provide emotional support and care during survivorship.
Depression is often underestimated in breast cancer sufferers, with prevalence varied between 10% and 25% (Qiu et al., 2012;Popoola and Adewuya, 2012).While depression was found to be highest at diagnosis (Odle, 2011), depression during survivorship has not been studied adequately.During the follow-up care, survivors' psychological needs were often neglected (Park and Hwang, 2012).A main theme of depression among breast cancer survivors was the fear of breast cancer recurrence and metastasis (Pan et al., 2010).Long term effects associated with chemotherapy include infertility, sexual dysfunction, menopause-associated health problems e.g.osteoporosis and cardiovascular disease (Azim et al., 2011); all of which can be associated with depression.In addition, tamoxifen, an anti-oestrogen for treatment and prophylaxis of hormone-sensitive breast cancer has also been associated with depression (Lorizio et al., 2012).Other main concern associated with depression among survivors was body image, particularly in those who were sexually active.A related factor was the concern about physical appearance which might predict the patients' psychosocial adjustment post-surgery (Begovic-Juhant et al., 2012).The other associated factors include impending loss of support, continual follow-up medical care and problems related to adjustment to changes in their normal lives.While there were many individual studies which considered several of the factors above were conducted; to the authors' knowledge, systematic review on the psychosocial variables associated with depression during survivorship has not been performed.
Potential sources in the gap and inconsistencies of current knowledge include differences in the study methods (eg.variable measures in screening or diagnosing depression), variable definition of study populations i.e. survivors, sample selection, sampling time points, and the quality of reporting.A good review of this topic was conducted systematically (Fann et al., 2008) but did not focus on depression during survivorship; rather it looked at depression at no specific point of the disease.
These discrepancies emphasized the need to examine the current literature on depression among the breast cancer survivors.The focus on breast cancer survivors would provide a greater understanding of the frequency of depression and psychosocial characteristics of those in the period after going through a life threatening disease and its treatment.This helps to define future research priorities and sustain evidence-based psychosocial care for cancer survivors.
This paper aims to systematically review the medical literature which measured the prevalence of depression and examined the associated psychosocial and clinical factors of depression in breast cancer survivors.

Materials and Methods
The review procedure was guided by PRISMA (preferred reporting items for systematic reviews and meta-analyses) statement (Moher et al., 2009).

Search strategy
Identification of relevant studies began with systematically searches of the electronic databases through PubMed, CINAHL, PsycINFO and Ovid) with the last search being up to 15 th November 2012.The search terms used were (Depression OR Dysthymia) and (Cancer OR Tumor OR Neoplasms).Searches were limited to studies published in the English language.Hand search was done according to the reference lists of all eligible studies to identify relevant studies.

Selection strategy
Title and abstract of each study were reviewed independently by two investigators (NZ & NR) to retrieve the full-text articles.To reinforce validity, the investigators then compared the papers and discrepancies resolved by consensus.Duplications or multiple publications were excluded to avoid double counting.Inclusion criteria were as follows: (i) adult patients (aged≥18); (ii) all or predominantly breast cancer survivors (survivor is defined as individual who has completed active treatment ie.surgery, adjuvant chemotherapy and radiation, and no evidence of disease recurrence); (iii) adequate data on prevalence of depression and (iv) all observational studies (cross-sectional, prospective and case-control).All methods of screening or diagnosing depression were allowed.However, quality-of-life subscale data were excluded because quality-of-life scales measure a wider construct than depression-specific rating scales.Clinical trials and experimental studies, reviews, commentaries, case reports, systematic review and meta-analyses were excluded.
One author (NZ) extracted the data from all the eligible studies, entered them into a pre-designed data extraction form.The place of study, study design, measurement tools for depression, description of the subjects, prevalence of depression and its associated factors were extracted.The results were synthesized in a narrative review.

Results
We identified 2250 titles from MEDLINE, CINAHL, PsyINFO, Ovid and reference lists hand-search strategies.Further screening for relevant articles based on title and abstract yielded 190 articles.One hundred and fifty-eight articles were excluded after reviewing and applying the inclusion criteria.Hence 32 studies were eligible for systematic review.The flowchart is shown in Figure 1.

Methods of studies
The study design and characteristics of the breast cancer survivors are listed in Table 1.Most of the studies were conducted in USA (thirteen), followed by in Europe (five) and four studies from United Kingdom.Seven    (Inagaki et al., 2004;Chen et al., 2009;Karakoyun-Celik et al., 2010;Vahdaninia et al., 2010;Alcalar et al., 2012;Hsiao et al., 2012;Jang et al., 2012).Nineteen of the studies were cross-sectional design and twelve prospective studies.Eight distinct methods were used to diagnose depression in the 32 included studies.The most frequent methods used were the dichotomous approach using patient-rated scales.Eleven studies used Center for Epidemiological Studies -depression (CES-D), 6 studies used Beck Depression Inventory (BDI) and 6 studies used Hospital Anxiety and Depression Scale (HADS).On the other hand, 5 studies diagnosed depression by using a structured clinical interview.Each of the studies conducted in the Asian countries used different type of scales for diagnosis.

Subjects
A total of 10,826 breast cancer survivors were included in the 32 studies (median sample size=129 patients, range=15-2208 patients).The mean age of the subjects ranged from 47-63 years.In studies that reported ethnicity and marital status, majority of those in the western studies were Caucasians or white (median=80%, range: 30-100%) and about three-quarter of the subjects were married (median=73%, range=33-100%).Most breast cancer survivors had stage 0-II during the diagnosis and mean/ median duration post-treatment was between 1-98 months.

Factors associated with depression
Factors that were found to be significantly associated with depression in breast cancer survivors are presented in Table 2. Depression was most frequently associated with fatigue (Bower et al., 2000;2011;Den Oudsten et al., 2009;Vahdaninia et al., 2010) followed by lower income/poor financial status (Eversley et al., 2005;Morrill et al., 2008;Chen et al., 2009), low education (Mehnert and Koch, 2008;Morrill et al., 2008;Alcalar et al., 2012) and younger age (Broeckel et al., 2000;Mehnert and Koch, 2008).Regarding marital status, being unmarried (Broeckel et al., 2000;Morrill et al., 2008), widowed or divorced (Chen et al., 2009) were significantly associated with depression although in contrast with one study which found being married (Pumo et al., 2012) was associated with depression.Deshields et al. (2006) found having more number of children at home were associated with depression.In terms of ethnic variation, only one study found that being Latina was associated with depression (Deshields et al., 2006).
Beside fatigue, other physical symptoms associated with depression were pain (Vahdaninia et al., 2010), sleep disturbance (Bower et al., 2011) and low sexual desire (Speer et al., 2005).Other illness-related factors included reduced chance of survival (Watson et al., 1999), increased cortisol level at 2100 hours (Hsiao et al., 2012) and high co-morbidity index (Chen et al., 2009).In terms of time since diagnosis, Broeckel et al. (2000) found more time elapsed since diagnosis to be associated with depression which contradicted with 'recent diagnosis of breast cancer' reported by another study (Khan et al., 2012).
Factors related to treatment of breast cancer reported to be significantly associated with depression included more time elapsed since the end of chemotherapy (Broeckel et al., 2000), receiving chemotherapy (Eversley et al., 2005;Morrill et al., 2008), had mastectomy (Eversley et al., 2005), had breast conserving type of surgery (Den Oudsten et al., 2009) and using an alternative treatment (Karakoyun-Celik et al., 2010).
One study found that depression was correlated with poor quality and dietary intakes (Tangney et al., 2002).Other significant lifestyle factors were poor sleep quality (Carpenter et al., 2004;Den Oudsten et al., 2009) and shorter sleep duration (Carpenter et al., 2004).Three studies found poor quality of life was significantly associated with depression (Deshields et al., 2006;Chen et al., 2009;Rabin et al., 2009) in which Rabin et al. (2009) found the significance was confined to the domains of physical, psychological and social relations (Rabin et al., 2009).Social support was examined in several studies which report a significant association with depression i.e. poor social support (Mehnert and Koch, 2008;Den Oudsten et al., 2009), being unaccompanied by spouse and requested for help from psychologist (Karakoyun-Celik et al., 2010).Other factors included having frequent cancer discussion and high level of detrimental interactions (Ullrich et al., 2008).

Discussion
The prevalence of depression among breast cancer survivors varies across the extreme of 1-56% according to how it was defined.The range of prevalence of depression reported by Fann et al. (2008) in a review of breast cancer population of any phases of treatment was rather narrower ie.10-25%.The study populations in this systematic review were rather homogenous in terms of their mean age, marital status, stage of breast cancer and have completed their treatment.However the tool for screening/ diagnosing depression varies from one study to another.CES-D returned about similar prevalence of depression (median=22%, range=13-56%) with BDI (median=22%, range=17-48%) while HADS returned a lower prevalence (median=10%, range=1-22%).CES-D has been reported as a valid and reliable tool to screen depression in breast cancer patients (Hann et al., 1999) and demonstrates the same general structure of responses in different ethnic groups (Makambi et al., 2009).Similarly BDI and HADS are other useful screening tools for depression in cancer patients (Wang et al., 2011;Warmenhoven et al., 2012).However these are symptom scales that were not designed to diagnose Major Depression.The used of diagnostic interview scales such as SCID, SADS and MINI in the studies included in this systematic review yielded prevalence of depression between 9-25%.Eventhough HADS seems to detect about similar prevalence of depression as compared to the diagnostic tools, it is not recommended to be used for case-finding instrument but suitable in screening programme (Mitchel et al., 2010).Analysis of the items in HADS-depression used in cancer patients, combination of "feeling down, depressed or hopeless" with "thoughts that would be better off dead, little interest or pleasure in doing things" gave high sensitivity but low specificity (Mitchell et al., 2012).
The prevalence of depression in breast cancer survivors reported in the Asian studies varied 12.5-31% (median=19%).The prevalence was not much different from those found in the western countries.Two of the studies used DSM criteria to diagnose Major Depression and five studies conducted in China, Turkey, Iran and Taiwan used self-report screening tools.
The factors associated with depression varied from one study to another depending on the primary objectives of the study.Some studies have reported correlation between psychosocial factors such as being women, single or divorcee, lower education, low income, poor social support with depression.Other study reported no association between socio-demographic variables and depression (Karakoyun-Celik et al., 2010).Illness factors such as pain and physical impairment have been associated with depression in medically ill patients (Johnson et al., 2006).Similarly pain could predict depression in breast cancer survivors (Vahdaninia et al., 2010).Common physical symptom in cancer patients such as fatigue which can be symptom of the disease progression but also the common adverse effects of radiotherapy or chemotherapy was frequently reported to be associated with depression (Bower et al., 2000;2011;Den Oudsten et al., 2009;Vahdaninia et al., 2010).In addition, mastectomy can cause scars and disfigurement which can lead to perception of becoming less attractive, less feminine and poor body image which have been associated with depression (Begovic-Juhant et al., 2012).Alcalar et al. (2012) reported the association between negative automatic thoughts and cognitive errors with depression in women in Turkey who had completed breast cancer treatment.Helplessness/hopelessness was more often expressed in severely depressed cohort that can lead to negative effect on coping with cancer (Schou et al., 2004).
There were several limitations to our review of these studies.While we attempted to minimize publication bias by searching multiple databases through electronic and hand search, we limited our analysis to studies in English and did not look for the unpublished papers.However the studies were from various parts of the regions including Asia-pacific countries though more numbers representing United States.There may be some cultural diversity within this region but we did not intend to examine on cultural issues.This is a systematic review of observational studies; therefore we could not ensure that all potential confounders were controlled for in those studies.
In conclusion, despite these limitations, the data suggest that breast cancer survivors are at risk for depression and identification of related factors (socio-demographic, illness, lifestyles, treatment and psychological) is important for practicing clinicians during routine psychooncology care.There may be a need for pharmacotherapy (Ng et al., 2011) and/or psychological interventions (Zainal et al., 2012).
For future research, these data implied that authors should attempt to use consistent screening tools so that cross-study comparisons evaluate similar definition of depression.In addition, more studies should include diagnostic assessments for depression.We hope the data from this review will help mental health professionals and hospital administrators to consider the clinical settings to effectively treat the psychological problems of cancer patients and to define future research priorities and sustain evidence-based psychosocial care for breast cancer survivors.

Figure 1 .
Figure 1.The Flow Diagram of Study Selection Figure 1.The flow diagram of study selection

Table 1 . Descriptive Characteristics of the Included Studies (N=32)
*SCID: Structured Interview for DSM (Diagnostic Statistical Manual); DASS: Depression, Anxiety and Distress Scale; MINI: Mini International Neuropsychiatric; CES-D: Center for Epidemiological Studies -Depression; BDI: Beck Depression Inventory Interview; HADS: Hospital Anxiety and Depression Scale; HDI: Hamilton Depression Inventory; SADS: Schedule for Affective Disorders and Schizophrenia

Table 2 . Prevalence of Depression among Breast Cancer Survivors and its Associated Factors
*SCID: Structured Interview for DSM (Diagnostic Statistical Manual), DASS: Depression, Anxiety and Distress Scale, MINI: Mini International Neuropsychiatric Interview, CES-D: Center for Epidemiological Studies -Depression, BDI: Beck Depression Inventory, SADS: Schedule for Affective Disorders and Schizophrenia, HADS: Hospital Anxiety and Depression Scale, HDI: Hamilton Depression Inventory