Distinct trajectories of emotional distress and physical health among informal dementia caregivers surrounding bereavement: a systematic review

Abstract

Background

Informal caregivers of individuals with dementia experience substantial psychological and physical burdens, with bereavement marking a critical transition. However, the longitudinal trajectories of their emotional responses remain poorly understood within conventional psychiatric frameworks. This study aimed to systematically synthesize longitudinal evidence on the emotional changes experienced by informal dementia caregivers before and after bereavement.

Methods

A comprehensive search of MEDLINE and EMBASE was conducted to identify prospective observational studies assessing emotional changes in informal dementia caregivers. Eligible studies included at least two assessment points (one pre-bereavement) and a minimum follow-up of six months. The quality of the included studies was evaluated using the modified Newcastle-Ottawa Scale.

Results

The analysis of seven longitudinal studies confirmed that caregivers’ experiences follow complex, multidimensional trajectories. Emotional outcomes were mixed: complicated grief tended to alleviate, depression/anxiety were heterogeneous, and loneliness tended to worsen. Notably, a trend of increased long-term depression was observed in women. The divergent trajectories between psychological recovery and physical burden were evident: role burden and mental quality of life recovered, while physical quality of life and immune function declined. Protective factors included the cessation of caregiving burden and a pre-bereavement social network.

Conclusion

These findings underscore that emotional adaptation is a multidimensional phenomenon requiring support strategies tailored to timing, gender, and individual resources. Strengthening protective factors in the pre-bereavement stage and addressing persistent physical burden in the long term are critical for effective preventive management.

Supplementary Information

The online version contains supplementary material available at 10.1186/s40359-026-03963-2.

Introduction

Informal caregivers are individuals who provide physical and emotional support to others based on personal relationships, such as family, friends, or neighbors, without professional training or financial compensation. In the United States, approximately 7 million individuals live with dementia, supported by more than 11 million informal caregivers [1]. Similarly, in the United Kingdom, where nearly one million people live with dementia, approximately 82% of those residing in the community rely on unpaid care from relatives or friends [2]. These figures underscore the critical role of informal care in the nations where most longitudinal evidence for this review originates.

The demands of informal dementia care are profound, often necessitating up to 24-hour daily assistance [3]. This relentless responsibility frequently leads to chronic exhaustion and a severe lack of personal time [4], while the underlying multidimensional burdens, including economic and role strains, increase a caregiver’s general susceptibility to health-related problems [5, 6]. Specifically, when the burden of caregiving exceeds an individual’s resources and capacity, caregivers can face severe consequences, including deterioration of mental and physical health, financial hardship, and psychological distress such as anxiety and depression [7]. In this study, emotional distress is conceptualized as a multidimensional state of psychological suffering encompassing depressive symptoms, anxiety, and subjective burden [8, 9]. This definition is particularly salient as the number of individuals with severe dementia, requiring the most intensive assistance, is projected to rise by 199% by 2040 [2], inevitably escalating the psychological toll on family carers. The emotional and physical responses arising from these complex challenges exhibit unique characteristics that distinguish them from conventional psychiatric classifications. Consequently, the academic community has termed this condition ‘caregiver stress syndrome,’ recognizing it as a distinct clinical entity [7].

Dementia has emerged as a major global public health concern, and as the number of patients continues to rise, the importance of informal care is growing [10]. Dementia is characterized by a progressive decline in cognitive function and unpredictable behavioral and psychological symptoms of dementia. These characteristics significantly intensify the emotional and physical stress on caregivers compared to those caring for individuals with other diseases [11]. Given the progressive nature of dementia, most caregivers inevitably experience bereavement, which marks a critical transition signifying the end of their long-term caregiving role [12].

Previous research indicates that the emotional responses of caregivers post-bereavement do not follow a uniform trajectory. While some experience persistent or worsening psychological distress, such as depression and anxiety, others report contrasting outcomes, including a sense of relief and psychological recovery following the cessation of the caregiving burden [13]. Despite the recognition that the trajectories of emotional change surrounding bereavement are complex and heterogeneous, a comprehensive and systematic understanding of them is still lacking.

The heterogeneous emotional trajectories experienced by caregivers after bereavement often do not fit neatly into existing psychiatric diagnostic classifications, highlighting the need for an alternative conceptual framework [14]. From a preventive medicine perspective, there is growing interest in ‘sub-health’ states—conditions not classified as distinct diseases but characterized by declining physical and psychological function [15]. The complex state of caregivers who have endured long-term caregiving and bereavement can therefore be viewed as a critical target for preventive health management, warranting proactive care before the onset of formal illness [16].

Therefore, this study aims to systematically review longitudinal studies on the emotional changes experienced by informal caregivers of individuals with dementia before and after bereavement. The goal is to elucidate the complex trajectories of these emotional changes and to establish an empirical basis for recognizing the post-bereavement state of caregivers as a condition requiring preventive and individualized management.

Methods

This systematic review was conducted in accordance with the Preferred Reporting Items for Systematic Reviews and Meta-Analyses 2020 guidelines [17] (Supplementary File 1). The study protocol was prospectively registered with the Open Science Framework (DOI: 10.17605/OSF.IO/K8G4S).

Search strategies

A comprehensive literature search was conducted using the MEDLINE (via PubMed) and EMBASE (via Ovid) databases to identify studies concerning the emotional changes in informal dementia caregivers who had experienced bereavement. The search strategy was developed by combining keywords with controlled vocabularies such as MeSH and Emtree. The final search was conducted on July 1, 2025, with no restrictions on language or publication date. To ensure maximum comprehensiveness and to identify any studies potentially missed by electronic database searches, we also performed manual citation tracking (snowballing) of all included articles. The entire search process was performed independently by two researchers (CYK, JWK), and the detailed search strategies for each database are presented in Supplementary File 2.

Inclusion criteria

The inclusion criteria for this review were as follows. Eligible studies were limited to prospective observational studies that targeted informal caregivers of patients with dementia and measured changes in their emotional distress using quantitative tools. Furthermore, included studies were required to have at least two assessment points, with at least one conducted pre-bereavement, and a minimum total follow-up period of six months. Conversely, studies were excluded if they were intervention, qualitative, cross-sectional, or retrospective in design. Editorials, case reports, letters, and conference abstracts without an available full text were also excluded. In the case of duplicate publications from the same study, only the most recent or comprehensive report was included.

Study selection

Two researchers (JWK, JWO) independently conducted the study selection process. In the first stage, titles and abstracts were screened to exclude clearly irrelevant studies. In the second stage, the full texts of the articles that passed the initial screening were reviewed to determine final eligibility. Any disagreements throughout the selection process were resolved through discussion; if a consensus could not be reached, a third researcher (CYK) was consulted for a final decision. All bibliographic data were managed using EndNote 20 (Clarivate Analytics, London, UK).

Data extraction

Data were extracted from the included studies using a predefined data extraction form. Four researchers (JWK, JWO, BMK, JEC) independently extracted the data, which were then cross-checked for accuracy. Any disagreements during the extraction process were resolved through discussion; if a consensus could not be reached, a third researcher (CYK) mediated the final decision. When data were unclear or missing, the corresponding authors of the respective studies were contacted for clarification. The extracted data included: (1) general study information (e.g., author, publication year, country, study design); (2) participant characteristics (e.g., age, sex, relationship to the patient, duration of care); (3) follow-up duration; (4) outcome measures; and (5) funding information.

The primary outcome of this review was caregiver emotional distress (e.g., depression, anxiety). The secondary outcomes were defined as caregiver burden, quality of life (QOL), complicated grief, resilience, risk and protective factors, physical changes, and qualitative descriptions of emotional distress.

Risk of bias assessment

The quality of the included studies was assessed using the modified Newcastle-Ottawa Scale (mNOS) [18]. The mNOS is a tool for evaluating the risk of bias in observational studies across three domains: selection of the study groups, comparability of the groups, and ascertainment of the outcome. Four researchers (JWK, JWO, BMK, JEC) independently assessed the quality of each study. Any disagreements were resolved through discussion to reach a consensus. If a consensus could not be reached, a third researcher (CYK) was consulted to make the final decision.

Data analysis

The data from the included studies were synthesized qualitatively. Although a meta-analysis was planned in the pre-registered study protocol if two or more studies used the same assessment tool, it was deemed inappropriate due to significant heterogeneity in outcome measures and assessment timepoints across the studies. Therefore, a meta-analysis was not performed; instead, a narrative synthesis of the findings was conducted.

Results

Search results

The initial database search yielded a total of 4,746 records. After removing duplicates, 3,756 records were screened based on their titles and abstracts, leading to the exclusion of 3,748 records. The full texts of the remaining eight articles were then assessed for eligibility. Following this review, one study was excluded due to the potential inclusion of intervention effects [19], resulting in a final selection of seven studies for this review [20–26]. The complete study selection process is illustrated in the flow diagram in Fig. 1.

Fig. 1.

Flow diagram of study selection process

Characteristics of included studies

The seven included studies, published between 2002 and 2024, utilized a prospective longitudinal observational design with follow-up periods of 8 months to 5 years (median: 34 months). Most studies (six) were conducted in the United States [20, 21, 23–26]. Caregiver demographics reflected typical trends: participants were predominantly older adults (average age: 71.6 years), female (average 73.4%), and largely Caucasian (average 86.6%). Care recipients generally had moderate-to-severe dementia, as evidenced by assessment scales like the Clinical Dementia Rating [21, 23] and Functional Assessment Staging [22]. Studies utilized a wide range of outcome measures across several domains, including negative emotions (e.g., complicated grief, depression, loneliness), caregiver burden, QOL (Mental and Physical components), psychosocial factors, and both objective (e.g., immune function, blood pressure) and subjective (e.g., self-rated health) physical health indicators (Tables 1 and 2).

Table 1.

Characteristics of included studies

Study	Country	Sample Size	Population	Duration of Care	F/U Duration	Outcome Measure	Analysis Method	COI
Mausbach 2007	US	126	Informal caregiver (Spouse)	NR (but CDR 2.6 ± 0.9 indicates advanced stage)	5 yrs	HAM-D; PROS; Seven-item personal mastery scale (developed by Pearlin and Schooler); Plasma D-dimer	Random regression model	NR
Tweedy 2007	US	87	Informal caregiver (Spouse)	NR (but patients had significant cognitive impairment)	4 yrs	CES-D (Eight-item version)	Within-subjects (Repeated measures) ANCOVA	NR
Moore 2017	UK	35	Informal caregiver (Family or Friend)	NR (but FAST stage ≥ 6E indicates advanced dementia)	16 mon ≥a	HADS; SF-12; ICG; SWC-EOLD; In-depth interviews	Proportion analysis; Descriptive statistics	No COI
Grant 2002	US	53b	Informal caregiver (Spouse)	104.6 ± 5.9 months	8 monc	HAM-D; Medical symptoms (Questionnaire based on the review of systems); BP	Repeated measures ANOVA	NR
Kotwal 2024	US	2098d	Informal caregiver (Spouse)	NR (long-term HRS cohort study)	4 yrs	Loneliness (Single-item loneliness question; 3-item UCLA Loneliness Scale for sensitivity analysis); CES-D (Eight-item version)	Multivariable logistic regression	NR
Wilson 2020	US	52	Informal caregiver (Spouse)	7.4 ± 3.1 years	7 yrs	BDI; SNI; Self-rated health; Lymphocyte proliferation to Con A; Lymphocyte proliferation to PHA	Piecewise linear mixed models	NR
Aneshensel 2004	US	291	Informal caregiver (Spouse or Adult child)	4.28–5.75 years (across subgroups)	5 yrs	HSC	Latent class mixture model	NR

Abbreviations: ANCOVA  Analysis of Covariance, BDI  Beck Depression Inventory, CDR  Clinical Dementia Rating, CES-D  Center for Epidemiologic Studies Depression Scale, CIND  Cognitive Impairment No Dementia, COI  Conflict of Interest, FAST  Functional Assessment Staging Test, F/U  Follow-up, HADS  Hospital Anxiety and Depression Scale, HAM-D  Hamilton Depression Rating Scale, HRS  Health and Retirement Study, HSC  Hopkins Symptom Checklist, ICG  Inventory of Complicated Grief, PHA  Phytohemagglutinin, NR  Not Reported, PROS  Pearlin’s Role Overload Scale, SF-12  12-Item Short-Form Health Survey, SNI  Social Network Inventory, SWC-EOLD  Self-Willingness to Care at End-of-Life for a Person with Dementia, UCLA  University of California, Los Angeles, UK  United Kingdom, US  United States

^aPre-bereavement: participants were followed for up to 9 months until the death of the care recipient

^bThe original study reported several subgroups; for this review, only the two relevant groups were combined and included in the table

^cThe original study reported follow-up durations for several subgroups; for this review, the duration was calculated as the mean of the two groups of interest

^dTotal sample size reported in the study; calculated as the sum of caregivers of normal (57%), CIND (25%), and dementia groups (18%)

Table 2.

Overall sample characteristics

Study	Age	Sex (M/F)	Ethnicity (n)	Patient Dementia Type	Patient Dementia Severity
Mausbach 2007	73.0 ± 8.7	38/88	Caucasian: 93.65% (118); Hispanic: 2.38% (3); Asian: 2.38% (3); African American: 0.8% (1); Other: 0.8% (1)	Alzheimer’s disease	CDR (0–3): 2.6 ± 0.9
Tweedy 2007	79.8 ± 6.8	20/67	Caucasian: 85.1% (74); Hispanic: 2.3% (2); African American: 12.6% (11)	Unspecified dementia	Composite cognition score (0–35): 12 ≥b; Proxy interview (Excellent-poor (5 stg)): fair, poora
Moore 2017	62	11/24	NR	Unspecified dementia	FASTc ≥ 6E
Grant 2002	NAe	17/36f	NR	Alzheimer’s disease	CDR: Mild: 6% (3/53) Moderate: 38% (20/53) Severe: 57% (30/53)
Kotwal 2024	72.3 ± 9.3	616/1482	White: 78.7% (1650); Black / African American: 10.5% (220); Latinx: 8.6% (180); Other: 2.3% (48)	Unspecified dementia	Composite cognition score (0–27): 6≥; proxy interview (0–11): 6–11d,a
Wilson 2020	71.3 ± 8.1	12/40	White 87% (45); Other 13% (7)	Unspecified dementia	BDS (NR)
Aneshensel 2004	Spousal caregiver: 71.36 ± 7.95; Adult child caregiver: 50.24 ± 8.60	132/159 (total sample)	Non-Hispanic white: 88.66% (258); Other: 11.34% (33)	Alzheimer’s disease	Cognitive difficulty (0–5): 2.82–3.14 by subgroups

Abbreviations: BDS Blessed Dementia Scale, CDR Clinical Dementia Rating, CIND Cognitive Impairment No Dementia, FAST Functional Assessment Staging Test, IADL Instrumental Activities of Daily Living, NA Not Applicable, NR Not Reported

^aWhen the patient was unable to participate

^bWord recall, serial sevens (100–7), backward counting, date, naming leaders, object naming

^ca Functional Assessment Staging Test 6e and above (e.g. doubly incontinent, loss of ability to speak more than six words, ambulatory ability lost or can’t hold up head independently)

^dProxy respondents rated memory (0–4) and IADL limitations (0–5); study informants rated interview difficulty due to cognitive impairment (0–2)

^eSubgroup ages presented in Supplementary Material 3; main table does not show combined value

^fThe original study reported several subgroups; for this review, only the two relevant groups were combined and included in the table

Risk of bias assessment

In the Selection domain, most studies used representative samples [20, 23–26], and two studies [22, 23] included a non-bereaved comparison group. Ascertainment of bereavement was conducted via reliable methods such as structured interviews in four studies [20, 23, 25, 26], while three relied on self-report [21, 22, 24]. Regarding Comparability, all studies controlled for basic confounders like age and sex, and five [15, 16, 18–20] enhanced their quality by additionally adjusting for caregiving-related variables. In the Outcome domain, a key strength was the inclusion of a pre-bereavement baseline and the use of validated tools across all studies. However, while five studies [20, 22, 23, 25, 26] had low attrition or provided clear attrition data, two studies [21, 24] were limited by insufficient reporting on this aspect. Based on this comprehensive assessment, five studies [20, 22, 23, 25, 26] were ultimately rated as ‘high quality’ and two [21, 24] as ‘medium quality,’ with detailed results available in Table 3.

Table 3.

The results of quality assessment among the included studies (Modified Newcastle-Ottawa Scale)

Study	Selection				Comparability	Outcome			Total	Quality
	Representativeness	Selection exposed cohort	Ascertainment	Result not present at start of the study	Comparability for confounders	Assessment of outcome	Follow-up duration	Adequacy of follow up
Mausbach 2007	0	0	0	1	1	1	1	0	4	Moderate
Tweedy 2007	1	0	1	1	1	1	1	1	7	High
Moore 2017	1	1	0	1	1	1	1	1	7	High
Grant 2002	1	1	1	1	1	1	1	1	8	High
Kotwal 2024	1	0	0	1	1	1	1	0	5	Moderate
Wilson 2020	1	0	1	1	1	1	1	1	7	High
Aneshensel 2004	1	0	1	1	1	1	1	1	7	High

Changes in emotional distress in caregivers

Complicated grief

Only one included study, Moore et al. (2017), measured complicated grief. In this study, grief symptoms tended to alleviate over time following bereavement [22]. The proportion of caregivers with clinically significant grief symptoms (Inventory of Complicated Grief [ICG] score ≥ 32) decreased from 38% at baseline to 22% at seven months post-bereavement [22]. Similarly, the median ICG score, an indicator of symptom severity, also decreased from 27 to 23 during the same period [22].

Negative emotions

The changes in caregivers’ negative emotions after bereavement showed complex patterns rather than a consistent trend. Several studies reported that symptoms of depression and anxiety tended to decrease over the long term. Grant et al. (2002) reported a statistically significant decrease in depression scores (Hamilton Depression Rating Scale) post-bereavement (p-value not presented) [23], and Moore et al. (2017) found that the proportion of caregivers with clinically significant depression and anxiety decreased seven months after bereavement [22]. Two other studies also observed a trend of decreasing depression scores, although the changes did not reach statistical significance (p = 0.05 [21] and p = 0.085 [25], respectively).

However, conflicting findings were also reported. Notably, a study by Tweedy et al. (2007) found that depression (the Center for Epidemiologic Studies-Depression Scale) significantly increased in the short term immediately following bereavement (p < 0.001) [20]. Furthermore, in a long-term follow-up of over two years, this study noted a significant gender difference (p = 0.01), with depression decreasing in male caregivers but increasing in females [20]. Another study found no statistically significant change (p = 0.18) in the proportion of caregivers with high depressive symptoms before and after bereavement [24].

In contrast to depression, loneliness tended to worsen after bereavement. In the study by Kotwal et al. (2024), the proportion of caregivers reporting loneliness increased from 26% pre-bereavement to 48% post-bereavement [24]. Conversely, a positive psychological change was also observed; in a study by Mausbach et al. (2007), caregivers’ sense of mastery significantly increased in the first six months post-bereavement (p < 0.001) [21].

The complexity of these emotional trajectories was clearly demonstrated in a study by Aneshensel et al. (2004), which classified caregivers into four heterogeneous groups [26]: (1) the Repeatedly asymptomatic group; (2) the Temporarily distressed group (sharp increase followed by significant recovery, p < 0.001); (3) the Repeatedly symptomatic group (persistent moderate symptoms with a slight decrease, p < 0.001); and (4) the Repeatedly distressed group (persistent or increased high symptoms, p < 0.001). These findings comprehensively demonstrate that the post-bereavement emotional response is not a uniform process but a varied phenomenon dependent on individual and contextual factors.

Caregiving-related outcomes

Indicators directly related to the caregiving role showed distinct changes following bereavement. According to a study by Mausbach et al. (2007), caregivers’ perceived role overload significantly decreased during the first six months post-bereavement (p < 0.001) [21]. This trend of decreased burden continued over the long term, with male caregivers reporting significantly lower role burden than female caregivers (p = 0.001) [21]. In contrast, satisfaction with end-of-life care (the End-of-Life in Dementia-Satisfaction With Care scale [SWC-EOLD]), as evaluated by Moore et al. (2017), remained stable before and after bereavement [22]. The median satisfaction score showed no significant change, moving from 30 at baseline to 31 at two months post-bereavement [22].

Quality of life

Caregivers’ QOL post-bereavement showed opposing trajectories for its physical and mental components [22]. The physical component score of the Short Form 12-item Health Survey (SF-12) showed a continuous decline, decreasing from 50.32 at baseline to 46.6 at seven months post-bereavement [22]. In contrast, the mental component score of the SF-12, after an initial dip from 47.57 pre-bereavement to 44.88 at two months post-bereavement, significantly rebounded to 52.52 by the seven-month mark [22]. These findings suggest that while bereavement may have a long-term positive impact on mental health, its negative effects on physical health may persist.

Physical changes in caregivers

Bereavement had complex effects on caregivers’ physical health, with objective physiological indicators and subjective health perceptions showing opposing trends.

Immune and cardiovascular indicators

Caregivers’ cell-mediated immune function significantly declined post-bereavement. In a study by Wilson et al. (2020), lymphocyte proliferation responses decreased significantly after the death of the care recipient (p = 0.014 and p = 0.009, respectively) [25]. Cardiovascular indicators showed conflicting results. While systolic blood pressure significantly increased post-bereavement in one study (p = 0.02) [23], another study found that plasma D-dimer levels, a risk marker related to blood clotting, significantly decreased in the post-bereavement period (p < 0.05) [21].

Subjective health perception

In stark contrast to the decline in some objective indicators, subjective health perceptions markedly improved. In the study by Wilson et al. (2020), self-rated health scores showed a highly significant improvement of approximately 33% post-bereavement (p < 0.001) [25]. Similarly, another study by Grant et al. (2002) reported a 30% reduction in self-reported medical symptoms [23].

Risk and protective factors

The included studies reported various risk and protective factors that influence the emotional trajectories of caregivers after bereavement. These factors could be categorized into social-relational and personal-internal factors (Fig. 2).

Fig. 2.

Risk and protective factor. Abbreviations. EOL = End of Life; HPD = Home-Placed-Deceased

Social-relational factors

A key protective factor was the size of the social network before bereavement. According to Wilson et al. (2020), caregivers with larger pre-bereavement social networks showed more pronounced health recovery and a slower decline in immune function post-bereavement [25]. Conversely, a poor quality of the end-of-life care experience acted as a risk factor. Qualitative interviews in the study by Moore et al. (2017) indicated that feelings of helplessness or guilt about the quality of end-of-life care were associated with negative post-bereavement emotions [22].

Personal-internal factors

Among protective factors, the relief from the burden of caregiving itself was the most direct. Grant et al. (2002) reported that the conclusion of chronic caregiving stress contributed to the alleviation of depression [23]. Furthermore, Aneshensel et al. (2004) found that being male, having a higher level of education, higher self-esteem, and abundant psychosocial resources were associated with resilient trajectories [26]. The same study identified risk factors for negative post-bereavement outcomes, including low income, older age, pre-existing depressive symptoms, and a sense of losing one’s self during the caregiving process [26].

Discussion

This systematic review analyzed seven longitudinal observational studies to comprehensively explore the complex emotional changes in informal dementia caregivers. Caregiver demographics reflected typical trends, with participants being predominantly older adults (mean age: 71.6 years) and female, as dementia is a disease of old age [27]. Furthermore, the severity of dementia in the care recipients was predominantly in the moderate-to-severe range [20–24, 26]. This clinical focus is significant as these findings primarily reflect the experiences of caregivers managing the escalating care demands typical of this stage, underscoring the importance of our results [28].

A key finding is that caregivers’ emotional changes post-bereavement follow complex, multidimensional patterns rather than a single trajectory. Specifically, complicated grief tended to decrease after bereavement [22]. Symptoms of depression and anxiety, which frequently co-occur due to shared physiological mechanisms like hypothalamic-pituitary-adrenal (HPA) axis hyperactivity [29, 30] and negative future expectancies [31], initially increased as a normal reaction to shock [22, 26] but tended to alleviate in the long term as caregivers adapted [22]. In sharp contrast, loneliness tended to worsen [24]. This suggests that the relief from the caregiving role may concurrently lead to a loss of purpose or social isolation and feelings of emptiness [32]. A simultaneous positive psychological shift occurred, as caregivers’ sense of mastery significantly increased, likely due to the ability to reinvest personal resources [21, 32].

Contrasting results emerged for caregiving-related indicators: While role burden significantly decreased with the cessation of caregiving tasks [21, 33], satisfaction with end-of-life care remained stable before and after bereavement [22]. This divergence is likely because satisfaction is a subjective, retrospective evaluation of past experiences, and thus less affected by the death event itself. However, a limitation is that the stability of the satisfaction scale (SWC-EOLD) may reflect a fixed perception of the ‘quality’ of past care rather than sensitively capturing the ‘change’ in a caregiver’s internal state post-bereavement [34].

QOL exhibited opposing trajectories in the mental and physical domains. Mental QOL recovered significantly in the long term, exceeding the general population average [22]. This improvement is interpreted as a combined result of reduced caregiver burden, alleviated negative emotions, and increased mastery [32]. In contrast, physical QOL showed a declining trend in the long term [22]. This decline is supported by the deterioration of physiological indicators, including significantly impaired cell-mediated immune function [25] and a tendency for systolic blood pressure to increase [23]. These physical changes may reflect somatization, where chronic emotional stress manifests as physical symptoms [35].

Meanwhile, cardiovascular indicators showed conflicting results because they reflect stress responses on different timelines. The rise in blood pressure [23] can be interpreted as an acute psychological stress response to the bereavement event itself [36]. In contrast, the decrease in plasma D-dimer [21] may reflect a long-term physiological stabilization resulting from the ‘relief’ from chronic physical burden and inflammatory stress that had persisted during caregiving [37]. In fact, inflammatory markers also tended to worsen initially but became more gradual over time [25], suggesting that while physiological conditions may deteriorate shortly after bereavement, they may gradually stabilize in the long term.

An intriguing finding is the discrepancy between the overall deterioration of objective physiological indicators [22, 23, 25] and the improving trend in subjective self-reported health outcomes [23, 25]. This contrast suggests that self-report measures are heavily influenced by psychological states. Specifically, the improvement in mental QOL following burden relief may lead caregivers to perceive their physical health more positively than it actually is [38]. This finding, however, also suggests that the cumulative physical toll of caregiving continues to affect caregivers’ physiological state for a period, even if their subjective mood improves [23].

Gender differences are a noteworthy finding: post-bereavement depression decreased in male caregivers but increased in female caregivers [20]. This aligns with the finding that being male acts as a protective factor associated with lower burden and resilient trajectories [26]. These differences are rooted in sociocultural factors, as female caregivers often assume primary responsibility and invest more resources, leading to a greater sense of loss [39], while male caregivers are more likely to utilize external resources, enabling a quicker recovery [40]. Furthermore, these variations may be exacerbated by neurophysiological differences: women tend to exhibit greater and more prolonged activation of the HPA axis in response to stress, which may be disadvantageous in adapting to long-term caregiving stress [41].

Various protective factors contribute to emotional recovery, including relief from caregiving burden, higher education and self-esteem, abundant psychosocial resources, and a broad pre-bereavement social network [22, 23, 25, 26]. The pre-bereavement network was more critical, acting as a stable buffer throughout the process [42], whereas the post-bereavement network had limited impact [25]. This suggests adaptation depends heavily on internal resources and the ability to reconstruct the loss experience, rather than the quantitative amount of social support [43]. Conversely, risk factors identified included poor quality end-of-life care, low income, older age, pre-existing depressive symptoms, and the experience of self-loss [22, 26]. Poor end-of-life care induces guilt [44]; low financial and physical resources exacerbate burden [45]; and pre-existing emotional vulnerability slows the recovery trajectory [46].

These findings emphasize that the emotional experience of dementia caregivers is a multidimensional trajectory, not simple ‘burden relief,’ requiring time-specific, tailored management. Caregivers’ responses vary across the pre-bereavement, immediate post-bereavement, and long-term phases. Interventions must be phase-appropriate: pre-bereavement support should focus on burden relief and education; immediate post-bereavement support should address the temporary spikes in distress and long-term strategies must address persistent physical stress despite mental recovery through comprehensive psycho-physical programs. Given the caregivers’ critical role [47] and frequent self-neglect [48], strengthening protective factors in the pre-bereavement stage is paramount to minimizing post-loss distress, particularly for this specialized group facing accumulated physical and emotional burden [13, 49].

Although the included studies were generally assessed as medium to high quality, supporting the reliability of the findings, this review has several limitations. First, the limited geographical scope—most studies were conducted in Western countries—restricts the generalizability of findings regarding cultural and ethnic differences. Second, the study population was heavily concentrated on spousal caregivers, limiting the exploration of differences across various family relationships. Third, the inherent scarcity of pre- and post-bereavement longitudinal studies and the lack of shared assessment tools across papers precluded quantitative synthesis through meta-analysis. Finally, our search strategy was limited to MEDLINE and Embase. Although these databases achieve a high overall indexing rate of 90.2% in health-related research [50], it is possible that some nursing-specific literature indexed in specialized databases like Cumulative Index to Nursing and Allied Health Literature was omitted.

Conclusion

This systematic review confirms that the emotional experiences of bereaved informal dementia caregivers follow complex, multidimensional trajectories rather than simple burden relief. Emotional outcomes were mixed: complicated grief tended to alleviate, depression and anxiety were heterogeneous, and loneliness tended to worsen. A notable finding was the trend of increased long-term depression in women. The divergent trajectories between psychological recovery and physical burden were evident: role burden decreased and mental QOL recovered, while physical QOL and immune function declined. Protective factors included the cessation of caregiving burden and a pre-bereavement social network, whereas risk factors included poor quality end-of-life care and pre-bereavement depression. These findings underscore the need for support strategies tailored to timing, gender, and individual resources. Future research should address limitations, such as conducting cross-cultural comparisons, and focus on developing novel assessment frameworks capable of explaining these complex, sub-clinical health changes not captured by conventional classifications.

Authors’ contributions

JK conceptualized the study and wrote the original draft. CK and JK developed the methodology. CK, JK, and JO conducted the investigation. Formal analysis was performed by CK, JK, JO, BK, and JC. CK, JK, and JO reviewed and edited the manuscript. CK provided supervision and acquired funding for the project. All authors read and approved the final manuscript.

Funding

This work was supported by the Institute of Information & Communications Technology Planning & Evaluation(IITP)-Innovative Human Resource Development for Local Intellectualization program grant funded by the Korea government (MSIT) (IITP-2026-RS-2020-II201791).

Data availability

Deidentified study data is available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

This study does not contain any studies with human or animal subjects performed by any of the authors. This study is a systematic review of existing literature and does not involve the recruitment of human participants or the collection of primary data.

Consent for publication

Not applicable.

Competing interest

The authors declare no competing interests.