Association between pain and behavioral and psychological symptoms of dementia (BPSD) in older adults with dementia: a systematic review and meta-analysis

Abstract

Background

Behavioral and Psychological Symptoms of Dementia (BPSD) are core and highly detrimental symptoms of dementia. Previous research has suggested a potential association between pain and BPSD, but pieces of evidence are lacking.

Objective

This study aimed to investigate the association between pain and BPSD in patients with dementia.

Methods

Seven databases were searched from inception to February 2024, including PubMed, Web of Science, Embase, CINAHL, Cochrane Library, China National Knowledge Infrastructure (CNKI), and WanFang Data. Cross-sectional and longitudinal studies were included. We included studies that involved older adults with dementia and assessed the pain and BPSD by using validated tools. The quality of cohort studies was assessed using the Newcastle–Ottawa Scale, and the quality of cross-sectional studies was evaluated using AHRQ criteria. Two researchers independently screened the articles, extracted the data, and assessed the quality of the studies. A meta-analysis was conducted using Stata 15.0. Data not suitable for meta-analysis was analyzed through a qualitative synthesis to provide a comprehensive overview of the findings.

Results

A total of 12 studies were included, comprising nine cross-sectional studies and three longitudinal studies. All studies were of medium to high quality. Studies reported that pain was associated with 13 different types of BPSD, including agitation, aggression, abnormal thought process, anxiety, care refusal, delirium, depression, delusions, hallucinations, sexual disinhibition, sleep disturbances, socially inappropriate behavior, and wandering. A pooled odds ratio (OR) of 1.25 (95%CI [1.17,1.33], p < 0.001) indicated a significant positive association between pain and BPSD. Specifically, pain was positively associated with aggression (OR = 1.07, 95%CI [1.00,1.13], p = 0.035), agitation (OR = 1.17, 95%CI [1.14, 1.21], p < 0.001), and depression (OR = 2.11, 95%CI [1.76,2.52], p < 0.001). However, pain was significantly negatively associated with wandering (OR = 0.77, 95%CI [0.73, 0.81], p < 0.001).

Conclusions

Pain was significantly positively associated with BPSD in patients with dementia, specifically with aggression, depression, and agitation. However, pain was negatively associated with wandering. This emphasizes the importance of further research in this area and improved interventions for pain and BPSD management.

Trial registration

www.crd.york.ac.uk CRD42023432320, registered 08/08/2024.

Supplementary Information

The online version contains supplementary material available at 10.1186/s12877-025-05719-w.

Background

Dementia is a clinical syndrome characterized by progressive cognitive decline that significantly impairs an individual's ability to perform activities of daily living (ADL), making it a major global health concern [1]. BPSD are among the core symptoms of dementia, encompassing a range of psychiatric symptoms, emotional symptoms, and behavioral symptoms such as delusions, depression, euphoria, agitation, anxiety, abnormal motor behavior, apathy, irritability, hallucinations, disinhibition, sleep disturbances, and changes in appetite [2]. These symptoms affect approximately 90% of dementia patients [3], placing a heavy burden on both patients and caregivers [4], and often leading to early institutionalization and diminished quality of life [5].

Research has demonstrated that BPSD has complex etiologies [6] and is influenced by numerous factors, including patient-related factors such as the severity of cognitive impairment [7] and their unmet needs [8], as well as environmental factors such as socioeconomic status [9] and caregiver interactions [10]. Among these factors, pain is considered to be one of the most important factors for BPSD [11]. Pain is a common issue among the elderly [12], with research indicating its high prevalence (range: 54.6–78.6%) across all subtypes of dementia [13]. Pain and cognitive impairment share similar pathological bases and are comorbid, both primarily involving brain regions such as the prefrontal cortex, hippocampus, and cingulate gyrus [14]. Experts in pain and geriatric medicine recommend considering BPSD as a potential indicator of pain [15]. Moreover, BPSD such as depression, agitation, and irritability, should not only be viewed as “pure” psychiatric sequelae of neurodegeneration but also as manifestations of unrecognized pain [16].

A cross-sectional study found a positive association between the level of pain and the occurrence of BPSD, and suggested the higher the intensity of pain, the greater the severity of BPSD, but it did not explore the association between pain and specific types of BPSD [17]. Only one systematic review explored the association between pain and sleep disturbances [18]. Conversely, one study found a correlation but no causal or directional association between pain and agitation, suggesting no significant association in terms of cause or effect [19]. Furthermore, given the possible association between pain and BPSD, researchers found that the use of analgesics had a moderate impact on reducing BPSD, suggesting that effective pain management could be crucial in improving BPSD [20]. Improvements were observed in the following types of BPSD: agitation, aggression, social behaviors, anxiety, depression, and aberrant motor behavior. However, one systematic review found the available studies did not support the hypothesis that pain management reduced agitation in patients with dementia [21]. The investigators also noted that the prevalence of delusions and hallucinations was not associated with the use of opioids [22].

Overall, previous studies have indicated potential association between pain and BPSD, but the results remain inconsistent, and several studies suffer from methodological flaws, such as small sample size, lack of longitudinal design, and limited exploration of pain with specific BPSD types. Therefore, we aimed to systematically evaluate the existing evidence on the association between pain and BPSD.

Methods

This systematic review was reported by following the Preferred Reporting Items for Systematic Reviews and Meta-Analysis (PRISMA) statement [23]. The protocol was registered in the PROSPERO database (CRD42023432320).

Objective

This study aimed to systematically review the existing evidence on the association between pain and BPSD in patients with dementia and analyze the findings.

Eligibility criteria

We employed the PICOS (Population, Intervention, Comparison, Outcome, and Study) framework [24] to define the eligibility criteria. The inclusion criteria were as follows: (1) P: Older adults with dementia (Individuals aged 60 years and above [25]). The diagnosis of dementia should be in line with the International Classification of Diseases (ICD-11) [26], the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) [27] or other recognized diagnostic criteria for dementia. The type of dementia was not restricted (including Alzheimer's disease, vascular dementia, mixed dementia, etc.); (2) I: Not applicable; (3) C: Not applicable; (4) O: Assessment results of pain and BPSD (obtained from recognized assessment tools or assessed by professionals), providing OR, risk ratio (RR), or hazard ratio (HR); (5) S: Cross-sectional and longitudinal studies published in English or Chinese.

The exclusion criteria were as follows: (1) the use of unreliable or unvalidated assessment tools for pain or BPSD; (2) unreliable data collection method: studies using unreliable or non-standardized data collection methods; (3) review articles or conference abstracts; (4) literature lacking full text.

Search strategy

Seven databases were searched from inception to February 2024, including PubMed, Web of Science, Embase, CINAHL, Cochrane Library, China National Knowledge Infrastructure (CNKI), and WanFang Data. The search terms included the use of Medical Subject Headings (MeSH) and free-text terms, and were combined with Boolean operators (OR/AND). Additionally, a manual search was conducted by reviewing the reference lists of the retrieved articles. Alerts for updated literature from these databases were set to receive monthly email notifications. The search strategy is shown in Table 1. The specific search results for each database are provided in the supplementary material, along with the search strategies.

Table 1.

Search strategy

Theme	Search Terms
#1 dementia	dement* OR alzheimer* OR cognitive impair* OR dementia OR AD OR VaD
#2 pain	pain OR hyperalgesia OR nociception OR neuralgia OR ache
#3 BPSD	BPSD OR behavioral and psychological symptoms OR eating disturbance OR agitation OR sleep disturbances OR apathy OR depression OR anxiety OR Irritability OR aberrant motor behaviour OR hallucinations OR delusions OR disinhibition OR euphoria OR aggression OR stereotyped behaviors OR wandering OR night-time behaviour disturbances OR dysphoria OR care refusal
#4	#1 AND #2 AND #3

Study selection

EndNote 20 was used to manage the study selection. Duplicates were first removed. Subsequently, two reviewers (two master students, both of whom have received evidence-based training) independently screened the remaining studies in two steps: (1) preliminary screening of titles and abstracts, and (2) secondary screening of full-text articles based on the eligibility criteria. Any disagreements were resolved through the involvement of a third researcher (a faculty member with a PhD degree), who evaluated the study and made the final decision.

Data extraction

Two independent reviewers employed a uniform data extraction template to gather information from the selected studies, and the third reviewer checked the accuracy. This template captured details like: author, year, country, sample size, study design, demographic information (age and gender), measurement methods for pain and BPSD, and reported outcomes (OR, HR, or RR, and 95% confidence intervals (CI)).

Risk of bias in individual studies

The risk of bias in the included studies was assessed by two reviewers independently, Cohort studies were evaluated using the Newcastle–Ottawa Scale (NOS). NOS includes three dimensions with eight items: selection of study subjects (four items), comparability between groups (one item), and measurement of outcomes or exposure factors (three items). Except for the comparability between groups, which is scored as 2, the rest of the items are scored as 1, with a total score ranging from 0 to 9 [28]. Cross-sectional studies were evaluated using quality assessment criteria recommended by the US Agency for Healthcare Research and Quality (AHRQ), which consist of 11 items, such as data sources, variables, time, samples, bias, statistical analysis, data collection, and follow-up. Each item is scored as 1, with a total score ranging from 0 to 11 [29]. Any disagreements were resolved by discussion with a third reviewer to reach a consensus.

Data analysis

We used OR and its 95%CI as the effect size and considered a P value less than 0.05 to be statistically significant. A meta-analysis was performed using Stata (version 15.0; Stata Corporation, College Station, TX, USA). The Q test (I²) was used to evaluate the heterogeneity between studies. If P ≥ 0.1 and I² ≤ 50%, it indicated no significant heterogeneity among the studies [30], a fixed effects model was chosen for the analysis. We used random effects models when significant heterogeneity was present (I² ≥ 50%). Given that each study may report associations between pain and more than one type of BPSD, we extracted and pooled ORs for the associations between pain and all types of BPSD reported in each cross-sectional study, as well as the ORs for the associations between pain and the same type of BPSD collected at different times in each longitudinal study. For associations between pain and specific types of BPSD, a meta-analysis was conducted only if three or more ORs could be extracted for that specific type. Conversion formulas [31, 32] were used to convert HR or RR to OR. A funnel plot was used to evaluate the publication bias.

Results

Study selection

Figure 1 presents the PRISMA flow-chart illustrating the process of study selection. In total, there were 3,855 articles retrieved, and 1,497 duplicates were excluded. There were 2,358 records excluded after screening the title and abstract, leaving 61 records for full-text screening. Finally, 12 articles were included.

Fig. 1.

PRISMA flow chart showing the study screening process

Study characteristics

The characteristics of the included 12 studies were summarized in Table 2. The studies were conducted between 2005 and 2021 in the USA (n = 5), Europe (n = 3), Norway (n = 1), Netherlands (n = 1), Italy (n = 1), and Canada (n = 1). Eleven studies were conducted in the nursing homes or long-term care facilities, and one study was conducted in the community. The sample sizes ranged from 155 [33] to 103,344 [34], with a total of 281,969 participants. The mean age ranged from 80.7 [35] to 86.3 [33] years old, and the proportion of females ranged from 65.9% [36] to 81% [37]. Three studies [38–40] were longitudinal, while the remaining nine [33–37, 41–44] were cross-sectional.

Table 2.

Characteristics of included studies

NO	Studies	Design	Setting	Sample size (female)	Mean age (SD)	Measurement methods (Pain)	Measurement methods (BPSD)	Outcomes
1	Ahn 2015 [41]	Cross-sectional study	nursing home, the USA	71,227 (73.96%)	85(7)	residents who could not communicate: MDS-PBS residents who could communicate: NRS and VRS	The behavioral items in the MDS	residents who could not communicate: pain (OR = 1.23, 95% CI [1.17, 1.29]) with verbal aggression pain (OR = 1.20, 95% CI [1.14, 1.26]) with physical aggression residents who could communicate: pain (OR = 1.12, 95% CI [1.04, 1.18]) with verbal aggression pain (OR = 1.05, 95% CI [0.96, 1.15]) with physical aggression
2	Tosato 2012 [43]	Cross-sectional study	nursing home, Europe	2822 (74.8%)	84.19(9.1)	interRAI LTCF assessment	behavioral and psychiatric symptoms in the SHELTER database	pain (OR = 0.74, 95% CI [0.55, 1.00]) with wandering pain (OR = 1.06,95%CI[0.80,1.41]) with verbal abuse pain (OR = 1.08,95%CI[0.75,1.55]) with physical abuse pain (OR = 1.37,95%CI[1.04,1.80]) with Socially inappropriate behavior pain (OR = 1.41,95%CI[1.08,1.83]) with care refusal pain (OR = 1.06, 95% CI [0.86, 1.32]) with 1 or more behavior symptoms pain (OR = 1.48,95%CI[1.16–1.90]) with Abnormal thought process pain (OR = 1.48, 95% CI [1.07, 2.03]) with delusions pain (OR = 1.37, 95% CI [0.90, 2.08]) with hallucinations pain (OR = 1.42, 95% CI [1.12, 1.80]) with 1 or more psychiatric symptoms
3	Erdal 2017 [40]	Cohort study	nursing home, Norway	931 (66.8%)	85.4(7)	Mobilisation-Observation-Behaviour-Intensity-Dementia-2 (MOBID-2) Pain Scale	the Cornell Scale for Depression in Dementia (CSDD)	pain (OR = 2.35, 95% CI [1.76, 3.12]) with depression
4	Hendriks 2015 [39]	longitudinal observational study	Long-term care facilities, the Netherlands	372 (70%)	84(7)	By physicians’ assessment		across the regular assessments: pain (OR = 1.2, 95% CI [0.95, 1.6]) with agitation across the last regular assessment and the after-death assessment: pain (OR = 1.4, 95% CI [0.86 2.3]) with agitation
5	Leonard 2006 [34]	cross-sectional study	nursing homes, the USA	103,344 (75.9%)	84	MDS item		Mild pain: (OR = 0.9, 95% CI [0.8, 1.0]) with physical aggression (OR = 1.1, 95% CI [1.0, 1.1]) with verbal aggression Moderate pain: (OR = 0.8, 95% CI [0.7, 0.9]) with physical aggression (OR = 1.1, 95% CI [1.0, 1.2]) with verbal aggression Severe pain: (OR = 0.93, 95% CI [0.6,1.3]) with physical aggression (OR = 1.0, 95% CI [0.7, 1.3]) with verbal aggression
6	Malara 2016 [35]	cross-sectional study	long term care facilities, Italy	181 (66.3%)	male:80.7 ± 9.3 Female:85.6 ± 7.3	Pain Assessment in Advanced Dementia Scale (PAINAD)	Cornell Scale for Depression in Dementia (CSDD) the Cohen-Mansfield Agitation Inventory (CMAI) and Neuropsychiatric Inventory (NPI)	pain (OR = 2.7129, 95% CI [1.333, 5.52]) with depression pain (OR = 3.2571, 95% CI [1.284, 8.2643]) with aggression
7	Mühler 2021 [44]	cross-sectional study	Long-Term Care Facilities, Europe	2822 (74.8%)	/	pain variables used in the SHELTER database	12 BPSD variables in interRAI LCTF	pain (OR = 0.81, 95% CI [0.60, 1.09]) with wandering pain (OR = 1.07, 95% CI [0.81, 1.42]) with verbal aggression pain(OR = 1.40,95%CI [1.08 – 1.81])with care refusal pain (OR = 1.08, 95% CI [0.76, 1.54]) with physical aggression pain (OR = 1.37,95%CI[1.05 – 1.78]) with Socially inappropriate behavior pain (OR = 1.22,95%CI[0.68–2.17]) with Sexual not inhibited pain (OR = 1.12, 95% CI [0.91, 1.38]) with 1 or more behavior symptoms pain (OR = 1.59, 95% CI [1.14, 2.11]) with delusions pain (OR = 1.34, 95% CI [0.89, 2.01]) with hallucinations pain (OR = 1.51,95%CI[1.18–1.92]) with Abnormal thought process pain (OR = 1.54, 95% CI [1.18, 2.01]) with sleeping disorder pain (OR = 2.23, 95% CI [1.80, 2.76]) with depression pain (OR = 1.49, 95% CI [1.16, 1.92]) with anxiety pain (OR = 1.87, 95% CI [1.53, 2.29]) with 1 or more psychiatric symptoms
8	Ahn 2013 [42]	cross-sectional study	nursing home, the USA	56,566 (67.7%)	84 (65–109)	MDS-Pain severity scale	The MDS-wandering item The MDS-Aggression Behavior Scale (MDS-ABS) The revised MDS-Challenging Behavior Profile (MDS-CBP) agitation subscale	pain (OR = 0 .77, 95% CI [0.73, 0.81]) with wandering behaviors pain (OR = 1.04, 95% CI [1.01, 1.08]) with aggression pain (OR = 1.17, 95% CI [1.13, 1.20]) with agitation
9	Gruber-Baldini 2005 [37]	cross-sectional study	nursing home, the USA	347 (81%)	84.5(7.1)	the Philadelphia Geriatric Center Pain Intensity Scale	the Cornell Scale for Depression in Dementia (CSDD)	pain (OR = 3.54, 95% CI [1.59, 7.85]) with depression
10	Giebel 2015 [36]	cross-sectional study	community and care home, Europe	414 (65.9%)	82(6.8)	items from the Minimum Data Set Residents Assessment Instrument (MDS/RAI)	the Cornell Scale for Depression in Dementia (CSDD)	pain (OR = 1.93, 95% CI [1.11, 3.41]) with depression
11	Voyer 2009 [33]	cross-sectional study	Long-Term Care Facilities, Canada	155 (73.6%)	86.3(6.9)	the DOLOPLUS-II	the confusion assessment methods (CAM)	pain (OR = 2.15, 95% CI [1.00, 4.59]) with delirium
12	Wei 2021	cohort study	Long-Term Care Facilities, the USA	Depression cohort (n = 27,131) Behavioral symptoms cohort (n = 15,657)	/	residents who were verbal: VRS/VRS residents who were verbal: CNPI	the Patient Health Questionnaire (PHQ)−9	pain (HR = 1.67, 95% CI [1.54, 1.81]) with depression pain (HR = 1.28, 95% CI [1.19, 1.37]) with behavioral symptoms

SD standard deviation

Nine methods were used to assess the pain of dementia patients, including the Checklist of Nonverbal Pain Indicator (CNPI) [38], the DOLOPLUS-II [33], Mobilisation-Observation-Behaviour-Intensity-Dementia-2 (MOBID-2) Pain Scale [40], Pain Assessment In Advanced Dementia Scale (PAINAD) [35], MDS-Pain severity scale [34, 41, 42, 44], the Philadelphia Geriatric Center Pain Intensity Scale [37], Numeric Rating Scale (NRS) or Verbal Rating Scale (VRS) [38, 41], interRAI LTCF assessment [43, 44], and physicians’ assessment [39]. Eight methods were used to assess BPSD, including the Cornell Scale for Depression in Dementia (CSDD) [35–37, 40], the Cohen-Mansfield Agitation Inventory (CMAI) [35], Neuropsychiatric Inventory (NPI) [35], the Confusion Assessment Methods (CAM) [33], the Patient Health Questionnaire (PHQ-9) [38], the behavioral items in the MDS [34, 41, 42], 12 BPSD variables in interRAI LCTF [43, 44], and physicians’ assessment [39].

Quality of included studies

The included cross-sectional studies were rated 6 to 9 points according to AHRQ criteria, and the included longitudinal studies were rated 8 points according to NOS, indicating that all studies were of medium to high quality (Tables 3, 4).

Table 3.

Critical appraisal of included cross-sectional studies

Studies	1	2	3	4	5	6	7	8	9	10	11	Score
Ahn 2015 [41]	Y	Y	Y	Y	U	N	Y	Y	N	N	N	6
Tosato 2012 [43]	Y	Y	Y	Y	U	Y	Y	Y	N	Y	N	8
Leonard 2006 [34]	Y	Y	Y	Y	U	N	Y	N	N	Y	N	6
Malara 2016 [35]	Y	Y	Y	Y	U	N	Y	N	N	Y	N	6
Mühler 2021 [44]	Y	Y	Y	Y	U	N	Y	Y	Y	Y	Y	9
Gruber-Baldini 2005 [37]	Y	Y	Y	Y	U	Y	Y	Y	N	Y	N	8
Giebel 2015 [36]	Y	Y	Y	Y	N	Y	Y	Y	N	Y	Y	9
Voyer 2009 [33]	Y	Y	Y	Y	U	N	Y	N	N	Y	N	6
Ahn 2013 [42]	Y	Y	Y	Y	U	N	Y	Y	N	Y	N	7

Agency for Healthcare Research and Quality, AHRQ: 1 = Define the source of information (survey, record review); 2 = List inclusion and exclusion criteria for exposed and unexposed subjects (cases and controls) or refer to previous publications; 3 = Indicate time period used for identifying patients; 4 = Indicate whether or not subjects were consecutive if not population-based; 5 = Indicate if evaluators of subjective components of study were masked to other aspects of the status of the participants; 6 = Describe any assessments undertaken for quality assurance purposes (e.g., test/retest of primary outcome measurements); 7 = Explain any patient exclusions from analysis; 8 = Describe how confounding was assessed and/or controlled; 9 = If applicable, explain how missing data were handled in the analysis; 10 = Summarize patient response rates and completeness of data collection; 11 = Clarify what follow-up, if any, was expected and the percentage of patients for which incomplete data or follow-up was obtained

Y yes, N no, U unclear

Table 4.

Critical appraisal of included cohort studies

Studies	Selection				Comparability	Outcome			Score
	1	2	3	4	5	6	7	8
Erdal 2017 [40]	1	0	1	1	2	1	1	1	8
Hendriks 2015 [39]	1	0	1	1	2	1	1	1	8
Wei 2021	1	0	1	1	2	1	1	1	8

the Newcastle–Ottawa Scale, NOS: 1 = Representativeness of the exposed cohort; 2 = Selection of the non-exposed cohort; 3 = Ascertainment of exposure; 4 = Demonstration that outcome of interest was not present at start of study; 5 = Comparability of cohorts on the basis of the design or analysis controlled for confounders; 6 = Assessment of outcome; 7 = Was follow-up long enough for outcomes to occur; 8 = Adequacy of follow-up of cohorts

Association between pain and BPSD among dementia patients

Cross-sectional studies [33–37, 41–44] reported associations between pain and all 13 types of BPSD, including aggression [34, 35, 41–44], depression [35–38, 40, 44], wandering [42–44], agitation [39, 42], socially inappropriate behavior [43, 44], care refusal [43, 44], abnormal thought process [43, 44], delusions [43, 44], hallucinations [43, 44] sexual not inhibited [43, 44], sleep disturbances [43, 44], anxiety [44], and delirium [33]. While longitudinal studies [38–40] mainly focused on the changes over time of associations between pain and agitation [39] and depression [38, 40]. Eleven studies reported ORs, while one study reported HR [24].

The meta-analysis of 12 studies [33–44] indicated that pain (OR = 1.25, 95%CI [1.17,1.33], p < 0.001) had a significant positive association with BPSD (Fig. 2). Specifically, the meta-analysis indicated that pain was significantly positively associated with aggression [34, 35, 41–44] (OR = 1.07, 95%CI [1.00,1.13], p = 0.035), depression [35–38, 40, 44] (OR = 2.11, 95%CI [1.76,2.52], p < 0.001), and agitation [39, 42] (OR = 1.17, 95%CI [1.14, 1.21], p < 0.001) (Figs. 3, 4, 5). However, the meta-analysis of three studies [42–44] indicated that pain was significantly negatively associated with wandering (OR = 0.77, 95%CI [0.73, 0.81], p < 0.001) (Fig. 6). A funnel plot was used to analyze the publication bias (Fig. 7). The funnel plot showed some asymmetry, which may suggest potential publication bias. We further discussed this issue in subsequent sections.

Fig. 2.

Forest plot of the association between pain and BPSD (all types)

Fig. 3.

Forest plot of the association between pain and aggression

Fig. 4.

Forest plot of the association between pain and depression

Fig. 5.

Forest plot of the association between pain and agitation

Fig. 6.

Forest plot of the association between pain and wandering

Fig. 7.

Funnel plot analysis of publication bias in the association between pain and BPSD

Data not suitable for meta-analysis was analyzed through a qualitative synthesis. Tosato [43] reported the lowest OR (0.74, 95% CI [0.55, 1.00]) for the association between pain and wandering, while Gruber-Baldini [37] reported the highest OR (3.54, 95% CI [1.59, 7.85]) for the association between pain and depression. Included studies reported that pain was positively associated with most types of BPSD, including aggression [34, 35, 41–44], depression [35–38, 40, 44], agitation [39, 42], socially inappropriate behavior [43, 44], care refusal [43, 44], abnormal thought process [43, 44], delusions [43, 44], hallucinations [43, 44] sexual not inhibited [43, 44], sleep disturbances [43, 44], anxiety [44], and delirium [33]. This is consistent with the results of our meta-analysis, suggesting that pain may exacerbate BPSD or serve as an underlying trigger. Two studies [34, 41] divided aggression into physical aggression and verbal aggression and reported the association between pain and them. However, one of the studies [34] found a positive association between pain and verbal aggression, while a negative association between pain and physical aggression. Additionally, only one study [41] investigated the association between pain and aggression in communicative versus non-communicative patients with dementia, and only one study [34] investigated the association between varying levels of pain and aggression.

Discussion

This systematic review revealed the significant association between pain and BPSD, with a particular focus on specific types of BPSD, such as aggression, agitation, depression, and wandering. At the same time, we found a negative association between pain and wandering. These findings are consistent with and extend previous research [13, 45].

Pain is significantly associated with depression [35–38, 40, 44]. This may be explained by chronic inflammation and neurobiological pathways [46], as well as the difficulties these patients face in expressing and managing their pain due to cognitive decline, which often leads to the overlook of pain and, consequently, exacerbates depression [47]. Additionally, the bidirectional association between pain and depression, known as the ‘pain-depression dyad’, has been identified in individuals without dementia [48, 49]. Studies have found that depression can increase chronic musculoskeletal pain in patients with Alzheimer's Disease (AD) [50], suggesting a possible bidirectional association between pain and depression.

We found that pain was positively associated with aggression [34, 35, 41–44] (OR = 1.07, 95% CI [1.00, 1.13], p = 0.035). However, the effect size was minimal, and the confidence interval only narrowly excluded no effect (1.00). This finding suggested a potential but weak association between pain and aggression. We considered some possible reasons. For instance, some patients with pain may become less mobile and show reduced aggressive behavior due to physical limitations, while others may react to pain with physical or verbal aggression [51]. This highlights the importance of considering clinical factors, such as patients' physical strength, energy levels, and whether bedridden, when applying these findings in practice. We also found that pain was positively associated with agitation [39, 42] (OR = 1.17, 95%CI [1.14, 1.21], p < 0.001). Pain causes discomfort and distress, and dementia patients may be unable to communicate effectively due to cognitive impairments [47]. This inability to express pain can result in frustration, which may manifest as agitation.

The negative association found between pain and wandering [42–44] is intriguing and needs further investigation. One possible explanation is that reduced wandering and other activities might act as self-soothing mechanisms or forms of physical coping strategies in response to pain [52]. A study found that higher pain scores were associated with less time spent on activities among patients with dementia (estimate 0.897, P = 0.043) [53]. Alternatively, it could have different underlying causes or may be influenced by factors other than pain alone.

From the perspective of biological mechanism, pain and BPSD may be interconnected through neurobiological pathways. Inflammatory responses in the brain are common in patients with dementia (particularly those with AD). The activation of microglia and astrocytes releases pro-inflammatory cytokines, such as IL-1β, IL-6, and TNF-α [54, 55]. These factors not only advance dementia pathology but may also influence pain perception and trigger BPSD. Additionally, patients with dementia often exhibit imbalances in neurotransmitters such as dopamine, serotonin, and glutamate, which are important for mood regulation and pain perception [56]. The phosphodiesterase (PDE) also plays a significant role in dementia pathology. PDEs regulate neuronal signal by specifically hydrolyzing cAMP and cGMP. PDE2, 4, 5, and 7 subtypes also have an impact on mood regulation and pain perception [57]. The decline in cholinergic function in dementia not only affects cognition but may also impact pain perception, leading to pain and behavioral changes in patients [58]. The association between pain and BPSD may also stem from dysregulation of the hypothalamic–pituitary–adrenal axis (HPA axis). Pain responses can trigger abnormal HPA axis activity, resulting in increased cortisol secretion, which may exacerbate anxiety, depression, and other behavioral symptoms in patients with dementia [59].

Given the association between pain and BPSD, we consider that effective pain control may alleviate some of the distressing BPSD, and advocate for integrating pain management into dementia care plans. Effective interventions should focus on timely and comprehensive pain assessment and individualized pain relief measures. Furthermore, pain management not only alleviates BPSD [60] but also significantly improves the patient's quality of life and reduce the caregiver's burden [61]. Caregivers, who have the most frequent contact with patients, should be vigilant when observing BPSD such as depression, agitation, and aggression, as these may indicate underlying pain. Consequently, dementia caregivers are advised to receive specialized training to improve their skills in assessing and managing pain in patients. Cognitive decline often impairs patients’ ability to effectively report their pain, leading to frequent under-recognition of pain [62]. To facilitate better pain management, a variety of non-verbal behavioral pain assessment scales(observational pain scale, OPS), electronic pain assessment tools [63, 64], and pain training programs [65–67] for dementia caregivers have been developed and used in the studies included in this reviews [33, 35, 38, 40]. In addition to pharmacological treatments, employing non-pharmacological pain management strategies is also necessary. Although medications can be effective in relieving pain, they usually have side effects, particularly in the elderly [68]. Non-pharmacological interventions, such as social robot [69], Tai Chi program [70], music therapy [71], and massage [72], may offer safer alternatives, addressing both the physiological and psychological aspects of pain.

Future research should broaden its scope to include Asian populations and hospitalized or home-based patients with dementia to generalize the findings. Additionally, incorporating cross-cultural and longitudinal approaches will lead to more effective care strategies [45] and provide a more comprehensive understanding of the association between pain and BPSD. Moreover, we advise to explore deeper mechanisms underlying the association between pain and BPSD through further induction experiments and animal studies, as well as assess the effectiveness of different pain management strategies in reducing BPSD. It is also important to explore whether these strategies can be customized based on individual characteristics, such as the type of dementia or the severity of cognitive impairment.

Limitations

This study offers valuable understanding into the association between pain and BPSD and guidance for dementia care and future research directions. However, several limitations should be acknowledged. First, differences in the assessment tools used for both pain and BPSD likely contributed to the variability. For example, the use of various pain assessment scales, ranging from self-reported measures to observational tools, may affect the consistency of reported associations. Second, since most studies were conducted in nursing homes or long-term care facilities in some European countries or the USA, there was a lack of studies focusing on patients with dementia in Asian populations and those who were hospitalized or home-based. Additionally, the funnel plot revealed some asymmetry, which may suggest potential publication bias. This asymmetry could be attributed to differences in study design or statistical methods as mentioned above, leading to inconsistent effect estimates.

It is worth noting that "neuropsychiatric symptoms" were not included in the search terms as we considered that neuropsychiatric syndrome refers to a broader range of behavioral and emotional symptoms across different stages of neurocognitive disorders. This includes mild behavioral impairment, which often precedes cognitive decline and increases the risk of developing dementia, while BPSD typically manifests at later stages. Lastly, our study aimed to investigate the association between pain and BPSD, specifically focusing on whether pain is a risk factor for BPSD. Considering the difficulties of obtaining original raw data from the studies reporting continuous variables and regression coefficients, we did not include these studies. However, we acknowledge that the use of continuous outcome measures and regression coefficients can offer valuable insights into the temporal or causal associations. This can be an interesting direction for future research.

Conclusions

This systematic review demonstrates a significant association between pain and BPSD. The results highlight the need to integrate effective pain management into dementia care to alleviate BPSD and enhance patients' quality of life.

Authors' contributions

Conceptualization, Y.D., Y.Y. and T.S.; methodology, Y.D., Y.Y.; software, T.S.; validation, T.S., Y.Y., Y.D., X.L., Y.X., Q.L., L.Z., H.J., K.Q. and S.S.; formal analysis, T.S. and Y.X.; writing—original draft preparation, Y.D., Y.Y. and T.S.; writing—review and editing, T.S., Y.Y., Y.D., X.L., Y.X., Q. L., L. Z., H.J., K.Q. and S.S.; supervision, Y.D. and Y. Y.; All authors have read and agreed to the published version of the manuscript.

Funding

This work was supported by Project of “Nursing Science” Funded by the 4th Priority Discipline Development Program of Jiangsu Higher Education Institutions (Jiangsu Education Department 〔2023〕No.11) and Natural Science Foundation of Jiangsu Province (BK20240527).

Data availability

Data sharing is not applicable to this article as no datasets were generated during the current study. All data supporting the findings of this study are derived from published literature.

Declarations

Ethics approval and consent to participate

Not applicable.

Consent for publication

All authors agree to publish this work.

Competing interests

The authors declare no competing interests.