Search
Close
Search
Advanced Search
Search Menu
AI Discovery Assistant

Abstract

Background

Previous observational studies have indicated a complex association between chronic pain and frailty. This study aimed to examine the bidirectional causal relationship between frailty and chronic pain and to quantify mediating effects of known modifiable risk factors.

Methods

A bidirectional two-sample Mendelian randomisation (MR) analysis was applied in this study. Summary genome-wide association statistics for frailty, as defined by both frailty index (FI) and Fried Frailty Score (FFS), pain at seven site-specific chronic pain (SSCP) (headache, facial, neck/shoulder, stomach/abdominal, back, hip and knee) and multisite chronic pain (MCP) were extracted from populations of European ancestry. Genetic instrumental variables strongly correlated with each exposure were selected. The inverse-variance-weighted method was the primary method used in the MR, supplemented by a range of sensitivity and validation analyses. Two-step MR analysis was undertaken to evaluate the mediating effects of several proposed confounders.

Results

Genetically predicted higher FI and FFS were associated with an increased risk of MCP and specific types of SSCP, including neck/shoulder pain, stomach/abdominal pain, back pain, hip pain and knee pain. In the reverse direction analysis, genetic liability to MCP was found to be associated with increased FI and FFS. These results remained consistent across sensitivity and validation assessments. Two-step MR suggested a mediating role for body mass index, smoking initiation, physical inactivity, educational attainment and depression.

Conclusions

Our research provided genetic evidence that the association between frailty and chronic pain was bidirectional where the coexistence of both conditions will exacerbate each other.

frailty, chronic pain, Mendelian randomisation study, mediation analysis, older people

Key Points

  • Genetically predicted frailty was positively associated with the risk of MCP.

  • Genetic liability to MCP was positively associated with higher FI and FFS.

  • There is a bidirectional causal relationship between frailty and chronic pain.

Introduction

Frailty, an emerging geriatric syndrome characterised by a decline in functioning across multiple physiological systems, presenting a substantial challenge to global healthcare systems [1]. The commonly used metric for assessing frailty status is the Frailty Index (FI) and the Fried Frailty Score (FFS), each capturing different aspects of this complex condition [2, 3]. Higher FI and FFS correlate with numerous adverse health consequences such as disability, mobility constraints, an array of chronic ailments, hospitalisation and increased mortality rates [3–6].

Chronic pain, manifesting in diverse forms, stands as another critical global health issue, impacting a notable proportion of the population [7]. The interrelation between chronic pain and frailty is complex and multi-dimensional, suggesting a potential mutual influence [8, 9]. Both conditions are predominant in older adults and exhibit overlapping challenges. Chronic pain could serve as a pivotal stressor in the ageing process, potentially instigating or hastening frailty [10]. Conversely, frailty might pose a risk factor affecting pain modulation through descending inhibitory pathways [11]. Shared biological mechanisms, such as inflammatory reactions, neural modifications and hormonal imbalances, might underlie both frailty and diverse chronic pain conditions [9]. Yet, the prevailing evidence concerning the intersection of frailty and chronic pain largely originates from observational studies [12–14]. Such studies can be susceptible to confounding biases and the pitfalls of reverse causation, thus complicating causal interpretations. The issue of which condition, frailty or chronic pain, occurs first or whether there is a bidirectional association is not well established. Furthermore, the ramifications of frailty on an extensive spectrum of chronic pain manifestations remain largely unexplored.

Mendelian randomisation (MR) has risen as a potent method to navigate the challenges of causal inference [15]. As genetic variants are randomly allocated at conception, MR tends to be more robust against confounding and reverse causation compared with traditional epidemiologic designs [16]. This inherent randomness in genetic assignment mirrors the attributes of a randomised controlled trial. As such, MR stands as a proficient tool to mitigate the influence of potential confounders and reverse causality.

In this study, we aim to employ a bidirectional MR approach to delve into the causal relationship between frailty, quantified by the FI and FFS, and multiple phenotypes of chronic pain. To elucidate potential mechanistic pathways, we further employed two-step MR analysis to assess the mediation effects of several common risk factors. Deciphering the causal interplay between frailty and the myriad forms of chronic pain holds promise for pioneering diagnostic, preventive and therapeutic strategies.

Methods

Study design

Figure 1 provides a comprehensive flowchart elucidating our study design. We first performed a two-sample bidirectional MR based on recent large-scale genome-wide association studies (GWAS) focusing on individuals of European ancestry. We then used a two-step MR approach to assess the potential mediations in the frailty and chronic pain relationship. The design and reporting of this study adhere to the guidelines of the Strengthening the Reporting of Observational Studies in Epidemiology using Mendelian Randomisation (STROBE-MR) checklist [17] to facilitate readers’ evaluation of our results. No additional ethical approval was required for the current analysis, as all GWAS statistics included in this study have been made publicly available, and had been previously approved by the corresponding ethical review boards.

(A) Schematic diagram of three assumptions for MR in this study: (i) Relevance: genetic variant is associated with the exposure; (ii) Independence: genetic variant is not related to any confounding factors of the exposure-outcome association and (iii)Exclusion restriction: genetic variant does not affect outcome except through its potential effect on the exposure. (B) Two-step MR analysis framework. Step 1 estimated the causal effect of the exposure on the potential mediators, and step 2 assessed the causal effect of the mediators on outcome. ‘Direct effect’ indicates the effect of exposure on outcome after adjusting for the mediator. ‘Indirect effect’ indicates the effect of exposure on outcome through the mediator. SNPs, single nucleotide polymorphisms; IVs, instrumental variables.
Figure 1

(A) Schematic diagram of three assumptions for MR in this study: (i) Relevance: genetic variant is associated with the exposure; (ii) Independence: genetic variant is not related to any confounding factors of the exposure-outcome association and (iii)Exclusion restriction: genetic variant does not affect outcome except through its potential effect on the exposure. (B) Two-step MR analysis framework. Step 1 estimated the causal effect of the exposure on the potential mediators, and step 2 assessed the causal effect of the mediators on outcome. ‘Direct effect’ indicates the effect of exposure on outcome after adjusting for the mediator. ‘Indirect effect’ indicates the effect of exposure on outcome through the mediator. SNPs, single nucleotide polymorphisms; IVs, instrumental variables.

Open in new tabDownload slide

GWAS data sources

Frailty data sources

Summary statistics for FI, which quantifies the accumulation of health deficits throughout an individual’s life, was obtained from a GWAS meta-analysis involving European descent participants from the UK Biobank and the Swedish TwinGene cohort [18]. In addition, summary data for the FFS, a validated and standardised definition of frailty phenotype, were obtained from a recent large-scale GWAS with 386,565 participants of European ancestry enrolled in the UK Biobank [19].

Chronic pain phenotypes data sources

Summary GWAS data for multisite chronic pain (MCP) were obtained from a recent study conducted by Johnston et al. [20], which encompassed 387,649 UK Biobank participants. Summary-level data for chronic pain (for 3+ months) across different body sites (headache, facial, neck/shoulder, stomach/abdominal, back, hip and knee chronic pain) were collected from the OpenGWAS database at MRC-IEU.

Data sources for possible mediators

Potential mediators for chronic pain [21] and frailty [22], encompassing factors such as body mass index (BMI), smoking initiation, physical inactivity, educational attainment and depression, were investigated. These factors were, respectively, extracted from publicly available GWASs. Detailed information and the source of GWAS data used for our analyses were presented in Supplementary Table 1.

Selection of genetic instruments

Single nucleotide polymorphisms (SNPs) were selected at the genome-wide threshold of 5 × 10−8. We relaxed the selection criteria to 5 × 10−6 in cases where the number of included SNPs was less than three. SNPs that exhibited high linkage disequilibrium (r2 > 0.001 or clump windows <10 Mb) were excluded based on the 1,000 Genomes European reference panel [23]. The strength of individual SNPs was also validated by the F-statistic to avoid weak instrument bias (only F-statistic value >10 were included).

Two-sample bidirectional MR

The inverse-variance weighted (IVW) method was applied as the primary analysis under a multiplicative random effects model. Whereas MR-Egger, weighted median and weighted mode method were adopted to improve the IVW estimates [24, 25]. For sensitivity analysis, the Egger-intercept test and the MR Pleiotropy Residual Sum and Outlier (MR-PRESSO) global test were used to assess the presence of horizontal pleiotropy. Cochrane’s Q-test was adopted to detect heterogeneity among the included SNPs.

Two-step MR

We conducted a two-step MR analysis [26] to explore the mediation effects of potential mediators mentioned above. We followed the approach described in Tyler et al. [27] to estimate the indirect effect and standard errors for the indirect effects were derived by using the delta method [28].

Statistical analysis

All analyses were performed using the packages TwoSampleMR (version 0.5.7) and MRPRESSO (version 1.0) under the environment of R (version 4.2.3). For assessing the significance at a global level, a two-sided P-value threshold of P < 0.05 was considered statistically significant. To address multiple testing, a conservative Bonferroni-corrected threshold (P < 0.002) by the IVW method was adopted to indicate statistical significance, because we assessed the associations between eight chronic pain related traits and frailty (FI and FFS) in both directions.

Results

Genetic instruments

All SNPs selected for inclusion or exclusion according to the criteria of genetic instrument selection are presented in Supplementary Tables 2–11 and for replication. Notably, all selected SNPs exhibited an F-statistic value exceeding 10, indicating a low risk of weak instrument bias.

The causal effect of frailty on various chronic pain phenotypes

Figure 2 illustrates the positive association between genetic liability to frailty, as assessed by both FI and FFS, and eight distinct chronic pain phenotypes. Importantly, these associations persisted even after corrections for multiple comparisons. Specifically, a genetic predisposition to frailty was linked with higher risk of chronic pain on the neck/shoulder (FI: odds ratio (OR) = 1.06, 95% confidence interval (CI) = 1.03–1.08, P = 1.33E-05; FFS: OR = 1.20, 95% CI = 1.11–1.30, P = 3.11E-6), stomach/abdominal (FI: OR = 1.11, 95% CI = 1.07–1.16, P = 3.63E-05; FFS: OR = 1.27, 95% CI = 1.10–1.46, P = 0.001), back (FI: OR = 1.08, 95% CI = 1.05–1.10,P = 1.96E-08; FFS: OR = 1.22, 95% CI = 1.13–1.32, P = 4.70E-07), hip (FI: OR = 1.05, 95% CI = 1.02–1.08, P = 0.001; FFS: OR = 1.13, 95% CI = 1.04–1.24, P = 5.00E-04), knee (FI: OR = 1.06, 95% CI = 1.03–1.08, P = 4.31E-06; FFS: OR = 1.17, 95% CI = 1.10–1.25, P = 3.58E-06), as well as MCP (FI: OR = 1.49, 95% CI = 1.33–1.68, P = 1.05E-11; FFS: OR = 1.94, 95% CI = 1.70–2.21, P = 1.59E-23). Conversely, no significant effects were observed linking the frailty with headache (FI: OR = 1.01, P = 0.357; FFS: OR = 1.06, P = 0.253) or facial pain (FI: OR = 1.08, P = 0.142; FFS: OR = 1.00, P = 0.984). Consistency in MR estimates across methods, including MR-Egger, weighted median and weighted mode, lends further credence to these causal associations, as detailed in Supplementary Table 12. For sensitivity analysis, although heterogeneity was detected by Cochran’s Q statistic for some associations between FFS and chronic pain phenotypes, the pleiotropy test by Egger intercept showed no obvious horizontal pleiotropy (all P > 0.05, Supplementary Table 13). MR-PRESSO detected three outliers (rs17501820, rs2726033 and rs28429148) in the analysis of FFS on MCP and the association persisted after removing these SNPs (corrected P = 1.61E-15) (Supplementary Table 13).

MR analysis of the bidirectional association between frailty and chronic pain. (A) The causal effect of frailty on various chronic pain phenotypes. (B) The causal effect of various chronic pain phenotypes on frailty. CI, confidence interval; OR, odds ratio. The error bars represent 95% CIs.
Figure 2

MR analysis of the bidirectional association between frailty and chronic pain. (A) The causal effect of frailty on various chronic pain phenotypes. (B) The causal effect of various chronic pain phenotypes on frailty. CI, confidence interval; OR, odds ratio. The error bars represent 95% CIs.

Open in new tabDownload slide

The causal effect of various chronic pain phenotypes on frailty

Considering genetic predisposition to chronic pain as the exposure, we conducted a reverse MR analysis to examine its potential causal effect on the frailty. We identified a positive association between genetic susceptibility to MCP with an elevated FI (beta = 0.684, 95% CI = 0.580–0.788, P = 5.77E-38) and FFS (beta = 0.396, 95% CI = 0.325–0.467, P = 1.52E-27) (Figure 2). Nevertheless, we found no evidence of causal relationships for the other seven site-specific chronic pain (Figure 2). MR estimates of MCP on FI and FFS were consistent in direction across various methods (Supplementary Table 14). Furthermore, no substantial pleiotropy was detected by the MR-Egger intercept tests (Supplementary Table 15). MR-PRESSO detected two outliers (rs10888692 and rs12765185) and one outlier (rs13164188) in the analysis for FI and FFS on MCP, respectively. However, the association persisted after the removal of the SNPs (FI: corrected P = 1.02E-17; FFS: corrected P = 1.04E-12) (Supplementary Table 15).

Mediation MR analysis

The two-step MR was further employed to conduct a mediation MR analysis. The results of the mediation MR analysis and the proportion mediated by individual factors were presented in Table 1 and Figure 3. Interestingly, our findings indicated that genetic liability to both frailty (FI and FFS) and MCP were associated with higher levels of BMI, physical inactivity and educational attainment (Figure 3A). In addition, associations were identified between these two conditions with a higher risk of smoking initiation and depression (Figure 3A). We also observed that these five common risk factors mediated a significant proportion of the effect in the relationship between frailty and MCP, ranging from 7.1 to 28.3% for the influence of frailty on MCP and 5.7 to 40.7% for the impact of MCP on frailty (Table 1).

Table 1
Open in new tab

Mediation MR analysis outcomes

Risk factorExposure-outcomeMediation effectMediated proportion (%)
β (95% CI)P(95% CI)
BMIFI-MCP0.067 (0.046, 0.091)2.36E-0817.0 (11.4, 22.6)
FFS-MCP0.187 (0.138, 0.238)7.43E-1428.3 (20.8, 35.9)
MCP-FI0.045 (0.069, 0.245)5.27E-0423.0 (10.1, 35.8)
MCP-FFS0.150 (0.088, 0.235)4.86E-0440.7 (22.2, 59.3)
Smoking initiationFI-MCP0.068 (0.042, 0.098)1.19E-0617.4 (10.4, 24.4)
FFS-MCP0.046 (0.005, 0.089)0.0287.1 (0.8, 13.4)
MCP-FI0.072 (0.019, 0.131)0.01325.7 (2.8,19.2)
MCP-FFS0.049 (0.013, 0.091)0.01413.1 (3.3, 22.9)
Physical inactivityFI-MCP0.040 (0.003, 0.079)0.03510.0 (0.7, 19.3)
FFS-MCP0.157 (0.103, 0.215)2.06E-0824.0(15.5, 32.4)
MCP-FI0.089 (0.049, 0.135)5.22E-0513.5 (7.2, 19.7)
MCP-FFS0.167 (0.107, 0.236)4.18E-0713.8 (27.0, 59.6)
Educational attainmentFI-MCP0.054 (0.029, 0.083)1.15E-048.2 (4.2, 12.1)
FFS-MCP0.124 (0.086, 0.168)3.53E-0919.1 (13.0, 25.3)
MCP-FI0.091 (0.058, 0.129)4.29E-0713.7 (8.5, 18.9)
MCP-FFS0.075 (0.046, 0.108)2.77E-0619.5 (11.6, 27.3)
DepressionFI-MCP0.055 (0.024, 0.090)1.22E-0314.2 (6.0, 22.4)
FFS-MCP0.050 (0.022, 0.086)1.78E-038.1 (3.3, 13.0)
MCP-FI0.038 (0.016, 0.066)3.47E-036.0 (2.3, 9.6)
MCP-FFS0.021 (0.008, 0.037)2.70E-035.7 (2.0, 9.3)
Risk factorExposure-outcomeMediation effectMediated proportion (%)
β (95% CI)P(95% CI)
BMIFI-MCP0.067 (0.046, 0.091)2.36E-0817.0 (11.4, 22.6)
FFS-MCP0.187 (0.138, 0.238)7.43E-1428.3 (20.8, 35.9)
MCP-FI0.045 (0.069, 0.245)5.27E-0423.0 (10.1, 35.8)
MCP-FFS0.150 (0.088, 0.235)4.86E-0440.7 (22.2, 59.3)
Smoking initiationFI-MCP0.068 (0.042, 0.098)1.19E-0617.4 (10.4, 24.4)
FFS-MCP0.046 (0.005, 0.089)0.0287.1 (0.8, 13.4)
MCP-FI0.072 (0.019, 0.131)0.01325.7 (2.8,19.2)
MCP-FFS0.049 (0.013, 0.091)0.01413.1 (3.3, 22.9)
Physical inactivityFI-MCP0.040 (0.003, 0.079)0.03510.0 (0.7, 19.3)
FFS-MCP0.157 (0.103, 0.215)2.06E-0824.0(15.5, 32.4)
MCP-FI0.089 (0.049, 0.135)5.22E-0513.5 (7.2, 19.7)
MCP-FFS0.167 (0.107, 0.236)4.18E-0713.8 (27.0, 59.6)
Educational attainmentFI-MCP0.054 (0.029, 0.083)1.15E-048.2 (4.2, 12.1)
FFS-MCP0.124 (0.086, 0.168)3.53E-0919.1 (13.0, 25.3)
MCP-FI0.091 (0.058, 0.129)4.29E-0713.7 (8.5, 18.9)
MCP-FFS0.075 (0.046, 0.108)2.77E-0619.5 (11.6, 27.3)
DepressionFI-MCP0.055 (0.024, 0.090)1.22E-0314.2 (6.0, 22.4)
FFS-MCP0.050 (0.022, 0.086)1.78E-038.1 (3.3, 13.0)
MCP-FI0.038 (0.016, 0.066)3.47E-036.0 (2.3, 9.6)
MCP-FFS0.021 (0.008, 0.037)2.70E-035.7 (2.0, 9.3)

CI, confidence interval; FI, Frailty Index; FFS, Fried Frailty Score; MCP, multisite chronic pain; BMI, body mass index.

Table 1
Open in new tab

Mediation MR analysis outcomes

Risk factorExposure-outcomeMediation effectMediated proportion (%)
β (95% CI)P(95% CI)
BMIFI-MCP0.067 (0.046, 0.091)2.36E-0817.0 (11.4, 22.6)
FFS-MCP0.187 (0.138, 0.238)7.43E-1428.3 (20.8, 35.9)
MCP-FI0.045 (0.069, 0.245)5.27E-0423.0 (10.1, 35.8)
MCP-FFS0.150 (0.088, 0.235)4.86E-0440.7 (22.2, 59.3)
Smoking initiationFI-MCP0.068 (0.042, 0.098)1.19E-0617.4 (10.4, 24.4)
FFS-MCP0.046 (0.005, 0.089)0.0287.1 (0.8, 13.4)
MCP-FI0.072 (0.019, 0.131)0.01325.7 (2.8,19.2)
MCP-FFS0.049 (0.013, 0.091)0.01413.1 (3.3, 22.9)
Physical inactivityFI-MCP0.040 (0.003, 0.079)0.03510.0 (0.7, 19.3)
FFS-MCP0.157 (0.103, 0.215)2.06E-0824.0(15.5, 32.4)
MCP-FI0.089 (0.049, 0.135)5.22E-0513.5 (7.2, 19.7)
MCP-FFS0.167 (0.107, 0.236)4.18E-0713.8 (27.0, 59.6)
Educational attainmentFI-MCP0.054 (0.029, 0.083)1.15E-048.2 (4.2, 12.1)
FFS-MCP0.124 (0.086, 0.168)3.53E-0919.1 (13.0, 25.3)
MCP-FI0.091 (0.058, 0.129)4.29E-0713.7 (8.5, 18.9)
MCP-FFS0.075 (0.046, 0.108)2.77E-0619.5 (11.6, 27.3)
DepressionFI-MCP0.055 (0.024, 0.090)1.22E-0314.2 (6.0, 22.4)
FFS-MCP0.050 (0.022, 0.086)1.78E-038.1 (3.3, 13.0)
MCP-FI0.038 (0.016, 0.066)3.47E-036.0 (2.3, 9.6)
MCP-FFS0.021 (0.008, 0.037)2.70E-035.7 (2.0, 9.3)
Risk factorExposure-outcomeMediation effectMediated proportion (%)
β (95% CI)P(95% CI)
BMIFI-MCP0.067 (0.046, 0.091)2.36E-0817.0 (11.4, 22.6)
FFS-MCP0.187 (0.138, 0.238)7.43E-1428.3 (20.8, 35.9)
MCP-FI0.045 (0.069, 0.245)5.27E-0423.0 (10.1, 35.8)
MCP-FFS0.150 (0.088, 0.235)4.86E-0440.7 (22.2, 59.3)
Smoking initiationFI-MCP0.068 (0.042, 0.098)1.19E-0617.4 (10.4, 24.4)
FFS-MCP0.046 (0.005, 0.089)0.0287.1 (0.8, 13.4)
MCP-FI0.072 (0.019, 0.131)0.01325.7 (2.8,19.2)
MCP-FFS0.049 (0.013, 0.091)0.01413.1 (3.3, 22.9)
Physical inactivityFI-MCP0.040 (0.003, 0.079)0.03510.0 (0.7, 19.3)
FFS-MCP0.157 (0.103, 0.215)2.06E-0824.0(15.5, 32.4)
MCP-FI0.089 (0.049, 0.135)5.22E-0513.5 (7.2, 19.7)
MCP-FFS0.167 (0.107, 0.236)4.18E-0713.8 (27.0, 59.6)
Educational attainmentFI-MCP0.054 (0.029, 0.083)1.15E-048.2 (4.2, 12.1)
FFS-MCP0.124 (0.086, 0.168)3.53E-0919.1 (13.0, 25.3)
MCP-FI0.091 (0.058, 0.129)4.29E-0713.7 (8.5, 18.9)
MCP-FFS0.075 (0.046, 0.108)2.77E-0619.5 (11.6, 27.3)
DepressionFI-MCP0.055 (0.024, 0.090)1.22E-0314.2 (6.0, 22.4)
FFS-MCP0.050 (0.022, 0.086)1.78E-038.1 (3.3, 13.0)
MCP-FI0.038 (0.016, 0.066)3.47E-036.0 (2.3, 9.6)
MCP-FFS0.021 (0.008, 0.037)2.70E-035.7 (2.0, 9.3)

CI, confidence interval; FI, Frailty Index; FFS, Fried Frailty Score; MCP, multisite chronic pain; BMI, body mass index.

Summary MR estimates derived from the IVW for the effect of the exposure on each mediator (A) and the effect of each mediator on outcome (B). CI, confidence interval; BMI, body mass index. The error bars represent 95% CIs.
Figure 3

Summary MR estimates derived from the IVW for the effect of the exposure on each mediator (A) and the effect of each mediator on outcome (B). CI, confidence interval; BMI, body mass index. The error bars represent 95% CIs.

Open in new tabDownload slide

Discussion

In the present study, we performed a comprehensive MR investigation to unravel the relationships between chronic pain phenotypes and the risk of frailty, utilising large-scale summary-level statistics from GWAS. We found that genetic susceptibility to frailty was associated with an increased risk of specific pain phenotypes, including neck/shoulder, stomach/abdominal, back, hip, knee and MCP, but not for headache and facial pain. In the reverse directional MR, potential causal associations were identified between MCP and an elevated FI and FFS.

The FI and FFS have emerged as potent tools, aptly suited to assess and monitor frailty conditions over time [3, 19]. As frailty-related brain changes might cause compromised descendent inhibitory pain modulation and dysfunction of pain gating mechanisms [11], it was expected that frailty could act as a significant risk determinant for pain. Reinforcing this notion, prior research has suggested that the evaluation of persistent pain could serve as an effective avenue to refine the precision of frailty assessments [29]. Despite the evident importance, scant research has been conducted to elucidate whether frailty acts as a precursor to the onset of chronic pain. In our study, we identified that a genetic predisposition to increased FI and FFS targeted pain in specific regions and MCP. Regarding the observed lack of significant associations between frailty and headache as well as facial pain, it is important to recognise that these two specific pain phenotypes may involve complex and multifaceted mechanisms that extend beyond the scope of our study. Furthermore, the prevalence and aetiology of headache and facial pain, including primary headache disorders or localised facial pain conditions can vary widely among individuals [30, 31], making it challenging to establish a uniform causal relationship with frailty. Moreover, genetic factors, lifestyle choices and other unmeasured variables may play unique roles in the development of headache and facial pain, which can lead to the absence of significant causal associations with frailty in our analysis. Future research may benefit from more specific and detailed investigations into these particular pain phenotypes, considering their distinct characteristics. In addition, Megale et al. [32] presented a prospective, population-based cohort study delving into the question of whether frailty, as measured by the frailty phenotype, escalates the risk of chronic pain manifestation. Their efforts, however, did not yield any significant correlation between the two conditions. This disparity might be attributable to the relatively modest sample size and the potential that frail individuals are more prone to be lost in follow-up in such a longitudinal study. Collectively, our research contributes new evidence to the understanding of the causal relationship between baseline frailty and the subsequent emergence of chronic pain.

Our reverse MR analysis aimed to explore the influence of chronic pain on frailty. Intriguingly, we discovered that genetic predispositions to MCP correlated with an increased risk of frailty, which indicate that an elevated number of chronic sites predict future frailty. Such findings underscore the notion that the consequences of persistent pain extend beyond mere discomfort, potentially catalysing homeostenosis and frailty. Results from this study are consistent with previous traditional observational studies [12, 32–38], which also suggest that chronic pain was independently associated with frailty. However, no significant effects were observed linking the other pain phenotypes with FI and FFS, primarily attributed to the limited instrumental variables and heightened statistical variability. Future studies would benefit from a renewed investigation into the implications of other site-specific pain on frailty. Multiple biological mechanisms are hypothesised to mediate the effect of chronic pain in frailty onset and progression, such as the systemic inflammation and the hypothalamic–pituitary–adrenal axis dysfunction [9]. Pain, in its persistent form, could act as an unremitting stressor, depleting reserves and compromising physiological systems. This could manifest as disturbances in stress hormone levels, sleep irregularities and nutritional imbalances, cumulatively amplifying frailty risks [33]. In addition, a concurrent perspective to consider is the shared roots of pain and frailty, with determinants potentially spanning genetic factors [39], ageing processes and co-morbidities.

In our mediation MR analysis, we further quantified the mediation effects focus on the role of five modifiable lifestyle exposures, including BMI, smoking initiation, physical inactivity, educational attainment and depression. Our findings suggest that the causal relationship between these two conditions may, in part, be mediated by these potentially modifiable common risk factors. Consistent with epidemiological studies [40–43], our findings underscore the significance of weight control and smoking cessation in reducing the risk of pain and frailty. In addition, prior evidence and our MR analysis also suggest that exercise-based interventions may improve physical functioning in frail older people [44] and provide pain relief [45]. Observational studies have previously noted the association between higher education and a reduced risk of frailty [22] and chronic pain [46]. Our results further strengthen the evidence that genetically predicted higher educational attainment is inversely associated with the risk of chronic pain and frailty. Furthermore, recent MR studies have also provided support for the bidirectional causal associations between depression and both chronic pain [47] and frailty [48].

In terms of implications for public health and clinical practice, our primary findings highlight a bidirectional causal relationship between frailty and chronic pain. This implies that interventions aimed at preventing or treating one of these conditions may offer protection against the other. First, given the observed causal relationship between frailty and chronic pain, it is essential to consider interventions targeting frailty as a means to manage pain. Second, continuous early screening and therapeutic approaches for chronic pain syndromes in older adults are imperative to prevent the onset of frailty and mitigate its adverse consequences. Furthermore, our study revealed that several potentially modifiable risk factors contribute causally to both of these conditions. Therefore, effective and multidimensional intervention strategies, including early screening for depression and the promotion of a healthy lifestyle, can be employed to reduce the high co-occurrence of pain and frailty. Taken together, these findings may provide valuable insights into reducing associated adverse outcomes and improving the quality of life for older people.

The major strength of the current study is the utilisation of two-sample bidirectional MR design, incorporating three MR methods, which effectively mitigated biases arising from confounding and reverse causality. Thorough checks of MR assumptions were conducted, and large-sample GWAS data sources were employed, further bolstering the study’s credibility. Moreover, we controlled for bias arising from population stratification by restricting participants to individuals of European ancestry. In addition, the application of MR-Egger and MR-PRESSO analyses yielded no evidence of pleiotropic effects. Furthermore, we also explored the mediating pathways by conducting two-step MR analysis, which deepened the mechanistic understandings and provided evidence supports for prevention strategies.

Nevertheless, several limitations should be noted when interpreting findings from this study. Firstly, although we have tried our best to avoid sample overlap, the use of summary data from the UKB for various components of our study introduces the potential for bias in our analysis. Unfortunately, detailed participant information necessary to resolve this sample overlap issue was not available in the public datasets. It’s important to acknowledge that this issue may affect the robustness of our MR results. Secondary, our study focused solely on individuals of European ancestry, which may limit the generalizability of the results to other populations with different genetic backgrounds. Thirdly, the choice of pain phenotypes and frailty indices may not capture the full spectrum of these complex conditions. Variability in how individuals report pain or experience frailty may not be fully accounted for in our study. Lastly, it is important to note that, despite our inclusion of several potential confounders, there might be additional unmeasured or confounding variables that could impact the associations observed in our analysis. Given the inherent complexities and the potential limitations highlighted, further research with a more diverse population and more comprehensive datasets is warranted to validate and expand upon our findings.

Conclusions

Our findings pave the way for a deeper understanding of the intricate relationship between frailty and chronic pain. BMI, smoking initiation, physical inactivity, educational attainment and depression appeared to mediate many of these associations. The results underscore the potential of early intervention strategies to reduce the risk of these conditions and enhance the quality of life for ageing populations. The recognition that frailty and pain coexist has the potential to improve patient care for both conditions. Research aimed at preventing or ameliorating frailty should consider pain management as a part of multi-component interventions.

Acknowledgements

We thank all the participants and investigators of the GWAS studies included in this study.

Declaration of Conflicts of Interest

None.

Declaration of Sources of Funding

None.

References

1.

Dent
 
E
,
Martin
 
FC
,
Bergman
 
H
,
Woo
 
J
,
Romero-Ortuno
 
R
,
Walston
 
JD
.
Management of frailty: opportunities, challenges, and future directions
.
Lancet
 
2019
;
394
:
1376
–
86
.

Google Scholar

Crossref
Search ADS
PubMed

 
2.

Shi
 
SM
,
Olivieri-Mui
 
B
,
McCarthy
 
EP
,
Kim
 
DH
.
Changes in a frailty index and association with mortality
.
J Am Geriatr Soc
 
2021
;
69
:
1057
–
62
.

Google Scholar

Crossref
Search ADS
PubMed

 
3.

Kojima
 
G
,
Iliffe
 
S
,
Walters
 
K
.
Frailty index as a predictor of mortality: a systematic review and meta-analysis
.
Age Ageing
 
2018
;
47
:
193
–
200
.

Google Scholar

Crossref
Search ADS
PubMed

 
4.

Williams
 
DM
,
Jylhävä
 
J
,
Pedersen
 
NL
,
Hägg
 
S
.
A frailty index for UK Biobank participants
.
J Gerontol A Biol Sci Med Sci
 
2019
;
74
:
582
–
7
.

Google Scholar

Crossref
Search ADS
PubMed

 
5.

Palliyaguru
 
DL
,
Moats
 
JM
,
Di Germanio
 
C
,
Bernier
 
M
,
de
 
Cabo
 
R
.
Frailty index as a biomarker of lifespan and healthspan: focus on pharmacological interventions
.
Mech Ageing Dev
 
2019
;
180
:
42
–
8
.

Google Scholar

Crossref
Search ADS
PubMed

 
6.

Damluji
 
AA
,
Chung
 
S-E
,
Xue
 
Q-L
  et al.  
Frailty and cardiovascular outcomes in the National Health and Aging Trends study
.
Eur Heart J
 
2021
;
42
:
3856
–
65
.

Google Scholar

Crossref
Search ADS
PubMed

 
7.

Cohen
 
SP
,
Vase
 
L
,
Hooten
 
WM
.
Chronic pain: an update on burden, best practices, and new advances
.
Lancet
 
2021
;
397
:
2082
–
97
.

Google Scholar

Crossref
Search ADS
PubMed

 
8.

Lin
 
T
,
Zhao
 
Y
,
Xia
 
X
,
Ge
 
N
,
Yue
 
J
.
Association between frailty and chronic pain among older adults: a systematic review and meta-analysis
.
Eur Geriatr Med
 
2020
;
11
:
945
–
59
.

Google Scholar

Crossref
Search ADS
PubMed

 
9.

D'Agnelli
 
S
,
Amodeo
 
G
,
Franchi
 
S
  et al.  
Frailty and pain, human studies and animal models
.
Ageing Res Rev
 
2022
;
73
:
101515
.

Google Scholar

Crossref
Search ADS
PubMed

 
10.

Saraiva
 
MD
,
Suzuki
 
GS
,
Lin
 
SM
,
Andrade
,
Jacob-Filho
 
W
,
Suemoto
 
CK
.
Persistent pain is a risk factor for frailty: a systematic review and meta-analysis from prospective longitudinal studies
.
Age Ageing
 
2018
;
47
:
785
–
93
.

Google Scholar

Crossref
Search ADS
PubMed

 
11.

Karp
 
JF
,
Shega
 
JW
,
Morone
 
NE
,
Weiner
 
DK
.
Advances in understanding the mechanisms and management of persistent pain in older adults
.
Br J Anaesth
 
2008
;
101
:
111
–
20
.

Google Scholar

Crossref
Search ADS
PubMed

 
12.

Wade
 
KF
,
Lee
 
DM
,
McBeth
 
J
  et al.  
Chronic widespread pain is associated with worsening frailty in European men
.
Age Ageing
 
2016
;
45
:
268
–
74
.

Google Scholar

Crossref
Search ADS
PubMed

 
13.

Coyle
 
PC
,
Sions
 
JM
,
Velasco
 
T
,
Hicks
 
GE
.
Older adults with chronic low back pain: a clinical population vulnerable to frailty?
 
J Frailty Aging
 
2015
;
4
:
188
–
90
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
14.

Coelho
 
T
,
Paúl
 
C
,
Gobbens
 
RJJ
,
Fernandes
 
L
.
Multidimensional frailty and pain in community dwelling elderly
.
Pain Med
 
2017
;
18
: 693–701.

Google Scholar

OpenURL Placeholder Text

 
15.

Davey Smith
 
G
,
Hemani
 
G
.
Mendelian randomization: genetic anchors for causal inference in epidemiological studies
.
Hum Mol Genet
 
2014
;
23
:
R89
–
98
.

Google Scholar

Crossref
Search ADS
PubMed

 
16.

Burgess
 
S
,
Thompson
 
SG
.
Mendelian Randomization: Methods for Using Genetic Variants in Causal Estimation
. vol.
xiv
.
Boca Raton
:
CRC Press Taylor & Francis Group
,
2015
;
210
 
(Chapman & Hall/CRC interdisciplinary statistics series)
.

Google Scholar

Crossref
Search ADS

 
17.

Skrivankova
 
VW
,
Richmond
 
RC
,
Woolf
 
BAR
  et al.  
Strengthening the reporting of observational studies in epidemiology using Mendelian randomization: the STROBE-MR statement
.
JAMA
 
2021
;
326
:
1614
–
21
.

Google Scholar

Crossref
Search ADS
PubMed

 
18.

Atkins
 
JL
,
Jylhävä
 
J
,
Pedersen
 
NL
  et al.  
A genome-wide association study of the frailty index highlights brain pathways in ageing
.
Aging Cell
 
2021
;
20
:
e13459
.

Google Scholar

Crossref
Search ADS
PubMed

 
19.

Ye
 
Y
,
Noche
 
RB
,
Szejko
 
N
  et al.  
A genome-wide association study of frailty identifies significant genetic correlation with neuropsychiatric, cardiovascular, and inflammation pathways
.
Geroscience
 
2023
;
45
:
2511
–
23
.

Google Scholar

Crossref
Search ADS
PubMed

 
20.

Johnston
 
KJA
,
Adams
 
MJ
,
Nicholl
 
BI
  et al.  
Genome-wide association study of multisite chronic pain in UK Biobank
.
PLoS Genet
 
2019
;
15
:
e1008164
.

Google Scholar

Crossref
Search ADS
PubMed

 
21.

Mills
 
SEE
,
Nicolson
 
KP
,
Smith
 
BH
.
Chronic pain: a review of its epidemiology and associated factors in population-based studies
.
Br J Anaesth
 
2019
;
123
:
e273
–
83
.

Google Scholar

Crossref
Search ADS
PubMed

 
22.

Hoogendijk
 
EO
,
Afilalo
 
J
,
Ensrud
 
KE
,
Kowal
 
P
,
Onder
 
G
,
Fried
 
LP
.
Frailty: implications for clinical practice and public health
.
Lancet
 
2019
;
394
:
1365
–
75
.

Google Scholar

Crossref
Search ADS
PubMed

 
23.

Auton
 
A
,
Brooks
 
LD
,
Durbin
 
RM
  et al.  
A global reference for human genetic variation
.
Nature
 
2015
;
526
:
68
–
74
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
24.

Yavorska
 
OO
,
Burgess
 
S
.
Mendelian randomization: an R package for performing Mendelian randomization analyses using summarized data
.
Int J Epidemiol
 
2017
;
46
:
1734
–
9
.

Google Scholar

Crossref
Search ADS
PubMed

 
25.

Burgess
 
S
,
Thompson
 
SG
.
Interpreting findings from Mendelian randomization using the MR-Egger method
.
Eur J Epidemiol
 
2017
;
32
:
377
–
89
.

Google Scholar

Crossref
Search ADS
PubMed

 
26.

Carter
 
AR
,
Sanderson
 
E
,
Hammerton
 
G
  et al.  
Mendelian randomisation for mediation analysis: current methods and challenges for implementation
.
Eur J Epidemiol
 
2021
;
36
:
465
–
78
.

Google Scholar

Crossref
Search ADS
PubMed

 
27.

Vander Weele
 
TJ
.
Mediation analysis: a practitioner's guide
.
Annu Rev Public Health
 
2016
;
37
:
17
–
32
.

Google Scholar

Crossref
Search ADS
PubMed

 
28.

Carter
 
AR
,
Gill
 
D
,
Davies
 
NM
  et al.  
Understanding the consequences of education inequality on cardiovascular disease: Mendelian randomisation study
.
BMJ
 
2019
;
365
:
l1855
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
29.

Lohman
 
MC
,
Whiteman
 
KL
,
Greenberg
 
RL
,
Bruce
 
ML
.
Incorporating persistent pain in phenotypic frailty measurement and prediction of adverse health outcomes
.
J Gerontol A Biol Sci Med Sci
 
2017
;
72
:
216
–
22
.

Google Scholar

Crossref
Search ADS
PubMed

 
30.

Sharma
 
TL
.
Common primary and secondary causes of headache in the elderly
.
Headache
 
2018
;
58
:
479
–
84
.

Google Scholar

Crossref
Search ADS
PubMed

 
31.

Ziegeler
 
C
,
Beikler
 
T
,
Gosau
 
M
,
May
 
A
.
Idiopathic facial pain syndromes–an overview and clinical implications
.
Dtsch Arztebl Int
 
2021
;
118
:
81
–
7
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
32.

Megale
 
RZ
,
Ferreira
 
ML
,
Ferreira
 
PH
  et al.  
Association between pain and the frailty phenotype in older men: longitudinal results from the Concord Health and Ageing in Men Project (CHAMP)
.
Age Ageing
 
2018
;
47
:
381
–
7
.

Google Scholar

Crossref
Search ADS
PubMed

 
33.

Shega
 
JW
,
Dale
 
W
,
Andrew
 
M
,
Paice
 
J
,
Rockwood
 
K
,
Weiner
 
DK
.
Persistent pain and frailty: a case for homeostenosis
.
J Am Geriatr Soc
 
2012
;
60
:
113
–
7
.

Google Scholar

Crossref
Search ADS
PubMed

 
34.

Sodhi
 
JK
,
Karmarkar
 
A
,
Raji
 
M
,
Markides
 
KS
,
Ottenbacher
 
KJ
,
Al Snih
 
S
.
Pain as a predictor of frailty over time among older Mexican Americans
.
Pain
 
2020
;
161
:
109
–
13
.

Google Scholar

Crossref
Search ADS
PubMed

 
35.

Wade
 
KF
,
Marshall
 
A
,
Vanhoutte
 
B
,
Wu
 
FCW
,
O'Neill
 
TW
,
Lee
 
DM
.
Does pain predict frailty in older men and women? Findings from the English Longitudinal Study of Ageing (ELSA)
.
J Gerontol A Biol Sci Med Sci
 
2017
;
72
: glw226  
403
–
9
.

Google Scholar

PubMed
OpenURL Placeholder Text

 
36.

Rodríguez-Sánchez
 
I
,
García-Esquinas
 
E
,
Mesas
 
AE
,
Martín-Moreno
 
JM
,
Rodríguez-Mañas
 
L
,
Rodríguez-Artalejo
 
F
.
Frequency, intensity and localization of pain as risk factors for frailty in older adults
.
Age Ageing
 
2019
;
48
:
74
–
80
.

Google Scholar

Crossref
Search ADS
PubMed

 
37.

Dapp
 
U
,
Minder
 
CE
,
Anders
 
J
,
Golgert
 
S
, von
Renteln-Kruse
 
W
.
Long-term prediction of changes in health status, frailty, nursing care and mortality in community-dwelling senior citizens—results from the Longitudinal Urban Cohort Ageing Study (LUCAS)
.
BMC Geriatr
 
2014
;
14
:
141
.

Google Scholar

Crossref
Search ADS
PubMed

 
38.

Veronese
 
N
,
Maggi
 
S
,
Trevisan
 
C
  et al.  
Pain increases the risk of developing frailty in older adults with osteoarthritis
.
Pain Med
 
2017
;
18
: 414–27.

Google Scholar

OpenURL Placeholder Text

 
39.

Livshits
 
G
,
Ni Lochlainn
 
M
,
Malkin
 
I
  et al.  
Shared genetic influence on frailty and chronic widespread pain: a study from TwinsUK
.
Age Ageing
 
2018
;
47
:
119
–
25
.

Google Scholar

Crossref
Search ADS
PubMed

 
40.

Narouze
 
S
,
Souzdalnitski
 
D
.
Obesity and chronic pain: systematic review of prevalence and implications for pain practice
.
Reg Anesth Pain Med
 
2015
;
40
:
91
–
111
.

Google Scholar

Crossref
Search ADS
PubMed

 
41.

Chen
 
Z
,
Chen
 
Z
,
Jin
 
X
.
Mendelian randomization supports causality between overweight status and accelerated aging
.
Aging Cell
 
2023
;
22
:
e13899
.

Google Scholar

Crossref
Search ADS
PubMed

 
42.

Robinson
 
CL
,
Kim
 
RS
,
Li
 
M
  et al.  
The impact of smoking on the development and severity of chronic pain
.
Curr Pain Headache Rep
 
2022
;
26
:
575
–
81
.

Google Scholar

Crossref
Search ADS
PubMed

 
43.

Jayanama
 
K
,
Theou
 
O
,
Godin
 
J
,
Mayo
 
A
,
Cahill
 
L
,
Rockwood
 
K
.
Relationship of body mass index with frailty and all-cause mortality among middle-aged and older adults
.
BMC Med
 
2022
;
20
:
404
.

Google Scholar

Crossref
Search ADS
PubMed

 
44.

Angulo
 
J
,
El Assar
 
M
,
Álvarez-Bustos
 
A
,
Rodríguez-Mañas
 
L
.
Physical activity and exercise: strategies to manage frailty
.
Redox Biol
 
2020
;
35
:
101513
.

Google Scholar

Crossref
Search ADS
PubMed

 
45.

Belavy
 
DL
,
van
 
Oosterwijck
 
J
,
Clarkson
 
M
  et al.  
Pain sensitivity is reduced by exercise training: evidence from a systematic review and meta-analysis
.
Neurosci Biobehav Rev
 
2021
;
120
:
100
–
8
.

Google Scholar

Crossref
Search ADS
PubMed

 
46.

Zadro
 
JR
,
Shirley
 
D
,
Pinheiro
 
MB
  et al.  
Does educational attainment increase the risk of low back pain when genetics are considered? A population-based study of Spanish twins
.
Spine J
 
2017
;
17
:
518
–
30
.

Google Scholar

Crossref
Search ADS
PubMed

 
47.

Tang
 
B
,
Meng
 
W
,
Hägg
 
S
,
Burgess
 
S
,
Jiang
 
X
.
Reciprocal interaction between depression and pain: results from a comprehensive bidirectional Mendelian randomization study and functional annotation analysis
.
Pain
 
2022
;
163
:
e40
–
8
.

Google Scholar

Crossref
Search ADS
PubMed

 
48.

Sang
 
N
,
Li
 
B-H
,
Zhang
 
M-Y
  et al.  
Bidirectional causal relationship between depression and frailty: a univariate and multivariate Mendelian randomisation study
.
Age Ageing
 
2023
;
52
: afad113.

Google Scholar

OpenURL Placeholder Text

 
© The Author(s) 2024. Published by Oxford University Press on behalf of the British Geriatrics Society. All rights reserved. For permissions, please email: [email protected].
This article is published and distributed under the terms of the Oxford University Press, Standard Journals Publication Model (https://academic.oup.com/pages/standard-publication-reuse-rights)
Topic:
Subject
Sarcopenia and Frailty
Issue Section:
Research Paper
Download all slides
  • Supplementary data

    • Supplementary data

    Supplementary_Table_afad256 - pdf file

    Comments

    0 Comments
    Comments (0)
    Submit a comment
    You have entered an invalid code
    Thank you for submitting a comment on this article. Your comment will be reviewed and published at the journal's discretion. Please check for further notifications by email.