Frailty, a common geriatric syndrome, is defined as an age-related state characterized by impaired biological reserves, a reduced ability to maintain physiologic homeostasis, and an increased vulnerability to adverse outcomes (Dent et al., 2019). With the increasing rate of population aging, concerns about frailty and its impact on the global health burden are intensifying. Frailty is commonly assessed using the frailty index (FI) (Atkins et al., 2021), defined as the proportion of accumulated health deficits (signs, symptoms, dysfunctions, and laboratory abnormalities) over a lifetime. FI is better at identifying frailty in its low to mid-range spectrum compared to the frailty phenotype (FP), rendering it more sensitive in younger individuals. Additionally, FI is considered predictive of the risk of falls, fractures, disability, and death (Shi et al., 2021). Insomnia, a prevalent sleep disorder, affects approximately 10% of adults chronically and an additional 20% experience occasional insomnia symptoms (Perlis et al., 2022). Insomnia prevalence is higher among women, the elderly, and individuals facing economic hardship. Growing evidence suggests that individuals with insomnia may experience multiple somatic and psychological disorders, impaired quality of life, and higher all-cause mortality rates (Ge et al., 2019). Evidence indicates that insomnia can result in sleep dysfunction, fatigue, unsteady gait, and decreased physical activity, all factors that elevate the risk of frailty (Fan et al., 2022). Furthermore, the variety of psychological and somatic disorders induced by insomnia may heighten vulnerability to frailty. Recently, there has been an increasing body of evidence suggesting the coexistence of frailty and insomnia. A recent meta-analysis encompassing 12 observational studies (Pourmotabbed et al., 2020) suggests an association between sleep disturbances and frailty. However, prior studies exploring the potential association between frailty and insomnia have produced inconsistent results (Nemoto et al., 2021; Tang et al., 2021; Wen et al., 2023), making it challenging to establish causality.

Frailty and insomnia both possess a substantial genetic basis. Studies have shown that frailty has a genetic foundation, with estimates of heritability varying from 30 to 45% (Livshits et al., 2018). A recent extensive GWAS study identified 202 common genetic loci associated with insomnia (Jansen et al., 2019). Few studies to date have explored the link between frailty and insomnia at a genetic level, and it remains unclear as to whether and how their genetic structures overlap. Based on the established link between frailty and insomnia in epidemiological studies, we hypothesize that frailty and insomnia could share a common genetic structure and that their relationship might entail a causal connection. Based on this hypothesis, it is anticipated that this study will yield new insights into the pathophysiological mechanisms underlying the co-morbidity of frailty and insomnia, and identify new therapeutic targets for the two diseases, whether occurring individually or in combination, thereby contributing to healthier, more independent aging.

Cross-trait analysis (Zhu et al., 2021) employs various meta-analytic methods to integrate summary statistics of distinct yet potentially related traits. This approach aims to identify specific loci with shared associations within a meta-analytic framework. It contributes significantly to the comprehension of the genetic architecture of complex traits, facilitating the exploration of genetic sharing, trait interactions, and potential biological mechanisms. Multi-trait analysis of GWAS (MTAG) (Turley et al., 2018) represents a methodology for concurrent multi-phenotypic analyses, utilizing GWAS generalized new data. This approach enables the leveraging of information from associated phenotypes to augment the statistical power of tests for target phenotypes, offering an advantage over single-phenotype GWAS. Cross-Phenotype Association (CPASSOC) (Zhu et al., 2021) employs aggregated single SNP-trait associations from GWAS to ascertain which variants are associated with at least one trait. It exhibits several benefits in identifying cross-phenotypic associations, including the ability to accommodate opposite risk effects and various types of phenotypic traits. Cross-trait analyses are currently being applied in the field of neurology. For example, Tian et al. (2023) have demonstrated a shared genetic framework between amyotrophic lateral sclerosis and Parkinson’s disease, encompassing 9 single-nucleotide polymorphisms, 3 risk loci, and 7 genes. These genes are linked to neuronal projection development pathways. In another study, Zeng et al. (2023) discovered a notable positive genetic link between body mass index and multiple sclerosis, indicated by 39 common risk SNPs. Such results provide new perspectives on the mechanisms driving their co-morbidities and guide upcoming treatment strategies.

Mendelian randomization has been utilized in studies of frailty and insomnia, leveraging its ability to mitigate the effects of confounders and diminish reverse causal inference (Emdin et al., 2017). For instance, Deng et al. (2023) demonstrated a bidirectional causal relationship between frailty and depression from a genetic standpoint via Mendelian randomization. This suggests that routine frailty screening for depressed patients is necessary, and treating depression may also moderately reduce the risk of frailty. Utilizing Mendelian randomization, researchers have determined that various sleep characteristics (short sleep, long sleep, continuous sleep duration) not only are associated with but also causally influence the prognosis of ischaemic stroke, potentially serving as a therapeutic target to enhance stroke prognosis (Zhang et al., 2023).

In our research, we explored the genetic links and the possibility of a causal connection between frailty and insomnia, using extensive GWAS data. Initially, to comprehend the shared genetic bases of frailty and insomnia, we established the genetic correlations and local genetic correlations. Subsequently, we utilized cross-trait GWAS meta-analyses for identifying common SNPs. Moreover, we implemented two-sample Mendelian randomization analyses to examine the causative link between frailty and insomnia. Furthermore, our study involved analyzing the tissue-specific enrichment of genetic associations linking frailty and insomnia. Summary data-based Mendelian randomization (SMR) (Zhu et al., 2016) utilizing pooled data, is a methodology that merges summary data from GWAS with data from expression quantitative trait loci (eQTL) studies in order to identify genes whose expression levels are linked to complex traits owing to pleiotropy. Utilizing SMR, we pinpointed genes with shared functionality between frailty and insomnia. The STROBE-MR guidelines (Skrivankova et al., 2021a,b) were followed throughout the course of the study. Figure 1 presents a flowchart depicting our analytical strategy.

To our knowledge, this study is the first genome-wide cross-trait analysis to systematically assess the shared genetic foundations of frailty and insomnia. Our research, leveraging a detailed GWAS dataset along with tissue-specific expression data, reveals potential genetic connections between frailty and insomnia. This is supported by several critical observations: Initially, we found that specific genomic regions and functional elements are linked to both frailty and insomnia. Second, bidirectional MR analyses indicate that frailty exerts a causal effect on insomnia, albeit not reciprocally. Finally, our investigation into genetic commonalities between frailty and insomnia, particularly in brain tissue, led to the identification of numerous potential functional genes impacting both phenotypes.

A path-analysis study revealed that poor sleep quality contributes to increased frailty susceptibility in older adults, potentially linked to poor physical functioning (Tang et al., 2021), and indicated that over half of pre-frail and frail older adults exhibit insomnia symptoms (Nemoto et al., 2021). The association between insomnia and aging is believed to initiate a chronic low-grade inflammatory state in older adults, leading to alterations in the immune system, a phenomenon termed inflammatory aging (Besedovsky et al., 2019; Irwin, 2019). Transitioning from a state of homeostasis to low-grade inflammation induces an inflammatory response, the severity of which is predictive of mortality in older adults (Ingiosi et al., 2013). Prolonged activation of a low-grade inflammatory state may impede the return to homeostasis during sleep repair, potentially leading to disease in older adults (Pak et al., 2020). Furthermore, extended sleep latency is known to elevate insulin resistance via the activation of inflammatory markers (Kim et al., 2016), thereby exacerbating protein synthesis impairment and muscle loss, potentially worsening limb movement.

To date, there has been no comprehensive analysis of the genetic link between frailty and insomnia in any study. Following a local genetic correlation analysis, the independent region 3p21.31 (chr3:47727212 … 0.49316972) was identified. One study reported a correlation between 3p21.31 and the incidence of gliomas and glioblastomas in women (Ostrom et al., 2018). Additionally, the TMIE gene located at 3p21.31 is believed to be significantly associated with sleep duration (Meng et al., 2022). Considering the notable gene-level correlations, a cross-trait GWAS analysis was conducted to identify 124 SNPs co-associated with frailty and insomnia, including two SNPs at 3p21.31: rs34290943 and rs10865954. The SNP rs34290943 was mapped to the DALRD3 gene, while rs10865954 was located in the MIR191 gene locus. The protein encoded by the DALRD3 gene possesses a DALR anticodon binding domain, akin to that found in the leucyl aminoacyl tRNA synthetase class. It has been shown that the expression level of DALRD3 in the brain tissue of patients with insomnia is significantly lower compared to controls, and significant dynamic changes in the expression level of this gene were observed in the mouse cerebral cortex during sleep and sleep deprivation states (Sun et al., 2020). Previous studies have suggested that significantly reduced plasma levels of MIR191 in elderly patients are associated with age-related frailty syndrome, with its notable age-related downregulation in plasma aligning with prior observations in monocytes (Serna et al., 2012). Furthermore, miR-191 is a component of a distinctive circulating 7-miRNA signature that differentiates Alzheimer’s disease patients from normal controls (Nagpal and Kulshreshtha, 2014).

Subsequent analyses of functional components uncovered 24 annotated regions linked to frailty and insomnia. It is suggested that the Coding_UCSCL2_0 region may have associations with neuropsychiatric disorders including Alzheimer’s disease, major depression, bipolar disorder, and sporadic Creutzfeldt-Jakob disease (de Souza et al., 2010). In the study by Hnisz et al. mutations in MAFbin1L2_0 were found to be associated with Alzheimer’s disease (Hnisz et al., 2013). Animal experiments demonstrated that increased H3K27ac enrichment in early gene promoters of neurons enhances memory capacity in mice, thereby influencing aging (Li et al., 2021). Intron_UCSCL2_0 is crucial for forebrain inhibitory neuronal differentiation and exhibits a strong association with restless legs syndrome (RLS) and insomnia (Lam et al., 2022). These functional components linked to frailty and insomnia require additional experimental validation.

We investigated the causal relationship between frailty and insomnia using a GSMR approach, an extension of the SMR approach that utilizes all SNPs of the highest genome-wide significance levels associated with the exposure as IVs for causal testing. The GSMR method excelled in efficacy for discovering causal effects, surpassing methods such as IVW and MR-Egger. It also demonstrated considerable robustness in the face of pleiotropy in the instrumental variables and linkage disequilibrium among instrumental variables (Zhu et al., 2018). Two-sample Mendelian randomization analyses disclosed a causal effect of frailty on insomnia and indicated horizontal pleiotropy in the causal effect of insomnia on frailty. This implies that older frail patients are more prone to insomnia symptoms, that the onset of insomnia leading to frailty could be impacted by other confounders, and that a shared genetic risk exists between frailty and insomnia. One meta-analysis suggested that insomnia independently correlates with frailty among older adults in the community or nursing homes; however, the inclusion of six studies using the FP to identify frailty partly accounts for the inconsistency in findings (Wen et al., 2023). Compared to randomized controlled trials (RCTs), Mendelian randomization offers the benefits of lower cost, diminished confounding, and lessened reverse causal inference, thus serving as a ‘natural’ RCT that mitigates traditional observational study biases and is superior in deducing causality (Emdin et al., 2017).

Tissue enrichment analysis for key genes associated with frailty and insomnia was conducted, revealing significant enrichment in brain tissues for both conditions. S-LDSC analysis demonstrated that SNPs co-associated with frailty and insomnia were enriched in 11 brain regions, mainly in the anterior cingulate cortex (BA24) and the amygdala. Similarly, MAGMA tissue-specificity analyses indicated enrichment in 9 brain regions, especially in the cerebellum and cerebellar hemispheres. This enrichment, particularly in the cerebellum and cerebellar hemispheres, suggests a potential mechanism for the co-morbidity of frailty and insomnia in brain structures. Aligned with our findings, a genome-wide analysis of insomnia (involving 593,724 cases and 1,771,286 controls) demonstrated that genes linked to insomnia were primarily enriched in brain tissues [including cerebellar hemispheres, cerebellum, frontal cortex (BA9), cerebral cortex, anterior cingulate cortex (BA24)] (Watanabe et al., 2022). A GWAS study with 386,565 participants of European ancestry from the UK Biobank revealed that genes related to frailty were predominantly enriched in specific brain areas (cerebellar hemispheres, frontal cortex BA9, cerebellum, anterior cingulate cortex BA24, nucleus ambiguus in the basal ganglia) (Ye et al., 2023). Neuroimaging meta-analyses on insomnia have suggested a critical role for the anterior cingulate cortex and amygdala in insomnia disorders (Reimann et al., 2023). Voxel-based studies have revealed that frailty is associated with multiple brain regions, potentially leading to reduced grey matter volume in areas such as the hippocampus, amygdala, fusiform gyrus, and several cortical regions (Nishita et al., 2019). Additionally, reduced grey matter in the cerebellum is recognized as a neurological feature of frailty (Chen et al., 2015). Through SMR exploration of functional genes linked to frailty and insomnia in specific enriched tissues, four genes were identified: RMB6, MST1R, RF123, and FAM212A. In line with our study, MST1R and RF123 are considered to be associated with brain function or psychiatric disorders, exhibiting upregulated expression in various brain regions (Qiu et al., 2022). The role of these identified genes in frailty and insomnia necessitates further experimental verification.

Our study presents several limitations. Firstly, the genetic data on frailty and insomnia in this study originated exclusively from European populations, potentially diminishing the heterogeneity of the results to some extent. However, this aspect also restricts the generalization of our results to other ethnic populations, and therefore, future analyses should incorporate data from diverse ethnicities. Secondly, the genetic data encompassed a large cohort of patients with insomnia symptoms, thereby enlarging the sample size of the data, but also introduced a certain level of heterogeneity in the findings. Incorporating additional GWAS data that integrate genetic information with comprehensive clinical characteristics (e.g., varying severities of insomnia, duration of symptoms, frequency of occurrence) in subsequent studies may yield a more nuanced comprehension of the role genetic factors play in the varied clinical manifestations of frailty and insomnia. Thirdly, our analyses were confined to the genetic characteristics of the traits, acknowledging that epigenetic influences on these traits are equally crucial.

In conclusion, our research revealed a notable genetic link between frailty and insomnia, highlighted by the identification of shared risk SNPs. This correlation was found to be more distinct in certain genomic areas. Partitioned genetic analyses revealed 24 functional elements significantly linked to both frailty and insomnia. Mendelian randomization was additionally employed to demonstrate a causal relationship between frailty and insomnia. The genetic correlation was found to be enriched in 11/9 brain regions, within which 4 functional genes were detected. These findings may offer insights into the shared genetic architecture of frailty and insomnia, contributing to an enhanced understanding of their pathogenesis and the development of therapeutic targets aimed at alleviating insomnia symptoms and potentially preventing or reversing aging.

Publicly available datasets were analyzed in this study. Data about the frailty index can be found at: https://www.ebi.ac.uk/gwas/downloads/summary-statistics GCST90020053. Data about insomnia can be found at https://gwas.mrcieu.ac.uk/datasets/ukb-b-3957.

Written informed consent was not required for any of the potentially identifiable images or data included in this study, as all statistical analyses were based on existing publicly available summary data and thus no additional ethical approval was required.

ZS: Visualization, Writing – review & editing. WL: Formal analysis, Writing – original draft. YH: Formal analysis, Writing – original draft. YX: Data curation, Writing – original draft. YW: Supervision, Writing – review & editing.