Causal relationship between subhealth status and pan-cancer risk: A Mendelian randomization study.

This study used publicly available GWAS data obtained through the Integrative Epidemiology Unit (IEU, https://gwas.mrcieu.ac.uk/ ). [ 32 ] HRV contains 3 exposure factors obtained from IEU OpenGWAS Project with more than 24088 individuals. [ 32 ] Mood swings were available from IEU OpenGWAS Project with 407,746 individuals. [ 28 ] Irritability data were available from IEU OpenGWAS Project with 442,169 individuals. [ 32 ] Health satisfaction data were extracted from IEU OpenGWAS Project with 152,420 individuals. [ 32 ] All samples were of European ancestry (Table 1 ). The GWAS summary statistics and transcriptomic data utilized in our investigation were sourced from publicly accessible websites. [ 32 ] Details of the GWAS data analyzed in this study. GWAS = genome-wide association studies, IEU = integrative epidemiology unit, pvRSA/HF = peak-to-valley respiratory sinus arrhythmia or high-frequency power, RMSSD = root mean square of successive differences in interbeat intervals, SDNN = standard deviation of normal-to-normal interbeat intervals, SNP = single-nucleotide polymorphism.

All MR analyses were performed using RStudio. Initially, single-nucleotide polymorphisms (SNPs) associated with a specific exposure variable ( P  < 5 × 10 −6 ) were selected to mitigate weak-instrument bias. Next, linkage disequilibrium was minimized using the PLINK clumping method ( r ² < 0.001, kb = 10,000). To harmonize effect alleles across the exposure and outcome GWAS datasets, SNPs with palindromic alleles and intermediate allele frequencies were removed. The MR Steiger filtering method was then applied to exclude SNPs where the association with the exposure was weaker than with the outcome, further refining the validity of the selected instruments. To account for multiple testing across numerous exposure–outcome pairs, false discovery rate (FDR) correction was performed using the Benjamini–Hochberg method. As the FDR-adjusted results were consistent with the nominal P -values, only the nominal values are presented in the main text for clarity (Fig. 1 ). Study design overview. Exposure to mediating variable causal effects: HRV (heart-rate variability), [ 32 ] Mood swings, [ 28 ] Irritability, [ 32 ] Health satisfaction. [ 32 ] Exposure to outcome variable causal effects: Colorectal cancer, Lung cancer, Basal cell carcinoma, Cervical cancer, Pancreatic cancer. pvRSA/HF = peak-to-valley respiratory sinus arrhythmia or high-frequency power, RMSSD = root mean square of successive RR-interval differences, SDNN = standard deviation of all NN intervals.

Applying a relaxed threshold for significance ( P  < 5 × 10 −6 ) and pruning for linkage disequilibrium yielded a set of independent SNPs. Ultimately, 157 SNPs were included as IVs in the analysis (Fig. 3 ). Univariable MR analysis using the IVW method (OR = 1.01; 95% CI = 1.00–1.01; P  = .046), the IVW radial method (OR = 1.01; 95% CI = 1.00–1.02; P  = .038), and the maximum likelihood method (OR = 1.01; 95% CI = 1.00–1.01; P  = .045) suggested a modest positive association between irritability and lung cancer. No evidence of directional pleiotropy or heterogeneity was detected in the MR-Egger and weighted median analyses (Table 3 ). Mendelian randomization results of irritability and mood swing with cancers. IVW = inverse variance-weighted, MR = Mendelian randomization, OR = odds ratio, SNP = single-nucleotide polymorphism. Scatter plot illustrating the causal association between irritability and lung cancer risk. Genetically predicted higher irritability was significantly associated with an increased risk of lung cancer (IVW: OR = 1.01; 95% CI = 1.00–1.01; P  = .046).

After applying a relaxed significance threshold ( P  < 5 × 10 −6 ) and adjusting for linkage disequilibrium, a set of independent SNPs was identified, leading to the inclusion of 95 SNPs as instrumental variables for the subsequent analysis (Fig. 4 ). Univariable MR analysis, employing the IVW method (OR = 1.00; 95% CI = 0.99–1.00; P  = .024), the IVW radial method (OR = 1.00; 95% CI = 1.00–1.00; P  = .015), the weighted median method (OR = 1.00; 95% CI = 1.00–1.00; P  = .025), and the maximum likelihood method (OR = 1.00; 95% CI = 1.00–1.00; P  = .024), all indicated a modest positive association between mood swings and the incidence of basal cell carcinoma. Moreover, MR-Egger analysis provided no evidence of directional pleiotropy or heterogeneity, further supporting the robustness of the observed associations (Table 3 ). Scatter plot illustrating the causal association between mood swings and the risk of basal cell carcinoma. Genetically predicted higher mood swings were significantly associated with a decreased risk of basal cell carcinoma (IVW: OR = 1.00; 95% CI = 0.99–1.00; P  = .024). By applying a relaxed significance threshold ( P  < 5 × 10 −6 ) and considering linkage disequilibrium, we identified a set of independent SNPs, resulting in 27 SNPs being retained as instrumental variables for the analysis (Fig. 5 A). Univariable MR analysis, employing the IVW method (OR = 1.01; 95% CI = 1.00–1.02; P  < .001), the IVW radial method (OR = 1.01; 95% CI = 1.00–1.02; P  < .001), the weighted median method (OR = 1.01; 95% CI = 1.00–1.02; P  = .011), and the maximum likelihood method (OR = 1.01; 95% CI = 1.00–1.02; P  < .001), suggested a modest positive association between health satisfaction and cervical cancer. MR-Egger analyses revealed no evidence of directional pleiotropy or heterogeneity (Table 4 ). Mendelian randomization results of Health satisfaction with cancers. IVW = inverse variance-weighted, MR = Mendelian randomization, OR = odds ratio, SNP = single-nucleotide polymorphism. Scatter plots illustrating the causal associations between health satisfaction and cancer risk. Genetically predicted higher health satisfaction was significantly associated with an increased risk of (A) cervical cancer (IVW: OR = 1.01; 95% CI = 1.00–1.02; P  < .001) and (B) pancreatic cancer (IVW: OR = 2.29; 95% CI = 1.17–4.50; P  = .015). Following the application of a relaxed significance threshold (( P  < 5 × 10 −6 ) and the adjustment for linkage disequilibrium, a set of independent SNPs was identified, resulting in the inclusion of 46 SNPs as instrumental variables in the subsequent analysis (Fig. 5 B). Univariable MR analysis, utilizing the IVW method (OR = 2.29; 95% CI = 1.17–4.50; P  = .015), the IVW radial method (OR = 2.29; 95% CI = 1.41–3.74; P  < .001), and the maximum likelihood method (OR = 2.35; 95% CI = 1.19–4.65; P  = .014), all demonstrated a modest positive association between health satisfaction and the incidence of pancreatic cancer. Furthermore, both MR-Egger and weighted median analyses yielded no evidence of directional pleiotropy or heterogeneity, supporting the robustness of the observed associations (Table 4 ).

This study employed a two-sample MR approach using data from the UK Biobank to investigate the potential causal relationships between SHS and cancer risk. Specifically, it examined the effects of 4 subhealth indicators – HRV, health satisfaction, irritability, and emotional instability – on the risk of various cancers. The results suggested that high HRV may have a protective effect against colorectal cancer, whereas irritability was associated with an increased risk of lung cancer. Additionally, health satisfaction showed a positive association with cervical cancer and pancreatic cancer, while emotional instability was linked to basal cell carcinoma. These findings indicate that certain subhealth conditions may contribute to cancer development, underscoring the importance of early intervention. Based on the identified associations, the study also proposes several potential strategies for improving these subhealth conditions to facilitate early cancer prevention. Nevertheless, the study acknowledges several limitations, including potential detection bias and the complex interplay between genetic and environmental factors. Future research should focus on strengthening genetic instrumental variables, employing multivariable MR analyses, and integrating additional biological data to further validate these associations and enhance understanding of the underlying causal mechanisms.

Subhealth, defined as an intermediate state between health and disease, is becoming increasingly prevalent within the Chinese population. [ 1 ] With advancements in quality of life and medical care, many diseases are now detected earlier and treated more effectively than ever before. However, the state bridging health and disease often remains overlooked. Characterized by psychological, behavioral, or physical disturbances, or abnormalities in certain medical indicators without definitive pathological features, [ 2 ] subhealth has garnered growing attention from researchers. The identification of subhealth is more complex than that of typical disease states due to the absence of definitive pathological features. To accurately identify individuals in a subhealthy state within the population, the use of quantifiable indicators and standardized questionnaires is required. Cancer remains a leading cause of morbidity and mortality worldwide, representing a threatening public health challenge. [ 3 ] Characterized by uncontrolled cell proliferation, invasion into surrounding tissues, and the potential for metastatic spread, cancer encompasses a heterogeneous group of diseases driven by genetic, epigenetic, and environmental factors. [ 4 ] Advances in molecular biology have elucidated the pivotal roles of oncogenes, tumor suppressor genes, and dysregulated cellular signaling pathways in tumor initiation and progression. Moreover, the tumor microenvironment, comprising immune cells, stromal components, and extracellular matrix, has emerged as a critical modulator of cancer pathophysiology, influencing tumor growth, angiogenesis, and immune evasion. [ 5 ] Despite remarkable progress in therapeutic strategies, including targeted therapies and immunotherapy, challenges such as drug resistance and disease heterogeneity persist. [ 6 ] Thus, a deeper understanding of the underlying mechanisms of carcinogenesis is essential for developing innovative diagnostic and therapeutic approaches. Emerging evidence suggests a multifaceted interplay between subhealth states and cancer development, underscoring the role of chronic, low-grade systemic dysregulation as a potential prelude to malignancy. As mentioned above, subhealthy states encompass a spectrum of conditions, including obesity, irritable bowel syndrome (IBS), and endocrine hormonal imbalances, all of which are characterized by physiological and psychological deviations from homeostasis without overt clinical disease. Obesity, for instance, has been linked to increased systemic inflammation [ 7 ] and adipokine dysregulation, [ 8 ] fostering a pro-tumorigenic microenvironment. Similarly, IBS is associated with altered gut microbiota and chronic gastrointestinal inflammation, which may influence carcinogenesis through microbial metabolite production and immune modulation. [ 9 ] These findings highlight the need for a deeper understanding of the pathophysiological continuum between subhealth status (SHS) and cancer to inform preventive strategies and early interventions. Certain subhealth indicators, such as obesity, endocrine disorders, and IBS have been extensively studied and are suggested to have potential associations with cancer development. However, subhealth encompasses a broader spectrum of conditions beyond these quantified indicators, and their potential links to cancer remain uncertain. To explore whether less studied but representative subhealth indicators might also be associated with cancer risk, we selected a subset of these indicators for further investigation. Heart rate variability (HRV) serves as a noninvasive and readily measurable parameter that reflects the intricate interplay between the brain and the cardiovascular system. [ 10 ] Three sets of instrumental variables were selected to characterize HRV as an exposure factor: the standard deviation of normal-to-normal interbeat intervals (SDNN), the root mean square of successive differences in interbeat intervals (RMSSD), and the peak-to-valley respiratory sinus arrhythmia or high-frequency power (pvRSA/HF). [ 11 ] Therefore, increased HRV is generally considered indicative of favorable health status, whereas reduced variability may signal underlying pathological changes. [ 12 ] Prolonged SHS can lead to excessive tension in the central nervous system and dysfunction of the autonomic nervous system. HRV, as a sensitive indicator of autonomic nervous system function, is thus proposed as a potential objective diagnostic marker for SHS. Health satisfaction serves as a metric for individuals’ subjective evaluations of their own health, encompassing emotional and cognitive assessments of their quality of life. [ 13 ] In 1946, the World Health Organization redefined “health” as a state of complete physical, mental, and social well-being, rather than merely the absence of disease or infirmity. [ 14 ] This redefinition underscores the notion that “health” extends beyond physical well-being to include individuals’ positive perceptions of their own health status. An increasing body of research indicates that life satisfaction is not solely influenced by physical health; rather, life satisfaction itself may exert a reciprocal influence on individual health outcomes, such as cardiovascular disease, [ 15 , 16 ] kidney function [ 17 ] and gastro-esophageal reflux disease. [ 18 ] Irritability, characterized by an exaggerated response to aversive stimuli with associated negative outcomes, is most commonly manifested through impulsivity, anger, and hostility. [ 19 ] Previous studies have demonstrated a potential association between irritability and a range of cardiovascular diseases. [ 20 – 22 ] Both cohort and prospective studies have reported positive correlations between irritability and the incidence of severe acute coronary syndrome, atrial fibrillation, [ 21 ] heart failure, [ 23 ] and hypertension. [ 24 ] Mood swings are a common personality trait characterized by frequent, sudden, and unpredictable mood changes. [ 25 ] Extensive research has established a strong association between mood swings and various psychiatric disorders, such as bipolar disorder. [ 26 ] However, emerging evidence suggests that mood swing is also closely linked to cerebrovascular, [ 27 ] cardiovascular, [ 28 ] gastrointestinal [ 29 ] and gynecological [ 30 ] diseases, highlighting its broader relevance to physical health. After reviewing the relevant literature, we found that, apart from the association between irritability and lung cancer, few systematic investigations have examined the relationships between other SHS indicators and cancer risk. [ 31 ] Therefore, we incorporated HRV, health satisfaction, irritability, and mood swings as representative indicators of SHS, and performed univariable Mendelian randomization (MR) analyses to explore their potential causal associations with multiple types of cancer. The results showed that genetically predicted higher HRV (SDNN and pvRSA/HF) was associated with a reduced risk of colorectal cancer, whereas higher irritability was causally linked to an increased risk of lung cancer. Mood swings exhibited a modest positive association with basal-cell carcinoma, while greater health satisfaction was associated with higher risks of cervical and pancreatic cancer. No directional pleiotropy or heterogeneity was detected for these associations. Collectively, these findings provide novel insights into the potential causal roles of SHS indicators in cancer susceptibility, highlighting their value for cancer prevention and risk stratification.

Conceptualization: Jing Liu. Data curation: Xin-Ning Yu, Hua-Tao Wu, Yan-Yu Hou, Jing Liu. Formal analysis: Xin-Ning Yu, Hua-Tao Wu. Funding acquisition: Jing Liu. Investigation: Xin-Ning Yu, Hua-Tao Wu, Yan-Yu Hou, Yang-Zheng Lan, Wen-Jia Chen, Jing Liu. Methodology: Xin-Ning Yu, Hua-Tao Wu, Yan-Yu Hou, Yang-Zheng Lan, Wen-Jia Chen, Jing Liu. Software: Xin-Ning Yu, Hua-Tao Wu, Jing Liu. Supervision: Jing Liu. Visualization: Xin-Ning Yu, Hua-Tao Wu, Yan-Yu Hou, Yang-Zheng Lan, Wen-Jia Chen, Jing Liu. Writing – original draft: Xin-Ning Yu, Hua-Tao Wu. Writing – review & editing: Jing Liu.

MR analysis was conducted based on public-aggregated statistics from genome-wide association studies (GWAS). We analyzed the overall causal relationship between selected subhealth indexes and pan-cancer after eliminating the interference of confounding factors. Ethical approval and informed consent were not required for this study because it utilized publicly available, de-identified genetic summary data.

Independent SNPs were identified by applying a relaxed significance threshold ( P  < 5 × 10 −6 ) and excluding those in linkage disequilibrium. Subsequently, traits influenced by these SNPs were extracted for further analysis. SNPs associated with potential confounders, including relevant indicators, were identified and excluded. Ultimately, 11 SNPs were retained for analysis as instrumental variables (IVs). Univariable MR analysis identified an association between pvRSA/HF and colorectal cancer risk. This relationship was observed using the inverse variance-weighted (IVW) method (OR = 0.87; 95% CI = 0.76–0.99; P  = .038), and was further supported by the weighted median approach (OR = 0.81; 95% CI = 0.69–0.96; P  = .017). Also, IVW radial (OR = 0.87; 95% CI = 0.76–0.99; P  = .038) and maximum likelihood (OR = 0.87; 95% CI = 0.76–0.99; P  = .029) methods gave the same result (Table 2 ). Next, 18 SNPs were retained for analysis as IVs. Univariable MR analysis identified an association between SDNN and colorectal cancer risk (Fig. 2 ). This relationship was observed using the IVW method (OR = 0.77; 95% CI = 0.58–1.09; P  = .043), and was further supported by the weighted median approach (OR = 0.61; 95% CI = 0.41–0.89; P  = .010). Also, IVW radial (OR = 0.76; 95% CI = 0.60–0.96; P  = .021) and maximum likelihood (OR = 0.77; 95% CI = 0.58–1.01; P  = .054) methods gave the same result (Table 2 ). Although 17 SNPs were identified as IVs, there is no statistically significant association between root mean square of successive differences in interbeat intervals and the risk of colorectal cancer (Table 2 and Fig. S1, Supplemental Digital Content, https://links.lww.com/MD/Q604 ). Mendelian randomization results of HRV traits with colorectal cancer. HRV = heart rate variability, IVW = inverse variance-weighted, MR = Mendelian randomization, OR = odds ratio, pvRSA/HF = peak-to-valley respiratory sinus arrhythmia or high-frequency power, RMSSD = root mean square of successive differences in interbeat intervals, SDNN = standard deviation of normal-to-normal interbeat intervals, SNP = single-nucleotide polymorphism. Scatter plots illustrating the causal associations between heart rate variability (HRV) indicators and colorectal cancer risk. (A) Peak-to-valley respiratory sinus arrhythmia or high-frequency power (pvRSA/HF) and (B) Standard deviation of normal-to-normal interbeat intervals (SDNN) were both inversely associated with colorectal cancer risk, as indicated by statistically significant results. Specifically, pvRSA/HF (IVW: OR = 0.87; 95% CI = 0.76–0.99; P  = .038) and SDNN (IVW: OR = 0.77; 95% CI = 0.58–1.09; P  = .043) showed protective associations against colorectal cancer.

Subhealth is defined as a state characterized by abnormalities in psychological behavior, physiological features, or certain medical examination indicators, in the absence of clear pathological features. [ 2 ] It is considered a potential predictor of overall health. In this study, we selected representative yet underexplored subhealth indicators – HRV, irritability, mood swings, and health satisfaction – and investigated their associations with various types of cancer, using MR analysis based on publicly available summary statistics from GWAS. According to univariable MR analysis, high HRV appears to be a protective factor against colorectal cancer (OR < 1). Previous studies have demonstrated an association between HRV and survival outcomes in patients with colorectal cancer. However, HRV has not been identified as an independent predictor of survival. [ 33 ] HRV reflects the complex interplay between sympathetic and parasympathetic influences on the autonomic regulation of sinoatrial node activity through both central and peripheral nervous systems. It serves as an important noninvasive marker for assessing autonomic nervous system function, particularly vagal activity. [ 34 ] The vagal nerve plays a critical role in regulating inflammation and preventing tissue damage caused by excessive inflammatory responses. Its primary mechanisms of action involve reducing the production of pro-inflammatory cytokines and inhibiting the migration of leukocytes to sites of inflammation. [ 35 ] Consequently, HRV is typically inversely correlated with levels of inflammatory markers. Notably, the association between inflammatory responses and colorectal cancer has been extensively documented in the literature. Overall, inflammation is intricately involved throughout the entire course of colorectal cancer, from tumor initiation to progression, and the development of therapeutic resistance. Inflammation not only contributes to tumorigenesis by inducing DNA damage and promoting oxidative stress, but also facilitates tumor growth and metastasis by remodeling the tumor microenvironment. [ 36 ] Specifically, it shapes the phenotypic polarization of immune cells and stromal cells within the tumor microenvironment, thereby creating a pro-tumorigenic milieu that supports cancer cell proliferation, invasion, and immune evasion. [ 37 ] In addition to mechanistic investigations, numerous retrospective studies have further substantiated the close association between inflammation and colorectal cancer by examining the relationship between inflammation-related biomarkers and patient prognosis. Biomarkers, such as neutrophil-associated markers, C-reactive protein-related markers, and platelet-associated markers, have been shown to correlate with clinical outcomes in colorectal cancer, highlighting the critical role of systemic inflammatory responses in influencing disease progression and survival. [ 38 ] Based on these findings, we cautiously hypothesize that enhancing HRV through lifestyle interventions may, to some extent, contribute to a reduction in the incidence of colorectal cancer. Supporting this notion, a systematic review has reported that short-term meditation training, such as integrative body-mind training, can improve attentional control and self-regulation abilities. [ 39 ] These improvements are associated with the activation of the vagal nerve and the enhancement of parasympathetic nervous system activity, ultimately leading to increased HRV. A meta-analysis has found that voluntary slow breathing can modulate parasympathetic nervous system control over cardiac function, thereby enhancing HRV. [ 40 ] As a low-technology, low-cost intervention, voluntary slow breathing practice holds potential for use in both preventive and adjunctive therapeutic strategies, with a relatively low risk of adverse effects. [ 40 ] Research has demonstrated a positive association between elevated irritability levels and an increased incidence of lung cancer (OR > 1). Previous studies have suggested that patients with lung cancer often exhibit higher levels of emotional distress and anxiety symptoms compared to patients with other types of cancer. [ 41 ] These emotional disturbances are associated with adverse physiological and psychosocial outcomes. However, the mechanisms by which irritability-related emotional disorders may increase the risk of developing lung cancer remain to be further elucidated. An MR study has identified a strong association between irritability and the risk of developing gastroesophageal reflux disease. [ 42 ] Irritability may act as a contributing factor by promoting the onset of gastroesophageal reflux disease, which in turn exacerbates systemic inflammatory responses. This inflammatory milieu may serve as a mediating pathway linking irritability to an increased incidence of lung cancer. [ 31 ] These findings suggest a potential indirect mechanism through which emotional dysregulation could influence cancer risk, warranting further investigation. A retrospective cohort analysis investigating the relationship between depression and cancer diagnosis indicated that individuals with depression exhibited an increased risk of being diagnosed with cancer during the follow-up period. Overall, depression was associated with an 18% increase in cancer risk, with lung cancer showing a particularly prominent association. [ 43 ] Additional studies have shown that depression and anxiety are not only associated with an increased risk of developing lung cancer, but are also linked to a higher risk of lung cancer-specific mortality and elevated all-cause mortality among lung cancer patients. [ 44 ] These findings suggest that emotional factors may play a more pronounced role in lung cancer development and may even serve as an independent risk factor for cancer onset. This highlights the need for further research to elucidate the underlying biological mechanisms linking emotional factors to cancer development. Based on these findings, we cautiously hypothesize that mitigating adverse emotional factors through lifestyle interventions may contribute, at least in part, to reducing the incidence of lung cancer and to improving the prognosis of lung cancer patients. Cognitive behavioral therapy has been widely applied in the emotional management of cancer patients. One study demonstrated that chemotherapy patients who received cognitive behavioral therapy interventions exhibited clinical improvements in anger control and death anxiety, suggesting that cognitive behavioral therapy may serve as an effective psychotherapeutic approach within psycho-oncology to enhance the emotional well-being of cancer patients. [ 45 ] Additionally, mindfulness meditation has been shown to reduce stress and anxiety, improve sleep disturbances, enhance immune system function, and increase overall quality of life. [ 46 ] One study further indicated that mindfulness meditation exerted positive effects on psychological symptoms among cancer patients, and contributed beneficially to their overall health outcomes. [ 47 ] Research data indicate that individuals with greater mood swings have a higher risk of developing basal cell carcinoma compared to those with more stable emotions (OR = 1), suggesting a positive association between mood swings and the incidence of basal cell carcinoma. A previous MR study investigating the relationship between mood swings and gynecological disorders demonstrated a positive association between mood swings and the risks of endometrial cancer, cervical cancer, and endometriosis, whereas no association was observed with the risk of ovarian cancer. [ 30 ] It is well established that psychological distress is closely associated with the development of various cancers, particularly gynecological malignancies. [ 48 ] Psychological distress may lead to dysregulation of immune and endocrine functions, such as triggering chronic stress responses and sustained activation of the hypothalamic-pituitary-adrenal axis. This prolonged activation can impair immune responses, potentially contributing to cancer development. [ 49 ] Although previous studies have suggested a strong association between mood swings and reproductive system tumors – which are often closely linked to sex hormones – there is currently no evidence supporting a genetic causal relationship between basal cell carcinoma and sex hormones. [ 50 ] This discrepancy may indicate that mood swings are unlikely to act as a direct promoting factor for basal cell carcinoma development, despite the association observed between the 2 in our study ( P  < .05). In this MR study, we observed a novel association between mood swings and basal cell carcinoma, yet the estimated odds ratio (OR) was exactly 1.00, suggesting no detectable causal effect. This apparent paradox prompts careful consideration of both methodological and biological explanations. One potential reason lies in the limited statistical power of the genetic instruments used to proxy mood swings. The genetic variants identified from GWAS may account for only a small proportion of the phenotypic variance in mood instability, thereby reducing the instrumental variable strength and attenuating the causal estimate. Furthermore, if the confidence interval around the OR is narrow but still includes 1.00, it implies a precise null effect. However, if the interval is wide, it may reflect imprecision due to insensitive instruments or sample size limitations. Another explanation involves the biological plausibility of the association. Mood swings have been linked to reproductive system cancers, which are known to be influenced by sex hormones. In contrast, current evidence suggests that basal cell carcinoma has little to no genetic causal connection with sex hormones. [ 50 ] This difference implies that mood swings, while potentially hormonally mediated, may not have a direct pathophysiological relevance to basal cell carcinoma development. The observed statistical association might therefore be confounded by unmeasured shared risk factors, or reflect pleiotropic effects of the selected genetic variants. Additionally, it is plausible that mood swings influence health-related behaviors – such as sun exposure, smoking, or stress management – that are themselves risk factors for basal cell carcinoma. However, if these behavioral mediators are not captured or adjustment is not made in the MR model, the analysis may fail to detect an indirect causal pathway. Multivariable MR or mediation analyses could help elucidate such mechanisms in future work. Taken together, our findings highlight the importance of interpreting MR results with caution, particularly when the point estimate is exactly null despite statistical significance. The absence of a clear causal link between mood swings and basal cell carcinoma suggests that their observed correlation in observational data may not be driven by direct biological mechanisms, but rather by complex, possibly indirect, interactions that warrant further investigation. Emerging data suggest a positive correlation between higher levels of health satisfaction and an increased risk of developing cervical cancer and pancreatic cancer (OR > 1). In previous studies exploring the relationship between health satisfaction and cancer risk, a prospective investigation into life satisfaction and cancer incidence reported no association after adjusting for key confounding factors. [ 51 ] A meta-analysis focusing on life satisfaction among cancer patients indicated that individuals diagnosed with cancer often experience a marked decline in life satisfaction during the early stages of the disease. This decline is largely attributed to the well-documented negative impacts of cancer, which hinder patients from living the life they desire. [ 52 ] However, the analysis also found that life satisfaction improved following appropriate exercise interventions. [ 53 ] However, in a meta-analysis examining life satisfaction, the overall quality of evidence for all outcomes was rated as very low. This assessment was driven by serious concerns regarding risk of bias, imprecision of results, and substantial indirectness of the evidence. [ 51 ] As such, previous studies suggesting a negative association between life satisfaction and cancer risk may lack reliability. Using MR analysis, our study is the first to reveal that individuals with higher health satisfaction may have an increased risk of developing cancer. This finding stands in contrast to most previous studies, which generally suggested that higher satisfaction is associated with lower cancer risk. There are several possible explanations for this discrepancy. One is detection bias: individuals with higher satisfaction may be more proactive in undergoing medical checkups and cancer screenings, which increases the likelihood of earlier or more frequent diagnosis, thereby inflating the apparent incidence. Another factor is the potential mismatch between subjective health satisfaction and actual health status. Some individuals may report feeling healthy despite having undiagnosed medical conditions. Health optimism bias may also play a role. People who are overly confident in their well-being may ignore early warning signs, which could delay diagnosis. It is also possible that people with high health satisfaction engage more in social activities, some of which may involve lifestyle or environmental exposures (such as alcohol consumption or pollution) that increase cancer risk. These factors may collectively influence the association between health satisfaction and cancer risk. We also highlight the importance of considering both objective health status and individual lifestyle behaviors when interpreting subjective perceptions of health. Encouraging regular health checkups and promoting engagement in activities that support physical well-being are essential steps toward the early detection and timely treatment of cancer. Next, we acknowledge the limitations inherent in the statistical methodologies employed in this study. First, only univariable MR analyses were conducted, as the associations between subhealth indicators and cancer risk were relatively modest, which made it challenging to establish a valid multivariable MR framework. Additionally, robust genetic instruments capable of simultaneously accounting for potential mediators or lifestyle-related confounders were not available. Second, all participants were of European ancestry, as the analyses were based on the UK Biobank dataset. This limits the generalizability of our findings to other ancestral populations, and further validation in non-European cohorts is warranted to confirm the robustness and applicability of our results. Third, certain exposure traits, such as irritability and mood swings, were self-reported, which may introduce recall bias and measurement error, potentially attenuating causal estimates. Fourth, several observed associations yielded marginal effect sizes (e.g., OR ≈ 1.00), suggesting that even if statistically significant, the underlying causal effects are likely weak and should be interpreted with caution. We also acknowledge that some of the selected SNPs may represent relatively weak instruments. Although these SNPs demonstrated statistically significant associations with the exposures, F -statistics were not calculated due to the lack of precise variance data for each SNP. Moreover, given the novelty of this research area and the limited availability of data, a slightly relaxed SNP selection threshold was applied to retain a sufficient number of instruments. In summary, while these limitations may have influenced certain findings, they also highlight areas for methodological improvement. Future studies will aim to optimize data quality, expand ancestral diversity, and employ more robust genetic instruments to enhance the reliability and generalizability of causal inferences.

We are thankful to Prof Stanley Lin for his critical and careful editing and proofreading of the manuscript. We are also thankful to Shantou Key Laboratory of Precision Diagnosis and Treatment in Women’s Cancer for the critical support.