Intracerebral hemorrhage (ICH), a non-traumatic form of spontaneous intracranial bleeding, constitutes the most severe and least amenable variant of stroke (Cordonnier et al., 2018). Accounting for approximately 10%–15% of all strokes, ICH presents a disproportionately high mortality rate, with nearly 40% of affected individuals succumbing within the first month (Fujii et al., 1994; Schrag and Kirshner, 2020). This condition afflicts an estimated two million individuals worldwide annually, eclipsing other intracranial haemorrhage subtypes such as subarachnoid and isolated intraventricular haemorrhages in both prevalence and gravity (Krishnamurthi et al., 2013). The societal ramifications are considerable, extending beyond mortality to include long-term disability and substantial healthcare expenditures (Ruiz-Sandoval et al., 1999).

While substantial strides have been made in mitigating the incidence of ICH through the rigorous management of established risk factors such as hypertension, smoking, and diabetes, it appears that we have reached an epidemiological plateau in these episode preventive endeavours (O'Donnell et al., 2010; Lovelock et al., 2007). To transcend this impasse and further attenuate the episode incidence of ICH, there is an emergent imperative to unearth novel risk determinants and acute triggers (Elkind, 2007). The concept of “triggers” delineated as acute precipitants capable of instigating an ICH episode, is increasingly being accorded pivotal importance. The elucidation of such triggers holds the promise of not only enhancing risk stratification algorithms but also fortifying predictive models and refining both short-term and long-term preventive strategies. Emerging evidence is starting to suggest that emotional states may influence and act as triggers for ICH (Aalbaek et al., 2017; Prasad et al., 2020; Smyth et al., 2022).

Current research on the triggering factors of ICH predominantly employs the case-crossover study design (Haynes, 2011; Liu et al., 2023). This approach is favored for its ability to capture transient exposures that may precipitate acute events like ICH (Maclure, 1991). Additionally, the design inherently controls for stable confounders by allowing each individual to serve as their own control, thereby enhancing the internal validity of the study. The case-crossover design is also more feasible and efficient compared to randomized controlled trials (RCTs) or cohort studies, especially when focusing on acute triggers.

However, this design is not without its limitations. One major drawback is the potential for residual confounding from time-varying factors that are not accounted for in the study. Furthermore, case-crossover studies often lack the rigor of higher-level evidence that could be obtained from controlled or prospective studies. This makes it challenging to definitively establish causality between potential triggers and ICH. Given the severity and complexity of ICH, there is an urgent need for high-quality evidence to conclusively identify its triggers. This would require innovative research methodologies that can eliminate all forms of confounding, thereby providing more reliable and actionable insights into the prevention of ICH.

Mendelian Randomization (MR) offers a robust alternative to traditional epidemiological methods for causal inference, minimizing confounding and reverse causation (Skrivankova et al., 2021; Verbanck et al., 2018; Thanassoulis and O'Donnell, 2009). MR has been widely applied in cardiovascular and cerebrovascular research, including studies on ICH (Larsson et al., 2019; Titova et al., 2020; Xiao et al., 2022; Yu et al., 2022; Wang et al., 2023). This makes it an ideal tool for exploring the yet-to-be-established causal relationship between various emotional states and ICH. Given the critical need for effective ICH prevention, MR could provide the high-quality evidence required to inform better preventive strategies.

In light of these considerations, the objective of this investigation is to employ MR analysis as a rigorous methodological approach to furnish elevated tiers of evidence, meticulously eradicate confounding variables, and discern the specific emotional states causally implicated in ICH.

Our MR study robustly demonstrated a significant causal relationship between the emotional trait “Worrier/anxious feelings” and ICH, supported by 195 instrumental variables (SNPs). Conversely, other emotional traits like “Nervous feelings,” “Sensitivity/hurt feelings,” and “Tense/'highly strung'” did not show significant associations, reinforcing the specificity of our primary finding. Additional tests for heterogeneity and pleiotropy further corroborated these results. For the first time, our Mendelian Randomization study provides compelling evidence that worry and anxiety can directly trigger ICH. This study adds a new dimension to our understanding of ICH triggers, potentially paving the way for more targeted preventive strategies.

In our study, the IVW, IVW Radial, Maximum Likelihood, and MR PRESSO methods all have p-values below 0.05, indicating a statistically significant association between the exposure and the outcome. In contrast, the Weighted Median method has a p-value greater than 0.05, suggesting a lack of statistical significance. The IVW method is commonly used in MR because it combines the effects of individual SNPs weighted by their inverse variance, assuming no horizontal pleiotropy, where a genetic variant affects the outcome through pathways other than the exposure of interest (Lin et al., 2021). The IVW Radial variant accounts for outliers and handles potential pleiotropy better. The Weighted Median method offers a valid causal estimate even if up to 50% of the information comes from invalid instruments, but it is less efficient than IVW and may be influenced by outliers, which could explain its non-significant result in this case (Scholz, 1978). In this study, the IVW method serves as the primary analysis because it generally provides efficient and reliable estimates, assuming no substantial horizontal pleiotropy. The Weighted Median method differs in its assumptions and is less efficient. The other methods offer complementary analyses to address different assumptions and identify potential pleiotropy.

The quest for acute triggers in ICH is becoming increasingly pivotal, especially as traditional risk management approaches appear to have hit a ceiling in reducing incidence rates (Puy et al., 2023). Identifying novel, non-traditional risk factors or preventable triggers, including acute triggers, opens up new avenues for intervention. It allows for the development of more targeted and personalized prevention strategies, which could include behavioral modifications, psychological interventions, or even pharmacological treatments specifically designed to mitigate these triggers. Moreover, the discovery of such triggers could enhance risk stratification models, enabling clinicians to more accurately identify individuals at heightened risk of ICH. This, in turn, could lead to more timely and effective interventions, potentially reducing the incidence and severity of ICH in vulnerable populations. The prospects are promising, offering hope for significant advancements in ICH prevention and management.

The INTERSTROKE study, a comprehensive international case-control investigation, has identified the leading risk factors for stroke. Notably, two of the top ten risk factors—psychosocial stress and depression—are related to emotional states, thereby offering indirect relevance to our own research (O'Donnell et al., 2010). A specific component of the INTERSTROKE study focuses on the role of anger, emotional upset, and heavy physical exertion as acute triggers for stroke. The authors reported that instances of anger or emotional upset occurring within the hour preceding a stroke were relatively common and exhibited a notable association with ICH. This finding is consistent with previous small-scale case-crossover studies (Koton et al., 2004).

In practical terms, case-control and crossover studies are often the most feasible designs for investigating triggers or risk factors for ICH. These study designs are particularly suited for capturing acute triggers and short-term risk factors that may precipitate ICH events (Maclure and Mittleman, 2000). However, to yield credible results, these studies require a large scale, often necessitating substantial investment in terms of human, financial, and time resources. Despite these considerable efforts, the evidence level generated from case-control and crossover studies is not considered the highest, leaving room for further methodological advancements to provide more definitive conclusions.

In the absence of high-quality clinical evidence, our MR study provides a compelling case for a causal relationship between the emotional traits of worry and anxiety and the incidence of ICH. This underscores the importance of considering emotional states as potential acute triggers, thereby offering a nuanced approach to further reducing ICH incidence.

Furthermore, clinical studies have established a link between anxiety or other negative emotions and ICH, although they do not elucidate directional causality. Animal studies, however, provide evidence that ICH can lead to negative emotional states. The study on thalamic ICH highlights its role in triggering Central Post-Stroke Pain (CPSP) and subsequent emotional comorbidities like anxiety and depression, suggesting therapeutic potential for stellate ganglion block (SGB) (Shi et al., 2023). Similarly, Zhu et al. also demonstrate that right hemispheric ICH in mice leads to changes in emotion-related behavior, adding another layer of complexity to the neurological and behavioral impact of ICH (Zhu et al., 2018). In contrast, our Mendelian Randomization study has the unique advantage of providing a unidirectional causal relationship between emotional states and ICH. This strengthens the case for emotional states as potential triggers for ICH and underscores the value of MR in advancing our understanding of these complex relationships.

Last in the list, previous research has robustly demonstrated that negative emotional states like anger can have a significant impact on cardiovascular systems, thereby increasing the risk of cardiovascular diseases (CVD) (Mittleman et al., 1995; Stalnikowicz and Tsafrir, 2002; Smyth et al., 2016). Given that CVD is a well-established major risk factor for ICH, the question arises whether the emotional traits directly trigger ICH or do so indirectly through their impact on CVD (Qureshi et al., 2009; Sheth, 2022). The intricacy of isolating the direct effects of emotional states on ICH is further complicated by the presence of numerous confounding factors in real-world settings. RCTs, although considered the gold standard for establishing causality, are particularly challenging to implement for studying acute triggers like ICH. Ethical considerations, logistical constraints, and the acute nature of the condition make RCTs nearly impractical for this line of research. Even if an RCT were feasible, it would still face the hurdle of controlling for a myriad of confounding variables, such as pre-existing CVD conditions, lifestyle factors, and medication use. Consequently, the challenge remains to design clinical studies that can effectively eliminate these confounders while still being both ethical and logistically feasible. In the absence of such high-level evidence, Mendelian Randomization serves as a valuable alternative for causal inference, albeit with its own limitations.

Our study, while pioneering in its identification of a causal relationship between the emotional traits of worry and anxiety and the incidence of ICH, is not devoid of limitations that warrant discussion. Primarily, the observational nature of the data, despite the methodological rigour afforded by MR, leaves room for residual confounding. Additionally, the paucity of SNPs serving as instrumental variables for certain emotional traits may compromise the statistical power of our analyses. The study’s external validity is also circumscribed by the demographic characteristics of the cohort, cautioning against overgeneralization. Moreover, the intricate nexus between emotional states, cardiovascular diseases, and ICH remains partially obscured, necessitating further research to elucidate potential intermediary mechanisms. It is also pertinent to note that while MR offers a higher echelon of evidence compared to traditional epidemiological methods, it does not supplant the gold standard of randomized controlled trials, which are logistically challenging in this domain. Lastly, the scope of our investigation was inherently constrained by the availability of suitable emotional traits within the genetic databases, limiting our ability to explore the full spectrum of negative emotions that may be implicated in ICH.

Our MR study could provide valuable insights into causal relationships but is subject to limitations. Pleiotropy, where genetic variants influence outcomes through multiple pathways, can confound results, necessitating sensitivity analyses. Population stratification, reflecting genetic diversity across populations, can bias MR results. Researchers should employ statistical techniques or focus on homogeneous populations to mitigate this bias. The strength of instrumental variables is crucial in MR. Weak instruments can lead to biased estimates or reduced statistical power, emphasizing the need for robust SNP selection. Temporal relationships can be ambiguous, especially in cases involving complex feedback loops. Time-lagged analyses and clear hypotheses regarding temporal sequences can help address this ambiguity. MR aims to address confounding but may not fully account for all forms, such as horizontal pleiotropy or genetic confounding. Robust MR methods like MR-Egger can help detect and adjust for these confounders. Generalizability is a concern, as MR studies often rely on specific populations. Validation in diverse populations enhances external validity. Data availability is another challenge, as MR studies depend on existing data sources. Leveraging comprehensive databases and transparent reporting can help mitigate this limitation.

From a clinical perspective, the identification of worry and anxiety as potential acute triggers for ICH bears significant implications for preventive medicine. This insight could catalyze the development of targeted interventions, including cognitive-behavioral therapy, mindfulness-based stress reduction, or pharmacological treatments. Such approaches hold promise in ameliorating the impact of negative emotional states on ICH risk, thereby offering a novel avenue for personalized healthcare. Furthermore, our findings extend to broader implications for public health, informing the design of regimen and educational programs aimed at emphasizing the importance of emotional wellbeing in stroke prevention. By elucidating the connection between mental health and stroke, heightened awareness and early interventions may be fostered, potentially reducing the incidence of ICH. Additionally, understanding the causal relationship between emotional traits and ICH stands to influence clinical practice by advocating for the integration of mental health assessments into routine care for at-risk stroke patients. This integrated approach holds potential for enhancing early detection and intervention strategies. In summary, the clinical and public health implications of our study are multifaceted. They have the potential to instigate a paradigm shift in the prevention and management of ICH, underscoring the necessity of holistic approaches that address both physical and emotional health.

In conclusion, our Mendelian Randomization investigation furnishes robust evidence supporting a causal nexus between emotional dispositions of worry and anxiety and the onset of ICH. This seminal finding augments the existing paradigm, which has predominantly focused on traditional risk factors, by introducing acute emotional triggers as a crucial element in the etiology of ICH. While acknowledging the constraints of our study, we posit that our results serve as an impetus for future multi-disciplinary research aimed at elucidating the complex interplay between emotional states and cardiovascular pathologies in the context of ICH. Ultimately, our insights could inform the development of more sophisticated risk stratification algorithms and targeted interventions, thereby enriching the armamentarium for comprehensive ICH management.