Insights Into Stroke Risk and Prediction for Underweight Individuals

Summary

This review explores the relationship between underweight status and stroke risk and outcomes. Although the impact of obesity on stroke is well-documented, the harmful effects of being underweight are often overlooked. Drawing on evidence from observational studies, case-control studies, and meta-analyses, this review highlights that underweight individuals face increased mortality, poorer functional recovery, and higher rates of major adverse cardiovascular events and recurrent strokes post-stroke. Challenges in stroke prediction for this population are examined, including the limited research focus, shortcomings of existing risk assessment tools such as the CHA₂DS₂-VASc score, and population-specific factors affecting underweight prevalence and outcomes. The review identifies key research gaps, emphasizing the need for longitudinal studies encompassing diverse populations to inform tailored preventive strategies and improve risk assessment. It concludes by advocating for individualized risk assessment approaches that integrate nutritional status and targeted interventions to reduce the elevated stroke risks associated with being underweight.

Introduction

Low body weight, specifically being underweight (body mass index [BMI] <18.5 kg/m²), is a critical factor influencing stroke risk and patient outcomes. Studies consistently show that underweight individuals face a heightened risk of stroke and worse post-stroke outcomes. A comprehensive review highlighted an increased risk of mortality and reduced functional recovery after strokes in underweight individuals, in contrast to the “obesity paradox” [1]. Additional research has established a correlation between the severity of being underweight and the incidence of stroke, myocardial infarction, and all-cause mortality [2]. Underweight patients without atrial fibrillation exhibit a significantly higher risk of major adverse cardiovascular events and recurrent stroke within 1 year post-stroke [3]. Furthermore, being underweight is associated with a higher risk of hemorrhagic stroke, particularly in men [4]. A meta-analysis confirmed that underweight individuals experience shorter post-stroke survival times compared to those with normal or higher BMI [5]. Maintaining a stable body weight is essential because both weight loss and gain are linked to increased ischemic stroke risk, suggesting that weight fluctuations may exacerbate stroke risk for underweight individuals [6]. Low body weight also poses challenges for stroke prevention in patients with non-valvular atrial fibrillation because it is a risk factor for adverse outcomes [7]. Moreover, underweight patients undergoing mechanical thrombectomy have poorer survival rates and functional outcomes, indicating that low BMI adversely affects recovery [8]. Finally, cortical infarcts appear to be more prevalent in underweight patients, potentially influencing stroke presentation and management strategies [9]. In summary, these findings underscore the importance of incorporating body weight, particularly underweight status, into stroke risk assessment and management protocols. The objectives of this review are to synthesize evidence regarding the predictors of stroke in underweight individuals, identify gaps in existing research, and propose directions for future studies.

Challenges in stroke prediction for underweight individuals

Why are underweight individuals less researched?

Several factors contribute to the relative lack of research on underweight individuals compared with overweight and obese populations. The lower prevalence of underweight individuals in developed nations, combined with the substantial attention devoted to the obesity epidemic and its associated health risks (e.g., cardiovascular diseases and diabetes), has resulted in a disproportionate focus on obesity [10]. Consequently, research funding and public health resources are often directed toward addressing obesity, overshadowing the health implications of being underweight [11]. Although studies indicate increased mortality risks for underweight individuals—particularly from external causes—these findings receive less emphasis than the widely recognized risks of obesity [12]. Demographic characteristics of underweight populations, such as younger age, higher smoking rates, and lower socioeconomic status, further complicate the analysis and interpretation of health outcomes, resulting in less definitive conclusions compared to other BMI categories [13]. The simultaneous challenges of underweight and overweight populations, especially among lower socioeconomic groups in developing countries, introduce additional complexity to public health strategies because these nations face the dual burden of malnutrition [14]. Vulnerable groups, including the elderly, individuals with chronic diseases, and those living in extreme circumstances, often dominate the underweight population but are not typically the primary focus of health research in wealthier nations [15]. Methodological challenges, such as reliance on self-reported BMI data, can also introduce inaccuracies in assessing the true health risks associated with being underweight, further contributing to the research gap [16]. Overall, the combined effects of lower prevalence, competing public health priorities, and methodological difficulties have led to the under-representation of research on underweight individuals compared with their overweight and obese counterparts.

Limitations of existing models

Stroke risk prediction tools, such as the CHA₂DS₂-VASc score, are commonly used to assess stroke risk in patients with atrial fibrillation, but they have notable limitations, particularly for underweight individuals. The CHA₂DS₂-VASc score, which considers factors such as congestive heart failure, hypertension, age, diabetes, stroke history, vascular disease, and sex, serves as a standard guide for anticoagulation therapy in patients with atrial fibrillation [17]. However, its predictive accuracy is moderate, with C-statistics ranging from 0.63 to 0.68, highlighting the need for improvement [18]. A significant limitation is the exclusion of BMI, an independent predictor of stroke risk, because lower BMI—often linked to being underweight—is associated with an increased risk of ischemic stroke, regardless of the CHA₂DS₂-VASc score [19]. This omission may lead to under-treatment with anticoagulants in underweight individuals because of an underestimation of their stroke risk. Furthermore, the score excludes other emerging risk factors, such as renal function, inflammation, and genetic predispositions, which could be particularly relevant for underweight patients with distinct physiological profiles [20]. While modifications like the CHA₂DS₂-VA and other variants have been proposed, they have not achieved widespread clinical adoption [21]. Developing more comprehensive risk assessment tools that incorporate additional factors, including BMI, is crucial for identifying high-risk patients, particularly those who are underweight, and for tailoring anticoagulation therapy more effectively [19]. Table 1 summarizes the most recognized models, their applications, and the gaps in addressing underweight-specific stroke risks.

Table 1.

Overview of key stroke risk prediction models: Their applications and gaps in addressing underweight-specific risks.

Model	Key features	Strengths	Limitations	References
CHA2DS2-VASc	Considers congestive heart failure, hypertension, age, diabetes, stroke history, vascular disease, and sex.	Widely used; simple and easy to apply; provides a baseline for anticoagulation therapy in atrial fibrillation patients.	Moderate predictive accuracy (C-statistics: 0.63–0.68); excludes BMI and emerging risk factors.	Abouzid et al. (2024) [17]; Basit et al. (2023) [18].
CHA2DS2-VA	A modified version of CHA2DS2-VASc excluding the sex category.	Attempts to refine prediction accuracy by simplifying risk factors.	Limited clinical adoption; does not fully address underweight-specific risks or incorporate BMI.	Rubanenko (2023) [21].
HAS-BLED	Assesses bleeding risk in patients undergoing anticoagulation therapy, considering hypertension, renal/liver disease, stroke, bleeding history, age, drugs/alcohol.	Effective in predicting bleeding risk; complements CHA2DS2-VASc for anticoagulation decision-making.	Not specifically designed to assess stroke risk; focuses on bleeding complications.	Harrington et al. (2023) [19].
Framingham Score	Evaluates 10-year risk of stroke using factors like age, sex, blood pressure, cholesterol, and smoking.	Longstanding tool for cardiovascular risk assessment; validated in diverse populations.	Does not include atrial fibrillation, BMI, or malnutrition-related factors relevant for stroke risk.	Kwon et al. (2021) [2]; Cho et al. (2019) [6].
ABC Stroke Score	A model incorporating age, biomarkers (e.g., NT-proBNP, troponin), and clinical history to assess stroke risk.	Integrates biomarkers for a more comprehensive evaluation.	Requires biomarker testing; limited adoption in routine clinical practice.	Rachieru et al. (2023) [20]; Cho et al. (2022) [22].
PROGRESS Model	Focuses on predicting recurrent stroke risk post-initial stroke using blood pressure control, lifestyle factors, and medication adherence.	Addresses secondary prevention of stroke; tailored for post-stroke populations.	Not designed for first-stroke prediction or underweight-specific risk factors.	Sun et al. (2017) [31]; Fauzi et al. (2024) [42].
NIHSS (Stroke Severity)	Assesses acute stroke severity based on clinical symptoms; commonly used in emergency settings.	Effective for acute stroke severity and outcome prediction.	Not intended for long-term stroke risk prediction or prevention.	Zhang et al. (2022) [5]; Liu et al. (2023) [38].

Population-specific challenges

The underrepresentation of underweight individuals in large cohort studies is a complex issue influenced by demographic characteristics, health risks, and societal perceptions. Despite evidence of significant health risks, underweight individuals (BMI of <18.5 kg/m²) receive less research attention than individuals with obesity. Studies indicate that underweight individuals face elevated mortality risks, particularly from external causes, and are more vulnerable to conditions such as tuberculosis, with hazard ratios markedly higher than those of normal-weight individuals [22]. The prevalence of underweight status varies across populations and is shaped by sociodemographic factors. For instance, in Korea, underweight status is more common among women, older adults, and low-income individuals, with disparities growing over time [23]. In Denmark, underweight status is more prevalent among girls than boys, and socioeconomic status appears to have minimal influence on its prevalence [24]. In developed countries, societal focus on obesity often overshadows the challenges of being underweight, while in developing regions, underweight prevalence leads to severe health consequences [15]. Additionally, underweight individuals often exhibit distinct demographic and lifestyle traits, such as younger age, higher smoking rates, and lower physical activity levels, making their inclusion in obesity-focused studies more challenging [25]. Societal pressures to maintain a thin physique, particularly among young women, exacerbate health risks such as anemia and osteoporosis, further complicating public health approaches [26]. Addressing the underrepresentation of underweight individuals in research requires a comprehensive approach that acknowledges their unique health risks and the sociodemographic factors contributing to their underweight status. This would enable the development of inclusive public health strategies to address the needs of all weight categories, ultimately improving overall population health outcomes [2]. Predictive factors for BMI vary considerably across BMI groups, reflecting diverse health outcomes and associated risk factors. Among individuals with a BMI of 35.0 to 39.9 kg/m², the risk of coronary heart disease is notably lower than in those with a normal BMI, suggesting a complex relationship between BMI and cardiovascular risk. In this group, age and diabetes are independent predictors of coronary heart disease, while alcohol consumption appears to reduce the risk [27]. In youth, psychological and parental factors, such as body weight perception and parental obesity, are key predictors of higher BMI percentiles, highlighting the impact of sociocultural influences [28]. Patients with higher BMI undergoing elective percutaneous coronary intervention tend to be younger and more frequently present with diabetes, hypertension, and dyslipidemia, indicating a distinct risk profile compared with normal-weight individuals [29]. BMI is a more reliable predictor of obesity-related risk factors than body fat percentage, with age and psychological factors, such as stress and compulsive eating, playing significant roles in BMI models [30]. In patients with acute pancreatitis, obesity is associated with increased mortality and complications, although BMI alone does not reliably predict disease severity. (Ince et al., 2022). These findings underscore the complexity of BMI as a predictive factor, demonstrating that risk factors and health outcomes vary across BMI groups. This situation necessitates tailored interventions and assessments that consider specific demographic and health profiles.

Key findings from existing studies

Observational studies consistently show that underweight individuals (BMI of <18.5 kg/m²) face increased risks of adverse outcomes following a stroke. Underweight individuals have higher mortality rates and poorer functional recovery than normal-weight individuals. A systematic review and meta-analysis found that underweight patients have a significantly higher risk of long-term mortality, with an adjusted hazard ratio of 1.65 (Mehta et al., 2022). Similarly, an umbrella review confirmed a higher post-stroke mortality risk for underweight individuals, in contrast to the reduced mortality observed in overweight and obese patients (a phenomenon known as the “obesity paradox”) [1]. Underweight status is also linked to an elevated risk of major adverse cardiovascular events (MACE) and recurrent strokes, particularly in patients without atrial fibrillation, with an adjusted hazard ratio for MACE of 1.66 [3]. The severity of the underweight status correlates with an increased incidence of stroke, myocardial infarction, and all-cause mortality, with hazard ratios rising alongside the severity of the underweight status [2]. In patients with acute ischemic stroke, being underweight is a predictor of unfavorable outcomes, including higher mortality and greater dependency [31]. Furthermore, low body weight is independently associated with increased ischemic stroke and major bleeding risks in patients with non-valvular atrial fibrillation [7]. Although anemia has been identified as a stronger predictor of mortality in some studies, underweight status remains a significant independent predictor of poor outcomes [32]. The nutritional status at the time of stroke, including undernutrition, is a critical determinant of long-term outcomes, with undernourished patients experiencing higher mortality and complication rates [33]. These findings underscore the need to consider body weight and nutritional status in stroke management and prognosis, emphasizing the importance of targeted interventions to address undernutrition [5]

Case-control studies and related research consistently highlight the significant impact of being underweight as a predictor of adverse outcomes in stroke patients. Underweight individuals (low BMI) face an elevated risk of poor post-stroke prognosis. For example, underweight patients without atrial fibrillation show a markedly higher risk of MACE and recurrent stroke within 1 year post-stroke, with adjusted hazard ratios of 1.66 for MACE and 1.50 for recurrent stroke compared with normal-weight individuals [3]. Undernutrition at admission is also a critical predictor of complications and poor outcomes in patients with acute stroke, underscoring the importance of assessing nutritional status early [33,34]. Findings from the FOOD trial (2003) support this, showing higher mortality rates among undernourished patients with stroke, with an odds ratio of 2.32 even after adjusting for confounding factors [33]. A meta-analysis further confirmed that being underweight is associated with increased mortality risk, with a pooled hazard ratio of 1.71 compared with normal-weight individuals [5]. Additionally, a non-linear relationship between BMI and functional outcomes was observed, with underweight patients facing a greater likelihood of poor functional recovery post-stroke [35]. A systematic review also demonstrated that malnutrition at the time of stroke significantly increases the risk of long-term mortality and poor functional outcomes, with an adjusted odds ratio of 2.38 for all-cause mortality [36]. These findings collectively highlight the urgent need for early nutritional assessment and targeted interventions in underweight stroke patients to improve outcomes and reduce the risk of recurrent cardiovascular events.

Meta-analyses and systematic reviews consistently demonstrate that being underweight (BMI <18.5 kg/m²) is a strong predictor of adverse outcomes after stroke, including increased mortality and poor functional recovery. A meta-analysis revealed significantly higher mortality hazard ratios for underweight patients than for those with normal weight, with a pooled hazard ratio of 1.71 for mixed-stroke patients and 1.53 for acute ischemic stroke patients [5]. Additionally, underweight status is associated with a higher incidence of MACE and recurrent strokes, particularly in patients without atrial fibrillation, with an adjusted hazard ratio of 1.66 for MACE [3]. The “obesity paradox,” where overweight and obese individuals exhibit lower post-stroke mortality, underscores the vulnerability of underweight individuals [1,5]. The nutritional status at the time of stroke is another critical predictor of outcomes because undernutrition independently correlates with complications, infections, gastrointestinal bleeding, and higher mortality [37]. The FOOD trial reinforced these findings, linking undernutrition with significantly increased mortality and complication rates. The literature highlights both consensus and debate regarding the impact of being underweight on stroke outcomes. Consistently, underweight individuals experience greater risks of long-term mortality, recurrent strokes, and adverse cardiovascular events, particularly in the absence of atrial fibrillation [1,3,36]. Worse outcomes have also been observed in underweight patients undergoing mechanical thrombectomy, with poorer survival and functional recovery reported below a certain BMI threshold [8]. Malnutrition, commonly accompanying underweight status, exacerbates risks by increasing susceptibility to infections, prolonging hospital stays, and raising mortality rates [38]. Pathophysiological mechanisms, such as heightened metabolic demands and protein breakdown following stroke, contribute to systemic complications and impaired recovery [39]. However, there is debate about the precise relationship between being underweight and functional recovery, as some studies report inconclusive findings [1]. Moreover, variability in nutritional assessment tools and their diagnostic accuracy complicates efforts to fully understand the role of malnutrition in stroke recovery, underscoring the need for standardized screening protocols [38,40]. Despite these challenges, a consensus exists on the importance of addressing nutritional status in stroke management, with evidence suggesting that targeted nutritional interventions could improve outcomes for underweight stroke patients [36,40]. While the detrimental effects of being underweight post-stroke are well-established, further research is needed to elucidate underlying mechanisms and develop effective risk mitigation strategies. Table 2 summarizes some key findings related to studies related to the association between underweight and stroke.

Table 2.

Findings from existing studies.

Key Area	Findings	Sources
Mortality Risks	Underweight individuals face significantly higher mortality risks after stroke, with hazard ratios ranging from 1.65 to 1.71.	Mehta et al. (2022) [36]; Zhang et al. (2022) [5].
Functional Recovery	Poorer functional recovery post-stroke is consistently observed among underweight individuals compared to those with normal BMI.	Holland et al. (2024) [1]; Wakisaka et al. (2023) [35].
Major Adverse Cardiovascular Events	Higher risks of MACE and recurrent strokes are reported, especially in underweight patients without atrial fibrillation. Hazard ratio for MACE: 1.66.	Ju et al. (2022) [3]; Mehta et al. (2022) [36].
Stroke Subtypes and Outcomes	Increased risks of hemorrhagic and ischemic strokes, with worse outcomes for underweight patients undergoing treatments like mechanical thrombectomy.	Shiozawa et al. (2021) [4]; Fecker et al. (2024) [8].
Nutritional Status	Malnutrition at the time of stroke predicts complications, infections, and increased mortality, emphasizing the importance of early nutritional assessment.	Choi et al. (2024) [37]; FOOD trial (2003).
Obesity Paradox	Contrasts with underweight risks, showing reduced mortality for overweight/obese patients, highlighting unique vulnerabilities of underweight individuals.	Holland et al. (2024) [1]; Zhang et al. (2022) [5].
Pathophysiology and Risks	Heightened metabolic demands and protein breakdown in underweight individuals exacerbate systemic complications and recovery challenges post-stroke.	Amalia et al. (2024) [39]; Di Vincenzo et al. (2024) [40].
Debates and Challenges	Variability in nutritional assessment tools complicates conclusions, underscoring the need for standardized methods to evaluate malnutrition and stroke recovery links.	Liu et al. (2023) [38]; Di Vincenzo et al. (2023) [40].

Implications for clinical practice

Importance of individualized risk assessment

Clinicians should account for several unique factors influencing outcomes when evaluating stroke risk in underweight patients (BMI of <18.5 kg/m²) because this population faces an elevated risk of adverse post-stroke outcomes, including higher mortality and poorer functional recovery. This increased risk is partly attributable to the strong association between low body weight and malnutrition, which is prevalent in patients with stroke and exacerbates outcomes [1,36,41]. Malnutrition, whether pre-existing or post-stroke, correlates with longer hospital stays, increased mortality, and poorer functional outcomes, emphasizing the importance of early nutritional assessment and intervention [41]. Post-stroke metabolic demands, particularly in cases of subarachnoid hemorrhage, require careful nutritional management to support recovery [41]. Additionally, underweight patients are at higher risk for MACE and recurrent strokes, particularly in the absence of atrial fibrillation, underscoring the need for vigilant cardiovascular monitoring [3]. The severity of being underweight is proportionally linked to increased risks of stroke, myocardial infarction, and all-cause mortality, indicating that even mild underweight should not be overlooked [2]. Furthermore, underweight patients undergoing mechanical thrombectomy often have worse survival and functional outcomes, suggesting that frailty and low BMI are significant predictors of poor prognosis [8]. Nutritional indices such as the prognostic nutritional index score and Controlling Nutritional Status score provide valuable prognostic insights and should be incorporated into routine clinical evaluations [5]. Given these findings, clinicians should prioritize comprehensive nutritional assessments and implement tailored nutritional interventions to address the heightened risks associated with being underweight in patients with stroke.

Need for tailored preventive strategies

Specific lifestyle interventions and modifications can effectively reduce stroke risk in underweight individuals by addressing modifiable risk factors and promoting overall health. Regular is vital because it helps manage blood lipid abnormalities, diabetes, and hypertension—major stroke risk [42]. Engaging in at least 30 minutes of moderately intense exercise on most days is recommended for substantial stroke risk reduction [43]. Dietary modifications also play a crucial role, with a Mediterranean diet—rich in fruits, vegetables, fish, fibers, and low-fat dairy, and supplemented with olive oil and nuts—being associated with reduced stroke risk [44]. This diet not only aids in weight management but also provides essential nutrients that lower stroke risk. Smoking cessation is critical because smoking is a well-established, independent risk factor for stroke [43,45]. Similarly, reducing alcohol consumption to low-to-moderate levels can be beneficial; while excessive intake raises stroke risk, moderate consumption may confer some protective effects [43,45]. Managing stress and minimizing chronic stress exposure are also important lifestyle modifications [44]. Innovative approaches, such as computerized phone-based lifestyle coaching systems, have shown promise in supporting healthier habits, improving blood pressure, lipid profiles, and dietary practices [46]. Collectively, these interventions target multiple modifiable risk factors and provide a comprehensive approach to significantly reduce stroke risk in underweight individuals while fostering a healthier overall lifestyle.

Research gaps and future directions

Current research highlights several gaps in understanding underweight populations, emphasizing the need for longitudinal studies to provide more comprehensive insights. While some studies have begun examining longitudinal factors associated with being underweight in children, focusing on maternal characteristics and early childhood influences [47], there remains a critical lack of data on how these factors evolve over time and across diverse populations. The absence of robust demographic and morbidity data for underweight individuals complicates health outcome assessments compared with other BMI categories [25]. In Korea, disparities in underweight prevalence across sociodemographic groups point to the necessity of longitudinal research to better understand these trends and their broader implications [23]. Additionally, the limited exploration of dietary patterns and their relationship to the double burden of malnutrition—including underweight status—underscores the need for longitudinal approaches to elucidate these dynamics [48]. Research on Danish adolescents has revealed a stable prevalence of underweight status over two decades; however, the underlying causes, such as malnutrition or eating disorders, remain inadequately explored, warranting further longitudinal investigations to track causative factors and the impact of interventions [24]. Collectively, these studies indicate that while some progress has been made, longitudinal studies are essential to uncover the life course and intergenerational factors affecting underweight populations and to inform the development of targeted, evidence-based interventions.

Integrating diverse populations into predictive model development is essential to improve accuracy, fairness, and generalizability across demographic groups. Traditional models often exhibit biases due to the over-representation of specific populations, particularly those of European ancestry, resulting in inaccurate predictions and exacerbating health disparities among underrepresented groups [49,50]. For instance, polygenic risk scores derived from multi-ancestry genome-wide association studies demonstrate better performance in diverse populations than do those from single-ancestry genome-wide association studies, underscoring the importance of incorporating varied genetic backgrounds in model training [51]. Diversity-aware population models, such as those from the ABCD cohort study, account for socio-spatial factors and intrinsic population stratification, yielding more precise and interpretable insights, particularly in fields like population neuroscience [52]. Including diverse populations also addresses algorithmic bias in clinical prediction models, where demographic factors, including race, can provide prognostically significant information and ensure equitable healthcare decisions [53]. Machine learning strategies such as the Diverse Class-Aware Self-Training (DCAST) method further mitigate selection bias and enhance model robustness by incorporating diverse datasets [54]. Additionally, the PRIMED Consortium emphasizes the use of appropriate population descriptors to avoid reinforcing biological determinism and to ensure ethical applications of genomic research [55]. By integrating diverse populations, predictive models achieve better performance and equity, addressing the intricate interplay of genetic, socioeconomic, and geographic factors that influence health outcomes.

In conclusion, underweight status is a strong, independent predictor of adverse stroke outcomes, including higher mortality, recurrent events, and poorer functional recovery. Current risk models often underestimate stroke risk in this group because they exclude BMI and nutritional status. Incorporating comprehensive nutritional assessment and individualized preventive strategies into stroke care is essential. Future research should focus on longitudinal studies and the development of inclusive, diversity-aware predictive models to better capture the unique vulnerabilities of underweight populations.

Acknowledgments

We would like to thank Proof-Reading-Service.com for their professional English language editing services.

Disclosure

Authors have no conflict of interest, and the work was not supported or funded by drug company.