Risk Factors and Acute-Phase Blood Markers in Stroke Among Young Adults: A Hospital-Based Case-Control Study from South India

Abstract

Background:

Early-onset stroke carries profound health and socioeconomic consequences and is an escalating concern owing to increased cardiovascular risk in young adults. This study explores lifestyle and comorbid risk factors in predicting stroke occurrence among young Indian adults. We also analyze post-admission blood markers to reveal acute-phase metabolic and inflammatory dysregulation.

Methods:

Unicentric case-control study of 40 cases of imaging confirmed acute stroke and 81 sex-matched controls all aged ≤45. Self-reported demographic, lifestyle and comorbidity histories, and laboratory reports within 6–8 hours post-admission were collected. Conditional and multivariate logistic regression estimated stroke odds associated with risk factors. Age- and sex-stratified analyses explored clustering of risk and pathophysiological derangements among subgroups.

Results:

Hypertension (OR: 22.0, P < 0.001) and regular alcohol consumption (OR: 6.7, P = 0.047) independently conferred the strongest stroke odds. The youngest subgroup (18–30 years) carried 17-fold higher odds with alcohol use (P = 0.024) and had elevated bilirubin (P = 0.014). Elevated white blood cells, creatinine, post-prandial blood sugar, and hyponatremia were found among cases, with females exhibiting higher platelets (P = 0.002) and depressed ALT (P = 0.010), and males, elevated hemoglobin (P = 0.027). Significant risk-factor clustering was found between hypertension and diabetes among male and female cases (P = 0.027, P = 0.041) with complete concomitance of smoking and alcohol use in male cases and controls (P = 0.031, P = 0.012).

Conclusions:

This study compellingly identifies hypertension and alcohol use as major, independent predictors of early-onset stroke, underscoring distinct pathogenesis in this cohort and the urgent need for early detection, aggressive management, and behavioral modification. Stratified analyses further unveiled novel insights on age- and sex-specific heterogeneity in risk profiles and pathophysiology.

INTRODUCTION

While stroke is a leading cause of death and disability worldwide, low- and middle-income countries (LMICs) continue to bear disproportionate burden of global incidence, with a notable increase in incidence of early-onset stroke among adults aged under 50 in Asia and Africa.[1,2,3,4] In India, urbanization and the concurrent adoption of Western lifestyles have significantly altered the cardiovascular risk landscape, particularly among the young, with rising prevalence of environmental factors like pollution, lifestyle factors like sedentary lifestyles, chronic stress, unhealthy diets, smoking or alcohol use, and comorbidities like hypertension and diabetes, all of which are traditionally linked with stroke.[5,6,7] As a consequence, the average stroke patient in developing nations is around 15 years younger than in developed countries, with studies from India reporting that 10–15% of stroke cases could occur in individuals aged below 40.[8,9]

Moreover, in contrast to stroke in the elderly, who typically experience stroke secondary to chronic vascular changes and multiple comorbidities, stroke in young adults manifests predominantly through distinct and more acute inflammatory, vascular, and metabolic dysregulation.[5,8,10] Emerging evidence hints at a contributory role of blood markers like white blood cell (WBC) and platelet counts, C-reactive protein, glucose, insulin, iron, and lipids in their pathogenesis and prognosis.[11]

Despite the growing recognition as a significant public health concern in young adults, a comprehensive analysis of factors predicting risk or prognosis in this cohort remains limited, with existing literature predominantly derived from Western populations that may not adequately represent the unique genetic, environmental, and lifestyle contexts of the Asian continent.[6] This case-control study contrasts conventional and emerging lifestyle and comorbid risk factors among first-ever stroke cases and selected controls, all under the age of 45,[12,13] to identify predictors of risk and quantify relative strengths of association. Beyond risk factors, the study also aims to leverage readily available post-admission acute-phase blood markers to investigate the acute pathophysiological derangements following stroke in this cohort.

METHODS

This unicentric, case-control study involving adult human participants received ethical approval from the Institutional Ethics Committee and was conducted in accordance with the Declaration of Helsinki (2013). In view of altered mental statuses following acute-stroke (cases) or other illnesses (controls), patients who were determined to have clinical decision-making capacity during the interview provided consent for participation while healthcare proxies consented on behalf of patients without capacity. Written informed consent was obtained from all participants and proxies prior to data collection, with the approved consent form read in the respondent’s own language.

Between September and October 2023, patients 18–45 years of age[12,13] presenting to the emergency department and subsequently admitted at Government Stanley Hospital in Chennai with stroke-like symptoms were identified via prospective consecutive sampling [Figure 1]. Following CT/MRI confirmation of acute CVA, 40 patients were recruited as cases. Concurrently, 94 inpatients admitted to the general medicine department within the same period with unrelated diagnoses were initially considered for control selection. Through detailed history-taking and meticulous review of personal medical records, 81 participants without prior history of stroke or other neurological diagnoses were recruited via purposive hospital-based sampling from the general medicine department; these were matched by sex to form the final control group at a 2:1 ratio, ensuring both groups originated from a common institutional catchment area to minimize selection bias. Based on an expected risk factor prevalence of 27.5% in cases and 5.9% in controls,[14,15,16] the sample size was calculated to detect a minimum odds ratio of 6.0 with 80% power and a 5% significance level (alpha = 0.05); accounting for 10% potential attrition, a final target of 40 cases and 81 controls was established.

Figure 1.

Study flow diagram depicting recruitment of the study population

Self-reported data on basic demographics, lifestyle, and comorbidities collected via structured questionnaire [Annexures ^{(2.5MB, pdf)} ], and results from hematological and biochemical tests performed on samples collected within 6–8 hours post-admission were recorded. All blood analytes were processed on industry-standard, fully automated clinical biochemistry and hematology analyzers as per institutional quality-control protocols. Lifestyle factors (smoking or alcohol consumption) were determined via detailed participant histories (current and/or regular consumption), while comorbidities (such as hypertension and diabetes) were established based on a combination of self-reported medical history and review of prior clinical records and medications.

Statistical analyses were conducted using Stata 18.5 (StataCorp, Texas). Descriptive characteristics were summarized using proportions with Chi-square or Fisher’s exact tests for between-group comparisons of categorical variables. Laboratory parameters were compared via independent t-tests and Wilcoxon rank-sum tests based on normality. Unadjusted association of lifestyle, comorbidity, and post-admission blood parameters with stroke incidence were evaluated using conditional logistic regression accounting for the 1:2 sex-matched design, with non-normally distributed data transformed via rank-based inverse normal transformation. Given case-control matching solely by sex, age was significantly different between groups and thus adjusted for all subsequent analyses. Comprehensive stratified analyses were performed across age and sex subgroups using logistic regression for categorical and independent t-test or Wilcoxon rank sum for continuous parameters to explore effect modification. Multivariable logistic regression models were constructed by incorporating significant covariates from bivariate analyses (P < 0.2) alongside age and sex as essential confounders. Exploratory analysis using Fisher’s exact test was finally conducted to investigate potential concomitance of key risk factors and pathophysiological derangements within specific strata of the study population.

RESULTS

The final sample comprised 40 cases and 81 controls aged 18–45 years, all included in subsequent univariate and multivariate analyses [Figure 2]. Post-hoc analysis confirmed no statistically significant difference in sex distribution between the two groups, with males constituting 60% of the cases and 52% of the controls (χ² =0.72). Cases were significantly older than controls, with median ages of 39 years (IQR: 34–43) versus 27 years (IQR: 21–35), respectively (Wilcoxon rank-sum: z = -4.45, P < 0.001).

Figure 2.

Demographic composition of study participants in both groups

Bivariate analyses using Chi-square or Fischer’s exact tests revealed substantial differences in lifestyle and comorbid risk factor prevalence between the groups [Table 1]. Hypertension demonstrated the most elevated prevalence in stroke cases compared to controls, with alcohol consumption and diabetes mellitus also significantly associated with stroke occurrence, Figure 3. While smoking demonstrated a weaker association, and other risk factors like ischemic heart disease, family history, and hypercholesterolemia were more prevalent in cases, these differences did not reach significance. No differences emerged for anemia, hypothyroidism, and the other risk factors. Most acute-phase analytes were within reference range and with prevalence of abnormal results comparable among most cases and controls [Table 2]. Fasting (FBS) and post-prandial (PPBS) blood sugars, creatinine, blood urea nitrogen (BUN), WBC counts, and high-density lipoproteins (HDLs) were most deranged among cases, Figure 4.

Table 1.

Demographic, lifestyle, and comorbid characteristics of study participants (Cases=40, Controls=81)

		Controls, n (%)		Cases, n (%)		P*
Sex
Male		42 (52)		24 (60)		0.397
Female		39 (48)		16 (40)
Age (years)
18–24		32 (39)		4 (10)		<0.001
25–34		28 (35)		6 (15)
35–45		21 (26)		30 (75)
Lifestyle
Smoking		2 (2.5)		4 (10.0)		0.092
Alcohol		5 (6.2)		11 (27.5)		0.003
Substance abuse		2 (2.5)		0 (0.0)		1.000
Comorbidities
Hypertension		3 (3.7)		19 (47.5)		<0.001
Diabetes		15 (18.5)		16 (40.0)		0.015
CAD/IHD		1 (1.2)		2 (5.0)		0.254
Congenital Heart Disease		1 (1.2)		1 (2.5)		1.000
Hypercholesterolemia		6 (7.4)		4 (10.0)		0.728
Anemia		22 (27.2)		9 (22.5)		0.662
Hypothyroidism		7 (8.6)		3 (7.5)		0.417
Hyperhomocysteinemia		0 (0.0)		2 (5.0)		0.107
Polycythemia		3 (3.7)		4 (10.0)		0.218
Family h/o stroke		0 (0.0)		1 (2.5)		0.331

*P-values derived using Fisher’s exact test based on cell counts

Figure 3.

Lifestyle and comorbid risk factor prevalence across groups

Table 2.

Postadmission acute-phase metabolic and inflammatory blood markers of study participants

		Controls, n (%)		Cases, n (%)		P*
Hemoglobin
Reduced		30 (37)		14 (35)		0.250
Normal		45 (56)		19 (47)
Elevated		6 (7)		7 (18)
WBC count
Normal		66 (84)		30 (75)		0.327
Elevated		14 (16)		10 (25)
Platelets
Reduced		4 (5)		1 (3)		1.000
Normal		70 (86)		35 (87)
Elevated		7 (9)		4 (10)
Sodium
Reduced		8 (10)		6 (15)		0.090
Normal		65 (80)		34 (85)
Elevated		8 (10)		0 (0)
Potassium
Reduced		4 (5)		3 (7)		0.718
Normal		75 (93)		37 (93)
Elevated		2 (2)		0 (0)
Albumin
Normal		54 (67)		25 (63)		0.688
Reduced		27 (33)		15 (37)
Total protein
Normal		77 (95)		37 (92)		0.683
Reduced		4 (5)		3 (8)
Triglycerides
Normal		24 (49)		22 (56)		0.525
Elevated		25 (51)		17 (44)
HDL
Normal		55 (68)		21 (53)		0.113
Reduced		26 (32)		19 (47)
BUN
Normal		40 (51)		19 (47)		0.846
Elevated		38 (49)		21 (53)
Creatinine
Normal		77 (96)		35 (87)		0.115
Elevated		3 (4)		5 (13)
AST
Normal		29 (83)		19 (90)		0.696
Elevated		6 (17)		2 (10)
ALT
Normal		41 (89)		32 (91)		1.000
Elevated		5 (11)		3 (9)
Indirect bilirubin
Normal		40 (98)		28 (85)		0.083
Elevated		1 (2)		5 (15)
Total bilirubin
Normal		46 (96)		30 (86)		0.126
Elevated		2 (4)		5 (14)
TSH
Reduced		0 (0)		1 (3)		0.854
Normal		26 (84)		34 (85)
Elevated		5 (16)		5 (12)
Fasting Blood Sugar
Normal		29 (73)		21 (54)		0.105
Elevated		11 (28)		18 (46)
Postprandial Blood Sugar
Normal		32 (80)		25 (66)		0.204
Elevated		8 (20)		13 (34)

*P-values derived using Fisher’s exact test based on cell counts

Figure 4.

Postadmission acute-phase blood markers of cases and controls. Reference ranges: WBC: White blood cell count (>11,000/mm³); HDL: High-density lipoprotein (<40 mg/dL); BUN: Blood urea nitrogen (>20 mg/dL); Creatinine (>1.3 mg/dL); FBS: Fasting blood sugar (>110 mg/dL); PPBS: Postprandial blood sugar (>200 mg/dL)

Bivariate associations and risk-factor clustering

Conditional logistic regression, matching for sex, further quantified the strengths of association of lifestyle, comorbid, and postadmission blood markers by yielding odds ratios (ORs) for stroke [Table 3]. Hypertension demonstrated the most profound association, exhibiting an OR of 22 and accounting for the highest explanatory variance in the model (Pseudo R² =0.2248, LR χ² =32.38). Regular alcohol consumption emerged another significant lifestyle factor, conferring sixfold increased odds of stroke, while diabetes mellitus demonstrated a threefold increase. Additional factors like smoking history, coronary artery disease, and polycythemia demonstrated elevated point estimates suggestive of increased risk but failed to reach significance along with hypercholesterolemia, anemia, and hypothyroidism.

Table 3.

Conditional logistic regression analysis (stratified by sex) for unadjusted association of lifestyle, comorbid, and acute-phase blood markers with stroke odds

Predictor		n*		Odds Ratio (OR)		95% Confidence Interval		P
Lifestyle and comorbid factors
Smoking		121		3.91		0.67 – 22.86		0.117
Alcohol Use		121		6.06		1.79 – 20.49		0.002
Hypertension		121		21.86		5.94 – 80.39		<0.001
Hypercholesterolemia		121		1.36		0.36 – 5.08		0.653
CAD/IHD		121		3.65		0.32 – 41.74		0.280
Diabetes		121		2.95		1.27 – 6.84		0.012
Anemia		121		0.88		0.34 – 2.29		0.800
Hypothyroidism		121		1.44		0.53 – 3.87		0.477
Congenital Heart Disease		121		1.77		0.11 – 28.91		0.692
Postadmission metabolic and inflammatory markers‡
Hemoglobin		121		1.21		0.75 – 1.96		0.435
WBC count		119		1.65		1.08 – 2.53		0.018
Platelet count		121		1.20		0.82 – 1.78		0.345
Sodium		121		0.86		0.78 – 0.96		0.004
Potassium		121		0.82		0.55 – 1.21		0.314
Albumin		107		1.56		0.70 – 3.44		0.269
Total Protein		106		0.97		0.65 – 1.44		0.870
Triglycerides		88		0.74		0.48 – 1.15		0.180
HDL		85		1.37		0.88 – 2.14		0.154
Urea		118		0.90		0.61 – 1.33		0.590
Creatinine		120		1.96		1.24 – 3.09		0.002
AST		56		0.59		0.32 – 1.10		0.090
ALT		81		0.70		0.44 – 1.11		0.118
Indirect Bilirubin		74		1.33		0.79 – 2.24		0.287
Total Bilirubin		83		1.41		0.86 – 2.31		0.164
TSH		71		0.68		0.41 – 1.11		0.116
FBS		79		1.37		0.86 – 2.18		0.174
PPBS		78		1.51		0.93 – 2.43		0.086

*Number of observations included in each specific conditional logistic regression model. ^‡All blood parameters except for serum sodium and albumin were transformed using Rank-based Inverse Normal Transformation (RINT) prior to analysis because of non-normal distribution of data and to yield ORs interpreted per 1-unit increase in the transformed value (=1 standard deviation increase of original value)

Acute phase blood parameters revealed complex dysregulation patterns, with significantly elevated WBC count, serum creatinine, and PPBS demonstrated in cases compared to controls [Table 3]. While each standard-deviation (2627/mm³) rise in WBC count was associated with a 65% increase in odds, a similar elevation of serum creatinine nearly doubled stroke odds, whereas PPBS demonstrated a weaker trend. Serum sodium levels were notably lower in cases, while alanine aminotransferase (ALT), indirect bilirubin, and FBS exhibited weaker positive associations precluding statistical significance. No significant differences were found with the other analytes. The analysis of rarer exposures like hyperhomocysteinemia, substance abuse, antipsychotic drug use, psychiatric illness, and family history yielded extreme odds ratios with convergence issues, while acute/chronic kidney disease was omitted from the model due to collinearity, precluding reliable interpretation of these specific associations.

Comprehensive pairwise analyses between the lifestyle factors and comorbidities to investigate clustering of key risk factors within specific strata of the study population revealed distinct patterns. A robust positive association between smoking and alcohol consumption was observed in both male cases (Fisher’s exact P = 0.031) and controls (P = 0.012) demonstrating complete concomitance; all smokers (2/2 in controls and 4/4 among cases) consumed alcohol, contrasting sharply with only 7.5% (3/40) and 35% (7/20) of controls and cases, respectively, who were nonsmokers. Hypertension and diabetes demonstrated significant linkage among both males and females, but exclusively within cases; males showed 58% (7/12) diabetes prevalence among hypertensive patients versus 8% (1/12) in nonhypertensive individuals (P = 0.027), while 86% (6/7) of hypertensive female cases had diabetes compared to 22% (2/9) of the normotensive (P = 0.041). Conversely, no significant associations emerged between smoking and hypertension, smoking and diabetes, alcohol and hypertension, and alcohol and diabetes across the two groups within both sexes. Stratified by age, no statistically significant associations emerged overall between any variable pairs among controls, while cases revealed clinically significant interactions within the older age strata. This older cohort aged 35–45 years exhibited robust clustering of smoking and alcohol (P = 0.003), with 4/4 smokers consuming alcohol versus 15% (4/26) of nonsmokers. Hypertension and diabetes were also synergistically linked (P = 0.033), with diabetes prevalence almost threefold higher in hypertensive individuals (65%, 11/17) versus normotensive counterparts (23%, 3/13). In contrast, alcohol consumption inversely correlated with diabetes (P = 0.039) as alcohol users had lower diabetes prevalence (12.5%, 1/8) compared to nonusers (59%, 13/22) in this cohort. To account for multiple hypothesis testing across the stratified analysis, raw P values were adjusted using the Benjamini–Hochberg (BH) procedure to control the False Discovery Rate. As shown in Supplementary Table 1, all pairwise clustering tests remain statistically significant after controlling for false discoveries using the BH-FDR procedure at q < 0.05.

Supplementary Table 1.

Benjamini–Hochberg (BH) False Discovery Rate (FDR) adjusted p-values at q<0.05 for stratified pairwise analysis of risk-factor clustering

Subgroup and Comparison		n		Unadjusted P		BH Rank (k/7)		BH-FDR q-value
Smoking vs. Alcohol
Cases, 35-45 years		30		0.003		1		0.021
Controls, All males		42		0.012		2		0.041
Cases, All males		24		0.031		4		0.041
Hypertension vs. Diabetes
Cases, All males		24		0.027		3		0.041
Cases, 35-45 years		30		0.033		5		0.041
Cases, All females		16		0.041		7		0.041
Alcohol vs. Diabetes
Cases, 35-45 years		30		0.039		6		0.041

Stratified and multivariate analyses

Age- and sex-stratified logistic regression analyses revealed distinct patterns of unadjusted association among subgroups. Hypertension and alcohol use were strong predictors overall, with cases having over 23-fold (OR: 23.5, 95% CI: 6.4-87.1; P < 0.001) higher odds of being hypertensive and 6-fold higher odds of consuming alcohol (OR: 5.8, 95% CI: 1.8-18.0; P = 0.003) compared to controls. This association with hypertension (OR: 13.0, 95% CI: 3.1-53.8; P < 0.001) and alcohol use (OR: 6.3, 95% CI: 1.8-21.5; P = 0.004) persisted in males but could not be estimated among females due to lack of variation. Most notably however, alcohol use was associated with over 17-fold (OR: 17.0, 95% CI: 1.3-216.7; P = 0.029) increase in odds among the younger group aged 18–30. Sex in itself was not independently associated with differing odds (OR: 0.72, P = 0.398), while other lifestyle and comorbid factors precluded significant associations within subgroups as well.

Concurrent stratification of blood parameters also revealed a nuanced pattern of metabolic and inflammatory dysregulation among groups [Supplementary Figure 1 ^{(3.8MB, tif)} ]. Female stroke cases exhibited significantly lower serum sodium levels (137.8 vs 141.2 mmol/L, P = 0.015), higher platelet counts (3.1 vs 2.6 × 10⁵/mm3 with Wilcoxon rank sum z = −1.98, P = 0.047), and elevated serum creatinine (0.8 vs 0.6 mg/dL; z = -3.10, P = 0.002), alongside lower ALT (15 vs 20 U/L; z = 2.57, P = 0.010) compared to controls. In contrast, males showed no such differences and instead demonstrated higher hemoglobin (15.2 vs 14.0 g/dL; z = −2.21, P = 0.027) among cases. On the other hand, age-dependent associations were also evident. The youngest cohort aged 18–24 years displayed significantly raised indirect (0.4 vs 0.1 mg/dL; z = -2.47; P = 0.014) and total bilirubin (0.6 vs 0.3 mg/dL; z = -2.26, P = 0.024), marginal creatinine elevation (0.9 vs 0.7 mg/dL; z = -1.94 P = 0.052), and a weak trend toward increased HDL cholesterol (40 vs 35 mg/dL; z = -1.87, P = 0.061) among cases, effects absent overall. While cases in the oldest subgroup (35–45 years) displayed marked leukocytosis (9650 vs 8400/mm³; z = −2.82, P = 0.005) and suppressed TSH (2.2 vs 3.6 mIU/L; z = 2.60, P = 0.009), persistently low sodium levels were observed in cases across the 25–34-year cohort (137 vs 140 mmol/L; z = 1.99, P = 0.046).

The final multivariate logistic regression model predicting stroke odds demonstrated robust statistical performance with excellent overall fit and appropriate calibration [Table 4]. Age emerged as a significant independent predictor, with individuals aged 35–45 years exhibiting substantially elevated stroke odds compared to those 18–24 years, while the intermediate age group (25–34) demonstrated no significant difference from the reference category. Although females showed increased odds relative to males, this association failed to achieve statistical significance in the adjusted model. Among modifiable risk factors, hypertension retained its position as the predominant stroke predictor, conferring a remarkable 22-fold increase in stroke odds, while alcohol consumption maintained its significant association with elevated odds. Notably, diabetes mellitus and smoking lost their significance following multivariate adjustment.

Table 4.

Multivariate logistic regression analysis stratified by age and sex to yield final adjusted associations of risk factors and stroke incidence (n=121)

Predictor		aOR		95% CI		P
Age
18-24 years		Reference
25-34 years		1.03		0.22 – 4.75		0.968
35-45 years		7.93		1.91 – 32.90		0.004
Sex
Male		Reference
Female		1.85		0.60 – 5.69		0.281
Alcohol
No		Reference
Yes		6.68		1.02 – 43.63		0.047
Smoking
No		Reference
Yes		0.45		0.03 – 6.95		0.568
Hypertension
No		Reference
Yes		22.07		4.66 – 104.64		<0.001
Diabetes
No		Reference
Yes		0.52		0.14 – 1.96		0.335

aOR: Adjusted Odds Ratio; CI: Confidence Interval. Model Fit Statistics: N=121; LR χ²(7)=54.49 (P<0.0001); Pseudo R²=0.3548; Area Under ROC Curve (AUC)=0.8508; Hosmer–Lemeshow test P=0.9193

DISCUSSION

The major strengths of this study are its focus on a relatively under-researched demographic and a comprehensive assessment of a wide array of potential risk factors and readily available post-admission laboratory parameters. The findings underscore an overwhelming dominance of established stroke risk factors even in this vulnerable demographic.

Hypertension and alcohol use among the young

Hypertension emerged as the most formidable independent predictor with multivariate analysis revealing a striking 22-fold increase in stroke odds, in alignment with global evidence such as the landmark INTERSTROKE study.[17] The sheer magnitude of this risk among the young, far exceeding other factors, suggests either a heightened vascular susceptibility in this demographic[18] or a combination of secondary (nonessential) causes of hypertension combined with neglect and prolonged underdiagnosis.[19,20,21] This signals an alarming need for enhanced, age-appropriate screening protocols and aggressive management strategies for hypertension especially since hypertension is detectable and manageable through cost-effective interventions.[22,23,24]

Regular alcohol consumption was also a significant independent contributor, consistent with literature[7,25] implicating it in stroke pathogenesis through arrhythmias, prothrombotic, or cardioembolic states, and also by inducing hypertension. Notably, the INTERSTROKE study also reported regional variations in association, highlighting an increased odds of stroke even with low alcohol intake in India.[26] A particularly novel and concerning revelation from our stratified analyses was the exceptionally strong association with alcohol use among the youngest subgroup (18–30), where it was linked to an over 17-fold increase in odds. This suggests possibly riskier consumption patterns, such as heavy episodic drinking, within this very young demographic, demanding urgent, age-specific awareness programs and interventions that address these unique susceptibilities. This strong association in the very young further elucidates their heightened vulnerability due to certain stroke mechanisms, such as cardio embolism secondary to alcohol-induced arrhythmias, or hemorrhagic events due to acute blood pressure fluctuations.[26,27,28]

Beyond these primary predictors, the study unveiled nuances in the risk profiles of other factors like diabetes and smoking. Diabetes, which demonstrated threefold increased odds in bivariate analyses, and smoking, which showed a weaker initial association, lost their statistical significance in the final multivariate model. This attenuation could suggest that their effects might be partly mediated through, or overshadowed by, the profound impact of hypertension and alcohol in this cohort. This observation does not diminish their established roles as cardiovascular risk factors but rather indicates that in younger individuals, stroke pathogenesis could be driven more by acute hemodynamic stresses (from hypertension) or direct toxic and arrhythmic effects (from alcohol) rather than the more chronic, gradual atherosclerotic processes typically associated with long-standing diabetes and smoking.

Furthermore, we identified significant clustering between hypertension and diabetes exclusively within cases (both males and females), and it was particularly synergistic in older cases (35–45 years). This pattern suggests that while they may exist independently in the general population, their coexistence could dramatically escalate the propensity for stroke, potentially acting as a ‘tipping point’ in susceptible young adults. This underscores the importance of a holistic approach, where clinicians encountering young patients with either hypertension or diabetes should maintain a high index of suspicion for the other and manage both conditions aggressively. Moreover, complete concomitance was observed between smoking and alcohol use among males (cases and controls), indicating they often coexist and should be addressed concurrently.

Acute-phase response

Post-admission blood markers provided a window into the acute pathophysiological disturbances accompanying stroke in this cohort. Significantly elevated WBC counts, serum creatinine, and PPBS in cases compared to controls are indicative of processes well established in stroke pathogenesis. Leukocytosis specifically reflects the acute inflammatory cascade triggered by cerebral ischemia and was most robust among the oldest (35–45) subgroup, who incidentally also had suppressed TSH levels, linked to increased risk of cardiovascular events including stroke even in euthyroid states.[29] While elevated creatinine may point toward underlying renal impairment or acute injury secondary to stroke itself, raised PPBS suggests glucose dysmetabolism potentially resulting from occult diabetes or stress hyperglycemia, both of which are deleterious in the acute stroke setting. Hyponatremia, a common electrolyte disturbance in neurological patients attributable to cerebral salt wasting or SIADH,[30] was also notably found in stroke cases, particularly among females and those aged 25–34.

Stratified analyses unveiled considerable heterogeneity. Female stroke cases exhibited particularly higher platelet counts, signifying a prothrombotic state or a more pronounced inflammatory response, and lower ALT levels. While ALT is primarily recognized as a liver enzyme, emerging evidence suggests that low levels might serve as a surrogate marker for frailty or poor nutritional status and conditions such as early-onset sarcopenia, which could in turn influence long-term prognosis and morbidity.[31,32] Such sex-specific findings intimate potential sex-dimorphic variations, indicating the need for targeted interventions such as early nutritional and rehabilitation support for young female patients. In contrast, male cases demonstrated higher hemoglobin, whose role in stroke is multifaceted, with both very low (anemia) and very high levels (polycythemia) associated with adverse outcomes, through effects on cerebral oxygen delivery and blood viscosity. Remarkably, the youngest stroke cohort (18–24 years) displayed significantly raised indirect and total bilirubin levels, further stressing hepatic derangement associated with heavy alcohol consumption, as predicted above in this cohort.

CONCLUSIONS

This investigation into early-onset stroke among young Indians contributes vital regional data to inform clinical practice and public health strategies in countries like India where rapid urbanization and lifestyle transitions are amplifying cardiovascular risk profiles among the young.[5,6] In summary, this study underscores the urgent necessity for early detection and prompt management of hypertension along with alcohol cessation in this vulnerable demographic.[25] Stratified analyses unveiled novel insights on age- and sex-specific heterogeneity and revealed acute pathophysiological disturbances within subgroups of this cohort, necessitating holistic management approaches extending beyond neurological care.

However, the methodological limitations of this study warrant careful interpretation of results. The case-control design is inherently susceptible to selection bias and also precludes establishment of causality. Moreover, cases might recall or report past exposures differently than controls, potentially inflating associations and underestimating socially sensitive behaviors. Potential differential ascertainment bias must also be acknowledged; stroke cases underwent more intensive clinical monitoring and emergency diagnostic workups than general ward controls, which may have led to a higher rate of detection for previously unrecognized hypertension or diabetes among cases. Additionally, the hospital-based sample drawn from a tertiary referral center may not reflect the broader community, limiting generalizability. Furthermore, while recruiting hospital-based, acutely ill inpatient controls helped mitigate Berkson’s selection bias, the presence of non-neurological acute illnesses in controls may have elevated baseline inflammatory and metabolic markers, potentially biasing acute-phase biomarker estimates toward the null. Certain significant factors like obesity (waist–hip ratio or BMI), detailed socioeconomic, diet, physical activity, and alcohol use histories, and serial or baseline blood markers were unmeasured. The modest sample size (121) also limits precision of OR estimates (wide confidence intervals) and precluded in-depth subgroup modeling especially among females and the rarer exposures (hyperhomocysteinemia, kidney disease, psychiatric illness, etc.).

These limitations underscore the necessity of future research with larger, multicentric, prospective cohort studies across diverse Indian settings to more definitively establish temporality, mitigate biases, and enhance external validity. The integration of such studies with the ICMR-NCDIR national stroke registry,[33] designed to standardize stroke surveillance across India, offers a powerful platform for robust data collection and epidemiological insight. These collective efforts should inform and reinforce existing public health initiatives in India like the NPCDCS and “Mission Brain Attack” by intensifying their focus on younger age groups and thereby crafting more effective prevention strategies for early-onset stroke.[33,34]

Ethical approval

This study involving adult human participants received ethical approval from the Institutional Ethics Committee at Government Stanley Medical College and Hospital, Chennai (EC/NEW/INST/2022/TN/0102) and was conducted in accordance with the Declaration of Helsinki (2013). Written informed consent was obtained from all participants and proxies (for patients deemed lacking capacity) prior to data collection, with the approved consent form read in the respondent’s own language.

Data confidentiality and Data availability statement

Participants’ health information relevant to the study was obtained and deidentified as per the approved protocol. Deidentified data is available upon request from the corresponding author.

Conflicts of interest

There are no conflicts of interest.

Supplementary Figure 1

Significant acute-phase laboratory derangements across age- and sex-stratified subgroups.

WBC: White blood count; HDL: High-density lipoprotein; ALT: Alanine Transaminase; TSH: Thyroid Stimulating Hormone

Funding Statement

Nil.