Outcomes of methylprednisolone as an adjunct to endovascular therapy in patients with diabetes and admission hyperglycemia

Introduction

As endovascular therapy (EVT) has become the standard treatment for acute ischemic stroke (AIS) with large vessel occlusion (LVO), approximately 50% of patients still experience poor functional outcomes, despite achieving recanalization rates exceeding 80% (1, 2). Therefore, how to further improve 90-day functional independence in patients following successful recanalization has emerged as a critical issue that urgently needs to be addressed (3). In recent years, neuroprotective and anti-inflammatory therapeutic strategies targeting reperfusion injury and post-ischemic inflammatory responses have become a frontier hotspot in stroke research, providing new directions for optimizing EVT efficacy (4, 5).

Methylprednisolone, a potent glucocorticoid with anti-inflammatory properties, has emerged as a promising adjunctive therapy for mitigating reperfusion injury (6, 7). The MARVEL (Methylprednisolone as Adjunct to Endovascular Thrombectomy for Large-Vessel Occlusion Stroke) randomized controlled trial demonstrated that intravenous methylprednisolone following EVT, compared with placebo, did not improve 90-day functional outcomes but significantly reduced mortality and symptomatic intracranial hemorrhage (sICH) (8). Nevertheless, methylprednisolone is known to exert considerable effects on glucose metabolism (9, 10). It can significantly elevate blood glucose levels by increasing hepatic gluconeogenesis, inhibiting peripheral glucose uptake, and inducing insulin resistance (9, 11). Moreover, admission hyperglycemia has been established as an independent risk factor for infarct expansion, hemorrhagic transformation, and poor functional outcomes in patients with AIS (12, 13). Building on this, diabetes mellitus, as an underlying condition characterized by chronic hyperglycemia and metabolic dysregulation, may not only exacerbate the aforementioned glycemic fluctuations induced by methylprednisolone, but also alter patients’ responses to glucocorticoid therapy through its distinct neurovascular pathological mechanisms (14, 15). However, the effect of methylprednisolone as adjunctive therapy to EVT in patients with with different glycemic states (acute or chronic hyperglycemia phase) remain unclear.

In this context, using the data from the MARVEL trial (8), this study aimed to evaluate the effect of methylprednisolone as adjunctive therapy to EVT in patients with different glycemic states.

Methods

Study design and participants

This study was a secondary analysis of the MARVEL trial, an investigator-initiated, randomized, double-blind, placebo-controlled clinical trial conducted at 82 stroke centers across China from February 9, 2022, to June 30, 2023, with follow-up for 3 months (8). The MARVEL trial aimed to assess the efficacy and adverse events of adjunctive intravenous low-dose methylprednisolone to endovascular thrombectomy for AIS secondary to LVO. Comprehensive information regarding the rationale, design, methods, and major findings of the MARVEL trial has been previously documented (8). This study followed the ethical principles of the Helsinki Declaration and was approved by the ethics committees of all participating centers. Written informed consent was provided by all patients or their legally authorized representative before randomization. This study was reported in accordance with the Consolidated Standards of Reporting Trials (CONSORT) guidelines for randomized trials.

Patients who met the following inclusion criteria were recruited to this study (1): age ≥ 18 years (2); The time from onset to randomization was within 24 hours (3); Anterior circulation ischemic stroke was preliminarily determined according to clinical symptoms or imaging examination (4); Baseline National Institutes of Health Stroke Scale (NIHSS) score ≥ 6 (5); Baseline Alberta Stroke Program Early CT Score (ASPECTS) ≥3 (6); Computed tomography angiography (CTA)/magnetic resonance angiography (MRA)/digital subtraction angiography (DSA) confirmed occlusion of the intracranial segment of the internal carotid artery or middle cerebral artery, and decision to undergo endovascular therapy. The exclusion criteria for this study included (1): pre-stroke mRS score ≥2 (2); Allergic to glucocorticoids (3); patients with severe or terminal illnesses; Detailed inclusion and exclusion criteria are provided in previously documented 8.

Intervention and definition

All patients underwent EVT. Eligible patients were randomized to receive intravenous methylprednisolone at a dose of 2mg/kg/d (up to a maximum dose of 160mg) for 3 days vs placebo. We defined acute admission hyperglycemia as blood glucose levels ≥7.8 mmol/L. Chronic hyperglycemia is defined as a history of diabetes mellitus in the patient.

Data collection

We collected the following clinical variables from the database: demographic characteristics, medical history, prestroke modified Rankin Scale (mRS) score, baseline glucose level, baseline stroke severity, extent of ischemic injury, stroke etiology, the site of occlusion, the use of intravenous thrombolysis, time metrics (last known to be well to puncture, last known to be well to randomization, last known to be well to recanalization, time from randomization to initial treatment) and blood pressure at hospital arrival. Baseline stroke severity was measured using the NIHSS score. The extent of ischemic injury was assessed with ASPECTS. Stroke etiology was categorized based on the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) criteria.

Outcomes

All participants underwent follow-up assessment at the 90-day timepoint, measured using the modified Rankin Scale (mRS) scores. Data collection was conducted either through in-person clinic visits or structured telephone interviews. To minimize assessment bias, mRS scoring was predominantly performed via centralized review of video or audio recordings by blinded outcome adjudicators. In circumstances where such recordings were not obtainable, face-to-face evaluation by blinded local investigators served as an alternative approach. The primary outcome was the distribution of the mRS score (ranging from 0 [no symptoms] to 6 [death]) at 90 days after randomization. Secondary outcomes included NIHSS score at 5 to 7 days or at early discharge; the following prespecified dichotomizations of the mRS score at 90 days: 0 to 3 vs 4 to 6, 0 to 2 vs 3 to 6 and 0 to 1 vs 2 to 6. Safety outcomes included all-cause mortality at 90 days, sICH within 48 hours, and any intracranial hemorrhage (ICH) within 48 hours. sICH was classified according to the modified Heidelberg Bleeding Classification.

Statistical analysis

The Kolmogorov-Smirnov test was applied to assess the normality of continuous variables. As all continuous variables failed to meet the assumption of normal distribution, they were described using medians and interquartile ranges (IQR). Differences between groups in continuous variables were compared using the Mann-Whitney U test. Categorical variables were presented as frequencies and percentages, and differences between groups were analyzed using the chi-square test or Fisher’s exact test, as appropriate.

Logistic regression analysis were performed to evaluate clinical outcomes. For the distribution of mRS score, ordinal logistic regression was applied to estimate the common odds ratio (cOR) for a shift in mRS distribution toward better outcome. Binary logistic regression was performed for dichotomous outcomes, including mRS 0-3, mRS 0-2, mRS 0-1, mortality, sICH and any ICH. The NIHSS scores was analyzed by the linear regression with robust standard errors. In order to reduce the imbalance in confounders between treatment groups, the outcomes were adjusted using the inverse probability treatment weighting (IPTW) method (Supplementary method 1). Sample imputation was applied for the covariate to perform IPTW. (Supplementary method 2) (16) Missing data in the regression analysis were imputed using multiple imputation by chained equations. Ten imputed datasets were generated, and analysis were based on the combined results in accordance with Rubin’s rules (17). The variables included in the multiple imputation model were: admission hyperglycemia, group, age, baseline NIHSS score, pre-stroke mRS score, baseline ASPECTS, intravenous thrombolysis, onset-to-randomization time, occlusion site, 90-day mRS score, NIHSS score at 5–7 days, death, sICH and any ICH (18). The IPTW was calculated for patients adjusted for 7 covariates: age, baseline NIHSS score, pre-stroke mRS score, baseline ASPECTS, use of intravenous thrombolysis, time from stroke onset to randomization, and occlusion location, which are consistent with the covariates adjusted for in the multivariable logistic regression analysis. The estimates of odds ratios (OR), cOR, β coefficients, and 95% CIs were reported. The treatment effect of the history of diabetes on the primary outcome among patients with EVT plus methylprednisolone was assessed across different subgroups.

A two-tailed P value of less than 0.05 was considered statistically significant. All statistical analysis were performed using R software, version 4.4.2 (R Foundation for Statistical Computing).

Results

Baseline characteristics

A total of 1,680 patients were initially enrolled from the MARVEL trial. (Supplementary Figure 1) After excluding one patient with basilar artery occlusion and one patient with anterior cerebral artery occlusion, 1,678 patients were finally included in this study. Stratified by the history of diabetes, among the 1,678 total patients (321 with diabetes, 1,357 without), no significant baseline differences were observed between the methylprednisolone and placebo groups in the diabetes subgroup. However, in the non-diabetes group, the methylprednisolone group was slightly younger, had lower glucose levels at hospital arrival, and a lower prevalence of coronary heart disease compared with the placebo group. Further baseline characteristics were summarized in Table 1. Stratified by the admission hyperglycemia, among the 1,470 total patients (540 with admission hyperglycemia, 930 without admission hyperglycemia), no significant baseline differences were observed between the methylprednisolone and placebo groups. Further baseline characteristics were summarized in Table 2. Between-group comparisons after IPTW are detailed in Supplementary Table 1 and Supplementary Table 2.

Table 1. Baseline characteristics of methylprednisolone versus placebo stratified by diabetes history before inverse probability of treatment weighting.

Comorbidities based on family or patient report.

Scores on the mRS of functional disability range from 0 (no symptoms) to 6 (death).

Scores on the NIHSS range from 0 to 42, with higher scores indicating more severe neurologic deficit.

The ASPECTS is an imaging measure of the extent of ischemic stroke. Scores range from 0 to 10, with higher scores indicating a smaller infarct core. Listed are values for the core laboratory assessment. Data were missing for 6 in patients without diabetes.

Stroke etiology was categorized according to the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification.

One patient had a basilar artery occlusion and 1 patient had an anterior cerebral artery occlusion in the methylprednisolone group.

ASPECTS, Alberta Stroke Program Early CT Score; ICA, Intracranial internal carotid artery; IQR, Interquartile range;

M1, M1 middle cerebral artery segment; M2, M2 middle cerebral artery segment; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; SMD, Standardized mean differences; TOAST, Trial of ORG 10172 in Acute Stroke Treatment.

Table 2. Baseline characteristics of methylprednisolone versus placebo stratified by admission hyperglycemia before inverse probability of treatment weighting.

Comorbidities based on family or patient report.

Scores on the mRS of functional disability range from 0 (no symptoms) to 6 (death).

Scores on the NIHSS range from 0 to 42, with higher scores indicating more severe neurologic deficit.

The ASPECTS is an imaging measure of the extent of ischemic stroke. Scores range from 0 to 10, with higher scores indicating a smaller infarct core. Listed are values for the core laboratory assessment. Data were missing for 2 patients with admission hyperglycemia and 4 patients with no admission hyperglycemia.

Stroke etiology was categorized according to the Trial of ORG 10172 in Acute Stroke Treatment (TOAST) classification.

One patient had a basilar artery occlusion and 1 patient had an anterior cerebral artery occlusion in the methylprednisolone group.

ASPECTS, Alberta Stroke Program Early CT Score; ICA, Intracranial internal carotid artery; IQR, Interquartile range;

M1, M1 middle cerebral artery segment; M2, M2 middle cerebral artery segment; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; SMD, Standardized mean differences; TOAST, Trial of ORG 10172 in Acute Stroke Treatment.

Clinical outcomes

Stratified by diabetes history(methylprednisolone versus placebo)

Among patients with diabetes, methylprednisolone was not associated with significant improvement in the 90-day mRS score compared with placebo (IPTW-adjusted OR, 1.07 [95% CI, 0.81–1.40]; P = 0.640), nor in any secondary functional outcome (all P > 0.05). Safety outcomes did not differ significantly for 90-day mortality (IPTW-adjusted OR, 0.89 [95% CI, 0.62–1.27]; P = 0.526) or sICH within 48 hours (IPTW-adjusted OR, 0.96 [95% CI, 0.56–1.66]; P = 0.880), although any ICH within 48 hours was significantly more frequent with methylprednisolone (IPTW-adjusted OR, 1.63 [95% CI, 1.17–2.26]; P = 0.004) (Table 3).

Table 3. Outcomes according to the treatment in patients with and without history of diabetes.

Values adjusted for age, baseline NIHSS score, prestroke mRS score, baseline ASPECTS, use of intravenous thrombolysis, time from onset to randomization, and occlusion location.

Common odds ratio; The common odds ratio was estimated from an ordinal logistic regression model. The mRS of functional disability ranges from 0 (no symptoms) to 6 (death). Data were not available for 1 patient in the diabetes group and 6 patients in the no diabetes group.

The β values were estimated from a multivariable linear regression model. The win proportion was calculated as the number of wins for the methylprednisolone group over the control group in NIHSS scores among all possible pairs of NIHSS scores (taking 1 patient from the methylprednisolone group and 1 patient from the control group), divided by the total number of pairs. The results were adjusted using the inverse probability of treatment weighting method. NIHSS scores range from 0 to 42; higher scores indicate more severe neurological deficits.

sICH was defined according to the Heidelberg bleeding classification (an increase in the NIHSS score of 4 points or an increase in the score for an NIHSS subcategory of 2 points with any intracranial hemorrhage on imaging results).

CI, confidence interval; ICH, intracranial hemorrhage; IPTW, inverse probability of treatment weighting; IQR, interquartile range; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; sICH, symptomatic intracranial hemorrhage.

Among patients without diabetes, methylprednisolone was associated with significantly better 90-day mRS scores (IPTW-adjusted OR, 1.16 [95% CI, 1.02–1.33]; P = 0.026) and higher rates of mRS 0–4 (IPTW-adjusted OR, 1.31 [95% CI, 1.11–1.54]; P = 0.001) and mRS 0–3 (IPTW-adjusted OR, 1.20 [95% CI, 1.03–1.40]; P = 0.019), as well as significantly lower 90-day mortality (IPTW-adjusted OR, 0.74 [95% CI, 0.62–0.88]; P = 0.001) and sICH within 48 hours (IPTW-adjusted OR, 0.66 [95% CI, 0.51–0.85]; P = 0.001). A significant treatment-by-diabetes-history interaction was observed only for any ICH within 48 hours (P for interaction = 0.036). No significant interactions were found for other outcomes (Table 3). The distribution of the modified Rankin Scale scores at 90 days by the history of diabetes, stratified by treatment arm was shown in Figure 1A.

Stratified by admission hyperglycemia(Methylprednisolone Versus Placebo)

Among patients with admission hyperglycemia, methylprednisolone was not associated with significant improvement in the 90-day mRS score compared with placebo (IPTW-adjusted OR, 1.17 [95% CI, 0.95–1.45]; P = 0.139), nor in any secondary functional outcome (all P > 0.05). Safety outcomes also did not differ significantly for 90-day mortality, sICH within 48 hours or any ICH within 48 hours (all P > 0.05). Among patients without admission hyperglycemia, methylprednisolone was not significantly associated with the 90-day mRS score (IPTW-adjusted OR, 1.13 [95% CI, 0.96–1.33]; P = 0.138), but was associated with significantly higher rates of mRS 0–4 (IPTW-adjusted OR, 1.24 [95% CI, 1.01–1.52]; P = 0.040) and mRS 0–3 (IPTW-adjusted OR, 1.21 [95% CI, 1.01–1.46]; P = 0.042). No significant differences were observed for safety outcomes, including 90-day mortality, sICH within 48 hours, and any ICH within 48 hours in patients with or without admission hyperglycemia. No significant treatment-by-admission-hyperglycemia interactions were observed for any outcome (all P for interaction > 0.05) (Table 4). The distribution of the modified Rankin Scale scores at 90 days by admission hyperglycemia, stratified by treatment arm was shown in Figure 1B.

Table 4. Outcomes according to the treatment in patients with and without admission hyperglycemia.

Values adjusted for age, baseline NIHSS score, prestroke mRS score, baseline ASPECTS, use of intravenous thrombolysis, time from onset to randomization, and occlusion location.

Common odds ratio; The common odds ratio was estimated from an ordinal logistic regression model. The mRS of functional disability ranges from 0 (no symptoms) to 6 (death). Data were not available for 1 patient in the admission hyperglycemia group and 5 patients in the no admission hyperglycemia group.

The β values were estimated from a multivariable linear regression model. The win proportion was calculated as the number of wins for the methylprednisolone group over the control group in NIHSS scores among all possible pairs of NIHSS scores (taking 1 patient from the methylprednisolone group and 1 patient from the control group), divided by the total number of pairs. The results were adjusted using the inverse probability of treatment weighting method. NIHSS scores range from 0 to 42; higher scores indicate more severe neurological deficits.

sICH was defined according to the Heidelberg bleeding classification (an increase in the NIHSS score of 4 points or an increase in the score for an NIHSS subcategory of 2 points with any intracranial hemorrhage on imaging results).

CI, confidence interval; ICH, intracranial hemorrhage; IPTW, inverse probability of treatment weighting; IQR, interquartile range; mRS, modified Rankin Scale; NIHSS, National Institutes of Health Stroke Scale; OR, odds ratio; sICH, symptomatic intracranial hemorrhage.

Sensitivity analysis

Among patients with different blood glucose status, subgroup analyses were performed to evaluate whether the association between treatment and 90-day mRS score differed across various subgroups. No significant interactions were identified across subgroups stratified by age, sex, baseline ASPECTS, baseline NIHSS score, occlusion site, stroke etiology, the use of intravenous thrombolysis and time from onset to randomization (Figure 2; Supplementary Figure 6-8). For patients with missing admission glucose, we performed a multiple-imputation sensitivity analysis, and the results were broadly consistent with those of the primary analysis (Supplementary Table 3).

Margin effect

Considering that the same baseline glucose level may be observed in both patients with and without a history of diabetes, we analyzed the association between baseline glucose and clinical outcomes (an mRS score of 0–2 and mortality) among patients stratified by the history of diabetes (Supplementary Figure 9). The results demonstrated the predicted probability of an mRS score of 0–2 progressively decreased with increasing baseline glucose, while the predicted probability of mortality progressively increased in patients with or without diabetes mellitus.

Discussion

In this secondary exploratory analysis of the MARVEL trial, we evaluated the effect of methylprednisolone as an adjunct to EVT in patients with and without diabetes mellitus, as well as between patients with and without admission hyperglycemia. The interaction between diabetes and treatment did not reach statistical significance. In patients without diabetes, a benefit trend was observed with adjunctive methylprednisolone. Moreover, in patients with diabetes, adjunctive use of methylprednisolone may be associated with a higher risk of any intracranial hemorrhage. In addition, the presence or absence of admission hyperglycemia did not significantly affect the effectiveness or safety of methylprednisolone.

Diabetes mellitus denotes a hyperglycemic state caused by various metabolic disorders, which feature either impaired insulin secretion, defective insulin action on target tissues, or frequently both (19). It has been previously established that a genetic predisposition to type 2 diabetes confers a higher risk of ischemic stroke, with particularly strong associations observed for large artery and small vessel stroke (20). Independent of hyperglycemia, a genetic tendency toward insulin resistance (but not β-cell dysfunction) correlated with an increased risk of large artery stroke, whereas a genetic predisposition involving both insulin resistance and β-cell dysfunction was linked to a higher risk of small vessel stroke (21). Acute hyperglycemia at admission was defined as a random plasma glucose level >140 mg/dL using the latest American Diabetes Association criteria (22). Among patients hospitalized for ischemic stroke, up to 40% present with elevated blood glucose levels, usually without a prior diagnosis of diabetes (23). Moreover, hyperglycemia is associated with a 3.3% increase in stroke-related mortality among non-diabetic individuals and correlates with more severe neurological impairments (24).

Our study revealed two principal findings regarding the primary outcome. First, among patients without a history of diabetes, adjunctive methylprednisolone to EVT may be associated with a trend toward favorable outcomes. In contrast, among patients with a history of diabetes, methylprednisolone did not demonstrate a similar trend toward benefit. This observation suggests that any potential benefit of methylprednisolone, may be attenuated or even negated by the presence of diabetes, thereby explaining the neutral result of the MARVEL trial, which showed no improvement in 90-day functional outcomes in the intention-to-treat population (8). Several mechanisms may account for this finding. Hyperglycemia in diabetic patients is associated with sustained low-grade inflammation, increased oxidative stress, and endothelial dysfunction (25, 26), which may render the anti-inflammatory effect of glucocorticoids less effective due to a “ceiling effect” of an already activated inflammatory cascade. Moreover, acute hyperglycemia at admission has been shown to exacerbate blood brain barrier disruption, promote mitochondrial dysfunction, and worsen reperfusion injury after EVT, pathways that are not readily reversed by methylprednisolone alone (27, 28).

Second, in both patients with and without admission hyperglycemia, no significant differences in outcomes were observed between the methylprednisolone and placebo groups. This result may be attributable to several potential explanations. Acute hyperglycemia exacerbates blood-brain barrier disruption, promotes mitochondrial dysfunction, and aggravates reperfusion injury, and these pathways cannot be readily reversed by methylprednisolone alone. Patients with admission hyperglycemia often already have a larger ischemic core volume and a higher core-to-penumbra ratio, suggesting that hyperglycemia itself may accelerate the conversion of the ischemic penumbra to the infarct core, thereby narrowing the window of “salvageable tissue” available for neuroprotective intervention (29–33). Admission hyperglycemia is independently associated with worse 90-day functional outcomes, higher mortality, and a higher risk of symptomatic intracranial hemorrhage. In such a population with an already poor baseline prognosis, the relative benefit potentially conferred by methylprednisolone may not be large enough in absolute terms to manifest statistically. The methylprednisolone regimen employed in this study may have been insufficient to meet the anti-inflammatory demands of patients with acute ischemic stroke, or may have failed to adequately cover the prolonged neuroinflammatory cascade that persists beyond the period of reperfusion. These mechanisms may collectively attenuate the neuroprotective anti-inflammatory effects of methylprednisolone, thereby rendering it less efficacious in this metabolically vulnerable population.

Moreover, in patients with diabetes, adjunctive methylprednisolone may be associated with a higher risk of any ICH, a phenomenon that can be jointly explained by the intrinsic cerebrovascular fragility of diabetes and the additional effects of glucocorticoids. In the diabetic state, cerebral microvessels already exhibit adverse structural remodeling, endothelial dysfunction, and impaired blood-brain barrier integrity, leaving the vascular bed inherently vulnerable (34). Meanwhile, diabetes can lead to persistently abnormal upregulation of vascular repair signals such as VEGF-R2, increasing blood-brain barrier permeability, and the direction of the response of these abnormalities to intervention may differ from that in the non-diabetic state (35). On this basis, methylprednisolone, through its well-established hyperglycemic effect, may further exacerbate glucose fluctuations (36), and both glycemic variability and admission hyperglycemia have been confirmed as independent predictors of symptomatic intracranial hemorrhage in patients after successful recanalization with EVT (37, 38). Therefore, in patients with diabetes, methylprednisolone may not only fail to effectively correct the diabetes-specific vascular signaling abnormalities, but may instead promote an increased risk of any ICH by aggravating metabolic disturbance and vascular fragility; whereas in patients without diabetes, this risk signal was not observed owing to the absence of this vulnerable background.

We note that the IPTW analysis yielded a statistically significant result (OR 1.16, 95% CI 1.02–1.33; P = 0.026) while the multivariable regression analysis did not (acOR 1.20, 95% CI 0.99–1.46; P = 0.061). We believe that this discrepancy may be attributable to several interrelated factors. First, the two methods target different estimates. IPTW estimates the marginal average treatment effect, whereas multivariable regression estimates the conditional treatment effect (39). For the odds ratio, a non-collapsible effect measure, these two quantities need not be numerically equal, particularly in the presence of effect modification. Second, within a subgroup of limited sample size, IPTW and regression adjustment, although asymptotically equivalent, may diverge in their finite-sample performance owing to differences in efficiency; in the present study, the narrower confidence interval yielded by the IPTW analysis suggests that weighting may have improved covariate balance within this subgroup.

Limitation

The present study has some limitations. First, this was a secondary analysis of the MARVEL trial. The findings should be interpreted as exploratory and hypothesis generating. They do not supersede the original trial conclusions. Second, while the MARVEL trial randomization process was not stratified according to the history of diabetes mellitus or admission hyperglycemia, we used the IPTW method to control for confounders. Third, since the MARVEL trial enrolled only patients from China, the study results should be interpreted with caution. Fourth, in this study, the definition of diabetes status relied on medical history records, and the definition of hyperglycemia relied on a single admission glucose value. Dynamic glycemic parameters (such as glycemic variability or duration of hyperglycemia) were not incorporated, which may affect a comprehensive assessment of the relationship between glycemic-related factors and prognosis.

Conclusion

In this exploratory secondary analysis of the MARVEL trial, the interaction between glycemia status and treatment did not reach statistical significance. A benefit trend was observed with adjunctive methylprednisolone in patients without diabetes. Moreover, adjunctive use of methylprednisolone in patients with diabetes may be with a higher risk of any intracranial hemorrhage. Admission hyperglycemia status did not significantly modify the effectiveness or safety of methylprednisolone. These findings warrant validation in further studies.

Generative AI statement

The author(s) declared that generative AI was not used in the creation of this manuscript.

Any alternative text (alt text) provided alongside figures in this article has been generated by Frontiers with the support of artificial intelligence and reasonable efforts have been made to ensure accuracy, including review by the authors wherever possible. If you identify any issues, please contact us.

Publisher’s note

All claims expressed in this article are solely those of the authors and do not necessarily represent those of their affiliated organizations, or those of the publisher, the editors and the reviewers. Any product that may be evaluated in this article, or claim that may be made by its manufacturer, is not guaranteed or endorsed by the publisher.

Supplementary material

The Supplementary Material for this article can be found online at: https://www.frontiersin.org/articles/10.3389/fendo.2026.1932813/full#supplementary-material

Supplementary material Supplementary material at the source

Footnotes

Edited by: Marta Iacobucci, Umberto 1 Hospital, Italy

Reviewed by: Martina Kolackova, Charles University, Czechia

Nat Na-Ek, University of Phayao, Thailand