Neurology · Medical Literature
Effect of Admission Diabetes History on Safety and Efficacy of Early Tirofiban Infusion after Intravenous Thrombolysis in Ischaemic Stroke: a post-hoc analysis of the ASSET-IT trial.
Abstract
This study aimed to investigate how diabetes mellitus (DM) influences the efficacy of early tirofiban administration after intravenous thrombolysis in patients with acute ischaemic noncardioembolic stroke. This was a post-hoc analysis of the ASSET-IT (Advancing Stroke Safety and Efficacy through Early Tirofiban Administration after Intravenous Thrombolysis) trial. Patients were categorised into DM and non-DM groups based on baseline diabetes history. The primary efficacy outcome was an excellent functional outcome at 90 days, defined as an mRS score of 0-1. Safety outcomes included sICH within 36 h, any ICH and all-cause mortality within 90 days. The effect of diabetes on treatment outcomes was assessed using multivariable regression models adjusting for relevant confounders. A total of 832 patients at 38 centres were randomised in the ASSET-IT trial (414 to tirofiban, 418 to placebo). The median age was 69 years (IQR, 59-76); 301 (36.2%) were women. Overall, 191 patients (23.0%) had diabetes mellitus. Among patients without diabetes, an excellent functional outcome at 90 days occurred in 69.0% of patients in the tirofiban group and 55.8% in the placebo group, whereas among patients with diabetes, the corresponding rates were 55.0% and 52.0%, respectively. The nominal P value for the treatment-by-diabetes interaction was .009. The 90-day mortality was higher in the DM group receiving tirofiban (6.6% vs 2.0%; RR, 3.79; 95% CI, 0.84-17.05; P = .082) but not in the non-DM group (3.4% vs 4.4%; RR, 0.82; 95% CI, 0.40-1.70; P = .597). Exploratory post-hoc findings from the ASSET-IT trial suggest that the association between early tirofiban administration after intravenous thrombolysis and 90-day functional outcomes may differ according to diabetes status. These hypothesis-generating observations require confirmation in prospective studies with pre-specified subgroup analyses.
Clinical relevance
Among patients with acute ischaemic noncardioembolic stroke treated with intravenous thrombolysis, does a history of diabetes mellitus modify the efficacy and safety of early adjunctive tirofiban infusion on 90-day functional outcomes? __SNAPSHOT__ {"studyDesign":"Post-hoc subgroup analysis of a phase 3, multicenter, double-blind, randomized, placebo-controlled trial (ASSET-IT)","setting":"38 comprehensive stroke centers in China","sampleSize":"n=832 total parent trial (tirofiban: 414, placebo: 418); DM subgroup: n=191 (23.0%); non-DM subgroup: n=641 (77.0%)","population":"Adults with acute ischaemic noncardioembolic stroke presenting within 4.5 hours of onset, treated with IV thrombolysis (alteplase ~75%, tenecteplase ~25%), not eligible for endovascular thrombectomy; median age 69 years (IQR 59–76); 36.2% women; stroke etiologies included large-vessel atherosclerosis (58%), small-vessel occlusion (35%), and undetermined (6%); 28% had prior stroke history","intervention":"IV tirofiban: loading infusion 0.4 µg/kg/min for 30 minutes followed by maintenance 0.1 µg/kg/min for 23.5 hours, initiated within 60 minutes of IV thrombolysis","comparator":"IV placebo infusion, identical dosing schedule","primaryOutcome":"Excellent functional outcome at 90 days defined as modified Rankin Scale (mRS) score 0–1; stratified by DM vs. non-DM status","followUp":"90 days","fundingSource":"Not reported in abstract; parent ASSET-IT trial was investigator-initiated (funding source for this post-hoc analysis not reported in available sources)"}
Key findings
Among patients WITHOUT diabetes (n=641): excellent functional outcome (mRS 0–1) at 90 days was achieved in 69.0% of tirofiban-treated patients vs. 55.8% of placebo-treated patients (RR 1.22; 95% CI 1.09–1.37; P=0.001). Absolute risk difference: +13.2 percentage points (calculated); NNT ≈ 8 (calculated). Among patients WITH diabetes (n=191): excellent functional outcome was achieved in 55.0% with tirofiban vs. 52.0% with placebo (RR 1.03; 95% CI 0.79–1.34; P=0.846), indicating no statistically significant benefit. Regarding 90-day mortality: in the DM group, mortality was 6.6% (tirofiban) vs. 2.0% (placebo) (RR 3.79; 95% CI 0.84–17.05; P=0.082)—a numerically concerning but not statistically significant signal. In the non-DM group, mortality was 3.4% (tirofiban) vs. 4.4% (placebo) (RR 0.82; 95% CI 0.40–1.70; P=0.597). Formal interaction p-value for DM × treatment was not reported in available sources. Safety outcomes (sICH within 36 hours and any ICH) by DM subgroup were collected but specific stratum-level rates were not reported in the available abstract. In the parent ASSET-IT trial overall, sICH occurred in 1.7% of tirofiban patients vs. 0% placebo; overall 90-day mortality was 4.1% vs. 3.8% (RR 1.07; 95% CI 0.55–2.09).
Practice implications
This post-hoc analysis suggests that diabetes mellitus status may modify the response to adjunctive tirofiban following IV thrombolysis in noncardioembolic ischaemic stroke. The treatment benefit observed in the parent ASSET-IT trial (RR 1.20 for mRS 0–1) appears driven primarily by the non-diabetic subgroup, while patients with diabetes showed no statistically significant efficacy benefit. A numerically elevated mortality signal in the DM tirofiban group (6.6% vs. 2.0%; RR 3.79) warrants attention but is non-significant, the confidence interval is wide, and the subgroup is small (n=191), so this finding cannot establish harm causally. The authors themselves note that diabetes-related endothelial dysfunction, impaired fibrinolysis, renal dysfunction, and longer onset-to-needle times may contribute to attenuated response, but causal mechanisms cannot be established from this data. Patients with diabetes were underrepresented (23%) relative to real-world stroke populations, and the DM subgroup was likely underpowered for safety endpoints. These findings apply specifically to patients with mild-to-moderate noncardioembolic stroke (median NIHSS 6) not eligible for thrombectomy, treated within 4.5 hours, at Chinese stroke centers—generalizability to other ethnicities, health systems, or more severe strokes is uncertain. Practice should NOT change based on this analysis alone; prospective validation with pre-specified stratification by diabetes status is required. Only the abstract was available for this review; full statistical details including the formal DM × treatment interaction test were not confirmable from available sources. __LIMITATIONS__ Post-hoc, unplanned subgroup analysis—diabetes stratification was not a pre-specified primary hypothesis of ASSET-IT; risk of type I and type II errors is elevated.,Small DM subgroup (n=191, 23%): markedly underpowered to detect realistic differences in safety endpoints including mortality and sICH.,Formal treatment × diabetes interaction p-value not reported in the available abstract; it is therefore not confirmed whether the efficacy difference between DM and non-DM subgroups is statistically significant.,Diabetes classified by self-reported 'baseline diabetes history'—no data on glycemic control (HbA1c), diabetes type, duration, or antidiabetic therapy, which may substantially modify vascular biology and treatment response.,Patients with DM had longer onset-to-needle times, introducing confounding that multivariable adjustment may not fully correct.,Trial conducted exclusively at Chinese stroke centers; results may not generalize to other populations with different stroke etiologies, genetic backgrounds, or comorbidity profiles.,Parent trial enrolled patients not eligible for thrombectomy, with mild strokes (median NIHSS 6)—findings do not apply to more severe strokes or those undergoing endovascular therapy.,The overall sICH rate in the placebo arm (0%) is unusually low and has been noted by independent experts as atypical, potentially reflecting selection bias or case-mix characteristics unique to this Chinese trial population.,Funding source for this post-hoc analysis is not reported, limiting assessment of financial conflict of interest.,Only abstract available for this review; full methods, adjusted models, and subgroup interaction statistics could not be fully verified. __BOTTOM_LINE_CATEGORY__ promising_but_preliminary __BOTTOM_LINE__ This post-hoc analysis of ASSET-IT suggests that the 90-day functional benefit of adjunctive tirofiban after IV thrombolysis in noncardioembolic ischaemic stroke appears limited to patients without diabetes (mRS 0–1: 69.0% vs. 55.8%; RR 1.22; P=0.001), with no statistically significant efficacy or safety signal established in the diabetic subgroup. A numerically elevated mortality in diabetic patients receiving tirofiban (6.6% vs. 2.0%) did not reach significance and must be interpreted with extreme caution given the small subgroup size and wide confidence interval. These hypothesis-generating findings should not change practice; prospective trials with pre-specified stratification by diabetes status are needed before tirofiban use can be individualized on this basis.