Lina Palaiodimou, MD
Hyperglycemia upon admission is a common phenomenon in acute ischemic stroke and is an independent predictor of poor outcome in both diabetic and non-diabetic stroke patients. More specifically, previous studies have shown that hyperglycemia is independently associated with infarct expansion, early hemorrhagic transformation, impaired recanalization and increased rates of symptomatic intracranial hemorrhage after intravenous thrombolysis. Increased admission and fasting glucose are associated with unfavorable short-term outcome in patients with large vessel occlusion treated with endovascular reperfusion therapies, especially in the subgroup of patients not achieving complete reperfusion following mechanical thrombectomy. However, aggressive serum glucose lowering in clinical studies failed to translate into improvements in functional outcomes, indicating heterogeneity of the biological effects of glucose and the different ways that can modify prognosis in different clinical settings. Accordingly, case-specific glucose management appears to be important.
Κim et al. conducted a post-hoc analysis using data from the Triple-S database, in order to explore the potential association between poor collaterals at presentation and hyperglycemia and the interaction between pretreatment collaterals and glucose upon admission on outcomes. They obtained data from 3 prospective clinical trials (SWIFT, SWIFT PRIME and STAR) and included 309 patients who had acute ischemic stroke with moderate to severe neurological deficit and angiographically confirmed large vessel occlusion and were treated with mechanical thrombectomy within 8 hours of onset. Pretreatment collaterals grades were scored according to the American Society of Interventional and Therapeutic Neuroradiology collateral grading (AGC) system, and reperfusion was evaluated using modified TICI score.
Τhe investigators concluded that there was no association between collateral grade and mean glucose level at hospital admission. Moreover, baseline glucose level was not associated with reperfusion status after mechanical thrombectomy. In addition, hyperglycemia was not associated with reduced likelihood of good functional outcomes at three months in the adjusted analyses of the overall study population. Interestingly, among patients with good pretreatment collaterals, higher glucose levels reduced the likelihood of 3-month functional independence. However, this association was not significant in the subgroup of patients with poor or partial collaterals.
The limitations of the study of Kim et al. are the following: First, sample size was moderate with inadequate representation of patients with very high or low glucose level and collateral status grade 0. Another methodological shortcoming is the possible underestimation of collaterals by pretreatment conventional angiography and the fact that analysis was based on one isolated value of glucose on admission without considering the subsequent evolution of glucose values or possible insulin administration. Most importantly, the study was based on post-hoc analysis and should be considered as hypothesis-generating only, as correctly suggested by the authors.
In conclusion, the study by Kim et al. underscores that pretreatment collateral status could modulate the effect of glucose on functional outcomes in acute ischemic stroke patients treated with endovascular reperfusion therapies. Consequently, large vessel occlusion patients with good collaterals could be a potential therapeutic target group for acute glucose-lowering treatment. Nevertheless, before embarking on a large randomized-controlled trial, independent validation of this association is required in larger datasets, including patients with extreme glucose values, multiple glucose recordings and very poor collaterals grades.