The relation between high on-treatment platelet reactivity (HPR), and the level of glucose intolerance and insulin resistance (IR) was studied in clopidogrel-treated patients with minor ischemic stroke or TIA. The cohort consisted of 66 patients, 11 of which had known type 2 diabetes mellitus (DM). Platelet aggregation in whole blood (Multiplate™) and metabolic variables were measured 1 month after acute onset of neurological symptoms. Glucose tolerance was measured by Oral Glucose Tolerance Test (OGTT). IR was estimated by homeostasis model assessment HOMA-IR. Patients were categorized as "responders" (R) or "non-responders" (NR) to clopidogrel according to an established cut-off in platelet aggregation induced by adenosine diphosphate (ADP). In total, 14/66 (21%) patients were NR. Impaired glucose tolerance (IGT), impaired fasting glucose (IFG) or DM was seen in 13/14 NR (93%), while for R this was the case in 25/52 (48%), p = 0.001. The percentage of NR was 33% in patients with DM and 35% in patients with IGT or IFG. In the group with normal glucose tolerance (NGT) the percentage of NR was low, 4% (1/28). Fasting plasma glucose (f-PG) was higher for NR than for R, 6.0 (5.5-6.7) mM vs. 5.3 (5.0-6.0) mM, p = 0.023. Glycated hemoglobin (HbA1c) did not differ between NR and R. NR also had higher arachidonic acid-induced platelet aggregation than R, and a tendency towards higher aggregation induced by thrombin receptor agonist peptide (TRAP), indicating that HPR reflects a global platelet hyper-reactivity. HOMA-IR was calculated for 52 of the patients above without known diabetes, 9 of which were NR (17%). NR were significantly more insulin resistant than R, with median HOMA-IR 4.5 (3.0-7.4) compared to 2.1 (1.5-3.2) for R, p = 0.001. HOMA-IR and fasting plasma insulin were the only metabolic variables with significant relationships to ADP-induced platelet aggregation. The results suggest that HPR develops in the pre-diabetic phase. A metabolic disturbance with glucose intolerance and/or high level of IR was a pre-requisite for HPR in the tested cohort. Conversely, normal glucose tolerance combined with normal or mildly elevated HOMA-IR excluded HPR. NR are likely to constitute a high-risk group among patients with ischemic cerebrovascular disease. Measurement of f-PG or HbA1c is insufficient to identify NR, while OGTT and HOMA-IR are more predictive.