tes unsatisfactorily controlled (HbA1c>7 % ) by combined treatment with insulin sensitizers (metformin and / or glitazones) and insulin secretagogues (sulfonylureas or repaglinide), for whom data from home glucose self-monitoring were available (at least 4 determinations of fasting blood glucose, and 4 determinations 2 hours after each meal, in the last 30 days prior to enrolment). The policy of our Clinic is to suggest monitoring of blood glucose after overnight fast and 2 hours from the beginning of each meal at least one day a week in all patients. The average of available values in the fasting state and after each meal was considered for analysis. Patients included in the analysis had a mean age of 69.3 ± 9.6 years, a duration of diabetes 11.4 ± 10.1 years, body mass index 30.4 ± 4.6 kg / m 2 . Fasting and post-prandial glucose and HbA1c are reported in Table 1 . Twelve patients (25 % ) showed a fasting glucose < 7.22 mmol / l, while 18 (37.5 % ) and 12 (25 % ) had fasting glucose of 7.22 -8.25 and 8.25 -10.0 mmol / l, respectively. Fasting glucose was >10.0 mmol / l in only fi ve cases. Among the 12 patients with fasting glucose < 7.22 mmol / l, 10 (76.9 % ) showed mean glucose after either breakfast, lunch or dinner, greater than 30 % in comparison with fasting glucose; the same occurred in 12 out of 30 (40.0 % ) patients with fasting glucose between 7.22 and 10.0 mmol / l. Both fasting and post-prandial glucose showed a signifi cant correlation with HbA1c (Pearson ' s r = 0.76 and 0.64, respectively; both p < 0.01). Interestingly, age showed a signifi cant inverse correlation with fasting glucose (r = − 0.42; p < 0.01), but not with HbA1c (r = − 0.14) or postprandial (r = − 0.23). Among patients aged more than 65 years (n = 28), correlation of fasting glucose and post-prandial with HbA1c was r = 0.71 and 0.93, respectively (both p < 0.01), while in younger subjects a signifi cant correlation was observed for fasting glucose (r = 0.94; p < 0.01), but not for post-prandial (r = 0.41).