The aim of this study was to examine the safety and reliability of a research-grade implantable pump for controlled delivery of insulin glargine in cats. For this purpose, a small telemetrically controlled drug delivery pump with a refillable reservoir was implanted into the subcutaneous tissues of the dorsal neck in 10 clinically healthy cats. The reservoir was filled with insulin glargine, and the pump was programmed to deliver four boluses of 0.25 IU/kg, 2-3 weeks apart. As a control, insulin glargine (0.25 IU/kg) was injected SC. Blood glucose and plasma insulin glargine concentrations were measured before each bolus and SC injection and for 8 h afterward. Cats were monitored for signs of discomfort. Pumps were easily implanted and well tolerated by all cats. The experiment was completed in five of 10 cats. In four, the pump failed because of technical reasons; another cat developed severe hypoglycaemia attributable to insulin leakage. Overall, plasma insulin glargine increased after six of eight (75%) initial boluses and after one of 16 (6%) successive boluses. Glucose decreased after seven of eight (88%) initial boluses and after four of 16 (25%) successive boluses. Only the first bolus significantly increased plasma insulin glargine (P = 0.008) and decreased glucose (P = 0.008). Of 20 SC injections, 10 (50%) increased plasma insulin glargine (P <0.001) and 12 (60%) decreased glucose (P <0.001). The pump did not cause discomfort in cats, but life-threatening hypoglycaemia occurred in one. Frequent device problems suggest that the pump needs improvements. Because successive boluses did not increase plasma insulin glargine, this type of insulin may not be appropriate with the pump.