Variation in absorption of NPH insulin due to intramuscular injection

To evaluate the importance of accidental intramuscular injection of NPH insulin, we measured disappearance rates of 125I-labeled NPH insulin (Protaphane) from subcutaneous and intramuscular injection sites in the thighs of 11 insulin-dependent diabetes mellitus patients. Both subcutaneous and intramuscular absorption rates were measured four times in each patient. NPH insulin was absorbed much faster when given intramuscularly than when given subcutaneously (T50% = 5.3 vs. 10.3 h, P less than 0.0001). The intrapatient (day-to-day) coefficient of variation (C.V.) of T50% values (C.V. T50%) for subcutaneously injected NPH insulin in this study, where all injections were guided by ultrasound determination of the subcutaneous fat layer, was 18.4%. Intrapatient variation of absorption was significantly lower for subcutaneously than intramuscularly injected NPH insulin (C.V. T50% = 18.4 vs. 29.8%, P less than 0.01) and was also lower than interpatient variation for subcutaneously injected insulin (C.V. T50% = 18.4 vs. 50%, P less than 0.0001). The faster absorption rate and shorter duration of action, together with the higher day-to-day variation in absorption, led us to conclude that intramuscular injection of NPH insulin should be avoided.