Heres a studie on different types of administration:
Medical Department M (Diabetes & Endocrinology), Aarhus University Hospital, Kommunehospitalet, Aarhus University, Denmark.

The current mode of growth hormone (GH) replacement therapy is daily subcutaneous (s.c.) injections given in the evening. This schedule is unable to mimic the endogenous pulsatile pattern of GH secretion, which might be of importance for the induction of growth and other GH actions. The present study was conducted in order to study the pharmacokinetics of different doses of GH following intranasal (i.n.) administration and the biological activity of GH after i.n. administration as compared with sc and intravenous (i.v.) delivery. Sixteen GH-deficient patients were studied on five different occasions. On three occasions GH was administered intranasally in doses of 0.05, 0.10 and 0.20 IU/kg, using didecanoyl-L-alpha-phosphatidylcholine as an enhancer. On the other two occasions the patients received an sc injection (0.10 IU/kg) and an i.v. injection (0.015 IU/kg) of GH, respectively. The nasal doses and the sc injection were given in random order in a crossover design. In a double-blinded manner the subjects received the three nasal doses as one puff in each nostril. The patients received no GH treatment between the five studies or during the last week before the start of each study. Intravenous administration produced a short-lived serum GH peak value of 128.12 +/- 6.71 micrograms/l. Peak levels were 13.98 +/- 1.63 micrograms/l after s.c. injection and 3.26 +/- 0.38, 7.07 +/- 0.80 and 8.37 +/- 1.31 micrograms/l, respectively, after the three nasal doses. The peak values of the 0.05 and the 0.20 IU/kg nasal doses were significantly different (p = 0.007). The mean levels obtained by the low nasal dose were significantly lower than those obtained with the medium (p < 0.001) and the high dose (p < 0.001), while there was no significant difference between the medium and the high doses. The absolute bioavailability of GH following s.c. relative to i.v. administration was 49.5%. The bioavailabilities of the nasal doses were: 7.8% (0.05 IU). 8.9% (0.10 IU) and 3.8% (0.20 IU). Serum insulin -like growth factor I (IGF-I) levels increased significantly after s.c. administration only. Mean levels were significantly higher after s.c. administration as compared with the i.v. and all three nasal does (p < 0.001). Serum IGF binding protein 3 (IGFBP-3) levels remained unchanged on all five occasions. Mean serum IGFBP-I levels were significantly lower after s.c. GH injection than after administration of the i.v. (p < 0.001) and the three nasal doses (p < 0.005). Subcutaneous GH administration resulted in significantly higher levels of serum insulin and blood glucose (p < 0.001). In conclusion, the bioavailability of nasal GH was low (3.8-8.9%). An i.v. bolus injection of, on average, 1 IU of GH induced no metabolic response. Only s.c. GH administration induced increased levels of IGF-I, insulin and glucose. These data reveal that a closer imitation of the physiological GH pulses than achieved by s.c. GH administration is of limited importance for the induction of a metabolic response to GH.