High-dose recombinant human GH (rhGH) has been shown to improve the nutritional status of malnourished older adults. It is uncertain whether low-dose rhGH is effective and whether its effect on nutritional status will lead to any improvement in physical function. There is also no data on the outcome after a short course of rhGH treatment. The objectives of this study were to determine the efficacy of low-dose rhGH treatment for 4 weeks in malnourished elderly patients, its effect on physical functions, and the intermediate term outcome after a 4-week rhGH treatment. The study design was a randomized, placebo-controlled, double-blind trial conducted in a university teaching hospital. The patients were 19 medically stable malnourished elderly subjects. Intervention in the rhGH group was as follows: rhGH (Saizen, Serono, Switzerland) 0.09 IU/kg body weight (BW) 3 times weekly were given together with appropriate dietary intervention as prescribed by the dietitian. In the placebo group, equal volumes of normal saline per kilogram BW were given 3 times weekly together with the dietary intervention. The baseline demographic, anthropometric, nutritional, and hematological variables, measures of physical function, and insulin-like growth factor I levels in both groups were comparable. Compared with the placebo group, the GH-treated group showed a more rapid gain in BW (after 3 weeks, +1.27 +/- 0.36 vs. -0.28 +/- 0.37 kg; P = 0.008), total lean body mass (change after 3 weeks by bio-impedance analysis, +1.45 +/- 0.36 vs. -0.37 +/- 0.48 kg; P = 0.009) and a faster improvement in 5-m walking time (decrease after 4 weeks, 23.79 +/- 9.41 vs. 0.45 +/- 4.62 sec; P = 0.047). The hemoglobin level rose more in the rhGH than the placebo groups (change at 8 weeks, +0.84 +/- 0.34 vs. -0.42 +/- 0.29 g/dL; P = 0.012). Serum albumin level also showed a greater delayed increase in the rhGH group than in the placebo group (change at 8 weeks, +5.1 +/- 0.8 vs. 1.6 +/- 1.2 g/dL; P = 0.023). There was no statistically significant difference for other nutritional variables. There was a greater rise in the mean serum insulin-like growth factor I level at 4 weeks in the GH than in the placebo groups (197 +/- 58 vs. 54 +/- 26 U/L; P = 0.034). The improvement in the rhGH group gradually diminished on follow-up and became statistically insignificant 8 weeks after stopping rhGH treatment. There were no GH-related adverse effects. Low-dose rhGH was an effective and safe adjuvant to dietary augmentation for stable malnourished elderly subjects. It led to a faster gain in total lean body mass, which was associated with greater improvement in walking speed when compared with dietary intervention alone. There were no apparent side effects.