In many centres, a test dose (TD) of octreotide is administered before commencing somatostatin analogue therapy (SAT), although the merits of this procedure are uncertain. We have analysed the value of the GH response to a TD in predicting the efficacy of subsequent SAT in 47 patients with acromegaly (25 male, median age 51 years, range 20-82). The primary goal of SAT was a mean GH of < 5 mU/l. Median baseline GH was 19.3 mU/l (2.2-233 mU/l) and with the TD fell by 78% (35-98%) to a nadir of 4.2 mU/l (< 0.3-85 mU/l). Optimal predictive power was observed when GH fell to < 5 mU/l after the TD. With this criterion, the TD had a positive predictive value (PPV) of achieving the primary goal on SAT of 82% and a negative predictive value (NPV) of 50%. However, baseline GH was also highly predictive of the likelihood of successful SAT (GH < 5 mU/l). The GH response to the TD had PPV of 83% and NPV of 61% of normalising IGF-I on SAT. In summary, baseline GH and nadir after a TD are highly predictive of a good response to SAT; however, a poor response to a TD does not exclude an optimal response to SAT. Furthermore, failure to achieve biochemical control does not equate to no benefit, as biochemical improvement was seen in every patient; therefore, no patient should be deprived of octreotide therapy because of the result of a TD. In conclusion, our data indicate that the octreotide TD has no place in selecting patients for SAT.