GH represents several molecular isoforms in addition to the main 22-kDa (22K) GH. There have been reports suggesting that circulating non-22K GH isoforms are increased in acromegaly, but the possible implications of such observations in the management of the disease have not been addressed. The aim of this study was to evaluate the proportion of circulating non-22K GH isoforms in acromegaly. In addition, the relationships between the amount of non-22K GH and tumor size, biochemical measurements, and body composition also were investigated. Samples with different GH levels were selected from 24-h GH profiles from 15 acromegalic men evaluated before and 1 yr after transsphenoidal surgery and from 13 healthy men. The serum non-22K GH levels, expressed as percentage of total GH concentration, were determined by the 22K GH exclusion assay, which is based on immunomagnetic extraction of 22K GH from serum and quantitation of non-22K GH using a polyclonal GH assay. The proportion of non-22K GH isoforms was fairly constant in different samples from the same patient, regardless of the GH level. However, a wide variation of values was observed among acromegalics, both before (14-51%) and after surgery (8-62%). The proportion of non-22K GH isoforms was increased in untreated patients, compared with controls (26.6 vs. 17.4%; P < 0.01), and the values correlated significantly to tumor size, mean 24-h GH concentration, serum PRL, and extracellular water. After surgery, patients not truly cured, with mean 24-h GH concentration of 1 microg/L or more, had an increased proportion of non-22K GH, compared with those with levels less than 1 microg/L (P < 0.01). In the former group, the median values were similar than those in untreated acromegalics (34 vs. 26.6%, respectively), whereas in the latter, they were comparable with those in the controls (15.2 vs. 17.4%, respectively). We conclude that acromegalics have an increased proportion of circulating non-22K GH isoforms. The values are fairly constant in different samples from an individual, regardless of GH level, but a large spectrum can be observed among patients. This variability suggests that different pituitary adenomas secrete GH isoforms in variable amounts. Our observation that a higher proportion of non-22K GH isoforms is present in patients not truly cured after surgery suggests that the evaluation of non-22K GH isoforms can be useful in the follow-up of acromegalic patients.