We performed the polyethylene glycol (PEG) precipitation test to determine the prevalence of macroprolactinemia in 172 hyperprolactinemic subjects with prolactin (PRL) levels > 700 mIU/l. PRL was measured by Architect i2000 (Abbott Diagnostics) both in serum and in supernatant obtained after PEG. We used the PRL recovery percentage ( R%) as interpretative criterion of PEG test and we estimated a prevalence of macroprolactinemia of 21.5% ( R% ≤ 40%), whereas 74.4% of subjects showed a true hyperprolactinemia ( R% ≥ 60%). In true hyperprolactinemic subjects PRL levels obtained before and after PEG showed a good correlation (Pearson coefficient, R = 0.995); in macroprolactinemic individuals no correlation was found. The PEG test was also performed in 50 normoprolactinemic subjects and a new PRL reference interval was obtained: 64–453 mIU/l. In this case the macroprolactinemic subjects (hyperprolactinemic presenting a normal PRL value after PEG) were 20.3% whereas 79.7% showed still elevated PRL levels. In 35 females we were also able to evaluate the symptoms and the pituitary imaging findings: no significant differences between macroprolactinemic ( N = 11) and true hyperprolactinemic ( N = 24) subjects were found. In conclusion, the PEG test appears to be simple, reproducible and absolutely necessary for diagnostics of hyperprolactinemia.