We were interested in the article by Sciarra et al.  suggesting that neuroendocrine markers might complement the PSA assay in selected cases of poorly differentiated tumours, and that an increase in chromogranin A expression, despite low PSA levels, may indicate progression and a poor prognosis. Tissue neuroendocrine markers including chromogranin A and neurone-specific enolase have been identified in all 18 patients with clinically progressive androgen-independent prostate cancer and low serum PSA levels , leading to the belief that most low serum-PSA metastatic prostate cancers were neuroendocrine in origin. Our current work with treatment-naive patients with metastatic prostate cancer and low serum PSA levels at presentation has shown no additional benefit from neuroendocrine markers. Tissue chromogranin A immunostaining was negative in 22 of 33 cases (66%), six of whom were also tissue PSA-negative. In five cases (15%) there was a 2+ intensity of immunostaining, but the pattern of staining was very focal. In the one case where there was 3+ intensity of chromogranin A immunostaining this was seen in very occasional cells, and not sufficient to describe the tumour as neuroendocrine . CD56, which has also been useful in detecting neuroendocrine differentiation , was unhelpful, being focally positive in three of 33 (9%) cases only. Our experience suggests that neuroendocrine markers are of no additional benefit in most metastatic prostate cancers presenting with low serum PSA levels.