Doubt has been thrown recently on the value of prostate specific antigen (PSA) tests in diagnosing prostate cancer. Some men have raised values but have no cancer, or only an extremely slow-growing one, while others have values below 4 ng/mL and still have prostate cancer. What's to be done? 'Watchful waiting' has been advocated for some men who have localized prostate cancer when it's agreed that the likely benefits of treatment are outweighed by the adverse effects of that treatment. The decision may be helped by a new analysis of repeated PSA determinations.

Most men with prostate cancer have a non-palpable tumor, and are diagnosed as possible cancer by a raised PSA level. This study was planned to see whether the rate of rise in the PSA during the year before diagnosis, the PSA level at diagnosis, the Gleason score (a way of expressing the severity of the cancer on microscopy) and the tumor stage could identify those men at risk of death from prostate cancer.

This sort of information was available for 1804 men treated by radical prostatectomy at the Barnes-Jewish Hospital in St Louis from January 1989 to June 2004. After exclusion of those who were found to have lymph node metastases, those who received radiotherapy, and those who had only one PSA level determination before surgery, there were 1095 men left for evaluation in the study. Their average age was 65?, and their average PSA was 4.3 ng/mL.

Clinical tumor staging was done pre-operatively, and Gleason scores were assigned on prostate biopsy and prostatectomy specimens.

Follow-up lasted, on average, for 5 years after prostatectomy. During this time PSA levels were measured every 6 months and digital rectal exams every year. The PSA values were used to calculate the rate of PSA increase during the year before diagnosis. The rates were compared with the lymph-node status, the clinical stage, the Gleason scores, the time to recurrence, death from prostate cancer, and death from any cause.

There were no changes in the time to recurrence, death from prostate cancer, or death from any cause when the rate of PSA increase was below 2.0 ng/mL per year. However, at rates above 2.0 ng/mL there was a significant increase in all three of these 'endpoints'.

In addition, a PSA increase rate of more than 2.0 ng/mL per year was linked to lymph node metastases, a higher clinical stage and a higher Gleason score, when compared with lower rates.

Men with a PSA that increases at more than 2.0 ng/mL a year who are otherwise healthy should consider radical prostatectomy along with some form of anti-hormonal therapy. Watchful waiting is not a particularly good option for these men.

In men who have had a prostatectomy, different guidelines apply; if there is a doubling of the PSA within a given period - 3 to 10 months, according to different authorities - the prognosis is significantly worse.

There have been recent reports suggesting that a PSA below 4.0 ng/mL is no longer a 'safe' level, carrying no risk of prostate cancer. The response to these new findings should be: have an annual check-up by your physician from 55 onwards, making sure it includes a digital rectal exam and a PSA test. If the level is above 4.0 ng/mL or it has increased 2.0 ng/mL over the previous year's value, have further investigations done by a qualified urologist.