The empirical use of antibiotics for asymptomatic patients with raised PSA levels confers no benefit and should be discouraged, the authors of a study in the BJUI conclude.

The trial aimed to compare the outcomes for patients who presented with borderline PSA who either had a repeat PSA after 45 days, or who received levofloxacin 500 mg daily for 30 days followed by a repeat PSA after the same interval.

This was not a randomised study, instead patients were offered antibiotic therapy, but if they declined were placed in the control arm. All subjects received a 12-core biopsy of the prostate regardless of the impact of the period of observation or antibiotics on their PSA level.

A total of 202 patients took part in the study with 145 in the antibiotic group and 57 in the control group.

Median baseline PSA was similar in both groups; 7.6 ng/mL and 7.7 ng/mL in the antibiotic and control groups respectively. Median follow up PSA was no different between the two groups, 6.8 ng/mL and 7.0 ng/mL respectively. The percentage PSA variation was also similar. There was no difference either in the number of responders, defined as those patients whose PSA level dropped by more than 20% during the study. The overall detection rate for prostate cancer was 15.8% and did not vary between the groups.

There has been much debate on the best management strategy for asymptomatic patients with borderline PSA and a normal digital rectal examination. Should we proceed direct to transrectal ultrasound and biopsy, simply arrange a repeat PSA after a short delay, or perhaps (more controversially) give the patient a course of antibiotics before repeating the PSA test.

An empirical course of antibiotics has been proposed as a mechanism to help differentiate between benign and malignant disease. The theory is that many patients with borderline PSA have a degree of prostatic inflammation causing a false elevation of their PSA level. Antibiotics would reduce the contribution of this inflammation to the total PSA and would therefore improve the pick up rate of cancer at subsequent biopsy in those patients whose PSA did not respond.

The practice of prescribing antibiotics to patients in this situation has been used to a degree by UK urologists. This study is by no means definitive, but makes it clear that the evidence for this practice is far from conclusive.

For GPs the message must be that this is not a technique we should implement ourselves before referring for a urological opinion. There is clear value in repeating the PSA of a patient with a borderline result, but prescribing of antibiotics is not as yet an established management practice.