We read with interest the present letter by Carlson et al. with regard to our study published in the journal. As these authors noted, our approach was to produce the balance of years of life gained and lost without discussing the other side effects of prostate screening, such as impotence and incontinence, that are associated with treatment. In such an evaluation, randomized trials can only give information on cancer-specific mortality.

Carlson et al. are almost suggesting that biopsy is an absolutely safe procedure, but prostate biopsy is not entirely safe. Although most urologists will rarely see adverse effects, this procedure has been associated with a 6.9% hospitalisation rate for severe complications at 30 days [1] with a twelvefold risk of dying from infectious complications. This side effect is real and should not be neglected or denied by the clinicians performing these biopsies.

The main criticism raised by Carlson et al. is that we use in our calculation an excessively high mortality rate for prostate biopsy. They cite as proof the published study that used data from the European Randomized Study of Screening for Prostate Cancer (ERSPC) [2]. A major limitation of the analysis of ERSPC data for assessing prostate biopsy mortality rates is that the ERSPC was not designed to investigate this phenomenon. This resulted in a lack of statistical power, with only 14 deaths reported <120 days after prostate biopsy. Furthermore, the ERSPC does not contain a proper control group to make a comparison to check whether the mortality rate after prostate biopsy is elevated.

In addition, death after prostate biopsy would occur after sceptic shock, infection, bleeding or cardiovascular complication. Hence, it is not possible to identify cause-specific data indicating that the biopsy could be the cause of death.

The population undergoing biopsy is naturally selected based on the patients' current health condition: using his or her personal judgment, the clinician will not expose an individual whose general health status is considered inadequate; therefore, we believe that only by investigating all-cause mortality with a properly constructed comparison group could information on mortality after prostate biopsy be produced.

To our knowledge, the only available study published on prostate biopsy mortality rates that has a reference group is that of Gallina et al. [3]. Because, this statistic is crucial in the estimation of years of life gained or lost after prostate cancer screening, we contacted the corresponding author of that study (1 February 2010) to request further information on the control group, but we did not receive any such clarification.

In the absence of a randomized trial, and given the difficulty in defining a proper control group and potential biases in the attribution of cause of death, our estimation of a 0.2% mortality rate after prostatic biopsy is valid and perfectly in line with estimates from studies by Loeb et al. [1], Carlson et al. [2], Gallina et al. [3] and Nam et al. [4], and with personal communications on French national insurance data [5].

The study by Gallina et al. should be replicated using a larger study population as we believe that mortality resulting from prostate biopsy is a key element in the evaluation of the efficacy of screening.


  1. Top of page
  2. References