Background: The number of prostate biopsies are estimated to approximately 1.000.000 each
year I Europe1. The infection rates after transrectal prostate biopsy are increasing
annually2. The most common antibiotic used for prophylaxis is Ciprofloxacin, however,
Trimethoprim/Sulfamethoxazole is also a feasible alternative3.
Aim: To evaluate the best antibiotic prophylaxis prior to trans rectal prostate biopsy in low
risk patients.
Outcome: Hospitalization for infection within 14 days from biopsy.
Method: By randomization of all eligible patients with low risk of infectious complications.
Patient and treating physician is blinded to treatment allocation. By using the nationwide
and full coverage national patient register (NPR), mandatory for inpatient care4. All
patients can be identified by the Swedish personal identification number5. Both patients
related baseline characteristics is gathered through register linkages. To validate the
outcome variables, for all patients admitted to hospital within 14 days will be subject for
chart review. Comparing all patients from the participating units, having a code specific to
trans rectal biopsy, and comparing to the excluded patients as well as included patients,
will make a separate exclusion analysis.
The standard operating procedure for randomization is that patients will fill a form and
informed consent of participation prior to the planned biopsy. If no exclusion criteria are
filled and informed consent obtained, patient will be randomized through a web based
randomization program, where only date of biopsy, earlier prostate biopsy, number of
biopsies, PSA and prostate size is collected. Other baseline variables will be obtained by
cross linkage to the NPR and the National Prostate Cancer Registry. A randomization number is
given to all patients at this stage and a key code is established at each participating unit
containing every patient randomization number and subsequent personal identification number.
After complete requirement, data will be collected at Västernorrlands Läns Landsting and
analyzed according to protocol.
Analysis will be made by logistic regression and point estimates with Single sided 97.5%
Confidence Interval. Sample size is calculated to require 2800 patients assuming a 0.5%
frequency of infections, able to detect a difference of 0.75% in absolute proportion of
infections.
Missing data on outcome variable, received treatment or date of biopsy will be excluded from
analysis.