Antibiotic Prophylaxis and the Transperineal Shift for Prostate Biopsy
A biopsy-naive man with an elevated PSA faces the exact comparison PREVENT was built to settle. PREVENT answered it — and then the two other randomized trials asking the same question did not reproduce the answer.
K.B., a 58-year-old man, manages a small fleet of delivery vans for a regional bakery supplier, work that has him behind the wheel or on the phone with drivers most of the day and has left him with little patience, he says, for anything that sounds like it might sideline him with a complication. An elevated PSA of 6.8 on a routine physical, followed by a prostate MRI showing a PI-RADS 4 lesion, has brought him to a biopsy decision that used to be more straightforward than it has recently become: transrectal biopsy under targeted antibiotic prophylaxis has been the default approach for decades, but a growing share of practices, including this one, now offer transperineal biopsy under local anesthesia as a genuine alternative — one that can be performed without any antibiotic prophylaxis at all.
PREVENT, a ten-center randomized trial, was built specifically to test this comparison head-to-head: transperineal biopsy without prophylaxis against transrectal biopsy with targeted, culture-guided prophylaxis. Its first report, at 658 randomized participants, found zero transperineal-arm infections against four — one point four percent — in the transrectal arm, a difference that fell just short of conventional significance at a p-value of 0.059. Its final intention-to-treat analysis, published in 2024 with the full enrollment of 875 patients, crossed that line: zero infections against six, one point six percent, a difference of negative one point six percentage points with a confidence interval excluding zero and a p-value of 0.02. Cancer detection was equivalent, and transperineal biopsy caused modestly more periprocedural pain that resolved within a week. What complicates K.B.'s question is not PREVENT's own result but the company it keeps: TRANSLATE, a larger randomized comparison, found the same direction at a smaller magnitude — one point two percent against zero point two percent — without reaching significance, and ProBE-PC found no significant difference either. The trial built to answer his question did answer it; the two trials on either side of it did not reproduce the answer, and one positive result among three is a different thing to hand a patient than a single trial read alone.
In clinic, weighing a trial that almost answered the question
I'd recommend transperineal biopsy under local anesthesia with no antibiotic prophylaxis at all. PREVENT is the actual randomized comparison built for this exact question, and in its final intention-to-treat analysis the result was zero infections in the transperineal-no-prophylaxis arm against six, one point six percent, in the transrectal-with-prophylaxis arm — a difference that reached statistical significance. That's the trial designed for this question, answering it. And it avoids fluoroquinolone exposure entirely, which he'd otherwise be taking for a procedure that, done this way, doesn't need it.
I'm not going to dispute the significance, because PREVENT's final analysis did reach it and I'd rather argue about what it weighs than pretend it didn't. My concern is that it's six events. Six, across three hundred seventy men — and PREVENT's own transrectal arm ran at one point six percent, not the five to seven percent that gets quoted from the empirical-prophylaxis era, which tells you culture-guided selection is already doing most of the work. Then look either side of it: TRANSLATE is larger, found the same direction, and didn't reach significance. ProBE-PC found no difference at all. Transrectal biopsy with proper culture-guided antibiotic selection has decades of validated use behind it.
One positive trial with six events, flanked by two neutral ones, is not the same evidentiary object as one positive trial considered on its own. You're quoting the member of that set that most favors your position and calling it the answer.
'Zero infections' sounds decisive, but with only four events total in a sample this size, the trial simply wasn't powered to detect a difference of that magnitude with confidence. A hair's breadth from significance is still not significance, and building a strong preference on a result that close to the line risks over-reading noise as signal.
You're both arguing about how much one significant result weighs against two neutral ones, and I don't think that argument resolves cleanly — the honest summary for K.B. is that the direction of effect is consistent across all three trials and the magnitude is small enough that only the trial built for the question detected it. But notice what the stewardship case doesn't need: it doesn't need that argument settled at all. Avoiding fluoroquinolone exposure entirely, for a procedure that can be done safely without it, has real value on its own terms — every prophylactic course avoided is one less contribution to population-level resistance pressure, whatever the three trials eventually converge on for his personal infection risk.
So my recommendation to him would be transperineal biopsy without prophylaxis, and I'd give him the fuller reason rather than the most persuasive one: cancer detection was equivalent between approaches, one randomized trial found significantly fewer infections while two others found the same direction without significance, and he avoids an antibiotic course entirely for a procedure this practice is equipped to do well without one. The stewardship argument is the part of that which doesn't depend on how the infection-rate question finally settles — and it is, on its own, sufficient here.
K.B. chose transperineal biopsy without antibiotic prophylaxis, after being told plainly that PREVENT's final analysis did find significantly fewer infections, that it rests on six events and was not reproduced at significance by either TRANSLATE or ProBE-PC, and that the recommendation rested as much on antimicrobial-stewardship grounds as on the infection-rate result itself.
Not agreed: whether this practice should adopt transperineal-without-prophylaxis as its new default for all biopsy-naive patients going forward, or continue offering transrectal-with-targeted-prophylaxis as an equally reasonable first option. The transrectal-experienced urologist wants the three trials reconciled — ideally by a pooled analysis rather than another single trial — before treating transperineal as the new standard; the transperineal-adopting urologist and pharmacologist believe the stewardship argument alone is sufficient to prefer transperineal now, in any practice equipped to perform it, whatever that reconciliation eventually shows.