Hyperbaric oxygen did not improve erectile function after radical prostatectomy

Clinical bottom line:

1. HBOT did not improve erectile function 18 months after radical prostatectomy compared to standard therapy.
2. There was no difference between groups for any outcome.

Citation: 1. Chiles KA, Staff I, Johnson-Arbor K, Champagne A, McLaughlin T and Graydon J. A Double-Blind, Randomized Trial on the Efficacy and Safety of Hyperbaric Oxygenation Therapy in the Preservation of Erectile Function after Radical Prostatectomy. The Journal of Urology 2018; 199:805-811.

Lead author's name: Tara McLaughlin, Urology Division at Hartford Healthcare Medical Group Three-part Clinical Question: For patients having radical prostatectomy by robot assisted surgery, does the addition of HBOT to sildenafil and other standard measures, compared to those standard measures alone, result in better preserved erectile function?

Search Terms: Prostatectomy, erectile dysfunction The Study: Double-blind concealed randomised controlled trial without intention-to-treat.

The Study Patients: Men aged 45-60 years old with no serious comorbidities post robot assisted bilateral nerve sparing radical prostatectomy. As part of standard postoperative rehabilitation, all patients had a 12 month course of 50mg sildenafil daily plus use of various erectile aids then sildenafil and/or aids PRN.

Control group (N = 54; 43 analysed): Sildenafil and a range of erectile aids plus sham HBO starting day one after discharge breathing air at 2.2ATA 70 minutes daily Mon-Fri to a total of 10 sessions. Experimental group (N = 55; 40 analysed): As above but breathing 100% oxygen on the same schedule as control. The Evidence:

Non-Event Outcomes Time to outcome Control group HBOT group P-value
Median and IQR for International Index of Erectile Function (ERIC-26)) (0 worst to 30 normal)
18 m

28 (12-29)

21 (8.5-30)



1. 10 treatments seems a short course for any expected therapeutic effect of HBOT for this clinical situation.

2. Authors powered this study to find a difference of six points between groups. They found a difference of seven points (in favour of the control group) but without statistical significance. Probably due to more variability than expected and use of Wilcoxon Rank Sum Test rather than t-test. It is not clear why they compared medians rather than means as other workers using this outcome have done.

3. In this group of relatively young and healthy men, erectile function was near-normal in most control patients (median 28/30). It does not appear likely that any intervention could improve this in a clinically important way using this outcome and comparison. Perhaps comparing the proportion of individuals with a poor outcome would be more fruitful.

Appraised by: K. Witheridge; Friday, 16 March 2018


Kill or Update By: March 2021