Introduction
Hyperbaric oxygen therapy (HBOT) is a treatment that delivers 100% oxygen in a pressurized chamber. HBOT enhances tissue oxygenation and neovascularization and proved clinical efficacy in chronic wounds, post-radiation proctitis and enteritis, perianal fistulizing Crohn’s disease, and acute severe ulcerative colitis. Refractory ulcerative colitis (UC) remains a major therapeutic challenge.
Aims & Methods
We evaluated the tolerability, preliminary efficacy, and mechanistic effects including perfusion kinetics measured with intestinal ultrasound (IUS) of 10 and 20 sessions of HBOT in patients with UC. PARADOX was an investigator-initiated, prospective, open label, uncontrolled, dose-finding pilot trial (phase 2a) in biologic-experienced patients with moderate-to-severe, highly refractory UC with a follow-up of 26 weeks (Mayo score >5, MES ≥2, failed ≥2 classes advanced therapies). Patients (n=16; 8 per arm) continued optimized, stable dose background therapy and received 10 or 20 sessions HBOT (2.4 ATA, 120 minutes per day with two 5-minute air breaks). The primary efficacy outcome was predefined composite clinical and endoscopic response (3-point and 30% decrease in Mayo score, RB = 0 and 1-point decrease in MES) at week (W) 12. Other outcomes included tolerability; clinical, symptomatic, endoscopic, and biochemical response, and transmural response assessed by IUS including contrast-enhanced ultrasound (CEUS) to evaluate bowel wall thickness (BWT), vascularity and perfusion kinetics. Endoscopy and IUS were centrally read and adjudicated.
Results
Baseline characteristics were comparable in both cohorts with median age: 48 vs 64 years; disease duration: 12 vs 9 years; ≥4 prior advanced therapies: 3/8 in both; Mayo score: 10.5 vs 9.
At W12, the predefined composite clinical and endoscopic response rate was met in 2/8 patients with 10 sessions and 4/8 patients with 20 sessions. Clinical, endoscopic, symptomatic, biochemical, and IUS parameters all improved over time in both HBOT cohorts with a trend towards larger improvement after 20 compared to 10 sessions (See Table).
At W12, clinical responders demonstrated a significant increase in peak enhancement on CEUS compared to baseline (12.1 dB, 95% CI: 0.8 to 34.3), while non-responders showed a numerical decrease (–5.4 dB, 95% CI: –20.9 to 4.0). This divergent trajectory was consistent across perfusion parameters related to both wash-in and wash-out (e.g., time to peak, wash-in and wash-out rate, area under the curve), suggesting a treatment-related vascular response. In contrast, early gains in non-responders had waned by W12. This effect was numerically larger after 20 sessions, supporting a potential dose–response effect.
HBOT was well tolerated with no dose-limiting toxicities.
| Week-12 efficacy results | 10 sessions HBOT
N = 8 | 20 sessions HBOT
N = 8 | (adjusted) Risk Ratio (95% CI) | P-value |
| Composite clinical and endoscopic response (W12), n (%) | 2 (25) | 4 (50) | 1.7 (0.4 - 4.2) | .31 |
| Clinical remission (W12), n (%) | 0 | 3 (38) | 5.2 (0.6 - 49) | .2 |
| Clinical response (W12), n (%) | 3 (38) | 6 (75) | 1.7 (0.4 - 4.2) | .31 |
| Endoscopic improvement (W12), n (%) | 0 | 3 (38) | 5.2 (0.6 - 50) | .2 |
| Symptomatic remission (W12), n (%) | 4 (50) | 4 (50) | 0.8 (0.2 - 4.2) | .68 |
| Symptomatic response (W12), n (%) | 4 (50) | 6 (75) | 1.3 (0.4 - 3.9) | .45 |
Fecal calprotectin at BL, µg/g, mean (95% CI)
change from BL-W12, EMM (95 % CI) | 1020 (312 - 3370)
–631 (–2086 to –194) | 735 (223 - 2444)
–542 (–1809 to –164) | - | .76 |
Bowel Wall Thickness of sigmoid (mm) at BL, mean (95% CI)
change from BL-W12, EMM (95% CI) | 4.2 (2.8 - 5.6)
–0.44 (–1.15 to 0.28) | 3.9 (2.8 - 5.1)
–0.25 (–0.93 to 0.43) | - | .69 |
Color Doppler Signal of sigmoid, mLimberg ≥2, n (%)
at least 1-point improvement from BL-W12, n (%) | 2 (25)
3 (38) | 4 (50)
3 (38) | 1.1 (0.6 – 2.0) | .66 |
Conclusion
HBOT is a feasible adjunctive therapy in treatment-refractory moderate-severe UC. CEUS showed improved intestinal perfusion in HBOT responders, which offers promising non-invasive tools for response assessment. These findings provide a rationale for a randomized dose-optimized phase 2 trial with potentially longer duration of treatment.
