P579 Randomised open-label controlled trial of ciprofloxacin/doxycycline/hydroxychloroquine combination compared with standard budesonide in active Crohn’s disease (APRICOT)

Abstract

Abstract Background Mucosal E. coli are increased in Crohn’s disease (CD). They replicate within macrophages where they are inaccessible to penicillins and gentamicin. Hydroxychloroquine is successfully used together with doxycycline, typically for 12 months, to treat Whipple’s disease. It raises macrophage intra-vesicular pH and inhibits replication of bacteria , including E. coli that require an acid pH. Ciprofloxacin and doxycycline are also effective against E. coli replicating inside macrophages. Methods Adult patients with active CD (CDAI>220 plus CRP≥5mg/l and/or faecal calprotectin >250 μg/g) were randomised to receive either standard oral budesonide (Entocort CR 9mg/day for 8 weeks, 6mg/day for 2 weeks, 3mg/day for 2 weeks) or antibiotics/hydroxychloroquine (AB/HCQ) - oral ciprofloxacin 500mg bd, doxycycline 100mg bd, hydroxychloroquine 200 mg tds for 4 weeks, followed by continuation of doxycycline 100 mg bd and hydroxychloroquine 200 mg tds for a further 20 weeks. The use of anti-TNF in the previous 3 months was an exclusion. Primary endpoint parameters were remission at 10 weeks, remission maintained through to 24 weeks, and remission maintained through to 52 weeks. Secondary endpoints included remission and/or response (fall in CDAI of >70) at 10 weeks and remission at 4 weeks. Patients who failed to respond by 10 weeks were invited to cross-over onto the alternative therapy. Results 59 patients were recruited across 8 sites, lower than target (100) because recruitment slowed due to widening access to biologics. Including cross-over, 39 patients received AB/HCQ and 39 received budesonide. No significant differences were seen comparing AB/HCQ with budesonide at 10, 24 or 52 weeks on either intention-to-treat or per-protocol analysis (see table). Withdrawals by 10 weeks due to adverse events were seen in 16 AB/HCQ, - including nausea (3), photosensitivity (2), Achilles pain (2) and activity of CD (4); and 7 budesonide. Although not significant when analysed per protocol, when patients on AB/HCQ who responded at 10 weeks and later remitted were included, 9/24 patients were in remission at 24 weeks and 4/23 at 52 weeks. No correlation was seen between response to AB/HCQ and ASCA/OmpC status. Conclusion The longer-term remissions seen in some patients receiving AB/HCQ are encouraging and a larger phase 3 study is justified.