We would like to thank Mullish et al. for their letter regarding our study [1,2]. They rightly point out that the amount of faecal material (15e30 g of stool) used in patients treated with capsulized faecal microbiota transplant (FMT) was low and below the minimum of 30 g recommended in the 2017 European Consensus Conference on Faecal Microbiota Transplantation in Clinical Practicefor the treatment of recurrent Clostridium difficile infection (CDI) [3]. At the time when the trial was planned (September 2012) there was still relatively little information available about capsulized FMT and a potential doseeefficacy relationship for the treatment of recurrent CDI. As mentioned in the Discussion section of our study we would encourage future studies to consider using more capsules or repeated administrations. It should, however, be pointed out that the evidence for the optimal dose in FMT remains weak, even for CDI, in the absence of a randomized trial comparing different amounts of faecal matter.The letter further points out that eradication seems to have been slightly higher at the 6-month follow up (visit 5) in centres using administration via the nasogastric tube (and so administering FMT derived from a larger amount of faecal matter). We specifically decided against highlighting this subgroup analysis as the small numbers preclude any meaningful interpretation. Furthermore, at visit 4 (the study visit for the primary outcome, about 4 weeks after FMT) 'success' was 2/5 (40%) for nasogastric and 7/16 (44%) for capsulized FMT, a finding hardly in favour of a superior efficacy of the nasogastric application.We do not believe that there is strong evidence to date to suggest that the FMT should be prepared under anaerobic conditions for the purpose of multidrug-resistant microorganism decolonization. Indeed, a recent study suggests that anaerobically prepared FMT may be more effective to induce remission in patients with ulcerative colitis than aerobically prepared FMT [4]. However, in this study, patients were randomized to receive either FMT prepared anaerobically from a pool of donors or autologous FMT prepared under aerobic conditions. There was no arm with aerobically prepared FMT from pooled donors and the direct effect of the preparative condition is therefore difficult to evaluate.We are also somewhat sceptical that all 'non-CDI indications' are similar, as suggested in the letter. What applies for inflammatory bowel disease may not be true for the purpose of decolonization of multidrug-resistant microorganisms. It should be noted that, using culture methods, we measured the level of extremely oxygensensitive bacteria in faecal suspensions prepared aerobically (n ¼ 6) used for the preparation of capsules and did not find significant variation compared with native stools (personal data, N. Kapel). We agree, however, that the role of anaerobes is definitely worth exploring, possibly in animal models or in simulators of the human intestinal microbial ecosystem such as the SHIME® model [5].Overall, we believe that...