Prevention of post‐dural puncture headache (PDPH) in parturients. Contributions from experimental research

of 72 + 72 was required. Being part of a PhD thesis, however, including 144 patients was not realistic in the available time frame. So, we intended to conduct a two-centre study and include 80 patients, well aware of the risk of a type 1 error, as mentioned in the discussion. Unfortunately, one of the centres was not able to include patients as planned. Including patients from more than one centre is advantageous in underlining the effect of an intervention. This, however, requires a representative number of cases from each centre. In this specific study, the number of patients, included in the second centre, was not representative. We chose to exclude these patients from analysis, primarily due to questionable quality of data; secondly due to obvious differences in local traditions, regarding anaesthetic and surgical treatment, as well as discharge criteria. Including these patients in analysis would, in our opinion, potentially add more bias than accuracy to the results. Data of the remaining 40 patients, however, were of high quality with limited missing data. Thus, to evaluate the effect of fluid-related complications (non-primary endpoint) we decided to use multiple regression with stepwise elimination in the analysis. Furthermore, the effect of non-matching demographic and intraoperative data was tested in this model. Fitting of the model was performed in collaboration with a professional biostatistician. We agree with Dr. Jammer that conclusions from this study must be drawn with caution. As stated in the title we believe that there may be ‘room for improvement’ in haemodynamic optimisation of patients for lower limb revascularisation. This specific group of patients is at a high risk of having impaired compensatory mechanisms for meeting increased oxygen demands, inflicted by anaesthesia and surgery. Thus, measuring flowrelated haemodynamic parameters and optimising circulation according to well-validated algorithms may very well contribute to improving outcome. We look forward to seeing larger randomised controlled trials investigating this particular subject in the future.