We read with interest the recently-published article by Winiszewski et al. 1 detailing the prevalence and risk factors associated with extracorporeal membrane oxygenation (ECMO) cannulae-related infections (CRIs). Winiszewski highlights a number of pertinent issues related to cannulae infections which, we believe, require further investigation to reduce CRI and improve our patients' outcomes.It is of interest that the definitions of CRI Winiszewski used were of the author's own derivation -primarily because there are currently no standardized definitions of CRI to guide diagnosis in clinical practice. This concurs with previous articles that use variable, nonstandardized or ECMO-specific, infection definitions, 2,3 reporting CRI prevalence of between 1.1% and 24%. [1][2][3][4][5][6][7] Factors such as the inability to routinely culture and replace cannulae, 5,6 artificial temperature regulation, 5,8 and the systemic inflammatory response to the extracorporeal circuit 9 mean that general nosocomial infection surveillance definitions 10,11 and central-line associated bloodstream infection guidelines, 12,13 are not appropriate in this population. Urgent consensus is required to develop ECMO-specific infection definitions, not only for CRI but all nosocomial infections on ECMO, to efficiently and reliably diagnose infectious complications, quantify the clinical issue, and improve patient care.Whilst difficult to diagnose, CRIs can be reduced with diligent care. Winiszewski et al. identify that cannula antisepsis and dressing practices were not protocolized during their study, and we have previously reported widespread cannulae insertion site dressing and antisepsis practice variation internationally. 14 Furthermore, unpublished point prevalence data from the Australia and New Zealand ECMO cohort demonstrate whereas 83% of ECMO centers have a local ECMO cannulae dressing and securement guideline, only 10% of cannulae were dressed in accordance with local guidelines. 15 This suggests further dressing and securement technique/guideline compliance research is needed to inform clinical practice and ensure uniformity in global practice.Although there is a lack of research around optimal cannulae securement techniques, emerging evidence suggests n-butyl-2-octyl cyanoacrylate tissue adhesive (TA) may be a useful adjunct for peripheral ECMO cannulae securement 16 and infection prevention, 17 as it has been successfully used for such indications in smaller intravascular devices. [18][19][20][21] Winiszewski's study highlights that coagulase-negative Staphylococci was one of the pathogens most frequently responsible for CRIs. Research indicates that TA inhibits bacterial growth and migration of Staphylococcus epidermidis, 17 particularly down cannulae insertion tunnels in vitro, whereas increasing the force required to dislodge cannulae compared with transparent, polyurethane dressings (P = 0.003). 16 The application of TA to peripherally-inserted ECMO cannulae insertion sites may therefore be a potential bedside strategy for p...