Randomized trial of near-infrared spectroscopy for personalized optimization of cerebral tissue oxygenation during cardiac surgery

Background. We assessed whether a near-infrared spectroscopy (NIRS)-based algorithm for the personalized optimization of cerebral oxygenation during cardiopulmonary bypass combined with a restrictive red cell transfusion threshold would reduce perioperative injury to the brain, heart, and kidneys. Methods. In a randomized controlled trial, participants in three UK centres were randomized with concealed allocation to a NIRS (INVOS 5100; Medtronic Inc., Minneapolis, MN, USA)-based 'patient-specific' algorithm that included a restrictive red cell transfusion threshold (haematocrit 18%) or to a 'generic' non-NIRS-based algorithm (standard care). The NIRS algorithm aimed to maintain cerebral oxygenation at an absolute value of> 50% or at> 70% of baseline values. The primary outcome for the trial was cognitive function measured up to 3 months postsurgery. Results. The analysis population comprised eligible randomized patients who underwent valve or combined valve surgery and coronary artery bypass grafts using cardiopulmonary bypass between December 2009 and January 2014 (n = 98 patient-specific algorithm; n = 106 generic algorithm). There was no difference between the groups for the three core cognitive domains (attention, verbal memory, and motor coordination) or for the non-core domains psychomotor speed and visuospatial skills. The NIRS group had higher scores for verbal fluency; mean difference 3.73 (95% confidence interval 1.50, 5.96). Red cell transfusions, biomarkers of brain, kidney, and myocardial injury, adverse events, and health-care costs were similar between the groups. Conclusions. These results do not support the use of NIRS-based algorithms for the personalized optimization of cerebral oxygenation in adult cardiac surgery.