Cerebral oximetry in adult cardiac surgery: A closer look at outcomes

Dear Editor,

We discussed Bennett et al.’s randomised controlled trial on the use of cerebral oximetry in cardiac surgery at our journal club, based in the North West School of Anaesthesia. ¹ This study found a reduced duration of intensive care admission and improved self-reported function at 6 months amongst 90 patients who received cerebral oximetry monitoring and interventions to maintain cerebral oxygen saturation above baseline, compared to 91 patients who received a sham intervention (monitoring attached but not connected). Although these outcomes are potentially important and suggest that the routine use of cerebral oximetry may have a role in cardiac surgery, we feel that the study may not have been appropriately designed to detect other important effects.

Bennett et al. specified five primary outcomes: early neurocognitive dysfunction, neurological dysfunction (i.e. stroke and delirium), functional status at 6 months, cognition at 6 months and hospital length of stay. However, their study was only powered to detect early neurocognitive dysfunction. In the power calculation, the authors assumed a 50% reduction in the incidence of early neurocognitive dysfunction and used data from a similar study by Slater et al. to predict a baseline incidence of 40%. ² However, we note that although Slater et al. randomised 265 patients to cerebral oximetry or control, they found no difference in early neurocognitive dysfunction between groups. ² This prior finding suggests that Bennett et al.’s power calculations were based on a magnitude of clinical benefit that they were highly unlikely to observe. ¹ Furthermore, we suggest that a less ambitious improvement would have more appropriately met the criteria of ‘minimal clinically important difference’.^3,4

A significant difference in self-reported function at 6 months was found by Bennett et al. ¹ Although several validated measures incorporating functional status exist, ⁵ their study used what appears to be a newly created scale. This involved asking patients to state how they were managing stairs, driving, cleaning, dressing, and cooking, in comparison to pre-operatively (i.e. better, the same or worse).^1,4 Whilst this scale has face validity for the assessment of activities of daily living, some activities (e.g. driving) may not be done by all, thereby limiting the options available to some patients. It is unclear how this scale handles these issues, and the authors do not explain the design or justify the use of this questionnaire in their study design, or comment on its reliability or validity.^1,4

The full potential of cerebral oximetry in cardiac surgery remains to be determined. Whilst there are sound theoretical foundations for its use, its potential benefits have not yet been shown to translate into practice. Whilst Bennett et al. have identified some encouraging results, we feel that the generalisability and applicability of their findings may have been hampered somewhat by the study design. To set their findings in context, we would be interested to learn more from Bennett et al. about the development of their follow-up scale, the selection of their primary outcomes and how they arrived at the threshold of 50% reduction in early cognitive dysfunction when powering their study.