Can Cognitive Training Improve Perioperative Brain Health?

See Articles, p 582, p 586 Perioperative neurocognitive disorders represent a major public health issue, as surgical populations continue to age, and associated complications threaten postoperative recovery.1 Strengthening cognitive reserve represents a candidate strategy for reducing the risk of cognitive impairment after surgery. Low preoperative cognitive performance is associated with postoperative delirium and related outcomes,2 and surgical patients who engage in cognitively stimulating activities preoperatively demonstrate a reduced incidence and severity of delirium.3 While these are observational associations, interventions that aim to enhance cognitive capacities can be tested in randomized controlled trials.4 Given the lack of investigation in this area, and the challenges inherent to conducting behavioral trials, preliminary feasibility studies of interventions designed to provide structured cognitive training are required.4 In this issue of Anesthesia & Analgesia, O’Gara et al5 examined the feasibility of perioperative cognitive training in cardiac surgery patients. The intervention was delivered in the homes of patients using tablet computers without structured supervision from study staff. The primary outcome of feasibility was assessed by enrollment patterns and protocol adherence, defined as enrollment ≥50% of all eligible patients and the proportion of prescribed minutes played, respectively. Overall, 45/69 (65%) patients approached enrolled in the study. However, 129 patients initially met eligibility criteria (Figure 1 in the associated manuscript5), reflecting that enrollment may be <50% in pragmatic, real-world settings. Disinterest, technological issues, and time commitment were the main reasons for declining participation. The highest median (interquartile range [IQR]) protocol adherence, 39% (20%–68%), was in the preoperative setting. Protocol adherence was 6% (0%–37%) in the immediate postoperative setting and 19% (0%–56%) after discharge. No differences were demonstrated between groups in terms of neurocognitive or clinical outcomes. However, as the authors acknowledge, this was a small-scale, feasibility trial that was not powered to detect meaningful differences between groups. Finally, a survey was administered to examine training barriers. Lack of energy, forgetting to train, and feeling overwhelmed were common reasons for not using the training program after enrollment. These findings align with a similar feasibility trial in noncardiac surgery patients, where <20% of participants successfully adhered to the training protocol.4 Similarly, time constraints, technical issues, and feeling overwhelmed were commonly cited, and those randomly assigned to training were more likely to withdraw from the trial. Given the feasibility issues demonstrated, and the labor-intensive efforts likely needed for future trials, is cognitive training a worthwhile endeavor for continued investigation? The answer may still be yes, given that (1) there are plausible biological mechanisms that underlie the cognitive reserve protective effect6,7 and (2) pharmacological trials have generally failed to prevent postoperative delirium and related complications.8 However, we should proceed cautiously and “with eyes wide open,” recognizing that such interventions, while well intended, may be neither practical nor effective for many patients. Moving forward with prehabilitation efforts would also require optimizing study design, participant selection, and training methods. A careful reflection on these considerations is warranted before designing future prehabilitation trials. First, we should acknowledge that many surgical patients would be either ineligible or unlikely to benefit from cognitive prehabilitation. Patients with urgent surgical pathology, preoperative anxiety, time constraints, inability to adhere to training protocols, inexperience with computers, or low expectations will be either ineligible or unlikely to benefit from training.4,9 Comorbidity burden may also preclude successful prehabilitation. Indeed, in the present study,5 patients with risk factors associated with postoperative delirium (eg, depression, anxiety, stroke) were excluded. Thus, cognitive prehabilitation may not be generalizable to high-risk patients. For those who are expected to experience difficulty adhering to training regimens, tailored cognitive training programs could be offered to facilitate adherence. However, this would also require additional, targeted efforts to design, implement, and test such strategies. Second, we need to consider the dose of cognitive training required to observe clinically meaningful effects. Although this was a small feasibility study,5 the median training time was 4 hours (245 minutes) preoperatively, and no outcome differences were demonstrated. To detect meaningful cognitive effects, cumulative training probably needs to reach at least 9 hours10,11; in fact, trials in older adults rarely provide <10 hours of training.12 To follow this study design, participants would require a lead-in time of multiple weeks before surgery. Moreover, it remains unknown how large the cognitive effects of such training would need to be to drive a meaningful reduction in delirium risk, bearing in mind that effect sizes generally decrease in real-world settings, particularly for behavioral interventions.13 Third, study design needs to be enhanced for generating cognitive gains while minimizing attendant risks (eg, training fatigue). Both the trial by O’Gara et al5 and a previous trial in noncardiac surgery patients4 followed an unsupervised design that was based on daily training and relatively brief sessions. A meta-analysis of 51 trials in older adults suggests that, for computerized cognitive training to confer cognitive gains, training efforts require supervision, scheduled breaks to avoid fatigue, and at least 30 minutes of training per session.12 A next step may then be to initiate trials that optimize study conditions and treatment fidelity by incorporating these design elements. Focus could also be shifted to the pre- and postoperative setting given the low inpatient adherence rates reported in this study (6%).5 Finally, attention should be paid to training supervision, defining robust adherence targets, and promoting bidirectional feedback between participants and the research team. Positive findings should then be tested for replication, and intention-to-treat analyses should be strictly followed for estimating externally valid effect sizes. A list of proposed requirements for future trials is outlined in Table 1. Table 1. - Proposed Requirements for Cognitive Training Trials in Surgical Patients Study design Provide all participants with a standardized training platform Adequately describe “dose” of training (eg, cumulative hours, session length) to be achieved Supervised, structured participant training with regular expert guidance Schedule training breaks to minimize anxiety and fatigue Harmonize training efforts—create manual for therapy protocols Select training platform designed for remote supervision and short-term cognitive gainsTable 1. Select cognitive and behavioral outcomes sensitive to symptomatology Monitoring training fidelity Ensure continued supervision of training efforts Monitor for “drift” in adherence to training protocols with interim corrections as needed Automated monitoring of training progress to ensure that prespecified “dose” thresholds are achieved Solicit and provide participant feedback to determine training barriers Consider augmenting training by providing postdischarge training or other behavioral interventions Plan for anticipated difficulties Interim analysis: prespecify rules for stopping a trial if key milestones are not met aDue to surgical scheduling logistics, the length of time available for preoperative training is unlikely to extend beyond a few weeks for many patients. Ultimately, if early findings demonstrate low efficacy or problems with feasibility, there should be a low threshold for stopping trials and pivoting to other candidate avenues of investigation for improving perioperative brain health. Cognitive training trials will be labor intensive and require considerable resources, and these efforts could be redirected to more promising strategies for optimizing perioperative cognitive function if initial findings are not encouraging. In addition, even if rigorous trials demonstrate benefit, computerized cognitive training development would remain incomplete until such interventions were not only effective in research settings, but also implementable—and effective—in a real-world, pragmatic environment. Overall, this preliminary study by O’Gara et al5 can inform the design of future trials of perioperative cognitive training. Given the limited number of effective prevention strategies for delirium and lack of high-quality evidence to support current guidelines,14 novel and cost-effective strategies are certainly desired. Although behavioral trials can be challenging to conduct, the ultimate benefits may still outweigh the risks. With aging surgical populations, delirium-related complications are likely to soon increase. The time is ripe for exploring novel and potentially effective solutions for mitigating the risk of postoperative delirium and related consequences. Does cognitive prehabilitation represent a promising—and realistic—method for improving perioperative brain health? Optimizing methods and conducting the necessary trials may be the only way to know. DISCLOSURES Name: Phillip E. Vlisides, MD. Contribution: This author helped conceptualize the content of the manuscript, write the initial draft, review the clinical and scientific content, and approve the final version of this manuscript. Conflicts of Interest: None. Name: Hannah A. D. Keage, PhD. Contribution: This author helped conceptualize the content of the manuscript, write the initial draft, review the clinical and scientific content, and approve the final version of this manuscript. Conflicts of Interest: H. A. D. Keage receives in-kind research support in the form of software free of charge from HAPPYneuron Inc for a trial of computerized cognitive training in older adults undergoing coronary artery bypass surgery (ACTRN12618000799257). Name: Amit Lampit, PhD. Contribution: This author helped conceptualize the content of the manuscript, write and revise the manuscript, review the clinical and scientific content, and approve the final version of this manuscript. Conflicts of Interest: A. Lampit codeveloped systems for remote delivery of computerized cognitive training as part of industry collaborations funded by the Australian National Health and Medical Research Council (NHMRC GNT1095097) and the German Federal Ministry of Education and Research (BMBF grant 13GW0212A) but has no financial interests in these or any other computerized cognitive training program. This manuscript was handled by: Robert Whittington, MD.