An Interview with Dr. Phillip Vlisides:
Feasibility of Home-based Cognitive Prehabilitation in Older Surgical Patients
Phillip E. Vlisides, MD
Assistant Professor, Department of Anesthesiology
University of Michigan
Interviewed by Arnoley S. Abcejo, MD
Assistant Professor, Department of Anesthesiology and Perioperative Medicine
|Phillip E. Vlisides, MD|
|Arnoley S. Abcejo, MD|
How did you get involved in this study? Also why did you decide to study this?
I’ve always been interested in cognitive impairment, because it can have such a distressing impact on patients and families. We have also learned that delirium is associated with adverse clinical outcomes, like prolonged hospital length of stay, cognitive and functional decline, and even increased mortality.
Cognitive training has been studied for several years in non-surgical settings, but we don’t really know whether this type of training is feasible – or effective – in surgical patients for improving postoperative cognitive function. As such, our main objective with this study was to determine the feasibility of such a training program in older surgical patients.
What is postoperative delirium? It seems like there are multiple possible synonyms for postoperative delirium: mental fog, postoperative cognitive dysfunction. How would you classify these terms? Do you think these are all different disease states or part of a spectrum that we don’t fully understand?
These cognitive disorders, like delirium, are largely based on the Diagnostic and Statistical Manual of Mental Disorders (DSM) criteria from the American Psychiatric Association. Delirium reflects an acute, fluctuating disturbance in cognition primarily involving attention and arousal. Per formal recommendations from the Nomenclature Consensus Working Group, the suggested time frame for characterizing postoperative delirium is up to one week postoperatively (or discharge, whichever is sooner). Postoperative Cognitive Dysfunction (POCD) is a research term used to describe cognitive dysfunction several months (or longer) after surgery. Across research groups, there has been considerable heterogeneity in terms of POCD definition, testing methodology and associated time frame. As such, the Nomenclature Consensus Working Group has proposed new nomenclature that aligns with DSM criteria and has a standardized, defined time frame. In brief, the term Postoperative Neurocognitive Disorder is now suggested and this refers to subjectively reported and objectively documented cognitive impairment occurring between 30 days and 12 months after surgery. The interested reader is referred to these publications for further information.
In terms of the neurobiologic basis for these disease states, this is a great question. Our pathophysiologic understanding of altered perioperative brain states very much remains incomplete. PACU delirium has been correlated with postoperative delirium in the days following surgery, so these altered brain states may certainly exist on the same pathophysiologic continuum. Neurotransmitter imbalances – paired with neuroinflammation – may result in dysfunction of key brain networks, and various groups are currently using neurophysiologic and neuroimaging to better understand these relationships.
What is the impact of postoperative delirium or cognitive dysfunction?
Postoperative delirium – and cognitive dysfunction more broadly – is associated with prolonged hospitalization, reduced functional independence, increased healthcare costs and even increased mortality. What isn’t entirely clear is the extent to which delirium causes these outcomes or reflects an underlying neurological and clinical vulnerability.
On a personal level, delirium can be a deeply distressing time for patients and families. Witnessing a family member or loved one experience a profound level of confusion can be quite upsetting and questions naturally arise from those experiences. Will this happen again? Does this mean my family member might have dementia? Or progress to dementia? Such questions present challenging answers, and as a field, I think we’re still trying to understand the underlying scientific causes of delirium and clinical implications, both in the short- and long-term.
In what context has cognitive prehabilitation been shown to be effective — how does that translate to the concept of a surgical home?
In healthy adults, cognitive training exercises have been associated with increases in cognitive performance (i.e., scores on unrelated cognitive function tests) over a short time span. However, performance gains can vary, and unsupervised home training appears to be ineffective. Over a longer period of time, there are some preliminary data to suggest that targeted, supervised training efforts may reduce the risk of dementia (from the NIH-funded ACTIVE Trial), though his remains an active area of investigation, and further research is needed.
In terms of the surgical home model, I think this an important and timely question. With this trial, we found that many patients were excited to participate and many voiced concern regarding postoperative confusion and associated risks. However, before considering implementation into the surgical home model, key questions need to be addressed. First, can supervised training feasibly be conducted – for older patients specifically – over the course of a few weeks leading up to surgery? Second, what are the cognitive gains expected from training over such a short time frame? Do they outweigh any feelings and consequences of stress that might be generated? Lastly, will gains be robust enough to improve clinical outcomes? I think these questions need to be rigorously tested prior to considering implementation.
If your colleagues in anesthesia, neuroanesthesia and critical care could take one thing away from your study, what would it be?
Unsupervised cognitive prehabilitation probably isn’t feasible for many older surgical patients, with factors such as inadequate time and preoperative anxiety serving as training barriers.
If a patient could take one thing away from your study, what would it be?
For “brain training” to be effective, it likely needs to occur over a dedicated period of time (several weeks or months, if possible) with expert supervision.
If you followed the patients you were able to examine and were able to re-do their cognitive testing, do you think you would see a benefit to prehabilitation — was your therapeutic evaluation of success too soon?
This might be possible, but the objective was to determine whether prehabilitation would improve cognitive function in the immediate postoperative setting, when patients might be most vulnerable. Delirium risk seems to peak in the first two postoperative days, when pain, inflammation, opioid use, and sleep deprivation all tend to be at their worst. As such, this time period is when we postulated that prehabilitation might be most helpful. We did perform post-hoc correlation analyses, in which we compared intensity of training with cognitive function scoring across different perioperative time points. We didn’t find any appreciable gains by postoperative day three compared to morning of surgery.
No daily computer access — do you believe if you were able to “lend” an iPad or other device so that participants would be able to engage in the tests, you would have found more success — what did you learn most about research, anesthesia and this complex patient population?
Yes, I do think providing a tablet, or related device, to participants would have helped with enrollment. This may have also helped with technical difficulties, particularly if we would have provided internet connection with the tablet and program training after enrollment.
Interestingly, I was surprised by the number of logistical challenges patients experience before surgery that I didn’t appreciate previously. From a clinical standpoint, after preoperative clinic visits, patients are often sent directly to the laboratory for bloodwork, a radiology suite for imaging, or additional preoperative appointments, and multiple research teams may try to enroll them in additional studies along the way. Patients might then spend several hours at various appointments before driving home (which may be a long commute). Once home, patients often have social obligations to which they need to attend (e.g., family, work) leading up to surgery. So, I think we realized that we’re asking a lot from patients before surgery and I could see how patients may easily become overwhelmed. All of these tasks and obligations exist in addition to the stress and anxiety someone may feel leading up to surgery.
This method of prehabilitation to improve acute cognitive outcome may not be feasible. But what small changes do you think you could have made in the methodology that would have led to success? What about large global changes?
Providing patients with standard tablets may have boosted enrollment and reduced the risk of technical difficulties. Large, global changes would involve supervised, scheduled training sessions with input from an expert neuropsychologist.
Did you find patients more willing to participate depending on the surgery type? Length of anesthesia (i.e. exposure)?
That is difficult to say, as we only recruited patients presenting for major, non-cardiac surgery with expected length of hospital stay at least three days. In that sense, all patients approached were presenting for similar cases with anticipated hospitalization for multiple days. I will say certain patients seemed inherently energetic and motivated to participate, and those patients tended to train for longer, more intense periods, with minimal reminding. Our post-hoc correlation analysis also revealed that baseline cognitive processing speed actually correlated with subsequent training time and intensity. So, baseline motivation and cognitive function may also predict training success and cognitive outcomes. I think these are important considerations for future study design.
Do you think this methodology can be improved to apply to a multicenter study?
I do think this is a possibility, yes. However, considerable resources would be needed to rigorously perform such a trial. First, the training platform (such as a tablet) would need to be provided to every patient. Next, training sessions would likely need to be supervised, either in-person or remotely. These supervised training sessions might not always be logistically feasible. For example, many of our patients had jobs or other responsibilities during the day, and it may be challenging to align patient schedules with those of research team members. Lastly, training probably shouldn’t occur more than a few days a week. This might help minimize anxiety and training stress. Unfortunately, several hours of cumulative training are probably also required to start seeing cognitive gains. So, creating a schedule that allows enough time for training – while minimizing the risk of burnout – might be challenging over a short time period. Our preoperative clinic at Michigan Medicine generally doesn’t see patients earlier than 3-4 weeks prior to surgery. With that type of time constraint, designing an effective and feasible training schedule might be challenging.
If such a multicenter trial were to be designed, investigators may need to consider specifying a stopping point a priori if recruitment is hindered by the above logistical requirements. Nonetheless, even if such a trial were to be stopped early, that might be still be a valuable message.
What can we expect from you and your group for your next study?
Our group’s main research goal is to better understand – and improve – neurocognitive recovery after surgery. We’re currently using electroencephalography to determine whether we can predict postoperative delirium based on advanced brainwave patterns. Additionally, we’re currently studying the effects of caffeine on pain, neuropsychological function and emergence profiles after surgery. We’re hoping to complete recruitment for that trial by fall 2019.