The use of dexmedetomidine and intravenous acetaminophen for the prevention of postoperative delirium in cardiac surgery patients over 60 years of age: a pilot study

Background: Delirium is associated with many negative health outcomes. Postoperative sedation and opioid administration may contribute to delirium. We hypothesize that the use of dexmedetomidine and Intravenous acetaminophen (IVA) may lead to reduced opioid consumption and decreased incidence of postoperative delirium. This pilot study aims to assess feasibility of using dexmedetomidine and IVA in cardiac surgical patients, and estimate the effect size for incidence and duration of delirium. Methods: A total of 12 adult patients >60 years of age undergoing cardiac surgery were recruited and randomized into 4 groups: Propofol only (P), Propofol with IVA (P+A), Dexmedetomidine only (D), Dexmedetomidine with IVA (D+A). Preoperative baseline cognition and postoperative delirium was assessed daily until discharge. The feasibility was assessed by the number of patients who completed the study. Results: All patients completed the study successfully. The total incidence of delirium in the study population was 42% (5/12): 67% (2/3) in the group P, and 67% (2/3) in the group D, 33% (1/3) in D+A group and 0%(0/3) P+A group. The incidence of delirium was 17% (1/6) in the group receiving IVA compared to 67% (4/6) that did not receive IVA. The mean range of duration of delirium was 0-1 days. One patient expired after surgery, unrelated to the study protocol. One patient in the D group experienced hypotension (systolic blood pressure <90 mm of Hg.) Conclusions: The feasibility of performing a project is ascertained by the study. Patients receiving IVA had lower incidence of delirium compared to patients not receiving IVA which suggests that IVA may have a role in reducing the incidence of delirium. A prospective randomized, placebo-controlled trial will be the next step in investigating the role of dexmedetomidine and IVA in reducing the incidence of delirium.


Introduction
Delirium is defined as a change in mental status, characterized by acute onset and fluctuating course, inattention, disorganized thinking, and altered level of consciousness 1,2 . Delirium increases the risk of mortality, readmissions, and accelerated cognitive decline 2-6 . Around 158 billion dollars of national healthcare cost is attributable to delirium 7 . The incidence of delirium in cardiac surgery is 11-46% 2 . Delirium is preventable in 30-40% of the cases 2 . Some of the modifiable risk factors for delirium include the choice of analgesic and sedatives 8 .
Currently used sedatives, like midazolam, act via GABA receptors and release deliriogenic mediators 9 . Dexmedetomidine is an alpha 2 adrenergic agonist, with no interaction with GABA receptors.
A recent study by Li et al. in 285 elderly cardiac surgical patients evaluated dexmedetomidine versus propofol and showed no difference in the incidence of delirium 10 . A major limitation of this study is that CAM-ICU is limited in delirium assessments and may miss the diagnosis. There are other tools that could help assess cognitive assessment more efficiently. A meta-analysis suggested that the use of dexmedetomidine for sedation in cardiac surgery patients may reduce the incidence of delirium 9 . The first study in the meta-analysis was a retrospective study by Corbett SM et al. that does not mention how delirium was assessed in individual hospitals 11 . The second study by Shehabi Y et al. was not powered adequately and used CAM-ICU for assessing delirium both in intubated and extubated patients which have reduced sensitivity for delirium in verbal patients and could have underestimated the incidence of delirium 12 . The third study by Maldonado JR et al. of recruited118 patients and 90 patients were finally analyzed. 28 randomized patients were excluded for protocol violations which introduce a significant selection bias 13 . It also showed a 94% reduction in the incidence of delirium and this effect size is almost implausibly large and has not been seen in other studies. The final study by Dasta JF et al. was not designed to identify patients with delirium with any specific tools such as CAM and thus could have gross underestimated delirium incidence 14 . The use of IV acetaminophen (IVA) has been shown in multiple studies to reduce the amount of opioids consumed by patients undergoing surgeries 4,15 . IVA has never been studied in the context of cardiac surgery and delirium prevention. So, we hypothesized that the use of dexmedetomidine and acetaminophen would provide adequate analgesia, leading to a decreased opioid consumption and thus decreased incidence of delirium.

Purpose
This is a pilot trial which ran from November 2013 to November 2014 to assess the feasibility of using IV dexmedetomidine and acetaminophen in the cardiac surgical intensive care unit. The study also aimed to obtain effect size estimates for primary study outcomes which will help power an ongoing large scale randomized trial (NCT02546765, registered on 13 th January 2015).

Study design
This is a single-centered, double-blinded, prospective, randomized controlled pilot trial.

Study population
A homogenous set of patients were chosen for this feasibility study. Patients who are 60 years of age or older who were undergoing coronary artery bypass grafting (CABG), and/or valve surgery were included in the study. Exclusion criteria were preoperative left ventricular ejection fraction < 30%, emergent and percutaneous procedures, aortic surgeries, preexisting cognitive impairment, recent seizures, patients on medications for cognitive decline, serum creatinine > 2 mg%, liver dysfunction, known history of alcohol or drug abuse, and hypersensitivity to any of the study drugs. Patients who might not get extubated in a reasonable amount of time (hence the EF, aortic surgery exclusion, etc) were excluded. Since Q6H IV acetaminophen was given, any clinical situation that can potentially set up patients for drug toxicity etc. were avoided. (Figure 1)

Randomization
Fourteen patients were recruited (the initial 2 patients to help train the investigators in the use of cognitive assessments). The remaining twelve patients were randomized into the following 4 groups, containing 3 patients each (1:1:1:1 allocation). (Table 1) Each patient was allotted a randomization number based on which the pharmacist assigned the study medications.

Study drug administration
Peri-operative anesthetic management was administered according to standard of care. Intra-operative propofol infusion was administered at the clinician's discretion for patients in propofol group. In the post-operative period, propofol infusion was titrated to 25-100 µg/kg/min. In the dexmedetomidine group, dexmedetomidine infusion was given after chest closure at a dose of 0.1-1.0 µg/kg/hr. The medications for sedation were continued in the post-operative period until extubation or for a minimal duration of 6 hours. For patients randomized to IVA group, IVA 1 gram was given every 6 hours for the first 48 hours postoperatively up to a total of eight doses.

Amendments from Version 1
The following changes, applied to the manuscript, are based on the referees' comments: • Emphasized that only the feasibility of the project is ascertained. We agree that a larger study is needed to draw a concrete conclusion.
• Reasons for the exclusion criteria were mentioned.

Patient assessment
Cognitive assessments were conducted using standard battery used for Mini Mental (MMSE) and Confusion Assessment Method (CAM).

Outcomes of the study
The primary outcome measured was the proportion of patients who completed the protocol successfully, which reflects the feasibility of the study. The incidence of delirium was measured to obtain effect size estimates for future studies. Secondary outcomes included hypotension, duration of delirium, breakthrough analgesic requirements, ICU days, and length of hospital stay.

Results
The baseline characteristics of the patients were comparable in all groups. The total incidence of delirium was 42% (5/12). The incidence of delirium was 67% (2/3) with propofol and 67% (2/3) with dexmedetomidine. Dexmedetomidine+Acetaminophen group had an incidence of 33% (1/3). The Propofol+Acetaminophen group had no occurrence of delirium. Interestingly, only 17% (1/6) of the subjects who received IVA were diagnosed with delirium compared to 67% (4/6) in the group who did not receive IVA. (Table 2) Also, the incidence of delirium was 33% (2/6) in the propofol groups as compared to 50% (3/6) in the dexmedetomidine groups. The mean duration of delirium ranged from 0 to 1 day. (Table 3) Secondary outcomes were similar

Group 3 (D) Dexmedetomidine No
Group 4 (D+A) Dexmedetomidine Yes *All groups will receive bolus doses of IV opioids (morphine and hydromorphone) as needed for breakthrough pain. I.V. Acetaminophen will be given to group 2 and group 4 every 6 hours for 48 hours, postoperatively.
between the groups. One patient expired after surgery, unrelated to the study protocol. One patient in the dexmedetomidine group experienced a significant hypotension with systolic blood pressure <90 mm of Hg.

Discussion
The use of IV dexmedetomidine and acetaminophen in the cardiac surgical patients was feasible. The study protocol was easily incorporated into patient care. All the enrolled patients completed the study protocol. There were no instances where the implementation of the study protocol was abandoned by the physicians. Sedation provided was useful. Apart from the incidence of hypotension in one patient with dexmedetomidine use, no other adverse events recorded directly related to the intervention.
No other statistical evaluation was done due to the small sample size. The overall incidence of delirium in our trial, at 42%, is close to previous reports 2 . The incidence of delirium in patients receiving IVA was 17% compared to 67% in the other group suggesting that IVA may be beneficial in reducing the incidence of delirium following cardiac surgery. The pilot study led to a larger ongoing trial (NCT02546765).  Table 3. Table 3 provides the incidence of delirium and secondary outcomes in the 4 groups in the study namely propofol, propofol + acemaninophen, dexmedetomidine, dexmedetomidine + acetaminophen.

Dexmedetomidine Dexmedetomidine + Acetaminophen
Incidence of Delirium % (n/N) 66.6 (2/3) 0 (2/3) 66.6 (2/3) 33.3 (1/3) A major limitation of our study is the small sample size, but this was a feasibility trial to study effect size for the design of future larger studies. Also, there was no blinding for the choice of sedatives used in the patients. The assessors were blinded to acetaminophen administration.

Mean Delirium duration [n days]
A multi-center randomized, controlled trial will be the next step in investigating the role of dexmedetomidine and IVA in reducing the incidence of delirium.

Data availability
F1000Research: Dataset 1. The datasets used and/or analyzed during the current study. No statistical evaluation was done due to the small sample size., 10.5256/f1000research.12552.d180828 16

Ethics and consent
This study has been approved by the Committee on Clinical Investigations at Beth Israel Deaconess Medical Center (IRB Protocol #: 2013-P-000149). Informed consent was obtained for all subjects prior to initiation of study procedures. The study was conducted from 13 th November 2013 to 9 th April 2015.

Competing interests
The authors declare that they have no competing interests of conflicts of interest.

Grant information
Balachundhar Subramaniam is supported by the National Institutes of Health Research Project Grant GM 098406.
The funders had no role in study design, data collection and analysis, decision to publish, or preparation of the manuscript. 1.

2.
3. The authors report results of a pilot study looking at postoperative administration of acetaminophen in reducing the incidence of postoperative delirium (POD) after cardiac surgery in elderly patients.

Open Peer Review
I believe that the authors are asking an important question whether adequate or superior pain relief is associated with reduced POD rates after cardiac surgery.

Concerns:
It is not clear to me why this pilot study is referred to as a feasibility trial. Both dexmedetomidine and acetaminophen are used routinely in many cardiac centres. Please comment.
I would strongly advise against quoting outcome measures as percentages of patients, you have 3 patients per group, e.g., quoting 2/3 (67%) is misleading. If I understood correctly DEX was given until extubation or for 6 hours postoperatively. Not a standard administration practice for DEX. Please comment on rational for this design.
With respect to pain relief after cardiac surgery, spinal/epidural analgesia has been associated with superior pain relief and significant opioid sparing effects, however, there are no studies that have shown a reduction in delirium rates with either of these techniques. Please comment if acetaminophen may have some intrinsic properties other than analgesia in reducing POD after cardiac surgery.

Is the study design appropriate and is the work technically sound? Yes
Are sufficient details of methods and analysis provided to allow replication by others?

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility? Yes Although this study is considered a pilot trial, the small number of patients 12 divided into 4 groups is a serious flaw in methodology. The interpretation of the data is completely meaningless not even as a feasibility study. feasibility study. I believe this study should not be more than a run-in phase for a proper pilot with reasonable numbers.
I do not recommend publication in F1000. Thanks.

If applicable, is the statistical analysis and its interpretation appropriate? No
Are all the source data underlying the results available to ensure full reproducibility? Partly

Are the conclusions drawn adequately supported by the results? No
No competing interests were disclosed. This is an interesting and important feasibility study, which seeks to determine whether it is possible to conduct a (large) study evaluating various combinations of sedation/analgesia post cardiac surgery, with the intention of preventing postoperative delirium. In general, with a very small feasibility study, the investigators demonstrate that conducting such a study might be possible. However this study has important constraints.
The investigators state, "The mean duration of delirium was 0-1 days." This appears to be a range.
The investigators conclude, "The feasibility of performing a large-scale project is ascertained by the study." I would rather argue that the feasibility of performing a project is ascertained. Whether or not a large-scale project is feasible is not necessarily established.
Contrary to the investigators, I do not believe that a 12 patient study with four groups supports the inference "that IVA may have a role in reducing the incidence of delirium." This is over-interpretive of very limited data.
I do not agree that using the CAM is a major limitation in delirium research (as the investigators suggest that it might lack sensitivity). I do agree that this might apply to the CAM-ICU.
More efficient tools might not necessarily be more sensitive.
I am not convinced that a 12 patient pilot trial will produce sufficiently precise estimates to power a large RCT appropriately. Although the absolute difference between proportions (delirium incidence) was 50% (67% minus 17%), the 95% confidence interval for this difference is -13% to 82%. Thus, it is not clear based on this small study what value should be used (estimated difference in delirium incidence between the groups [acetaminiphen and placebo]) for the sample size calculation.
From the registration site (for NCT02546765), it appears that 120 patients will be included in the (large) trial. This number is likely to be much too small to answer the research question with sufficient precision.
The reasons for the exclusion criteria are not clear.

2.
With the modified factorial design, was interaction between interventions considered? Specifically, both dexmedetomidine and acetaminophen might have an impact on the outcome of interest (delirium).
The (secondary) results of the study were as follows: "The Propofol+Acetaminophen group had no occurrence of delirium. Interestingly, only 17% (1/6) of the subjects who received IVA were diagnosed with delirium compared to 67% (4/6) in the group who did not receive IVA." The investigators should be cautious about over-interpreting these findings, or using these estimates to justify sample size for a large trial. The problem with this would be that 67% is a very high incidence, and is not likely to reflect the incidence with current standard of care (i.e., propofol or dexmedetomidine sedation, without intravenous acetaminophen). Similarly, the incidence of 17% is actually likely to be pretty close to the current delirium incidence in the type of patients who were enrolled to this study (but who receive standard of care).

If applicable, is the statistical analysis and its interpretation appropriate? Not applicable
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? No
No competing interests were disclosed.

Competing Interests:
I have read this submission. I believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above. Thank you for reviewing the paper. We have thoroughly reviewed the comments. We are addressing below, point-by-point all the suggestions.
The investigators state, "The mean duration of delirium was 0-1 days." This appears to be a range.Thank you. Yes. We reported the mean duration as a range.
The investigators conclude, "The feasibility of performing a large-scale project is 2. 3.
The investigators conclude, "The feasibility of performing a large-scale project is ascertained by the study." I would rather argue that the feasibility of performing a project is ascertained. Whether or not a large-scale project is feasible is not necessarily established. Thank you. We agree that only the feasibility of performing a project is ascertained. Contrary to the investigators, I do not believe that a 12 patient study with four groups supports the inference "that IVA may have a role in reducing the incidence of delirium." This is over-interpretive of very limited data.Thank you. We agree this is a pre-pilot study indeed. It was a feasibility project and we certainly do not want to overinterpret. Thus, we wrote "IVA may be beneficial in reducing the incidence of delirium following cardiac surgery. " Since this is a feasibility study we agree that a larger study is needed to draw a concrete conclusion. I do not agree that using the CAM is a major limitation in delirium research (as the investigators suggest that it might lack sensitivity). I do agree that this might apply to the CAM-ICU.Thank you. Agreed. More efficient tools might not necessarily be more sensitive.Thank you. Agreed. I am not convinced that a 12 patient pilot trial will produce sufficiently precise estimates to power a large RCT appropriately. Although the absolute difference between proportions (delirium incidence) was 50% (67% minus 17%), the 95% confidence interval for this difference is -13% to 82%. Thus, it is not clear based on this small study what value should be used (estimated difference in delirium incidence between the groups [acetaminiphen and placebo]) for the sample size calculation.Thank you. We planned this pilot study as a means to assess the feasibility of performing a large-scale project. However, there is no other preexisting data to power the next larger trial. Since this pilot study showed a 66% reduction in the incidence of delirium, for the power calculation, a 50% reduction in the delirium incidence was used. Each group needed 56 patients (alpha=0.05 and 1-Beta=0.80).
From the registration site (for NCT02546765), it appears that 120 patients will be included in the (large) trial. This number is likely to be much too small to answer the research question with sufficient precision.Thank you. A multicenter trial is definitely required to answer with precision. However, it will be a major leap to go from a pilot study to a multicenter trial. Thus, we chose to do the second step of testing in a larger sample size with what is known so far. Depending on what this study shows, the obvious next step will be a larger multicenter trial.
The reasons for the exclusion criteria are not clear.Thank you. We chose a homogenous set of patients to determine the drug efficacy. We wanted to avoid patients who might not get extubated in a reasonable amount of time (hence the EF, aortic surgery exclusion, etc). As we chose to give Q6H IV acetaminophen, we wanted to avoid any clinical situation that can potentially set up patients for drug toxicity etc. Preexisting cognitive dysfunction, etc. could muddle the picture in a 120 patient study. In a larger multicenter trial if and when launched, the heterogeneity can be accounted for and thus need not necessarily be an exclusion criteria.
With the modified factorial design, was interaction between interventions considered? Specifically, both dexmedetomidine and acetaminophen might have an impact on the outcome of interest (delirium).Thank you. Yes, these interactions will be analyzed in the ongoing trial The (secondary) results of the study were as follows: "The Propofol+Acetaminophen group had no occurrence of delirium. Interestingly, only 17% (1/6) of the subjects who received IVA were diagnosed with delirium compared to 67% (4/6) in the group who did not receive IVA." The investigators should be cautious about over-interpreting these findings, or using these estimates to justify sample size for a large trial. The problem with this would be that 10.
these estimates to justify sample size for a large trial. The problem with this would be that 67% is a very high incidence, and is not likely to reflect the incidence with current standard of care (i.e., propofol or dexmedetomidine sedation, without intravenous acetaminophen). Similarly, the incidence of 17% is actually likely to be pretty close to the current delirium incidence in the type of patients who were enrolled to this study (but who receive standard of care).Thank you for the caution. We agree with all the comments.
The authors report no competing interests. In this single center randomized controlled trial (RCT), Susheela and coworkers investigate the use of dexmedetomidine and intravenous acetaminophen for the prevention of postoperative delirium in elderly patients undergoing cardiac surgery. This pilot study primarily aimed to assess the feasibility of a larger RCT with identical interventions.
As a pilot project, the study is methodically sound. Overall, the manuscript is well-written and pleasant to read. The authors applied CONSORT criteria which is a definite plus.
Postoperative delirium has specific features that may be related to surgery (and anesthesia) and has a probable neuroinflammatory component. I therefore do not agree with your statement that 30-40% of delirium cases could be prevented, at least not in your sample of patients undergoing cardiac surgery. So far, multicomponent interventions to prevent delirium have only been proven efficacious in a general population of elderly hospitalized patients.
Please also discuss the opioid-sparing effects of intraoperative dexmedetomidine. Next to the missing interaction with GABA receptors, this might be an additional mechanism for delirium prevention when using dexmedetomidine.
Please discuss the reasons to exclude patients with preexisting cognitive impairment from your study. Preoperative cognitive impairment was found to be one of the most important risk factors for postoperative delirium (odds ratio ranged from 6.3 (95% CI 2.9 to 13.7) up to 11.5 (95% CI 6.1 to 20.1)). By excluding these patients (also in the follow-up RCT), you likely decrease the number of patients affected by the primary outcome measure (postoperative delirium) in your study. It would be interesting to investigate the effects of dexmedetomidine and intravenous acetaminophen in these high-risk patients. I find it difficult to blind the intravenous administration of propofol and dexmedetomidine. At best, you may try to conceal the study drug from patient, surgeon, ICU caregivers, and investigators by using opaque syringes and lines. Was study drug administration really double-blind?
Was there any kind of standardization in anesthesiologic management? If yes, state drugs and dosage.
At what frequency did you assess patients for postoperative delirium in the ICU/on the ward? Who 1 2 1.

2.
At what frequency did you assess patients for postoperative delirium in the ICU/on the ward? Who assessed the patients?

Response
In this single-center randomized controlled trial (RCT), Susheela and coworkers investigate the use of dexmedetomidine and intravenous acetaminophen for the prevention of postoperative delirium in elderly patients undergoing cardiac surgery. This pilot study primarily aimed to assess the feasibility of a larger RCT with identical interventions.
As a pilot project, the study is methodically sound. Overall, the manuscript is well-written and pleasant to read. The authors applied CONSORT criteria which is a definite plus.

Thank you
Postoperative delirium has specific features that may be related to surgery (and anesthesia) and has a probable neuroinflammatory component. I, therefore, do not agree with your statement that 30-40% of delirium cases could be prevented, at least not in your sample of patients undergoing cardiac surgery. So far, multicomponent interventions to prevent delirium have only been proven efficacious in a general population of elderly hospitalized patients.1 Thank you. These estimates are from the pilot study result. The ongoing study will inform us better about the effect size.
Please also discuss the opioid-sparing effects of intraoperative dexmedetomidine. Next, to the missing interaction with GABA receptors, this might be an additional mechanism for delirium prevention when using dexmedetomidine.
Thank you. Agreed. It is possible that dexmedetomidine can reduce the analgesic requirements. The ongoing study will inform us of this effect size as well due to the crossover design.
Please discuss the reasons to exclude patients with preexisting cognitive impairment from your study. Preoperative cognitive impairment was found to be one of the most important risk factors for postoperative delirium (odds ratio ranged from 6.3 (95% CI 2.9 to 13.7) up to 11.5 (95% CI 6.1 to 20.1)).2 By excluding these patients (also in the follow-up RCT), you likely decrease the number of patients affected by the primary outcome measure (postoperative delirium) in your study. It would be interesting to investigate the effects of dexmedetomidine and intravenous acetaminophen in these high-risk patients.
Thank you. As this is the first study of this size, we would like to get a homogenous group examined first and definitely the next step is to do the patients with preexisting cognitive . dysfunction I find it difficult to blind the intravenous administration of propofol and dexmedetomidine. At best, you may try to conceal the study drug from patient, surgeon, ICU caregivers, and investigators by using opaque syringes and lines. Was study drug administration really double-blind?
Thank you. We only double-blinded the Acetaminophen arms.
Was there any kind of standardization in anesthesiologic management? If yes, state drugs and dosage.
Cardiac surgery is the most protocolized intraoperative setting. Patients were induced with 500mcg of IV fentanyl, 50-100mg of Propofol, and endotracheal intubation was facilitated with 1mg/kg intravenous rocuronium. Anesthesia was maintained with isoflurane (0.6-1.0 MAC) in 100% oxygen with intermittent fentanyl up to a maximum of additional 500 mcg of fentanyl and intermittent rocuronium. All patients were taken intubated to the ICU At what frequency did you assess patients for postoperative delirium in the ICU/on the ward? Who assessed the patients?
Assistants trained from Edward Marcantonio's lab (with initial half a day training session, witnessing the assessments for a few patients and also calibration with gold standard assessor). Postoperative delirium was assessed by a trained study team member with the help of cognitive assessments such as MMSE, CAM-ICU, and CAM starting from the day after the surgery. If cognitive assessments are negative for 3 consecutive days (i.e days 5, 6 & 7), assessments were completed every other day until the date of discharge.
I wish the investigators of this pilot project the best of luck with the larger ongoing trial!

Thank you
The authors report no competing interests.

Competing Interests:
The benefits of publishing with F1000Research: Your article is published within days, with no editorial bias You can publish traditional articles, null/negative results, case reports, data notes and more The peer review process is transparent and collaborative Your article is indexed in PubMed after passing peer review Dedicated customer support at every stage For pre-submission enquiries, contact research@f1000.com