1. Introduction
Enhanced recovery after surgery (ERAS) is a multimodal program that reduces postoperative pain and complications induced by the stress of surgery [
1]. It has shown its advantages in general surgery, including a reduction in the length of hospital stay and total cost [
2,
3]. ERAS involves multidisciplinary participation, in which the role of anesthesia remains essential. The ERAS philosophy is increasingly implemented in various surgical disciplines, and the ERAS
® cardiac society was created in 2017 [
2].
One of ERAS’s pillars is a multimodal, opioid-sparing pain management plan (recommendation class I level B) [
4]. This is particularly important in cardiac surgery, where high-dose opioid strategies are associated with adverse effects [
5]. Dexmedetomidine is an intravenous selective α-2 receptor agonist that is used for reducing opioid requirements in cardiac surgery patients. Its use for postoperative sedation after cardiac surgery showed significant benefits [
6,
7,
8,
9]. However, data are scarce regarding its feasibility for intraoperative administration as part of the ERAS pathway in elective cardiac surgery with cardiopulmonary bypass (CPBV) where early extubation is one of the objectives.
Since 2015, our Swiss University Hospital has been implementing the ERAS pathway in cardiac surgery patients using an opioid-sparing technique with dexmedetomidine as an adjunct to general anesthesia combined with a bilateral multi-level parasternal nerve block.
The aim of this retrospective study was to share our past years’ experience with dexmedetomidine use as an adjunct to anesthesia and its potential impact on patient outcomes following elective cardiac surgery.
2. Materials and Methods
2.1. Study Design
We performed an observational retrospective cohort study that involved all the adult patients who underwent elective cardiac surgery with CPB and the ERAS pathway during a 14-month period at the University Hospital of Lausanne (Switzerland).
2.2. Patient Characteristics
After obtaining approval from the Ethics committee (CER-V n°2019-00167, 21 February 2019), data from 327 adult patients who underwent elective cardiac surgery between 1 June 2017 and 31 August 2018 were collected and analyzed.
2.3. ERAS Pathway
The ERAS pathway included standardized anesthesia protocol regardless of the type of surgery and comorbidities:
No premedication, except if requested by the patient (benzodiazepine).
A bolus of dexmedetomidine of 0.5 mcg/kg over 10 min before the induction of general anesthesia and immediately followed by a continuous infusion of 0.5 mcg/kg/h.
Induction of anesthesia: Propofol and sufentanil, with rocuronium or cisatracurium, and:
- o
A total of 4 g of intravenous magnesium sulfate before the surgical incision.
- o
A bolus of 10 mg/kg of tranexamic acid before surgical incision, followed by a continuous infusion of 1 mg/Kg/h until the end of surgical hemostasis.
Maintenance of anesthesia:
- o
Sevoflurane titrated to a BIS value between 40 and 60 before and after the cardiopulmonary bypass (CPB).
- o
Continuous propofol infusion monitored by BIS (value between 40 and 60) during CPB.
- o
Top-up doses of sufentanil if needed, up to a total maximum dose of 1 mcg/Kg.
- o
The dexmedetomidine infusion is stopped after weaning from cardiopulmonary bypass.
Parasternal nerve blocks:
Bilateral multi-level parasternal blocks were performed by the surgeon using 40 mL of bupivacaine 5% with epinephrine 1/200,000 at the end of the surgery.
Restrictive fluid therapy, guided by hemodynamic monitoring to avoid volume overload. Transesophageal echocardiography is the preferred technique for fluid guidance, performed by European certified cardiac anesthesiologists.
The systematic intraoperative use of antifibrinolytics and cell salvage, and permissive transfusion targets, i.e., hemoglobin concentration < 70–80 g/L according to clinical judgment.
2.4. Extubation and ICU Admission
All the patients were considered candidates for extubation in the operating room irrespective of comorbidities and the type of cardiac surgery with the use of CPB. Specifically, the patients with severe contractile dysfunctions, pre-existing respiratory conditions, or combined surgeries were not excluded. The decision to extubate the patient at the end of the surgery in the operating room was made by the responsible cardiac anesthesiologist based on hemodynamic stability, the absence of significant bleeding, metabolic or respiratory acidosis, the presence of an adequate neurological state, proper oxygenation, and well-controlled pain.
All the patients were transferred to the intensive care unit. Non-invasive mechanical ventilation commenced and continued in the ICU for two hours for the extubated patients.
Other non-medical reasons could be the cause of the delay in extubation such as logistical issues involving the unavailability of a respiratory therapist or time constraints.
2.5. ICU Discharge Criteria
The patients were discharged from the ICU if they had hemodynamic stability without the need for continuous intravenous medications, except for very low norepinephrine doses; respiratory stability and satisfactory oxygenation levels (normal for the patient); and normal neurological status, and they were awake and alert without significant neurological deficits, with a well-controlled pain with oral medications.
2.6. Data Collection
Demographic, intraoperative, and postoperative data were retrospectively retrieved from the institute’s administrative database AXYA (Cerner France, Puteaux, France) the electronic patient files Soarian (Cerner corp. North Kansas City, MO, USA) and Metavision 5 (IMDsoft, Tel Aviv, Israel), and the anesthesia record.
The following variables were collected:
2.6.1. Preoperative
General: age, sex, BMI, comorbidities, treatment, cardiovascular risk factors, and ASA classification.
Clinical and Biological: hemoglobin, renal function, left ventricular ejection fraction (LVEF), and NYHA class.
2.6.2. Intraoperative
The duration of the procedure, duration of CBP, duration of aorta cross-clamping, and nadir temperature during CBP.
Blood loss and transfusion requirements.
Opioids administered and the duration of dexmedetomidine infusion.
Hypotension (<55 mmHg mean arterial pressure) and hypertension (BP > 25% above the baseline).
Bradycardia (<50 beats per minute (bpm)) and tachycardia (HR > 25% above the baseline).
Insulin requirement (insulin drip started if blood glucose > 10 mmol/L).
Inotropic or vasopressor support.
Sternal infiltration by local anesthetics (mL).
2.6.3. Postoperative
Extubation time.
Blood loss and transfusion requirement within the first 24 h.
If reintubation is needed (for patients extubated in the OR).
If surgical re-exploration is needed.
Postoperative complications: cardiac, renal, neurological, infectious, and respiratory.
Postoperative opioid consumption calculated in Morphine Milligram Equivalent.
(MME = iv morphine (mg) + iv fentanyl (mcg)/10 + oral oxycodone (mg)/2)
Length of stay in ICU (ICU LOS).
Length of stay in hospital (HOSP LOS).
Mortality at 30 days.
Total hospitalization costs.
2.6.4. Endpoints
Intraoperative safety endpoints: the incidence of hypotension, hypertension, hyperglycemia, cardiac arrythmias, and need for vasoactive drugs. Postoperative safety endpoints: the incidence of postoperative complications, time of extubation, ICU and hospital LOS, and total cost of hospitalization.
2.7. Statistical Analysis
The distribution of data was evaluated using the Kolmogorov–Smirnov test. Normal distributed data are represented as mean (±SD, range), and non-normal distributed data are represented as median (IQR, range). Frequencies are represented as numbers (%).
Three groups were identified based on mechanical ventilation time: group 0: extubated in the OR, group < 6: extubated in the ICU within 6 h postoperatively, and group > 6: extubated in the ICU after 6 h postoperatively.
Group comparisons were made using the Wilcoxon rank sum and Kruskal–Wallis tests for continuous variables and Pearson’s chi 2 for continuous variables. p < 0.05 was considered statistically significant after correction for multiple comparisons using the Bonferroni method. Statistical analysis was performed using the JMP 15 software.
4. Discussion
The advantages of using dexmedetomidine in perioperative cardiac surgical care are well established, but almost all the studies concentrate on its use as a sedative in the postoperative period, or on postoperative outcomes. Limited data are published on the intraoperative effects of dexmedetomidine as an adjunct to general anesthesia.
In two meta-analyses comprising 52 studies [
10,
11], only 23 studies reported on the intraoperative use of dexmedetomidine. In the remaining 29, dexmedetomidine was administered only for postoperative sedation in the ICU.
A total of 18 of the 23 studies, where dexmedetomidine was used intraoperatively, only had postoperative outcome measures such as the markers of myocardial injury, renal injury, inflammation, conduction block, and cognitive function.
Of the remaining five, two used dexmedetomidine only as a bolus at the induction of anesthesia to blunt the sympathetic response to endotracheal intubation, and two concerned off-pump cardiac surgery. Only one study dating from 1997 [
12] describes its use as an adjunct to general anesthesia in cardiac surgery. In this randomized, double-blind trial, 40 patients received a placebo and 40 received dexmedetomidine at 0.5 mcg/Kg/min for 30 min prior to the induction of anesthesia and 0.07 mcg/Kg/min from the induction to the end of surgery. Further, high dose opioid (fentanyl 30 mcg/Kg) was used for induction and maintenance (0.15 mcg/Kg/min), and completed with enflurane inhalation anesthesia. Their study population was highly selective to exclude severe cardiac (ventricular dysfunction, severe valvular dysfunction, and > 50% left main stem stenosis) and other systemic illnesses.
They found that dexmedetomidine use resulted in lower blood pressure and heart rate at induction, during intubation, and throughout surgery. The incidence of hypotension, defined as a drop of >30% from baseline, was similar in both the groups, yet systolic blood pressure drop below 90 mmHg was more frequent in the treatment group. Tachycardia was more frequent in the placebo group and bradycardia was similar, with eight patients (20%) in each group requiring pharmacological intervention. Post CPB pacing for AV conduction block was more frequent in the treatment group, but the difference did not reach statistical significance.
In a much more recent trial, Aguerreche et al. [
13] retrospectively compared 40 patients in whom an opioid-free anesthetic protocol was used during cardiac surgery with 40 patients who were treated with an opioid-based anesthetic regimen. The opioid-free strategy comprised dexmedetomidine (bolus and continuous infusion), magnesium sulfate, dexamethasone, lidocaine (bolus and continuous infusion), and ketamine (bolus and continuous infusion). In the opioid-based strategy, remifentanil and propofol were used as target-controlled infusions. The comparison focused on postoperative opioid consumption and pain scores. They also compared intraoperative hemodynamic variables and postoperative complications. In terms of opioid reduction, through the use of Dexmedetomidine intraoperatively, Coustrouglo et al. reached the same results [
14].
The dexmedetomidine group required less morphine in the first 24 h after surgery (15 mg vs. 30 mg) while having lower pain scores (significant only during coughing). They did not find a difference in the incidence of bradycardia, nor in the intraoperative use of norepinephrine; yet, more patients in the opioid-based group required ephedrine boli.
Hemodynamic changes that are reportedly associated with dexmedetomidine infusion are hypertension, hypotension and bradycardia, atrial fibrillation, and cardiac conduction blocks.
We evaluated the effect of dexmedetomidine on blood pressure and heart rate during the pre-CPB phase of surgery, since bypass, bleeding, and the surgical procedure itself are major confounding factors. Hypertension and hypotension after the onset of the dexmedetomidine infusion occurred in 6.2% and 13.6% of the patients, respectively, but were easily manageable in all the cases using short-acting vasodilators or vasopressors. Bradycardia is a more frequent side effect (25% of the patients) easily managed by the administration of anticholinergic drugs in 8.6% or even the interruption of the infusion in 6% of the patients.
Postoperative atrial fibrillation was more frequent in the patients undergoing valve surgery (32.1%) than in the patients undergoing CABG surgery (24.5%). These incidences are consistent with the reported data [
15,
16,
17] and indicate that the intraoperative use of dexmedetomidine is unlikely to influence the occurrence of atrial fibrillation. Atrioventricular second- and third-degree conduction blocks were much more frequent in surgical procedures including valve repair or replacement than in isolated CABG, requiring permanent pacing in 5.4% of the cases. Again, these results are consistent with the reported data [
18], and suggest that dexmedetomidine does not increase the risk of permanent high-grade conduction block.
Taking into consideration the analgesic effect of dexmedetomidine, opioid use was minimal. An opioid-sparing technique is an essential prerequisite for early extubation.
Only 39 patients (12%) received more than 1 mcg/Kg of sufentanil during surgery, and the maximum dose administered was 1.6 mcg/Kg. However, due to its relatively short elimination half-life, the intraoperative administration of dexmedetomidine does not provide postoperative analgesia. The median 24 h Morphine Milligram Equivalent (MME) consumption in our subgroups of patients who were extubated immediately or within 6 h were 28 mg (IQR 15.5–51) and 30 mg (IQR 19.5–59.25), respectively. These results are similar to the results reported by Barr et al. [
19] who compared bilateral multi-level parasternal blocks using ropivacaine 0.75% with a placebo. They found reduced pain scores and opioid consumption in the treatment group which had a 24 h MME consumption of 30.8 mg, identical to our results. This suggests that the addition of dexmedetomidine during surgery does not reduce opioid consumption beyond its intraoperative effect.
Postoperative delirium is one of the most frequent complications after cardiac surgery [
20], and has been a research topic for improvement for decades. While many studies describe a significant reduction in delirium when using dexmedetomidine in the ICU [
21], this has been questioned based on the results of some recent trials [
22,
23,
24]. We diagnosed a very low number of patients (6.7%) suffering from postoperative delirium (CAM-ICU score >3), compared to the published rates going from 17 to 24% in the patients receiving dexmedetomidine. This discrepancy is likely due to the different definitions of delirium used. Most of our patients (57.2%) were extubated in the operating room (OR) after surgery, but logistical factors were the most common reason for non-extubation. In fact, the implementation of non-invasive mechanical ventilation (NIV) following surgery has effectively mitigated respiratory complications such as atelectasis and CO
2 retention. However, the restricted availability of NIV to anesthetists beyond 5 pm at our center introduces a potential confounding factor in our dataset. This limitation could obscure any advantages associated with early extubation, as no significant differences in the ICU length of stay (LOS) or cost were observed between the patients extubated in the operating room or shortly after ICU admission.
Additionally, our hospital discharges patients from the ICU in the morning, even if they meet the criteria for transfer to the ward the night before. Also, our stepdown unit cannot accommodate cardiac surgery patients immediately after surgery. While several of the patients extubated in the OR probably did not require ICU surveillance, the mandatory transit through the ICU increases their hospitalization cost and masks any benefits of early extubation.
This retrospective study has many limitations. The most obvious is the absence of a control group, making it difficult to determine if the observed outcomes are due to the intervention or other confounding factors. As for all the retrospective analysis, data collection relies on existing records, which may be incomplete or biased. The quality of the data might have affected the results as much as the intervention studied. The studied population includes patients with very diverse characteristics, requiring caution when generalizing the results to specific subgroups and contexts.