Retrospective Comparison of Intramuscular Admixtures of Ketamine and Dexmedetomidine Versus Ketamine and Midazolam for Preoperative Sedation

Abstract

Precooperative children and patients with intellectual disabilities often require intramuscular (IM) sedation prior to the induction of general anesthesia (GA). Ketamine is an effective preinduction sedative but can produce significant adverse side effects. Dexmedetomidine, a sedative with sympatholytic and analgesic properties, may provide advantages when used in combination with ketamine. This retrospective study evaluated the efficacy and safety of IM ketamine with dexmedetomidine for preoperative sedation. We conducted a chart review of all patients (n = 105) treated for dental rehabilitation who received either IM ketamine and dexmedetomidine (study group, n = 74) or IM ketamine and midazolam (control group, n = 31) prior to induction of GA. No significant difference (p = .14) was observed in the time interval from IM administration to operating room entry (median [interquartile range]) between the study and control groups (5 [4–8] vs 5 [2–7] minutes). Patients who received IM dexmedetomidine exhibited significantly lower mean arterial pressures throughout the induction (p = .004) and had lower heart rates (p = .01) throughout the intraoperative period compared with patients who did not receive dexmedetomidine. The combination of dexmedetomidine and ketamine may provide effective and safe IM sedation prior to the induction of GA.

Providing anesthesia to precooperative children and patients with intellectual disabilities can be very challenging. The decision to proceed with general anesthesia (GA) for dental rehabilitation often occurs after 1 or more failed attempts to deliver dental care using behavioral management techniques, mild or moderate sedation, or physical restraint. Significant anxiety, fear, and agitation commonly interfere with acceptance of oral premedication, intravenous (IV) access, and transport to the operating room (OR). In extreme circumstances, combative and violent behavior may pose a physical danger to the patient, caretakers, and the anesthesia care team.

The use of intramuscular (IM) sedative agents, which does not require patient cooperation, can facilitate an efficient transfer to the OR and induction of GA. Although IM ketamine is effective and commonly used for the management of special need patients, it has numerous undesirable side effects including dysphoria, hypersalivation, hyperreactive airway reflexes, hypertension, tachycardia, muscle hypertonia, hallucinations, increased incidence of nausea and vomiting, and emergence delirium.¹ Many of these side effects are dose dependent.^1,2 It is common practice at our institution to coadminister IM ketamine with midazolam to reduce the ketamine dose, and, therefore, the incidence and severity of ketamine-related negative side effects.

Dexmedetomidine, a centrally-acting α₂ adrenergic agonist with sedative and analgesic properties, is most frequently administered intravenously. However, alternative routes of administration, including oral,³ intranasal,⁴ and IM⁵ have been reported. Coadministration of IM dexmedetomidine and ketamine might produce optimal sedation and diminish the negative side effects of ketamine. There are limited data to support this practice⁶; therefore, we conducted a retrospective chart review to evaluate the efficacy and safety of IM ketamine admixed with dexmedetomidine as a premedication prior to the induction of GA. We hypothesized that IM ketamine and dexmedetomidine (KD) is an effective and safe preoperative sedative in comparison with IM ketamine and midazolam (KM).

MATERIALS AND METHODS

Following institutional review board approval with waiver of informed consent, a retrospective chart review was conducted. All patients who were treated by dentist anesthesiologists at Stony Brook University Hospital from January 1, 2016, to January 1, 2019, and who received an IM injection of ketamine (with or without any additional agents) prior to induction of GA were screened for eligibility. The initial study population was filtered to group patients who received IM KD as the study cohort or IM KM as the control cohort.

Relevant patient demographic data obtained from the electronic medical record (EMR) included age, sex, height, weight, American Society of Anesthesiologists (ASA) physical status classification, cognitive/behavioral diagnoses, concurrent use of oral α₂ adrenergic agonist, and preoperative vital signs.

Temporal data obtained from the EMR included time of IM administration, arrival into the OR, IV catheterization, induction, intubation, surgical start and end, extubation, OR discharge, postanesthesia care unit (PACU) discharge, and hospital discharge. The route of anesthesia induction, administration of vasopressors or anticholinergics, and perioperative adverse events were also recorded.

Efficacy was defined as the time elapsed from administration of the IM injection to OR entry. This was selected as a reasonable surrogate for effective sedation because timely transfer to the OR implies adequate sedation. Safety was defined as the absence of adverse clinical events or unwanted side effects. Adverse effects were defined as follows: (1) clinically significant hypotension, as inferred from the use of a vasopressor (phenylephrine, ephedrine, or epinephrine) within 60 minutes of induction; (2) clinically significant bradycardia, as inferred from the use of an anticholinergic within 60 minutes of induction; and (3) prolonged duration of the recovery period in the PACU, as determined by the time elapsed from OR discharge to PACU discharge. All adverse events that were documented in the anesthesia record were collected.

Data extracted from the EMR were recorded on paper case report forms and transcribed into a study specific RedCAP database. The study's statistician (co-author JR) exported these data into SAS(c) 9.4 Software (Cary, NC). Categorical variables were compared using either χ² or Fisher's exact tests as appropriate. For continuous variables, Shapiro-Wilk tests were used first to test for normality, and comparisons were tested using either Student's t tests (equal or unequal variance) for normally distributed variables or Wilcoxon rank sum tests for nonnormally distributed variables. A Kruskal-Wallis test was used for the additional efficacy-related subgroup analysis, and Dunn's test for multiple comparisons was used to further examine these differences between subgroups. All statistical analyses were performed using an α = 0.05 except for the Dunn's test, which utilized an α = 0.0125.

RESULTS

A total of 115 patients were identified who received IM ketamine during the study period. Of these, 10 were excluded from analysis: 1 patient who received only IM ketamine and 9 patients who received IM ketamine, dexmedetomidine, and midazolam. The remaining 105 patients met eligibility criteria and were included in the analysis. Demographic data are shown in Table 1. Of note, 59% of patients were younger than 13 years old. The majority (64%) had at least 1 behavioral diagnosis, with autism spectrum disorder (ASD) being the most common.

Table 1.

Demographic, Medication, and Surgical Time Data*

	Ketamine and Dexmedetomidine (n = 74)	Ketamine and Midazolam (n = 31)	P Value
Demographics
Age <13	42 (57%)	20 (65%)	.46
Male	44 (60%)	21 (68%)	.43
ASA III	36 (49%)	14 (45%)	.53
Preoperative behavioral diagnosis	45 (61%)	22 (71%)	.32
Autism spectrum disorder	37 (50%)	16 (52%)	.88
POS BP recorded	49 (66%)	19 (61%)	.63
Medications administered
Ketamine administered (median, IQR), mg/kg	2.5 (2.0–2.9)	2.5 (2.1–3.0)	.71
Dexmedetomidine administered (median, IQR), mcg/kg	0.80 (0.77–0.9)	n/a	n/a
Midazolam administered (median, IQR), mg/kg	n/a	0.10 (0.08–0.1)	n/a
Fentanyl (median, IQR), mcg/kg	50 (35–100)	58 (38–100)	.94
Surgical times
Surgery duration (mean, SD), min	145 (56)	130 (51)	.02
PACU length of stay (median, IQR), min	61 (39–94)	45 (29–73)	.06
Time from IM to PACU discharge (median, IQR), min	273 (232–319)	240 (207–267)	.007†

^*ASA, American Society of Anesthesiologists; POS, preoperative services; BP, blood pressure; IQR, interquartile range; PACU, postanesthesia care unit; IM, intramuscular.

^†Denotes statistical significance.

Drug Dosages

The median (interquartile range [IQR]) IM ketamine dose administered was 2.5 (2.0–2.9) mg/kg in the KM group and 2.5 (2.1–3.0) mg/kg in the KD group. The median (IQR) dose of IM dexmedetomidine in the KD group was 0.80 (0.77–0.90) mcg/kg. The median (IQR) dose of IM midazolam in the KM group was 0.1 (0.08–0.1) mg/kg. Differences between the 2 groups regarding intraoperative fentanyl use were not statistically significant.

Efficacy

The median (IQR) time from IM administration to OR entry was 5 (2–7) minutes for the KM group and 5 (4–8) minutes for the KD group, which was not a statistically significant difference (p = .14). There was no significant difference in the percentage of patients in whom GA was induced by the IV route (p = .36) (Table 2).

Table 2.

Efficacy, Safety, and Hemodynamics*

	Ketamine and Dexmedetomidine (n = 74)	Ketamine and Midazolam (n = 31)	P Value
Efficacy
Time elapsed from IM to OR entry (median, IQR), min	5 (4–8)	5 (2–7)	.14
IV induction	31 (42%)	16 (52%)	.36
Safety
Anticholinergic administered	39 (53%)	19 (61%)	.42
Vasopressor administered	7 (9%)	1 (3%)	.27
Intraoperative adverse event	3 (4%)	0 (0%)	.55
Hemodynamics
Highest induction MAP (median, IQR), mm Hg	77 (71–86)	85 (77–98)	.004†
Lowest induction MAP (mean, SD), mm Hg	61 (10)	74 (13)	.0002†
Highest induction HR (median, IQR)	92 (73–106)	98 (81–115)	.19
Lowest induction HR (mean, SD)	83 (21)	92 (24)	.12
Highest intraoperative MAP (median, IQR), mm Hg	82 (72–90)	83 (77–100)	.20
Lowest intraoperative MAP (mean, SD), mm Hg	59 (9)	66 (9)	.0002†
Highest intraoperative HR (median, IQR)	85 (72–97)	103 (79–126)	.01†
Lowest intraoperative HR (mean, SD)	80 (16)	89 (20)	.01†

^*IM, intramuscular; OR, operating room; IV, intravenous; MAP, mean arterial pressure; HR, heart rate; IQR, interquartile range. IV induction is the number of cases in which GA was induced through an IV. Induction is defined as time from IM injection until intubation, and intraoperative is defined as the period from the time of intubation plus the subsequent 60 minutes.

^†Denotes statistical significance.

Safety

No statistically significant differences were observed between the 2 groups regarding rates of anticholinergic or vasopressor administration or adverse events. Although the reason for administration was not explicitly stated in the anesthesia record, the dentist anesthesiologists who were part of this study often administer IV glycopyrrolate to patients receiving IM ketamine due to its antisialagogue effect. No patients in this study received anticholinergic drugs for the treatment of bradycardia. No skin or IM reactions were identified at the IM injection site in either the study or control groups.

Hemodynamics

This study evaluated the highest and lowest mean arterial pressures (MAPs) and heart rates (HRs) during the induction (time from IM injection to intubation) and the immediate intraoperative periods (postintubation plus 60 minutes). During the induction period, the maximum and minimum MAP measurements in the KD group were significantly reduced compared with the KM group (p = .004 and p = .0002, respectively). During the immediate intraoperative study period, the KD group demonstrated significantly reduced minimal MAP, minimal HR, and maximum HR compared with the KM group (p = .0002, p = .01, and p = .01, respectively; Table 2).

Prolonged Recovery

Time from IM injection to PACU discharge was prolonged in the KD group (273 minutes vs 240 minutes, p = .007). Median duration of surgery was prolonged in the KD group (145 minutes vs 130 minutes) but failed to reach statistical significance (p = .20). Patients in the KD group had longer post anesthesia care unit (PACU) stays (61 minutes vs 45 minutes), which was not a statistically significant difference (p = .06) but was close.

Prolonged Recovery: Supplemental IV Dexmedetomidine

Several patients enrolled in this study received IV dexmedetomidine beyond the intraoperative study period (after the first 60 minutes following intubation), either as part of GA maintenance or prior to extubation as prevention for emergence agitation. A subgroup analysis examined the effect of IV dexmedetomidine on prolonging PACU discharge times. Patients in the study cohort who also received intraoperative IV dexmedetomidine (KD + IV dexmedetomidine) had the longest overall PACU stays (median, 100 minutes), followed by patients in the same cohort (median, 62 minutes) and patients in the control cohort (median, 58 minutes), neither of which received IV dexmedetomidine (KD – IV dexmedetomidine and KM – IV dexmedetomidine, respectively). Patients in the control cohort who also received intraoperative IV dexmedetomidine (KM + IV dexmedetomidine) had the shortest PACU duration (median, 43 minutes; Table 3). There was an overall difference in PACU discharge times between the 4 groups (Kruskal Wallis p-value = .02), but when examining differences between the groups in a multiple comparisons analysis, only the groups with the shortest and longest PACU times (KM + IV dexmedetomidine and KD + IV dexmedetomidine, respectively) were statistically significantly different from one another (Dunn's p-value < .0125).

Table 3.

Length of Recovery*

Group	n	Median (IQR)
IM ketamine with midazolam + IV Dex	12	43 (24–77)†
IM ketamine with midazolam	19	58 (40–79)
IM ketamine with Dex	63	62 (40–88)
IM ketamine with Dex + IV Dex	11	100 (67–119)†

^*Dex, dexmedetomidine; IQR, interquartile range; IM, intramuscular; IV, intravenous. Subgroup analysis of time, in minutes, from OR departure to PACU discharge based on intraoperative use of IV dexmedetomidine. IV dexmedetomidine was only administered beyond the studied intraoperative period (after the first 60 minutes following intubation). The overall Kruskal Wallis test showed that the length of recovery was significantly different between the 4 groups (p = .02). Dunn's test for multiple comparisons revealed that the group with the shortest recovery time (IM ketamine with midazolam + IV dex group) was significantly different from the group with the longest recovery time (IM ketamine with dex + IV dex group). No other differences were significant.

^†Denotes statistical significance.

Unwanted Side Effects and Adverse Events

Three adverse events occurred in the KD group, with 2 of the 3 events being deemed unlikely related to IM dexmedetomidine. One patient with a history of reactive airway disease developed bronchospasm during induction, which resolved with albuterol administration. A second patient was agitated during the emergence and immediate postoperative period, requiring sedation in the PACU. One of the additional sedative drugs administered to that patient was dexmedetomidine. Both patients were discharged home on the day of surgery with no subsequent adverse effects.

The third patient experienced hypotension characterized by sustained low blood pressure (MAP of 50–55 mm Hg), requiring administration multiple vasopressors, including a bolus of epinephrine and an infusion of phenylephrine to achieve acceptable hemodynamics. The patient was discharged home on the same day after an uneventful recovery in the PACU. The hypotension noted in this case was attributed to the administration of 2 α₂ adrenergic agonists on the morning of surgery: the patient's regularly scheduled oral clonidine for behavior modification (0.2 mg administered 2 hours prior to anesthesia) followed by a preoperative IM KD admixture.

DISCUSSION

In this retrospective analysis, we investigated the efficacy and safety of IM ketamine with dexmedetomidine for sedation prior to dental rehabilitation under GA. The results of this analysis suggest that coadministration of IM KD may provide effective sedation to facilitate safe patient transfer to the OR and induction of GA. Patients receiving IM KD displayed no difference in either time from IM injection to OR entry or in the percentage of patients in whom GA was induced by the IV route. In addition, this study did not identify a significant difference between the 2 groups in the incidence of adverse events. The incidence of bradycardia and hypotension as inferred from the use of anticholinergic or vasopressor medications was not found to be significantly different between the study and control cohorts. Neither were there any local skin or muscular injuries identified resulting from the IM administration of the admixture of KD. Patients who received dexmedetomidine displayed a favorable hemodynamic profile throughout the induction and immediate intraoperative periods.

Ketamine anesthesia produces a distinctive “dissociative” state characterized by catatonia, amnesia, and analgesia.¹ Respiratory and cardiovascular function is preserved.¹ IM dosing regimens ranging from 4 to 10 mg/kg have been described.⁷ Due to 93% bioavailability⁸ and rapid absorption following IM administration, acceptable anesthetic conditions result within approximately 5 to 8 minutes.⁹ Although satisfactory outcomes using IM premedication in individual patients is most reliable when larger doses of ketamine are chosen, this must be weighed against the risks. Ketamine is associated with several well-described and dose-dependent adverse side effects, including hypertension, tachycardia, dysphoria, hallucinatory reactions, postoperative nausea and vomiting, and emergence agitation.¹ Intellectually disabled adult patients experiencing dysphoria, hallucinations, and/or emergence phenomena can be potentially injurious to themselves and clinical staff. In addition, this may be particularly distressing to caregivers who witness such reactions.

An admixture of IM KM, with a concomitant reduction in total ketamine dose, has been reported to achieve desired sedation levels and reduce postinjection dysphoria.¹⁰ Indeed, at this institution, it has been typical to coadminister IM midazolam (0.1 mg/kg) with ketamine (2–3 mg/kg), and thereby reduce the dose of ketamine. Most patients in the control cohort of our study received this IM combination. Due to the pharmacokinetics of IM midazolam, patients may still experience emergence reactions because midazolam can produce paradoxical excitation, especially in patients with ASD. These patients can remain physically combative and express anxiety and fear both verbally and nonverbally. In addition, midazolam has minimal effect on blood pressure and HR and thus, no ability to modulate the known hypertension, tachycardia, and poor hemodynamic control commonly seen during induction of GA following IM ketamine sedation.¹¹

Dexmedetomidine is known to be effective for pediatric procedural sedation, has been used as the sole sedative for procedures such as MRI,¹² and is also well-characterized as an adjunct to GA.¹³ Dexmedetomidine reduces the doses required of other general anesthetic agents,¹³ decreases postoperative opioid consumption,¹⁴ and decreases the incidence of emergence agitation.¹⁵ Levanen et al¹¹ showed that IM dexmedetomidine as a single agent, in comparison with IM midazolam, produced better HR and blood pressure control during intubation when used as a premedication prior to IV ketamine induction. In addition, IM dexmedetomidine more effectively reduced the incidence of ketamine-related side effects (ie, hallucinations) in comparison with IM midazolam. When admixed with IM ketamine, dexmedetomidine may reduce both postinjection dysphoria and emergence reactions.

The hemodynamic effects of dexmedetomidine depend on the route of administration. IV dexmedetomidine produces a well-characterized biphasic hemodynamic response with a rapid increase in MAP accompanied by a reflexive bradycardia.¹⁶ As the dexmedetomidine redistributes from plasma to other compartments, MAP quickly drops below baseline levels. HR rises from the initial nadir but takes several hours to return to baseline. The biphasic hemodynamic profile exhibited with IV dexmedetomidine is not observed with IM dexmedetomidine, which produces a more gradual reduction in both HR and MAP without the initial hypertension and reflexive bradycardia due to the movement/absorption into the systemic vasculature. In contrast to IM midazolam, which has minimal effect on MAP and HR, it is possible that IM dexmedetomidine may counteract the unwanted tachycardia and hypertension seen immediately following IM ketamine administration through dexmedetomidine's intrinsic sympatholytic effect.¹¹ Dexmedetomidine may also indirectly reduce the side effects of ketamine by achieving adequate sedation with a lower ketamine dose. Because patients in the control group of our study received IM midazolam, there was no comparative reduction in ketamine doses between study cohorts.

Tammam⁶ compared the use of IM KD combination for pediatric MRI sedation. This study randomly assigned 162 patients into 1 of 3 groups: patients receiving 3.0 mcg/kg dexmedetomidine, 4.0 mg/kg ketamine, and patients receiving an admixture of 1.5 mcg/kg dexmedetomidine and 2.0 mg/kg ketamine. Patients who received the IM admixture of dexmedetomidine and ketamine demonstrated an onset of sedation in 4.8 minutes, which approximated the onset of those receiving ketamine alone (4.6 minutes) but was much faster than those receiving dexmedetomidine alone (16.8 minutes). There were fewer sedation failures and less rescue midazolam administered in the admixture group. In addition, time to discharge was significantly faster in the admixture group (30 minutes vs 37 minutes for dexmedetomidine and 54 minutes for ketamine). Hemodynamic variables, such as HR and MAP, were more stable and remained closer to baseline values in the admixture group compared with those receiving dexmedetomidine (lower HR and MAP) or ketamine alone (higher HR and blood pressure).

Zor et al¹⁷ reported similar efficacy for an IM KD admixture in adult patients requiring burn dressing changes. In this study, 24 patients were randomly assigned to receive IM ketamine alone (2 mg/kg); IM tramadol (1 mg/kg), ketamine (2 mg/kg), and dexmedetomidine (1 mcg/kg); or IM tramadol (1 mg/kg), ketamine (2 mg/kg), and midazolam (0.05 mg/kg). The group receiving dexmedetomidine showed better analgesia, cardiovascular stability, and patient satisfaction in comparison with the group receiving midazolam. Both the Zor et al¹⁷ and Tammam⁶ studies utilized IM KD for procedural sedation. To our knowledge, there is no prior literature reporting the use of an IM KD admixture for sedation prior to induction of GA.

Our results suggest that a preoperative IM admixture of KD may be an effective and safe means of facilitating patient transfer to the OR, IV cannulation, and induction of GA. We selected time elapsed from IM administration to OR entry as a reasonable surrogate for IM efficacy. In order to enter the OR after administration of an IM sedative, the patient must be separated from their caregivers by the anesthesia staff, positioned safely on a hospital stretcher, and transferred to the OR. This transition can only efficiently occur following successful IM sedation. Patient separation was initiated once the anesthesia provider judged it both safe and feasible to do so. In our study, the difference in times from IM injection to OR entry between the study and control groups was not statistically significant. We also recorded the primary route of induction (IV or inhalational) as a secondary measure of the effectiveness of preoperative sedation with the following presumptions: (1) a more sedated patient will allow IV cannulation; (2) an anesthesiologist who can establish IV access will elect to do so; and (3) a patient with IV access will receive an IV induction. In our study, we found no significant difference between the study and control groups in the percentage of cases beginning with an IV induction, as opposed to an inhalational induction. Patients in the KD group displayed lower HR and blood pressure measurements throughout the induction and immediate intraoperative periods when compared with the KM group. The avoidance of hypertension and tachycardia during the induction of GA and endotracheal intubation may be more significant in older disabled patients or those with cardiovascular comorbidities.

It should be noted that the PACU discharge times in our study do not necessarily reflect when the patient was ready for discharge. Dexmedetomidine produces a state of sedation remarkably close to natural sleep, in which a patient who appears sedated can be easily aroused. PACU staff may be reluctant to disrupt a quiet or sleeping patient, particularly one who was combative prior to anesthesia. Furthermore, in view of the pharmacokinetics of IM dexmedetomidine, it is unlikely that the clinical effect of residual plasma drug concentrations would delay discharge 4 hours or more after drug administration, the average length of time in our study from IM administration to PACU discharge. Nevertheless, it is interesting that although IV dexmedetomidine appears to hasten recovery in the KM group, it appears to prolong recovery in the KD group. The impact of IV dexmedetomidine on recovery time and recovery behavior warrants further investigation in future studies.

There is some theoretical benefit to the use of an α₂ adrenergic agonist in patients with ASD, who comprised half of the patients in this study. Many patients with ASD and attention-deficit/hyperactivity disorder benefit from regular use of other centrally acting α₂ adrenergic agonists, such as guanfacine¹⁸ or clonidine,¹⁹ for improved control of sleep and behavior disorders. The behavioral impact of dexmedetomidine in patients with ASD may be most notable during emergence from GA, where dexmedetomidine is reported to reduce emergence agitation.²⁰

Due to its retrospective nature, our study had several limitations. Due to a lack of randomization, those who received dexmedetomidine may have differed from those who did not in ways not captured by the demographic analysis. The anesthesia EMR used in this study did not reliably capture patient behavior before the IM injection or the level of restraint required to administer the IM injection. In addition, an accurate assessment of the unwanted side effects of ketamine or patient behavior in the PACU (including emergence delirium and postoperative nausea and vomiting) was not available. Similarly, it was not possible to control for the drugs and doses used for maintenance anesthesia, analgesics/antiemetics given, or the contents and timing of local anesthetics.

It is likely that these findings are generalizable to dental patients receiving treatment in other hospital ORs and ambulatory facilities. The patient population of this study is reflective of patients who require GA for dental rehabilitation. Approximately 50% of patients had a diagnosis of ASD and 47% were ASA III, highlighting the behavioral and medical challenges often seen when providing dental care in a hospital OR. This patient population included a balanced mix of children and young adults. Only 2 supervising dentist anesthesiologists treated the patients in this study, improving the consistency of patient management. Another strength of our study is that all data were abstracted from the EMR, not handwritten records.

CONCLUSION

In summary, our data suggest that this novel application of IM ketamine with dexmedetomidine may be an effective and safe means to facilitate the induction of GA when compared with IM ketamine with midazolam. Additional randomized clinical trials should be performed to adequately characterize the effects of this admixture and to assess optimal dosing protocols.