Comparison of midazolam and dexmedetomidine on postoperative delirium in older patients undergoing total knee replacement under spinal anesthesia

Abstract

Background

Postoperative delirium (POD) is a frequent and serious complication in older patients undergoing orthopedic surgery. This study aimed to compare the effects of midazolam and dexmedetomidine on the incidence of POD in elderly patients undergoing total knee arthroplasty under spinal anesthesia.

Methods

In this prospective, double-blind, randomized, parallel group comparative trial, a total of 175 patients aged 65 years or older were allocated to receive either midazolam or dexmedetomidine for intraoperative sedation. The primary outcome was the incidence of POD, evaluated using the confusion assessment method for five postoperative days. Secondary outcomes included the quality of sedation, incidence of sedation failure, postoperative analgesic profiles, and complication rates.

Results

The incidence of POD was significantly lower in the dexmedetomidine group compared to the midazolam group (13.9% vs. 27.4%, P = 0.045). In addition, dexmedetomidine was associated with superior sedation quality and a significantly lower rate of sedation failure (P = 0.007 and P = 0.018, respectively). Postoperative pain intensity and complication rates were comparable between the two groups.

Conclusion

Dexmedetomidine significantly reduced the incidence of POD and improved sedation quality compared to midazolam. These findings suggest that dexmedetomidine may be a preferable sedative for elderly patients undergoing total knee arthroplasty under spinal anesthesia.

Trial registration

This study was registered with the WHO International Clinical Trials Registry Platform (Trial number: KCT0006587, Registration date: 16th September 2021).

Background

Postoperative delirium (POD) is a serious complication in older patients that can occur after major orthopedic surgery because it is associated with increased morbidity and mortality, prolonged hospital and intensive care unit stays, and long-term cognitive decline (Jin et al. 2020; Inouye et al. 2014). The reported prevalence of POD ranges from 5 to 50% and varies considerably according to the patients’ characteristics and the type of surgery (Zywiel et al. 2015; Li et al. 2022). Nevertheless, advanced age (≥65 years) and the existence of preoperative cognitive impairment are constantly recognized as significant predisposing factors for the development of POD (Yang et al. 2017; Rong et al. 2021). Major orthopedic surgeries, such as total knee or hip arthroplasty, are recognized as high-risk surgical categories for the occurrence of POD comparable to cardiac and major thoracic and abdominal surgeries (Zywiel et al. 2015; Urban et al. 2020).

The multifactorial etiology of POD makes it challenging to predict and prevent. This complexity highlights the necessity of identifying key risk factors and developing targeted, patient-specific strategies. Previous studies have investigated various preventive strategies, including reducing modifiable perioperative risks and employing both pharmacologic and non-pharmacologic interventions (Jin et al. 2020; Lee et al. 2021). Dexmedetomidine, a highly selective α2-adrenergic agonist, has gained attention for its potential to prevent POD. Recent studies have reported that dexmedetomidine administration as a sedative in the intensive care unit and during general anesthesia reduces the incidence of POD (Pereira et al. 2020; Duan et al. 2018; Zeng et al. 2019; Norden et al. 2021). While previous studies have demonstrated that dexmedetomidine sedation during spinal anesthesia can reduce the incidence of POD in older patients undergoing lower-extremity orthopedic surgery (Shin et al. 2023; Lee et al. 2024), direct comparison between dexmedetomidine and midazolam remains limited. Midazolam, a widely used short-acting benzodiazepine due to its excellent anxiolytic effect, has been associated with a potential drawback of increased incidence of POD (Weinstein et al. 2018; Memtsoudis et al. 2019).

The current study aimed to compare the effect of dexmedetomidine and midazolam on the occurrence of POD when used as sedatives during total knee arthroplasty under spinal anesthesia in older patients. The authors hypothesized that dexmedetomidine would be superior to midazolam in preventing POD. Additionally, the study evaluated the potential benefits of dexmedetomidine, including enhancing the quality of sedation and the postoperative analgesic profile, compared to midazolam.

Methods

Study setting and participants

This prospective, double-blinded, randomized, parallel-design comparative trial was approved by the Institutional Review Board of Jeonbuk National University Hospital (IRB number: CUH 2021-05-049, Approval date: 7th September 2021) and registered with the WHO International Clinical Trials Registry Platform (Trial number: KCT0006587, Registration date: 16th September 2021). This study was conducted in the orthopedic operating room setting of the single university hospital from October 2021 to September 2024. Informed consent was obtained from all participants before enrollment.

Older patients, aged 65 years or above, who were scheduled for elective total knee arthroplasty at single institution, were screened for the current study. This study enrolled 175 participants with American Society of Anesthesiologists physical status (ASA/PS) I–III who agreed to receive spinal anesthesia combined with intraoperative sedation during the surgery. The exclusion criteria were as follows: (1) patient refusal; (2) severe bradycardia (heart rate ≤ 45 beats per minute), sick sinus syndrome, or second-degree or higher atrioventricular block; (3) a history of neuropsychiatric diseases including Parkinson’s disease, schizophrenia, or dementia; (4) patients with a recent cerebrovascular accident within the last 6 months; (5) patients with a preoperative mini-mental status examination (MMSE) score of 23 or lower, indicating cognitive impairment; (6) severe hepatorenal impairment; (7) patients who had a problem of communication such as visual, language, or hearing impairment.

The baseline cognitive function was assessed using the Korean version of MMSE (K-MMSE) on the day before surgery for all participants. The K-MMSE is a widely used screening tool for diagnosing cognitive impairment and dementia. The test consists of six subdomains: orientation, registration, attention and calculation, memory recall, language, and visual-spatial skills. It is scored on a 30-point scale, with higher scores indicating better cognitive function. A score of 23 or lower is generally considered indicative of cognitive impairment (Kang et al. 2016).

Group allocation, randomization, drug preparation, and blinding

This study employed a block-randomized, parallel-design trial. Participants were randomly assigned to the midazolam or dexmedetomidine groups in a 1:1 ratio using a web-based random number generator with a permuted block size of 4, ensuring balanced group allocation. An anesthesiologist, who was not involved in anesthetic management and outcome assessment, performed the randomization. The allocation results were delivered in sealed, concealed envelopes to a senior anesthetic nurse responsible for preparing the study drug.

The study drugs were prepared as follows: midazolam (10 mg) was diluted with normal saline to a total volume of 50 ml, resulting in a concentration of 200 µg/ml for the midazolam group. Dexmedetomidine (Precedex®, Hospira, USA, 200 µg/1 ml) was diluted with 49 ml of normal saline to achieve a final concentration of 4 µg/ml in a total volume of 50 ml for the dexmedetomidine group. Both study drugs, when diluted in normal saline, were transparent, visually indistinguishable, and equal in volume. To maintain blinding, both the anesthetic care provider and the outcome assessor were unaware of group allocation and study drug identity until the end of study. Patients also remained blinded to group allocation throughout the procedure, as the sedative agents were indistinguishable in appearance, volume, and subjective effect under spinal anesthesia. At the end of sedation, we recorded the total volume of administered sedatives.

Spinal anesthesia and intraoperative sedation

No premedication was administered to any participant. Upon arrival in the operating room, standard anesthetic monitoring including non-invasive blood pressure, electrocardiogram, pulse oximetry, and bispectral index (BIS) was applied. Patients were then positioned in the lateral decubitus position with the operating leg down. The interspinous space at the L4/5 vertebral level was identified using the Tuffier’s line palpation method. A spinal needle was then inserted using a midline approach at the L3/4 or L4/5 interspace. Upon confirming positive cerebrospinal fluid flow, 10–15 mg of 0.5% hyperbaric bupivacaine (Marcaine heavy injection®, AstraZeneca AB, Seoul, Korea) was administered intrathecally. The success of spinal anesthesia was determined through evaluating pin-prick sensation, and adjustments were made to achieve the desired final level of anesthesia at T8–T10. Prior to initiating patients’ sedation, hemodynamic parameters were closely monitored for a minimum of 15 min. If the systolic blood pressure (SBP) fell below 90 mmHg or decreased by 30% or more from the preoperative baseline, 5–10 mg of ephedrine was given intravenously. Bradycardia, defined as a heart rate of less than 45 beats per minute (bpm), was treated with 0.5 mg of atropine.

After ensuring hemodynamic stability, the study drug was delivered using an infusion pump (Agilia®, Fresenius Vial, France) for the patients’ sedation. Initially, a loading dose of 0.175 ml/kg (equivalent to 35 µg/kg of midazolam or 0.7 µg/kg of dexmedetomidine, based on equipotent doses reported in previous studies (Kang et al. 2023; Eren et al. 2011)) was administered over a 10-min period. Subsequently, a maintenance infusion of 0.05–0.25 ml/kg/h (equivalent to 10–50 µg/kg/h of midazolam or 0.2–1.0 µg/kg/h of dexmedetomidine) was followed until the skin closure. During the surgical procedure, the infusion rate was adjusted at 3-min intervals to achieve moderate sedation, as indicated by a Ramsay sedation score (RSS) 4–5. Participants who did not reach the appropriate depth of sedation were considered as the sedation failure and were dropped from the study. An experienced anesthetic care provider managed the drug infusion and assessed the quality of sedation using a 4-point Likert scale, classifying it as excellent, good, fair, or poor. At the skin closure, sedative infusion was discontinued and the elapsed time to eye opening on verbal command was recorded. In the event of hypotension with a SBP below 90 mmHg or bradycardia with a heart rate less than 45 bpm, it was promptly treated using intravenous ephedrine (5–10 mg) or atropine (0.5 mg), respectively.

Management of postoperative pain

Multimodal analgesia for managing acute postoperative pain were as follows: (1) at the conclusion of the knee arthroplasty, a total of 30 ml of 0.5% ropivacaine with 1:200,000 epinephrine and 8 mg of morphine was administered directly into the joint space (intra-articular injection) after skin closure; (2) intravenous patient-controlled analgesia (PCA) contained fentanyl 600 μg, nefopam 120 mg, and ketorolac 120 mg, prepared in a total volume of 60 ml. PCA was initiated at skin closure following a 4 ml loading dose; (3) 10 ml of 1% lidocaine was administered via the femoral catheter either at the patient’s demand or during physical stretching. The femoral catheter was placed in the operating leg under ultrasound guidance the day before surgery.

Assessment of patient outcome

The primary outcome of the current study was the incidence of POD. The occurrence of POD was evaluated using confusion assessment method (CAM). The evaluations were conducted twice daily, at 8:00 AM and 6:00 PM, throughout the five postoperative days. The assessment findings were cross-referenced with current medical records and involved interview with caregivers and nursing staff.

Secondary outcomes included the severity of POD, the type of POD, the antipsychotic usage, intraoperative hemodynamic variables, analgesic profiles, and postoperative complications. The symptom severity of delirium was evaluated using the Korean version of the memorial delirium assessment scale (K-MDAS). The MDAS is a well-validated and reliable tool for both assessing delirium symptom severity and diagnosing delirium (Kang et al. 2019). In this study, the K-MDAS was administered to patients who yielded positive results on the CAM. For patients unable to complete the K-MDAS assessment due to severe psychomotor agitation or uncooperativeness, the Richmond agitation sedation scale (RASS) was used as alternative. The K-MDAS grading ranges from 0 to 30, with scores of 9–13 classified as mild delirium, 14–18 as moderate delirium, and 19 or higher as severe delirium (Klankluang et al. 2020). The RASS grading ranges from −5 to +4: a score of +1 or −2 was classified as mild, +2 or −3 as moderate, and scores of +3 or higher or −4 or lower as severe delirium (Han et al. 2015). Regarding the type of POD, patients can be categorized as hypoactive with a RASS score of −2 or lower, hyperactive with a RASS score of +1 or higher, or mixed if they exhibit both positive and negative RASS scores. The anesthesiologist received individual training from a board-certified psychiatrist before conducting delirium assessments.

Intraoperative vital signs, including mean arterial pressures and heart rates, and BIS value were recorded at 5-min interval from the initiation of anesthesia until the end of surgery. The analgesic profile included assessments of pain intensity (by numeric rating scale; 0–10) at rest and during movement, as well as cumulative analgesic consumption, which were recorded over the five postoperative days. The occurrence of postoperative complications, including cardiovascular, cerebrovascular, pulmonary, hepatic, and renal complications, were also evaluated. Additionally, the elapsed time to first micturition and the incidence of urinary difficulties requiring catheterization were recorded.

Sample size estimation and statistical analysis

The primary endpoint of the study was the incidence of POD between the midazolam and dexmedetomidine groups. The sample size calculation was based on the assumption of a 15% difference in the incidence of POD between the two groups, which was derived from our preliminary results. Using IBM SPSS statistics for Windows, version 26 (IBM Corp., Armonk, NY, USA), the required sample size was estimated to be 76 patients per group, with a significant level of 0.05 (α = 0.05) and a power of 80% (β = 0.20). To account for a 15% dropout rate, the total sample size was increased to 175 participants.

The incidence of POD, as the primary outcome, was compared between the groups using the chi-square test. Continuous demographic data and analgesic profile were analyzed using either two-tailed independent-samples t-test or Mann-Whitney rank-sum test, depending on the results of normality testing. Categorical variables, including the severity and the type of POD, were analyzed using the Mann-Whitney rank-sum test. Vital signs and BIS values obtained within the study were analyzed using one-way repeated measures analysis of variance (RM ANOVA).

All descriptive data are expressed as the numbers, numbers (percentages), mean ± standard deviations, or median (interquartile range). A P value < 0.05 was considered statistically significant.

Results

The subject flow diagram is presented in Fig. 1. Of the 184 participants assessed for eligibility, 6 individuals (3.3%) were excluded from the study due to a preoperative K-MMSE score of ≤23, which is indicative of cognitive impairment. A total of 175 patients were enrolled in the present study. Of these, 73 patients in the midazolam group and 72 patients in the dexmedetomidine group completed the five postoperative day follow-up and contributed data for statistical analysis. Post-randomization exclusions (n = 30) occurred due to failed regional anesthesia, sedation failure, or patient refusal of sedation or PCA, resulting in these participants not receiving the allocated intervention or providing outcome data. Given this, a per-protocol analysis was conducted. Patient demographics and clinical characteristics were comparable between the midazolam and the dexmedetomidine groups (Table 1).

Fig. 1.

Subject flow diagram

Table 1.

Patients’ demographics and clinical characteristics

	Midazolam group (n = 73)	Dexmedetomidine group (n = 72)	P value
Age (years)	74.6 ± 5.7	73.9 ± 5.0	0.423
Female (%)	58 (79.5%)	49 (68.1%)	0.119
BMI (kg/m2)	27.0 (24.8–29.5)	25.8 (23.8–28.1)	0.061
ASA/PS (I/II/III)	0/40/33	1/44/27	0.305
Preoperative MMSE	26.0 (25.0–28.0)	26.0 (24.0–28.0)	0.798
Education level (year)	6.0 (6.0–9.0)	6.0 (6.0–9.0)	0.685
Underlying disease (n)
Hypertension	48	42	0.683
Diabetes mellitus	15	14	0.868
Coronary arterial disease	10	6	0.303
Cerebrovascular disease	10	8	0.637
Chronic kidney disease	4	2	0.414
Respiratory disease	6	1	0.055
Others	10	10	0.973
Anesthesia time (min)	119.4 ± 30.7	111.2 ± 21.0	0.064
Operation time (min)	83.8 ± 30.3	76.1 ± 18.3	0.069
Total volume of administered iv fluid (ml)	397.3 ± 210.8	366.9 ± 146.3	0.316

Values are mean ± SD, number (percentage), number, or median (interquartile range)

BMI body mass index, ASA/PS American Society of Anesthesiologists Physical Status, MMSE mini-mental state examination

Table 2 summarizes the POD profiles, including the primary endpoint of the study—the incidence of POD between the two groups. The incidence of POD was significantly lower in dexmedetomidine group compared to the midazolam group (27% vs. 13%, 95% confidence interval; 0.01–0.27, P = 0.045). However, no significant differences were observed between the groups in the severity or type of POD, as well as in the use of antipsychotic medications, with the majority of cases in both groups classified as mild.

Table 2.

Postoperative delirium

	Midazolam group (n = 73)	Dexmedetomidine group (n = 72)	P value
Incidence of POD [n (%)]	20 (27%)	10 (13%)	0.045*
Severity of POD [n (%)]			0.132
Mild	13 (17%)	6 (8%)
Moderate	6 (8%)	2 (3%)
Severe	1 (1%)	2 (3%)
Type of POD [n (%)]			0.145
Hypoactive	6 (8%)	1 (1%)
Hyperactive	12 (16%)	8 (11%)
Mixed	2 (3%)	1 (1%)
Antipsychotic usage [n (%)]	5 (7%)	5 (7%)	0.982

Values are number (percentage)

POD postoperative delirium

^*P < 0.05 by chi-square test

Sedation profiles of the midazolam and the dexmedetomidine groups are presented in Table 3. There was a statistically significant difference in the quality of sedation between the midazolam and dexmedetomidine groups (P = 0.007, by rank-sum test). Furthermore, the incidence of sedation failure was significantly higher in the midazolam group compared to the dexmedetomidine group (11.4% vs. 2.3%, P = 0.018 by chi-square test). The depth of sedation, assessed using the RSS, is presented in Fig. 2A. Sedation levels were comparable between the dexmedetomidine and midazolam groups at all recorded time point. However, BIS values were significantly lower in the dexmedetomidine group at 20 and 40 min following skin incision and at the time of skin closure (P < 0.001) (Fig. 2B).

Table 3.

Sedation profile

	Midazolam group	Dexmedetomidine group	P value
Total volume of administered sedatives (ml)a	19.8 ± 7.2	17.6 ± 7.3	0.069
Quality of sedation (excellent/good/fair/poor)	16/46/8/3	31/35/4/2	0.007†
Incidence of sedation failure [n (%)]	10/88 (11.4%)	2/87 (2.3%)	0.018‡
The elapsed time to obey command (min)	5.0 (2.5–7.0)	4.0 (1.0–7.0)	0.338

Values are mean ± SD, number, or number (percentages)

^†P < 0.05 by Mann-Whitney rank-sum test

^‡P < 0.05 by chi-square test

^aThe sedatives (study drugs) were prepared at concentrations of midazolam 10 mg in 50 ml (200 μg/ml) and dexmedetomidine 200 μg in 50 ml (4 μg/ml). The total volume of administered sedatives was compared between the two groups, reflecting these standardized concentrations

Fig. 2.

The depth of sedation between the midazolam and the dexmedetomidine groups. A Ramsay sedation scale. B Bispectral index

Hemodynamic parameters including mean arterial pressure, heart rates, and the usage of rescue vasoactive drugs are shown in Table 4. Mean arterial pressures were significantly lower in the dexmedetomidine group compared to the midazolam group at baseline, 60 min after skin incision, and at the time of skin closure (P = 0.007, P < 0.001, and P = 0.011, respectively). Heart rates were also significantly lower in the dexmedetomidine group than in the midazolam group at most intraoperative time points (P < 0.05). Ephedrine was administered to 6 participants in the midazolam group and 4 participants in the dexmedetomidine group, while atropine was required in 4 participants in the dexmedetomidine group during the surgery.

Table 4.

Hemodynamic parameters

	MAP (mmHg)		HR (bpm)		Vasoactive usage (ephedrine/atropine)
	Midazolam group (n = 73)	Dexmedetomidine group (n = 72)	Midazolam group (n = 73)	Dexmedetomidine group (n = 72)	Midazolam group (n = 73)	Dexmedetomidine group (n = 72)
Baseline	113.0 ± 15.8	106.1 ± 15.0*	75.7 ± 12.5	75.3 ± 14.3	1/0	2/1
SI	86.7 ± 13.8	90.8 ± 15.5	69.2 ± 10.9	64.4 ± 15.9*	4/0	2/2
SI-20	85.9 ± 12.9	83.9 ± 12.9	67.6 ± 11.8	62.8 ± 12.2*	2/0	1/1
SI-40	87.0 ± 11.9	83.6 ± 11.7	67.6 ± 11.5	63.2 ± 12.1*	1/0	0/0
SI-60	92.0 ± 12.2	80.8 ± 11.9*	67.9 ± 9.8	62.8 ± 12.0*	0/0	0/0
SI-80	84.9 ± 10.2	86.7 ± 18.0	73.0 ± 12.4	60.0 ± 12.3*	0/0	0/0
SI-100	88.6 ± 12.1	84.0 ± 13.1	75.1 ± 11.7	65.8 ± 16.6*	0/0	0/0
SC	90.2 ± 14.4	84.2 ± 13.3*	72.0 ± 11.8	64.4 ± 12.8*	0/0	0/0

Values are mean ± SD or number

MAP mean arterial pressure, HR heart rate, SI skin incision, SC skin closure

^*P < 0.05 based on inter-group analysis of repeated measures analysis of variance

The occurrence of postoperative complications and analgesic profiles including pain intensity (NRS scores) and cumulative analgesic consumption are presented in Tables 5 and 6, respectively. All of these outcomes were similar between the two groups.

Table 5.

Postoperative complications

	Midazolam group (n = 73)	Dexmedetomidine group (n = 72)	P value
Postoperative complications [n (%)]	18 (24.7%)	22 (30.6%)	0.427
Cardio/cerebrovascular	2	2
Hematologic	3	5
Pulmonary	2	5
Hepatic	6	2
Renal	5	8
Elapsed time to 1st micturition	5.0 (3.0–7.0)	5.0 (4.0–8.0)	0.200
Incidence of urinary difficulties [n (%)]	14 (19.2%)	16 (22.2%)	0.651

Values are number (percentage), number, or median (interquartile range)

Table 6.

Postoperative analgesic profile including pain intensity and cumulative analgesic consumption

	Midazolam group (n = 73)	Dexmedetomidine group (n = 72)	P value
Pain intensity (NRS) at rest
Postoperative 6 h	2.0 (0.0–4.0)	1.0 (0.0–3.3)	0.114
Postoperative day 1	3.0 (2.0–4.0)	3.0 (2.0–4.0)	0.996
Postoperative day 2	3.0 (2.0–4.0)	3.0 (2.0–4.0)	0.996
Postoperative day 3	2.0 (2.0–3.0)	2.0 (2.0–3.0)	0.872
Postoperative day 4	2.0 (1.0–3.0)	2.0 (1.0–2.0)	0.369
Cumulative analgesic consumption (ml)
Postoperative 6 h	7.2 (5.9–8.7)	7.0 (5.8–8.1)	0.382
Postoperative day 1	15.4 (12.3–16.7)	15.1 (12.3–16.7)	0.839
Postoperative day 2	29.2 (24.5–33.2)	28.1 (25.2–32.6)	0.614
Postoperative day 3	33.7 (30.9–38.5)	34.0 (30.6–38.0)	0.898
Postoperative day 4	47.3 (43.7–54.0)	45.2 (42.8–51.2)	0.594

Values are median (interquartile range)

NRS numeric rating scale

Discussion

This study provides important insights into the comparative effects of midazolam and dexmedetomidine, two commonly used sedatives, on the incidence of POD in older patients undergoing total knee arthroplasty under spinal anesthesia. Notably, it also offers a comprehensive evaluation of POD characteristics, including the severity and the subtype of POD. The authors hypothesized that dexmedetomidine would be more effective than midazolam in preventing POD. Consistent with our hypothesis, the incidence of POD was significantly lower in patients sedated with dexmedetomidine compared to those receiving midazolam (27.4% vs. 13.9%, P = 0.045), although no significant differences were observed in POD severity or subtype.

The study results also highlighted another advantage of dexmedetomidine: superior anesthesiologist-assessed quality of sedation compared to midazolam. Achieving and maintaining an appropriate level of sedation during spinal anesthesia can be clinically challenging. Excessive sedation may result in respiratory depression or hemodynamic instability, whereas insufficient sedation may leave patients feeling anxious and uncomfortable. In this study, anesthesiologists rated sedation quality significantly higher in the dexmedetomidine group compared to the midazolam group. Furthermore, the incidence of sedation failure was also significantly lower in the dexmedetomidine group compared to the midazolam group (P = 0.018 by chi-square test). The 11.4% (10/88) of patients in the midazolam group who failed to achieve adequate sedation despite adjustment up to the maximum maintenance dose is likely attributable to the pharmacodynamic properties of midazolam, which is known to exhibit significant inter-individual variability in its sedative effects (Park et al. 2019). In contrast, dexmedetomidine appears to provide a more consistent and predictable sedative effect.

Dexmedetomidine is a preferred sedative due to its unique properties, including cooperative sedation, minimal respiratory depression, and potential analgesic properties (Lee 2019). Additionally, it exhibits the potential neuroprotective effect in several experimental studies by reducing neuroinflammation, suppression of oxidative stress, inhibition of apoptosis, and maintenance of blood-brain barrier integrity (Hu et al. 2022; Chen et al. 2022; Guo et al. 2022; Wang et al. 2019; Liaquat et al. 2021). These properties may favor dexmedetomidine as a pharmacologic prophylaxis for POD. Several studies have reported that perioperative administration of dexmedetomidine during general anesthesia effectively prevented POD after both cardiac and non-cardiac surgeries (Duan et al. 2018; Zeng et al. 2019; Norden et al. 2021). However, most studies have focused on its effects in the context of general anesthesia. Although dexmedetomidine is considerably more often employed for sedation during regional anesthesia than as an adjunctive agent in general anesthesia, the effect of dexmedetomidine administered for sedation during spinal anesthesia on the incidence of POD has been understudied.

Interestingly, although the depth of sedation of both group was adjusted to achieve moderate sedation, as defined by a RSS 4–5, BIS values were significantly lower in the dexmedetomidine group at specific intraoperative time points compared to the midazolam group. This discrepancy may be partially explained by the different pharmacologic mechanisms of the two agents. Dexmedetomidine, which acts on α2-adrenergic receptors in the locus coeruleus, is known to induce a natural sleep-like state, whereas midazolam exert its sedative effects by potentiating GABAergic inhibition in the cerebral cortex (Haenggi et al. 2009). The different sedation mechanisms might account for the variations in processed EEG parameters, such as BIS values, at the similar sedation level (Xi et al. 2018).

The overall incidence of POD in the current study was 20.7% even after spinal anesthesia. Regional anesthesia is often recommended and believed by some physicians to be a preferable anesthetic choice for reducing the incidence of POD or postoperative cognitive dysfunction compared to general anesthesia (Li et al. 2025). However, in this study, the incidence of POD in patients undergoing total knee arthroplasty under spinal anesthesia was similar to that reported in studies using general anesthesia. This suggests that the choice of anesthetic technique may not influence the incidence of POD, and that high-risk patients for POD remain susceptible even with spinal anesthesia, thus necessitating the consideration of effective preventative strategies.

Meanwhile, the postoperative analgesic profiles were similar between the midazolam and dexmedetomidine groups. Previous studies have suggested that dexmedetomidine may enhance the quality of regional anesthesia and prolong postoperative analgesia (Hong et al. 2019; Abdallah et al. 2016; Hussain et al. 2017). The discrepancy between the current study findings and previous results may be partly caused by differences in institutional settings for postoperative pain management. In this study, postoperative pain was effectively managed to ≤3 at rest in all patients in both groups using multimodal analgesia, including patient-controlled analgesia, femoral catheter local anesthetic administration, and intra-articular morphine injection.

There are limitations of the study. First, the study population was predominantly female in both groups (79.5% and 68.1%), which may limit the generalizability of the findings to a broader population, particularly male patients. Additionally, the study was conducted in the orthopedic theater of a single institution, and variations in clinical practice and institutional perioperative protocols could influence the results. Second, there is a lack of objective diagnostic measures for POD. Although the methods such as CAM-IV are widely accepted for diagnosing delirium worldwide, the accuracy of the diagnosis can still depend on the outcome assessors’ evaluations. Third, the short follow-up period of only five postoperative days may not capture all instances of POD, particularly if symptoms emerged later. Additionally, long-term outcomes for each group were not evaluated, which could offer valuable insights into the persistence or recurrence of POD and its association with postoperative cognitive dysfunction. Finally, our study had a slight reduction in sample size from the initial calculation due to participant exclusions after randomization. However, the final sample size remained sufficient to preserve statistical power and support the validity of our findings.

Conclusions

In conclusion, dexmedetomidine used for sedation in older patients undergoing total knee arthroplasty under spinal anesthesia reduced the incidence of POD compared to midazolam. Additionally, dexmedetomidine demonstrated superior sedation quality and a lower incidence of sedation failure compared to midazolam. Therefore, we suggest that dexmedetomidine is a preferable sedative to midazolam during spinal anesthesia in older patients undergoing total knee arthroplasty.

Abbreviations

POD

Postoperative delirium

ASA PS

American Society of Anesthesiologist Physical Status

MMSE

Mini-mental status examination

BIS

Bispectral index

SBP

Systolic blood pressure

BPM

Beats per minutes

RSS

Ramsay sedation scale

PCA

Patient-controlled analgesia

CAM

Confusion assessment method

MDAS

Memorial delirium assessment scale

RASS

Richmond agitation sedation scale

RM ANOVA

Repeated measures analysis of variance

Authors’ contributions

A.D. helped perform the research and wrote the original draft. J.J. helped collect data. S.K. helped design the research and edit and revise the manuscript. All authors reviewed the final manuscript.

Funding

This study was supported by fund from the Biomedical Research Institute of Jeonbuk National University Hospital, Jeonju, Republic of Korea.

Data availability

The data analyzed during the current study are available from the corresponding author on reasonable request.

Declarations

Ethics approval and consent to participate

The study was performed in accordance with the Good Clinical Practice regulations and guidelines and in conformity with the ethical principles of the Declaration of Helsinki. Institutional Review Board of Jeonbuk National University Hospital approved the study (IRB number: CUH 2021-05-049) and waived the requirement to obtain informed consent based on the Good Clinical Practice regulations and guidelines.

Consent for publication

Not applicable.

Competing interests

The authors declare no competing interests.