This study adhered to the principles outlined in the "Declaration of Helsinki" by the World Medical Congress and received approval from the Chinese Ethics Committee of Registering Clinical Trials under the ethics number ChiECRCT20210592. Additionally, it obtained approval and was documented with the Scientific Research Management of the First Hospital of the University of Science and Technology of China. The study was registered with the Chinese Clinical Trials Registry Centre at www.chictr.org.cn, with the registration number ChiCTR2100043968 (06/03/2021). All participants provided written informed consent before participating in the study.

The inclusion criteria were as follows: (1) patients who were aged between 60 and 75 years; (2) those classified as American Society of Anaesthesiologists grade III–IV; (3) those scheduled for cardiac surgery, such as off-pump coronary artery bypass grafting and valve or aortic root surgery, from 7 January 2022 to 30 November 2022; and (4) those with sleep disorders (defined as having a Pittsburgh Sleep Quality Index [PSQI] score of ≥8). The exclusion criteria were as follows: (1) patients who had a baseline Mini-mental Status Examination (MMSE) score of ≤21 points; (2) those who declined or were unable to participate in the study, communicate, or speak; (3) those with hearing disorders; (4) those with sleep apnoea syndrome; and (5) those who had an ejection fraction of ≤50% after admission. Criteria for withdrawal and termination of the study included patients who (1) had a preoperative hospital stay of <6 days; (2) experienced third-degree atrioventricular block; (3) were allergic to DEX; (4) were lost to follow-up; and (5) had data analysis failure.

The flowchart representing the technical steps of the clinical research involving the patients is presented in Fig. 1.

The technical flowchart of the clinical research on patients. EF ejection fraction, DEX dexmedetomidine

The patients were randomly assigned in a 1:1 ratio using computer-generated numbers to form Groups A and B. The group allocations were sealed within sequentially numbered opaque envelopes, and these envelopes remained sealed until the patients provided informed consent. The patient, outcome assessor, and anaesthesiologist were blinded to the group assignment.

Patients received either DEX or saline nasal drip before bedtime for 6–8 days before the surgery, with a minimum of 1 day interval between each of the two or three interventions. The administration of the intranasal intervention was performed by either the sole anaesthetist or a clinically registered sedation doctor, directly supervised by the anaesthesiologist in the ward. The syringe used for administration comprised either 2 mL of normal saline or 0.3 µg/kg of DEX. Each syringe was used twice (1 mL for each nostril, administered 5 min apart). The patients were observed for 60 minutes, and if their Ramsay score was 5 points> Ramsay score ≥3 points (Kaur et al. 2016) and the bispectral index (BIS) was <70 (Kang et al. 2019) (indicating a sedation level comparable to natural sleep), the intervention was concluded. If not, an additional treatment dose was administered, and the patient was observed for another 60 min. No further treatment was administered after the third intervention, irrespective of the Ramsay score or the BIS. In terms of intranasal DEX dosage in adults, Yuen et al. (2007) reported DEX at doses of 1.0 and 1.5 µg/kg produced significant sedative effects compared to a placebo, causing reductions in blood pressure and heart rate (HR, p<0.05). Barends et al. (2020) used 0.5–2.0 μg/kg DEX nasal drips to achieve sedation. Even with the lowest dose (0.5 µg/kg DEX), the mean arterial pressure decreased by 30% for 5 min in 10% of patients. Our aim was to induce a state resembling “natural sleep” rather than general anaesthesia or deep sedation. The dose was titrated to improve the patient’s sleep while minimising the impact on circulation and avoiding respiratory depression. Standard anaesthesia procedures were implemented for all patients upon entering the operating room to mitigate the impact of the anaesthesia protocol on the study outcomes. Both groups received conventional treatment for sleep disorders, including the use of earplugs and maintaining a regular wake-sleep rhythm.

Upon the patient’s entry into the operating room for cardiac surgery, monitoring was initiated for the HR, invasive arterial pressure, central venous pressure, electrocardiogram, pulse oxygen saturation, and BIS. Anaesthesia induction involved administering 0.2–0.3 mg/kg etomidate, 0.5–0.9 mg/kg rocuronium, and 1.0–1.5 μg/kg sufentanil to all patients. Mask-assisted breathing ventilation was employed for 2 min, ensuring adequate muscle relaxation. The insertion of a tracheal tube with an appropriate inner diameter was performed using a visual laryngoscope (Insight iS2 Shenzhen Sultan Said Khan Medical Technology Co., Ltd.) when the BIS was maintained between 40 and 60. A fibreoptic bronchoscope was used to position the tube tip 2–3 cm distal from the tracheal carina. Mechanical ventilation parameters were set as follows: tidal volume of 6–10 mL/kg, an adjustable breathing rate to maintain partial pressure of end-tidal carbon dioxide of 35–45 mmHg, and an inhalation:exhalation ratio of 1:2. Inhalation of 1–2% of sevoflurane was administered, with intravenous anaesthetics maintained at 4–8 mg/(kg·h) of propofol, 0.1–0.3 mg/(kg·h) of cisatracurium, and a single supplement of 20–100 μg of sufentanil as appropriate.

The assessment of POD in the cardiac surgery intensive care unit (CSICU) was conducted every 12 h using the Confusion Assessment Method for the ICU (CAM-ICU) (Guenther et al. 2010), while delirium assessment at the ward employed the CAM. POD evaluation was performed 1-5 days postoperatively. A singular researcher administered the CAM-ICU or CAM measurement. Given the variations in specificity and sensitivity, the Diagnostic and Statistical Manual of Mental Disorders, fifth edition, was concurrently used for delirium screening. The secondary outcomes in this study encompassed postintervention anxiety, postintervention clinical insomnia, mechanical ventilation duration exceeding 12 h, occurrences of myocardial infarction, stroke, CSICU stay, hospital stay, and in-hospital mortality. Postintervention anxiety status was assessed a day before the surgery using the Self-rating Anxiety Scale (SAS). Postintervention sleep status (clinical insomnia) was evaluated for sleep quality approximately a day before the operation based on the insomnia severity index. Mechanical ventilation duration exceeding 12 h was reported as the percentage of time spent on ventilation from CSICU admission to ventilator withdrawal. The CSICU stay duration was recorded in hours, while the hospital stay was recorded in days. In-hospital mortality was defined as the number of patients who died from admission to discharge divided by the total number of patients in the study group.

In the pretrial phase, preoperative intervention with intranasal DEX drip resulted in a 10% incidence of POD, where one out of 10 patients exhibited this outcome. In contrast, recent studies (Enomoto et al. 2023; Nakamura et al. 2023) reported POD incidences ranging from 14.1% to 53.3%. Taking the median value of 33.7%, a two-tailed type I error rate of 5% and a test efficacy of 80% (α of 0.05 and β of 0.20) was assumed. Using Power Analysis Software 15, the sample size was calculated, indicating a requirement of 44 patients in each group. Considering a drop-out rate of approximately 10%, a sample size of 49 cases per group was suggested.

For continuous variables adhering to a normal distribution, the mean (standard deviation) was used for description. Non-normally distributed continuous variables were described as medians (interquartile ranges), while categorical variables were described as percentages. Between-group comparisons of continuous variables were compared using the independent sample t-test or Mann–Whitney U test. Categorical variables across different groups were compared using the chi-square test, and a two-sided p-value of <0.05 was considered statistically significant. Baseline variables influencing POD (p<0.05) were screened through a multivariate logistic regression model. All statistical analyses were performed using the Statistical Package for Social Science (SPSS) 22.0 (Chicago, IL, USA).