A 65-year-old woman (weight: 57 kg; height: 148 cm) was scheduled for a robotic single-port hysterectomy under general anesthesia. She had a medical history of interstitial lung disease diagnosed one year before presentation, for which she used an indacaterol and glycopyrronium bromide combination inhaler. There were no acute exacerbations post-diagnosis. Laboratory test results, including renal and hepatic functions, were within normal ranges. Preoperative electrocardiography revealed a normal sinus rhythm, with a heart rate of 75 beats/min. The Institutional Review Board of our institution granted ethical approval (approval no. 2023-07-038).

Thirty min before surgery, glycopyrrolate (0.2 mg) and famotidine (20 mg) were administered intramuscularly. Upon arrival to the operating room, standard monitoring commenced, including electrocardiography, non-invasive blood pressure monitoring, and pulse oximetry. Additionally, the bispectral index (BIS™, Medtronic), neuromuscular monitoring (train of four, TOF), and a warm blanket were applied. The patient’s preoperative vital signs remained within normal ranges.

General anesthesia was induced and maintained using remimazolam (Byfavo™ Inj, Hana Pharm Co. Ltd) and remifentanil (Remiva™ Inj, Hana Pharm Co. Ltd). Remimazolam was initiated at a loading dose of 6 mg/kg/h until loss of consciousness, then maintained at 1 mg/kg/h. During the operation, remifentanil was adjusted to the target site concentration of 2.5-4.5 ng/ml (Minto model). Rocuronium (Rocumeron™ Inj, Il Sung Pharm) was used for neuromuscular blockade, with a bolus of 50 mg followed by maintenance at 50 mg/h administered intravenously.

During the operation, mean arterial pressure was maintained above 70 mmHg, and no arrhythmia episodes occurred. Adequate anesthesia depth was maintained, with a BIS of 40-60 (Fig. 1). The operation proceeded uneventfully; anesthesia lasted 140 min. The administration of rocuronium ceased upon removal of the robotic surgical device. To prevent postoperative pain, nausea, and vomiting, fentanyl (30 μg), ramosetron (0.3 mg), and ketorolac (30 mg) were intravenously injected as the incision site closed. Remimazolam was discontinued when the skin sutures were complete. In total, 140 mg of rocuronium and 160 mg of remimazolam were injected.

Sugammadex (300 mg) was administered intravenously at the end of anesthesia to reverse neuromuscular blockade. At that time, the TOF count was 0. Five min after sugammadex administration, TOF showed a count of 4, and the ratio was 99%. However, 15 min post-remimazolam discontinuation, consciousness did not return. Intravenous flumazenil (0.2 mg) was administered, leading to spontaneous breathing and response to verbal commands, with a sustained BIS > 90. Subsequently, the patient was extubated and transferred to the post-anesthesia care unit (PACU) under intravenous patient-controlled analgesia (IV-PCA). Normal saline was mixed with fentanyl (970 μg), ramosetron (0.6 mg), and ketorolac (150 mg) to make 100 ml of solution; the continuous injection amount was 0.5 ml/h, the bolus dosage was 1 ml, and the lockout interval was set at 10 min. Upon PACU arrival, the patient was fully awake and normothermic. Thirty-five min post-admission, vital signs remained stable, and she was transferred to the general ward with a modified Aldrete score of 10.

However, 6 h post-transfer to the general ward, the patient's consciousness gradually declined, leading to deep drowsiness. According to the caregiver, the patient complained of pain and pressed the button to receive bolus IV-PCA. Opioids were initially suspected as the cause; therefore, IV-PCA was clamped, and she was closely observed for an hour. Approximately 29.1 µg of fentanyl was administered through IV-PCA. Her consciousness further declined to a stuporous state, responsive only to a strong sternal rub, with a Modified Observer's Alertness/Sedation scale (MOAA/S) score of 1. She showed no upper airway obstruction, a respiratory rate of 12 breaths per min, and a peripheral oxygen saturation of 95% on room air. Both pupillary reflexes and deep tendon reflexes were normal, and both Babinski and Hoffmann's signs were absent. Neck stiffness was not observed. Although there was insufficient evidence in the physical examination to suspect an opioid-related adverse effect, naloxone was empirically administered considering that opioids are the most likely to change mental status among the drugs administered after recovery from anesthesia. Despite intravenous naloxone (200 μg) administration to reverse opioid effects, her MOAA/S score remained at 1.

A neurologist was consulted considering the possibility of cerebral infarction, although no neurological findings suggestive of a brainstem lesion were observed. To further investigate, brain magnetic resonance imaging (MRI) was planned. During the preparation for brain MRI, flumazenil (0.2 mg) was administered, considering the residual effect of remimazolam. After administering flumazenil, the patient’s mental status normalized immediately; the MOAA/S score was 5. Fig. 2 illustrates the timeline post-remimazolam discontinuation.

As consciousness was sustained, brain MRI was canceled, and no further testing was performed. The patient was discharged three days postoperative. A follow-up two weeks postoperative revealed no specific findings, and no further neurological examinations were conducted.

Remimazolam is an ultra-short-acting IV benzodiazepine (BDZ) that is rapidly hydrolyzed to a pharmacologically inactive metabolite (CNS 7054) via non-specific tissue esterase activity, mainly liver carboxylesterase, and excreted predominantly in the urine. This process leads to the rapid onset and recovery of sedation and a predictable duration. Remimazolam shows a high clearance, small steady-state distribution volume, and short terminal half-life (70 ± 10 min). The simulated context-sensitive half-life after a 4 h infusion was 6.8 ± 2.4 min. Previously, loss of consciousness was observed 5 ± 1 min after the start of infusion; full alertness was regained 19 ± 7 min after ending the infusion [1]. The remimazolam pharmacokinetics did not differ between patients with normal renal function (estimated glomerular filtration rate [eGFR] ≥ 80 ml/min/1.73 m²) and end-stage renal failure (eGFR < 15 ml/min/1.73 m²) [1].

As a reversible agent that enables rapid recovery of consciousness, remimazolam offers hemodynamic stability, unlike propofol or inhalational anesthetics. Despite its short half-life and widespread use, several cases of patients not awakening after general anesthesia with remimazolam have been reported. In most cases, delayed emergence was observed in patients with severe hepatic dysfunction [2-4]. Flumazenil was administered to restore consciousness. However, re-sedation may occur after using flumazenil [5,6]. In these cases, re-sedation occurred within an hour, possibly caused by the flumazenil treatment duration. Therefore, closely monitoring re-sedation one hour after emergence is advisable. In our case, re-sedation occurred 6 h post-emergence in a healthy patient with normal renal and hepatic functions.

Possible reasons for re-sedation include an overdose of remimazolam and the use of flumazenil as a reversal agent. During remimazolam administration, electroencephalography showed an increased β-wave, the criterion for a high BIS value. BIS values during remimazolam administration at appropriate doses were greater than those during administration of other anesthetics; some patients showed a BIS > 60. Furthermore, BIS values correlate weakly with the depth of anesthesia when BDZs are used instead of propofol [7,8]. High concentrations of untitrated remimazolam caused delayed recovery; the reversal agent, flumazenil, was used for rapid emergence. Unlike sugammadex, which neutralizes rocuronium, flumazenil reverses the effects of BDZs through competitive antagonism, and it has a short half-life of 40-80 min [9,10]. Therefore, remimazolam’s sedative effects may reappear as the blood flumazenil concentration decreases. However, this likelihood remains minimal given the recommended dose of remimazolam (1 mg/kg/h) and the administered dose (1 mg/kg/h).

Second, inter-individual variability and drug interactions affect carboxylesterase enzymatic activity [11]. However, determining this is impossible because enzyme activity tests were not performed; the test also requires liver tissues. Previous studies reported individual differences in sensitivity to BDZs [11], possibly a causal factor in our case. However, this hypothesis does not explain the re-sedation occurring after several hours. A drug interaction is also not likely, given the patient’s medication history.

Third, the dose of flumazenil may have been insufficient. However, the appropriate flumazenil dose that reverses the effects of remimazolam remains unknown. Aya et al. [12] indicated that the amount of flumazenil needed to reverse remimazolam was related to sex, height, weight, and body mass index; it was not related to the total dose of remimazolam or the average rate of remimazolam administration. When calculating the amount of flumazenil our patient needed, the formula presented in this study yielded 0.273 mg, which is more than the amount administered. However, even if the amount of flumazenil was initially insufficient, explaining why re-sedation occurred after 6 h is difficult, considering the context-sensitive half-life of remimazolam.

Fourth, re-sedation and delayed emergence due to remimazolam extravasation have been reported [13]. In our case, no problems occurred with the IV site; achieving sedation would have been difficult after several hours in a completely awake state.

Fifth, a transient ischemic attack (TIA) may have occurred. Temporary neurological symptoms of decreased consciousness were observed. Despite the recovery, a TIA could not be completely excluded. However, because the patient had no risk factors that could trigger a TIA and showed improvement immediately after flumazenil administration, the likelihood of TIA seems low. However, dismissing this idea is impossible because we did not perform brain imaging.

Sixth, given the pharmacological profile of CNS 7054, the metabolite of remimazolam, it is highly unlikely that it would cause sedation. The molecule has significantly reduced affinity for the benzodiazepine receptors—approximately 300 times lower than remimazolam [9]. Due to this significantly reduced affinity, CNS 7054 is not generally considered capable of causing sedation under normal circumstances and dosages. However, like with any pharmacokinetic and pharmacodynamic considerations, individual variations in metabolism and sensitivity could theoretically influence outcomes, albeit such an occurrence would be exceptionally rare and not supported by the known properties of CNS 7054.

In our patient, 160 mg of remimazolam was administered intravenously over 140 min; re-sedation occurred 6 h later. Considering remimazolam’s pharmacokinetics, it was not initially suspected to be the cause. However, the mental status returning immediately after flumazenil administration suggests that remimazolam caused the re-sedation. Multiple factors contribute to post-anesthesia re-sedation, including opioid side effects, postoperative stroke, and hypoglycemia resulting from extended nothing by mouth. Beyond these recognized factors, individual sensitivity to remimazolam may also be important [14]. Fatal side effects can occur at any instance. If a patient’s mental state is abnormal, determining the cause through additional evaluation is important.

Remimazolam is increasingly used due to its significant benefits, including rapid recovery time and an excellent safety profile. Its use is particularly growing in outpatient anesthesia and short-term procedures. Although hospital protocols vary, most patients are discharged on the same day, making it difficult to monitor them for more than 4 to 6 h. Consequently, while extremely rare, re-sedation may occur after the patient returns home. Although remimazolam typically causes minimal respiratory depression, this can become life-threatening if spontaneous breathing disappears without close observation. Furthermore, patients experiencing re-sedation may face an increased risk of falls and related injuries due to impaired consciousness and motor coordination. This is especially concerning in settings where patients may attempt to walk or perform tasks soon after procedural sedation. Therefore, it is crucial for patients, their caregivers, and healthcare providers to be aware of the potential for re-sedation and its associated risks when using remimazolam. If unexpected sedation occurs after the patient awakens from anesthesia, it is necessary to review the anesthesia records to ascertain the efficacy of the medications administered during the procedure.

In conclusion, re-sedation or mental changes after surgery can have fatal consequences; therefore, identifying and resolving these issues is important. Based on this case report, even with the high clearance, short context-sensitive half-life, and ultra-short-acting properties of remimazolam, re-sedation can occur anytime, depending on the patient. We suggest extending observation periods beyond the expected duration of the drug's action to quickly identify and manage re-sedation, and to establish protocols that prevent unexpected accidents more effectively.