A 69-year-old female patient (American Society of Anesthesiologists physical status III) underwent right-sided video-assisted thoracoscopic surgery with mediastinal lymph node dissection under general anesthesia. This study was approved by the Institutional Review Board (IRB) of Kosin University Gospel Hospital (IRB No. KUGH 2025-07-008), which waived the requirement for informed consent. Preoperatively, a thoracic epidural catheter was inserted at the T6-7 interspace for postoperative analgesia. The patient was 154 cm tall and weighed 54 kg. The catheter was advanced 6 cm into the epidural space using the loss-of-resistance technique without fluoroscopic guidance. A standard test dose of 3 ml of 1.5% lidocaine with epinephrine 15 μg was administered to exclude inadvertent intrathecal or intravascular placement, with no adverse findings. The surgical procedure lasted for 1 h and 40 min with a total anesthesia time of 2 h and 50 min.

When the surgery was nearing completion, the primary anesthesia nurse briefly stepped away and delegated epidural analgesia administration to the covering anesthesia nurse. The intended loading dose (ropivacaine 0.75%, 2 ml; morphine, 2 mg; diluted with 6 ml of normal saline) was prepared in a labeled syringe. However, a medication error occurred when a syringe containing sugammadex (Breathon^®, Boryung) 200 mg and ramosetron (Nasea^®, Daiichi Sankyo) 0.3 mg—intended for IV use—was inadvertently selected and administered via the epidural catheter. The drugs were given in the left lateral decubitus position. An error was identified within 5 min, and the anesthesiologist was notified immediately. As no immediate adverse effects were observed, we proceeded with emergence under close monitoring. Neuromuscular monitoring revealed a train-of-four (TOF) count of 2, and 200 mg of sugammadex was administered intravenously. The patient was successfully extubated and transferred to the post-anesthesia care unit (PACU), where she remained hemodynamically stable and neurologically intact for the first 30 min. The patient was subsequently transferred to the surgical intensive care unit (SICU) for close observation.

Six hours later, neurological examination revealed mild right triceps weakness (Medical Research Council grade 3) and bilateral weakness in the C8 dermatomes. No lower extremity deficits or sphincter dysfunction were observed. The motor and sensory symptoms partially improved on postoperative day (POD) 1. However, the patient developed transient confusion and memory disturbances, prompting neurological consultation. Magnetic resonance imaging showed no acute lesions. On POD 2, both the neurological and cognitive symptoms had almost completely resolved. The epidural catheter was removed without incident, and the patient was discharged on POD 6 without neurological sequelae (Fig. 1).

This case report describes the inadvertent epidural administration of two IV medications, sugammadex and ramosetron, during the perioperative period, which resulted in transient neurological impairment. To the best of our knowledge, this is the first reported case involving the delivery of these drugs into the epidural space. The outcomes were benign and limited to transient nerve root symptoms without permanent neurological sequelae. Nonetheless, this incident underscores the potential dangers of wrong-route medication errors in anesthetic practice and highlights the need for robust preventive strategies. In cases of epidural drug misadministration, clinical effects are determined by a combination of factors: (1) dose and volume relative to neuraxial tolerability, (2) anatomical spread dynamics, (3) physicochemical properties such as pH and osmolality, and (4) preservative content. Therefore, the potential neuraxial effects of sugammadex and ramosetron were evaluated.

Sugammadex is a modified cyclodextrin designed to encapsulate neuromuscular blocking agents such as rocuronium in the bloodstream [5]. Its large molecular weight (approximately 2178 Da) and hydrophilic structure greatly limit its ability to cross lipid membranes such as the blood-brain barrier [5]. In the present case, sugammadex administered via epidural injection likely had minimal direct action on the spinal cord or nerve roots. It is primarily absorbed into the systemic circulation through the epidural veins over time. Notably, the patient did not exhibit signs of dense motor or sensory block immediately after the injection, suggesting that sugammadex does not act as a local anesthetic. Moreover, although sugammadex is generally considered safe and well-tolerated when administered intravenously, data on its neurotoxicity remain limited. An in vitro study suggested that extremely high concentrations of sugammadex could induce neuronal apoptosis in cell cultures; however, this has not been observed clinically [5]. Given the single dose used in this case (200 mg), it is unlikely that sugammadex caused direct neural injury. Although sugammadex administered via the epidural route may be absorbed into the systemic circulation over time, it did not result in a clinically meaningful reversal of neuromuscular blockade in our patient. Persistently low TOF counts (≤ 2) necessitated additional IV sugammadex administration, indicating that epidural absorption alone was insufficient to achieve therapeutic plasma concentrations for effective rocuronium antagonism.

Ramosetron is a 5-HT₃ antagonist, similar to ondansetron, that is used to prevent nausea and vomiting. There is limited literature on the neuraxial administration of ramosetron. Ondansetron, however, has been studied and even intentionally administered via neuraxial techniques because of its effects on spinal-induced shivering and pruritus [6]. Notably, a previous case report documented an accidental epidural injection of ondansetron (4 mg) in a laboring patient, with no neurologic complications observed [3]. Ondansetron’s poor central nervous system penetration and lack of significant effects on neural tissue likely account for this benign outcome [6]. By extension, the inadvertent epidural administration of ramosetron in our patient was not expected to cause direct neurotoxicity.

In the present case, no neurological symptoms were observed immediately after the medication error. Inadvertent intrathecal drug administration was excluded based on a negative response to the epidural test dose administered at the time of catheter insertion. Neurological symptoms were first noted approximately 6 hours after the epidural injection. Transient weakness in the triceps muscles and C8 dermatomal tingling suggest an effect on the lower cervical nerve roots.

Inadvertent epidural injection involved approximately 4 ml of IV formulations (sugammadex and ramosetron) at the T6-7 level. With this limited volume, direct cephalad spread to the C8 nerve roots is anatomically unlikely; therefore, we considered other possible mechanisms. The most likely mechanism is chemical irritation of the neuraxial structures due to the physicochemical properties of the injected solutions. Sugammadex is an IV-only preparation with a near-neutral pH (7.0-8.0) and moderately high osmolality (300-500 mOsm/kg) [7], whereas ramosetron has an acidic pH (4.0-5.0) and an osmolality approximately equivalent to normal saline [8]. The relatively high osmolality of sugammadex and the low pH of ramosetron may have contributed to neural irritation [9]. In addition, preservatives such as benzyl alcohol and polyethylene glycol are known to be neurotoxic when inadvertently introduced into the intrathecal space [10,11]. Both sugammadex and ramosetron are preservative-free and lack benzyl alcohol, chlorobutanol, or other potentially neurotoxic antimicrobial agents [7,8]. Supporting this interpretation, meningitis-like presentations have been reported after epidural steroid injections using non-neuraxial formulations, in which sterile CSF findings and spontaneous recovery suggested chemical meningeal irritation rather than infection or direct trauma [12,13].

The patient experienced transient postoperative delirium that resolved within 24 h. While the timing might raise questions about a possible relationship with the epidural medication errors, this complication was more likely attributable to established risk factors, including advanced age and ICU stay. Neither sugammadex nor ramosetron has been associated with significant central nervous system effects at therapeutic doses. Postoperative cognitive dysfunction is a well-documented phenomenon in elderly patients undergoing major surgery. The patient’s neuroimaging findings were unremarkable, and her cognitive status returned to baseline without any specific intervention. Therefore, delirium was considered a common postoperative complication rather than a direct consequence of epidural injection error.

There are no established antidotes or reversal agents for drugs inadvertently injected into the epidural or intrathecal spaces. Management is primarily supportive and situation-dependent. Key steps include immediate recognition of the error, discontinuation of further administration, and prompt neurological assessment. If detected early (during or immediately after injection), some experts suggest aspirating the epidural catheter to remove residual drug and possibly flushing the epidural space with preservative-free saline to dilute the substance [14]. In our scenario, by the time the error was recognized (5 min post-injection), the drugs had likely already been distributed within the epidural space; therefore, we limited management to close monitoring. Large-volume lavage of the epidural space is not well studied and could potentially promote cranial spread of the drug [14]; thus, we did not attempt it. Supportive care involved continuous hemodynamic and respiratory monitoring (given the risk of hypotension or respiratory depression), serial neurological examinations, and ensuring readiness to provide advanced supportive measures (intubation, mechanical ventilation, or vasopressor administration) if needed.

In this case, the administration error was multifactorial, arising from task delegation without verification, similar syringe appearance, and lack of route-specific safeguards. The intended epidural bolus volume was 8 ml, whereas the mistaken syringe contained 4 ml. The drug was prepared without remeasuring its volume immediately before administration. All syringes were identical, unlabeled 10-ml syringes, and the covering anesthesia nurse administered the drug based solely on visual estimation, without a cross-check procedure. Consequently, this discrepancy was not detected at that time. In addition to these human factors, structural and system-related issues must be emphasized, highlighting the need for structured, system-level interventions. Recommended strategies include standardized labeling practices, the use of color-coded or route-specific connectors, mandatory double-check or read-back protocols during medication handoffs, and consideration of NRFit^®-compatible epidural equipment to prevent misconnections. Implementing perioperative safety protocols that account for human errors, such as cognitive overload and communication lapses, may further reduce the likelihood of recurrence. Following this event, our institution conducted a root cause analysis and revised the PACU handover process, accompanied by targeted staff reeducation on medication administration procedures.

This case highlights the potential risks of wrong-route administration of IV medications via an epidural catheter, even when the outcome is benign. Although no permanent neurological injury occurred, transient symptoms and diagnostic uncertainty underscore the importance of vigilance in perioperative drug handling. Ultimately, prevention remains the most effective strategy. Clear labeling, standardized protocols, and strict verification processes are essential for minimizing the risk of similar incidents in clinical practice.