Feasibility of remimazolam rescue sedation for neonatal MRI: a case report

Abstract

Background

High-quality neonatal imaging often necessitates sedation or anesthesia to minimize motion artifacts. However, procedural sedation for non-painful pediatric examinations outside the operating room presents unique challenges due to their distinct cardiopulmonary physiology and drug metabolism. Although remimazolam is associated with minimal hemodynamic and respiratory effects in adults, evidence supporting its use in neonatal MRI remains limited.

Case description

We report the first successful application of remimazolam for MRI sedation in an early-term neonate (37 ⁴⁄₇ weeks’ gestational age, 2.9 kg). After unsuccessful non-pharmacologic measures and intranasal dexmedetomidine, intravenous remimazolam (0.1 mg/kg) produced adequate sedation within 1 min, allowing completion of a 17-minute MRI examination while preserving hemodynamic and respiratory stability.

Conclusions

This case highlights potential utility of remimazolam s a rescue sedative for neonatal MRI sedation. Comprehensive multicenter studies are warranted to evaluate its safety, efficacy, and role in clinical practice.

Supplementary Information

The online version contains supplementary material available at https://doi.org/10.1186/s12871-026-04004-6.

Background

Detailed and high-quality imaging is frequently essential for accurate pediatric diagnosis and optimal treatment planning [1]. Performing imaging in infants and neonates is particularly difficult without the aid of sedation, especially magnetic resonance imaging (MRI). The combination of extended scan duration (15–20 min), loud acoustic environment, and high susceptibility to motion artifacts creates substantial barriers when imaging non-cooperative pediatric patients who cannot maintain the necessary stillness without pharmacological intervention [2].

Neonatal sedation necessitates tailored protocols owing to several physiological considerations, including smaller airways, distinct drug metabolism [3], and immature cardiopulmonary systems that differ substantially from older pediatric or adult patients. Consequently, optimal neonatal sedation must balance adequate sedation depth with preservation of airway integrity and hemodynamic stability, while facilitating prompt post-procedural recovery [4, 5].

Remimazolam, a novel benzodiazepine with an ultrashort duration of action, provides quick sedation onset and rapid recovery while minimizing common side effects like hypotension and respiratory depression. Although approved for adult procedural sedation in 2020, its use in pediatric populations remains limited [6].

This case report is the first to explore the use of remimazolam as a sedative agent for MRI in early-term neonates, potentially expanding options for non-painful pediatric examinations if supported by future studies.

Case presentation

A 2.9 kg female early-term neonate born at 37 ⁴⁄₇ weeks’ gestational age was referred for evaluation of a palpable sacrococcygeal mass. The mother first detected a fetal sacrococcygeal mass measuring approximately 26 × 11 mm during an ultrasound at 26 ⁵⁄₇ weeks’ gestation. By 37 ³/₇ weeks’ gestation, a follow-up ultrasound showed that the mass had enlarged to a 43 × 32 mm hypoechoic lesion with a prominent feeding vessel. Additionally, the ultrasound revealed intermittent elevation of the middle cerebral artery flow velocity, with a pulsatility index below the 5th percentile. Following multidisciplinary discussion among obstetricians, pediatric surgeons, and neonatologists, a cesarean section was scheduled, with postnatal surgical evaluation planned after further anatomical assessment.

After delivery, the newborn was immediately transferred to the pediatric intensive care unit (PICU) for monitoring and preoperative evaluation. On examination, the neonate was clinically stable, alert, and without significant associated comorbidities. All vital signs and standard biochemical parameters remained within age-appropriate reference ranges. Because the exact extent of the lesion, possible deep extension, and its relationship to adjacent structures could not be fully characterized clinically, the surgical team requested early MRI after initial stabilization to assist operative planning. Thus, MRI was obtained for preoperative anatomic delineation rather than because of hemodynamic compromise or another immediate life-threatening indication.

Accompanied by family members, the neonate was transported to the MRI suite at 5 h post-delivery. The PICU team provided intravenous fluid therapy but withheld feeding in anticipation of possible sedation and subsequent surgical decision-making. Three attempts at non-pharmacologic imaging using a swaddle-and-sleep approach were unsuccessful because the infant awoke due to scanner noise. The radiology and anesthesia teams therefore concluded that high-quality diagnostic MRI could not be achieved without pharmacological sedation. Consequently, a procedural sedation protocol was initiated after obtaining informed parental consent.

Continuous multiparameter cardiorespiratory monitoring was maintained throughout the MRI procedure, including heart rate (HR), oxygen saturation (SpO₂), blood pressure (BP) and respiratory rate (RR). Baseline vital signs of the neonate before drug sedation were recorded as follows: HR 141 bpm, SpO2 100%, BP 79/45 mmHg, RR 35 breaths/min. Initially, intranasal dexmedetomidine (2 µg/kg) was administered together with parental comforting measures. After 30 min, the neonate attained a Modified Observer’s Assessment of Alertness/Sedation (MOAA/S) [7] score of 4, with stable hemodynamic and respiratory parameters (HR 126 bpm, SpO2 100%, BP 71/40 mmHg, RR 28/min). Despite meeting conventional sedation benchmarks, the initial MRI attempt failed due to motion artifacts upon transfer. A rescue dose of dexmedetomidine (1 µg/kg) was administered, achieving comparable sedation (MOAA/S score of 4) within 15 min and consistent physiological stability (HR 126 bpm, SpO₂ 100%, BP 78/49 mmHg, RR 26/min). However, recurrent agitation during positioning again precluded successful imaging. Given that the neonate already had a 24G intravenous cannula in the right hand and had been appropriately fasted, the anesthesia team prepared to implement an intravenous rescue sedation plan after assessing the neonate’s increasing agitation.

With all MRI-compatible equipment in place and emergency preparations complete (including immediate availability of flumazenil and size-appropriate airway devices), the attending anesthesiologist administered remimazolam as a slow intravenous push (0.1 mg/kg over 60 s) while positioned adjacent to the MRI scanner. Adequate sedation (MOAA/S score of 3) was achieved over one minute, with stable vital signs (HR 121 bpm, SpO₂ 100%, BP 73/38mmHg, RR 25/min). Following confirmation of adequate sedation depth and airway patency, the 17-minute MRI examination was completed uneventfully without the need for additional remimazolam (Fig. 1). During sedation, the following predefined adverse events were specifically assessed: apnea, oxygen desaturation below 95%, bradycardia requiring intervention, clinically significant hypotension requiring treatment, airway repositioning, suctioning, bag-mask ventilation, supraglottic airway placement, tracheal intubation, and flumazenil administration. None of these events occurred. Hemodynamic changes during sedation are summarized in Fig. 2.

Fig. 1.

MRI of the sacrococcygeal region. The sacrococcygeal mass measured approximately 4.6 × 2.3 × 1.8 cm. A: Axial view. B: Coronal view

Fig. 2.

Hemodynamic changes during sedation. dBP, diastolic blood pressure; sBP, systolic blood pressure; HR, heart rate

At the end of the scan, the neonate showed prompt return of spontaneous movement and age-appropriate responsiveness. After a standard 30-minute post-procedural observation period in the recovery area during which no immediate clinically apparent adverse events were observed, the infant met institutional recovery criteria and was returned to the PICU, where continuous monitoring was maintained for an additional 20 h. No delayed respiratory depression, desaturation, bradycardia, hypotension, airway intervention, or pharmacologic reversal was required during this period.

The neonate was then transferred to the operating room, where resection of the sacrococcygeal mass was performed under general anesthesia with tracheal intubation. The operation lasted 80 min, and two lesions were removed: a 5 × 3 cm sacrococcygeal mass and a 1.5 × 1 cm lesion located between the sacrum and coccyx. After tracheal extubation, she was returned to the PICU. The postoperative course was uneventful, and enteral feeding was resumed 4 h later. Histopathological examination showed that the larger sacrococcygeal lesion was a mature cystic teratoma containing mature brain tissue, whereas the smaller lesion consisted of cartilaginous tissue.

Discussion and conclusion

As the number of diagnostic procedures performed outside the traditional operating room setup for infant and neonate continues to increase, it is of utmost necessity to improve the efficacy of procedural sedation while ensuring safety [8]. While deeper sedation levels may improve behavioral control in uncooperative children, they paradoxically increase the risk of respiratory depression and desaturation events that can ultimately compromise procedural success [9]. This balance is especially delicate in neonates because airway patency, ventilatory reserve, and drug disposition are developmentally immature. In the present case, MRI was obtained early after initial stabilization not because of hemodynamic instability or another absolute life-threatening indication, but to better define the mass’s precise dimensions, neural involvement, and vascular relationships with surrounding tissues for operative planning. The neonate subsequently underwent surgical resection 20 h after MRI, with an uneventful postoperative course. Histopathology demonstrated a mature cystic teratoma in the main sacrococcygeal lesion and cartilaginous tissue in the smaller lesion between the sacrum and coccyx. This clinical course supports the role of MRI in preoperative anatomical assessment in this individual case, although it should not be interpreted as supporting routine emergent MRI within the first hours of life for all similar lesions.

Non-pharmacologic strategies should remain the first-line approach when feasible in neonates. In this patient, however, three swaddle-and-sleep attempts were unsuccessful because the infant awoke in response to scanner noise, and two administrations of intranasal dexmedetomidine, although associated with acceptable sedation scores and stable vital signs, did not provide sufficiently reliable immobility for MRI acquisition. This observation highlights an important practical point: clinically apparent sedation does not necessarily translate into successful motion-free imaging, particularly when transfer, positioning, and acoustic stimulation are involved. The limited success of dexmedetomidine in this case therefore likely reflects the procedural context and the known variability of intranasal absorption, rather than a complete lack of sedative effect.

Several sedatives have been recommended for pediatric procedural sedation, such as propofol, midazolam, and dexmedetomidine [10, 11]. Propofol is a widely used intravenous anesthetic agent for pediatric MRI sedation due to its rapid onset, short duration of action, and low incidence of postoperative nausea. However, the occurrence of injection pain and hypotension persist as significant drawbacks particularly in low-birth-weight infants, even with formulation improvements [12, 13]. Midazolam, a benzodiazepine acting on gamma-aminobutyric acid-A (GABA-A) receptors, provides sedation along with anxiolytic, anticonvulsant, amnestic, and muscle relaxant effects. Yet its clinical utility in neonates is limited due to risks of prolonged awakening times, especially in preterm infants with immature hepatic metabolic pathways that delay drug clearance [14]. Dexmedetomidine generally preserves spontaneous ventilation, making it attractive for imaging sedation, but bradycardia, prolonged sedation, and variable intranasal bioavailability may complicate its use [15]. In the present case, intranasal dexmedetomidine had already been administered twice, precluding further intravenous use to avoid potential bradycardia. Available evidence demonstrate that an ideal sedative should possess attributes including simplicity in application, rapid onset and recovery, negligible injection-associated discomfort and minimal cardiorespiratory depressive effects [16, 17].

Remimazolam, classified as a short-acting benzodiazepine, is rapidly metabolized by tissue esterases to inactive metabolites (independent of liver and kidney function). Its favorable pharmacokinetic features, including high clearance, small volume of distribution, rapid onset, quick recovery and combined with painless administration makes it an attractive option for procedural sedation [18]. Published pediatric experience remains sparse and has largely involved older infants and children undergoing procedural sedation or anesthesia; neonatal data are extremely limited, and remimazolam is not currently supported by standardized neonatal dosing guidelines.To the best of our knowledge, after a literature search of [PubMed, Embase, and Web of Science] through 2018–2026, no previous report describing remimazolam use for MRI sedation in an early-term neonate was identified. Published pediatric cases are summarized in Table 1 [19–22].

Table 1.

Literature review of remimazolam administration for procedural sedation during imaging examinations in pediatric patients

Author, year	Country	Population	Application	Remimazolam dosing and outcome	Reference
Zheng, 2026	China	N = 150, age = 3 months to 6 years	MRI	Initial intranasal dexmedetomidine (3 µg/kg) and administered remimazolam via Dixon’s up-and-down method. Remimazolam for successful sedation maintained consistent ED50 values of 0.12 mg/kg across all paediatric age groups.	[19]
Fu, 2024	China	N = 60; mean age = 2.8 ± 1.3 years;	CT/MRI examinations	Initial dose 0.2 mg / kg and maintenance dose 1–2 mg / kg / h. Compared with propofol, remimazolam offers effective sedation with a faster recovery time and exerts minimal effects on children’s cardiovascular and respiratory functions.	[20]
Hirano, 2023	Japan	N = 48; mean age = 7 years;	CT/ MRI, angiography, and radiation therapy	Initial administered as a continuous infusion (12 mg / kg / h) and decreased to 1–2 mg / kg / h once the desired level of sedation was achieved. Remimazolam was a safe and effective agent for procedural sedation when supplemented with propofol or ketamine.	[21]
Yang, 2021	China	N = 23; age = 5–18 months;	echocardiographic examination	Initial dose 0.3 mg / kg and rescue dose 0.1 mg / kg. Compared with dexmedetomidine, intravenous remimazolam exhibits significantly better sedative effects with shorter recovery time for pediatric echocardiography.	[22]

CT Computed tomography, MRI Magnetic resonance imaging

This case therefore adds preliminary clinical experience, but it should be interpreted in the context of very limited age-specific evidence.

There is currently no validated intravenous remimazolam dose for neonatal procedural sedation. The 0.1 mg/kg dose used in this patient was selected empirically as a conservative rescue dose rather than as a recommended neonatal regimen. In the limited pediatric literature, clinically effective sedation or anesthesia has been achieved with low intravenous bolus doses or titrated infusions in older infants and children [19], and 0.1 mg/kg lies at the lower end of reported pediatric bolus dosing. Because remimazolam has not been approved for pediatric indications in many settings, reported regimens vary substantially, further limiting extrapolation to neonates. Several considerations supported starting at the low end in this case. First, neonatal pharmacokinetics and pharmacodynamics cannot be assumed to mirror those of older children because esterase activity, protein binding, body water composition, and drug sensitivity change rapidly during early life [23]. Second, remimazolam was administered only after two intranasal dexmedetomidine doses, so a residual sedative effect may have reduced the amount of additional drug required. Third, the decision was made in a rescue context, with the goal of using the smallest effective dose and avoiding repeat boluses or infusion in the absence of neonatal pharmacokinetic data. For these reasons, the dose should be viewed as a cautious empirical choice rather than evidence of an optimal neonatal regimen.

Because neonatal safety cannot be inferred from older pediatric populations, we adopted a structured adverse-event framework rather than relying solely on a general description of “stable vital signs.” Continuous monitoring of HR, SpO₂, BP, and RR was maintained throughout sedation. Prespecified clinically relevant adverse events included apnea, oxygen desaturation below 95%, bradycardia requiring intervention, clinically significant hypotension requiring treatment, airway repositioning, suctioning, bag-mask ventilation, supraglottic airway placement, tracheal intubation, and flumazenil administration. None of these events occurred during the 17-minute MRI examination. In addition, the infant recovered promptly, and no delayed respiratory or hemodynamic instability was observed during the 30-minute post-procedural recovery period or the subsequent 20-hour PICU monitoring period before surgery. These observations support short-term clinical tolerability under close supervision in this individual patient.

However, the implications of this case are necessarily limited. Remimazolam use in this neonate was off-label, and neonatal pharmacokinetic and pharmacodynamic data remain lacking. No plasma drug concentrations were measured; therefore, no pharmacokinetic inferences can be drawn. In addition, the observed sedative effect may have been influenced by prior dexmedetomidine exposure, and follow-up was limited to the documented in-hospital observation period. Although no immediate or delayed clinically apparent adverse events were identified, subclinical respiratory events, delayed adverse drug reactions, or dose-related accumulation cannot be completely excluded. More broadly, the absence of adverse events in a single clinically stable early-term neonate cannot establish safety, cannot be generalized to preterm infants or critically ill neonates, and should not be taken as support for routine neonatal use.

In summary, this report describes the use of a single low-dose intravenous remimazolam bolus as rescue sedation for MRI in a carefully monitored early-term neonate after failure of non-pharmacologic measures and intranasal dexmedetomidine. Its organ-independent metabolism via tissue esterases may represent a theoretical advantage in neonates with immature hepatic or renal function. Prospective studies incorporating pharmacokinetic/pharmacodynamic assessment, standardized adverse-event reporting, and procedure-specific outcome measures are needed before broader neonatal application can be considered.

Abbreviations

MRI

Magnetic resonance imaging

PICU

Pediatric intensive care unit

HR

Heart rate

SpO2

Oxygen saturation

BP

Blood pressure

RR

Respiratory rate

MOAA/S

Modified Observer's Assessment of Alertness/Sedation

GABA-A

Gamma-aminobutyric acid-A

Author contributions

Q., L. C., J. C. and C. Y. experienced the case and performed anesthetic management. D. W. contributed to manuscript preparation and critical revisions. All authors approved the final manuscript.

Funding

Medical Empowerment Fund of China Red Cross Foundation (No.2022-51).

Data availability

The data supporting the findings of this case report are available from the corresponding author upon reasonable request.

Declarations

Ethics approval and consent to participate

All methods were performed in accordance with the relevant guidelines and regulations. Written informed consent for publication of the clinical details and clinical images was obtained from the patient.

Consent for publication

Written consent was obtained from the patient’s parents for publication of the case report.

Competing interests

The authors declare no competing interests.