Pediatric sedation is not adult sedation in small: lessons from a PK/PD analysis of remimazolam

Introduction

Each year, millions of children undergo diagnostic and therapeutic procedures that require sedation to ensure technical success while minimizing psychological trauma, behavioral distress, and procedural failure (1). Yet, reliable pediatric sedation remains difficult because children vary substantially in developmental stage, procedural tolerance, and vulnerability to respiratory compromise. Propofol provides dependable hypnosis but carries dose-dependent respiratory and cardiovascular depression without a specific antagonist (2); conventional midazolam has variable onset and duration and may cause paradoxical agitation or disinhibition (3); and sevoflurane is associated with emergence agitation or delirium (4). These limitations underscore the need for agents with rapid onset, predictable offset, hemodynamic stability, and pharmacologic reversibility. Remimazolam, an ultra-short-acting benzodiazepine designed for rapid ester hydrolysis to the inactive metabolite CNS7054 while retaining antagonism by flumazenil, has gained approval for adult procedural sedation (5). The approved adult regimen is relatively simple, using a 5 mg induction bolus with 2.5 mg top-up boluses (6). However, adult approval does not define pediatric dosing.

Translation into pediatric practice remains constrained by limited developmentally informed pharmacokinetic (PK) and pharmacodynamic (PD) data. In the absence of pediatric-specific models, clinical practice has often depended on extrapolation from adult regimens. Prior data in children aged 2 to 6 years undergoing general anesthesia suggested that, after allometric scaling, remimazolam elimination clearance was relatively similar to that of adults (7). However, linear weight-based scaling and adult-derived dose caps may produce inconsistent effects across pediatric populations. Children differ from adults in body composition, distribution volumes, organ blood flow, protein binding, enzyme maturation, and clearance, and these differences may alter both exposure and clinical response (8). Thus, pediatric dosing should be aligned not only with systemic exposure, but also with developmental pharmacology, the intended clinical endpoint, and safety monitoring. In this context, Colin et al. provide an important PK/PD analysis of remimazolam for procedural sedation in children and adolescents aged 6 years and older (9). Their study asks whether “adult-equivalent exposure” is the appropriate target when the pediatric endpoint, procedural conditions, and rescue workflow differ from adult sedation trials. Because the cohort was small and the dosing implications were derived from simulations, the findings should be interpreted as exploratory and hypothesis-generating rather than definitive clinical recommendations.

The trial reality: study design, dose-capping, and the rescue cascade

Colin et al. reported the first cohort of a phase 2/3, prospective, open-label U.S. study evaluating intravenous remimazolam for procedural sedation in children and adolescents aged 6 to 18 years (9). Eligible patients were American Society of Anesthesiologists physical status I to III and underwent elective diagnostic or therapeutic procedures expected to require less than two hours of sedation with preserved spontaneous ventilation. Exclusion criteria included emergency procedures, need for advanced airways, active respiratory or hepatic failure, and obstructive sleep apnea. The analyzed cohort included 31 patients across four treatment arms: bolus (n=5), infusion (n=5), bolus plus fentanyl (n=11), and infusion plus fentanyl (n=10) (9). The original dosing scheme was designed to match adult remimazolam exposures, using weight-based dosing capped at adult-equivalent absolute limits, including maximum doses of 7.0 mg without fentanyl and 5.0 mg with fentanyl (9). Sedation was assessed with the University of Michigan Sedation Scale (UMSS), targeting a score of 2 to 3, representing moderate to deep sedation (9,10).

The trial itself was clinically informative. After the first five participants, enrollment was paused because of a higher-than-expected incidence of hypotension (9). The investigators attributed much of this signal to propofol rescue administered when capped remimazolam doses produced inadequate sedation. The protocol was amended to allow higher infusion rates and top-up doses, but a second pause occurred after 31 patients because failure remained high; rescue sedatives were required in 29% of cases (9 of 31 patients), and rescue use was higher in the two infusion groups (6 of 15 patients; 40%) than in the two bolus groups (3 of 16 patients; 19%) (9). These events illustrate how regimen design and co-medication strategy shaped procedural feasibility.

This rescue cascade is central to interpreting both efficacy and safety. When initial dosing fails to achieve the intended depth, clinicians often escalate to familiar rescue agents to salvage the procedure. Rescue therapy then influences procedural success and adverse-event attribution. In this cohort, 19 of 31 participants experienced at least one adverse event, and transient hypotension—defined as a systolic or diastolic blood pressure decrease of at least 30% from baseline—was the most frequent event, with 20 total incidents reported (57.1%) (9). This does not establish remimazolam as the direct cause of the hemodynamic signal. Rather, the definition is sensitive to transient fluctuations that may occur with rescue sedation, including propofol. Fixed adult-derived dose caps may therefore function as operational drivers of under-sedation, rescue medication use, and a safety narrative shaped by rescue therapy. The trial is consequently informative not only as a PK/PD study, but also as an example of how pediatric dosing assumptions can fail operationally when insufficiently calibrated to the clinical endpoint.

Unraveling the PK/PD mismatch: from size correction to endpoint-driven dosing

To examine this mismatch, Colin et al. modeled 170 remimazolam and 160 CNS7054 concentration samples (9). Remimazolam disposition was best described by a three-compartment model, while delayed CNS7054 formation was captured using a one-compartment metabolite model with five transit compartments. The principal PK finding was lower size-normalized remimazolam clearance in this older pediatric cohort. After allometric scaling, clearance was estimated at 0.70 L·min⁻¹·70 kg⁻¹, compared with adult estimates of approximately 1.18 L·min⁻¹·70 kg⁻¹ for procedural sedation and 1.03 L·min⁻¹·70 kg⁻¹ for general anesthesia (9,11,12). This suggests clearance in children aged 6 to 18 years was approximately 32% to 41% lower than in adults, challenging the assumption that body weight correction alone is sufficient.

Developmental regulation of hepatic carboxylesterase 1 (CES1) offers a biologically plausible contributor. Ontogenic data indicate lower CES1 expression and hydrolytic activity in children than in adults, which may partly explain reduced clearance (13). Remimazolam is not primarily cleared by cytochrome P450 metabolism; it undergoes rapid hydrolytic inactivation predominantly through hepatic CES1 (14). This interpretation should be made cautiously because CES1 activity, expression, and genotype were not measured in the Colin et al. cohort, and the study population included older children and adolescents rather than only younger children.

The PK finding creates an apparent clinical paradox. Lower size-corrected clearance would usually imply higher systemic exposure at a given mg/kg dose and, consequently, lower dose requirements. Yet the trial experience suggested that more permissive dosing was needed to address under-sedation (9). The explanation lies in the broader exposure-response framework. From a PD perspective, adult remimazolam trials used the Modified Observer’s Assessment of Alertness/Sedation scale, often targeting mild to moderate sedation (6,15), whereas this pediatric study targeted UMSS 2 to 3, requiring greater behavioral suppression (9,10). Therefore, the same nominal exposure target may not produce the same procedural condition. From a PK perspective, developmental differences in body composition, including higher total body water and changing lean/fat proportions, may alter distribution volumes and early post-bolus concentrations. A larger apparent central volume per kilogram could lower peak concentrations after a fixed mg/kg bolus and delay effect-site attainment, particularly when dose caps and minimum redosing intervals restrict induction titration.

The PD model clarifies the clinical implications. Using proportional odds logistic regression, the authors identified a steep concentration-response relationship, with a Hill coefficient of 3.56 (9). The remimazolam effect-site concentration producing a 50% probability of UMSS ≥3 was 777 ng/mL without fentanyl. Target attainment increased sharply, from 17% at 500 ng/mL to 71% at 1,000 ng/mL (9). Clinically, this means that the transition from inadequate sedation to deeper-than-intended sedation may occur over a narrow exposure range. This finding does not simply justify higher dosing; it supports incremental response-guided titration rather than large empiric dose increases.

Fentanyl shifted the exposure-response relationship leftward. In Colin et al., the opioid effect was incorporated through a simplified Minto-type additive interaction framework; the remimazolam EC50 (effect-site concentration producing a 50% probability of response) decreased to 655, 533, and 287 ng/mL at predicted fentanyl steady-state concentrations of 1, 2, and 4 ng/mL, respectively (9,16). However, fentanyl concentrations were predicted rather than measured, and the trial lacked a fentanyl-only arm. Thus, the model could not distinguish additivity from synergy or infra-additivity (9,17). If the true interaction is synergistic, regimens derived under an additive assumption may overestimate remimazolam requirements during coadministration. Conversely, inadequate accounting for procedural stimulation or analgesic requirements could underestimate the exposure needed in painful procedures. Opioid coadministration strategies should therefore remain candidates for cautious titration and prospective validation.

The simulations translated these PK/PD findings into practical dosing implications. Under the revised but still constrained trial regimen, 9.2% to 22.0% of patients were predicted to fail to achieve UMSS ≥3 by the end of the 15-minute induction period (9). Model-informed optimization recommended higher per-kilogram bolus doses, faster infusion rates, extended continuous titration, and removal of fixed maximum absolute dose caps. The optimized regimens predicted improved target attainment, with 88% to 97% of patients achieving target sedation depth (9). This contrast, summarized in Table 1, illustrates how model-informed dosing can identify when conservative-appearing regimens may fail operationally and can propose strategies better aligned with the clinical endpoint. Importantly, these simulations should be read as defining a titration space rather than prescribing a fixed regimen. In practice, the optimized scheme would need to be tested as a protocolized clinical strategy in which dose escalation is conditional on observed sedation depth, elapsed time from the previous dose, procedural stimulation, and evolving respiratory or hemodynamic status. This framing preserves the value of model-informed design while avoiding the impression that simulated target attainment alone is sufficient to justify routine higher dosing.

However, dose-cap removal should not be interpreted as unrestricted escalation. The steep exposure-response relationship makes protocolized titration essential. Any higher-exposure regimen should incorporate incremental dosing, appropriate redosing intervals, predefined stopping rules, and continuous respiratory and hemodynamic monitoring. Safety assessment should include over-sedation, airway obstruction, hypoventilation or apnea, oxygen desaturation, capnographic abnormalities, hypotension, bradycardia, delayed recovery, and re-sedation. Capnography is particularly important because oxygen saturation may lag behind hypoventilation when supplemental oxygen is used. Hemodynamic interpretation should also account for rescue propofol, procedural stimulation, fasting status, intravascular volume, and concomitant opioids. Thus, model-informed dosing should replace rigid adult-derived caps with structured response-guided titration, not open-ended dose escalation. The goal is not to increase exposure for its own sake, but to reach the intended clinical endpoint at the lowest exposure that is safe and effective.

Limitations and future directions

Several limitations should temper interpretation. First, the cohort was small and open-label, limiting covariate discovery and generalizability (9). Second, fentanyl concentrations were not directly measured, and reliance on an external pediatric fentanyl model may bias estimation of the opioid-benzodiazepine interaction (9,16). Third, the absence of a fentanyl-only arm prevented formal discrimination among additive, synergistic, and infra-additive effects (9,17). Rescue therapy should also be interpreted as both an efficacy signal and a potential confounder of safety attribution, particularly when rescue agents have their own respiratory or hemodynamic effects.

CNS7054 adds another layer of uncertainty. Volunteer studies suggest that CNS7054 may act as a competitive antagonist to remimazolam and contribute to acute tolerance under some exposure conditions (18). Colin et al. attempted to incorporate this tolerance into the UMSS model, but estimates were unstable, likely because median CNS7054 concentrations in this brief procedural cohort (1,539 ng/mL) were much lower than in the volunteer study (9,176 ng/mL) (9,18). This issue may matter more during longer procedures, repeated dosing, prolonged infusions, or reduced hydrolytic capacity. The model also could not characterize UMSS 4 because excessively deep sedation was not observed, and children younger than 6 years were not represented. This absence is important because younger children may differ substantially in airway vulnerability, respiratory control, protein binding, organ maturation, distribution volumes, and CES1 activity.

Pharmacogenetic variability should also be incorporated into future precision-dosing frameworks. Functional CES1 variants may impair remimazolam hydrolysis; the CES1 G143E variant, in particular, has been shown to markedly reduce remimazolam metabolic capacity in experimental systems (14). CES1-mediated hydrolysis may also be affected by drug-drug interactions involving CES1 substrates or inhibitors (14). Although the clinical relevance of these variants and interactions in pediatric sedation remains to be established, this variability further supports individualized, response-guided titration rather than fixed extrapolated dosing.

A logical next step is prospective validation of pediatric-specific, response-guided regimens. Future trials should include younger children, standardize rescue algorithms, directly measure opioid concentrations where feasible, and prespecify efficacy and safety endpoints. Respiratory endpoints should include capnography-defined hypoventilation, apnea, airway obstruction, airway maneuvers, oxygen desaturation, delayed recovery, re-sedation, unplanned escalation of care, and need for airway intervention. Larger cohorts should also evaluate variability across procedure type, painful versus nonpainful stimulation, opioid coadministration, sedation depth targets, body composition, hepatic function, CES1 ontogeny, CES1 genotype, and concomitant CES1-interacting drugs. Prospective studies should distinguish induction failure from maintenance failure. Early under-sedation after bolus dosing may reflect distribution kinetics, dose caps, or delayed redosing, whereas later instability may reflect procedural stimulation, analgesic requirements, metabolite accumulation, or tolerance. Treating these events as a single endpoint could obscure why a regimen fails and which component of the dosing strategy requires revision.

Conclusions

The analysis by Colin et al. yields three lessons. First, adult exposure targets cannot be assumed to produce adequate pediatric procedural conditions when sedation depth goals, developmental pharmacology, and distribution kinetics differ. Second, model-informed dosing can identify when adult-derived constraints limit target attainment, but its recommendations should remain exploratory and hypothesis-generating until prospectively validated. Third, safety signals must be interpreted within the context of sedation depth targets and rescue workflows. Inadequate primary sedation may trigger rescue therapy, including propofol, which can alter hemodynamic interpretation and obscure attribution of adverse events.

The central message is therefore not simply that higher remimazolam exposure may be required in some children. Rather, pediatric sedation development should move from adult-exposure matching toward pediatric endpoint-driven, response-guided titration. Remimazolam may ultimately have an important role in pediatric procedural sedation, but optimal use will require alignment among clinical endpoints, developmental pharmacology, exposure-response modeling, predefined stopping rules, and respiratory and hemodynamic monitoring. This alignment should be tested prospectively before candidate regimens are generalized to routine practice.

Acknowledgments

The authors used ChatGPT (OpenAI) and NotebookLM (Google) during manuscript preparation and revision to support language editing, refinement of wording and structure, organization of source materials, and preparation of the summary table. The tools were not used to generate original data nor to perform statistical analyses. All AI-assisted text and table content were critically reviewed, revised, and verified against the original sources by the authors, who take full responsibility for the accuracy, integrity, and final content of the manuscript.

Footnote

Provenance and Peer Review: This article was commissioned by the editorial office, Translational Pediatrics. The article has undergone external peer review.

Peer Review File: Available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0233/prf

Funding: None.

Conflicts of Interest: All authors have completed the ICMJE uniform disclosure form (available at https://tp.amegroups.com/article/view/10.21037/tp-2026-0233/coif). The authors have no conflicts of interest to declare.

Ethical Statement: The authors are accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Open Access Statement: This is an Open Access article distributed in accordance with the Creative Commons Attribution-NonCommercial-NoDerivs 4.0 International License (CC BY-NC-ND 4.0), which permits the non-commercial replication and distribution of the article with the strict proviso that no changes or edits are made and the original work is properly cited (including links to both the formal publication through the relevant DOI and the license). See: https://creativecommons.org/licenses/by-nc-nd/4.0/.

References

1. Cravero JP, Blike GT. Pediatric sedation. Curr Opin Anaesthesiol 2004;17:247-51. [Crossref] [PubMed]
2. Chidambaran V, Costandi A, D'Mello A. Propofol: a review of its role in pediatric anesthesia and sedation. CNS Drugs 2015;29:543-63. [Crossref] [PubMed]
3. Nordt SP, Clark RF. Midazolam: a review of therapeutic uses and toxicity. J Emerg Med 1997;15:357-65. [Crossref] [PubMed]
4. Costi D, Cyna AM, Ahmed S, et al. Effects of sevoflurane versus other general anaesthesia on emergence agitation in children. Cochrane Database Syst Rev 2014;2014:CD007084. [Crossref] [PubMed]
5. Keam SJ. Remimazolam: First Approval. Drugs 2020;80:625-33. [Crossref] [PubMed]
6. U.S. Food and Drug Administration. BYFAVO (remimazolam) for injection, for intravenous use [prescribing information]. 2020. Available online: https://www.accessdata.fda.gov/drugsatfda_docs/label/2020/212295s000lbl.pdf
7. Gao YQ, Ihmsen H, Hu ZY, et al. Pharmacokinetics of remimazolam after intravenous infusion in anaesthetised children. Br J Anaesth 2023;131:914-20. [Crossref] [PubMed]
8. Fideler F. Pharmacokinetics in Children - What is Important for Correct Drug Dosage? Anasthesiol Intensivmed Notfallmed Schmerzther 2019;54:496-506. [Crossref] [PubMed]
9. Colin PJ, Bichajian LH, Curt VR, et al. Pharmacokinetics and Pharmacodynamics of Remimazolam for Procedural Sedation in Children and Adolescents. Anesthesiology 2025;143:368-82. [Crossref] [PubMed]
10. Malviya S, Voepel-Lewis T, Tait AR, et al. Depth of sedation in children undergoing computed tomography: validity and reliability of the University of Michigan Sedation Scale (UMSS). Br J Anaesth 2002;88:241-5. [Crossref] [PubMed]
11. Zhou J, Curd L, Lohmer LL, et al. Population Pharmacokinetics of Remimazolam in Procedural Sedation With Nonhomogeneously Mixed Arterial and Venous Concentrations. Clin Transl Sci 2021;14:326-34. [Crossref] [PubMed]
12. Zhou J, Leonowens C, Ivaturi VD, et al. Population pharmacokinetic/pharmacodynamic modeling for remimazolam in the induction and maintenance of general anesthesia in healthy subjects and in surgical subjects. J Clin Anesth 2020;66:109899. [Crossref] [PubMed]
13. Yang D, Pearce RE, Wang X, et al. Human carboxylesterases HCE1 and HCE2: ontogenic expression, inter-individual variability and differential hydrolysis of oseltamivir, aspirin, deltamethrin and permethrin. Biochem Pharmacol 2009;77:238-47. [Crossref] [PubMed]
14. Wang Z, McCalla Z, Lin L, et al. Impact of genetic polymorphisms and drug-drug interactions mediated by carboxylesterase 1 on remimazolam deactivation. Drug Metab Dispos 2025;53:100023. [Crossref] [PubMed]
15. Chernik DA, Gillings D, Laine H, et al. Validity and reliability of the Observer's Assessment of Alertness/Sedation Scale: study with intravenous midazolam. J Clin Psychopharmacol 1990;10:244-51.
16. Ginsberg B, Howell S, Glass PS, et al. Pharmacokinetic model-driven infusion of fentanyl in children. Anesthesiology 1996;85:1268-75. [Crossref] [PubMed]
17. Zhou J, Curd L, Lohmer LRL, et al. A population pharmacodynamic Markov mixed-effects model for determining remimazolam-induced sedation when co-administered with fentanyl in procedural sedation. Clin Transl Sci 2021;14:1554-65. [Crossref] [PubMed]
18. Vellinga R, Koomen JV, Eleveld DJ, et al. Target-controlled Infusion of Remimazolam in Healthy Volunteers Shows Some Acute Tolerance. Anesthesiology 2024;140:207-19. [Crossref] [PubMed]