Early Remimazolam Sampling: What a Population Clearance Estimate Can Establish

A population clearance estimate summarizes a fitted pharmacokinetic model. It is not a direct measurement of how quickly every patient eliminates a drug, and its interpretation depends on when samples were collected and which model parameters were estimated.

A 2025 remimazolam study by Ueshima and colleagues illustrates these limits. The investigators measured drug and metabolite concentrations during the early stages of infusion in 20 Japanese surgical patients.

Three early samples per patient

The researchers collected blood at one, three and five minutes after loss of consciousness. They measured remimazolam and its metabolite CNS7054 in separated plasma using liquid chromatography with tandem mass spectrometry. Remifentanil or fentanyl was added only after the final study sample.

The dataset contained 60 plasma sampling observations from 20 patients. Fourteen patients were at least 65 years old, but the full age range extended from approximately 42 to 89 years. Severe liver cirrhosis was excluded.

These observations characterize a short early infusion interval. They are not a series of post-infusion samples documenting each patient’s complete terminal decline.

Measured concentrations and modeled parameters differ

The investigators used a three-compartment model for remimazolam and a two-compartment model with a transit compartment for CNS7054. Several parameters were fixed at values reported in earlier research because the available individual data were insufficient to estimate them.

Consequently, the model combines current concentration measurements with structural assumptions and information from previous work. The number of compartments should not be mistaken for direct anatomical measurements, nor should a fitted parameter be described as an independently measured laboratory result.

How body weight entered the clearance estimate

Body weight was retained as a covariate of remimazolam clearance. The reported population mean clearance for the reference patient weighing 60.4 kg was 1.38 L/min, with a reported 95% confidence interval of 1.12 to 1.62 L/min.

The final model scaled clearance with body weight raised to the power of 0.75. Within that model, clearance therefore increased less than proportionally with body weight. This is a population relationship under the study conditions, not a universal individual dosing equation.

Other tested patient characteristics were not retained as pharmacokinetic covariates. Failure to retain a covariate in this small dataset does not establish that it has no effect in every clinical population.

What the early window cannot settle

The authors acknowledged limited patient numbers and sparse data. They did not assess longer administration periods or the later effects of concomitant anesthetic drugs. Albumin was not collected under a predefined protocol, preventing assessment of its pharmacokinetic contribution.

The study does not directly establish an individual’s terminal half-life, time to full recovery or concentration after a prolonged infusion. Those questions require observations and validation appropriate to the time interval and population being assessed.

The practical lesson is to keep the assay result, sampling window, fixed assumptions and population estimate distinct. A useful early-infusion model can still have clear limits when applied outside the data that supported it.

Source: Ueshima S and colleagues. Pharmacokinetic and pharmacodynamic characterization of remimazolam in older Japanese adults who underwent general anesthesia at early stage of infusion. Scientific Reports. 2025;15:24979.