Fixed-Dose Nivolumab Costs More Than Simulated Weight-Based Dosing in Japan

A new analysis of nationwide Japanese claims data suggests that fixed-dose nivolumab, the immune checkpoint inhibitor used across a range of cancers, carried a substantially higher drug acquisition cost per administration than a simulated body weight-based alternative. Among 3,243 patients drawn from a large administrative database, the median difference between what was actually billed under fixed dosing and what would have been billed under weight-based dosing was 89,540 Japanese yen per administration, with a bootstrap 95 percent confidence interval of 86,945 to 92,135 yen. At the exchange rate of roughly 155 yen per US dollar cited in the study, that gap amounts to several hundred dollars per infusion at the median. The finding, published in Drugs – Real World Outcomes, quantifies a structural cost consequence of dosing policy in a population whose body weight distribution differs markedly from the Western populations where most prior economic analyses were conducted.

Nivolumab is a fully human IgG4 monoclonal antibody that blocks the programmed cell death-1 receptor, and it was first approved in Japan in July 2014. Initially, patients received 3 milligrams per kilogram of body weight every two weeks. Later, population pharmacokinetic analyses showed that fixed dosing produced comparable benefit-risk profiles across tumor types, and fixed-dose regimens of 80, 240, 360, and 480 milligrams were approved in Japan beginning in August 2018. The shift simplified drug preparation, reduced the risk of preparation-related errors, and improved workflow efficiency. But those operational gains came with an open economic question: in a country where the median body weight in this study cohort was just 59.6 kilograms, how much extra drug does fixed dosing effectively purchase?

From 17,161 patients who received nivolumab, the team excluded 13,474 who had no health checkup data, two who lacked body weight despite available checkup records, 372 whose first observed administration predated August 2018, and 70 whose dose did not match an approved fixed regimen. The remaining 3,243 patients formed the analytic cohort, with a median age of 72 years, 75.1 percent male, and only 4.6 percent weighing 80 kilograms or more. Body weights came from health checkup records, and the measurement closest in time to the index administration was used.

The study’s design deserves careful attention because it is a counterfactual comparison, not a head-to-head trial. For each patient, the researchers compared the billed milligram quantity of nivolumab under fixed dosing with a simulated weight-based dose calculated from that patient’s recorded body weight and the corresponding approved regimen: 1 milligram per kilogram for the 80-milligram dose, 3 for 240, 4.5 for 360, and 6 for 480. No separate group of patients actually received weight-based dosing. The index administration was defined as the first nivolumab administration observed during the database observation period, and because no washout period was applied, it did not necessarily represent a patient’s first-ever dose. All comparisons were restricted to this single administration to avoid the influence of later dose modifications, interruptions, or regimen changes.

Costs were valued using official Japanese National Health Insurance drug prices. In the primary analysis, both the observed fixed-dose quantities and the simulated weight-based quantities were priced at the same per-milligram rate derived from the 240-milligram vial, an approach that deliberately ignored vial-size constraints and wastage in order to isolate the dose-proportional cost difference arising from the dosing definitions themselves. The median cost per administration was 311,443 yen for fixed dosing, with an interquartile range of 311,443 to 467,165 yen, and 251,750 yen for the simulated weight-based counterfactual, with an interquartile range of 212,820 to 323,122 yen. The identical lower quartile and median for fixed dosing reflect the concentration of patients at the 240-milligram dose, which 2,237 patients, or 69.0 percent of the cohort, received. Mean costs were 369,000 yen for fixed dosing and 277,520 yen for the counterfactual.

The median of individual per-administration cost differences was 89,540 yen, and the median individual relative increase in cost for fixed dosing over the simulated counterfactual was 34.1 percent. Because cost variables were skewed, the authors summarized results with medians and interquartile ranges and performed no hypothesis testing; the bootstrap resampling with 10,000 iterations served only to characterize uncertainty in the median estimate. The authors are explicit that these differences should not be read as evidence of pharmacological overdosing. The fixed-dose quantities came from billed milligram amounts, which reflect reimbursed drug use under the Japanese insurance system, and the differences may arise from dosing definitions, drug prices, and billing structures rather than from clinical decisions about how much drug a patient needs.

Two sets of sensitivity analyses tested the robustness of the estimate. First, the researchers restricted the interval between the body weight measurement and the index administration to plus or minus 6 months and plus or minus 12 months, windows chosen to accommodate the annual health checkup structure of the database. The median cost difference was identical across both windows at 94,730.64 yen, and the median relative dose difference was 36.4 percent in both. Only a small fraction of cases, 3.9 percent under the 12-month window and 4.5 percent under the 6-month window, showed fixed dosing costing less than the simulated weight-based alternative. Second, the team systematically adjusted body weight values by minus 10, minus 5, zero, plus 5, and plus 10 percent to simulate measurement error. The median incremental cost fell from 115,494 yen under the minus 10 percent scenario to 63,586 yen under the plus 10 percent scenario, but it remained positive in every case, indicating that the direction of the estimated difference was stable even as its magnitude shifted with the assumed error.

A supplementary vial-level analysis added a layer of realism by valuing both dose quantities using the minimum-cost combination of the available 20, 100, 120, and 240 milligram vials at their official prices. Under those assumptions, the median individual cost difference shrank to 72,394 yen, with an interquartile range of 47,822 to 103,400 yen, but remained positive. The same vial-level model estimated a median residual drug amount of 10 milligrams for the simulated weight-based comparator, a median residual-to-required-dose ratio of 4.5 percent. The authors note that this supplementary analysis could not capture institution-specific practices such as vial sharing, inventory management, procurement arrangements, or preparation workflows, and the claims database did not allow identification of synchronized administrations within the same facility or on the same day.

The economic stakes are considerable. According to drug price-based sales data published by IQVIA cited in the paper, nivolumab ranks among the top-selling prescription medicines in Japan, so even modest per-administration differences can translate into meaningful population-level expenditure. Prior studies in Western countries had already reported that fixed dosing of immune checkpoint inhibitors can lead to higher administered drug quantities and higher acquisition costs, and budget impact analyses using claims and electronic health record data have reached similar conclusions. Pharmacokinetic modeling studies suggest that alternative strategies, including dose banding and pharmacokinetically guided dosing, might achieve comparable drug exposure at lower cost under certain assumptions. But the authors argue that Western findings cannot be directly extrapolated to Japan, where patient body weight is generally lower and healthcare systems, reimbursement structures, and preparation practices differ.

The study’s limitations frame what its findings can and cannot say. The weight-based comparator was simulated for the same patients rather than observed in a separate treatment group, so the results are counterfactual estimates. Clinical outcomes, safety, medication errors, workflow efficiency, and overall healthcare costs were not assessed, meaning the analysis cannot determine the clinical superiority, cost-effectiveness, or overall value of either dosing strategy. Body weight came from health checkup records rather than measurements taken contemporaneously with the infusion, and although sensitivity analyses restricted the interval, residual misclassification may remain. The primary per-milligram costing ignored vial-size constraints and wastage. The authors position their contribution as quantifying the expected acquisition cost difference under Japanese real-world reimbursement conditions, not as evidence favoring one strategy, and they call for future research incorporating longitudinal treatment patterns, dose modifications, clinical outcomes, and institution-level practices such as dose rounding and vial sharing.

Subject of Research: Acquisition cost comparison of fixed-dose versus simulated body weight-based nivolumab dosing using Japanese claims data

Article Title: Estimated Acquisition Cost Differences between Fixed and Simulated Body Weight-Based Nivolumab Dosing Using Real-World Claims Data in Japan

Article References: Harada, T., Noguchi, Y., Ohnishi, A., Masuda, R., Sugimoto, T., & Yoshimura, T. (2026). Estimated Acquisition Cost Differences between Fixed and Simulated Body Weight-Based Nivolumab Dosing Using Real-World Claims Data in Japan. Drugs – Real World Outcomes. https://doi.org/10.1007/s40801-026-00582-6

Image Credits: AI Generated

DOI: 10.1007/s40801-026-00582-6

Keywords: nivolumab, fixed dosing, body weight-based dosing, drug acquisition cost, Japanese claims data, DeSC database, immune checkpoint inhibitor, pharmacokinetics, sensitivity analysis, health economics, National Health Insurance drug prices, real-world evidence