Exploring individualized crovalimab dosing in PNH through in silico modelling: Potential for improved convenience and cost efficiency

Abstract

Aim

Paroxysmal nocturnal haemoglobinuria (PNH) is a rare, acquired haematopoietic stem cell disorder. Crovalimab, a complement C5‐inhibitor, is approved for PNH and can be self‐administered subcutaneously every 4 weeks, offering a more convenient route than intravenous C5‐inhibitors. Since exposure often exceeds pharmacokinetic targets and varies with body weight, we aimed to explored individualized dosing of crovalimab using clinical trial simulations to optimize therapy.

Methods

We compared standard, weight‐based and therapeutic drug monitoring (TDM)‐guided regimens in treatment‐naïve PNH patients over 52 weeks by means of clinical trial simulations based on the population pharmacokinetic model developed by the market authorization holder.

Results

Maintenance‐phase crovalimab concentrations were generally supratherapeutic with substantial variability. Weight‐based dosing reduced this variability, while TDM‐guided dosing further minimized it and allowed dose intervals to be prolonged in 24% of patients to a mean of 5.0 weeks, while maintaining exposure comparable to the standard regimen.

Conclusion

Individualized dosing of crovalimab may reduce exposure variability and enable less frequent dosing, supporting the feasibility of more flexible cost‐effective and patient‐tailored treatment regimens.

What is already known about this subject

• Crovalimab is a novel monoclonal antibody for treatment of complement‐mediated diseases.

• Treatment with crovalimab requires life‐long subcutaneous administration in a one‐dose‐fits‐all paradigm.

What this study adds

• Both the loading and maintenance dose of crovalimab can be optimized based on weight or pharmacokinetics.

• Dose optimization can improve patient convenience by prolonging the dose interval.

• Dose optimization can reduce costs.

1. INTRODUCTION

Paroxysmal nocturnal haemoglobinuria (PNH) is a rare, acquired haematopoietic stem cell disorder characterized by complement‐mediated haemolysis, thrombosis and bone marrow failure. ¹ Over recent years, several complement inhibitors have been introduced for the treatment of PNH, including C5‐inhibitors eculizumab, ravulizumab and crovalimab, C3‐inhibitor pegcetacoplan, factor D inhibitor danicopan and factor B‐inhibitor iptacopan. ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ These therapies have demonstrated significant efficacy in reducing haemolysis and thrombotic complications. ² , ³ , ⁴ , ⁵ , ⁶ , ⁷ However, their widespread use is limited by substantial treatment costs, even for biosimilars, restricting patient access worldwide. ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹²

Crovalimab is a complement C5‐inhibitor that is administered subcutaneously every 4 weeks by self‐administration at home during the maintenance phase, offering a convenient dosing route compared to the intravenous administration required for the other C5‐inhibitors. Furthermore, crovalimab binds a distinct C5‐epitope, enabling effective inhibition in patients with C5‐polymorphisms not responsive to eculizumab and ravulizumab. ¹³ Crovalimab shows a robust and predictable concentration‐dependent inhibition of terminal complement activity, with trough concentrations above ~50–100 μg/mL required to achieve maximal complement blockade, as evidenced by low CH50 and free C5‐levels. ¹⁴ , ¹⁵ , ¹⁶ Crovalimab is administered in a fixed‐dose interval regimen, with an increased dose for patients weighing >100 kg. ¹⁷ The mean steady‐state trough concentration of crovalimab was shown to 238 ± 72 μg/mL in patients weighing <100 kg and 195 ± 63 μg/mL in patients weighing >100 kg, despite administration of a higher dose, indicating a significant impact of body weight on drug exposure. ¹⁷ Given that this observed exposure substantially exceeds the pharmacokinetic target for full target inhibition and considering the effect of body weight, dosing intervals (and thus patient‐friendliness) could be optimized through further individualization based on weight or observed pharmacokinetics. Therefore, we performed an in silico evaluation of the potential benefits of individualized crovalimab dosing with regard to therapeutic exposure, patient‐friendliness and drug costs.

2. METHODS

We conducted a clinical trial simulation comparing the standard crovalimab dosing regimen with two alternative strategies in treatment‐naïve PNH‐patients over a 52‐week period (Table 1). The simulations were performed using the population‐pharmacokinetic model for crovalimab, as developed by the licence holder and published in the Food and Drug Administration (FDA) review documents. ¹⁷ In summary, the model was a two‐compartmental disposition model with a subcutaneous absorption compartment and first‐order elimination. Interindividual variability was included on clearance and volume parameters, as well as on the subcutaneous bioavailability and absorption rate constant. A combined proportional and additive residual error model was used. ¹⁷ The model parameters were previously published. ¹⁵

TABLE 1.

Evaluated dosing strategies for crovalimab.

	Standard of care	Weight‐based	TDM‐based
Loading dose
Day 1	<100 kg: 1000 mg	<75 kg: 1020 mg	<75 kg: 1020 mg
Loading dose (IV)	≥100 kg: 1500 mg	≥75 kg: 1360 mg	≥75 kg: 1360 mg
Days 2, 8, 15 and 22 (subcutaneous)	340 mg	‐	‐
Maintenance phase
Subcutaneous	Start day 28 <100 kg: 680 mg Q4W ≥100 kg: 1020 mg Q4W	Start day 14 <75 kg: 680 mg Q5W ≥75 kg: 680 mg Q4W	Start day 14 <75 kg: 680 mg Q5W ≥75 kg: 680 mg Q4W TDM at the third dose Ctrough Advise <75 μg/mL −1 week >200 μg/mL +2 weeks

The aim was to optimize both the loading and maintenance phase to improve patient convenience and reduce costs without compromising effective drug exposure.

2.1. Loading dose

We explored an alternative regimen with a single weight‐based intravenous loading dose on day 1 and subsequent subcutaneous maintenance dosing form the next dose onwards with the first maintenance dose timed so that the large majority of patients (>99%) maintained therapeutic exposure before the first maintenance dose (>75 μg/mL). Several weight‐based intravenous loading doses, rounded to whole‐vial amounts, were evaluated with two simulation‐derived weight categories (<75 kg and ≥75 kg).

2.2. Maintenance phase

For the maintenance phase, two dosing strategies were developed to ensure applicability in regions with and without access to therapeutic drug monitoring (TDM). (1) Weight‐based dosing strategy: the maintenance regimen was optimized by extending the dosing interval based on body weight, using the same simulation‐derived weight categories as applied for the loading dose assessment. We evaluated several interval extension scenarios in the simulations, and the selected scheme provided target attainment comparable to the standard‐of‐care dosing regimen. All patients received the same fixed 680 mg subcutaneous dose and only the dosing interval was modified. (2) TDM‐guided dosing strategy: The weight‐based dosing strategy in the maintenance phase was further adjusted based on crovalimab trough levels measured after the third dose. The dose interval was shortened by 1 week if the trough concentration was below the efficacy target of <75 μg/mL and extended by 2 weeks if it was >200 μg/mL (Table 1). These simulations were designed as a pragmatic approach and interval adjustments were, therefore, rounded to whole weeks. The threshold of 200 mg/L was selected after evaluating several potential cut‐offs and represented the lowest value that allowed interval extension while maintaining exposure comparable to the standard‐of‐care regimen (data not shown). For this scenario, first a simulation was performed using the weight‐based dosing strategy. Individually predicted pharmacokinetic parameters were extracted from the simulation, as well as the trough concentration, including residual error, at the third dose. A new simulation was then performed using the TDM‐based dosing schedule and the individually predicated pharmacokinetic parameters to determine the pharmacokinetics after TDM.

2.3. In silico evaluation

For this in silico evaluation the virtual population was derived from 1000 individuals in the publicly available NHANES database, with the following demographics (median, interquartile range): age: 45 (31–59) years; weight: 73 (62–82) kg; sex: 51% female. ¹⁸

All simulations were performed with the software package NONMEM v7.5 (Icon PLC, Dublin, Ireland) with and Pirana 2.9.9 as a graphical interface. ¹⁹ Data preparation and generation of the virtual population were performed in R 4.4.3 using RStudio 2025.05.01 as an interface. ¹⁹ , ²⁰ , ²¹

For each dosing scenario, we predicted the crovalimab concentrations during the first year of treatment. Furthermore, we calculated the percentage of patients with subtherapeutic crovalimab concentrations after the loading dose phase and during steady state. We then also calculated the fraction of patients on each dosing interval in the TDM‐based strategy group. Lastly, we calculated the average annual drug cost of crovalimab per patient, assuming costs of £9500 (excluding VAT) per 340 mg vial ²² and the potential drug costs savings.

2.4. Justification of the therapeutic threshold for crovalimab

The concentration of crovalimab leading to 90% of maximal reduction of LDH synthesis was previously estimated to be 34.9 μg/mL (90% CI, 15.9–67.5 μg/mL). ¹⁵ Conservatively, we defined a crovalimab trough concentration <75 μg/mL as subtherapeutic, as this corresponds approximately to the concentration required to achieve 95% of maximal LDH reduction (EC₉₅ ≈ 73.4 μg/mL, assuming a Hill‐coefficient of 1). Although a target trough concentration of 100 μg/mL has been postulated, previous clinical and pharmacokinetic analyses demonstrate that crovalimab concentrations between 50 and 100 μg/mL still effectively inhibit complement activity and maintain LDH within controlled levels, ¹⁴ , ¹⁷ aligning well with the chosen therapeutic threshold of >75 μg/mL.

3. RESULTS

In our evaluation, therapy was initiated with a weight‐based intravenous loading dose of 1020 mg for patients <75 kg and 1360 mg for patients ≥75 kg, followed by the subcutaneous maintenance dose of 680 mg on day 14. With this strategy, target attainment was achieved in 99% of the patients at the start of the maintenance phase, indicating effective target attainment with this strategy. The alternative loading dose regimen eliminated the need for weekly 340 mg subcutaneous crovalimab during the loading phase, simplifying the dosing strategy without compromising therapeutic exposure.

For the weight‐based maintenance strategy, the dosing interval was 5 weeks for patients <75 kg and 4 weeks for patients ≥75 kg. In the TDM scenario, the dosing interval could be prolonged in 24% of the patients, with a mean (SD) dosing interval of 5.0 (1.0) weeks.

The predicted concentrations of crovalimab for each dosing regimen at steady state are shown in Figure 1, and the concentration–time curves of crovalimab for each dosing regimen are shown in Figure 2.

FIGURE 1.

Tukey box‐and‐whisker plots for crovalimab trough concentrations at a single time point during steady state. From left to right: approved dosing regimen, alternative: weight‐based dosing regimen; and alternative: therapeutic drug monitoring (TDM)‐based dosing regimen. The dashed line represents a crovalimab concentration of 75 μg/mL (upper threshold for efficacy).

FIGURE 2.

Predicted concentration‐time profiles for crovalimab for each dosing approach. The upper panel (A) displays the concentration‐time profile for standard of care, the middle panel (B) displays the weight‐based dosing regimen and the lower panel (C) displays the TDM‐based dosing regimen. The dashed red line reflects the efficacy target of 75 μg/mL as defined by the manufacturer. The solid curve represents the median of the simulated population. The shaded areas represent the 50% (dark blue) and 90% (light blue) prediction intervals.

As illustrated in Figure 1, crovalimab concentrations during the maintenance phase were generally supratherapeutic, with substantial interindividual variability. Compared with the approved regimen, the weight‐based dosing strategy reduced interindividual variability, and the TDM‐guided regimen further reduced interindividual variability, with both strategies maintaining therapeutic exposure.

The percentage of patients with predicted trough levels below 75 μg/mL at steady state was 2.2% with the standard dosing regimen and was equivalent with fractions of 2.4% with the weight‐based dosing strategy and 2.2% when combining weight‐based dosing with TDM‐based dosing. The percentage of patients with predicted trough levels above 200 μg/mL at steady state was 57.6% with the standard dosing regimen, 44.0% with weight‐based dosing and 26.2% with TDM‐based dosing.

Mean annual drug costs per patient, excluding VAT, were £318 000 for the standard regimen, £258 000 for the weight‐based strategy (19% reduction), and £245 000 for TDM‐guided dosing (23% reduction), demonstrating not only the potential of patient benefit but also the economic benefit of individualized dosing.

3.1. Nomenclature of targets and ligands

Key protein targets and ligands in this article are hyperlinked to corresponding entries in http://www.guidetopharmacology.org and are permanently archived in the Concise Guide to PHARMACOLOGY 2021/22. ²³

4. DISCUSSION AND CONCLUSION

In our analysis, we explored the benefits of alternative dosing regimens for crovalimab in PNH patients. We predicted that with the use of an individualized strategy, dosing intervals of crovalimab can be extended in 24% of the patients, while maintaining therapeutic exposure and drug costs reductions up to 23% are achievable, translating to a reduction of £73 000 per patient per year. This approach is expected to ensure therapeutic effectiveness and to enhance patient‐friendliness, while yielding significant costs savings.

An interesting finding from our analysis is that the approved loading dose for crovalimab can be made more patient friendly at a lower cumulative dose (a single intravenous dose followed by the maintenance dose 14 days thereafter) This adjustment ensures effective target attainment from the start while simplifying the dosing strategy. The findings from treatment‐naïve patients cannot be directly extrapolated to individuals previously treated with other C5‐inhibitors. In this population, dose adjustments may be necessary due to the risk of type III hypersensitivity reactions. ¹⁷ We did not explore a scenario in which the standard loading dose is followed by either weight‐based or TDM‐based dosing because of the benefits of our loading dose strategy.

All approved complement inhibitors have a clear exposure‐response relationship and target drug concentrations can be relatively easily identified from drug approval data. ¹⁷ , ²⁴ , ²⁵ , ²⁶ , ²⁷ Given that exposure to these agents generally greatly exceeds the therapeutic targets, model‐informed precision dosing (MIPD) represents a promising strategy to optimize treatment by enabling dosing interval prolongation and reducing drug costs, while maintaining sustained therapeutic efficacy. Pharmacokinetically‐guided dosing not only prevents unnecessary high exposure but also enables the identification of the smaller subset of patients who would otherwise remain underexposed, ensuring they receive effective therapy. We have previously demonstrated through clinical studies and simulation studies that individualized dosing regimens based on drug levels for eculizumab, ravulizumab and pegcetacoplan can maintain comparable efficacy while enabling longer dosing intervals and reducing drug costs. ²⁸ , ²⁹ , ³⁰

In our analysis, we introduced a pragmatic TDM based dosing strategy with only one TDM‐based dose adjustment which was restricted to whole week steps. In theory, even more personalized (i.e., intervals rounded to days) dosing intervals could yield larger theoretical savings and serial measurements and dose adaptations may improve target attainment. When implementing TDM‐guided dosing, it should be noted that implementation of an assay and expertise for interpretation are a prerequisite The drug approval data of crovalimab describe a suitable assay that was used during clinical development of crovalimab. ¹⁷ Based on recent pharmacoeconomic analyses of monoclonal antibody TDM, the combined costs of ELISA‐based testing, laboratory staff time and clinical interpretation amount to approximately £50 per sample. It should be noted that effective implementation of TDM in clinical practice, especially for rare diseases, relies on an appropriate framework, including centralized laboratories with validated analytical platforms, as well as adequately trained staff to ensure reliable sample handling, timely analysis and clinical interpretation, which might be challenging. Given the clear benefits of such an approach for both patients and society, establishing and maintaining this infrastructure is highly valuable. Because crovalimab is self‐administered subcutaneously in an outpatient setting, TDM does not alter the number of hospital visits or the intensity of clinical monitoring. As a result, there are no expected changes in healthcare utilization that contribute to costs or savings. The economic impact, therefore, depends almost exclusively on two components: the cost of TDM (~£50 per test) ³¹ and the cost of the drug itself. With crovalimab priced at approximately £19 000 per administration, even modest reductions in drug use would generate savings that far exceed the costs of TDM. Thus, for crovalimab, dose reduction is the primary driver of cost reduction. In case TDM is not available, our proposed weight‐based dosing algorithm provides a pragmatic approach for dose interval prolongation.

Prospective studies should help to confirm the suitability and efficacy of the proposed crovalimab dosing regimens, thereby supporting broader implementation in clinical practice. Furthermore, the use of TDM can help prevent subtherapeutic exposure, particularly during periods of altered pharmacokinetics, for example, during pregnancy or in case of augmented complement synthesis, like infections. In settings where pharmacokinetic monitoring may not be available, weight‐based dosing may be an attractive alternative to TDM‐based dosing where intensified CH50 monitoring can be employed to monitor the efficacy of this strategy.

Overall, our results demonstrate the potential for a personalized crovalimab regimen, simplifying the loading phase, enabling dose interval prolongation in most patients and generating drug cost reductions up to 23%.

AUTHOR CONTRIBUTIONS

M.t.A., S.E.K., and R.t.H. designed the study. M.t.A. and S.E.K. analyzed and interpreted the data, and M.t.A. and R.t.H. wrote the manuscript. All authors reviewed drafts of the manuscript and approved the final version.

CONFLICT OF INTEREST STATEMENT

NvdK received consultancy fees from Alexion, Sobi, Samsung, Novartis, and Roche Pharmaceuticals and is a sub‐investigator in the APL2‐C3G trial, Apellis. Rob ter Heine has received consulting fees from Samsung Bioepis (paid to the institution). The remaining authors declare that they have no conflicts of interest to disclose.

ter Avest M, Koele SE, van de Kar NCAJ, Langemeijer SMC, ter Heine R. Exploring individualized crovalimab dosing in PNH through in silico modelling: Potential for improved convenience and cost efficiency. Br J Clin Pharmacol. 2026;92(10):3460‐3465. doi: 10.1002/bcp.70651

DATA AVAILABILITY STATEMENT

All relevant data were generated by means of computer simulations. The model code is available upon request.