Pharmacokinetic Comparability and Safety Between Original and Citrate-Free Mirikizumab Formulations for Subcutaneous Injections: Results from Three Clinical Trials

Abstract

Introduction

Mirikizumab, a p19-directed antibody against interleukin-23 (IL-23), is administered by subcutaneous (SC) injection. Injection site pain (ISP) associated with citrate buffers may negatively affect patient adherence to SC-administered treatments. We assessed the bioequivalence and safety of the citrate-free (CF) and original formulations of mirikizumab.

Methods

The formulations were assessed in three phase 1, two-arm, randomized, single-dose, parallel design studies in healthy participants: study A (NCT04548219), study B (NCT05515601), and study C (NCT05644353). Participants were randomized 1:1 to either formulation, then further randomized to injection site locations of abdomen, arm, or thigh. The relative bioavailability (RBA) study A had a primary objective of assessing the RBA of a single 200 mg dose. Bioequivalence (BE) studies B and C had the primary objective of assessing the BE of a 200 and 300 mg dose, respectively. In all studies, the primary endpoints were C_max, AUC(0–∞), and AUC(0–t_last). The secondary objective was to assess safety and tolerability by treatment-emergent adverse events and serious adverse events. In study A, ISP was quantified prospectively using the 100-mm validated visual analogue scale (VAS) assessment form.

Results

The primary objective was met in all studies. The RBA study found no significant difference in exposure between the formulations. BE was demonstrated between CF and original mirikizumab in both BE studies, with the 90% confidence intervals of the ratios of geometric least squares means within the pre-specified equivalence limits of 0.80 and 1.25. The frequency of ISP and injection site reactions (ISRs) was lower for CF than original mirikizumab in all studies. Furthermore, a significant difference in mean VAS score was observed in study A.

Conclusion

Mirikizumab CF and original formulations were bioequivalent. The CF formulation was associated with less pain and fewer ISRs, with no other notable differences in safety profiles.

Trial Registration

ClinicalTrials.gov identifier NCT04548219, NCT05515601, NCT05644353.

Supplementary Information

The online version contains supplementary material available at 10.1007/s12325-025-03158-y.

Key Summary Points

Why carry out this study?

Injection site pain associated with citrate buffers may negatively affect the patient experience with biologic treatments.

The original commercial formulation of mirikizumab was compared to a citrate-free formulation in an initial relative bioavailability study and two subsequent bioequivalence studies.

What was learned from the study?

The citrate-free and original formulations were proven to be bioequivalent in the two bioequivalence studies.

Although the two formulations had similar safety profiles, the citrate-free formulation was associated with fewer injection site reactions across the three studies, in addition to lower visual analogue scale (VAS) rated pain in the relative bioavailability study.

Introduction

Subcutaneous (SC) injection is the most frequently used method for administering biologics for several reasons, including ease of injection compared to intravenous or intramuscular administration, and potential for rapid onset of action [1]. However, injection site pain (ISP) associated with SC-administered medications can negatively affect the patient experience, increase fear and stress associated with dosing, and affect patient adherence [2]. ISP can be influenced by numerous factors including formulation parameters, delivery factors, and device parameters. ISP can be strongly affected by solution composition and formulation buffer excipients, with both citrate and sodium chloride (NaCl) previously shown to play a strong role [2]. Citrate-free (CF) formulations have been associated with less injection-associated pain when administering biologic drugs [3, 4].

Mirikizumab is a humanized immunoglobulin G4 (IgG4) monoclonal antibody which selectively binds to the p19 subunit of human interleukin-23 (IL-23). Following an initial induction phase via intravenous infusion, maintenance doses of mirikizumab are administered by SC injection from week 12 [5, 6]. Mirikizumab is currently approved for the treatment of ulcerative colitis (UC) in several major geographies, including the USA, the European Union, the UK, and Japan, and it is approved for the treatment of moderately to severely active Crohn’s disease (CD) in the USA and the European Union. Although mirikizumab has a well-accepted, long-term safety profile [6], injection site reactions (ISRs) were reported in 8.7% of adult patients with moderate-to-severe UC during the mirikizumab phase 3 maintenance trial. In addition, 4.4% of patients experienced ISP during the maintenance trial [7]. Separately, in a phase 3 trial studying mirikizumab in patients with moderately-to-severely active CD, the frequency of reported ISRs reached 10.8% [8].

A CF formulation of mirikizumab, containing no citrate buffer and a reduced concentration of NaCl, was developed to reduce the incidence and severity of ISP. Here, we report the results of three phase 1 studies comparing the original commercial formulation and the CF formulation of mirikizumab: one pilot study designed to evaluate the relative bioavailability (RBA) of both formulations, and two pivotal, registrational bioequivalence (BE) studies. All studies assessed the safety and tolerability of both formulations. The primary objective of these studies is to show the pharmacokinetic (PK) comparability of these two formulations in terms of PK parameters such as maximum observed drug concentration (C_max) and area under the concentration versus time curve (AUC) to abide with the regulatory guidance. The secondary objective of these studies is to show the safety and tolerability in terms of visual analogue scale (VAS) pain scale and treatment-emergent adverse events (TEAEs).

Methods

Study Design

All three studies were phase 1, two-arm, randomized, single-dose, participant-blind, parallel design studies in healthy participants. All studies were designed to assess the PK and safety of the original and CF formulations of mirikizumab.

Table 1 summarizes the main elements of the study designs for the three studies, including participant parameters, formulation randomization, and dosing. Study A (I6T-MC-AMBV, ClinicalTrials.gov identifier NCT04548219) was an investigator-blind RBA study for 200 mg mirikizumab which informed the study design for the later BE studies. Study B (I6T-MC-AMBT, ClinicalTrials.gov identifier NCT05515601) and study C (I6T-MC-AMBY, ClinicalTrials.gov identifier NCT05644353) were BE studies for 200 mg and 300 mg mirikizumab, respectively. The 200 mg and 300 mg doses are proposed for UC and CD, respectively. In studies B and C, participants may have had chronic, stable medical conditions which, in the investigator’s opinion, did not pose an increased risk to the participant. Participants were screened for 28 days (study A) or 35 days (studies B and C) prior to day 1 of the study. In all studies, participants completed 4-h post-dose safety assessments on day 1, and returned to the clinical research unit for PK sampling, immunogenicity sampling, and safety assessments at pre-specified times up to 12 weeks post dose. Table S1 (supplementary material) details the components of the CF and original mirikizumab formulations used in the studies.

Table 1.

Study design summary

Study	Subjects	Age (years)	BMI (kg/m2)	Weight stratification	Formulation randomization	Injection site location randomization	Total mirikizumab dose
A	Healthy male or female participants	18–75	18.0–32.0	NA	1:1 original:CF	1:1:1 abdomen:arm:thigh	200 mg: two 1-mL injections of 100 mg/mL mirikizumab via PFS
B	Healthy male or female participants	18–65	18.0–34.0	Either < 75.0 kg, 75.0–85.0 kg, or > 85.0 kg	1:1 original:CF	Approximately equal distribution of abdomen, arm, and thigh	200 mg: two 1-mL injections of 100 mg/mL mirikizumab via an AI
C	Healthy male or female participants	18–65	18.0–34.0	Either < 75.0 kg, 75.0–85.0 kg, or > 85.0 kg	1:1 original:CF	Approximately equal distribution of abdomen, arm, and thigh	300 mg: one 1-mL and one 2-mL dose of 100 mg/mL mirikizumab via PFS

All doses delivered subcutaneously. For studies B and C, weight stratification occurred on study day 1 at the clinical research unit. Participants were randomized to formulation and injection site location using an interactive web-response system

AI autoinjector, BMI body mass index, CF citrate-free, NA not applicable, PFS pre-filled syringe

Exclusion criteria included (a) significant allergies to humanized monoclonal antibodies or mirikizumab, (b) prior or concomitant therapies such as anti-IL-12p40 or anti-IL-23p19 antibody therapy, (c) current enrollment in an incompatible clinical study or prior study participation within the last 30 days or 5 half-lives, whichever is longer, increasing to at least 3 months or 5 half-lives for biologic agents, (d) incompatible diagnostic assessment such as prior severe infections within 12 weeks, and (e) other relevant exclusions such as lactation or pregnancy.

Ethics/Ethical Approval

The studies were conducted in accordance with the study protocols and with consensus ethics principles derived from international ethics guidelines, including the Declaration of Helsinki and Council for International Organizations of Medical Sciences International Ethical Guidelines, applicable International Council for Harmonisation Good Clinical Practice Guidelines, and applicable laws and regulations. The study protocols were approved by the institutional review boards before all studies were initiated. Ethical review boards were in place for all studies: the WCG – Midlands Independent Review Board and WCG Independent Review Board were the overall (“master”) ethics committees for study A, and the Salus Institutional Review Board was the overall (“master”) ethics committee for studies B and C. Informed consent forms, required to provide details about why the research was being done and what it would involve, were approved by the respective review boards and provided to participants. Subsequent dated written informed consent was required by each participant before the participant could participate in the study. The informed consent forms also noted that study data which did not directly identify participants could be published.

Objectives

Table 2 outlines the study objectives and endpoints. Endpoints were identical in all studies.

Table 2.

Study objectives and endpoints

Study	Objectives	Endpoints
A	Primary: evaluate the relative bioavailability of a single 200 mg SC dose (2 × 1-mL PFS injections) of mirikizumab CF compared to the original formulation	Cmax, AUC(0–∞), and AUC(0–tlast)
	Secondary: evaluate the safety and tolerability of a single 200 mg SC dose (2 × 1-mL PFS injections) of mirikizumab CF compared to the original formulation	TEAEs and SAEs
B	Primary: evaluate the bioequivalence of a single 200 mg SC dose of mirikizumab CF compared to the original formulation using a 2 × 1-mL AI	Cmax, AUC(0–∞), and AUC(0–tlast)
	Secondary: evaluate the safety and tolerability of a single 200 mg SC dose of mirikizumab CF compared to the original formulation using a 2 × 1-mL AI	TEAEs and SAEs
C	Primary: evaluate the bioequivalence of a single 300 mg SC dose of mirikizumab CF compared to original formulation using a 1-mL and 2-mL PFS	Cmax, AUC(0–∞), and AUC(0–tlast)
	Secondary: evaluate the safety and tolerability of a single 300 mg SC dose of mirikizumab CF compared to the original formulation using a 1-mL and 2-mL PFS	TEAEs and SAEs

AI autoinjector, AUC(0–∞) area under the concentration versus time curve from time zero to infinity, AUC(0–t_last) area under the concentration versus time curve from time zero to time t, where t is the last time point with a measurable concentration, CF citrate-free, C_max maximum concentration, PFS pre-filled syringe, SAE serious adverse event, SC subcutaneous, TEAE treatment-emergent adverse event

Statistical Analyses

For continuous data, summary statistics included the arithmetic mean and arithmetic standard deviation (SD). For categorical data, frequency count and percentages were presented. Summary statistics were presented by treatment formulation and injection site. A combination of SAS (version 9.4. or greater; SAS Institute Inc. Cary, NC, USA) and R (version 4.0.0 or greater; R Foundation for Statistical Computing, Vienna, Austria) was used for the statistical analyses.

Pharmacokinetic Analyses

The PK population consisted of all enrolled participants who received a dose of mirikizumab and who had evaluable PK data. Figure S1 shows that dosing and PK sampling times were the same across all studies. PK samples were collected up to day 85 ± 3 post dose in the absence of early discontinuation.

PK data were subject to data handling considerations. In study A, PK data exceeding the protocol-specified windows were excluded from the mean concentration–time profiles. Positive pre-dose samples and anomalous below the lower limit of quantitation post-dose concentrations were excluded from the PK analysis and mean concentration–time profiles. Study A was a small RBA study with insufficient sample size to achieve BE. Studies B and C were larger studies designed to have approximately 90% power to achieve BE.

Some subjects were excluded from PK statistical analysis because of insufficient concentration data for estimating PK parameters. For study A, no subjects were excluded, with a total of 60 subjects in the analysis. For study B, 8 subjects were excluded as a result of the area under the concentration versus time curve from time zero to time t, where t is the last time point with a measurable concentration, AUC(0–t_last), not meeting the minimum requirement of three consecutive concentrations above the lower limit of quantitation, with at least one concentration following C_max, due to missing mirikizumab concentration values. This resulted in a total of 388 subjects in the analysis. For study C, 34 subjects were excluded because of insufficient concentration data (1 participant with only a day 1 pre-dose sample and 2 participants who had insufficient concentration data were excluded) and incorrectly administered doses (17 participants receiving 300 mg original mirikizumab and 14 receiving 300 mg CF mirikizumab), with a total of 416 subjects in the analysis.

The C_max, area under the concentration versus time curve from time zero to infinity, AUC(0–∞), and AUC(0–t_last) were log-transformed and were evaluated in the linear fixed-effects model, with mirikizumab formulation (original vs CF) and injection location as fixed-effects. The least squares mean (LSM) differences between the CF and original formulations were backtransformed to present the ratios of geometric LSM and the corresponding 90% confidence interval (CI). Bioequivalence was concluded if this 90% CI falls within a pre-specified interval of 0.80 to 1.25. The time of maximum observed drug concentration (t_max) was analyzed using a Wilcoxon rank sum test. Estimates of the median difference based on the observed means, 90% CI, and p values from Wilcoxon rank sum test were calculated. Comparisons were also made for three injection locations, with all possible comparisons. The analysis assessing the difference by injection location was not statistically powered, so formal bioequivalence for three injection locations was not determined. PK parameter estimates were determined using non-compartmental procedures in validated software program (Phoenix WinNonlin Version 8.1. or later) for the serum concentration of mirikizumab.

Safety Analyses

The safety population, which consisted of all participants who were exposed to mirikizumab, was used when reporting summary statistics for both safety and demographics. Safety assessments included clinical laboratory parameters, vital signs assessments, electrocardiogram, hepatic monitoring, and immunogenicity assessments.

All adverse events (AE) were listed. TEAEs were summarized related to study treatment and severity. The number of AEs, the number of participants experiencing an AE, and the percentage of participants experiencing an AE were summarized by treatment, using the Medical Dictionary for Regulatory Activities (MedDRA) preferred term.

ISRs and ISP were assessed differently for study A compared to studies B and C. For study A, prospective assessments were conducted, with spontaneous subject-reported ISRs also recorded. Events relating to injection site were captured as a study exploratory endpoint and were not recorded as an AE unless that event was a serious adverse event (SAE). For studies B and C, there was no prospective collection of ISR information; however, spontaneously reported ISRs were recorded as AEs. ISR findings for a specific injection were captured as a single AE, with severity recorded as the highest severity. Frequency and type of ISR were also recorded.

In study A, VAS analyses were performed as follows: after the first injection, the investigator or designee quantified ISP for subjects at 1, 5, and 15 min using the 100-mm validated VAS Assessment form [9]. Additional scheduled timepoints were at 30, 60, 120, and 240 min. VAS pain assessment was not performed for pain reported spontaneously. In the event of reported ISP, pain intensity was quantified using the following VAS pain score categories: (a) no pain, VAS pain score of 0 mm, (b) mild pain, VAS pain score of 30 mm or less, (c) moderate pain, VAS pain score more than 30 mm and at most 70 mm, and (d) severe pain, VAS pain score more than 70 mm. Data were summarized by treatment, treatment subset by injection location, and timepoint. The 1-min post-injection pain score was also evaluated in a linear fixed-effects model with fixed-effects for treatment formulation and injection site location. The LSM, differences between the formulations, and corresponding 90% CI for the difference were presented.

Bioanalytical Methods

Mirikizumab concentrations in the serum samples collected during these studies were analyzed via a validated enzyme-linked immunosorbent assay. The validated assay range was 100 to 10,000 ng/mL, which comprised the lower and upper limits of quantitation, respectively. Quality control concentrations were 250, 1000, and 7500 ng/mL.

Results

Baseline Characteristics

Table 3 lists the baseline demographics for the three studies. Baseline characteristics were similar among studies and between treatment groups.

Table 3.

Baseline demographics

Parameter	Study A		Study B		Study C
	Original (N = 30)	CF (N = 30)	Original (N = 199)	CF (N = 197)	Original (N = 226)	CF (N = 224)
Age, years, mean (SD)	49.3 (15.6)	41.4 (16.1)	40.7 (13.0)	41.1 (12.2)	40.9 (12.7)	40.8 (12.4)
Sex, n (%)
Male	11 (36.7)	8 (26.7)	101 (50.8)	93 (47.2)	97 (42.9)	115 (51.3)
Female	19 (63.3)	22 (73.3)	98 (49.2)	104 (52.8)	129 (57.1)	109 (48.7)
Weight, kg, mean (SD)	75.5 (12.8)	78.2 (15.6)	78.86 (14.39)	78.71 (15.01)	77.14 (12.93)	77.31 (12.67)
BMI, kg/m2, mean (SD)	26.3 (3.1)	27.1 (3.6)	27.01 (3.51)	26.94 (3.59)	27.64 (3.52)	27.17 (3.37)
Race, n (%)
American Indian or Alaska Native	0 (0.0)	1 (3.3)	1 (0.5)	2 (1.0)	2 (0.9)	2 (0.9)
Asian	0 (0.0)	0 (0.0)	11 (5.5)	6 (3.0)	3 (1.3)	7 (3.1)
Black or African American	1 (3.3)	0 (0.0)	29 (14.6)	30 (15.2)	33 (14.6)	34 (15.2)
Native Hawaiian or Other Pacific Islander	0 (0.0)	0 (0.0)	1 (0.5)	1 (0.5)	0 (0.0)	0 (0.0)
White	29 (96.7)	29 (96.7)	150 (75.4)	152 (77.2)	180 (79.6)	173 (77.2)
Multiple	NA	NA	7 (3.5)	6 (3.0)	4 (1.8)	2 (0.9)
Not reported	NA	NA	NA	NA	4 (1.8)	6 (2.7)
Injection site location
Abdomen, n	10	10	63	66	74	75
Arm, n	10	10	69	66	76	75
Thigh, n	10	10	67	65	76	74

Includes participants from the safety population

BMI body mass index, CF citrate-free, N number of participants, n number of participants with ≥ 1 event, NA not applicable, SD standard deviation

Pharmacokinetics

Table S2 shows that there was no statistically significant difference in exposure to mirikizumab between the two formulations in study A, as assessed by AUC(0–t_last), AUC(0–∞), and C_max. The 90% CIs for the ratios of geometric LSM included unity for each comparison. The point estimates of the geometric mean ratio were within the 0.80 to 1.25 limits. Although study A was an RBA study, which was not powered to demonstrate BE, the results were useful for informing the later BE studies. Figure S2 shows that the mean serum concentration profiles were similar for the two formulations. Table S3 also shows no statistically significant difference in t_max between the two formulations.

The primary objectives were met for the BE studies, with bioequivalence demonstrated between the original and CF formulations in both studies B and C. Table 4 demonstrates similarity between AUC(0–t_last), AUC(0–∞), and C_max for both formulations. In both studies, bioequivalence was demonstrated between the formulations as the 90% CIs of the ratios of geometric LSM were within the pre-specified equivalence limits of 0.80 and 1.25. Figure 1 shows similarity between mean serum concentration profiles for both formulations across both studies. Table 5 shows that there were no statistically significant differences in t_max between CF and original mirikizumab.

Table 4.

Statistical summary of the pharmacokinetic parameters of mirikizumab in studies B and C

Parameter	Formulation	n	Generalized LSM	Ratio of generalized LSM (CF:original)	90% CI for the ratio (lower, upper)
Study B
AUC(0–tlast) (μg·day/mL)	Original	193	222
	CF	187	227	1.02	(0.960, 1.09)
AUC(0–∞) (μg·day/mL)	Original	195	226
	CF	192	228	1.01	(0.948, 1.08)
Cmax (μg/mL)	Original	196	12.9
	CF	192	12.7	0.986	(0.929, 1.05)
Study C
AUC(0–tlast) (μg·day/mL)	Original	205	320
	CF	208	313	0.979	(0.926, 1.03)
AUC(0–∞) (μg·day/mL)	Original	205	325
	CF	208	318	0.980	(0.927, 1.04)
Cmax (μg/mL)	Original	207	20.5
	CF	208	20.5	0.998	(0.939, 1.06)

Study B: 2 × 1-mL (200 mg mirikizumab) original or CF formulation via SC AI; study C: 1 × 1-mL (100 mg mirikizumab) and 1 × 2-mL (200 mg mirikizumab) original or CF formulation via SC PFS. Model: log(PK) = solution formulation + injection location + weight stratification + random error. Population: pharmacokinetic

AI autoinjector, AUC(0–∞) area under the concentration versus time curve from time zero to infinity, AUC(0–t_last) area under the concentration versus time curve from time zero to time t, where t is the last time point with a measurable concentration, CF citrate-free, CI confidence interval, C_max maximum observed drug concentration, log(PK) log transformed pharmacokinetic parameters, LSM least squares mean, n number of participants per treatment subgroup, PFS pre-filled syringe, SC subcutaneous

Fig. 1.

Arithmetic mean serum concentration profiles of mirikizumab for study B and study C. Data shown as arithmetic mean with one-sided SD. a Study B: linear scale. b Study C: linear scale. c Study B: semi-logarithmic. d Study C: semi-logarithmic. CF citrate-free, h hour, miri mirikizumab, SD standard deviation

Table 5.

Statistical summary of the t_max of mirikizumab in studies B and C

Parameter	Formulation	n	Median	Median of differences (CF − original)	Approx. 90% CI for the difference (lower, upper)	p value
Study B
tmax (day)	Original	196	3.98
	CF	192	3.98	0	(0.00, 0.02)	0.4011
Study C
tmax (day)	Original	207	3.98
	CF	208	3.97	− 0.01	(− 0.03, 0.00)	0.2113

t_max was analyzed using the procedure PROC NPAR1WAY and p values reported from the Wilcoxon rank sum test. Population: pharmacokinetic

CF citrate-free, CI confidence interval, n number of participants per treatment subgroup, t_max time of maximum observed drug concentration

The studies were not powered for comparisons by injection site location. However, Table S4 shows that the geometric LSM for AUC(0–t_last), AUC(0–∞), and C_max were similar between both formulations for each separate injection site location. The 90% CIs of the ratios of geometric LSM were again contained within the equivalence limits.

Safety

Table 6 shows similar safety profiles in terms of reported TEAEs for the original and CF formulations for study A, with no difference in the number or severity of TEAEs. Table 7 shows a lower overall incidence of TEAEs for the CF formulation compared to the original for studies B and C, with a larger difference observed for study C. The majority of TEAEs were mild in severity for studies B and C. There were no SAEs or discontinuations due to AEs reported in any of the reported studies. The higher incidence of treatment-related TEAEs for the original formulation, as observed for studies B and C, was deemed mainly due to the higher incidence of ISRs associated with the original formulation.

Table 6.

Summary of TEAEs for study A, all causalities

	Original (N = 30)	CF (N = 30)
Number of subjects with ≥ 1 TEAE, n (%)	3 (10.0)	3 (10.0)
TEAEs by severity, n (%)
Mild	1 (3.3)	1 (3.3)
Moderate	2 (6.7)	2 (6.7)
Severe	0 (0.0)	0 (0.0)
Subjects with ≥ 1 SAE, n (%)	0 (0.0)	0 (0.0)
Subjects with ≥ 1 TEAE related to study treatment, n (%)	2 (6.7)	1 (3.3)
ISRs, n (%)	23 (76.7)	15 (50.0)
AEs leading to discontinuation from study	0	0

When categorizing by severity, only the most severe TEAE for each participant is shown. ISRs were not included in the number of TEAEs. Population: safety

AE adverse event, CF citrate-free, ISR injection site reaction, N number of participants, n number of participants with ≥ 1 event, SAE serious adverse event, TEAE treatment-emergent adverse event

Table 7.

Summary of TEAEs for studies B and C, all causalities

	Study B		Study C
	Original (N = 199)	CF (N = 197)	Original (N = 226)	CF (N = 224)
Number of subjects with ≥ 1 TEAE, n (%)	59 (29.6)	57 (28.9)	110 (48.7)	70 (31.3)
TEAEs by severity, n (%)
Mild	52 (26.1)	41 (20.8)	93 (41.2)	57 (25.4)
Moderate	6 (3.0)	16 (8.1)	17 (7.5)	12 (5.4)
Severe	1 (0.5)	0 (0.0)	0 (0.0)	1 (0.4)
Subjects with ≥ 1 SAE, n (%)	0 (0.0)	0 (0.0)	0 (0.0)	0 (0.0)
Subjects with ≥ 1 TEAE related to study treatment, n (%)	32 (16.1)	18 (9.1)	75 (33.2)	36 (16.1)
AEs recorded as ISRs, n (%)	32 (16.1)	19 (9.6)	80 (35.4)	30 (13.4)
AEs leading to discontinuation from study	0	0	0	0

When categorizing by severity, only the most severe TEAE for each participant is shown. ISRs were included in the number of TEAEs. Population: safety

AE adverse event, CF citrate-free, ISR injection site reaction, N number of participants, n number of participants with ≥ 1 event, SAE serious adverse event, TEAE treatment-emergent adverse event

Table 8 shows the breakdown summary of the types of ISRs reported across the three studies. The number of participants reporting one or more ISRs, as well as the total number of ISRs reported, was lower for the CF formulation than the original for all three studies. The number of participants who experienced ISP was also lower for CF mirikizumab in all studies. In general, there was a lower or equal proportion of participants experiencing the other subtypes of ISRs, those being injection site edema, erythema, induration, and pruritus, for participants treated with CF mirikizumab than for original mirikizumab, except for one case in study B for injection site induration, where one participant reported for CF, and no participants reported for original.

Table 8.

Summary of ISRs for all studies, all causalities

ISR parameter	Study A		Study B		Study C
	Original (N = 30)	CF (N = 30)	Original (N = 199)	CF (N = 197)	Original (N = 226)	CF (N = 224)
Number (%) of subjects reporting ≥ 1 ISR, n (%)	23 (76.7)	15 (50.0)	32 (16.1)	19 (9.6)	80 (35.4)	30 (13.4)
Injection site edema, n (%)	0	0	1 (0.5)	1 (0.5)	7 (3.1)	2 (0.9)
Injection site erythema, n (%)	9 (30.0)	6 (20.0)	1 (0.5)	1 (0.5)	11 (4.9)	4 (1.8)
Injection site induration, n (%)	1 (3.3)	0	0	1 (0.5)	2 (0.9)	1 (0.4)
Injection site pain, n (%)	22 (73.3)	11 (36.7)	32 (16.1)	16 (8.1)	74 (32.7)	26 (11.6)
Injection site pruritus, n (%)	1 (3.3)	0	1 (0.5)	0	8 (3.5)	2 (0.9)
Total number of reported ISRs	47	20	62	31	142	51

Population: safety

CF citrate-free, ISR injection site reaction, N number of participants, n number of participants with adverse events

Figure 2 shows reduced LSM VAS score for the CF formulation in study A. The LSM VAS pain score for the original formulation was 26.07 within 1 min post dose compared to 12.63 for the CF formulation. The difference in LSM, − 13.43, was statistically significant, with the 90% CIs of the difference in geometric LSM excluding unity. Table 9 shows that the majority of VAS pain reported between minutes 1 and 15 was mild in severity, with severe pain only reported by two subjects at 1 min, both of whom received the original formulation. Figure 3 highlights the similar difference observed in VAS pain for studies B and C. Fewer subjects were reported in each of the VAS pain categories for the CF formulation compared to the original formulation.

Fig. 2.

Difference in VAS pain score LSM at 1 min for study A. The difference of LSM is represented as CF − original. The VAS scores range from 0 mm (no pain) to 100 mm (worst imaginable pain). CF citrate-free, CI confidence interval, LSM least squares mean, miri mirikizumab, N number of participants per treatment subgroup, VAS visual analogue scale

Table 9.

Summary of VAS pain at different timepoints for study A

Presence of pain	Pain category	Formulation	1 min n (%)	5 min n (%)	15 min n (%)
No pain	No pain	Original	0 (0.0)	11 (36.7)	17 (56.7)
		CF	3 (10.0)	11 (36.7)	17 (56.7)
Pain	Mild	Original	21 (70.0)	17 (56.7)	13 (43.3)
		CF	25 (83.3)	19 (63.3)	13 (43.3)
	Moderate	Original	7 (23.3)	2 (6.7)	0 (0.0)
		CF	2 (6.7)	0 (0.0)	0 (0.0)
	Severe	Original	2 (6.7)	0 (0.0)	0 (0.0)
		CF	0 (0.0)	0 (0.0)	0 (0.0)

Data shown as n (%). Original formulation N = 30, CF formulation N = 30. Pain categories are defined as: no pain = VAS pain score = 0, mild pain = VAS pain score > 0 and ≤ 30, moderate pain = VAS pain score > 30 and ≤ 70, severe pain = VAS pain score > 70. At time points 30, 60, 120, and 240 min, VAS pain assessment occurred only if pain was reported as “yes” on the injection-site assessment form

CF citrate-free, N number of participants, n number of participants per category, VAS visual analogue scale

Fig. 3.

Difference in VAS pain categories for study B (a) and study C (b). Pain categories are defined as: no pain = VAS pain score = 0 mm, mild pain = VAS pain score > 0 mm and ≤ 30 mm, moderate pain = VAS pain score > 30 mm and ≤ 70 mm, severe pain = VAS pain score > 70 mm. Participants received two injections. Where a participant had an ISR of pain in both injections and the severities were different, the most severe pain score was summarized. Only participants who reported an ISR of pain are included in this graph. Population: safety AI autoinjector, CF citrate-free, ISR injection site reaction, N number of participants studied, n number of participants in specified category, SC subcutaneous, VAS visual analogue scale

Discussion

These studies demonstrate BE between the original and CF formulations of mirikizumab for both the 200 mg and 300 mg dose. Following the initial RBA study, the results from studies B and C showed that the 90% CIs for all comparisons were fully within the confidence limits of 0.80 and 1.25. In addition, although there was no formal requirement to show BE by injection site, the data also shows similarities between the two formulations across the injection site locations, with comparisons across injection site location within the equivalence limits for studies B and C.

These studies followed a standard parallel-group design for monoclonal antibodies as a result of the long half-lives and a potential to trigger immunogenicity if a crossover design was employed. Serum samples were collected up to day 85 post dose to account for the typical estimated half-life of 9.5 days for mirikizumab [10]. The intersubject variability in pharmacokinetics in these studies may have been somewhat mitigated by the body weight-based stratification employed during the study enrollment, since body weight is a statistically significant but not clinically relevant covariate that may influence mirikizumab pharmacokinetics [10].

The efficacy and safety of mirikizumab has previously been demonstrated for the original formulation in pivotal studies for both UC [11] and CD [8]. The original formulation has a good safety profile, and the label rates of ISR are based upon this formulation. In the studies presented here, we observed that the CF formulation of mirikizumab was associated with a lower total number of ISRs, a lower proportion of participants reporting ISRs, and a lower overall prevalence of ISP, for all three studies. Apart from ISRs and ISP, study A showed similar safety profiles for the original and CF formulations, with studies B and C revealing a lower overall incidence of TEAEs for the CF formulation compared to the original, with a larger difference observed for study C.

The CF formulation was also associated with an improved VAS pain score. In study A, a statistically significant difference in LSM of − 13.43 was reported at 1 min. Notably, a change of 10 mm for the 100-mm VAS pain score has previously been identified as a clinically important difference [12]. This is important given that ISP for SC products is a significant issue which can worsen patient experience, increase fear and stress associated with dosing, and affect patient adherence [2].

The reduction in ISP observed for the CF formulation has also been shown for other biologic treatments. A pharmacokinetic study on SB5, an approved adalimumab biosimilar, found a lower mean VAS score for a CF high-concentration formulation compared to original formulation in healthy male volunteers, with the largest difference immediately after injection (3.1 ± 4.98 mm CF, 9.2 ± 12.98 mm original) [13]. A separate study in pediatric patients with inflammatory bowel disease (IBD) showed significantly less ISP in the low-volume, CF formulation compared to the original adalimumab formulation, in addition to a significant mean difference in the patient-reported Faces Pain Scale-Revised VAS scale for pain perception [14]. Pain perception has been shown to negatively influence treatment adherence for patients with IBD who were administered the original, citrate-containing formulation of adalimumab by SC injection, as assessed by patient questionnaire. In this study, the low-volume, CF formulation of adalimumab considerably reduced the frequency of injection-related pain [15]. This may affect the patient experience, with a separate study showing a significant improvement in adherence and persistence for adalimumab-naïve patients treated with CF adalimumab compared to the original formulation [16]. These studies, and our own, show that the drug matrix is a major influencer of ISP and that changing components is a primary way of improving injection experience for the patient.

An advantage of the studies described here was that mirikizumab was administered by trained study personnel, which reduced the risk of potential confounding factors. An additional advantage was the large sample sizes of studies B and C, which led to a combined analysis population of over 800 participants from those two studies. Furthermore, conducting the studies in healthy participants provided the benefits of (a) fewer confounding variables among the study population, such as concomitant medications, and (b) allowed for the collection of PK samples for a prolonged period following a single dose of mirikizumab.

One potential limitation of conducting studies in healthy participants is that they should not be directly compared with studies performed in patients. However, the bioequivalence in PK demonstrated the CF formulation would be equally effective in a clinical setting. Although there are limitations assessing the effect of immunogenicity on PK as these were single-dose administration studies conducted in healthy populations, the BE of CF was clearly shown for both 200 mg and 300 mg dosages, providing useful clinical data for comparative evaluation between the CF and original formulations of mirikizumab. An additional limitation is that measurements of clinical efficacy were not collected as these studies were conducted in healthy volunteers. However, the equivalent PK observed across these studies suggest that the CF formulation would be equally effective. One limitation for safety analyses is that the exact VAS pain score was not calculated in millimeters for all three studies, but rather only pain categories were reported, for studies B and C.

Conclusion

The CF formulation of mirikizumab was bioequivalent with the original formulation. Both formulations had similar safety profiles; however, the CF formulation was associated with less pain and fewer ISRs than the original formulation, which may improve treatment experience for patients.

Supplementary Information

Below is the link to the electronic supplementary material.

Author Contributions

Yuki Otani: Analysis and interpretation of data, drafting the manuscript, and critical review of the work for important intellectual content. Brian G. Feagan: Interpretation of data, and critical review of the work for important intellectual content. Geert R. D’Haens: Interpretation of data, and critical review of the work for important intellectual content. Rodrigo Escobar: Interpretation of data, and critical review of the work for important intellectual content. Nathan J. Morris: Design of the work, analysis and interpretation of data for the work, and critical review of the work for important intellectual content. Christopher D. Payne: Conception and design of the work, interpretation of data, drafting the manuscript, and critical review of the work for important intellectual content. Michelle Ugolini Lopes: Conception and design of the work, interpretation of data, and critical review of the work for important intellectual content. Xin Zhang: Conception and design of the work, analysis and interpretation of data, drafting the manuscript, and critical review of the work for important intellectual content.

Funding

Eli Lilly and Company funded the studies, the Rapid Service Fee, and the Open Access fee.

Data Availability

Lilly provides access to all individual participant data collected during the trial, after anonymization, with the exception of pharmacokinetic or genetic data. Data are available to request 6 months after the indication studied has been approved in the US and EU and after primary publication acceptance, whichever is later. No expiration date of data requests is currently set once data are made available. Access is provided after a proposal has been approved by an independent review committee identified for this purpose and after receipt of a signed data sharing agreement. Data and documents, including the study protocol, statistical analysis plan, clinical study report, blank or annotated case report forms, will be provided in a secure data sharing environment. For details on submitting a request, see the instructions provided at http://www.vivli.org.

Data are also available on clinicaltrials.gov: NCT04548219, NCT05515601, NCT05644353.

Declarations

Conflict of Interest

Yuki Otani, Rodrigo Escobar, Nathan J. Morris, Christopher D. Payne, Michelle Ugolini Lopes, and Xin Zhang are employees and shareholders of Eli Lilly and Company. Brian G. Feagan has received consulting fees from AbbVie, Abivax, Adiso, AgomAB Therapeutics, Akros, Alira Health, Ally Bridge Group, AnaptysBio, Apini Therapeutics, Argenx, Avoro Capital Advisors, Belmore Law, BioFactura, BioJamp, Biora Therapeutics, Blackbird Laboratories, Boehringer Ingelheim, Boxer Capital, Celsius Therapeutics, Celgene/BMS, Celltrion, Clarivate, Connect BioPharma, Disc Medicine, Duality, EcoR1, Eli Lilly, Ensho Therapeutics, Evida, Enveda, Faes Farma, First Wave, Forbion, Galapagos, Galen Atlantica, Genentech/Roche, General Atlantic, Genesis Therapeutics, Gilead, Gossamer Pharma, GSK, Imhotex, ImmiDomics, Immunic Therapeutics, Intercept, Janssen, Japan Tobacco Inc., Klick Health, LifeMine Therapeutics, Mage Biologics, Merck, Mestag, Mirador Therapeutics, Mobius, Monte Rosa Tx, Morphic Therapeutics, Nexys Therapeutics, Nighthawk Therapeutics, Nimbus Therapeutics, Novartis, OncoC4, OrbiMed, Orphagen, Pendopharm, Pfizer, Protagonist,32 Bio, REDX, Roche, Roivant/Televant, Sanofi, Sobi, Sorriso, Spyre Therapeutics, Surrozen Inc., Sun Pharma, Synedgen, Takeda, Teva, Triastek, Trex Bio, TR1X Inc. TVM Lifesciences, Ventyx Biosciences, Versant Ventures, Vida Ventures, Zagbio; payment or honoraria for lectures, presentations, speakers bureaus, manuscript writing or educational events from Abbvie, Takeda, Janssen, Pfizer, Eli Lilly; payment for expert testimony from Belmore Law; participation on a Data Safety Monitoring Board or Advisory Board for AbbVie, AnaptysBio, Boehringer Ingelheim, Celgene/BMS, Eli Lilly, Genentech/Roche, Janssen, Merck, MiroBio, Origo BioPharma, Pfizer, REDX Pharma, Sanofi, Takeda, Teva, Ecor1Capital, Morphic, GSK; and stock options from Connect BioPharma, EnGene. Geert R. D’Haens has served as adviser and/or speaker for Abbvie, Alimentiv, Bristol Meiers Squibb, Boehringer Ingelheim, Celltrion, Eli Lilly, Galapagos, Glaxo Smith Kline, Immunic, Index Pharmaceuticals, Johnson and Johnson, Merck, Polpharm, Prometheus biosciences, Prometheus Laboratories, Procise diagnostics, Protagonist, Sandoz, Takeda, Tillotts, and Ventyx.

Ethics/Ethical Approval

The studies were conducted in accordance with the study protocols and with consensus ethics principles derived from international ethics guidelines, including the Declaration of Helsinki and Council for International Organizations of Medical Sciences International Ethical Guidelines, applicable International Council for Harmonisation Good Clinical Practice Guidelines, and applicable laws and regulations. The study protocols were approved by the institutional review boards before all studies were initiated. Ethical review boards were in place for all studies: the WCG – Midlands Independent Review Board and WCG Independent Review Board were the overall (“master”) ethics committees for study A, and the Salus Institutional Review Board was the overall (“master”) ethics committee for studies B and C. Informed consent forms, required to provide details about why the research was being done and what it would involve, were approved by the respective review boards and provided to participants. Subsequent dated written informed consent was required by each participant before the participant could participate in the study. The informed consent forms also noted that study data which did not directly identify participants could be published.