FormalPara Key Summary Points

Why carry out this study?

Greater exposure to secukinumab in real-world use adds to the safety observed in clinical trials and demonstrates favorable safety profile.

The objective of this analysis was to summarize the updated safety experience of secukinumab.

What was learned from the study?

Safety risks are described contextually with adverse events frequencies to aid in benefit–risk assessment while prescribing secukinumab.

Prescribers are informed about uncommon paradoxical skin reactions such as dyshidrotic eczema and pyoderma gangrenosum.

Introduction

Secukinumab is a fully human anti-interleukin (IL)-17A monoclonal antibody approved to treat moderate-to-severe plaque psoriasis (PsO), psoriatic arthritis (PsA), axial spondyloarthritis (axSpA) including ankylosing spondylitis (AS) and non-radiographic axSpA, pediatric PsO, juvenile psoriatic arthritis (JPsA), and enthesitis-related arthritis (ERA) in more than 100 countries [1], and was recently approved for moderate-to-severe hidradenitis suppurativa (HS) [2].

Modulation of proinflammatory cytokines, including IL-17A, with biologic therapies have potential safety implications, including infections and immune-related conditions [3, 4]. Due to multiorgan effects of IL-17 pathway activation, patients with certain underlying diseases may have comorbidities that can complicate causality analyses of the adverse events (AEs) reported for IL-17 inhibitors including secukinumab [5,6,7,8].

The safety of secukinumab has been published with respect to malignancy, latent tuberculosis, and inflammatory bowel disease (IBD) [9,10,11]. Long-term, 5-year, pooled safety data were reported from 28 clinical trials (CTs) involving 12,637 patients, along with post-marketing setting (PMS) exposure of 285,811 patient-years (PY) as of 25 December 2018 [12].

From 25 December 2018 through 25 June 2022, secukinumab post-marketing exposure more than quadrupled to 1,159,260 PY plus patients from CTs totaling 27,765 PY. The objective of this analysis was to summarize the updated safety experience of secukinumab [12].

Methods

AEs reported for secukinumab in PMS and pooled CTs were reviewed cumulatively through 25 June 2022 (unless otherwise specified).

  • PMS reflects safety monitoring by Novartis based on real-world evidence, is regulated by health authorities (HA), and is recorded in the Novartis safety database. PMS AEs can originate from healthcare professionals, patients, caregivers, literature, or social media. AEs received by HA and forwarded to Novartis are also included. Although details such as demographics, comorbidities, treatment indications, and dosing details may be lacking, all PMS AEs (irrespective of the reported indications) were analyzed.

  • AEs in CTs were pooled from 47 phase II/III/IV studies in adults administered secukinumab 150 mg and/or 300 mg for at least 16 weeks for PsO, PsA, and axSpA. At the time of the data cutoff, the HS pivotal studies were not available for pooling but were published after the data cutoff for this report [2]. Safety of the pediatric studies were published separately [13, 14]

This analysis aimed to provide contextual descriptions of the safety experience with secukinumab primarily in PMS in conjunction with the data from pooled CTs. The AEs associated with the safety experience as described in this report are relevant for secukinumab and are called AEs of special interest (AESI). Additionally, paradoxical skin reactions identified from PMS (outside CTs) were also described in this report. All AEs were coded using the Medical Dictionary for Regulatory Activities (MedDRA; version 25.0) with individual AEs presented in MedDRA Preferred Terms (PT). The search for AESI was broader than the targeted medical concepts to include all potential events, e.g., “rash” is included in the search for hypersensitivity, although it may also be an underlying condition.

AEs and serious AEs (SAEs) definitions, AESI search criteria, estimation of post-marketing exposure, and descriptions of the CTs pool are described in Supplementary Material, Supplementary Table 1, and Supplementary Fig. 1.

Statistical Analysis

AE assessments were descriptive with reporting rate (RR) calculated as number of PMS reports divided by PMS exposure per 100 PY. Post-marketing exposure was estimated on the basis of the sold volume of drug substance and defined average daily dose (Supplementary Material). Pooled AEs in CTs were analyzed using exposure-adjusted incidence rates (EAIR) per 100 PY.

Ethical Approval

All AEs were assessed in accordance with Good Pharmacovigilance Practice Module VII and International Council for Harmonization (ICH) Guideline E2C [15]. All CTs were conducted in compliance with the Declaration of Helsinki [16], ICH Guidelines for Good Clinical Practice, and local country regulations. The protocols for the clinical trials were reviewed and approved by ethical review committees and authorities for each clinical site; all patients provided written informed consent.

Results

In PMS, 169,248 case reports were received (RR 14.60/100 PY, decreased from 24.18/100 PY in December 2018, Fig. 1), and of these, 18% were serious (17% serious in December 2018).

Fig. 1
figure 1

Post-marketing reporting rate per 100 PY. The bar graph demonstrates the decreasing reporting rate of safety events (left y-axis) as the cumulative exposure of secukinumab increases (right y-axis), as reported by periodic safety update reports from 25 December 2018, 25 December 2020, and 25 June 2022. PY patient-years, RR reporting rate

Full size image

In CTs, patients’ baseline characteristics are summarized in Supplementary Table 2. AEs are summarized by indication, with secukinumab exposure up to 5.4 years (Table 1). Similarly, these AEs were mainly non-serious: the EAIR for AEs versus SAEs was 181.59 [95% confidence interval (CI) 178.46, 184.77]/100 PY versus 7.04 (95% CI 6.72, 7.37)/100 PY.

Table 1 Summary of safety data from secukinumab clinical trials (entire treatment period)
Full size table

AESI are summarized for PMS and pooled CTs in Table 2. Changes in PMS RR since December 2018 are plotted in Supplementary Figs. 2 and 3, showing decreases in RR for all but 2 AESI (mycobacterial infections and malignancies, which had no change in RR).

Table 2 Summary of AEs of special interest in post-marketing setting and pooled clinical trials
Full size table

Infections were the most frequent AESI (RR 3.19/100 PY) with 71% being non-serious. In both PMS and CTs, nasopharyngitis (RR 0.59/100 PY, EAIR 16.08/100 PY) and pneumonia (RR 0.14/100 PY, EAIR 0.17/100 PY) were the most common infection and serious infection, respectively (Table 2). Coronavirus disease 2019 (COVID-19) infections were further tracked through 25 September 2022, totaling 5816 cases in PMS and CTs (Supplementary Table 3); 81 (1.4%) of these cases were reported as COVID-related deaths.

Fungal, herpes, mycobacterial, and Staphylococcus infections were mainly localized or non-disseminated in nature (Table 2). Candida infections (RR 0.20/100 PY, EAIR 2.16/100 PY) were the most common fungal infection, with oral candidiasis (RR 0.09/100 PY, EAIR 1.06/100 PY) as the most common Candida infection. Serious fungal infections were reported in 15% (576/3883, RR 0.05/100 PY) of the PMS fungal infection cases, most of which were esophageal candidiasis, oral candidiasis, and other unspecified fungal infections. History of diabetes was reported in 3.3% (76/2317, PMS) and 9% (54/582, CTs) cases with Candida infections that were mainly non-serious and superficial. In CTs, EAIR of Candida infections in patients with diabetes versus without diabetes was 2.31 (1.74, 3.02)/100 PY versus 2.14 (1.96, 2.33)/100 PY, respectively. Mycobacterial infections (RR 0.03/100 PY) included tuberculosis (RR 0.02/100 PY) and latent tuberculosis (0.008/100 PY), accounting respectively for 64% and 25% of PMS cases. In contrast, in CTs, AEs of tuberculosis were mainly latent tuberculosis (EAIR 0.04/100 PY).

Of the potential opportunistic infections (RR 0.05/100 PY, EAIR 0.17/100 PY, Supplementary Table 4), infrequent but potentially clinically significant events were identified in 48 out of the 559 PMS cases (RR 0.004/100 PY) and an additional 4 cases were identified from CTs. A majority of the 52 cases were either insufficiently documented or were likely related to comorbidities (Supplementary Material).

Deaths (eight in PMS, one in CT) reported with fungal, herpes, mycobacterial, or Staphylococcus infections were noted with limited or confounded information (Supplementary Material).

Hepatitis B virus (HBV) reactivation was reported in 13 cases (all in PMS, with 8 from Taiwan where HBV is endemic) as of 31 March 2023; of these, in 11 cases either there was no pertinent HBV diagnostic data or there was historical use of adalimumab, etanercept, or ustekinumab, which all share a risk of HBV reactivation as per prescribing information. In the remaining two cases, a potential association with secukinumab may be suspected, one with concomitant risks of hepatitis C and no administered HBV prophylaxis, and the other resulting in interruption of secukinumab treatment and initiation of nucleoside analogues. Secukinumab was reintroduced 6 months later, while nucleoside analogues were continued without recurrence.

Hypersensitivity was reported with onset dates available in 31% (4332) of the 14,113 PMS cases; 70% (3024/4332) occurred ≥ 2 weeks after the first dose of secukinumab. Unspecified rash was most common, which could also be an underlying condition (Table 2). Non-immunoglobulin (IgE)-mediated events were identified in 36 cases (RR 0.003/100 PY, Supplementary Table 5), none of which could be confirmed as related to secukinumab (Supplementary Material).

Anaphylaxis was identified in 239 PMS cases (RR 0.02/100 PY), 27 (11%, RR 0.002/100 PY) with reported onset within 2 weeks of secukinumab initiation without alternative etiology. All 27 cases resulted in secukinumab discontinuation. An additional 115 PMS reports without an explicit anaphylaxis description were reviewed for a combination of AEs involving any two of the following: skin/mucosal tissues, respiratory compromise, and hypotension (Supplementary Table 6). However, none of the 115 cases met criteria for anaphylaxis, per standardized diagnostic criteria [17].

IBD was reported with RR 0.14/100 PY and EAIR 0.26/100 PY (Table 2). In PMS, prior history of IBD was reported in 14% (227/1578) IBD cases. In CTs, 53 patients were new-onset IBD cases while 19 had IBD flare, accounting for 24% of all 78 patients with prior history of IBD, i.e., 76% patients with prior IBD did not have IBD flare (Table 3). Gastrointestinal bleeding, perforation, or diarrhea with dehydration were reported as life-threatening in 12 PMS cases and were reported with fatal outcome in another 3 cases (all 3 had cardiovascular risks). Another patient having a history of ulcerative colitis without reported flare died of a duodenal ulcer perforation.

Table 3 Summary of pooled clinical trial data describing reports of IBD history
Full size table

Malignancies (RR 0.23/100 PY, EAIR 0.78/100 PY) are summarized in Tables 2 and 4. Prior history of malignancies was found in 659 (25%) of the 2620 PMS cases with AEs of malignancy. Onset dates were reported in 1009/2620 (39%) PMS cases; one-fourth of these (245, or 9% of the 2620 cases) reported malignancies (mostly carcinomas and non-melanoma skin cancers) within 6 months after the first dose of secukinumab. Additionally, there were malignancies reported ≥ 6 months after being treated with secukinumab for ≤ 6 months.

Table 4 Summary of post-marketing data describing reports of malignancy
Full size table

Major adverse cardiovascular events (MACE, RR 0.12/100 PY, EAIR 0.39/100 PY, Table 2) were commonly reported with preexisting risk factors. In PMS, preexisting cardiovascular risks were observed in 613 of 1435 (43%) cases reporting MACE. In CTs, relevant medical histories included hypertension (33.6%), hyperlipidemia (18.4%) and diabetes mellitus (8.3%).

Suicidal ideation and behavior (SIB) were reported in PMS (RR 0.02/100 PY) and CTs (EAIR 0.08/100 PY) with death by suicide in 17 cases [15 in PMS (RR 0.001/100 PY), 2 in CTs (EAIR 0.01/100 PY)] (Table 2). Relevant medical history was identified in 75 of 234 PMS cases including depression, bipolar disorder, post-traumatic stress disorder (n = 45), suicidal ideation or behavior (n = 14), concomitant anti-depressive agents (n = 8), concurrent alcoholism (n = 2), and unspecified mental disorders (n = 6).

Dyshidrotic eczema (DE) and pyoderma gangrenosum (PG) were identified by Novartis as paradoxical skin reactions based on the well-documented post-marketing case reports. At data cutoff for this report, there were 74 PMS cases with DE (RR 0.006/100 PY) and 39 PMS cases with PG (RR 0.003/100 PY). In CTs, the EAIR for DE was 0.28 (95% CI 0.22, 0.34)/100 PY. By contrast, PG was reported in only one secukinumab-treated patient in CTs [EAIR 0 (95% CI 0.0, 0.02)/100 PY].

Discussion

As secukinumab exposure has surpassed more than 1 million PY through PMS and CTs, it is important to understand whether the previously reported safety profile remains applicable. The presented AEs were retrieved from more extensive post-marketing exposure and a larger CT pool. A majority of these AEs were non-serious in both PMS and CTs. Additionally, although the HS pivotal studies were not available for pooling in this report, the separately published HS safety data did not reveal any new safety findings [2].

Infections were reported consistently in PMS and CTs with nasopharyngitis, pneumonia, and oral candidiasis as the most common infection, serious infection, and fungal infection, respectively. EAIR of Candida infections was similar between patients with and without diabetes. COVID-19-related mortality rate was 1.4%, comparable to the background COVID mortality rate (1.1%) in the USA [18]. No increased risk for COVID-19 infection was found. Infrequent but clinically significant fungal, herpes, mycobacterial or Staphylococcus infections remained rare and were reported with either limited or confounded information precluding valid inference. In a meta-analysis, the incidence of opportunistic infections was similar between secukinumab and ixekizumab but was lower than bimekizumab [19].

Hypersensitivity events after 2 weeks of exposure to secukinumab could not be induced by secukinumab unless they were non-IgE-mediated. In PMS, 70% of the cases with available onset dates occurred ≥ 2 weeks after initiating secukinumab. Non-IgE-mediated events (based on the reported event terms) were rare without evidence of association with secukinumab. Anaphylaxis remained rare.

IBD was previously analyzed on the basis of 21 pooled CTs across 3 indications (PsO, PsA, and axSpA) through 25 June 2017, showing ulcerative colitis (EAIR up to 0.2/100 PY), Crohn’s disease (EAIR up to 0.4/100 PY), and unspecified IBD (EAIR up to 0.1/100 PY). Of these patients with IBD, 30 were new-onset cases; 11 had IBD flares accounting for 23% of the total 48 patients with a history of IBD, i.e., 37 (77%) of the patients with a history of IBD did not experience IBD flare [11]. In the current analysis, as presented in Table 1, the EAIR for IBD were either similar to or lower than previously reported. Cumulative CT data continued to affirm that 76% of the patients with a history of IBD did not report IBD flares. Additionally, the EAIR of IBD for secukinumab is comparable to the IBD rates reported in the general PsO, PsA, and axSpA populations [20,21,22], although differences in data collection and analysis methodologies are noted between CTs and observational studies, which may preclude a direct comparison.

Other targeted safety monitoring included malignancy, HBV reactivation, MACE, and SIB. In PMS reporting, one-fourth of malignancy cases with reported onset dates (9% of all PMS malignancy cases) occurred within 6 months after initiating secukinumab, which could suggest that secukinumab was not a causative factor [23]. The malignancy EAIR in pooled CTs for secukinumab is comparable to the rates reported in other CTs of IL-17A inhibitors (ixekizumab and brodalumab) [24, 25]. Relevant risk factors were commonly present when reporting HBV reactivation, MACE, and SIB in PMS. Underlying cardiovascular or metabolic conditions were also common in CTs. HBV reactivation in one patient did not reoccur after resuming secukinumab treatment along with anti-viral prophylaxis, which is reflective of the importance of mitigating the risk of adverse outcomes.

Paradoxical skin reactions were proactively detected during PMS monitoring. As a paradoxical skin reaction, PG was only reported in one CT patient. Due to its rarity, incidence of PG may not be determined. Psoriasis is predominantly driven by T helper 17 (Th17) cells, while atopic dermatitis is largely driven by Th2 cells. When Th17 cells are inhibited, the balance may shift toward Th2 cells, resulting in an eczematous paradoxical reaction. Paradoxical psoriasiform and eczematous eruptions have been identified as the most common type of paradoxical reactions and have been previously associated with exposure to tumor necrosis factor-α inhibitors (91.2% of all cases), IL-17A/17R inhibitors (3.5%), IL-4R-α inhibitors (2.7%), IL-12/23 inhibitors (2.4%), and IL-23 inhibitors (0.01%) [26].

Limitations to PMS data include insufficient information, lack of adjustment for confounders, medically unconfirmed data sources, and duplicate reports. Lower PMS RR may result from the recent COVID-19 pandemic causing limited access to healthcare professionals and reduced AE reporting [27]. Additionally, PMS RR are not equivalent to EAIR in CTs, given differences in the reliability of reported cases (numerator) and the overall at-risk population (denominator). Despite these limitations, monitoring RR over years provides insights into safety profiles over time. PMS data also serve as an effective mechanism for ongoing safety monitoring and identification of new safety findings not observed in CTs, due to limited sample sizes, compared with post-marketing use.

Conclusion

PMS analyses with more than 1 million PY of exposure were consistent with the safety profile of secukinumab observed in an increasing number of pooled CTs in different dermatologic and rheumatologic populations. Additionally, routine PMS of cumulative patients treated with secukinumab have identified paradoxical skin reactions, which due to their nature and very low frequency do not alter the previously established benefit–risk balance of secukinumab.