Post‐Marketing Safety of Ustekinumab Based on 14‐Year Follow‐Up in Danish National Patient Data Authors: Sejun Kim, Andreas Jensen, Alexander Egeberg, Lone Graff Stensballe Published: 2024 ABSTRACT ABSTRACTPurposePsoriasis (PsO), a chronic inflammatory skin disorder affecting a substantial proportion of populations globally, often necessitates systemic treatment including biologics. This 14‐year cohort study, based on Danish national register data, aimed to investigate the enduring safety profile of ustekinumab compared to other systemic psoriasis treatments.MethodsUsing comprehensive Danish national register data, this study scrutinized patients diagnosed with psoriasis or psoriatic arthritis (PsA) who received ustekinumab. The treatment group comparators were non‐biological systemic treatment (non‐biologic), tumor necrosis factor α inhibitor medicine groups (TNF‐α), interleukin (IL)‐17 inhibitors (IL‐17), and IL‐23 inhibitors (IL‐23). The study periods for comparisons were 2009–2022 for non‐biologic and TNF‐α, 2015–2022 for IL‐17, and 2018–2022 for IL‐23. Outcomes were malignancies, cardiovascular events, serious infections, and serious hypersensitivity reactions. Cox proportional hazards regression models were employed to analyze two estimands: a standard intention‐to‐treat (ITT) estimand and a continuous‐index‐treatment (CIT) estimand, which considered switch and re‐initiation of treatments within individuals.ResultsUsers of ustekinumab were found to be younger on average, with an average age of 45.1 years compared to 51.6, 47.2, 49.0, and 48.4 years in the non‐biologic, TNF‐α, IL‐17, and IL‐23 groups, respectively. Also, 57.3% of the ustekinumab users were male, compared to 46.7%, 48.9%, 50.9%, and 58.3% for the non‐biologic, TNF‐α, IL‐17, and IL‐23 groups, respectively. Although the hazard ratio estimates varied across comparators, ustekinumab was found to be safe: regardless of PsA status, no discernible safety signals in terms of malignancy, MACE, severe infections, or severe hypersensitivity reactions were observed for ustekinumab when compared to the treatment comparators.ConclusionsThe present study corroborated the enduring safety of ustekinumab in the context of PsO treatment.Keywords: drug safety, estimand, MACE, malignancies, psoriasis, register‐based study, ustekinumab Summary The long‐term safety of ustekinumab was good in terms of malignancy, MACE, or severe infections, and no serious hypersensitivity reactions were observed. The long‐term safety of ustekinumab was good in terms of malignancy, MACE, or severe infections, and no serious hypersensitivity reactions were observed. Ustekinumab patients were characterized by younger age and a higher proportion of males. Ustekinumab patients were characterized by younger age and a higher proportion of males. Ustekinumab was safe regardless of psoriatic arthritis (PsA) status. Ustekinumab was safe regardless of psoriatic arthritis (PsA) status. Introduction Psoriasis (PsO) is a chronic and recurrent inflammatory skin condition that affects approximately 7.9% of the Danish population [1, 2, 3]. The most prevalent form of PsO is plaque PsO (Psoriasis vulgaris), which affects around 90% of patients and often necessitates systemic therapy for effective management [4]. To address this persistent inflammatory ailment, various biological treatments have been developed. These treatments are designed to target specific cytokine signaling pathways recognized as pivotal in the pathogenesis of PsO. Among the available biological treatments for PsO, tumor necrosis factor α inhibitors (TNF‐α) and interleukin (IL) inhibitors are the most common drug classes employed to modulate the pathogenic inflammatory processes underlying PsO [5]. Notably, the IL‐17 and IL‐23 immune pathways have been recognized as pivotal in the pathogenesis of PsO [6, 7]. Thus, a range of IL‐17 and IL‐23 targeted antibodies were introduced and made accessible to the Danish population from 2009 onward. Ustekinumab, functioning as an IL‐12 and IL‐23 p40 inhibitor, entered the Danish market in 2009 and has since seen widespread use in the management of PsO [8, 9]. Emerging as a “first‐line” therapeutic option for PsO from 2012 and onwards, it necessitates a thorough investigation into its safety profile based on real‐world data [10]. While the safety profiles of these treatments have been assessed for certain adverse events, such as malignancies, major adverse cardiovascular events (MACE), and serious infections, these evaluations have primarily been based on other cohort reports with limited follow‐up periods [11, 12, 13]. The long‐term safety of IL inhibitors, especially in real‐world settings remains uncertain. This study, based on Danish national register data, aims to investigate the safety of ustekinumab and other medicines for PsO within the Danish population for a 14‐year period. Materials and Methods Data Sources This study was based on multiple data sources, including The National Patient Register [14], The National Prescription Registry [15], The Danish Cancer Registry [16], The Register of Causes of Death [17], and the population registers in Denmark [18]. One key advantage of the Danish data system is the assignment of a unique Central Person Register number (CPR‐number) to every Danish resident at birth or upon immigration [19]. This unique identifier facilitates the cross‐linking of data across the national registers within the country. The stability of the CPR‐number throughout an individual's lifetime or until emigration allowed for extended longitudinal follow‐up. The population registers were the source of information on the region of residency, occupations, emigration, and immigration. The national patient registers contained comprehensive details on all hospitalizations and outpatient visits to specialist care, including primary and secondary diagnoses recorded at discharge and procedure codes. Diagnoses of PsO and other approved indications, medical history, comorbidities, and outcomes were documented following the International Classification of Diseases, version 10 coding system (ICD‐10 code) [20]. The national cancer register includes information on all cancer diagnoses, verified through morphological criteria. To ascertain the date and cause of death, data were extracted from the cause of death registers. Finally, the national prescribed drug registers contained records on the formulations and dispensing dates of all prescriptions, with drugs categorized according to the WHO Anatomical Therapeutic Chemical classification system (ATC code) [21]. Due to Danish legislation, biologic treatments for PsO can only be prescribed from specialized hospital clinics in Denmark. Consequently, all biological drug treatments are exclusively administered and managed within the hospital setting. Population This study utilized data from nationwide health, administrative, and socio‐economic registers in Denmark, representing a population of approximately 6 million individuals. The study focused on individuals with hospital diagnoses of PsO between 1994 and 2022, identified by ICD‐10 codes: L40, M07.0, M07.1, M07.2, M07.3, and M09.0, including all subcodes. The development of psoriatic arthritis (PsA) was identified using ICD‐10 codes: L40.5, M07.0, M07.1, M07.2, M07.3, and M09.0, including all subcodes [22]. The presence or absence of PsA (on top of the initial PsO diagnosis) was assessed to highlight the excess severity of this co‐morbidity, and the different treatment patterns, that is, expressed by the fact that PsA‐diagnosed patients may also be treated by rheumatologists and not only dermatologists. The study population comprised the total population of Danish individuals with hospital diagnoses of PsO, irrespective of PsA status. In sub‐population analyses, the patients were categorized based on their PsA diagnosis. This resulted in two subgroups: those diagnosed with PsA (henceforth “With PsA population”) and those not (yet) diagnosed with PsA (henceforth “Without PsA population”). All patients diagnosed with PsA were confirmed to have received a diagnosis of PsO either before or on the same day as their PsA diagnosis. Also, individuals first diagnosed with PsA after inclusion in the study would switch to the With PsA population. Exposure The exposure encompassed treatments for PsO that were utilized or introduced in Denmark during the study period 2009–2022. Distinct exposure groups were defined based on their targeted molecules, encompassing the following categories: the IL‐12/23 targeted group (ustekinumab, IL‐12/23), non‐biologic systemic treatment (non‐biologic), tumor necrosis factor α inhibitors (TNF‐α), IL‐17 targeted group (IL‐17), and IL‐23 targeted group (IL‐23), including all approved biological treatments for PsO in Denmark. Each drug within these groups was defined by its ATC code or procedure code, that is, code of specific health service procedures for example in‐hospital administration of certain treatments like biologicals [18]. The complete list of ATC and procedure codes for each treatment group can be found in the supplementary information (Table S1). Study Design This study was a nationwide population‐based cohort study with 14 years of follow‐up from 2009 to 2022 (Figure 1). The study employed an active‐comparator new‐user (ACNU) design [23], comparing the safety defined by subsequent events of malignancies, MACE, severe infections, and severe hypersensitivity reactions among ustekinumab users with users of the other treatment groups. Individuals were defined as active comparators if they initiated treatment with a drug substance from one of the exposure groups and had not received the same drug substance in the preceding 365 days. The drug substance the individual filled a prescription for or had a registration of a hospital procedure code for was denoted the first‐exposed treatment and the first‐exposed date was defined as the date of filling the first prescription/procedure code of the first‐exposed treatment. Flowchart showing the reduction in the number of individuals by each criterion imposed before arriving at the statistical analysis set of ustekinumab‐exposed patients. The first box within each comparator indicates the relevant follow‐up period depending on when the treatments were introduced in Denmark (ustekinumab was introduced in 2009). Exposure Period The study's exposure period commenced in 2009, coinciding with the introduction of ustekinumab in Denmark. At that point in time the TNF‐α and non‐biologic treatment groups were already introduced on the market. After 2009, the study period for each treatment group was determined by the date of approval for the specific drug class (Figure 1). Notably, the study period of the IL‐17 group was initiated in 2015, while the IL‐23 study period began in 2018. The study followed individuals until the end of 2022. If a patient in the treatment group experienced an outcome event, death, or emigration before the end of the study period, their individual follow‐up time was censored. Outcomes This study considered four specific outcomes, representing a broad spectrum of PsO‐relevant frequently used potential‐adverse‐events markers: Malignancies excluding non‐melanoma skin cancer (NMSC), MACE, and serious infections [24, 25, 26, 27]. Serious hypersensitivity was included as an outcome of interest [28], but no events of serious hypersensitivity were captured during the study period. Consequently, this outcome was not further considered in the analysis. These outcomes were defined based on ICD‐10 codes, and a comprehensive list of these codes can be found in the supplementary information (Table S2). Note that events were only considered if the patient received a diagnosis of the specific outcome after exposure to a drug in the treatment group. Outcome Risk Windows To mitigate potential biases in the outcome data, this study employed outcome risk windows (Table S3). For malignancy outcomes, patients were excluded if the event occurred before exposure to their treatment or within 12 months after the first exposure. Patients diagnosed with MACE before exposure were also excluded. The outcome risk windows served as exclusion criteria to ensure that the observed disease outcome was likely attributable to drug exposure. For instance, malignancies often take a long time to develop before diagnosis, making it difficult to conclusively attribute them to the most recent drug exposure. Estimands This study defined two estimands: Intention‐to‐Treat (ITT) and Continuous‐index‐treatment (CIT). The ITT estimand assessed the effect of the first‐exposed treatment group on the incidence of outcome events, regardless of individual treatment changes. Each individual was allowed a single first‐exposed date. In ITT analyses, individuals were followed until the occurrence of the event, death, emigration, or the end of the study. The CIT estimand evaluated the effect of each treatment episode, considering the repetition of treatments within individuals. This estimand accommodated multiple episodes with different first‐exposed dates per individual, enabling the consideration of treatment repetition. In CIT analyses, the episode was followed until the occurrence of the event, death, emigration, end of the treatment episode, switching to other comparator drugs, or end of the study. Treatments were censored upon exposure group switching or the end of the treatment episode. Additionally, the CIT estimand was adjusted with a 365‐day grace period between treatment episodes to minimize biases associated with prior treatments. Matching New users of the ustekinumab group were pairwise matched between the ustekinumab and active comparator group based on the propensity score [29, 30]. Propensity score helps in reducing bias and confounding by creating a summary measure of the likelihood of receiving a particular treatment based on individual characteristics, it was calculated using a logistic regression model that incorporated the user's baseline characteristics, including sex, age, individual's past diagnosed disease history, and the most recent treatment group. The score was updated annually. Active comparator users were considered both before and after 30 days from the first‐exposed date of the ustekinumab user for matching. The matching aimed to ensure comparable follow‐up periods and similar backgrounds between the groups, thus mitigating the influence of unknown confounders. The matching procedure aimed for a maximum of four comparator matches per ustekinumab user and was carried out using the “nearest neighbor matching” approach, with exact matching for the user's history of PsA diagnosis and their most recent treatment group before the first‐exposed date (Figure 1, Figure S1). Further, for the CIT estimand, individuals were matched on their prior number of first‐exposed dates in the analysis. Statistical Modeling and Data Interpretation Cox proportional hazards regression models were utilized to evaluate the risk of outcome incidence after matching. The hazard ratio was employed as a measure of the outcome occurrence risk between ustekinumab and each comparator group. Specifically, the hazard ratio was interpreted as the event rate in the ustekinumab group relative to that of the comparator. Thus, a hazard ratio below 1 indicated that ustekinumab was associated with a lower rate of the outcome and vice versa. Statistical Programs and Packages The statistical programming and data analysis for this study were performed using R [31] (version 4.3.1) in a Windows 10 environment. Packages such as tidyverse [32] (version 2.0.0) and haven [33] (version 2.5.3) were employed for basic data manipulation. For matching and statistical modeling, the study utilized MatchIt [34] (version 4.5.4) and survival [35] (version 3.5–5). Data visualization was conducted using the ggplot2 [36] (version 3.4.2) package. Results Baseline Characteristics and Treatment Groups In the full study period, 2009–2022, 2078 PsO patients became new users of ustekinumab, while 20 398 and 9565 PsO patients were exposed to non‐biologic systemic treatment and TNF‐α inhibitors, respectively (Table S4, Figure 2). Bar plot of the total number of new users in each treatment group during the follow‐up period. The ustekinumab, non‐biologic, and TNF‐α groups were followed from 2009 to 2022. The IL‐17 group was followed from 2015 to 2022. The IL‐23 group was followed from 2018 to 2022. The actual number of new users in each exposure group is shown. From 2015 to 2022, 2140 PsO patients were exposed to IL‐17 inhibitors, and 1307 to ustekinumab. From 2018 to 2022, 312 PsO patients were exposed to IL‐23 inhibitors, and 563 to ustekinumab. The complete list of ATC and procedure codes for each treatment group can be found in the supplementary information (Table S1). The mean age of individuals exposed to ustekinumab was 45.1 years (standard deviation 15.5). Thus, the ustekinumab group exhibited the lowest average age among all treatment cohorts, since the mean ages in the comparator groups ranged between 47.2 and 51.6 years (Table S4, Figure 3). Bar plot of the number of new users by sex and treatment group. The ustekinumab, non‐biologic, and TNF‐α groups were followed from 2009 to 2022. The IL‐17 group was followed from 2015 to 2022. The IL‐23 group was followed from 2018 to 2022. Red bars represent the number of female patients in each treatment group, while blue bars represent the number of male patients in each exposure group. The actual number of new users is shown. Demographic characteristics revealed a predominance of male patients in the ustekinumab and IL‐23 exposed groups, whereas the TNF‐α and non‐biologic groups exhibited a predominance of female patients (Table S4, Figure 4). The IL‐17 exposed group demonstrated an equal distribution of sexes. Boxplot of the age distribution by treatment group. The ustekinumab, non‐biologic, and TNF‐α groups were followed from 2009 to 2022. The IL‐17 group was followed from 2015 to 2022. The IL‐23 group was followed from 2018 to 2022. The boxes represent the interquartile ranges (IQR) of the age distributions. The horizontal line in the box represents the median and the dot represents the mean. Outliers are shown as red dots. Safety Results No events of serious hypersensitivity were captured during the study period thus indicating no relevant safety signal. Hence, it was not relevant to further study the risk of serious hypersensitivity after exposure to PsO treatment in the present study. Further, comparing PsO patients exposed to ustekinumab with PsO patients exposed to IL‐23 was only meaningful for the serious infection outcome due to a limited number of events captured between 2018 and 2022 (Table 1). Specifically, there was one reported malignancy event (with no occurrence among ustekinumab‐treated patients), three events of MACE (none among ustekinumab‐treated patients), and 14 events of serious infections (10 among ustekinumab‐treated patients). TABLE 1The table of the hazard ratio of specific outcomes between ustekinumab and each comparator.Outcome a Estimand b Comparators c Non‐biologicTNF‐αIL‐17IL‐23HR (95% CI)Events/patients (ustekinumab)Incidence (ustekinumab)HR (95% CI)Events/patients (ustekinumab)Incidence (ustekinumab)HR (95% CI)Events/patients (ustekinumab)Incidence (ustekinumab)HR (95% CI)Events/patients (ustekinumab)Incidence (ustekinumab)Malignancies excl. NMSCITT1.36 (0.95, 1.94)114/6166 (40/1582)1.85% (2.53%)1.32 (0.86, 2.02)77/3728 (29/1056)2.07% (2.75%)0.69 (0.23, 2.05)17/1387 (4/461)1.23% (0.87%)NA (NA, NA)1/269 (0/151)0.37% (0%)CIT0.14 (0.02, 1.09)17/7229 (1/1797)0.24% (0.06%)0.22 (0.03, 1.79)15/4317 (1/1270)0.35% (0.08%)1.92 (0.12, 31.03)1/1486 (1/539)0.07% (0.19%)NA (NA, NA)0/253 (0/150)0% (0%)NMSCITT0.97 (0.61, 1.53)94/6326 (23/1572)1.49% (1.46%)0.76 (0.43, 1.35)64/3726 (14/1050)1.72% (1.33%)1.92 (0.54, 6.8)6/1387 (4/461)0.43% (0.87%)NA (NA, NA)0/265 (1/151)0% (0.66%)CIT0.76 (0.21, 2.81)11/7384 (3/1790)0.15% (0.17%)0.89 (0.26, 3.09)15/4311 (3/1264)0.35% (0.24%)NA (NA, NA)0/1485 (1/536)0% (0.19%)NA (NA, NA)0/248 (0/146)0% (0%)MACEITT0.95 (0.69, 1.3)197/6267 (48/1591)3.14% (3.02%)0.94 (0.62, 1.42)103/3727 (28/1064)2.76% (2.63%)1.56 (0.66, 3.68)15/1388 (8/465)1.08% (1.72%)NA (NA, NA)3/263 (0/154)1.14% (0%)CIT0.39 (0.11, 1.38)15/7335 (3/1814)0.2% (0.17%)0.29 (0.04, 2.26)13/4304 (1/1279)0.3% (0.08%)NA (NA, NA)2/1487 (0/540)0.13% (0%)NA (NA, NA)0/247 (0/151)0% (0%)Serious infectionsITT1.12 (0.98, 1.29)874/6485 (248/1633)13.48% (15.19%)0.92 (0.77, 1.09)606/3835 (162/1088)15.8% (14.89%)1.1 (0.8, 1.5)145/1449 (54/480)10.01% (11.25%)1.21 (0.54, 2.72)14/276 (10/158)5.07% (6.33%)CIT0.94 (0.69, 1.29)147/7587 (59/1868)1.94% (3.16%)0.86 (0.61, 1.23)173/4558 (38/1311)3.8% (2.9%)1.25 (0.68, 2.31)33/1578 (15/560)2.09% (2.68%)1.5 (0.4, 5.58)5/261 (4/155)1.92% (2.58%) a Outcome—malignancies excl. NMSC, malignancies excluding non‐melanoma skin cancer (NMSC); NMSC, non‐melanoma skin cancer; MACE, major adverse cardiovascular events. b Estimand—ITT, Intention‐to‐treat; CIT, Continuous‐index‐treatment. ITT was matched by the individual and CIT was matched by treatment episode. c Non‐biologic and TNF‐α groups were followed from 2009 to 2022, the IL‐17 group was followed from 2015 to 2022, and IL‐23 group was followed from 2018 to 2022.Abbreviations: CI, confidence interval; events/patients, the number of events divided by the number of individuals at risk (ITT) or the number of events divided by the number of episodes at risk (CIT); HR, hazard ratio; NA, Not applicable. The table of the hazard ratio of specific outcomes between ustekinumab and each comparator. Outcome—malignancies excl. NMSC, malignancies excluding non‐melanoma skin cancer (NMSC); NMSC, non‐melanoma skin cancer; MACE, major adverse cardiovascular events. Estimand—ITT, Intention‐to‐treat; CIT, Continuous‐index‐treatment. ITT was matched by the individual and CIT was matched by treatment episode. Non‐biologic and TNF‐α groups were followed from 2009 to 2022, the IL‐17 group was followed from 2015 to 2022, and IL‐23 group was followed from 2018 to 2022. Abbreviations: CI, confidence interval; events/patients, the number of events divided by the number of individuals at risk (ITT) or the number of events divided by the number of episodes at risk (CIT); HR, hazard ratio; NA, Not applicable. Based on the ITT estimand, PsO patients exposed to ustekinumab did not exhibit elevated hazards of the two malignancy outcomes when compared to PsO patients exposed to the other treatment groups (Table 1 and Figure 5). Specifically, the hazard ratios (HR) for malignancies (excluding NMSC) and NMSC were 1.36 (95% Confidence Interval (CI) 0.95, 1.94) and 0.97 (CI 0.61, 1.53) in the non‐biologic group comparison. The corresponding HR estimates for malignancies (excluding NMSC) and NMSC were 1.32 (CI 0.86, 2.02) and 0.76 (CI 0.43, 1.35) when comparing PsO patients exposed to ustekinumab with those exposed to TNF‐α, whereas the HRs were 0.69 (CI 0.23, 2.05) and 1.92 (CI 0.54, 6.8) when comparing PsO patients exposed to ustekinumab with those exposed to IL‐17. Hazard ratios by estimand, outcome, and comparator group. The dot represents the point estimation of the hazard ratios between ustekinumab and each comparator. The black horizontal lines represent the 95% confidence interval (CI). The figure displays the CI range up to 3, while the precise CI range is detailed in Table 1. An empty space for a specific outcome of each comparator indicates non‐applicability due to a lack of events. The black vertical dashed line indicates the point of identical hazards between ustekinumab and each comparator. The analyzed outcomes were malignancies excluding non‐melanoma skin cancer (NMSC), NMSC, major adverse cardiovascular events (MACE), and serious infections. The plot was divided by the two estimands: Intention‐to‐treat (ITT) and continuous‐index‐treatment (CIT). Also, for the outcomes of MACE and serious infections, when comparing PsO patients exposed to ustekinumab with PsO patients exposed to non‐biologics, TNF‐α and IL‐17, no evidence of increased rates was observed (Table 1 and Figure 5). The analyses based on the CIT estimand yielded similar conclusions (Table 1 and Figure 5). However, comparisons with the group of PsO patients exposed to IL‐17 were not meaningful due to an insufficient number of events. Regarding serious infections, there were no significant differences across all comparator groups including IL‐17 and IL‐23 (Table 1 and Figure 5). Subgroup Analyses by PsA Status As presented in Table S5 and Figure S2, the present study among hospital‐diagnosed PsO patients found high proportions of patients who were also diagnosed with PsA: 46.2% of patients in the ustekinumab group and 49.7% in the IL‐23 group were diagnosed with PsA at first‐exposed date, while at least 62% of the patients were diagnosed with PsA in the other treatment groups. Among PsO patients with PsA, the patients treated with non‐biologic, TNF‐α, and IL‐17 were predominantly female patients (Table S5, Figure S3). Furthermore, individuals without a PsA diagnosis were generally younger than those diagnosed with PsA across all treatment groups, except for the non‐biologic group in which the age distribution was the same (Table S5, Figure S4). No increased rate of malignancies, MACE, or serious infections were observed among PsO patients with PsA when comparing the patients exposed to ustekinumab with those exposed to non‐biologic, TNF‐α, or IL‐17 treatment (Tables S6 and S7, Figure S5). Discussion In this population‐based national cohort study with 14 years of follow‐up among all Danish patients diagnosed with PsO in hospitals, we found that when compared to other PsO treatments, exposure to ustekinumab was safe in terms of malignancies, MACE, serious infections, and serious hypersensitivities. No events of serious hypersensitivity reactions were observed. Baseline Characteristics and Treatment Groups In the present study, when compared to PsO patients receiving other treatments, the ustekinumab‐treated patients were found to be younger and with a higher proportion of males, aligning with previous studies on the demographic profile of ustekinumab patients [10]. The observed age distribution pattern can be attributed to ustekinumab being established as a first‐line treatment for PsO in Denmark since 2012 [10]. Additionally, the higher proportion of male patients among ustekinumab recipients is consistent with previous findings suggesting that male patients tend to exhibit more severe symptoms compared to their female counterparts [37]. Safety Results The safety profile of ustekinumab compared to that of other biological treatments has previously been studied within limited follow‐up periods [9, 10, 11, 12, 38, 39]. Our analysis corroborates previous studies indicating that ustekinumab is safe in terms of malignancies, MACE, and serious infections. However, comparisons between ustekinumab and the other IL‐23 treatment group were not feasible due to a low number of events during the limited follow‐up period of 5 years. This underscores the need for extended monitoring to capture additional events for this specific comparison. The additional analyses based on the CIT estimand facilitated a comprehensive evaluation of treatment effects at the episode level, enabling the consideration of treatment re‐initiations within individuals. By censoring episodes when the patient switched to other biological treatment groups, the CIT estimand was adjusted with a 365‐day washout period between treatment episodes to minimize biases associated with prior treatments. However, these strict follow‐up and censoring criteria designed to mitigate carryover effects, posed challenges in capturing events occurring between defined treatment episodes, potentially leading to underestimation of treatment effects. Subgroup Analyses by PsA Status The subgroup analyses by PsA status both showed that ustekinumab is safe compared to other PsO treatments in terms of malignancies, MACE, and serious infections. In total, 63.7% of the enrolled patients were diagnosed with PsA, a much larger proportion than what has been reported in the literature (Table S5) [7, 22, 40]. This high proportion of PsA diagnoses is explained by the fact that only the most severely affected PsO patients are diagnosed in the hospitals, and these patients may be characterized by a higher occurrence of symptoms and comorbidities including PsA. The time‐dependent nature of PsA development occurring subsequent to the diagnosis of moderate‐to‐severe PsO contributed to the older age profile observed among patients also diagnosed with PsA [41]. Also, the lower number of patients without PsA resulted in fewer events and thus a higher uncertainty of the results within that sub‐population. Strength One notable strength of the present study, based on real‐world data, was the extensive 14‐year follow‐up period of ustekinumab patients. This prolonged duration allowed for a thorough examination of the safety profile of ustekinumab. Also, the study was based on administrative register data which guaranteed no loss to follow‐up. Further, the safety outcomes were analyzed using two estimand approaches and evaluated in the subgroups of PsO patients by PsA status. Limitation One limitation of the study was related to the data collection process in terms of the reliance on hospital‐based PsO diagnoses and the administration of biological drugs in Denmark. Patients accessing care at the hospital level often present with more severe symptoms, which could potentially inflate the observed proportion of patients with PsA on top of their PsO diagnosis. Further, it is possible that different patterns of registration across the different Danish dermatology departments affected the results in either direction, but even non‐differential misclassification could bias the results toward the null. Also, no serious hypersensitivity reactions were recorded; however, there is a possibility that serious hypersensitivity occurrences may be underestimated due to underreporting within short outcome windows. This lack of occurred numbers also made statistical analyses unfeasible. Additionally, due to the limited number of outcome events within the brief follow‐up period for IL‐23, statistical analyses, aside from serious infections, were not applicable. Conclusion In conclusion, our study demonstrates that ustekinumab is safe compared to other PsO treatments in terms of malignancies, major adverse cardiovascular events, and serious infections. Ustekinumab posed a lower or equivalent rate of the considered outcomes when compared to non‐biologic, TNF‐α, IL‐17, and IL‐23. These findings provide valuable insights into the comparative safety profiles of different biological treatments, aiding clinicians in making informed decisions regarding the most suitable therapeutic options for their patients. Ethics Statement The authors have nothing to report. Consent The authors have nothing to report. Conflicts of Interest Kim and Dr. Jensen have no conflict of interest. Prof. Egeberg has received research funding from Almirall, Pfizer, Eli Lilly, Novartis, Bristol‐Myers Squibb, AbbVie, Janssen Pharmaceuticals, Boehringer Ingelheim, the Danish National Psoriasis Foundation, the Simon Spies Foundation, and the Kgl Hofbundtmager Aage Bang Foundation, and honoraria as consultant and/or speaker from Amgen, AbbVie, Almirall, Leo Pharma, Zuellig Pharma Ltd. Galápagos NV, Sun Pharmaceuticals, Samsung Bioepis Co. Ltd., Pfizer, Eli Lilly and Company, Novartis, Union Therapeutics, Galderma, Dermavant, UCB, Mylan, Bristol‐Myers Squibb, McNeil Consumer Healthcare, Horizon Therapeutics, Boehringer Ingelheim, and Janssen Pharmaceuticals, and is currently employed by LEO Pharma. Prof. Stensballe has received research funding from AbbVie. Supporting information Data S1. Data S2. Funding: The authors received no specific funding for this work. References Associated Data Supplementary Materials Data S1. Data S2.