Treatment patterns and drug survival for generalized pustular psoriasis: A patient journey study using a Japanese claims database

Generalized pustular psoriasis (GPP) is a potentially life‐threatening skin disease. Although several medications are approved for treating GPP in Japan, there are limited data on real‐world treatment patterns or drug survival (the number of prescribed days of treatment). This retrospective cohort study describes drug survival and treatment patterns of patients with newly diagnosed GPP (International Classification of Diseases, 10th Revision code L40.1), and ≥1 year of follow‐up, using de‐identified claims data (Medical Data Vision Co., Ltd.) from January 2016 to August 2021. Most (97.0%) of the 434 Japanese patients received first‐line therapy of etretinate (26.4%), topical medications (14.7%), or cyclosporin (14.3%); 80.0% and 60.1% of patients received a second and third line of therapy (LOT), respectively. Use of etretinate (12.6%) and cyclosporin (5.9%) decreased in second‐line therapies, whereas use of biologics (interleukin [IL]‐17, 14.3%; IL‐23 inhibitors, 7.6%) and topical medications (22.1%) increased or remained consistent. Approximately 50% of biologics were prescribed in combination with systemic medications or systemic corticosteroids. Median (range) time to next therapy (TTNT) was 2.8 (0.03–48.07) months for first‐line therapy and 3.3 (0.03–52.97) months for all other LOTs. TTNT was longer for combination therapies (up to 16.5 months) compared with monotherapies (up to 7.5 months). Biologics exhibited longer drug survival with fewer treatment episodes compared with non‐biologic systemic medications. Among frequently used therapies, the median (95% confidence interval) drug survival was 8.8 (5.8–11.8) months for etretinate, 4.3 (2.2–6.9) months for systemic corticosteroids, and 19.6 (16.1–26.7) months for secukinumab. Treatment patterns varied considerably, highlighting the need for treatment algorithms and effective, well‐tolerated medications to support patients to help them remain on long‐term therapy.


| INTRODUC TI ON
2][3][4] The clinical course of GPP is unpredictable, presenting as recurrent flares or a persistent disease with intermittent flares, and a broad spectrum of symptom severity between patients and individual flares. 2,4Therefore, some patients require a long-term treatment plan.
6][7][8] Given that only a few treatments have been evaluated in randomized controlled trials, no strong evidence is available to establish treatment algorithms or guidance on the optimal order of optimal treatment; consequently, GPP treatment varies considerably. 2,4,9Available management guidelines and recommendations generally suggest cyclosporin, retinoids, infliximab, or methotrexate as first-line therapies. 2,9r diseases requiring long-term treatment, prolonged drug survival (the length of time a patient remains on a specific drug) can provide insight into treatment effectiveness and safety. 10Although some reports indicate that patients with GPP are more likely to be treated with combination therapy compared with psoriasis vulgaris (PsV), 11,12 detailed analysis of GPP treatment combinations and treatment patterns by line of therapy (LOT) in clinical practice in Japan has not been conducted. 2Similarly, although GPP is considered a poor prognostic factor for drug survival in treating PsV, drug survival data for GPP-specific treatment options are needed. 10,13is study was conducted to explore treatment patterns and drug survival of GPP treatments in Japan to provide further insight into the LOT for GPP and biologic treatment switching.

| Study design
A retrospective cohort study was conducted using a large-scale hospital-based database (Medical Data Vision Co., Ltd, Tokyo, Japan).The database covers approximately 20% of hospitals using the Diagnosis Procedure Combination (DPC) system/Per-Diem Payment System in Japan, comprising >30 million people as of December 2020.De-identified data include patient demographics, diagnosis, laboratory tests, procedures, prescriptions, admissions, and discharge information.The study protocol was approved by the ethics committee of AMC Nishi-Umeda Clinic (Osaka, Japan; AMC-BI-21-010).

| Patients
Patients newly diagnosed with GPP (International Classification of Diseases, 10th Revision code L40.1) between January 1, 2016, and August 31, 2021, were eligible for the study.Patients were excluded if they only had one confirmed GPP diagnosis, could not be followed for >1 year after diagnosis (except those who died within 1 year), or had unclear or no information on GPP diagnosis history.
The date of initial GPP diagnosis was defined as the index date; the baseline period was a year before the index date.Patients were followed up until the end of the study (August 31, 2021), death, or loss to follow-up, whichever came first.Full details are provided in Methods S1.

| Outcomes and measurements
Demographic characteristics were recorded at the index date, or most recent record before the index date, during the baseline period.
The Charlson Comorbidity Index (CCI) scores for comorbidity evaluation were calculated based on the Quan method. 14 there are no established algorithms to identify the LOT for GPP treatment, we created and optimized a complex algorithm to capture treatment patterns in the real-world setting, based on simulation results and expert opinions.All medications for GPP were classified into main or supplementary medications.
Medication switch, add-on, and discontinuation were considered for monotherapy and combination therapy. 15Supplementary medications in combination therapy did not change the LOT; supplementary medications remained as one LOT in monotherapy.
Due to the number of specific regimens, these were categorized into general regimens according to medication category and defined prioritization (Table 1).Prioritization order was decided according to which drugs were considered the main drugs in combination therapy.Time to next therapy (TTNT) and drug survival analyses were performed as proxies for clinical outcomes.TTNT was defined as the start date of one LOT to the start date of the subsequent LOT.Drug survival analysis investigated patient adherence to a specific medication.We defined drug discontinuation as a 3-month interval for biologics and a 2-month interval for oral drugs.Frequency of switching to other biologics was measured according to the last episode of each biologic.A full list of treatments and study definitions are provided in Tables S1 and S2, respectively.

| Statistical analysis
Descriptive statistics were calculated for baseline characteristics and treatment patterns.TTNT and drug survival analyses were performed using the Kaplan-Meier method.For TTNT, the last LOT for each patient was defined as "censor," whereas other LOTs were defined as "event."For drug survival, when multiple episodes of one medication were recorded by one patient, the end date of the last episode was defined as "censor," whereas other episodes were defined as "event."Drug survival of all episodes was estimated among recipients of that medication.All statistical analyses were performed using SAS version 9.4 (SAS Institute Inc., Cary, NC, USA).

| Time to next therapy
Median (range) TTNT was 2.8 (0.03-48.07) months for first-line therapy and 3.3 (0.03-52.97) months for all LOTs (Table S4).Across all LOTs, most GPP treatments were used in combination, typically a general regimen (e.g., etretinate) plus supplementary drugs (e.g., tocilizumab); the resulting combination therapies had a longer TTNT than monotherapies (Table 3).Among the most frequent specific regimens, across all lines, median TTNTs were shorter for etretinate (2.1 months) and cyclosporin (2.7 months) alone, but longer for etretinate plus a supplementary drug (8.1 months) and cyclosporin plus a supplementary drug (9.9 months).Similarly, median TTNT for IL-17 inhibitors was 5.7 months for monotherapy and 11.4 months for IL-17 inhibitor plus a supplementary drug.In addition, systemic corticosteroids were frequently used as monotherapy, but had a short median TTNT.

| Drug survival
Generally, biologics exhibited longer drug survival with fewer treatment episodes compared with non-biologic systemic medications (

| Subgroup analysis: Elderly patients
This subgroup analysis included 185 (42.6%) patients aged ≥65 years and 87 (20.0%) patients aged ≥75 years.Use of biologic-based therapies as first-line therapy was less common among elderly patients than in the overall population (Figure 5; Figure S7).Notably, TNFα inhibitor-based regimens were not used as first-line therapy in the ≥75-year-old cohort (Figure S7b).Use of biologics increased from second-line therapy onwards, but it remained lower than in the overall population; the exception was IL-23 inhibitors, which were more commonly prescribed for elderly patients than in the overall population (Figure 5; Figures S7 and S8).The proportion of elderly patients treated with biologics, with a median drug survival of >1 year, was similar to the overall population (Table S6), suggesting that biologics were well tolerated by elderly patients with GPP.For patients aged ≥65 years with GPP, ustekinumab had the highest switch frequency (75.0%), whereas secukinumab was the most frequently switched biologic in patients aged ≥75 years (66.6%).Bio-switching was relatively common among both elderly cohorts (Figure S9a,b).Switching to biologics may reflect a need to improve patients' symptoms through selective suppression of GPP-associated inflammatory pathways, using targeted treatments. 22docaine hydrochloride/adrenaline (a local anesthetic) was the most frequent concomitant medication, probably because it is used during biopsies that are performed for diagnostic testing. 23,24mbination therapies, including supplementary drugs, generally had a longer TTNT than monotherapies, indicating that current GPP monotherapies are not as efficacious in managing GPP in the long term.Approximately 30% of systemic medications were combined with topical medications, whereas approximately 50% and 16-21% of biologics were combined with systemic medications and systemic corticosteroids, respectively.Dermatologists may consider biologic monotherapies to be inadequate, concluding that combinations with antihistamines, for example, are needed to control specific symptoms, such as itchiness.Further insight is required to understand which skin symptoms are the most challenging to treat.

| DISCUSS ION
Biologics generally had a longer drug survival with fewer treatment episodes than oral medications, which supports findings from Germany. 25 Drug survival data of patients with psoriasis, including GPP, from two university hospitals in Japan (Fukuoka University and Jichi Medical University) have shown similar results to our study. 13,26,27Of these, Kishimoto et al. identified GPP as a negative predictor for biologic drug persistence, 13 potentially due to the intermittent nature of GPP flares or treatment inefficacy.
Based on our data, no treatment had a substantially long drug survival (0.8-22.1 months) reflecting the dynamic nature of GPP in which symptom resolution following treatment may reduce the likelihood of patients using the treatment as a long-term therapy.
Additionally, patients frequently switched biologics, which may be associated with the tendency of practitioners to switch from older biologics, such as infliximab and ustekinumab, to newly available biologics.Conversely, there are also some physicians who continue to use biologics as GPP maintenance therapy, in the same way as for treatment of PsV.In conclusion, this analysis identified wide variations in treatment strategies, reflecting the complexity of managing GPP and the need to establish more detailed and up-to-date treatment algorithms.Despite biologics exhibiting the longest drug survival, IL-17 and IL-23 inhibitors were largely reserved for subsequent LOTs, potentially reflecting failed symptom control with nonbiologics.This may be due to the rarity of GPP, thus limiting empirical evidence for their efficacy, or may be due to their cost or burden on patients.In this study, oral drugs were commonly used as first-line therapy, followed by biologics.Based on our experience, some patients achieve better control with oral drugs than currently available biologics.Our results demonstrated that not all patients treated with biologics had a drug survival exceeding 1 year and switching occurred with all biologics.The relatively short drug survival and frequent switching observed

F I G U R E 1
Sankey plot of the treatment pathways of the first three lines of therapy in patients with newly diagnosed GPP (cohort number = 421).Each line represents the treatment pathway at each line of therapy.GPP, generalized pustular psoriasis; IL, interleukin; TNFα, tumor necrosis factor alpha.F I G U R E 2Combination patterns of GPP treatments: Combinations of biologics or apheresis/plasma exchange with other systemic medications.GPP, generalized pustular psoriasis; IL, interleukin; TNFα, tumor necrosis factor alpha.
Abbreviations: IL, interleukin; TNFα, tumor necrosis factor alpha. a Supplementary drugs are topical medications, other oral medications, and arthritis treatments.

| 397 TADA
et al. algorithm, and treatment selection dependent on the physician.

F I G U R E 3 F I G U R E 4
for biologics may be intrinsic to the recurrent and cyclic nature of GPP, reflecting the need to manage chronic and acute symptoms with different treatments.Investigating treatment patterns and drug survival within the context of GPP disease severity may help evaluate the efficacy of these treatments throughout the disease course.However, the variable treatment patterns also highlight that there are still considerable unmet needs for effective and well-tolerated GPP treatments that optimally manage GPP throughout the disease course, enabling patients to remain on long-term therapy where necessary.The identified treatment Patterns in switching GPP treatment from one biologic to another.GPP, generalized pustular psoriasis; IL, interleukin; TNFα, tumor necrosis factor alpha.Treatment pathway in patients with newly diagnosed GPP (a) with PsV and (b) without PsV.GPP, generalized pustular psoriasis; IL, interleukin; PsV, psoriasis vulgaris; TNFα, tumor necrosis factor alpha.
with PsV were treated with IL-17 inhibitors, IL-23 inhibitors, cyclosporin, and apremilast-based therapies than those without PsV across all LOT.Conversely, more patients without PsV were treated with TNFα inhibitors and systemic corticosteroid-based therapies (Figure4; FiguresS4 and S5).Median drug survival was generally longer for patients without versus with PsV for both biologics (e.g., 3.7 | Subgroup analysis: Patients with versus without PsVThis subgroup analysis included 159 (36.6%) patients with GPP with PsV and 275 (63.4%) patients with GPP without PsV.More patients with GPP Time to next therapy of frequently used specific regimens across all lines.
2bbreviations: CI, confidence interval; IL, interleukin; TNFα, tumor necrosis factor alpha. a N represents the total number of patients receiving each drug.bImmunosuppressantsaretacrolimus, dapsone, minocycline, or azathioprine.providedawider set of treatment options in this difficult-to-treat disease.Conversely, patients without PsV were more frequently treated with systemic corticosteroid and TNFα inhibitorbased therapies, which are used to treat systemic symptoms.2Differences in treatment pathways may be due to the variation in on patients treated with GPP-related therapies, miscoding situations have most likely been excluded.Strengths of this study are that the data were derived from a large, deidentified claims database that covers over 30 million people.Patient treatment patterns and LOTs were determined using a novel algorithm, which considered switching, add-on and discontinuation of treatment, and supplementary and main medications.This algorithm can be adapted and applied to other diseases in the future.Although data were available from 2011 onwards, we focused on data from 2016 to 2021 to reflect current treatment practices more accurately.