Clinical characteristics and health‐care resource utilization in patients with generalized pustular psoriasis using real‐world evidence from the Japanese Medical Data Center database

Abstract Little is known about the disease burden, health‐care resource utilization (HCRU), or treatment of patients with generalized pustular psoriasis (GPP) in Japan. This retrospective cohort study used data from the Japanese Medical Data Center database to compare the demographics, comorbidities, and medication use of patients with GPP and plaque psoriasis and estimate their all‐cause HCRU. The patient selection period was from January 1, 2015 to December 31, 2019, and patients must have had at least one confirmed inpatient claim or outpatient claim for GPP or plaque psoriasis. During the 12‐month follow‐up period, 110 patients with GPP and 20,254 patients with plaque psoriasis were identified. An age‐ and sex‐matched (4:1) comparator control cohort, including members of the general population without a diagnosis of psoriasis (but allowing for a diagnosis of psoriatic arthritis), GPP, or palmoplantar pustulosis, was used. The most prevalent comorbidities in patients with GPP included allergic rhinoconjunctivitis, hypertension, and peptic ulcer disease. Patients with GPP were more likely to experience more comorbidities than those with plaque psoriasis, including asthma, chronic obstructive pulmonary disease, interstitial pneumonia, hyperuricemia and gout, tonsillitis, psoriatic arthritis, other psoriasis, and osteoporosis. Patients with GPP were more likely to be treated with a combination therapy than those with plaque psoriasis (65.5% vs 21.7%, respectively) and less likely to be treated with a topical medication alone (20.9% vs 50.8%). Patients with GPP had more outpatient visits than patients in the plaque psoriasis or matched control cohorts (mean [standard deviation], 14.8 [8.3] vs 11.0 [7.6] and 7.8 [7.2], respectively). They were also more likely to require inpatient hospitalization (24.5% vs 6.4% and 5.0%, respectively). Despite study limitations, patients with GPP in Japan were found to have a higher disease burden, including presence of comorbidities and medication use, than those with plaque psoriasis.


| INTRODUC TI ON
Generalized pustular psoriasis (GPP) is a rare, intractable, potentially life-threatening neutrophilic skin disease characterized by widespread diffuse, erythematous patches with visible pustules. [1][2][3][4] Historically, GPP has been classed as a subtype of psoriasis; 5 it has only recently been established as a phenotypically and genetically separate disease entity to plaque psoriasis. 3,4,6,7 Due to its rarity, the clinical course of GPP is not well reported, 8 and there are limited data on disease management and patient experiences of living with GPP compared with plaque psoriasis. In contrast to GPP, plaque psoriasis is the most common subtype of psoriasis. Plaque psoriasis is an inflammatory skin disease that is characterized by patches of erythematous scaly, papular, or plaque skin that may be painful or tender. 9 It is also associated with comorbidities, including psoriatic arthritis, cardiometabolic diseases, and depression. 10 For mild plaque psoriasis, treatments usually include topical medications, with biologics (specifically tumor necrosis factor [TNF] inhibitors) used as first-line treatment for more severe cases of psoriasis. 10 First-line treatments for GPP usually include similar therapies, despite GPP being genetically and phenotypically distinct from plaque psoriasis, representing a potential unmet need for this patient population. 3 There are limited data on the disease burden, health-care resource utilization (HCRU), and treatment of patients with GPP in Japan. Understanding the disease burden in this population, including the associated comorbidities, is important from both the patient and health-care system perspective. This will aid in the understanding of the different needs of patients with different skin diseases, and allow insights into how patients with different diagnoses are treated.
Because GPP is a rare disease affecting a relatively small proportion of the Japanese population, there are limited data pertaining to the economic burden of the disease on the Japanese health-care system.
The potential major economic impact of rare diseases on health-care systems has previously been reported, 11 and understanding the HCRU and associated costs of managing GPP will be important to understand the impact of HCRU on health-care budgets.
Here, we describe the characteristics and HCRU of employed patients with GPP in Japan using data from the Japanese Medical Data Center (JMDC) database, a commercially available claims database in Japan. The JMDC database is a real-world, epidemiological receipt database that accumulates receipts from multiple health insurance associations submitted from inpatient, outpatient, dispensing pharmacy settings, and hospitals that use the Japanese diagnostic procedure combination payment system, as well as health check-up data. All claims received are collected by the insurer. Patients enrolled in the JMDC database are employed individuals, and their dependents, from relatively large companies who are commercially insured, representing approximately 7.3 million patients in Japan. 12 However, there are insufficient data for the population aged ≥65 years, and the data for those aged >75 years are also lacking.
Outcome measures in this study included comorbidities, dermatological medication use, concomitant medication use, and all-cause HCRU and health-care costs during the 12-month follow-up period.

| ME THODS
This retrospective cohort study used data on the characteristics (including demographics, comorbidities, medication use, and HCRU) of patients with GPP and plaque psoriasis from the JMDC database, an administrative claims database that holds accumulated prescription receipts (inpatient, outpatient, and dispensing) and health check-up data from multiple health insurance associations.
The patient selection period was from January 1, 2015 to December 31, 2019, and patients were enrolled into the study at any time during this period once they fulfilled the study criteria, with a minimum baseline period of 6 months of continuous enrollment preceding the index date. The index date was the date of the first inpatient or outpatient claim with a confirmed diagnosis code of L40.0 (plaque psoriasis) or L40.1 (GPP), respectively. A data collection follow-up period of 12 months started on the index date ( Figure 1).
The International Classification of Diseases, 10th revision (ICD-10 as dictated by the World Health Organization) codes and standard disease codes (named by the Japanese Ministry of Health, Labor, and Welfare) were used to identify patients with GPP (ICD-10 code L40.1) and patients with plaque psoriasis (ICD-10 L40.0).
Patients included in each of these cohorts had at least one inpatient or outpatient confirmed claim code, according to standard criteria for identifying patients in claims databases. The third cohort was an age-and sex-matched (4:1) comparator control cohort, matched to patients with GPP (denoted throughout as the "matched control cohort"). This cohort included members of the general population without a diagnosis of psoriasis (excluding ICD-10 codes L40.0, L40.2, L40.4, L40.8, or L40.9), GPP (ICD-10 code L40.1), or palmoplantar pustulosis (ICD10 code L40.3); however, inclusion criteria allowed for a diagnosis of psoriatic arthritis (L40.5). Patients with GPP and concomitant plaque psoriasis were included in the GPP cohort.
F I G U R E 1 Study design. † Patients were enrolled into a cohort when they fulfilled all study criteria (index date). GPP, generalized pustular psoriasis Only patients with a confirmed diagnosis (i.e., with one inpatient or outpatient ICD-10 diagnosis code) of GPP or plaque psoriasis and at least 6 months of continuous enrollment prior to the index date were eligible for inclusion. All patients in the GPP and plaque psoriasis cohorts must have had 12 months of continuous enrollment post the index date to be included in the 12-month follow-up analyses. ICD-10 and standard disease codes were also used to identify comorbidities.
All analyses were conducted using the Instant Health Data platform (Panalgo) and R software, version 3.5.2 (R Foundation for Statistical Computing). No comparative statistical analysis was undertaken; descriptive analyses were used throughout. This included the mean and standard deviation (SD), median and interquartile range for continuous variables, and frequencies and percentages for categorical variables. For HCRU analyses, all-cause visits were included. All comorbidities and medications analyzed can be found in Tables S1-S4.
This study was approved by the ethics committee of the Yoyogi Mental Clinic (approval no. NBI207-1). The study was also conducted in accordance with the Declaration of Helsinki. 13 Table 1) and the mean age of patients in the GPP, plaque psoriasis, and matched control cohorts was 45.0, 42.6, and 44.7 years, respectively. Patients with GPP were more likely to visit larger hospitals (≥20 beds, public or non-public) than patients with plaque psoriasis (39.2% vs 12.5%) as well as university hospitals (13.5% vs 2.9%), whereas patients with plaque psoriasis were more likely to visit a smaller clinic (≤19 beds) than patients with GPP (83.2% vs 46.6%) ( Table 1).

| Comorbidities during the 12-month followup period
The most prevalent comorbidities in patients with GPP included allergic rhinoconjunctivitis, hypertension, peptic ulcer disease, obesity, psoriatic arthritis, asthma, tonsillitis, sinusitis, other psoriasis, osteoporosis, hyperlipidemia, hyperuricemia or gout, chronic obstructive pulmonary disease (COPD), and interstitial pneumonia. In contrast to published literature, 14 thyroid disorders were not among the most commonly reported comorbidities in those with GPP, despite being slightly more prevalent in those with GPP compared with those with plaque psoriasis and the matched cohort (4.6% vs 2.9% and 2.1%, respectively). During the 12-month follow-up period, patients with GPP were more likely to experience the majority of these comorbidities than those with plaque psoriasis and individuals in the matched control cohort; however, hyperlipidemia was more commonly diagnosed in both the plaque psoriasis and matched control cohorts compared with the GPP cohort ( Figure 3). Obesity was more common among individuals in the matched control cohort than in patients with GPP or plaque psoriasis; however, not all patients in each cohort had a measure for body mass index, which may influence this finding.

12-month follow-up period
Patients with GPP were more likely to receive treatment for their disease during the 12-month follow-up period than patients with plaque psoriasis (96.4% vs 78.5%, respectively) ( Table 2). Patients with GPP were also more likely to be treated with a biologic monotherapy than patients with plaque psoriasis (1.8% vs 0.2%, respectively), and were F I G U R E 2 Study population. GPP, generalized pustular psoriasis; ICD-10, International Classification of Diseases 10th revision; PPP, palmoplantar pustulosis also more likely than patients with plaque psoriasis to be treated with non-biologic systemic therapy alone (8.2% vs 5.9%) ( Table 2). Of the patients who received systemic therapies, more patients with GPP were treated with combination therapies (topical steroids + non-biologic systemics; topical steroids + biologics; topical steroids + non-biologic systemics + biologics; and biologic + non-biologic systemic) than those with plaque psoriasis (65.5% vs 21.7%, respectively), and patients with GPP were less likely to be treated with a topical steroid alone compared with patients with plaque psoriasis (20.9% vs 50.8%) ( Table 2).
Of the non-biologic therapies, the most commonly prescribed medications in patients with GPP were etretinate and cyclosporin, which were both more commonly prescribed in patients with GPP than in those with plaque psoriasis (etretinate, 17.3% vs 1.2%; cyclosporin, 28.2% vs 3.6%, respectively) ( Figure 4). Of the biologic therapies, interleukin (IL) inhibitors were more commonly prescribed in patients with GPP than in those with plaque psoriasis (13.6% vs 0.6%, respectively), as were TNF inhibitors (10.9% vs 0.9%, respectively) ( Figure 4). T-cell inhibitors were not commonly prescribed in any of the cohorts (<0.1% for all). Of the topical medications, tacrolimus was received by 9.1% of patients with GPP, compared with 5.4% with plaque psoriasis. This is in contrast with systemic tacrolimus, which was received by just 0.9% of patients with GPP and 0.2% of patients with plaque psoriasis. Vitamin D 3 was received by 10.9% of patients with GPP compared with 1.2% of patients with plaque psoriasis.

12-month follow-up period
During the 12-month follow-up period, patients with GPP generally received more frequent concomitant medication for their comorbidities than patients with plaque psoriasis and individuals in the matched control cohort ( Figure 5). The most frequently prescribed medication across all cohorts was antibiotics; however, the proportion of patients with GPP requiring antibiotics was higher than the proportion of individuals in both the plaque psoriasis and control cohorts who required antibiotics (71.8% vs 50.4% and 44.5%, respectively).
Patients with GPP were nearly 2-times as likely to be prescribed psychiatric medication than those with plaque psoriasis, and nearly 2.5-times as likely than individuals in the matched control cohort.
Patients with GPP were also more likely to receive antihypertensive medication, type 2 diabetes medication, and opioid pain medication than those with plaque psoriasis and individuals in the matched control cohort (though overall use of opioid pain medication was low in both GPP and plaque psoriasis cohorts at 2.7% vs 1.2%, respectively).
Patients with plaque psoriasis and individuals in the matched control cohort were more likely to be prescribed statins, hyperlipidemia medication, asthma medication, and COPD medication than patients in the GPP cohort ( Figure 5); however, these differences were small, and no large differences were observed between all three cohorts.

| All-cause HCRU during the 12-month followup period
Compared with patients with plaque psoriasis and individuals in the matched control cohort, the mean number of outpatient visits for patients with GPP was higher (mean [standard deviation [SD]], 14.8 [8.3] vs 11.0 [7.6] and 7.8 [7.2], respectively). Patients with GPP were also more likely to require inpatient hospitalization (24.5% vs 6.4% and 5.0%) ( Table 3). In addition, the mean duration of hospitalization was between 2 and 4 days longer on average for patients in the GPP

| DISCUSS ION
This study summarizes the clinical characteristics of patients with GPP in a real-world setting in Japan and compares them with those of patients with plaque psoriasis and individuals in an age-and sexmatched control (general population) cohort.
Across all cohorts in this study, prevalent comorbidities included allergic rhinoconjunctivitis, hypertension, and peptic ulcer disease.
The comorbidities reported for patients with plaque psoriasis in this study are consistent with a previous study of patients in Japan with plaque psoriasis using the Japanese National Database of Health Insurance Claims. 15 In both GPP and plaque psoriasis, defects in inflammatory mediators that are involved in allergic rhinoconjunctivitis, including pro-inflammatory IL (specifically IL-6) and TNFα, 16,17 also contribute to the manifestation of GPP and plaque psoriasis.
However, approximately 44% of the overall population in Japan experience allergic rhinitis specifically; 18 21 Another prevalent respiratory comorbidity was interstitial pneumonia. Interstitial lung disease has previously been associated with biologic use in patients with different types of psoriasis in Japan. 22 The use of biologics may therefore contribute to the increased prevalence of interstitial pneumonia in patients with GPP in Japan; however, this association is not clear from this database. Steroids have been associated with development of osteoporosis, specifically after corticosteroid initiation. 23 Given that corticosteroids were among the medications with the highest prescription claims in both GPP and plaque psoriasis cohorts, this may have contributed to increased prevalence of osteoporosis in these cohorts, given that prevalence was particularly high in this database, despite the cohort of patients being relatively young.  Notes: Prescription costs are calculated using standard drug price. Standard drug price is designated by Medical Remuneration in Japan and is renewed every 2 years. The limitation of using standard cost is the cost is not necessarily the reimbursed cost, and this cost will not vary between facilities in the same time for the same medication code. suggesting that patients with GPP may experience prolonged complications of their disease.
The JMDC database is one of the largest claims databases in Japan for both industry and academic use; 24 however, this study was not without limitations. The database includes data from receipts for prescriptions registered to an administrative claims database, and these data were collected for reimbursement purposes rather than scientific research. Whilst we can be certain that biologics administered intravenously. in hospital were received, it is more difficult to determine whether self-administered medications were taken as prescribed after the prescription was filled. In addition, the study population is limited to individuals insured through specific  be male, rather than female, as has been previously reported) 3,5 as males in Japan are more likely to be employed. 25 In addition, this also means that there are very few patients captured who are older than 65 years of age, making it difficult to evaluate the disease burden in this age group. Additionally, the lack of a validation algorithm with positive predictive value to accurately identify patients with GPP, as well as a lack of consensus on the diagnostic criteria for a diagnosis of GPP specifically, presents a limitation in claims database studies; however, the stratification used in this study was based on standard procedures for retrospective claims database studies in Japan.
Given the employment status, overall disease severity of those in the database may be milder than that of the general population, which could include patients with disease so severe that they cannot work at all. A measure of work absenteeism, indicating a level of severity, can be inferred from the measure of inpatient visits, which was higher and of longer duration in patients with GPP than in those with plaque psoriasis; however, this may also be underrepresented because absenteeism may also include those who take days off sick from work, and these are not captured in the database. Finally, the JMDC database does not account for, or measure, the reasons underlying, specific treatment-related decisions, such as treatment discontinuation, switching, and augmentation, and it does not fully cover, or measure, mortality data, including cause of death.
Despite the limitations stated above, the data included in this analysis represent the real-life management of patients with GPP in Japan, and the results show that patients with GPP in Japan have a greater disease burden than those with plaque psoriasis. This study allowed for a detailed characterization of younger patients with GPP, which, to our knowledge, has not previously been performed. Ingelheim sponsored this study and provided funding for the conduct, data analysis, and medical writing assistance for the study's publication. Gao, and K. Saito are employees of Boehringer Ingelheim.

DATA AVA I L A B I L I T Y
We conducted a retrospective study using patient data from the JMDC database, a commercially available database of health claims and administrative data from health insurance associations. All patient data were anonymized and included no identifying information; therefore, informed consent was not necessary. Restrictions do apply to the availability of these data due to a contract between JMDC and Boehringer Ingelheim and are thus unavailable to the public. For inquiries on the database analyzed in this study, please contact JMDC (https://www.jmdc.co.jp/en/index).