Biologic initiation rates in systemic-naive psoriasis patients after first-line apremilast versus methotrexate use

This study employed a retrospective cohort design of administrative claims data (2014–2019) from the IBM^® MarketScan^® Commercial and Medicare Supplemental databases to examine biologic initiation rate in patients with PsO who newly initiated APR and MTX. The MarketScan data comprise health services for more than 39.7 million patients through privately insured fee-for-service, point-of-service or capitated health plans. This database contains enrollment information and administrative claims data with healthcare utilization information (e.g., inpatient and outpatient services, and prescription drug claims). This study used deidentified patient records and did not involve the collection, use or transmittal of individually identifiable data; therefore, institutional review board approval to conduct this study was not necessary.

Prevalent patients with PsO were identified based on having at least one diagnosis code for PsO (International Classification of Diseases, 9th Revision, Clinical Modification [ICD-9-CM] code 696.1x or ICD-10-CM code L40.0, L40.8, L40.9) from a dermatologist or rheumatologist visit during the study period between 1 January 2014 to 31 March 2019 [15,16]. Patients were included if they initiated APR or MTX during the identification period (1 January 2015 to 31 March 2018). The date of the first claim for APR or MTX during the identification period was assigned as the index date. Patients were required to be at least 18 years of age on the index date, have continuous enrollment for at least 1 year prior to (baseline period) and 1 year after (follow-up period) the index date, and have at least one of the diagnosis claims for PsO in the baseline period or on the index date. Patients were excluded if they had claims for any systemic treatment agents (systemic-naive, including APR and MTX but not including symptomatic treatments) in the baseline period, biologic-indicated autoimmune conditions (ulcerative colitis, Crohn’s disease, rheumatoid arthritis and other inflammatory polyarthropathies, ankylosing spondylitis, juvenile idiopathic arthritis) or cancer (malignant neoplasms excluding nonmelanoma skin cancer) [17] in the baseline and follow-up periods, or had multiple systemic medications administered on the index date. A subset with 2 years of follow-up was also identified for a subgroup analysis.

Analyses were based on intention to treat, with individuals analyzed as part of their index treatment group regardless of subsequent changes in therapy. Demographic characteristics, prescriber specialty (defined as the specialty on the medical claim closest in time to the index date) and comorbidities, including the Charlson Comorbidity Index, as well as healthcare utilization and costs, were measured in the baseline period [18–20]. The primary outcomes were biologic initiation rate and time to biologic initiation, reported for the 1-year follow-up period in the main analysis and for the 2-year follow-up period in the subgroup analysis. Biologic initiation was defined as having a claim for a biologic therapy during the follow-up period, regardless of whether it was in addition to (add-on) or switch from the index therapy. The secondary outcomes were treatment patterns. Particularly, index therapy adherence was measured as the proportion of days covered during the follow-up period, defined as the number of days with index therapy available divided by the length of the observation period (365 days). Index treatment discontinuation was also reported for the follow-up period (defined as a ≥60-day gap in days’ supply). Restart of the index therapy was also measured (defined as reinitiating the index treatment following discontinuation). Descriptive statistics including means, standard deviations, and relative frequencies and percentages were reported for continuous and categorical data.

As the descriptive analyses used retrospective data rather than data from randomized trials, modeling was further performed to control for differences in observed characteristics of the two cohorts that may confound the findings. Logistic regression models were conducted to estimate the likelihood of biologic initiation during the 1-year follow-up period. Cox regression models were used to evaluate the risk of biologic initiation. All models were adjusted for the following: age group, gender, region, prescriber specialty, comorbid psoriatic arthritis (PsA), Charlson Comorbidity Index, index year, nonalcoholic fatty liver disease, serious infection, pain medication and glucocorticoid utilization, baseline healthcare utilization (in both inpatient and outpatient settings) and baseline healthcare costs (per US$1000). The odds ratio (OR) and 95% confidence interval (CI), as well as adjusted rates and 95% CI, were reported for the logistic regression model, whereas the hazard ratio (HR) and 95% CI were reported for the Cox regression model. In a subgroup analysis, estimations were replicated for a subgroup with 2 years of follow-up.

All data transformations and statistical analyses were performed using SAS^© version 9.4.

Among the total of 5860 systemic-naive PsO patients identified meeting the study criteria between 1 January 2015 and 31 December 2018, 3288 initiated APR and 2572 initiated MTX. Table 1 provides baseline characteristics for the study cohort. The mean age of APR initiators was 49.2 years versus 49.5 years for MTX initiators (p = 0.289). The percentage of females was similar among APR (53.9%) and MTX (55.1%) initiators (p = 0.36). Approximately 90% of each group was commercially insured. The prescriber specialty was significantly different between groups, with 22.2% of the MTX initiators receiving the index prescription from a rheumatologist compared with 2.5% of the APR initiators (p < 0.001). The mean number of Charlson comorbidities was similar among the two groups (APR vs MTX: 0.6 vs 0.5; p = 0.121), and 13.1% of APR users had a comorbid PsA diagnosis compared with 19.6% of the MTX group (p < 0.001). APR users were less likely than MTX users to have depression (9.9% vs 11.8%; p = 0.022) and hypertension (17.6% vs 20.0%; p = 0.017); however, for every other comorbidity of interest, the differences between the two groups were not statistically significant (Supplementary Table 2).

APR users were less likely than MTX users to be on pain medications (41.2% vs 49.0%; p < 0.001) and glucocorticoids (26.3% vs 36.4%; p < 0.001) at baseline (Table 1 & Supplementary Table 3). Last, mean baseline healthcare costs were higher among APR users than among MTX users (US$10,509 vs $8882; p = 0.005).

Fewer APR users initiated biologic treatment than MTX users throughout the 1-year follow-up period (Figure 1). During the first 3 and 6 months of follow-up, the unadjusted biologic initiation rates in the APR and MTX cohorts were 2.4% versus 10.7% (p < 0.001) and 6.8% versus 21.3% (p < 0.001), respectively. At the end of the 1-year follow-up, fewer APR patients (14.2%) than MTX patients (30.5%) initiated biologic treatment (Table 2).

The mean time to biologic initiation during the 1-year follow-up was 188.4 days in the APR cohort compared with 138.7 days in the MTX cohort (p < 0.001) (Table 2). The most frequently used first biologic treatment was adalimumab for both the APR and MTX cohorts (Supplementary Tables 6 & 7).

APR users were more compliant to their index therapies. The mean proportion of days covered for the index therapy was 0.57 for the APR cohort and 0.46 for the MTX cohort (p < 0.001). The discontinuation rate of the index therapy was 56.7% for APR users and 71.7% for MTX users during the 1-year follow-up period (p < 0.001). Among patients who discontinued their index therapy, 12.6% of APR users and 6.0% of MTX users restarted after an index treatment gap of more than 60 days within the follow-up period (p < 0.001) (Table 2).

After adjusting for potential confounders, patients treated with APR maintained a lower risk of biologic initiation when compared with patients treated with MTX (hazard ratio: 0.45; 95% CI: 0.40–0.51; p < 0.001) (Table 3 & Supplementary Table 8).

The logistic model showed that the likelihood of biologic initiation was statistically significantly lower with APR treatment even after adjusting for potential confounders (odds ratio: 0.42; 95% CI: 0.37–0.48; p < 0.001) (Supplementary Table 4) with adjusted rates of biologic initiation being 14.4% (95% CI: 13.2–15.7%) for the APR cohort compared with 28.6% (95% CI: 26.8–30.5%) for the MTX cohort.

To evaluate the robustness of the findings, the analyses were repeated for a subgroup of patients who had at least 2 years of continuous enrollment after their index date. Findings were similar in this subgroup to ones in the main analysis.

This study in an adult, systemic-naive PsO population with primary commercial or Medicare supplemental insurance found that patients treated with APR had a lower rate of, as well as longer time to, biologic initiation when compared with patients treated with MTX. These results were robust in patients with 1 and 2 years of follow-up after APR or MTX initiation, with differences being observed as early as 3 months post-index.

APR and MTX are both indicated for PsO, and real-world evidence shows that they are common first-line treatments for that disease. However, there are no randomized trials directly comparing APR to MTX in PsO. An indirect comparison based on data from clinical trials did not find a statistically significant difference in the efficacy of APR versus MTX [21].

Biologic initiation as an outcome may serve as a suitable proxy for suboptimal disease control in PsO, and it can easily be assessed from administrative claims databases. Previous studies demonstrated that higher disease activity is a predictor for biologic therapy initiation in PsO [22] and other autoimmune diseases such as rheumatoid arthritis [23–25].

Use of biologic initiation in PsO can be particularly helpful. Fluctuating symptoms, frequent dose adjustments and medication hoarding are typical in this condition. Therefore, other traditional treatment pattern measures used in claims studies, such as adherence or persistence, may be harder to interpret. Although low adherence is usually negatively associated with adequate control of disease, it is possible that patients with PsO experience periods of symptom control and temporarily pause treatment. This is consistent with the higher restart rate observed with APR.

In a cohort of patients with PsO, this study found that treatment with APR was associated with extended time to biologic initiation – the next line of treatment – when compared with MTX treatment. Future research using clinical measures could investigate whether biologic initiation is associated with disease progression, suboptimal disease control or toxicity in PsO. Research could also examine patient-specific and clinical factors associated with progression to biologic therapies.

Our study has several limitations. First, this study was a retrospective, observational analysis using large administrative claims data, thus randomized treatment assignment was not possible. Generally, administrative claims data lack possibly important clinical details providing information on disease severity and symptoms and thus does not allow to control for all potentially confounding variables. In particular, it is plausible that MTX users in our cohort have more severe disease than APR users and thus are more likely to switch to biologics. However, the modeling analyses included several measurable proxies of disease severity, such as prescriber specialty, pre-index healthcare costs and resource utilization, use of glucocorticoids and pain medications, and comorbidities/PsA. In addition, administrative claims data do not reflect whether medications are taken as prescribed. Thus, in studying adherence, only information regarding medication fills was considered. Finally, this study is limited to patients with commercial and Medicare supplemental insurance. Results may not be generalizable to other populations.

Systemic-naive adult PsO patients who initiated APR were more adherent to their index therapy and had a lower rate of biologic initiation when compared with patients initiating MTX. In addition, considering only patients who initiated a biologic during the follow-up period, the time to biologic initiation was longer among APR users than MTX users. APR use may delay biologic initiation in patients with PsO, suggesting better symptom control and outcome relative to MTX.

Direct clinical comparisons between APR and MTX in the treatment of PsO could be an important area of future research to better improve patient outcomes.

All authors made substantial contributions to the conception or design of the work, or the acquisition, analysis or interpretation of data for the work; drafted the work or revised it critically for important intellectual content; provided final approval of the version to be published; and made agreement to be accountable for all aspects of the work.

The study was funded by Amgen, Inc. The authors have no other relevant affiliations or financial involvement with any organization or entity with a financial interest in or financial conflict with the subject matter or materials discussed in the manuscript apart from those disclosed. E Chang, M Broder, C Paydar and K Bognar are employees of PHAR, which was hired by Amgen to conduct this study. J Yan was an employee of PHAR at the time of this study. S Richter was an employee of Amgen at the time of this study. P Desai was an employee of Amgen at the time of this study and owned stock in the company. I Khilfeh was an employee of Amgen at the time of this study.

No writing assistance was utilized in the production of this manuscript.

This study used deidentified patient records and did not involve the collection, use or transmittal of individually identifiable data; therefore, institutional review board approval to conduct this study was not necessary.

This work is licensed under the Attribution-NonCommercial-NoDerivatives 4.0 Unported License. To view a copy of this license, visit http://creativecommons.org/licenses/by-nc-nd/4.0/