Available Alternative Biologics and Disease Groups Influence Biologic Drug Survival in Patients with Psoriasis and Psoriatic Arthritis

The introduction of biologic agents over the past two decades has led to a significant shift in psoriasis and psoriatic arthritis (PsA) treatment paradigm1. However, biologic treatment costs have been increasing as biologic therapies became the standard of care for these patients2.

Better and more comprehensive methods to measure biologic drug effectiveness are needed to evaluate their cost-effectiveness. Improvement in clinical outcomes, such as the psoriasis area and severity index (PASI), has been the most frequently used effectiveness measure. However, considering only the clinical outcomes as a measure of biologic agent effectiveness in daily practice might not reflect the patients’ perception of treatment success3.

Drug survival is an increasingly used measure in daily practice to access biologic therapy effectiveness for psoriasis3. Drug survival was defined as the time interval between treatment initiation and discontinuation4. Drug survival is considered an overall marker for treatment success, as it informs on the effectiveness and safety of biologics in the real world3.

There are still significant gaps in understanding the long-term efficacy of biologics. Open-label extension studies have indicated that most patients are likely to maintain the effect for years5, 6, 7. However, patients treated with biologic agents in the real world often have to discontinue treatment or switch biologic agents over time due to loss of effectiveness or the development of adverse events8.

Dermatologic drug survival studies were used for head-to-head comparisons between biologic agents for psoriasis treatment5, 9. Head-to-head comparisons are still a major target of drug survival studies; however, the results should be carefully interpreted because many factors affecting drug survival could change over time4. For example, introducing new drugs superior to old ones might influence drug survival. With limited treatment options to switch to, physicians maximize their efforts to keep patients on the current treatment4. However, the presence of better alternatives might make physicians less hesitant to switch to other drugs. Nevertheless, this hypothesis has not been investigated clearly.

Additionally, it is unknown whether biologic drug persistence differs between treatments for PsA and psoriasis. Comorbid PsA is usually associated with biologic persistence in patients with psoriasis10. However, a comparison between PsA and psoriasis as the primary diseases for prescription has not been investigated yet. In Korea, interleukin (IL)-23/IL-17A and tumor necrosis factor (TNF)-α inhibitors received insurance benefits for PsA and psoriasis.

This study used the Health Insurance and Review Assessment (HIRA) data in Korea to investigate whether alternative availability and different disease groups could influence the drug survival of biologic agents approved for psoriasis and PsA.

We identified 4,852 patients with 6,053 biologic episodes of psoriasis or PsA during the study period. For ETA, ADA and IFX, 59 biologic episodes were excluded from the drug survival analysis because they had already received biologic treatment before the study period. We also excluded 360 biologic episodes with a single visit during the follow-up period. Finally, 5,634 biologic episodes were included in the drug survival analysis.

Patient characteristics

Table 1 presents the baseline characteristics of the 5,634 biologic episodes. UST was the most frequently prescribed biologic (n=2,488, 44.2%), followed by ADA (n=1,049, 18.6%). The overall mean age of the cohort was 44.8±13.8 years and males predominated (n=3,843, 68.2%). Three TNF-α inhibitors (ETA, ADA, and IFX) were prescribed to treat PsA more frequently than UST, SEC, GUS, and IXE. UST had the longest follow-up, with a median of 87.0 weeks (IQR 40.3~177.0 weeks), followed by ADA with a median of 66.0 weeks (22.1~136.6 weeks).

Table 1
Demographics based on biologic episodes at the time of biologic therapy

• Click for larger image
• Click for full table
• Download as Excel file

Predictive factors of biologic drug survival

Of the total 5,634 biologic episodes, 1,642 were terminated during the follow-up period. Table 2 presents the results of the univariable and multivariable analyses investigating predictors of drug survival. UST was considered the reference standard to which the other biologics were compared because it was the most frequently prescribed biologic in our cohort.

Table 2
Univariable and multivariable analyses for biologic drug discontinuation

• Click for larger image
• Click for full table
• Download as Excel file

The univariable time-dependent Cox regression analysis showed that female sex, patients with psoriasis alone, concomitant systemic therapy (cyclosporine, MTX, or acitretin), previous experience with biologics, no comorbidity, and being treated with a TNF-α inhibitor (ETA, ADA, or IFX) were significantly associated with discontinuation of the biologic treatment (Fig. 1). In the multivariable analysis, female sex, concomitant systemic therapy (cyclosporine or acitretin), treatment with a TNF-α inhibitor (ETA, ADA, or IFX), previous experience with biologics, and no comorbidity were still predictors of biologic discontinuation. Higher age was a predictor of biologic discontinuation and PsA (PsA alone, psoriasis with PsA) was a predictor of drug survival after adjusting for covariates. Concomitant MTX use was not associated with drug discontinuation after adjusting for covariates, although concomitant cyclosporine or acitretin use was associated with drug discontinuation. The insurance type was insignificant in the univariable and multivariable analyses in our cohort.

Fig. 1
Biologic drug survival in patients with psoriasis (PSO) or psoriatic arthritis (PsA). (A) Crude biologic drug survival in patients with PSO alone. (B) Crude biologic drug survival in patients with PSO and PsA. (C) Crude biologic drug survival in patients with PsA alone. FU: follow-up.

• Click for larger image
• Download as PowerPoint slide

We compared the drug survival of biologics in association with the disease groups in search of a difference in drug survival among diseases (Fig. 2). ADA showed longer drug survival in psoriasis with PsA (p<0.0001; hazard ratio [HR], 0.567; 95% CI, 0.444~0.726) and PsA alone (p<0.0001; HR, 0.555; 95% CI, 0.428~0.720) than in psoriasis alone (Fig. 2B). ETA and IFX showed similar results (Fig. 2A, C). UST showed higher drug discontinuation in PsA alone (p=0.0016; HR, 5.833; 95% CI, 2.160~15.751) than in psoriasis alone (Fig. 2D); although, the number of patients with PsA alone in UST was sparse (n=9). In contrast, SEC, GUS, IXE showed no significant differences in drug survival among disease groups (Fig. 2E~G). These findings suggested that the effects of disease groups might be different according to the different modes of action (TNF-α inhibitor vs. non-TNF-α inhibitor). Next, we analyzed data adding an interaction term, mode of action (TNF-α inhibitor [ETA, ADA, IFX] vs. non-TNF-α inhibitor [UST, SEC, GUS, IXE])×disease group. There was a significant interaction effect of mode of action×disease group after adjusting for covariates (p<0.0001). For non-TNF-α inhibitors, there was no significant difference in drug survival according to disease groups (p=0.1569). For TNF-α inhibitors, psoriasis with PsA (HR, 0.550; 95% CI, 0.452~0.670; p<0.0001) and PsA alone (HR, 0.591; 95% CI, 0.483~0.724; p<0.0001) were associated with longer drug survival compared with psoriasis alone.

Fig. 2
Drug survival for each biologic agent according to the primary disease. (A~C) The three tumor necrosis factor-α inhibitors showed longer drug survival in patients with psoriatic arthritis (PsA) (psoriasis with PsA, PsA alone) than in patients with psoriasis alone. (D~G) The interleukin (IL)-23/IL-17A inhibitors showed similar drug survival in psoriasis with PsA and psoriasis alone. The solid lines indicate drug survival in psoriasis alone, the dashed lines indicate drug survival in psoriasis with PsA, and the dotted lines indicate drug survival in PsA alone. FU: follow-up.

• Click for larger image
• Download as PowerPoint slide

Effects of alternatives’ availability on adalimumab and ustekinumab drug survival

We compared the ADA drug survival rate between before and after the availability of UST (June 2012) and the UST drug survival between before and after the availability of SEC (August 2017). The discontinuation rate of ADA started after the introduction of UST was higher than ADA started before UST became available (log-rank test, p=0.0097). The difference in ADA drug survival according to the availability of UST was insignificant at one year (p=0.3431; difference 6.2%; 95% CI, –6.6%~19.0%), but significant at two years (p=0.0377; difference 15.5%; 95% CI, 0.8%~30.2%).

We found the same trend in UST drug survival when comparing the periods before and after SEC became available (Fig. 3B). SEC is the first IL-17A inhibitor and superior to UST in efficacy14. The UST drug survival rate after the introduction of SEC was lower than before SEC introduction (log-rank test, p=0.0001). This trend was sustained for one year (p=0.0082; difference 5.1%; 95% CI, 1.3%~8.9%) and two years (p=0.0273; difference 6.8%; 95% CI, 0.8%~12.8%).

Fig. 3
The effect of alternative availability on biologic drug survival. (A) Adalimumab drug survival before and after the introduction of ustekinumab (June 2012) were compared. (B) Ustekinumab drug survival was evaluated according to the availability of secukinumab (as of August 2017). FU: follow-up.

• Click for larger image
• Download as PowerPoint slide

This study investigated drug survival of biologics for psoriasis or PsA using the HIRA data in Korea. We found that higher age, female sex, no comorbidity, concomitant cyclosporine or acitretin use, biologic-experienced, and use of TNF-α inhibitors were predictors of drug discontinuation. PsA was a predictor of drug persistence, particularly for TNF-α inhibitors. Drug discontinuation of UST and ADA significantly increased after the introduction of superior alternatives.

ADA and UST drug survival rates were compared in this study according to the availability of alternatives. We found that drug survival dropped after the introduction of a superior alternative. A recent meta-analysis of biologic drug survival suggested that UST showed a longer drug survival than SEC and TNF-α inhibitors15. UST was the first biologic to directly inhibit the IL-23/IL-17A pathway that received the US FDA approval (September 2009)16, followed by SEC (January 2015)16. In other words, UST was the only biologic targeting the IL-23/IL-17A pathway available for psoriasis for more than five years. This suggests that UST drug survival in previous reports might have been overestimated.

Comparisons of drug survival between biologics are the main purpose of drug survival analysis4. We found that TNF-α inhibitors had a lower drug survival rate than UST, consistent with previous studies8, 17. However, it was difficult to compare the drug survival of GUS and IXE because their follow-up was relatively short than that of other biologics. When biologics are switched in Korea, continuous treatment for over 6 months is strongly recommended under the insurance coverage.

We found an association between female sex and biologic discontinuation as in previous studies8, 10. A recent meta-analysis compared males and females and showed that the female sex had a pooled HR of 1.22 for discontinuation10. The reason for lower drug persistence rates in female patients is currently unknown, although adverse events seem to play a role in biologic discontinuation in female patients9.

Our results showed that concomitant MTX use was not associated with drug discontinuation although concomitant cyclosporine or acitretin use was associated with biologic drug discontinuation. Physicians may prescribe nonbiologic systemic agents during biologic therapy when they believe the patients are particularly recalcitrant. Unlike cyclosporine or acitretin, concomitant use of MTX is known to be a positive predictor for drug survival in patients with psoriasis probably due to its ability to reduce the immunogenicity of biologic agents18, 19, 20.

To the best of our knowledge, this study was the first to compare drug survival between patients with psoriasis and PsA. Previous studies on biologic drug survival were performed independently in patients with psoriasis or PsA. We only included biologics approved to treat both diseases in Korea and compared their drug survival rates. Little is known about the comparative drug effectiveness in different disease groups. Comparing drug survival between disease groups could provide insight into whether knowledge could be exchanged between those groups4.

Our results demonstrated that patients with PsA (PsA alone, psoriasis with PsA) showed longer biologic drug survival than patients with psoriasis alone. These findings were distinct in TNF-α inhibitors. It can be explained by the fact that non-TNF-α inhibitors were classified as second-line drugs for PsA during the study period. However, the patterns differed depending on individual biologics although all non-TNF-α inhibitors were second-line drugs during the study period. UST showed higher drug discontinuation in PsA alone, whereas the biologic survival rates of SEC, GUS, and IXE among disease groups were not significantly different.

PsA might influence drug survival of biologics because only some of the biologics were more efficient in PsA than psoriasis. A recent meta-analysis found few statistically significant differences between most biologic treatments of PsA21, while TNF-α inhibitors consistently showed a higher drug discontinuation rate than UST in patients with psoriasis15, 22. Additionally, patients with PsA might be more frequently motivated by their physicians for treatment persistence or recognize the importance of treatment persistence better than patients with psoriasis in whom only the skin is involved10.

Our results showed that age and number of comorbidities predicted biologic drug persistence, in contrast to previous studies5, 9. This difference could be partly explained by differences in the biologics included in the analysis. Previous studies included only TNF-α inhibitors and UST, whereas our cohort primarily consisted of IL-23/IL-17A inhibitors rather than TNF-α inhibitors. Comorbidities such as hypertension, renal disease, and hepatic disease are contraindications for conventional systemic therapy23. Thus, patients with comorbidities and their physicians would adhere to biologic therapy unless serious adverse events occur. The incidence of severe adverse effects in patients with psoriasis receiving IL-23/IL-17 inhibitors is lower than in patients with psoriasis receiving TNF-α inhibitors24. Therefore, patients with comorbidities receiving IL-23/IL-17A inhibitors might be less likely to stop biologic therapy due to adverse events than those treated with TNF-α inhibitors.

Similar to other studies using claims data, this study had some limitations. First, we did not know the reasons for discontinuation. However, previous studies showed that most biologic drug discontinuations in patients with psoriasis were due to a lack of efficacy5, 9. Second, the data used in this study did not include information about disease severity, such as PASI or the Disease Activity Score-28. However, biologics receive reimbursement benefits for patients with severe and recalcitrant psoriasis or PsA. For example, to be eligible for the benefits, patients with psoriasis must have a body surface area of ≥10% and PASI of ≥10, even after a 12-week continuous treatment with systemic therapy or phototherapy. The situation is similar for PsA. In other words, the prescription of biologics for patients with psoriasis or PsA covered by the NHI in Korea means that these patients have severe and recalcitrant disease.

Finally, biological naivety in our cohort may have been different from actual biological naivety. Some biologic-experienced patients with PsA may have been misclassified as biologic-naïve because our operational definition of naivety was patients who had not been prescribed biologics since 2009. This misclassification may have occurred only for PsA because the insurance coverage of biologics for psoriasis in Korea started after 2009. Patients with PsA treated with biologics before 2009 were classified as biologic-naïve. However, the proportion of such patients was minimal in this study, and misclassification of biologic-experienced as biologic-naïve would have decreased the statistical difference between the groups. Thus, misclassification did not alter our primary results.

We found that drug survival in patients with PsA was longer than in patients with psoriasis, particularly for TNF-α inhibitors, and the availability of alternatives decreased drug survival of biologics. Drug survival is a representative real-world outcome measure of the use of biologics for psoriasis. However, a comparison of drug survival between biologics should be carefully considered if their introduction times differ. Our results also suggested that attitudes toward treatment persistence might affect drug survival.