Abstract
Introduction
Psoriatic arthritis (PsA) is a chronic inflammatory arthritis that occurs in people affected by the autoimmune disease psoriasis. The cost effectiveness of secukinumab in PsA has not been evaluated in Germany.
Objective
The purpose of this study was to conduct a cost-utility analysis of secukinumab in three adult populations with active PsA in Germany: biologic naïve without moderate or severe plaque psoriasis, biologic naïve with moderate or severe plaque psoriasis, and biologic experienced. Comparators included other disease-modifying antirheumatic drugs (DMARDs), including biosimilar versions as well as standard of care.
Methods
The analysis took the viewpoint of the German statutory health insurance. We adapted a decision analytic semi-Markov model to evaluate the cost effectiveness of secukinumab over a lifetime horizon. Treatment response was assessed based on PsA Response Criteria at 12 weeks. Nonresponders or patients discontinuing the initial-line DMARD were allowed to switch to subsequent-line DMARDs. Model input parameters (Psoriasis Area Severity Index, Health Assessment Questionnaire (HAQ), withdrawal rates, costs, and resource use) were collected from clinical trials, published literature, and official reports. Health benefits were expressed as quality-adjusted life-years. An annual discount rate of 3% was applied to costs and benefits. The robustness of the study findings was evaluated via sensitivity analyses.
Results
In the biologic-naïve population without moderate or severe plaque psoriasis, secukinumab 150 mg either strictly dominated other DMARDs (certolizumab pegol, golimumab, and ustekinumab) or yielded favorable incremental cost-effectiveness ratios (ICERs) (vs. etanercept, adalimumab, and infliximab). In the biologic-naïve population with concomitant moderate to severe plaque psoriasis and in the biologic-experienced population, secukinumab 300 mg was more effective and had a lower ICER than other DMARDs, thus leading to extended dominance. Deterministic sensitivity analyses indicated that the results were most sensitive to the discount rate for costs and health outcomes as well as the HAQ score as an input to utility values.
Conclusions
Secukinumab appears to be cost effective compared with other DMARDs for the treatment of active PsA in biologic-naïve and biologic-experienced populations in Germany.
We’re sorry, something doesn't seem to be working properly.
Please try refreshing the page. If that doesn't work, please contact support so we can address the problem.
In the treatment of active psoriatic arthritis in Germany, secukinumab appears to be cost effective compared with other drugs. |
This also appears to hold compared with biosimilar versions. |
1 Introduction
Psoriatic arthritis (PsA) is a chronic inflammatory arthritis that occurs in people affected by the autoimmune disease psoriasis. The main symptoms of PsA are joint arthritis and skin psoriasis. Active disease has been defined as one or more tender and inflamed joints and/or tender enthesitis points and/or dactylitic digits and/or inflammatory back pain [1]. In Germany, the prevalence of PsA has been estimated at 0.28% in a sample of statutory health insurance (SHI) enrollees aged > 18 years [2]. The prevalence of active PsA requiring systematic treatment has been estimated at 0.12% [2].
The management of PsA comprises both nonpharmacological and pharmacological measures. Pharmacological therapies for PsA include nonsteroidal anti-inflammatory drugs (NSAIDs) and the so-called disease-modifying antirheumatic drugs (DMARDs). According to the most recent European League Against Rheumatism guideline for the management of PsA, published in 2015 [1], conventional synthetic DMARDs (csDMARDs) are recommended as therapy after failure of NSAIDs and local therapy for active disease. Methotrexate is suggested as the first-choice csDMARD. Although evidence on the efficacy of csDMARD combinations is scarce, they may still be considered. If necessary, these are followed by a biological DMARD (bDMARD) or a targeted synthetic DMARD (tsDMARD). The first bDMARD would usually be a tumor necrosis factor (TNF)-α inhibitor such as infliximab, etanercept, or adalimumab. bDMARDs targeting interleukin-12/23 (ustekinumab) or interleukin-17 pathways (secukinumab) may be used in patients for whom TNF inhibitors are inappropriate. If the first bDMARD strategy fails, another bDMARD or a tsDMARD such as the phosphodiesterase 4-inhibitor apremilast may be prescribed.
According to its summary of product characteristics, secukinumab alone or in combination with methotrexate is indicated for the treatment of active PsA in adult patients for whom the response to previous DMARD therapy has been inadequate [3]. For TNF inhibitor-naïve patients with concomitant moderate to severe plaque psoriasis or patients with inadequate responses to anti-TNFα agents, the recommended dose of secukinumab is 300 mg. For other patients, the recommended dose is 150 mg, which can be increased to 300 mg according to clinical response. The efficacy and tolerability of secukinumab in adult patients with active PsA has been demonstrated in both TNF inhibitor-naïve and -experienced patients compared with placebo (FUTURE 1 and FUTURE 2 randomized phase III clinical trials) [4, 5].
For the purpose of the mandatory early benefit assessment in Germany, the German Federal Joint Committee [6] defined a TNF-α inhibitor (i.e., etanercept or adalimumab or infliximab or golimumab, each with or without methotrexate) as an appropriate comparator of secukinumab in biologic-experienced patients with PsA. In biologic-naïve patients with PsA, it defined adalimumab or certolizumab or etanercept or golimumab or infliximab as appropriate comparators [7].
Cost-effectiveness analyses of secukinumab have recently been published for both biologic-naïve and biologic-experienced adults with active PsA who responded inadequately to prior csDMARDs and/or bDMARDs [8,9,10]. These analyses were based on Canadian, Finnish, and UK data. Given that the efficacy and tolerability of secukinumab has not been demonstrated compared with active comparators, the analyses were based on an indirect treatment comparison using a network meta-analysis (NMA) with placebo as a bridging comparator (published by McInnes et al. [11] or a variation thereof). To the best of our knowledge, the cost effectiveness of secukinumab in PsA has not been evaluated in Germany.
In Germany, the TNF-α inhibitors adalimumab, infliximab, and ustekinumab and the bDMARDs secukinumab and etanercept were among the top 20 drugs by gross sales in 2018 [12]. The annual acquisition costs of these drugs are around €20,000 per annum in Germany. For these reasons, it is of public interest to elucidate their value in terms of potential cost offsets and health gains. An evaluation is also relevant from a European standpoint as the cost effectiveness of the 300-mg dose has thus far only been assessed in the Finnish PsA population [10]. The purpose of this study was therefore to conduct a cost-utility analysis (CUA) of secukinumab in three adult populations with active PsA in Germany: (1) biologic naïve without moderate or severe plaque psoriasis, (2) biologic naïve with moderate or severe plaque psoriasis, and (3) biologic experienced. Comparators included other currently licensed bDMARDs and—for the biologic-experienced population—apremilast in addition (the German Federal Joint Committee [7] did not consider apremilast to be an appropriate comparator in the biologic-naïve population). The analysis took the viewpoint of the German SHI and therefore only accounted for direct medical costs.
2 Methods
2.1 Patient Population and Setting
The CUA was conducted in biologic-experienced (at least one bDMARD) and biologic-naïve (with or without moderate or severe plaque psoriasis) adults, representing the target population of secukinumab according to its label. Based on the demographic characteristics of the randomized trial population of FUTURE 2 [4], both populations were assumed to have an average age of 48 years, a mean weight of 87.11 ± standard deviation 19.66 kg, and a baseline utility weight of 0.479; 48% of the cohort was male. The setting was primary and secondary outpatient care.
2.2 Model Structure
We used the same model as Goeree et al. [9] and Purmonen et al. [10] for the Canadian and Finnish healthcare setting, respectively (see Fig. 1 for the model structure). The model was reviewed by the UK National Institute for Health and Care Excellence [13] and the Swedish Dental and Pharmaceutical Benefits Agency [14]. Both assessments resulted in positive decisions for secukinumab. We adapted the model to the German SHI setting using relevant medical and cost inputs. Unlike the other studies, we used data on the comparative effectiveness of various DMARDs from a recently published NMA [15] (see “Clinical Inputs” for details). At entry into the model, patients received a 3-month induction treatment (see Table 1 for administration frequencies). Thereafter, their transition depended on the clinical response based on the PsA Response Criteria (PsARC), the discontinuation rate, and the occurrence of serious infections, malignancy, or death. Once patients entered the malignancy state, they could transition to death, with the probabilities of death depending on the time since malignancy. At 5 years post-malignancy, mortality rates reverted back to those of the general population.
Reproduced from Purmonen et al. [10]
Model structure. *The efficacy parameter is based on the Psoriatic Arthritis Response Criteria. tx treatment.
Once patients discontinued from an initial bDMARD or apremilast, they were allowed to switch to another bDMARD or apremilast. Subsequent-line treatment was modeled using the mean efficacy, costs, and adverse event (AE) incidence rates of all biologics and apremilast. That is, subsequent DMARDs were assigned an equal probability of being chosen.
Health states for subsequent treatment lines replicated those for the initial modeled line of treatment. Patients were assumed to continue on the subsequent treatment for the remainder of the model time horizon or to switch to standard of care (SoC), i.e., methotrexate plus sulfasalazine and leflunomide. In the base case, patients entering SoC treatment were assumed to be ineligible to switch back to biologics or apremilast. In a sensitivity analysis, the discontinuation rate was lowered by 20%, thus indirectly assuming a certain fraction of back transitions.
Given that survival probabilities for patients with malignancy depended on the time since the initial occurrence of a malignancy, the model took the form of a semi-Markov model. The cycle length was 3 months, given that PsARC and Psoriasis Area and Severity Index (PASI) responses (see “Clinical Inputs”) are based on 3-month data.
2.3 Model Inputs
2.3.1 Clinical Inputs
The main clinical inputs for the model were the primary efficacy measure PsARC, the PASI response distribution, and the change in the Health Assessment Questionnaire (HAQ) (each at 3 months). PsARC is a composite measure that requires improvement in two (with at least one being a joint score) and worsening in none of the following four factors [16]: patient and physician global assessments (improvement defined as decrease by ≥ 1 unit; worsening defined as increase by ≥ 1 unit) and tender and swollen joint scores (the sums of all joints scored; improvement defined as decrease by ≥ 30%; worsening defined as increase by ≥ 30%).
In the absence of head-to-head comparisons between bDMARDs and/or apremilast, we searched PubMed on 21 April 2019 for meta-analyses reporting the comparative effectiveness of bDMARDs and apremilast based on the PsARC as a clinical endpoint. We obtained 14 hits with the search algorithm “secukinumab psoriatic arthritis response meta-analysis”. We chose the meta-analysis by McInnes et al. [15] because of its comprehensiveness in terms of comparators and endpoints. This Bayesian NMA was based on a systematic review of literature published up to November 2015 and included 20 randomized controlled trials conducted in biologic-naïve and biologic-experienced patients with possible background exposure to methotrexate and corticosteroids in the active and placebo arms. In trials that reported data from biologic-experienced populations, disparities in inclusion criteria meant that patients receiving certain agents might have experienced several previous anti-TNF therapy failures, whereas other populations were permitted to have received only one previous anti-TNF therapy. For trials that reported data from biologic-naïve populations, potential heterogeneity with regard to prior therapies (e.g., systematic therapies or NSAIDs) was not further elucidated. In addition, it was not specified what percentage of patients received concomitant methotrexate. Using AMSTAR, a measurement tool for the “assessment of multiple systematic reviews” [17], the methodological quality was not found to be adequate or adequately described (only 1 of 11 items could be clearly confirmed based on the main text and the supplementary materials), but it was still an improvement over the meta-analyses used in the abovementioned economic evaluations.
The NMA assessed the PsARC response at 12–16 weeks. PsARC response with secukinumab compared with placebo was lower in the biologic-naïve population than in the mixed biologic-naïve and biologic-experienced population but did not achieve statistical significance. Therefore, we applied PsARC responses from the mixed population to both populations. As the NMA only presented relative and not absolute risks, we took the probability of PsARC response under placebo (i.e., SoC) at 3 months from a Bayesian mixed-treatment comparison [18] and modeled responses with DMARDs based on the relative risks compared with placebo (i.e., SoC). See Table 2 for the PsARC response by comparator.
For the PASI response, which captures severity of skin lesions and the extent of the affected area, we took 12-week data on secukinumab and placebo (i.e., SoC) from the FUTURE 2 trial [4], similarly to Buchanan et al. [8]. Data on other biologics and apremilast were modeled based on the relative risks for the PASI 50, 75, and 90 response published in the NMA by McInnes et al. [15]. Given that this publication did not separately report the PASI response in the biologic-experienced population, we followed the approach for the PsARC response and applied the response rate in the mixed population. For each PASI level, we assumed patients would achieve at least the minimum (i.e., PASI 75 responders would achieve at least a 75% reduction in baseline PASI, as in Rodgers et al. [18]). See Table 3 for the PASI response by comparator. Depending on the PASI response, in the biologic-naïve population with moderate or severe plaque psoriasis, patients may or may not have achieved the same plaque psoriasis severity after treatment as patients with no moderate or severe plaque psoriasis in the first place. As the PsARC and PASI responses are not necessarily independent, their correlation was incorporated [18].
The impact of treatment on the arthritis component of PsA was modeled via changes in the HAQ score [4, 19]. The baseline HAQ score (from the FUTURE 2 trial [4]) represented the average HAQ score for the patient cohort at the start of treatment. Patients were assumed to experience a different change in HAQ based on the presence or absence of a PsARC response (see Table 4). Data on the effectiveness of secukinumab and SoC were taken from the FUTURE 2 trial [4]. For patients who discontinued biologic treatment or apremilast and thus entered the SoC health state, we applied the same “withdrawal” rates as did Purmonen et al. [10, Table S3]. In the base case, we conservatively assumed that the HAQ score returned to the natural history course. This presupposes a “catch-up” effect such that, after withdrawal, the HAQ score is the same irrespective of whether the patient received active treatment. In a sensitivity analysis, we assumed that, for patients who withdrew, the level of disability (in terms of HAQ score) deteriorated by the same amount as it improved when they initially responded to treatment (i.e., the rebound was equal to the initial gain). Whereas the Finnish study [10] applied the same base-case assumption as in our study, in the Canadian study [9], the base-case and alternative assumption were swapped.
2.3.2 Cost and Resource Use
The model only incorporated direct medical costs, which included drug acquisition costs, disease-related costs, laboratory costs, and AE costs. All costs were inflated to €, year 2018 values based on the general German Consumer Price Index.
We did not exclude drug copayments, which cover a portion of the (rebated) drug price, because they depend on factors such as insurance tariff, comorbidities, and income. By including copayments, our analysis, strictly speaking, holds the perspective of SHI members, which remains in line with recommendations for German cost-effectiveness analyses [20].
Drug acquisition costs were extracted in July 2018 from the German Drug Directory (LAUER-TAXE®) [21], a comprehensive database offering pharmacy retail prices and rebates of prescription drugs in Germany. Given the SHI perspective of our analysis, we calculated pharmacy retail prices net of mandatory rebates for the SHI (Table 5). We separately considered acquisition costs of biosimilars for etanercept and infliximab. Biosimilars for adalimumab have been available on the German market only since October 2018.
Disease-related costs were disaggregated into the arthritis and psoriasis components of PsA (Table 6). Arthritis-related costs were estimated via a linear regression based on HAQ scores [18]. Psoriasis-related costs were modeled based on PASI response, incorporating the costs for uncontrolled psoriasis, defined as the proportion of patients not reaching the PASI 75 response, and the cost for controlled psoriasis (i.e., patients reaching PASI 75). Based on the best available evidence from Germany (an observational study over 12 months conducted in 2002 [23]), the difference in SHI costs between patients with moderate and severe disease severity (defined as 10–20% and 20% affected body surface area, respectively) was not significant. Information on the PASI itself was not collected in this study. Therefore, we applied the costs (net of medication costs) of the mixed population to both PASI 75 responders and nonresponders.
The laboratory tests included full blood count, erythrocyte sedimentation rate, liver function tests, urea and electrolytes test, chest radiograph, tuberculosis immunoassay, antinuclear antibodies, and anti-double-stranded deoxyribonucleic acid antibodies (Table 7). The corresponding costs were obtained from the German ambulatory physician fee schedule (year 2018 data [24]).
Modeled AEs included malignancy, tuberculosis, and other serious infections (including inflammatory bowel disease, septicemia, bronchopneumonia, kidney or unitary tract infection, lower respiratory disease, and bronchitis). Costs of AEs are reported in Table 8 and were based on the German diagnosis-related group reimbursement rates (2018) and cost-of-illness data from 2015.
2.3.3 Utility Weight Inputs
This CUA used quality-adjusted life-years (QALYs) as a measure of health outcome. QALYs combine life-years and strength of preference for different health states. In the base case, utility values were derived from a linear mixed model that used patient-level EuroQol 5-Dimensions, 3-Levels (EQ-5D-3L) data from the FUTURE 2 trial [4], which was conducted at 76 centers across Asia, Australia, Europe, and North America. FUTURE-2 EQ-5D responses were converted to utility weights using the UK value set [27]. Utility values during follow-up were estimated based on absolute HAQ and PASI scores, baseline EQ-5D, age, sex, and prior TNF inhibitor status. In an alternative scenario, we applied the parameters obtained from Rodgers et al. [18] (see Table 9).
2.3.4 Mortality Inputs
Three types of mortality inputs were considered in the analysis: disease-specific mortality, AE-related mortality, and all-cause mortality. Patients with PsA have a higher mortality rate than the general population. We took a mortality ratio of 1.65 for males and 1.59 for females from a cohort study by Wong et al. [28]. Hazard ratios of death due to infection and malignancy were taken from a study of patients with psoriasis [29]. Hazard ratios for malignancy are available for up to 5 years. We assumed that mortality increased in the first year post-malignancy and then returned to baseline levels after 5 years. Data on all-cause mortality in the general population were obtained from the most recent (2015/2017) German life table data at the time of preparing the manuscript [30].
2.4 Base-Case Analysis
The analysis evaluated the cost effectiveness of secukinumab 150 mg (in the biologic-naïve population with no moderate or severe plaque psoriasis) and secukinumab 300 mg (in the biologic-naïve population with concomitant moderate to severe plaque psoriasis and in the biologic-experienced population) compared with other licensed biologics and apremilast. To evaluate the cost effectiveness of secukinumab, costs and QALYs relative to other modeled biologics and apremilast were assessed for strict dominance, i.e., secukinumab having higher QALYs at a lower cost. When secukinumab did not strictly dominate the comparator, incremental cost-effectiveness ratios (ICERs) were calculated. “As options ruled out by extended dominance should be identified” [31] (p. 357), we assessed whether secukinumab had higher QALYs and ICERs relative to other biologics and apremilast. To this end, ICERs were calculated compared with SoC (while realizing that, for an official cost-effectiveness analysis in Germany, the appropriate comparators would need to be determined by the German Federal Joint Committee according to § 35b Social Code Book V).
The primary analysis was conducted over a lifetime horizon (i.e., disease duration of 60 years), which allowed us to consider a continuation of treatment given the chronic nature of the disease. For the base-case analysis, we discounted both costs and effects at the official rate of 3% per annum [20].
2.5 Sensitivity Analysis
In one-way deterministic analyses, we assessed parameter uncertainty by varying input parameters that are subject to variation one at a time using the limits of the 95% confidence interval or, if unavailable, a 20% percentage decrease and increase (Tables 10, 11, 12, 13). In addition, we assessed structural uncertainty in four alternative scenarios. In the first alternative scenario, we assumed that patients who withdrew from active treatment had an HAQ deterioration in the cycle following treatment withdrawal that corresponded to the initial HAQ gain rather than the base-case assumption that HAQ rebounded back to natural history. In the second alternative scenario, utility weights were obtained from the York model [18], whereas, in the base-case analysis, they were obtained from the FUTURE 2 trial [4]. In the third alternative scenario, we assumed that average efficacy declined by 50% for subsequent-line treatment, in agreement with observational data from Denmark and Norway [32, 33]. In the fourth scenario, we assumed that 47% of patients received concomitant methotrexate (thus following Goeree et al. [9]), instead of the base-case assumption of 100%. Moreover, we assessed methodological uncertainty (1) using a 5- and 10-year time horizon instead of 60 years in the base-case analysis and (2) varying the discount rate between 0 and 5% [20].
In a probabilistic sensitivity analysis (PSA), we randomly sampled input parameters from their assigned distribution during simulations.
3 Results
3.1 Base-Case Results
In the biologic-naïve population without moderate or severe plaque psoriasis, secukinumab 150 mg achieved 8.67 QALYs at a cost of €318,469 over a lifetime horizon (Table 14). Secukinumab 150 mg strictly dominated certolizumab pegol, adalimumab, golimumab, and ustekinumab, as it produced more QALYs at a lower cost (Table 15). Compared with the brand-name versions of etanercept and infliximab, secukinumab 150 mg yielded ICERs in the range of several hundreds of thousands of euros but each for a QALY loss, indicating that the QALY loss was associated with substantial savings. When compared with biosimilar versions of etanercept and infliximab, ICERs still fell within that range (Table 15).
In contrast, in biologic-naïve patients with moderate or severe plaque psoriasis, secukinumab 300 mg did not strictly dominate any of the other biologics (Table 15). Still, as the 300-mg dose was more effective and had a lower ICER than SoC (Table 16), it extendedly dominated these drugs and thus can be interpreted as cost effective. This result also holds against the biosimilar versions of etanercept and infliximab.
Results are comparable between the biologic-experienced and the biologic-naïve population with moderate or severe plaque psoriasis: Secukinumab 300 mg was more effective and had a better ICER (Table 16). This result also held when compared with biosimilar versions of etanercept and infliximab.
3.2 Sensitivity Analysis
In one-way sensitivity analyses across all patient groups, the discount rate for costs and health outcomes as well as the HAQ score as an input to utility values were the input parameters with the largest influence on the ICER (Fig. 1 in the Electronic Supplementary Material).
The impact of structural and methodological uncertainty on the base-case results is shown in Table 17. Assuming that patients who withdrew rebounded by the initial treatment gain led to lower ICERs of all drugs compared with SoC but also increased the ICER of secukinumab relative to that of other drugs in the biologic-experienced population and hence resulted in a partial loss of extended dominance in that population. Similar results were obtained when applying 5- and 10-year time horizons. In contrast, the other scenarios confirmed the direction of the base-case results.
Scatterplots (Fig. 2) show the distribution of incremental costs and QALYs for each subpopulation using 1000 Monte-Carlo iterations. The cost-effectiveness acceptability curves indicated that secukinumab had a higher probability of being cost effective than any other DMARD across the three patient populations (Fig. 3).
Scatterplots. a Biologic-naïve population with no moderate or severe plaque psoriasis, b biologic-naïve population with moderate or severe plaque psoriasis, c biologic-experienced population
Cost-effectiveness acceptability curves for the following populations: a biologic naïve with no moderate or severe plaque psoriasis; b biologic-naïve with moderate or severe plaque psoriasis; c biologic experienced. ADA adalimumab, APR apremilast, CER P certolizumab pegol, ETN etanercept, GOL golimumab, INF infliximab, SEC secukinumab, SoC standard of care, UST ustekinumab
4 Discussion
This CUA investigated the cost effectiveness of secukinumab compared with other biologics (adalimumab, certolizumab pegol, etanercept, golimumab, ustekinumab, infliximab) and the nonbiologic apremilast in patients with active PsA over a lifetime horizon. To this end, the analysis built upon a recently published NMA on biologic-naïve and biologic-experienced patients [15]. In summary, our analysis shows that, compared with other branded bDMARDs, secukinumab 150 mg appeared clearly cost effective. A dose of 300 mg also appeared cost effective when applying the concept of extended dominance. However, when using a threshold ICER of, for example, €50,000 per QALY gained or less, bearing in mind that no official threshold exists in Germany, secukinumab 300 mg would not be considered cost effective, but neither would the comparator drugs. Given that secukinumab’s comparators are implicitly cost effective because they are official comparators (in agreement with §92 section 1 of Social Code Book no. 5), by deduction, secukinumab must also be considered cost effective. This also holds for biosimilar versions of comparators.
Still, we observe that changing the level of HAQ decline after treatment withdrawal had a relevant impact on treatment ICERs and the interpretation of cost effectiveness in the biologic-experienced population. As in the real world, some patients may rebound and others may not [34], it needs to be acknowledged that neither the base-case nor the alternative scenario is likely to completely capture reality, with the “true” scenario most likely falling in between the two we analyzed.
Similar to our study, the Canadian CUA [9] showed strict dominance of secukinumab 150 mg over other bDMARDs in the biologic-naïve population and a favorable ICER compared with an infliximab biosimilar (other biosimilars were not assessed). In contrast to our study, the 300 mg dose was also strictly dominant over bDMARDs, even including an infliximab biosimilar (apremilast was not assessed). This result was confirmed in a sensitivity analysis assuming that HAQ returned to the natural history course. In the Canadian setting, the default assumption was that 47% of patients received concomitant methotrexate; however, the Finnish study [10] and our study used a rate of 100%. Changing this rate had little impact over ICERs in our model.
Again, similar to our results, in the Finnish CUA [10], secukinumab 150 mg either strictly dominated its comparators or showed a favorable ICER (below the assumed, unofficial threshold of €30,000 per QALY gained). Comparators included the biosimilar versions of adalimumab and etanercept. Similar to our model, the 300 mg dose demonstrated a considerable change in cost effectiveness in both biologic-naïve and biologic-experienced populations according to the assumption of the level of HAQ decline after treatment withdrawal. In the experienced population, secukinumab 300 mg may or may not be cost effective compared with certolizumab pegol, etanercept, adalimumab, golimumab, and apremilast depending on this assumption.
According to the UK model [8], which only tested secukinumab 150 mg, secukinumab also either strictly dominated its comparators or showed a favorable ICER (vs. the official threshold of £20,000 per QALY gained). Comparators included biosimilar versions of etanercept and infliximab.
4.1 Limitations
One limitation of our analysis lies in the NMA of McInnes et al. [15], which only included short-term efficacy data over 12–16 weeks. This period may be too short to evaluate the long-term effects of secukinumab over other drugs. In addition, the NMA was limited by the heterogeneity of patient populations in terms of background exposure to methotrexate or corticosteroids, disease activity, and baseline PASI scores. This limitation also affects other meta-analyses in the treatment of active PsA (e.g., the recent meta-analysis by Strand et al. [35]). Therefore, even using the best available evidence, heterogeneity is unavoidable.
Moreover, for the purpose of conducting the PSA, correlations between the relative risks of secukinumab and other drugs were not available from the NMA by McInnes et al. [15]. This information could be important, as relative risks show a large spread, thus potentially leading to reverse orderings of relative risks in the PSA.
Adaptation of the model to the German SHI setting was limited by some of its structural assumptions. For example, the model did not include tests required by label such as annual skin cancer tests (€22.80, ambulatory fee no. 01745) or one-time hepatitis B test (€5.50, ambulatory fee no. 32781). Yet, they cannot be assigned to other tests required for all drugs because both skin cancer and hepatitis B testing is required for adalimumab, infliximab, and golimumab, but only hepatitis B testing is required for etanercept. In addition, some tests are required to be conducted more or less frequently than preset by the model, e.g., tuberculosis testing (€58.00); however, strictly speaking, this is not included in label for secukinumab, and other drugs include it at different frequencies. On the other hand, as the costs of these tests are comparatively small, the bias is also likely to be small.
Finally, an assessment of the cost effectiveness of secukinumab, strictly speaking, also needs to take into consideration its other labelled indication, plaque psoriasis. In Germany, the price of a drug does not vary by indication, therefore, when it comes to pricing and reimbursement, lower cost effectiveness in one indication can be “compensated” for by increased cost effectiveness in another.
5 Conclusion
For the treatment of active PsA in biologic-naïve and biologic-experienced populations, secukinumab appears to be cost effective compared with other DMARDs from a German payer’s perspective.
Data Availability
The datasets used for the cost-effectiveness analysis are not publicly available for proprietary reasons.
References
Gossec L, Smolen JS, Ramiro S, de Wit M, Cutolo M, Dougados M, Emery P, Landewé R, Oliver S, Aletaha D, Betteridge N, Braun J, Burmester G, Cañete JD, Damjanov N, FitzGerald O, Haglund E, Helliwell P, Kvien TK, Lories R, Luger T, Maccarone M, Marzo-Ortega H, McGonagle D, McInnes IB, Olivieri I, Pavelka K, Schett G, Sieper J, van den Bosch F, Veale DJ, Wollenhaupt J, Zink A, van der Heijde D. European League Against Rheumatism (EULAR) recommendations for the management of psoriatic arthritis with pharmacological therapies: 2015 update. Ann Rheum Dis. 2016;75(3):499–510.
Elsevier Health Analytics. Übersichtsbericht der Ergebnisse: Versorgungsforschung Psoriasis (Pso), Psoriasisarthritis (PsA) und axiale Spondyloarthritis (AxSpA) in Deutschland; eine retrospektive Analyse auf Routinedaten; 2017.
Cosentyx. Summary of product characteristics. Last updated: 25/10/2018. https://www.ema.europa.eu/documents/product-information/cosentyx-epar-product-information_en.pdf (accessed 10 Aug 2019).
McInnes IB, Mease PJ, Kirkham B, Kavanaugh A, Ritchlin CT, Rahman P, et al. Secukinumab, a human anti-interleukin-17A monoclonal antibody, in patients with psoriatic arthritis (FUTURE 2): a randomised, double-blind, placebo-controlled, phase 3 trial. Lancet. 2015;386(9999):1137–46.
Mease PJ, McInnes IB, Kirkham B, Kavanaugh A, Rahman P, van der Heijde D, et al. Secukinumab inhibition of interleukin-17A in patients with psoriatic arthritis. N Engl J Med. 2015;373(14):1329–39.
Gemeinsamer Bundesausschuss. Tragende Gründe zum Beschluss des Gemeinsamen Bundesausschusses über eine Änderung der Arzneimittel-Richtlinie (AM-RL): Anlage XII - Beschlüsse über die Nutzenbewertung von Arzneimitteln mit neuen Wirkstoffen nach § 35a SGB V—Secukinumab (neues Anwendungsgebiet). Gemeinsamer Bundesausschuss; 2. Juni 2016.
Gemeinsamer Bundesausschuss. Tragende Gründe zum Beschluss des Gemeinsamen Bundesausschusses über eine Änderung der Arzneimittel-Richtlinie (AM-RL): Anlage XII – Beschlüsse über die Nutzenbewertung von Arzneimitteln mit neuen Wirkstoffen nach § 35a SGB V—Ixekizumab (neues Anwendungsgebiet: Psoriasis-Arthritis). Gemeinsamer Bundesausschuss; 16. August 2018.
Buchanan V, Sullivan W, Graham C, Miles L, Jugl SM, Gunda P, Halliday A, Kirkham B. Cost effectiveness of secukinumab for the treatment of active psoriatic arthritis in the UK. Pharmacoeconomics. 2018;36(7):867–78.
Goeree R, Chiva-Razavi S, Gunda P, Graham CN, Miles L, Nikoglou E, Jugl SM, Gladman DD. Cost-effectiveness analysis of secukinumab for the treatment of active psoriatic arthritis: a Canadian perspective. J Med Econ. 2018;21(2):163–73.
Purmonen T, Puolakka K, Bhattacharyya D, Jain M, Martikainen J. Cost-effectiveness analysis of secukinumab versus other biologics and apremilast in the treatment of active psoriatic arthritis: a Finnish perspective. Cost Eff Resour Alloc. 2018;16(16):56.
McInnes I, Nash P, Ritchlin C, et al. THU0437 Secukinumab for the treatment of psoriatic arthritis: comparative effectiveness results versus licensed biologics and apremilast from a network meta-analysis. Ann Rheum Dis. 2016;75:348–9.
Spitzenverband Bund der Krankenkassen. GKV-Arzneimittel-Schnellinformation für Deutschland GKV-Arzneimittel-Schnellinformation für Deutschland. Berlin: GKV-Spitzenverband; 2018.
NICE Multiple Technology Appraisal—Certolizumab pegol and secukinumab for treating active psoriatic arthritis following inadequate response to disease modifying anti-rheumatic drugs, London; 2017. https://www.nice.org.uk/guidance/ta445/documents/html-content-3 (accessed 10 Aug 2019).
The Dental and Pharmaceutical Benefits Agency (TLV): Cosentyx ingår i högkostnadsskyddet med begränsning, Stockholm; 2016. https://www.tlv.se/beslut/beslut-lakemedel/begransad-subvention/arkiv/2016-05-31-cosentyx-ingar-i-hogkostnadsskyddet-med-begransning.html (accessed 10 Aug 2019).
McInnes IB, Nash P, Ritchlin C, Choy EH, Kanters S, Thom H, Gandhi K, Pricop L, Jugl SM. Secukinumab for psoriatic arthritis: comparative effectiveness versus licensed biologics/apremilast: a network meta-analysis. J Comp Eff Res. 2018;7(11):1107–23.
Mease PJ, Goffe BS, Metz J, VanderStoep A, Finck B, Burge DJ. Etanercept in the treatment of psoriatic arthritis and psoriasis: a randomised trial. Lancet. 2000;356(9227):385–90.
Shea BJ, Grimshaw JM, Wells GA, Boers M, Andersson N, Hamel C, Porter AC, Tugwell P, Moher D, Bouter LM. Development of AMSTAR: a measurement tool to assess the methodological quality of systematic reviews. BMC Med Res Methodol. 2007;15(7):10.
Rodgers M, Epstein D, Bojke L, Yang H, Craig D, Fonseca T, Myers L, Bruce I, Chalmers R, Bujkiewicz S, Lai M, Cooper N, Abrams K, Spiegelhalter D, Sutton A, Sculpher M, Woolacott N. Etanercept, infliximab and adalimumab for the treatment of psoriatic arthritis: a systematic review and economic evaluation. Health Technol Assess. 2011;15(10):i–xxi, 1–329.
Cawson MR, Mitchell SA, Knight C, Wildey H, Spurden D, Bird A, Orme ME. Systematic review, network meta-analysis and economic evaluation of biological therapy for the management of active psoriatic arthritis. BMC Musculoskelet Disord. 2014;20(15):26.
Institut für Qualität und Wirtschaftlichkeit im Gesundheitswesen. General methods. Version 5.0. Köln: IQWiG; 2017.
Lauer-Fischer GmbH. Lauer-Taxe Arzneimittelpreise. http://www.lauer-fischer.de (accessed 1 June 2018).
Rychlik R. Kosten der Unterversorgung mit Arzneimitteln in Deutschland. Februar: Gutachten für den Verband Forschender Arzneimittelhersteller e.V; 2008.
Berger K, Ehlken B, Kugland B, Augustin M. Cost-of-illness in patients with moderate and severe chronic psoriasis vulgaris in Germany. J Dtsch Dermatol Ges. 2005;3(7):511–8.
Kassenärztliche Bundesvereinigung. Einheitlicher. Bewertungsmaßstab. Stand: 1. Quartal 2018. Berlin; 2018.
Robert Koch Institut. Bericht zum Krebsgeschehen in Deutschland 2016, Berlin; 2016.
Statistisches Bundesamt (Destatis). Krankheitskosten 2015 in Deutschland nach Diagnosen, Einrichtungen, Alter und Geschlecht. Statistisches Bundesamt (Destatis); 2017.
Dolan P. Modeling valuations for EuroQol health states. Medical Care; 1997. p. 1095–1108.
Wong K, Gladman DD, Husted J, Long JA, Farewell VT. Mortality studies in psoriatic arthritis: results from a single outpatient clinic I Causes and risk of death. Arthritis Rheumatol. 1997;40(10):1868–72.
Abuabara K, Azfar RS, Shin DB, Neimann AL, Troxel AB, Gelfand JM. Cause-specific mortality in patients with severe psoriasis: a population-based cohort study in the UK. Br J Dermatol. 2010;163(3):586–92.
Statistisches Bundesamt (Destatis). Sterbetafel (Periodensterbetafel): Deutschland, Jahre, Geschlecht, Vollendetes Alter. Statistisches Bundesamt (Destatis); 2018.
Prosser LA, Neumann PJ, Sanders GD, Siegel JE. Reporting cost-effectiveness analyses. In: Neumann PJ, Sanders GD, Russell LB, Siegel JE, Ganiats TG, editors. Cost-effectiveness in health and medicine. Oxford: Oxford University Press; 2016.
Fagerli KM, Lie E, van der Heijde D, Heiberg MS, Kalstad S, Rødevand E, Mikkelsen K, Lexberg AS, Kvien TK. Switching between TNF inhibitors in psoriatic arthritis: data from the NOR-DMARD study. Ann Rheum Dis. 2013;72(11):1840–4.
Glintborg B, Ostergaard M, Krogh NS, Andersen MD, Tarp U, Loft AG, Lindegaard HM, Holland-Fischer M, Nordin H, Jensen DV, Olsen CH, Hetland ML. Clinical response, drug survival, and predictors thereof among 548 patients with psoriatic arthritis who switched tumor necrosis factor α inhibitor therapy: results from the Danish Nationwide DANBIO Registry. Arthritis Rheumatol. 2013;65(5):1213–23.
Harrold LR, Stolshek BS, Rebello S, Collier DH, Mutebi A, Wade SW, Malley W, Greenberg JD, Etzel CJ. Rebound in measures of disease activity and symptoms in Corrona Registry patients with psoriatic arthritis who discontinue tumor necrosis factor inhibitor therapy after achieving low disease activity. J Rheumatol. 2018;45(1):78–82.
Strand V, Elaine Husni M, Betts KA, Song Y, Singh R, Griffith J, Beppu M, Zhao J, Ganguli A. Network meta-analysis and cost per responder of targeted Immunomodulators in the treatment of active psoriatic arthritis. BMC Rheumatol. 2018;12(2):3.
Author information
Authors and Affiliations
Contributions
AG was involved in model conceptualization and adaptation and wrote the first draft of the manuscript. DO was involved in model conceptualization and commented on the draft version of the manuscript.
Corresponding author
Ethics declarations
Conflict of interest
DO is the founder and director of WifOR, which was contracted by Novartis Deutschland GmbH to undertake this work. AG is a consultant for WifOR.
Funding
This work was contracted by Novartis Deutschland GmbH.
Electronic supplementary material
Below is the link to the electronic supplementary material.
Rights and permissions
About this article
Cite this article
Gandjour, A., Ostwald, D.A. Cost Effectiveness of Secukinumab Versus Other Biologics and Apremilast in the Treatment of Active Psoriatic Arthritis in Germany. Appl Health Econ Health Policy 18, 109–125 (2020). https://doi.org/10.1007/s40258-019-00523-1
Published:
Issue Date:
DOI: https://doi.org/10.1007/s40258-019-00523-1