Microsimulation model of the cost-effectiveness of anifrolumab compared to belimumab in the United Arab Emirates

Abstract

Introduction

SLE imposes a significant morbidity and mortality as well as a substantial burden on the healthcare system. The model aimed to measure the cost-effectiveness of anifrolumab implementation against belimumab as an add-on-therapy to the standard of care (SoC) over a lifetime horizon for Emirati patients.

Methodology

A microsimulation model was used to assess the cost-effectiveness of anifrolumab against belimumab (IV/SC) as an add-on therapy to SoC in a hypothetical cohort of adult Emirati patients with systemic lupus erythematosus (SLE) over a lifetime horizon. The clinical data was captured from published clinical trials as; TULIP-1, TULIP-2, BLISS-52, BLISS-76 and BLISS‐SC. Health utility scores were constructed according to a linear regression model from the pooled data of the two TULIP Phase III trials of anifrolumab. Our model captures direct SLE-related medical costs from the Dubai Health Authority. Sensitivity analyses were conducted to assess model uncertainty.

Results

Using BICLA as a response criterion in the Johns Hopkins cohort, anifrolumab was found to be more effective than belimumab (IV/SC; the incremental discounted QALY of anifrolumab against belimumab was 0.42). The incremental cost-effectiveness ratio (ICER) of anifrolumab against belimumab IV and belimumab SC were AED 466,371 ($209,135) and AED 252,612 ($113,279), respectively, these ICERs are below the cost-effectiveness threshold in the United Arab Emirates (UAE) (three times gross domestic product capita; AED 592,278). In the Toronto lupus cohort, the ICER of anifrolumab against belimumab IV and belimumab SC were AED 491,403 ($220,360) and AED 276,642 ($124,055), respectively (anifrolumab was a cost-effective option vs. belimumab IV and belimumab SC).

Conclusion

The addition of anifrolumab to SoC is a cost-effective option versus belimumab for the treatment of adult patients with active, autoantibody-positive SLE, despite being allocated to SoC. Cost-effectiveness was demonstrated by a reduction in complications and organ damage, which reflected costs and outcomes.

Keywords:

• Cost-effectiveness
• SLE
• anifrolumab
• United Arab Emirates

JEL Classification Codes:

• C51
• C5
• C
• H51
• H5
• H

Introduction

Systemic lupus erythematosus (SLE) is a chronic, severe autoimmune disease with multisystemic involvement^1,2. SLE imposes a significant morbidity and mortality as well as a substantial burden on the healthcare system^1,2. Moreover, previous studies reported the costs incurred by SLE patients and pointed out its high resource utilization and economic burden^3–5.

The prevalence of SLE is higher in the United Arab Emirates (UAE) than in any other countries in the Gulf⁶. Dhanhani et al. pointed out that the SLE crude incidence rates per 100,000 person-years in 2009, 2010, 2011, and 2012 were 3.5, 1.1, 2.1, and 2.1, respectively⁶. The epidemiological burden of SLE in the UAE is increasing⁷. According to a recently published global epidemiological study (2023), the UAE has the highest prevalence of SLE⁷. The results of a global epidemiological study showed that the prevalence was 166.92 per 100,000 persons⁷, which was higher than the SLE prevalence in 2012 (103/100,000 population)⁶.

Due to the clinical course of the disease, the psychological effect of SLE on both patients and caregivers is an important aspect to consider⁸. Patients diagnosed with SLE experience pain, anxiety, or stress, and may need psychosocial support for their health complains⁸. Moreover, SLE affects caregivers of patients with SLE, both socially and financially⁹. For instance, employed caregivers reported missing up to 12.8% of their paid work time due to their responsibilities as caregivers, while 49.4% reported that caregiving influenced their ability to socialize with friends⁹. In addition, responsibility as caregivers made almost all caregivers (97.6%) experience stress and anxiety⁹.

SLE costs comprise SLE flare management, controlling SLE-related organ damage, and productivity loss due to the mental and physical impact of the disease on patients’ health. Therefore, SLE (especially with nephritis) is among the most costly chronic diseases¹⁰. Grabich et al. evaluated real-world data of SLE respondents over a period of 3-years (2016–2018) in comparison to non-SLE respondents in the United States (US)¹¹. In the SLE cohort, higher resource utilization was reported, such as the frequency of emergency room (ER) utilization and outpatient and inpatients visits¹¹. Consequently, annual all-cause healthcare costs in the SLE cohort (US$17,270) were approximately 2-folds higher, as compared to non-SLE (US$8350)¹¹. Furthermore, patients with SLE experience greater quality of life (QoL) deterioration due to depression, poor mental health, and physical limitation¹¹. Due to the disease, the number of missed working days among SLE patients was nine, while non-SLE patients had only five missed working days¹¹.

The course of SLE is characterized by the occurrence of exacerbations/flares¹², which are known to increase the risk of organ damage and the financial burden of SLE^13–15. For instance, they accounted for 77.5% of hospitalizations among SLE patients, with a mean length of hospital stay of 8.79 d¹⁶. Furthermore, severe SLE flares lead to an increase in annual SLE medical expenditures¹⁷. Compared to patients with no flares, the annual expenditure of patients with moderate and severe flares was 1.6 and 4.3 and times higher, respectively¹⁷. In addition, a high frequency of flares increases the risk of loss of work productivity through both absenteeism and presenteeism¹⁴. Compared with those without flares, the mean work-related activity impairment in patients with flares was 12%–32% higher¹⁴.

Management of SLE starts from targeting low level disease activity state (LLDAS) and minimizing dependency or chronic use of glucocorticoid (GC) to avoid their adverse events (AEs), up to decreasing numbers of flares and improving adherence to treatment^18–20. In many chronic conditions, treat-to-target (T2T) therapeutic approaches, in which treatment adjustments are made to reach a clinically defined goal, have been well established^21,22. For SLE management, LLDAS is accepted as a valid T2T endpoint²¹. According to experts, the current standard of care (SoC), which consists of antimalarials, GC, and immunosuppressive agents, fails to adequately control the disease.

Anifrolumab, a first-in-class type I interferon (IFNs) receptor antagonist, is used as an add-on therapy for adult patients with moderate-to-severe active autoantibody-positive SLE²³. Approximately 60-80% of adult SLE patients express high levels of type I IFN, which are associated with increased severity and activity of disease²⁴.

The efficacy and safety of anifrolumab has been demonstrated through two trials; the TULIP Phase III trials and the MUSE Phase II trial^25–27. In these trials, anifrolumab showed promising outcomes compared with a group allocated to the same standard therapy with a placebo. For instance, patients receiving anifrolumab experience a reduction in organ-specific disease activity (skin and joints) with a sustained reduction in oral GC dose^25,26. Furthermore, anifrolumab demonstrated earlier, more frequent, and more prolonged and sustained LLDAS than the placebo²¹. Therefore, anifrolumab can be used in the T2T approach in patients with moderate-to-severe SLE²¹.

Objective

This economic evaluation aimed to provide evidence regarding the cost-effectiveness of anifrolumab for the treatment of adult patients with active, autoantibody-positive SLE, despite being allocated to the SoC, omitting those with severe active lupus nephritis (LN) or severe active central nervous system (CNS) lupus. Anifrolumab was compared with belimumab (both intravenous (IV) and subcutaneous (SC) injection routes) as an add-on therapy to SoC in SLE patients over a lifetime horizon in the Emirati context.

Methodology

A micro-simulation model was established using Microsoft Excel 365 to simulate the heterogeneous and complicated disease course patterns of SLE. The model aimed to demonstrate the economic consequences of anifrolumab as an add-on therapy to SoCs (mainly unlicensed for SLE) for managing active, autoantibody-positive adult patients with SLE, omitting those with severe active LN or CNS lupus.

The baseline demographics data was obtained from pooled data of two TULIP Phase III trials^25,26. Subsequently, the eligibility criteria in our model were active, autoantibody-positive, adult patients with SLE with an average age of 41 years and ongoing treatment. According to our expert panel, the average adult weight in the UAE was 76 kg. The model excluded patients with systemic lupus erythematosus and severe active LN or CNS lupus.

The belimumab eligible cohort was considered to be adult, active, autoantibody-positive SLE patients, who showed a high degree of disease activity (e.g. low complement with positive anti-double-stranded deoxyribonucleic acid (anti-dsDNA) antibodies) and safety of estrogen in lupus national assessment (SELENA)– systemic lupus erythematosus disease activity index (SLEDAI) ≥10 in spite of standard therapy.

A base case analysis was conducted from the perspective of the Emirati payer. The demonstrated SLE-related direct medical costs in our models were SoC, treatment administration, monitoring, oral GC, organ damage, and AEs management costs, as measured by the United Arab Emirates Dirham (AED) 2023. Direct medical costs were simulated based on disease activity, flares, and organ damage at each simulated point. This economic evaluation is reported in accordance with the Consolidated Health Economic Evaluation Reporting Standards (CHEERS) statement²⁸. The inputs in our model were captured from a literature review of clinical trials conducted on available treatments for SLE and validated by a reputable Emirati Delphi panel. This Delphi panel composed of five experts; two rheumatologists affiliated to Gargash Hospital and Clemenceau Medical Center (CMC), HMS Mirdif Hospital in Dubai, two health economists and one payer from Dubai Health Authority. Those experts were selected from different backgrounds to represent the real clinical practice. We collected insights from all 5 experts through three rounds meetings by using the quasi-Delphi panel approach (Table 1). Delphi’s insights included the most standardized practices and treatment courses for these patients within Emirati settings. Furthermore, all assumptions were validated using a Delphi panel to decrease the uncertainty within the model.

Table 1. The Delphi panel questionnaire.

SLE-management questions:
Mention the most standardized treatment for management of adult patients with moderate to severe SLE upon diagnosis. Mention the most common immunosuppressant used (AZA, MTX, MMF), or a combination. Are there any concomitant medications? As Anifrolumab / Belimumab are a monoclonal antibody, is there any medication required before administration or specified monitoring? Does administration require hospitalization? (If yes, outpatient or a ward day)
Medication-related AEs questions:
Mention the management of the following AEs, if it requires hospitalization and its duration: Infusion-related reaction Herpes zoster Serious infection Serious psychiatric events
Consequences of organ damage questions:
1. Mention the management and follow-up among SLE patients who showed the following complications:
Affected organ	Management and follow-up during 1st Year	Management and follow-up during subsequent years
Ocular
Neuropsychiatric
Renal
Pulmonary
Cardiovascular
Peripheral vascular
Gastrointestinal
Musculoskeletal
Skin
Premature gonadal failure
Diabetes
Malignancy
Flares management questions according to severity:
1. Mention the required management of SLE patients with flares:
	Non-severe flares	Severe flares
Physician visit (frequency)
Laboratory test
Hospitalization
Medications

Abbreviations. AEs, adverse events; SLE, systematic lupus erythematosus; MMF, mycophenolate mofetil; AZA, azathioprine; MTX, methotrexate.

During the 1^st year of treatment (anifrolumab/belimumab), the extrapolated clinical outcomes (e.g. disease activity, flares, and organ damage) were treatment-dependent and extracted from the pooled data of two TULIP Phase III trials of anifrolumab^25,26 and three BLISS Phase III trials of belimumab^29–31. Regarding the 2nd year and beyond, a disease activity model was established utilizing the Johns Hopkins and Toronto Lupus cohorts’ natural history data to inform the efficacy parameters. Toronto Lupus cohort data were utilized in the reimbursement review of anifrolumab by the Canadian Agency for Drugs and Technologies in Health (CADTH)³², whereas our model simulated data from the Johns Hopkins cohort, as it is one of the most informative and largest datasets for patients with SLE. Watson et al. evaluated a sample of 1,354 patients from the Hopkins Lupus cohort and demonstrated an association between risk factors and long-term outcomes among patients with SLE³³. Figure 1 shows the structure of patients with SLE utilized in our model.

Figure 1. The model structure for SLE patients. Abbreviations. SLEDAI, Systemic Lupus Erythematosus Disease Activity Index; SLICC/ACR, Systemic Lupus International Collaborating Clinics/American College of Rheumatology.

Clinical parameters

At the end of the 1st year (52 weeks of treatment), patient responses were measured using the British Isles Lupus Assessment Group (BILAG) index-based combined lupus assessment (BICLA) response. The definitions of BICLA response are as follows: (a) improvement in baseline BILAG scores of all the body organs (improvement of BILAG A to BILAG B/C/D or of BILAG B to C/D); (b) no single new BILAG A or more than one new BILAG B scores; (c) no worsening of baseline SLEDAI total score; (d) no worsening in physician global estimate (≤10% worsening vs. baseline); and (e) no treatment failure/initiation of non-protocol treatment³⁴. The justification behind using the BICLA as a response criterion in our model was its association with clinical benefits in the assessment of SLE, patient-reported outcomes (PROs), and utilization of medical resources³⁵. Furthermore, the BICLA response has been utilized as an endpoint in other SLE-related clinical trials³⁶.

Our model included the treatment discontinuation rates in both arms. During 1^st year, the reasons for treatment discontinuation were natural attrition (e.g. non-adherence or AEs), not fulfilling the response criterion, or death, whereas the reason for discontinuation thereafter was mainly natural attrition or death. In the case of treatment discontinuation due to failure to fulfill the response criterion (non-responder), patients were shifted to receive SoC and reverted to the same clinical outcomes as the standard therapy. The model simulates whether the patient survives or dies annually. Notably, the age-and gender-specific death rate was extracted from UAE (2019) life tables sourced from the World Health Organization (WHO)³⁷.

We assumed that the long-term treatment effects of standard therapy remained constant throughout the duration of treatment administration (lifetime, zero waning). Belimumab IV vial waste was excluded from the model based on local practices. Due to the absence of a head-to-head clinical trial, an indirect treatment comparison (ITC) was performed to identify data on the comparators of anifrolumab³⁸. An ITC conducted by Bruce et al. assessed the efficacy of anifrolumab 300 mg versus belimumab 10 mg/kg in adults with moderate-to-severe SLE³⁸. The ITC was built using individual patient data from two TULIP Phase III trials and summary-level data obtained from BLISS clinical trials³⁸. Expert clinicians in ITC identified 14 Treatment effect modifiers (TEMs; patient characteristics at baseline): sex, race, age, SLEDAI score at baseline, organ domain score of one or more BILAG A items or two or more BILAG B items, antinuclear antibody titer ≥1:80, abnormal low C3, abnormal low C4, anti-dsDNA ≥30 IU/ml, antimalarial use, azathioprine use, methotrexate use, mycophenolate use, and OCS dose ≥7.5 mg at baseline. All these identified TEMs were ranked based on their relative degree of effect modification for each individual outcome and, based on the clinician’s opinion. Several TEMs were similar between TULIP and BLISS. Bruce et al. simulated the comparative efficacy between two treatments using simulated treatment comparison (STC) and matching-adjusted indirect comparisons (MAIC)³⁸. However, the ITC method applied in our model was built on the STC to compare the efficacies of anifrolumab and belimumab. This approach has been adopted previously in multiple National Institute for Care and Health Excellence (NICE) technology appraisals for estimating comparator treatment effects and is widely accepted by many health technology assessment (HTA) authorities.

For each treatment arm, costs and health-related quality of life (HRQoL) were assigned to each health state. As the model progressed for the assigned time horizon, the cost and HRQoL data were summed for each treatment arm, thus allowing for the calculation of the total costs and effectiveness of each comparator at model completion and the incremental cost-effectiveness ratio (ICER).

The cycle length was 1 year; a half-cycle correction was applied based on the International Society for Health Economics and Outcomes Research (ISPOR) modeling good research recommendations³⁹ to adjust the distribution of costs and effects that occurred throughout the cycle. The lifetime horizon of both scenarios was considered to be sufficiently long to capture the long-term clinical and economic impacts of SLE, which requires ongoing management until death. Costs and health-related outcomes were discounted at a rate of 3% in the base case (mean value), in accordance with another economic evaluation conducted in the UAE⁴⁰.

Treatment administration

For each treatment arm, our model assessed whether the patients survived or died annually. Within 1st year of treatment, discontinuation could be due to natural attrition (non-adherence or AEs), a lack of response to treatment, or death. Thereafter, discontinuation was primarily due to natural attrition or death. The risk of annual discontinuation of anifrolumab (7.8%) was extracted from pooled data from two TULIP Phase III trials^25,26. The same discontinuation rates for belimumab (IV and SC) were used in this study.

Within 1^st year of treatment, the extrapolated clinical outcomes (e.g. disease activity, flares, and organ damage) were treatment-specific and retrieved from the pooled data of two TULIP Phase III trials of anifrolumab^25,26 and three BLISS Phase III trials of belimumab^29–31. Thereafter, our study utilized the natural history model of the Johns Hopkins and Toronto Lupus cohorts to determine the efficacy parameters. In addition, our model included the most significant AEs from the available clinical trials (TULIP and BLISS), then validated by Delphi panel^{25,26,29–31}. Table 2 shows the input parameters of the model.

Table 2. The input parameters of the model.

	Base case	Low value	High value	References
(A) Patients’ characteristics:
Age (years)	41.25	33	49.5	^25^,^26
Female	92.3%	0.7384	1
Weight (kg)	76	60.8	91.2	Delphi panel
(B) Discontinuation:
Annual discontinuation with anifrolumab (%)	7.8%	0.0624	0.0936	^25^,^26
Annual discontinuation with belimumab (IV and SC; %)	7.8%	0.0624	0.0936	Assumption
(C) Aes
Anifrolumab
Herpes Zoster %	6.10%	0.0488	0.0732	^25^,^26
Serious infection %	5.25%	0.042	0.063
Infusion/Hypersensitivity reactions (mild) %	15.00%	0.12	0.18
Serious psychiatric events %	0.30%	0.0024	0.0036
Belimumab IV
Herpes Zoster %	0.60%	0.0048	0.0072	^29–31
Serious infection %	5.68%	0.04544	0.06816
Infusion/Hypersensitivity reactions (mild) %	17.00%	0.136	0.204
Infusion/Hypersensitivity reactions (severe) %	1.20%	0.0096	0.0144
Serious psychiatric events %	0.70%	0.0056	0.0084
Belimumab SC
Herpes Zoster %	0.60%	0.0048	0.0072
Serious infection %	5.68%	0.04544	0.06816
Infusion/Hypersensitivity reactions (severe) %	0.20%	0.0016	0.0024
Injection site reactions %	6.10%	0.0488	0.0732
Serious psychiatric events %	0.70%	0.0056	0.0084
(D) Discount rate:
Discount rate %	3%	0.028	0.042	^40

Abbreviations. SLE, Systemic lupus erythematosus; AEs, Adverse events; IV, intravenous; SC, subcutaneous.

Treatment efficacy

Treatment efficacy was captured from the TULIP-1 and TULIP-2 trials, which were conducted to assess the safety and efficacy of anifrolumab versus placebo as an add-on-therapy to SoC among moderate to severe active, autoantibody-positive, SLE patients while omitting those with severe active LN or CNS/neuropsychiatric lupus^25,26. The total number of randomized eligible SLE patients at TULIP-1 and TULIP-2 were 457 and 365, respectively^25,26. Regarding TULIP-1, the primary efficacy outcome of SLE responder index-4 (SRI-4) response at week 52 was not met (36% for anifrolumab 300 mg, while 40% for placebo (difference −4.2 (95% confidence interval (CI = −14.2 to 5.8)²⁵. However, anifrolumab led to significant improvements in multiple secondary endpoints such as sustained reduction of GC dose, BICLA response, and organ-specific measures (skin and joints)²⁵.

In contrast to TULIP-1, TULIP-2 utilized a primary outcome that was a secondary endpoint in TULIP-1, thus favoring anifrolumab 300 mg and showing its superiority over the placebo²⁶. The BICLA response at week 52, the primary endpoint of TULIP-2, was defined as a reduction in disease activity with no worsening of another organ system or disease activity, no treatment discontinuation, or utilization of restricted medications beyond the protocol-allowed limit²⁶. The percentage of patients with BICLA response at week 52 among anifrolumab 300 mg and placebo was 47.8% and 31.5%, respectively (difference (percentage points) = 16.3 (95% CI = 6.3–26.3))²⁶. Furthermore, 300 mg of anifrolumab reduced the severity of skin disease and GC dose over a time horizon of 52 weeks²⁶.

BLISS-52, BLISS-76 and BLISS‐SC are phase III, randomized, placebo‐controlled clinical trials aimed to evaluate the safety and efficacy of different injection routes of belimumab against placebo as an add-on-therapy to standard therapy^29–31. In the BLISS-52 trial, 867 adult patients with active SLE were randomized and showed clinical improvement with belimumab (e.g. higher SRI response rates or SELENA-SLEDAI score reduction)²⁹. BLISS-76 confirmed clinically significant improvements with belimumab (e.g. reduction of disease activity, reduction in severe flares, and improvement of SRI response rate) after belimumab administration in 826 randomized eligible patients³⁰. In BLISS‐SC, the safety and efficacy of belimumab SC among 839 randomized SLE patients demonstrated tolerability and reduction in disease activity³¹.

Bruce et al. established an ITC to evaluate the comparative efficacy of anifrolumab versus belimumab in adults with moderate-to-severe SLE receiving ongoing standard therapy³⁸. The study pointed out that patients treated with anifrolumab were more likely to demonstrate an improvement in disease activity than those treated with belimumab³⁸. For instance, anifrolumab-treated patients had more than double the chance to achieve a reduction of ≥ 4 points in SLEDAI score and SRI-4 response at 52 weeks, in comparison with patients treated with belimumab³⁸.

Outcomes

Our model simulates the direct effect of treatment on HRQoL. Subsequently, the parameters for treatment response/non-response in the HRQoL equations were used. Therefore, the HRQoL equations were treatment specific for each chosen response assessment. Our study developed a linear regression model for health utility scores built on pooled data from two TULIP Phase III trials of anifrolumab using the EuroQol five-dimension five-level (EQ5D-5L) scale, with utilities assigned annually according to treatment response, number of severe and non-severe flares, mean SLEDAI score, patient age, and body mass index (BMI)³². Furthermore, an equation was applied such that health-related utility scores decreased in cases in which patients developed or continued to experience organ damage³². The utilities associated with organ damage are presented in Table 3.

Table 3. The utilities used in the model.

	Base case	Low value	High value
(A) utility in 1st year:
Ocular	0.974	0.7792	1
Neuropsychiatric	0.713	0.5704	0.8556
Renal	0.972	0.7776	1
Pulmonary	0.713	0.5704	0.8556
Cardiovascular	0.779	0.6232	0.9348
Peripheral vascular	0.863	0.6904	1
Gastrointestinal	0.786	0.6288	0.9432
Musculoskeletal	0.655	0.524	0.786
Skin	0.943	0.7544	1
Premature gonadal failure	1	0.8	1
Diabetes	0.91	0.728	1
Malignancy	0.837	0.6696	1
(B) utility in 2nd year:
Ocular	0.992	0.7936	1
Neuropsychiatric	0.772	0.6176	0.9264
Renal	0.955	0.764	1
Pulmonary	0.719	0.5752	0.8628
Cardiovascular	0.806	0.6448	0.9672
Peripheral vascular	0.873	0.6984	1
Gastrointestinal	0.906	0.7248	1
Musculoskeletal	0.729	0.5832	0.8748
Skin	0.943	0.7544	1
Premature gonadal failure	1	0.8	1
Diabetes	0.91	0.728	1
Malignancy	0.837	0.6696	1

Costs

From the perspective of Emirati payers, the costs measured in our model were SLE-related direct medical costs. The costs include SoC, treatment administration, monitoring, oral GC, organ damage, flares, and AEs management costs. Organ damage costs were measured for each organ system, while flare costs were simulated based on the type of flare (non-severe/mild-moderate flares and severe flares). AEs management costs were calculated based on the average resources used and the length of hospital stay (if needed) to treat a single event.

The total direct medical costs were measured by multiplying resource use by the unit costs of each item. Resource use during routine follow-ups was informed by the Delphi panel and was based on local clinical practice in Emirati healthcare settings. All unit costs were measured in the AED and obtained from the Dubai Health Authority in the financial year 2023. Local currency conversions to the international dollar were performed using the purchasing power parity rate, captured from World Bank. Table 4 lists all the unit costs utilized in our model.

Table 4. The unit costs utilized in our model.

Cost parameters	Base case	Low value	High value	References
(1) SoC cost
Hydroxychloroquine 200 mg	1.55	1.24	1.86	Dubai Health Authority
Azathioprine 50 mg	0.56	0.448	0.672
Methotrexate 10 mg	4.87	3.896	5.844
MMF 500 mg	15.67	12.536	18.804
Prednisone 5 mg	0.44	0.352	0.528
(2) Biologics cost
Anifrolumab 300 mg/1 vial	5,128.5	4,102.8	6,154.2
Belimumab IV 120 mg/Vial	510	408	612
Belimumab IV 400 mg/Vial	1,649.5	1,319.6	1,979.4
Belimumab SC 200 mg/1 ml	1,227.38	981.904	1,472.856
ER visit	134.25	107.4	161.1
A-6-hour hospital admission (SC admission)	471.78	377.424	566.136
Chlorphenamine 4 mg tab.	0.325	0.26	0.39
Blood smear	85.79	68.632	102.948
(3) AEs
Acyclovir topical	73.5	58.8	88.2
Acyclovir 800 mg	7.15	5.72	8.58
Ward day	493	394.4	591.6
Azithromycin 250 mg	7.75	6.2	9.3
Azithromycin 500 mg	17.66	14.128	21.192
Ciprofloxacin 500 mg	5.5	4.4	6.6
Chlorphenamine 50 mg IV	5	4	6
Dexamethasone 4 mg IM	1.7	1.36	2.04
Dexamethasone 4 mg tab.	1.29	1.032	1.548
Epinephrine 1 mg/mL solution	6.06	4.848	7.272
Physician visit	575	460	690
Citalopram 20 mg	3.89	3.112	4.668
(4) Flares
Methylprednisolone IV (500 mg)	61	48.8	73.2
CBC	100	80	120
Anti-dsDNA	112.5	90	135
C4 and C3	92.5	74	111
CRP	170.88	136.704	205.056
ESR	64.62	51.696	77.544
KFT	951.59	761.272	1141.908
LFT	1012.6	810.08	1215.12
(5) Long-term costs associated with organ damage
ICU day	334.21	267.368	401.052
Renal biopsy	8,040.98	6,432.784	9,649.176
Ramipril 2.5 mg	1.4	1.12	1.68
Calcium/vitamin D supplementation	1.38	1.104	1.656
Bronchoscopy	3,089.05	2,471.24	3,706.86
BAL	849.39	679.512	1,019.268
CT scan of the chest with contrast	12,000	9,600	14,400
Nintedanib 150mg	260.325	208.26	312.39
Isosorbide mononitrate 20mg	1	0.8	1.2
ECHO	784.38	627.504	941.256
ECG	980.09	784.072	1,176.108
Angiography	3,463.34	2,770.672	4,156.008
Capillaroscopy	160.53	128.424	192.636
Nifedipine 30 mg	0.78	0.624	0.936
Aspirin 81 mg	0.6	0.48	0.72
Sildenafil 50 mg	9.63	7.704	11.556
Sildenafil 100 mg	15.50	12.4	18.6
Triamcinolone 10 mg	10.73	8.584	12.876
Triamcinolone 40 mg	11	8.8	13.2
Estradiol 10 mg	61.1	48.88	73.32
Insulin	1,776.75	1,421.4	2,132.1
HbA1C	155.56	124.448	186.672
Alendronate 5mg	49.875	39.9	59.85
Bone density assessment	560.12	448.096	672.144

Abbreviations. SoC, Standard of care; MMF, Mycophenolate mofetil; ECHO, echocardiogram; ICU, intensive care unit; KFT, kidney function test; CBC, complete blood test; C3, Complement component 3; C4, Complement component; ESR, erythrocyte sedimentation rate; CRP, c-reactive protein; Anti-dsDNA, Anti-double stranded DNA; HbA1C, Hemoglobin A1C; BAL, bronchoalveolar lavage; CT, computed tomography; ECG, electrocardiogram; LFT, liver function test; IV, intravenous; SC, subcutaneous; ER, emergency room; AEs, adverse events; IM, intramuscular.

Treatment regimens

SoC includes the following drugs in combination: immunosuppressants, antimalarials (hydroxychloroquine), and steroids (prednisone). According to the Delphi panel, patients received only one immunosuppressant: azathioprine, methotrexate, or mycophenolate mofetil (MMF). Table S1 lists the interventions and their doses.

For anifrolumab or belimumab IV, patients were hospitalized as outpatients for treatment. However, there has been no specific monitoring of anifrolumab. For the 1st time for belimumab SC administration, patients were required to be admitted (a 6-h hospital admission). According to our Delphi panel, patients allocated to belimumab therapy should be assessed monthly using blood smears. In addition, IV belimumab requires prophylactic premedication⁴¹ with oral antihistamines⁴². However, there are insufficient data to determine whether pre-medications can help prevent infusion reactions⁴³.

According to the Delphi panel, the most significant treatment-related AEs in patients with SLE were herpes zoster, serious infections (pneumonia), infusion/hypersensitivity reactions, injection site reactions, and serious psychiatric events (depression).

The flares were categorized as non-severe/mild-moderate or severe. According to the severity of the flares, patients were managed with prednisone, which was then tapered to the lowest allowed daily dose (5 mg daily). The duration of the initial prednisone dose was 2 weeks and was tapered once every 2 weeks (5 mg decrease). As 36.4% of the anifrolumab-treated patients were GC-free (0 mg/day)⁴⁴, we assumed that prednisone would be discontinued once the patients reached the lowest daily dose.

Sensitivity analyses

Deterministic sensitivity analysis (DSA) and probabilistic sensitivity analysis (PSA) were performed to ensure the robustness of the results. The various parameters varied within the plausible case values. The parameters tested were clinical data, probabilities, utilities, and unit cost data for all treatments.

Results

Using BICLA as a response criterion in the Johns Hopkins cohort, anifrolumab was found to be more effective than belimumab (IV/SC; the incremental discounted QALY of anifrolumab against belimumab was 0.42). The ICER of anifrolumab against belimumab IV and belimumab SC were AED 466,371 and AED 252,612, respectively, these ICERs are below the cost-effectiveness threshold in the UAE (three times gross domestic product (GDP)/capita; AED 592,278)⁴⁵. Among the Toronto lupus cohort, the ICER of anifrolumab against belimumab IV and belimumab SC was AED 491,403 and AED 276,642, respectively (cost-effective option; Table 5).

Table 5. The model results (BICLA response criterion).

Treatment Strategy	Total Costs (AED)	Difference in Costs (AED)	Total QALYs	Difference in QALYs	Total LYs	Difference in LYs	ICER of QALYs	Interpretation
(A) Johns Hopkins Lupus Cohort
Anifrolumab	1,778,713 ($797,629)	106,716 ($47,855)	9.32	0.42	16.65	0.54	252,612 ($113,279)	Anifrolumab generates a better QALYs/LYs when compared to belimumab in addition of being a cost-effective option
Belimumab SC	1,671,998 ($749,775)		8.90		16.12
Belimumab IV	1,581,695 ($709,280)	197,018 ($88,349)	8.90	0.42	16.12	0.54	466,371 ($209,135)
(B) Toronto Lupus Cohort
Anifrolumab	841,232 ($377,234)	102,790 ($46,094)	7.82	0.37	12.99	0.45	276,642 ($124,055)	Anifrolumab generates a better QALY/LY when compared to belimumab in addition of being a cost-effective option
Belimumab SC	738,442 ($331,140)		7.45		12.54
Belimumab IV	658,645 ($295,357)	182,587 ($81,878)	7.45	0.37	12.54	0.45	491,403 ($220,360)

Abbreviations. AED, United Arab Emirates dirham; QALY, quality adjusted life year; LY, life year; ICER, incremental cost-effectiveness ratio, the local currency was converted to international dollar using the 2022 purchasing power parity conversion factor (2.23), sourced from the World Bank.

Scenario analyses

Using the SLEDAI as a response criterion in the Johns Hopkin cohort, we found that anifrolumab was more effective than belimumab (IV/SC; the incremental discounted QALY of anifrolumab against belimumab was 0.42). The ICER of anifrolumab against belimumab IV and belimumab SC were AED 492,573 and AED 249,484, respectively. In the Toronto lupus cohort, the ICER of anifrolumab against belimumab IV and SC were AED 495,161 and AED 260,087, respectively. Similarly, both scenarios demonstrated that anifrolumab is a cost-effective add-on therapy for SoCs. Table S2 shows the results of using SLEDAI as the response criterion.

Sensitivity analyses

DSA showed the most sensitive parameter that affected the results (Figure 2). A cost-effectiveness plane was used to graphically demonstrate the variations in the incremental costs and QALYs of anifrolumab (Figure 3). Most difference pairs were found in the northeast quadrant of the cost-effectiveness plane, which indicates that anifrolumab use in patients with SLE is more effective. A cost-effectiveness acceptability curve was generated to illustrate the probability that each regimen was cost-effective at varying WTP levels of willingness to pay (Figure 4).

Figure 2. One-way sensitivity analysis of anifrolumab vs belimumab IV. The purple red color indicates low value of the output while the blue color indicates the high value of the output. Abbreviations. QALY, quality adjusted life year; SLE, Systemic lupus erythematosus; SLEDAI, Systemic Lupus Erythematosus Disease Activity Index.

Figure 3. A cost-effectiveness plane of anifrolumab vs belimumab IV. Abbreviations. QALY, quality adjusted life year; IV, intravenous.

Figure 4. A cost-effectiveness acceptability curve of anifrolumab vs belimumab IV. Abbreviations. QALY, quality adjusted life year; IV, intravenous.

Discussion

Anifrolumab was found to be cost-effective in comparison with belimumab (IV/SC) for the treatment of adult patients with active, autoantibody-positive SLE, despite being allocated to SoC. The ICER of anifrolumab against belimumab IV and belimumab SC were AED 466,371 and AED 252,612, respectively, these ICERs are below the cost-effectiveness threshold in the UAE.

SLE is a non-organ-specific autoimmune disease. Persistent disease activity over time, with the chronic use of GC and immunosuppressants, increases the risk of morbidity and mortality. In addition, patients with SLE experience daily challenges and several unmet needs, such as disease symptoms, complications of developing organ damage, or fear of treatment-related AEs, all of which eventually affect their QoL and productivity. Furthermore, multiple studies have highlighted the importance of reducing the occurrence of flares and slowing the progression of organ damage, as they cause a substantial cost to patients and society^15,46. Anifrolumab is the 2nd biologic approved for the management of patients with moderate-to-severe SLE⁴⁷. The efficacy of anifrolumab as an add-on-therapy has been evaluated in different clinical trials^25–27.

Bruce et al. conducted an ITC to evaluate the comparative efficacy of anifrolumab 300 mg against belimumab 10 mg/kg in adult patients with SLE with ongoing standard therapy³⁸. The study pointed out that the anifrolumab-treated cohort had more than double the chance to achieve a reduction of ≥ 4 points in SLEDAI score and SRI-4 response at 52 weeks, in comparison with the belimumab-treated cohort³⁸. Therefore, the study concluded that patients with moderate-to-severe SLE treated with anifrolumab had a higher chance of achieving an improvement in disease activity than those treated with belimumab³⁸.

Steroid tapering is a key goal for SLE management. Long-term administration of GC has a significant impact on patients with SLE. For instance, patients are prone to infection, accrual organ damage, and treatment-related AEs, such as osteoporosis. In addition, a 1 mg daily increase in prednisone dose increases the risk of organ damage by 2.8%⁴⁸. A post hoc analysis of the two TULIP Phase III trials demonstrated the favorable benefit-risk profile of anifrolumab, as it was linked with an early, maintained improvement in overall disease activity and GC tapering⁴⁹. From week 8, a greater percentage of patients in the anifrolumab group achieved a BICLA response than did those in the placebo group⁴⁹. Furthermore, this response was maintained for up to week 52⁴⁹. Subsequently, the early and greater improvement in overall disease activity with anifrolumab was interpreted as a higher reduction in GC dose with anifrolumab at week 20 and at all time points thereafter, in comparison with placebo⁴⁹. Furthermore, a reduction in prednisone dose by 1 mg/d leads to a 3% reduction in organ damage risk⁴⁸.

The T2T approaches have positively affected the management of many diseases^21,50. The T2T approach is recommended by an international task force for the management of SLE⁵¹. Furthermore, LLDAS is the T2T goal in patients with SLE²¹. In addition, LLDAS attainment is associated with favorable outcomes²¹. For instance, different studies pointed out that SLE patients with LLDAS attainment experienced a reduction of accrual organ damage, QoL improvement, and reduction in mortality^20,52,53. A post hoc analysis of pooled data from the two TULIP Phase III trials showed that the anifrolumab group had frequent, earlier, and more prolonged and sustained LLDAS than the placebo²¹. At week 52, the percentage of patients with LLDAS (responders) in the anifrolumab and placebo group was 30.0% and 19.6%, respectively²¹. Furthermore, patients with LLDAS attainment (in both the anifrolumab and placebo arms) experienced a three-fold improvement in patient global assessment scores compared to baseline scores²¹. Owing to early and frequent LLDAS attainment and remission, anifrolumab can be used as an optimal T2T approach for patients with moderate-to-severe SLE to achieve long-term remission²¹.

Yeo et al. assessed the effect of LLDAS attainment on cost reduction and resource utilization⁵⁴. The study highlighted that the LLDAS is a clinical outcome linked to a decrease in annual direct medical costs⁵⁴. After a median observation period of 2.1 years, SLE patients who spent ≥50% of the observation time in LLDAS had a decrement in annual direct cost of 25.9% (annual decrease = US$1,604)⁵⁴. As anifrolumab treatment was associated with more sustained and prolonged LLDAS²¹, it can be proved that anifrolumab can reduce the total cost of SLE.

CADTH, a reputable HTA body in Canada, recommends the reimbursement of anifrolumab as an added therapy to SoC for the treatment of adult patients with moderate to severe active autoantibody-positive SLE⁵⁵. The justification for reimbursement is that anifrolumab can fulfill some of the unmet needs of patients with SLE, as it reduces both the disease activity and utilization of GC⁵⁵. Due to the need for long-term disease management, patients seek treatments that can control their overall disease activity and improve their QoL. Anifrolumab addresses this unmet need⁵⁶.

Our model included several strengths that needs to be mentioned. First, all the included parameters were captured from clinical trials and validated using the Delphi panel. Furthermore, disease activity for 2nd year of treatment and beyond was captured from both the Johns Hopkins and Toronto Lupus cohorts’ natural history data. In addition, the general mortality data used in our model are local data from the Emirati population. Another strength, a sensitivity analyses were conducted to report any uncertainty in results. However, our model has some limitations that need to be addressed. Country-specific evidence and parameters are lacking. Nevertheless, all inputs were validated using the Delphi panel. The lack of head-to-head comparison between anifrolumab and belimumab was a great challenge in our study, however ITC was built using individual patient data from two TULIP Phase III trials and summary-level data obtained from BLISS clinical trials to report evident pooled data. As a possible confounder, BLISS 52 uses SELENA-SLEDAI while TULIP uses BICLA as their primary endpoints. However, ITC identified these assessments as TMEs from 14 TMEs and ranked them based on their relative degree of effect modification for each individual outcome and, based on the experts opinion and a literature review of the predictors of SLE treatment outcomes³⁸. Furthermore, an extensive sensitivity analyses were conducted to explore the impact of adjusting for these TEMs. Future research is recommended to collect real world data about resource utilization in SLE treatment for both treatment arms in UAE.

Conclusion

The addition of anifrolumab to SoC is a cost-effective option versus belimumab for the treatment of adult patients with active, autoantibody-positive SLE, despite being allocated to SoC. Cost-effectiveness was demonstrated by a reduction in complications and organ damage, which reflected costs and outcomes.

Transparency

Author contributions

GE: performing the analysis, and writing manuscript. AW and WS: interpretation of data, and revision of manuscript. MF, AW and WS: retrieving data, local clinical practice validation. All authors contributed to the article and approved the submitted version.

Reviewer disclosures

The peer reviewers on this manuscript have received an honorarium from JME for their review work. One of the reviewers has received a speaker’s honorarium from both GSK and AZ, the other reviewers have no other relevant financial relationships or otherwise to disclose.

Supplement statement

This article is part of a supplement sponsored by AstraZeneca. All articles within this supplement have been rigorously peer reviewed by experts in the field, as per Journal of Medical Economic’s peer review policy. Any conflicts of interest are stated in the “Declaration of financial/other relationships” section.

Acknowledgements

The authors gratefully acknowledge Mariam Elattar for the writing assistance utilized in the production of this manuscript and Neveen Kandil for her efforts in organizing the Delphi panel meetings.

Declaration of financial/other relationships

GE was employed by HTA Office, LLC. GE is a speaker for Janssen, Merck, Novartis, AstraZeneca, Roche, Eva pharma and Pfizer. The authors have no other financial relationships to disclose.

The experts did not receive any compensation for their participation in the Delphi panel.

Correction Statement

This article has been corrected with minor changes. These changes do not impact the academic content of the article.

Additional information

Funding

This study was funded by AstraZeneca, who was not involved in the study design, analysis, interpretation of results or manuscript writing. The funding received was used to pay for the submission and open access fees.