Has upadacitinib (UPA), a Janus kinase (JAK) inhibitor, been compared with other treatment options for rheumatoid arthritis (RA)?

With the growing number of treatment options in RA, understanding the differences in efficacy and safety of available RA therapies is becoming increasingly challenging, yet important. How do we evaluate whether one treatment is superior to another? What can head-to-head trials tell us regarding the comparison of treatment options?

UPA is an oral JAK inhibitor indicated for the treatment of adults with moderately to severely active RA who have had an inadequate response or intolerance to methotrexate (MTX). Use of UPA in combination with other JAK inhibitors, biologic disease-modifying antirheumatic drugs (bDMARDs), or with potent immunosuppressants such as azathioprine and cyclosporine is not recommended.1 As part of the Phase 3 clinical program for UPA, encompassing six trials, UPA was shown to be superior (for select ranked secondary endpoints) over 2 biologic therapies in separate head-to-head trials.2,3

Let's take a closer look at the endpoints in SELECT-COMPARE and SELECT-CHOICE that were evaluated for noninferiority and superiority of UPA against both adalimumab (ADA) and abatacept (ABA).

1. RINVOQ™ [package insert]. North Chicago, IL: AbbVie Inc; 2020. 2. Fleischmann R, et al. Arthritis Rheumatol. 2019;71(11):1788-1800. 3. Rubbert-Roth A, et al. N Engl J Med. 2020;383(16):1511-1521.

How are treatment comparisons made in head-to-head trials?

Head-to-head trials allow for the direct comparison of the efficacy of one treatment vs another. The trial can test for superiority, where the aim is to demonstrate that one treatment has a significantly greater effect than the other. It can also test for noninferiority, where the goal is to show that the investigational treatment is not clinically worse in a specific endpoint than the active comparator by a prespecified amount called the noninferiority margin.1,2

Testing for both noninferiority and superiority in a single head-to-head trial is possible if the trial has been properly designed and adequately powered. If an endpoint is to be assessed for both noninferiority and superiority, noninferiority must first be achieved before superiority can be assessed.

1. Fleischmann R, et al. Semin Arthritis Rheum. 2016;46(3):279-285. 2. Non-Inferiority Clinical Trials to Establish Effectiveness, Guidance for Industry. US Food and Drug Administration Web site. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/non-inferiority-clinical-trials. Updated November 2016. Accessed December 11, 2020.

SELECT-COMPARE evaluated UPA + MTX vs ADA + MTX for noninferiority in the proportion of patients achieving ACR50 at week 12. The noninferiority margin was set at 10%. The observed difference was 16.1%, with a 95% CI of 9.9% to 22.3% in favor of UPA + MTX. Since the 95% CI does not overlap with the noninferiority margin, we can conclude that UPA + MTX is noninferior to ADA + MTX in ACR50 at week 12.1,2

In SELECT-CHOICE, UPA + csDMARDs was evaluated for noninferiority to ABA + csDMARDs in the primary endpoint of the change from baseline in DAS28-CRP at week 12. The noninferiority margin was set at 0.6. The observed difference between the 2 treatment arms was -0.52, with a 95% CI of -0.69 to -0.35 in favor of UPA + csDMARDs. Since the 95% CI does not overlap with the margin of noninferiority, we can conclude that UPA + csDMARDs is noninferior to ABA + csDMARDs in the change from baseline in DAS28-CRP at week 12.1,3,4

1. Non-Inferiority Clinical Trials to Establish Effectiveness, Guidance for Industry. US Food and Drug Administration Web site. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/non-inferiority-clinical-trials. Updated November 2016. Accessed December 11, 2020. 2. Fleischmann R, et al. Arthritis Rheumatol. 2019;71(11):1788-1800. 3. Rubbert-Roth A, et al. N Engl J Med. 2020;383(16):1511-1521. 4. Supplement to: Rubbert-Roth A, et al. N Engl J Med. 2020;383(16):1511-1521.

UPA superiority to active comparators was evaluated for multiple endpoints across SELECT-COMPARE and SELECT-CHOICE

Clinical trials often include multiple endpoints. However, as the number of endpoints analyzed in a single trial increases, the likelihood of obtaining a false positive result also increases. This is an issue referred to as multiplicity. Failure to appropriately adjust and control for this can potentially lead to a false positive statistical conclusion.1

The issue of multiplicity can be controlled by prespecifying a fixed sequence (ranking) in which the endpoints will be analyzed.1 This method was employed in the SELECT-COMPARE and SELECT-CHOICE studies as shown below.2,3

1. Multiple Endpoints in Clinical Trials, Draft Guidance for Industry. US Food and Drug Administration Web site. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/multiple-endpoints-clinical-trials-guidance-industry. Updated January 2017. Accessed December 11, 2020. 2. Supplement to: Fleischmann R, et al. Arthritis Rheumatol. 2019;71(11):1788-1800. 3. Rubbert-Roth A, et al. N Engl J Med. 2020;383(16):1511-1521.

As previously discussed, UPA + MTX was noninferior to ADA + MTX in ACR50 at week 12 in SELECT-COMPARE. Evaluating for superiority in this same endpoint, we see that the difference between the treatment arms was 16.1% in favor of UPA + MTX, with a 95% CI (9.9, 22.3) that did not overlap with the line of no difference. Thus, we can further conclude that UPA + MTX is not only noninferior but also superior to ADA + MTX in ACR50 at week 12.1,2

Two other ranked endpoints compared UPA + MTX vs ADA + MTX in SELECT-COMPARE. These were pain reduction (treatment difference -6.5, 95% CI [-9.7, -3.3]) and change in HAQ-DI (treatment difference -0.11, 95% CI [-0.18,-0.03]) at week 12. In both endpoints, we see that UPA + MTX is superior to ADA + MTX.2

In SELECT-CHOICE, UPA + csDMARDs is noninferior to ABA + csDMARDs in the change from baseline in DAS28-CRP at week 12. Moreover, the difference between the treatment arms demonstrates that UPA + csDMARDs is superior to ABA + csDMARDs for this endpoint (difference -0.52, 95% CI [-0.69, -0.35]).3

Looking at another ranked endpoint in SELECT-CHOICE, we see that UPA + csDMARDs is superior to ABA + csDMARDs in the proportion of patients achieving DAS28-CRP <2.6 at week 12. The difference between the treatment arms was 16.8%, with a 95% CI (10.4, 23.2) that does not overlap with the line of no difference.3

1. Non-Inferiority Clinical Trials to Establish Effectiveness, Guidance for Industry. US Food and Drug Administration Web site. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/non-inferiority-clinical-trials. Updated November 2016. Accessed December 11, 2020. 2. Fleischmann R, et al. Arthritis Rheumatol. 2019;71(11):1788-1800. 3. Rubbert-Roth A, et al. N Engl J Med. 2020;383(16):1511-1521.

UPA safety profile across the SELECT Phase 3 program

Head-to-head trials are helpful in assessing clinical efficacy; however, we must also examine the safety of a treatment when evaluating treatment options. Below is a brief summary of safety data from the SELECT Phase 3 program of UPA with up to ≈3.5 years of exposure and over 4500 patient-years.1-3

1. Cohen SB, et al. Poster THU0197. European League Against Rheumatism 2020 E-Congress; June 3-6, 2020; Virtual. 2. Data on File. AbbVie. ABVRRTI70151. 3. Data on File. AbbVie. ABVRRTI70194.

During the PBO-controlled portion of SELECT-COMPARE (26 weeks), serious infections were reported in similar proportions of patients receiving UPA + MTX (1.8%) and patients receiving ADA + MTX (1.5%), both of which were higher than the rate of serious infections in patients receiving PBO + MTX (0.8%). There were 8 cases of herpes zoster (5 in UPA + MTX, 1 in ADA + MTX, and 2 in PBO + MTX) and 1 case of Varicella zoster in PBO + MTX; none were serious, and most affected a single dermatome. No adjudicated MACEs were reported in UPA + MTX, while 2 adjudicated MACEs were reported in ADA + MTX, and 3 in PBO + MTX. Six adjudicated VTEs were reported (1 DVT and 1 PE in UPA + MTX, 3 PEs in ADA + MTX, and 1 PE in the placebo group). All patients that experienced VTEs had preexisting risk factors.1

Throughout the 24-week trial period of SELECT-CHOICE, serious infections occurred in 3 patients (1.0%) receiving UPA + csDMARDs and 1 patient (0.3%) receiving ABA + csDMARDs. Opportunistic infections occurred in 4 patients (1.3%) receiving UPA + csDMARDs (3 patients with oral candidiasis and 1 patient with esophageal candidiasis) and 1 patient (0.3%) receiving ABA + csDMARDs (oral candidiasis). Herpes zoster infection occurred in 8 patients (4 in each treatment group); no cases were serious, and most affected a single dermatome. Hepatic disorder occurred in 23 patients (7.6%) receiving UPA + csDMARDs and 5 patients (1.6%) receiving ABA + csDMARDs. There were no Hy's law cases identified in either treatment group. One MACE was reported in UPA + csDMARDs. Two VTEs were reported in UPA + csDMARDs (1 DVT and 1 PE; both had preexisting risk factors).2,3

Note that the trials in the SELECT program, including SELECT-COMPARE and SELECT-CHOICE, were not powered to compare the safety outcomes between treatments. Refer to the full Prescribing Information to review the complete safety data.

1. Fleischmann R, et al. Arthritis Rheumatol. 2019;71(11):1788-1800. 2. Rubbert-Roth A, et al. N Engl J Med. 2020;383(16):1511-1521. 3. Data on File. AbbVie. ABVRRTI70702.

Find My Medical Science Liaison

To learn more, please enter your zip code below to find your local Medical Science Liaison.

*Required

{{result.firstName}} {{result.lastName}}
No results found.

Glossary

False positive

concluding that a treatment has a beneficial effect when it does not1

Head-to-head trials

studies that are appropriately designed and powered to allow formal direct comparison of different therapies2

Multiplicity

increased risk of obtaining a false positive result when testing multiple endpoints or making multiple comparisons1

Ranking

prespecified sequence for testing multiple endpoints. This fixed order of analysis controls the issue of multiplicity1

Noninferiority comparison

the goal of a noninferiority comparison is to show that the investigational treatment is not clinically worse than the active comparator. A treatment is considered to be noninferior to a comparator if the difference in effect between the 2 is in favor of the treatment and does not overlap with the noninferiority margin2,3

Noninferiority margin

clinically acceptable difference between the investigational treatment and active comparator in a noninferiority trial. For the investigational treatment to be considered noninferior to an active comparator, the investigational treatment’s effect must not be clinically worse than the noninferiority margin2,3

Superiority comparison

the goal of a superiority comparison is to show that the effect of the investigational treatment is significantly greater than the comparator. A treatment is considered to be superior to a comparator if the difference in effect between the 2 is in favor of the treatment and does not overlap with the line of no difference2,3

1. Multiple Endpoints in Clinical Trials, Draft Guidance for Industry. US Food and Drug Administration Web site. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/multiple-endpoints-clinical-trials-guidance-industry. Updated January 2017. Accessed December 11, 2020. 2. Fleischmann R, et al. Semin Arthritis Rheum. 2016;46(3):279-285. 3. Non-Inferiority Clinical Trials to Establish Effectiveness, Guidance for Industry. US Food and Drug Administration Web site. https://www.fda.gov/regulatory-information/search-fda-guidance-documents/non-inferiority-clinical-trials. Updated November 2016. Accessed December 11, 2020.