Efficacy and Safety of Belimumab in a Subgroup of Systemic Lupus Erythematosus (SLE) Patients with Higher Disease Activity (anti-dsDNA positive and low complement): A Pooled Analysis of the HGS1006-C1056 (BLISS-52) and HGS1006-C1057 (BLISS-76) Studies

Studies C1056 and C1057 followed very similar protocols, were of nearly identical design, had common inclusion and exclusion criteria, and were conducted over the same time period. Nevertheless, given the heterogeneous presentation of SLE disease and the fact that the Phase III program was run globally, variation in the patient population, both within the studies (e.g., between different centres) and between the studies (analogous to differences between centres within the same study) should be expected.

Since it has been established that the conduct of the studies was effectively the same, it then must be determined whether the relative treatment effect is different in one study compared with the other study when evaluating whether two studies are similar enough to pool. Each of these Phase III studies achieved statistical significance for belimumab 10 mg/kg on the pre-specified primary endpoint of SRI response at Week 52; therefore, these nearly identical studies provide independent replication of results. While pooling is not necessary to establish the effectiveness of belimumab, it was considered appropriate in order to evaluate treatment effects in the high disease activity subgroup of interest, given that the individual studies were not designed to provide sufficient power to demonstrate effectiveness within subgroups. Thus, statistical evaluation pooling the studies and testing for a treatment-by-study interaction was undertaken. A significant treatment-by-study interaction would indicate that the relative treatment differences were statistically different in the two studies and pooling would not be justified. Conversely, the lack of a treatment-by-study interaction would indicate the studies resulted in a similar treatment response and pooling would be justified.

When the two Phase III studies were pooled for the SRI analysis, the treatment-by-study interaction was >0.5. Likewise, for the target population of high disease activity, the treatment-by-study interaction was >0.7 suggesting that the high disease activity subgroup may be more homogenous and therefore have a more similar treatment effect between the studies than the population as a whole.

Given these considerations, it is reasonable and valid to pool the two studies and allows better precision for evaluation of subgroups.

Inclusion Criteria:

• Clinical diagnosis of SLE by ACR criteria.
• Active SLE disease.
• Autoantibody-positive.
• On stable SLE treatment regimen.

Exclusion Criteria:

• Pregnant or nursing
• Have received treatment with any B cell targeted therapy.
• Have received treatment with a biological investigational agent in the past year.
• Have received IV cyclophosphamide within 180 days of Day 0.
• Have severe lupus kidney disease.
• Have active central nervous system (CNS) lupus.
• Have required management of acute or chronic infections within the past 60 days.
• Have current drug or alcohol abuse or dependence.
• Have a historically positive test or test positive at screening for HIV, hepatitis B, or hepatitis C.

Subjects who were anti-dsDNA positive and had low C3 and/or C4 at baseline (Low C+anti-dsDNA Population) Subjects with baseline SELENA SLEDAI score >= 10 (SS >=10 Population)