Second Time's the Charm for Novel SLE Drug?

ATLANTA -- The second phase III trial with anifrolumab, an investigational type 1 interferon receptor antagonist for treating systemic lupus erythematosus (SLE), turned out better than the first, a researcher said here.

Unlike the drug's first pivotal trial, TULIP-1, which ended with a missed primary endpoint, anifrolumab met the primary endpoint in TULIP-2 as well as showing efficacy on a host of secondary outcomes, reported Eric Morand, MD, PhD, of Monash University in Melbourne, at the American College of Rheumatology (ACR) annual meeting.

With a so-called BICLA response rate of 47.8% with anifrolumab at week 52, compared with 31.5% for placebo (adjusted P=0.0013), the drug established superiority on an accepted measure of efficacy, Morand said at an ACR late-breaking abstract session, as well as demonstrated acceptable safety.

But whether that's enough to make the drug approvable is anyone's guess. And when pressed during the session's question-and-answer period, Morand declined to attempt a guess.

In large part, the uncertainty stems from the investigators' decision to change TULIP-2's primary endpoint partway through, to one that had a favorable outcome in TULIP-1, while downgrading TULIP-2's original primary endpoint -- the one that failed in TULIP-1 -- to sub-secondary status. That's in addition to the usual questions that arise whenever similarly-designed trials end with conflicting results.

On the other hand, Morand declared, "TULIP-2 was a positive phase III trial, and there aren't many times that sentence has been spoken."

'The Totality of the Data'

In introducing his talk, Morand noted that only one other lupus drug has been approved in the last 60 years. That's belimumab (Benlysta), which was dogged with disappointing trial results before finally winning approval in 2011.

Hopes were high for anifrolumab after phase IIb results strongly supported efficacy. The primary endpoint in that study was a composite of an SLE Responder Index score of 4 (SRI4) along with reduction in steroid dose to <10 mg/day. For TULIP-1, the investigators went with SRI4 alone as the primary endpoint (SRI4 is itself a composite, encompassing four different measures of reduced or stabilized disease activity).

TULIP-2 began while the first trial was still underway, and also had SRI4 as its primary endpoint. But when the TULIP-1 failure was discovered, the TULIP-2 leaders switched to the BICLA measure, a secondary endpoint in TULIP-1 that was easily met. This was before TULIP-2 assignments were unblinded, Morand said, so the investigators couldn't know at that point whether it would succeed.

BICLA -- the BILAG-based Composite Lupus Assessment, and thus an acronym built on another acronym -- was a six-part measure including clinical improvement on BILAG assessment, change in other organ system involvement, change in SLE Disease Activity Index (SLEDAI) score, change in patient global assessment, use of rescue medications not in the study protocol, and continuance on assigned therapy.

Meanwhile, SRI4 was reclassified to a status indicating that "no inferences can be drawn" from its results -- which did indicate success, albeit with a less stringent statistical analysis than the primary and secondary endpoints (For those, criteria for statistical significance included adjustment for multiple comparisons). In TULIP-2, SRI4 was achieved by 55.5% of patients on anifrolumab versus 37.3% of the placebo group (unadjusted P<0.001).

Three of five "key secondary endpoints" reported by Morand were met by the stricter criteria. These included week-52 BICLA response in patients with high interferon levels at baseline, sustained reduction in oral dosing to 7.5 mg/day prednisone equivalent or less, and at least 50% reduction in CLASI score for skin involvement.

Joint count responses at week 52 and annualized flare rate trended in favor of anifrolumab but didn't meet the significance threshold.

Line graphs illustrating the time course of various response measures, including BICLA, indicated that anifrolumab clearly separated from placebo by about week 20. For example, at week 24, some 30% of the placebo group had experienced a first lupus flare, compared with about 18% of those on anifrolumab.

On one key biomarker measurement, the trial demonstrated that anifrolumab hit its target hard, with type 1 interferon gene signature values dropping to about 5% of baseline by week 12. But some other biomarkers showed less effect, with no real differences seen in antibodies to double-stranded DNA or in complement levels.

Finally, Morand addressed safety. The only adverse event (AE) clearly more common with anifrolumab was herpes zoster activation (shingles), seen in 7.2% of patients receiving the active drug versus 1.1% of the placebo group. Serious AEs were more common with placebo, driven by higher rates of pneumonia and worsening lupus, though the numbers were still small. One patient in the trial died -- a case of community-acquired pneumonia in a patient on anifrolumab.

When, following his talk, an ACR attendee asked whether he thought anifrolumab is approvable, given the disparate findings in the two phase III trials, Morand said "the totality of the data" indicate the drug works to improve lupus outcomes.

Design Details

The study randomized 362 patients 1:1 to the two treatments. About one-third of patients were enrolled in North America and one-quarter from Europe; the rest were from Latin America, "Asia-Pacific" and "other."

Anifrolumab was infused at 300 mg every 4 weeks through week 52. A long-term extension is now underway, Morand noted.

To be eligible, patients needed SLEDAI scores of at least 6, at least one BILAG A item or two B items, patient global assessment of at least 1, positive test for common lupus autoantibodies, and current receipt of standard care.

For patients coming into the study on corticosteroid doses of 10 mg prednisone equivalent or more -- about half of each group -- the protocol required attempts to reduce doses to 7.5 mg or less.

Of the roughly 180 patients in each group, 130 in the placebo group and 153 assigned to anifrolumab completed the trial. The most common reasons for dropping out were AEs, worsening condition or lack of improvement, and withdrawal of consent. All three of those were more frequent with placebo than anifrolumab.

Comment