乌鸦传媒

COPENHAGEN -- A randomized trial reported here appears to confirm that, contrary to guidance from major rheumatology societies, years-long use of corticosteroids in older rheumatoid arthritis (RA) patients is both effective and safe.

RA patients ages 65 and older receiving low-dose prednisolone (5 mg/day) for 2 years achieved significantly greater reductions in disease activity, as measured by the and American College of Rheumatology (ACR) response criteria, than those assigned to placebo in the closely watched , said Maarten Boers, MD, PhD, of Amsterdam University Medical Center in the Netherlands.

Statistically significant improvements, or trends pointing that way, were also found for a host of more specific efficacy measures, Boers told attendees at the European Alliance of Associations for Rheumatology (EULAR) annual meeting.

Results were also published this week in .

And with regard to safety, with one exception -- infections, which were "mostly not severe," Boers said -- adverse event rates were similar between treatment groups.

Boers explained that, despite guidelines from the and EULAR that recommend steroids only for temporary "bridge" treatment, clinicians routinely prescribe them indefinitely under the belief that they add to the benefit obtained from disease-modifying drugs such as methotrexate and newer targeted therapies.

But those guidelines were based on relatively weak evidence. Until GLORIA began in 2013, long-term steroid use in RA had never been tested in an adequately powered, randomized, placebo-controlled trial.

Now that it has, and was found to be clearly effective, it seems inevitable that the guidelines will be revisited before long.

Boers and colleagues enrolled 451 older patients with DAS28 scores of at least 2.60 (thus showing more than minimal disease activity) without conditions that could be worsened with chronic low-dose prednisolone. Patients were randomized 1:1 to steroid or to placebo, and remained on whatever other RA treatments their clinicians deemed appropriate.

Assessments included DAS28 score (the primary outcome) as well as ACR20/50/70 response rates (20%, 50%, and 70% reductions in symptoms, respectively, by ACR criteria) and radiographic measures of joint damage. Boers and colleagues also analyzed specific types of symptoms such as tender and swollen joint counts and laboratory measures including C-reactive protein and erythrocyte sedimentation rate.

Mean patient age at enrollment was 73, and about 70% were women. Baseline DAS28 scores averaged 4.5. Roughly 80% of patients were taking standard disease-modifying agents, mostly methotrexate; about 15 were using biologic agents. One-quarter were using nonsteroidal anti-inflammatory drugs as well.

Within a few months, mean DAS28 scores fell markedly in both the prednisolone and placebo groups, but the magnitude was greater in the former, and stayed greater through the 2 years of the trial. At the final follow-up, the mean DAS28 score was lower by 0.37 points with prednisolone versus placebo, with a lower bound to the 95% confidence interval of 0.23 (P<0.0001).

The prednisolone group also saw less joint space narrowing during the trial (mean 0.2 vs 1.2 points by Sharp/van der Heijde scoring), as well as fewer erosions (mean 0.1 vs 0.7). Added together, that represented an advantage of 1.7 points for prednisolone in preventing joint damage (95% CI lower bound 0.7, P=0.003).

Response rates by ACR criteria were as follows:

• ACR20: 53% prednisolone, 33% placebo
• ACR50: 30% prednisolone, 13% placebo
• ACR70: 12% prednisolone, 2% placebo

Another point in prednisolone's favor came in a modified "per protocol" analysis focusing on participants who not only adhered strictly to their assigned regimen but also avoided potential complicating factors such as changes in background treatment or receiving intra-articular injections. In this group of about 300, evenly divided between prednisolone and placebo, the difference in DAS28 score at month 3 was 0.62 points (95% CI lower bound 0.44), or nearly twice the 0.37-point difference seen in the primary intent-to-treat analysis.

Serious adverse events (SAEs) occurred in 25 patients in each group. Boers and colleagues also tracked certain adverse events "of special interest" that might be expected from long-term steroid treatment, even at low doses. These included all SAEs plus any event leading to discontinuation, cardiovascular events, infections, and bone fractures, plus new-onset hypertension, diabetes, or ocular conditions.

No difference was seen for any of these except infections. These occurred at rates of 42 per 100 patient-years with prednisolone versus 30 for placebo. Patients receiving prednisolone had a mean 0.01 g/cm³ decrease in lumbar spine bone mineral density (BMD), whereas the placebo group saw an increase of 0.03 g/cm³ (P<0.001), but no significant difference between groups for total hip BMD. Fracture rates were similar, with 12 and 10 in the prednisolone and placebo groups, respectively.

Discontinuation rates were also similar, at about 80 in both groups. Boers commented that the COVID-19 pandemic played a role there, as some participants were unwilling or unable to come in for all the evaluations required in the trial. About half of discontinuations were for adverse events or worsened RA, and these, too, were nearly identical in both groups.

Boers summed up the GLORIA results as demonstrating "powerful long-term effects" from the steroid therapy with a "tradeoff of 24% increase in patients with mostly non-severe" adverse effects.

"These results are immediately applicable in clinical practice," he told the EULAR audience.