A massive Swedish cohort study suggesting that biologic agents reduce the known excess risk of lymphoma in patients with rheumatoid arthritis was hailed as one of the high points of the last year by speakers at the 2021 Rheumatology Winter Clinical Symposium.
The speakers gave additional shout-outs regarding other key developments in RA during the pandemic year, including two huge studies showing that even low-dose corticosteroids at 5 mg/day or less carry the price of increased risk of cardiovascular disease and serious infections; Swedish data showing that RA disease activity can be used for venous thromboembolism (VTE) risk stratification; evidence from the venerable BeST study showing at 17 years of follow-up that treat-to-target strategies didn’t prevent an increased mortality risk of RA compared to the general population; and a signal that artificial intelligence using machine learning will become useful in clinical practice sooner than many realize.
Also, the speakers offered kudos to the investigators in the SEAM-RA trial, whose results they expect to be practice changing. In addition, they praised the SELECT-CHOICE trial for providing insight into the relative safety and efficacy of upadacitinib (Rinvoq) and abatacept (Orencia) as second- or third-line therapy in RA. And they warned their colleagues of rocky upcoming months for tofacitinib (Xeljanz) as the Food and Drug Administration reviews concerning new data from a long-term postmarketing safety trial.
Swedes Show Biologics May Reduce Excess Lymphoma Risk in RA
Decades ago, Swedish investigators harnessed their nation’s comprehensive linked health registries to demonstrate that patients with RA are inherently at greater risk for developing lymphoma. Now the Swedes have used the registries to identify a strong signal that this excess risk of lymphoma is reduced in RA patients treated with biologic drugs.
The study included 16,392 RA patients who initiated treatment with a biologic during the study period, 55,253 who were biologic-naive, and 229,047 age- and sex-matched controls from the general population. During the years 2006-2016, there was an adjusted 31% reduction in the risk of incident lymphoma in RA patients who started a first biologic agent, compared with those who did not, and a 54% reduction in lymphoma in biologic-treated patients, compared with those who switched from one conventional synthetic disease-modifying antirheumatic drug (DMARD) to another.
Moreover, while RA patients with 0-2 years of accumulated time on biologics were at a statistically significant 3.42-fold increased risk of lymphoma, compared with the matched general population, that risk shrunk with longer exposure to these medications, dropping to a nonsignificant 1.53-fold increase with 3-4 years of biologic exposure and a 1.04-fold risk with 5 years or more. The lymphoma risks didn’t vary significantly among the TNF inhibitors, nor with the use of TNF inhibitors versus other biologics.
“I think this is really quite encouraging,” said John J. Cush, MD, professor of internal medicine at the University of Texas, Dallas. “If the lymphoma risk was related to inflammation, I think you’d see results like this. It really does speak to the fact that control of inflammation over time does lessen one’s risk of cancer.”
Arthur Kavanaugh, MD, observed that patients with RA continue to ask their rheumatologists if going on a TNF inhibitor will increase their risk of lymphoma, so these Swedish data are going to be useful in conversations in the clinic.
“I thought this study was very exciting,” said Kavanaugh, professor of medicine at the University of California, San Diego, and director of RWCS. “Maybe we can make the case now that the longer you’re on biologic therapies and the more recently you’re being treated, the better you do in terms of lymphoma risk.”
Beware Underappreciated Hazards of Very-Low-Dose Glucocorticoids
The increased risk of osteoporotic fractures associated with corticosteroid therapy gets tons of attention. But the past year brought two cautionary studies that newly identified significantly increased risks of cardiovascular events and serious infection as well, even at doses of less than 5 mg/day, which have traditionally been believed safe long term.
Investigators at the University of Leeds (England) conducted a population-based cohort study in which they used the U.K. Clinical Practice Research Datalink to analyze the oral corticosteroid dose-dependent cardiovascular risk in 25,324 patients with RA and 62,470 with five other immune-mediated inflammatory diseases: inflammatory bowel disease, giant cell arteritis, polymyalgia rheumatica, systemic lupus erythematosus, and vasculitis. None had known cardiovascular disease at baseline. The analysis accounted for changes in oral steroid dose over time and for the use of potentially confounding medications.
Over a median 5 years of follow-up, a strong steroid dose-dependent risk of acute MI, heart failure, atrial fibrillation, cerebrovascular disease, and abdominal aortic aneurysm was evident in patients with any of the six of the immune-mediated inflammatory diseases. The 1-year adjusted cumulative risk of major adverse cardiovascular events jumped from 1.4% during periods of nonuse to 8.9% at a daily dose of 25 mg or more. Notably, when patients were on oral steroids at doses less than 5 mg/day, their risk of one or more of these major adverse cardiovascular events was 74% greater than when off steroids.
A separate red flag regarding low-dose glucocorticoids was raised by a retrospective cohort study of the 1-year risk of hospitalization for serious infection in 172,041 RA patients in the Medicare claims database for the years 2005-2015. After 6 months of stable use of DMARDs, 47.1% of those patients were on corticosteroids. Their 1-year cumulative incidence of severe infection resulting in hospitalization was 11% if their dose was 5 mg/day or less, 14.4% at >5 to 10 mg/day, and 17.7% at >10 mg/day, all significantly greater than the 8.6% rate in patients not on steroids. The same pattern held true when the investigators separately scrutinized 44,118 RA patients in the Optum Clinformatics Data Mart database for the years 2006-2015.
“This is a really important paper telling us what you’ve always known, which is that there’s a dose-related increase in serious infections in people on steroids. Everybody knows that, but usually people think of 10 mg/day and above,” Cush said. “This is a little bit sobering for those of you who think, ‘Well, I’m doing OK using 2.5 or 4 mg/day.’ In this study, even at doses of 5 mg/day or less, there’s a significant increase in serious hospitalizable infections.”
Taken together, the message of these two huge studies is clear, he added: “The idea is we probably shouldn’t be using steroids.”
Given the large array of therapies now available to address various aspects of the inflammatory and immune responses in RA, Kavanaugh asked, why do so many rheumatologists still maintain their patients on low-dose steroids?
“Because when we find something that works, we stick to it,” Cush replied.
RA Disease Activity Predicts VTE Risk
The risk of VTE climbs with RA disease activity such that patients with high disease activity – defined using the erythrocyte sedimentation rate-based Disease Activity Score in 28 joints (DAS28-ESR) – have double the VTE risk of those in disease remission, according to a Swedish nationwide cohort study using data from the Swedish Rheumatology Quality Register. The investigators concluded that RA disease activity can be used in clinical practice as an additional tool for VTE risk stratification, alongside such established risk factors as immobilization, age, surgery, and comorbid conditions.
The study included 46,316 patients with RA who collectively experienced 2,241 VTE events within 1 year after one of their collective 322,601 visits to their rheumatologist. The overall cumulative 1-year incidence of VTE in the RA population was 0.71%, versus 0.36% in a control group composed of 215,843 randomly selected age- and sex-matched individuals drawn from the general Swedish population. That translates to an adjusted 1.88-fold increased risk in the RA group.
The VTE incidence was 0.52% in the year following a rheumatologist visit at which RA patients were found to be in DAS28-ESR remission. The adjusted risk for VTE climbed significantly in stepwise fashion with increasing disease activity: 12% greater risk with DAS28-ESR low disease activity, 48% greater risk with moderate disease activity, and 2.03-fold greater risk with high disease activity.
SEAM-RA Shows Which Drug to Withdraw to Maintain Remission Gained on Combo Therapy
The SEAM-RA trial presented at the 2020 annual meeting of the American College of Rheumatology (ACR) and published shortly afterward in Arthritis & Rheumatology provides guidance on how to best implement a drug tapering strategy in patients who have achieved sustained remission on combination therapy with etanercept (Enbrel) and methotrexate. The purpose of such a strategy is to reduce patients’ medication burden and exposure to safety concerns inherent in continuing therapy. SEAM-RA included 371 RA patients in sustained disease remission during 24 weeks of open-label treatment with etanercept plus methotrexate. At that point, they were randomized 2:2:1 to 48 weeks of double-blind methotrexate monotherapy with etanercept withdrawal, etanercept monotherapy, or continued combination therapy. Those who experienced disease worsening on double-blind monotherapy were eligible for combination rescue therapy with both drugs.
The primary endpoint was maintenance of remission without disease worsening at week 48. This was accomplished in 52.9% of the combination therapy group, with a similar 49.5% success rate in patients on etanercept monotherapy, both significantly better than the 28.7% rate with methotrexate monotherapy. The inference is that biologic monotherapy might be an advantageous strategy for maintenance after achieving prolonged remission on combination therapy.
“This study could impact guidelines,” Cush predicted.
Among the key study findings, in his view, was that disease worsening occurred early following the switch to methotrexate monotherapy: typically within the first 2-4 weeks. And combination rescue therapy worked: Almost 75% of patients were able to recapture disease remission regardless of which drug they’d been on as monotherapy, and almost 90% of patients on rescue therapy achieved a low-disease-activity state.
Still, just over half of patients were able to maintain long-term remission after downshifting to etanercept monotherapy, Kavanaugh noted. “The bedeviling thing is we don’t know which patients can do this,” he said.
This was made abundantly clear in PREDICTRA, a phase 4, double-blind study in which 122 RA patients in sustained remission on 40 mg of adalimumab (Humira) every 2 weeks were randomized 5:1 to adalimumab taper or withdrawal for 36 weeks. Overall, 36% of the taper group and 45% of those who halted the biologic flared within 36 weeks. Nothing on an extensive list of candidate baseline predictors predicted flare in either group.
“I think the interesting thing here is that nothing predicted it, not drug levels, antidrug antibodies, not even MRI evidence of inflammation while in remission. The lack of predictors of who’s going to do well off therapy makes us crazy,” Kavanaugh said. Also noteworthy was that when full-dose, open-label adalimumab was reinstituted as rescue therapy in patients who flared, only half of them regained remission during the following 4 months, he observed.
Cush didn’t mince words regarding his view of the concept underlying PREDICTRA.
“Why are we even talking about this? We spend our whole lives trying to get these patients into remission, we’re heavily invested in combination therapy, we’re very proud of our great successes, and now we want to find data to condone, ‘Let’s get off drug therapy?’ ” he argued.
“There are untold consequences to this,” the rheumatologist continued. “There are cardiovascular consequences, and we’ve seen the x-ray studies that show that patients may be doing clinically better but still have x-ray worsening because you’ve withdrawn the TNF inhibitor. I think this is all nonsense, and I’m totally against it.”
Making a SELECT-CHOICE in Refractory RA
In patients whose RA has proved refractory to one or more biologics, opting for the oral Janus kinase inhibitor upadacitinib as next-line therapy is significantly more effective than turning to the T-cell costimulation modulator abatacept, but at the cost of a doubled risk of severe adverse events, according to the findings of the phase 3, double-blind, 24-week, randomized SELECT-CHOICE trial.
In this 612-patient study, the week-12 clinical remission rate as defined by a C-reactive protein (CRP)-based DAS28 below 2.6 was 30% in the upadacitinib arm, compared to 13.3% with abatacept. At week 24, the remission rates were 46% and 31%, respectively. However, the severe adverse event rate was 6.3% in the upadacitinib group, including one death, one stroke, and two cases of VTE, compared to 3.2% with abatacept.
“Maybe this is a win for upadacitinib, but abatacept is a win when it comes to safety,” Cush said. “So the question is, when you’re making your second or third change in a biologic DMARD, are you going for safety or are you going for efficacy?”
Make Way for Machine Learning
Studies of machine learning and artificial intelligence aimed at developing a more individualized treatment approach are popping up with increasing frequency at the major rheumatology meetings.
“This is really, I think, the future because for most of us, in most of our choices, as great as all of us are in managing RA, it is a crystal ball and a coin toss as to whether our next therapy is going to work,” Cush said.
He highlighted a study presented at the 2020 ACR annual meeting as an example of how machine learning can generate better evidence for selecting therapies. An international team of investigators used artificial intelligence to develop a simple rule to predict an RA patient’s likely response to the interleukin-6 receptor inhibitor sarilumab (Kevzara). The researchers developed a list of 42 candidate predictors of an ACR20 response, including demographics, biomarkers, and disease activity scores, then applied machine learning to data from the 1,197-patient, phase 3 MOBILITY trial of sarilumab versus placebo in order to generate a predictive rule of better outcomes.
Out of the 42 candidate predictive values, the best predictive rule turned out to be the combination of the presence of anti-citrullinated protein antibodies and a baseline CRP level greater than 12.3 mg/L. This rule was then put to the test using data from four phase 3 clinical trials in which 34%-51% of RA patients randomized to sarilumab were rule-positive. The rule-positive patients had worse baseline prognostic factors and more severe RA. Yet rule-positive patients had a better outcome than that of those who were rule-negative. For example, a week-24 ACR70 response was achieved in 34% of rule-positive patients, nearly fivefold greater than the 7% rate in the rule-negative group.
Treat-to-Target Strategies Don’t Tame Excess Mortality in RA
At a mean of 17 years of follow-up in the Dutch BeST study – a pioneering treat-to-target study in RA – patients with early RA who were assigned to one of four treat-to-target strategies for 10 years showed excess mortality compared to the general age- and sex-matched Dutch population, Johanna M. Maassen, MD, reported at the 2020 annual meeting of the European Alliance of Associations for Rheumatology (EULAR). Indeed, mortality was 37% higher in the RA patients than in the comparator population, with no significant differences in survival curves evident between the four tight-control strategies.
“This raises the question of whether survival in patients not strictly treated to target would have been even lower. Are we not doing as well as we think we’re doing?” Kavanaugh said.
“This is very sobering. It tells us we really do need to be better at managing RA. And I would say based on this data we need to be better at managing RA really early to get better outcomes really late,” said Cush. “The damage that led to cardiovascular risk could have been incurred earlier on and then carried forward when you look out to 15-20 years.”
More Trouble for Tofacitinib
Tofacitinib faces an uncertain future following Pfizer’s late-January press release announcing that the JAK inhibitor failed to meet noninferiority compared to anti-TNF therapy in terms of adjudicated major adverse cardiovascular events (MACE) and adjudicated malignancies in an FDA-mandated postmarketing safety study. The study, known as ORAL Surveillance, or Study A3921133 – ‘1133’ for short – randomized 4,362 RA patients aged 50 years or older with at least one additional cardiovascular risk factor to tofacitinib at 5 or 10 mg twice daily or to a TNF inhibitor.
The Pfizer press release was soon followed by an FDA safety warning and announcement that the agency is reviewing the full study data before making any final recommendations. The FDA already requires black box warnings for tofacitinib regarding increased risks of blood clots and death, but only at the 10-mg twice-daily dose. In the ORAL Surveillance study, however, the primary endpoint was noninferiority of the combined tofacitinib doses, and there was no evidence of a difference in endpoints between the 5- and 10-mg twice-daily doses.
Only the topline results released in the Pfizer press release are publicly available so far. The combined tofacitinib doses were associated with a 1.33-fold greater risk of adjudicated MACE and a 1.48-fold increase in adjudicated malignancies other than nonmelanoma skin cancer, compared to adalimumab or etanercept.
The jury is still out pending release of the full study data, in Kavanaugh’s view.
“When I get to look at the data, I want to see who are the people who had the events, their characteristics,” he said.
Alexis R. Ogdie-Beatty, MD, said the study’s topline results “make me a little nervous.
“The effect size is pretty large. It’s in the 1.4 range, which means there might be something there for real,” said Ogdie-Beatty, a rheumatologist and epidemiologist at the University of Pennsylvania, Philadelphia.
In contrast, Eric M. Ruderman, MD, said he’s not concerned “just yet” about the results.
“Frankly, the FDA was just plain wrong in their press release because the confidence intervals for the MACE events crossed 1.0 meaningfully [0.91, 1.94], so there’s nothing there. And the confidence intervals hit 1.0 for the malignancies,” observed Ruderman, professor of medicine at Northwestern University, Chicago.
This was a nearly 4,400-patient study, and given that it couldn’t convincingly show a safety risk, such risk – if it even exists – would seem to be tiny, he added.
“It’s probably not going to change the way we prescribe these drugs. So I don’t see that trial taking away from the big run for JAK inhibitors we’re going to see in the next year,” the rheumatologist said.
The speakers reported having financial relationships with numerous pharmaceutical companies.
This article originally appeared on MDedge.com, part of the Medscape Professional Network.