Experts Discuss the Quandary of Whether to Discontinue or Taper Treatment and When

For the last few decades, clinicians have been trying to figure out how to decrease therapy in patients with rheumatoid arthritis who have reached stable remission. Various strategies for tapering or discontinuing treatment altogether have been explored to inconclusive results. This is not to say tapering therapy has been unsuccessful or discontinuing absolutely leads to immediate or uncontrollable flare. Flares happen and re-initiation of treatment is often necessary. Therefore, the questions at hand pertain to timing, dosage levels, clinical decision-making, and the interplay between steroid and biologic therapies.

Paul Emery, MD, professor of rheumatology, director of The Leeds Institute of Rheumatic and Musculoskeletal Medicine and director of the Leeds Musculoskeletal Biomedical Research Unit at Leeds Teaching Hospital NHS Trust, outlined the basics of the discussion.

“The first study looking at discontinuing therapy was conducted in the 1990s,” he said. “The idea was simple: Get them into deep remission, then you can stop therapy. The quicker you can do it, the better.”

That is about where the simplicity stopped, according to Joshua F. Baker, MD, MSCE, assistant professor of rheumatology and epidemiology at the University of Pennsylvania.

“The way to approach discontinuing or tapering may depend on the population and the clinical context,” he said. “There is a substantial risk for relapse with discontinuation.”

The issue is complicated by a few factors, according to Baker.

“One question is about how to taper down to minimal effective therapy and to find out if and when it is possible to completely withdraw,” he said. “Another question is about inducing remission. There are data showing it is possible to withdraw biologic therapy after remission has been induced during a clinical trial, but a number of these patients did not need biologic therapy in the first place. I am therefore skeptical of some of the available data.”

Baker said the key question is more about the aggressiveness of tapering or discontinuing treatment rather than a cut-and-dried do-or-do not binary. Emery added there are separate conversations about methotrexate (MTX) and biologic therapies.

“At least for biologic therapies, of course it is cheaper if you do not have to give them long term,” Emery said.

The pathway forward demands more sophisticated analysis of disease state in remission and accurate predictors of sustained remission, according to Emery.

“This would begin to allow us to select individuals, as opposed to populations, for dose reduction and possible subsequent discontinuation,” he said.

Before digging more deeply into the issues, however, it may be useful to review current American College of Rheumatology (ACR) and EULAR recommendations.

ACR/EULAR Recommendations

Once stable clinical remission is achieved, disease-modifying antirheumatic drug (DMARD) therapy may be tapered, according to the authors. They noted DMARD withdrawal is contraindicated in early rheumatoid arthritis (RA) among MTX-naïve patients. Recommendations also suggested that most patients with established RA flare when tumor necrosis factor (TNF) inhibitors are withdrawn.

DMARD tapering may be considered when DAS28 of less than 2.6 is sustained in both early and established RA, according to the ACR/EULAR document. In a 2016 update, the panel wrote, “If a patient is in persistent remission after having tapered glucocorticoids, one can consider tapering [biologic disease-modifying antirheumatic drugs] bDMARDs, especially if this treatment is combined with a [conventional synthetic disease-modifying antirheumatic drug] csDMARD.”

These recommendations only carry clinicians so far. Most decisions are made on a case-by-case basis and involve multiple factors ranging from patient history to the specific drugs involved.

“The only consensus that people would come to involves the patients that we all agree are doing well,” Baker said. “But there is a huge group of people who would cause argument among rheumatologists about whether there is still active disease. They should be doing well in many, if not all, aspects before we withdraw. If there is any confusion at all, then definitely do not withdraw, and think hard about how best to taper.”

Emery noted two of the landmark studies in the field provide no concrete conclusions. “Current data from the OPTIMA and PRIZE studies show variable results,” he said.

Heterogeneity

T. Martijn Kuijper, MD, MSc, of the Department of Rheumatology at the Erasmus Medical Center, and colleagues aimed to investigate the patient characteristics clinicians used to determine whether de-escalation of therapy is safe and feasible in individual patients. They also aimed to assess the level of heterogeneity of this clinical decision-making and to determine an optimal strategy for de-escalation. The study included results from 156 clinicians in The Netherlands who completed a questionnaire. The following five patient characteristics underwent analysis: number of swollen joints; presence of DAS remission or low disease activity; patient history; duration of remission or low disease activity; and patient willingness to de-escalate DMARD therapy. Results showed the number of swollen joints and patient history were the most important factors in the decision to de-escalate. Further analysis indicated there were five subgroups of clinicians with different levels of willingness to taper therapy and utilization of patient experience as a factor: those who always taper; those who taper in absence of swollen joints; those who taper in the absence of swollen joints and presence of favorable patient history; those who taper in DAS remission and favorable patient history; and those who consider all factors. De-escalation of TNF inhibition, as opposed to discontinuation of MTX, was preferred.

“Rheumatologists are not uniform in their decision on whom to de-escalate,” the researchers concluded.

“The aim of this study was to gain insight into the relative importance of patient characteristics that a rheumatologist should consider in their decision to taper DMARDs in RA patients in remission or with low disease activity,” Kuijper told Healio Rheumatology.

He noted the study was conducted using the methodology of a discrete choice experiment, in which rheumatologists were asked to complete a series of choice tasks. Each choice task consisted of two hypothetical patients for whom characteristics varied.

“Rheumatologists were asked to choose the patient he or she deemed most eligible for treatment de-escalation or indicate if neither patient was eligible,” Kuijper said. “Based on the choice patterns, the relative importance of varying patient characteristics in decision-making can be derived.”

Kuijper stressed the importance of the five groups. “However, subgroup analysis of rheumatologists by latent class model revealed heterogeneity among rheumatologists with respect to the relative importance of characteristics,” he said. “This was exemplified by calculating for each hypothetical patient, the probabilities rheumatologists would decide to de-escalate DMARDs based on their class.”

The findings suggested that probabilities were high for all five classes for the patients having the most favorable characteristics, according to Kuijper. “However, among other patients, class probabilities were highly varying,” he said. “This shows little consensus exists among rheumatologists about which patients are eligible. This is likely because accurate prediction of which patients may successfully de-escalate treatment is not yet possible and hence, there is also a lack of guidelines on this topic. Although current ACR/EULAR guidelines do state that de-escalation of biologics may be considered in patients in remission whose glucocorticoids have been tapered.”

Kuijper concluded that his research also has not provided answers about discontinuing MTX. “My best answer at this point is to follow ACR/EULAR guidelines,” he said.

Kuijper highlighted another data set from his group. This set of findings indicated de-escalation of TNF-blockers in RA patients in low disease activity (LDA)/remission resulted in a 1-year flare rate of about 34%. “Hence treatment de-escalation appears to be feasible in a large part of patients without short-term flare,” he said. “Of patients who do flare, most studies indicate that the majority regains remission within 6 months after treatment intensification. Evidence on radiographic progression is scarce, but appears to be limited in the studies that did investigate this and not significantly higher in the de-escalation group.”

Small sample sizes may impact the level of evidence, Kuijper added, and noted some data indicate halving the dose of TNF inhibitors may be non-inferior to continuing a full dose in terms of flare risk. “This could, of course, lead to great cost-reductions if the majority of patients could be safely switched to use half of the dose,” he said. “However, these findings should be confirmed first by other trials.”

A sundry component to the discussion is that all available therapies are not curative. They simply control the disease. “When you are dealing with the immune system, you can never be comfortable saying the disease is gone,” Baker said. “I keep coming back to the idea that there is always a risk. So, that leaves us the challenge of predicting those factors that will lead to relapse and identify patients that have T cells floating around. That will help us make better clinical decisions.”

Patient Population

Alverini and colleagues evaluated the possibility of using clinical and ultrasonography-based remission to select patients who may be eligible for tapering or discontinuation of TNF inhibitors. The study included 42 patients with RA who were in sustained remission, which the researchers defined as DAS28 of less than 1.6 at three visits 3 months apart, according to the findings. Participants underwent synovial hypertrophy and power Doppler signal presence. Five patients with positive synovial hypertrophy and negative power Doppler results underwent ultrasound-guided synovial tissue biopsy.

In these patients, CD68+ cells were found in the lining layer, while CD3+ and CD20+ cells also infiltrated at the time of sustained clinical remission. The relapse rate was 30.9% after tapering of TNF therapy. Among patients in the positive synovial hypertrophy and negative power Doppler group, 69.1% discontinued treatment after sustained remission. Of these 29 patients, 26 patients (89.7%) maintained in remission at 6 months. The previous biologic was used to treat relapses and yielded a EULAR response within 3 months of follow-up. “[Ultrasound] evaluation using [power Doppler] signaling allows the identification of patients with RA in clinical and histological remission after tapering and discontinuing biologics,” the researchers concluded.

These findings confirm DAS28 numerical remission alone may be insufficient to predict who will be able to successfully taper and/or discontinue biologic therapy, according to Allan Gibofsky, MD, JD, professor of medicine at Weill Cornell Medicine and attending rheumatologist at Hospital for Special Surgery. “Inflammation may still be present in the synovium of patients who achieve DAS28 numerical remission,” he said. “While DAS28 remission has been the gold standard in clinical trials that look at tapering and discontinuation, it is not the disease activity measure most widely or routinely used by U.S. rheumatologists in clinical practice who utilize disease activity measures.”

Emery added that patients with the same DAS28 score may have different levels of active inflammation, swollen joints or CRP. “DAS28 requires some interpretation,” he said.

The most commonly used activity scores used in the United States are the CDAI and the RAPID 3, Gibofsky said. “These are more easily administered and do not require the incorporation of the results of a laboratory test — [erythrocyte sedimentation rate] ESR or CRP — as these results are not available when the patient is being seen. Further, we are learning that patients who are in ‘numerical’ remission may still have synovitis on ultrasound or MRI and persistent inflammation. Thus, tapering or discontinuing biologic therapy may result in incomplete control of subclinical inflammation,” Gibofsky said.

Another factor in the discussion is the disparity between low disease-activity states, according to Emery. “Patients with the same DAS may be completely different from a disease perspective,” he said.

He added that serological status and history of how they reached LDA also are factors. The parameters of remission may be defined clinically, immunologically or through imaging tests, according to Emery.

“Patients with less immunological drive are more likely sustained remission on dose reduction,” he said.

Bouman and colleagues assessed whether the multi-biomarker disease activity (MBDA) score was useful as a predictive tool in a cohort of 180 patients from a registry in The Netherlands. In a randomized controlled trial, tapering of TNF inhibitors until discontinuation or flare was compared with standard care in the patient population, which included patients with long-standing disease and stable disease activity. The researchers defined flare as DAS28-CRP increase greater than 1.2 or greater than 0.6 if the current DAS was 3.2 or greater. A major flare was defined as one lasting longer than 3 months. The final analysis included 115 patients in the tapering group and 56 in the usual care group. Baseline MBDA failed to predict clinical outcomes among patients who were tapered, according to results of the AUROC analysis. However, the score predicted major flares among patients treated with standard care (AUROC = 0.72). MBDA also failed to predict radiographic progression, which was minimal in the cohort.

“In this disease activity-guided strategy study of TNFi tapering in RA patients with low disease activity, baseline MBDA score was not predictive for successful tapering, discontinuation, flare, major flare or radiographic progression in patients who tapered TNFi,” the researchers concluded.

Cyclic citrullinated peptide (CCP) positivity or negativity should be considered, as well, as should the way the patient reached low disease activity, according to Emery. Patients may have achieved this state through some combination of DMARD or MTX or induction with MTX and a TNF inhibitor. “All of the above will influence likelihood of flare on tapering,” Emery said.

Baker summed up the challenges clinicians face. “There are so many different ways of defining remission, and they all have flaws,” he said. “When you cannot define the patient population, it makes it difficult to treat them.”

The catch-22 is both tapering and ceasing treatment can come with pitfalls. Edwards and colleagues suggested the steps after remission or LDA remain uncertain for many clinicians. They conducted a meta-analysis of 52 papers to assess the clinical utility of tapering biologic therapy. They concluded that remission is not typically sustained among patients who discontinue treatment. Relapse and flare rates range from 48% to 54% for early RA, from 2% to 84% for established RA, 11% to 53% for axial spondyloarthritis, and 44.9% for psoriatic arthritis. They noted that re-treatment can lead to acceptable disease activity. “More research is needed to understand the long-term impacts of these strategies on efficacy, safety and cost,” they concluded.

“There are data to support tapering therapy,” Baker said. “Relapse risk ranges between 20% and 60%, depending on the study.”

Emery built on this point. “Even in the case of a flare, if a patient is in good remission, it is possible to regain the remission after you stop or taper medication.”

Patient’s Role

Markusse and colleagues recorded interviews with 20 patients with RA about treatment and discontinuation. Transcriptions recorded reported emotions surrounding discontinuation, from hope, happiness and relief, to fear and disappointment. While some patients expected they or other patients would be able to discontinue therapy, others did not have such expectations. At the prospect of increased disease activity after discontinuation, respondents expected a decrease once medication was re-initiated, but that the decrease would take time. Patients simultaneously felt positive about the option to taper or discontinue, but felt negatively about the possible increase in disease activity.

“In particular, patients expect that disease activity will flare and that improvement upon restarting medication will take time,” the researchers wrote. “Patients’ expectations and feelings should be addressed before drug tapering is attempted in a clear strategy of continued monitoring of disease activity.”

Gibofsky offered a broader perspective on the patient’s role in the equation.

“Often, it is the patient who seeks to taper or discontinue biologic therapy, usually for reasons of cost or convenience,” Gibofsky said. “We have an obligation to make them aware of the risks and benefits. While many patients may remain in remission, some will not and there is no good way of predicting this population. Further, if a patient flares after tapering or discontinuation, there is a chance that he or she may not be able to re-achieve their previous state of control. The patient should be made fully aware of these risks before tapering is attempted. The extent of this dialogue and how often it actually occurs is unknown at present.”

Baker’s view is that both the physician and patient should have “buy-in” when deciding how to proceed.

“If both agree, that is the best-case scenario,” he said. “But, it is often the case that the patient feels ready to stop taking medication but the physician is worried about subclinical activity. The other scenario is that the physician thinks the patient is doing well, but the patient is not ready because he or she has stiffness or triggers. Either scenario is not the right time to taper therapy.”

Something else to consider is the patient needs to be prepared for the risks of withdrawing, according to Baker. “This includes a certain degree of optimism,” he said. “We as clinicians cannot force the issue. Some physicians may disagree with that concept. While it is possible that some people do not need as aggressive therapy, we need to make sure the conversation takes place and everyone involved feels it is the right time.”

Tofacitinib, Leflunomide, MTX

Kubo and colleagues investigated the possibility of discontinuation in a cohort of 64 patients treated with tofacitinib. The phase 3, multicenter, non-randomized, open, prospective, observational study was conducted with the primary outcome measure of the proportion of patients not receiving the study drug at 52 weeks after treatment. Ten patients continued therapy and 54 patients discontinued. At the study’s end, 37% of patients who discontinued were free of the drug with no flares. Other results from 52 weeks after treatment, however, indicated disease activity was higher among patients who discontinued than in those who continued. Looking closer at the discontinuation group, baseline rheumatoid factor (RF) titer was 40 U/mL in patients who remained tofacitinib-free and 113 U/mL in patients who did not. A higher ratio of patients with lower RF at baseline were drug-free at study end compared with patients who reported a higher baseline RF, according to the results. For patients who failed to get off the drug entirely, re-initiation of therapy with tofacitinib or another biologic therapy yielded clinical improvement. “It is possible to discontinue tofacitinib without flare in about a third of patients with RA,” the researchers concluded. “A low RF predicts maintenance of LDA after discontinuation of tofacitinib.”

“We have gotten better at discontinuing these therapies,” Emery said. “Also, patients do not want to take drugs they do not need.”

Emery added that tapering biologics dramatically reduces the cost to the health care system. Yet, there is also a clinical component.

“While it is probably safer to be on therapy, even if you do not need it, to reduce flare. Logically, if you got rid of the mediating cytokine, then you do not need to block it with the same amount of drug,” Emery said.

Li and colleagues conducted a multicenter, randomized, parallel trial in 346 patients with active RA to determine whether prolonged DMARD therapy yielded high remission and low relapse rates. Eligible participants had a DAS28-ESR greater than 5.1. In the first phase of the trial, intense MTX, leflunomide and hydroxychloroquine therapy was administered for as long as 36 weeks with the goal of reaching DAS28 of no greater than 2.6 or LDA, which the researchers defined as DAS28 between 2.6 and no greater than 3.2.

The second phase was the tapering phase, with patients randomly assigned to either leflunomide plus hydroxychloroquine combination or leflunomide monotherapy. EULAR response and a decrease of DAS28 by at least 1.2 served as the primary endpoint. At 12 weeks, 18.7% of patients reached a strong EULAR response, according to the findings. This rate increased to 36.9% at 24 weeks and to 54.1% by 36 weeks. Good or moderate EULAR responses were reported in 75.4% of patients at the 36-week mark. “Compared with those achieving low disease activity and a high health assessment questionnaire (HAQ > 0.5), patients achieving remission (DAS28 2.6) and low HAQ ( 0.5) had a significantly higher retention rate when tapering the DMARDs treatment,” the researchers wrote. They added that no advantage was reported for tapering to combination as opposed to monotherapy. It was concluded that LDA when tapering begins may reduce flare risk. “Leflunomide is a good maintenance treatment as single treatment.”

Gibofsky offered perspective on the Li data set. “There appeared to be no advantage on tapering to combination rather than monotherapy,” he said. “Remission was achieved in a proportion of patients with RA receiving prolonged intensive DMARD therapy. Low disease activity at the start of disease taper leads to less subsequent flares. Leflunomide is a good maintenance treatment as single treatment. Of note, none of these patients were on biologic therapy, so the findings cannot be extrapolated to patients who are on those agents.”

Manders and colleagues investigated tapering or discontinuation rates for MTX in a cohort of patients treated with a TNF inhibitor from the Dutch Rheumatoid Arthritis Monitoring registry. Methotrexate tapering was reported in 458 patients, or 34% of the overall population. Ten percent of this group discontinued MTX, while 56% continued the drug at the same dosage level. An average improvement in DAS28 score was reported after tapering and discontinuing MTX at 6 months and 12 months, according to the findings. The relapse rate — defined as a DAS28 increase greater than 0.6 — was 21% in the taper group at 6 months and 12 months. Among patients who discontinued, the rate was 21% at 6 months and 24% at 12 months. The researchers noted tapering or discontinuing did not impact long-term survival on TNF inhibitors.

“Patients who taper and discontinue MTX have a similar DAS28 score over time as patients who continue MTX,” they wrote.

Emery noted CCP positivity correlates with a strong MTX response. “The other thing to consider with methotrexate is that it is cheap,” he said. “There is not a big cost to society if you continue it. At least part of our clinical decisions is dictated by the cost of therapies.”

Existing Data

Until recommendations become more concrete, experts recommend paying close attention to available evidence for help. In the PRIZE study, Emery and colleagues outlined factors that may be useful for predicting successful tapering or withdrawal. They include early treatment, speed of reaching remission, duration of remission, absence of stress-induced cytokines, low or normal HAQ, good quality of live and low baseline DAS.

Emery also highlighted the C-EARLY, PRESERVE and OPTIMA studies as sources of possible information. However, he stressed caution when interpreting data. “New data sets remain the same as past data sets,” he said. “No study cured a patient.”

Kuijper said results of the tREACH trial showed DAS44 may be a useful marker for remission, as well.

Baker summed up the arguments for and against the data and the decision-making. “My opinion is that we should be tapering carefully, and only among people who are in deep remission,” he said. “At the moment, this is only possible in a minority of patients.” – by Rob Volansky

• References:
• Alivernini S, et al. Arthritis Res Ther. 2016;doi:10.1186/s13075-016-0927-z.
• Bouman CA, et al. Rheumatology (Oxford). 2017;doi:10.1093/rheumatology/kex003.
• Edwards CJ, et al. Rheumatology (Oxford). 2017;doi:10.1093/rheumatology/kew464.
• Kubo S, et al. Rheumatology (Oxford). 2017;doi:10.1093/rheumatology/kex068.
• Kuijper TM, et al. Arthritis Res Ther. 2017;doi:10.1186/s13075-017-1287-z.
• Kuijper TM, et al. Ann Rheum Dis. 2016;doi:10.1136/annrheumdis-2016-209272.
• Li R, et al. Medicine (Baltimore). 2016;doi:10.1097/MD.0000000000003968.
• Manders SH, et al. RMD Open. 2015;doi:10.1136/rmdopen-2015-000147.

• For more information:
• Joshua F. Baker, MD, MSCE, can be reached at 8 Penn Tower Building, Civic Center Blvd. at 34th St., Philadelphia, PA 19104; email: joshua.baker@uphs.upenn.edu.
• Paul Emery, MD, can be reached at 2nd Floor, Chapel Allerton Hospital, Chapeltown Rd., Leeds. LS7 4SA. United Kingdom; email: p.emery@leeds.ac.uk.
• Allan Gibofsky, MD, JD, can be reached at Hospital for Special Surgery, 535 East 70th St., New York, NY 10021; email: gibofskya@hss.edu.
• T. Martijn Kuijper, MD, can be reached at Room Na-609, PO Box 2040, 3000 CA, Rotterdam, The Netherlands; email: t.kuijper@erasmusmc.nl.

Disclosures: Baker and Kuijper report no relevant financial disclosures. Emery reports he has undertaken clinical trials and has provided expert advice to Pfizer, MSD, AbbVie, Bristol-Myers Squibb, UCB, Roche, Novartis, Samsung, Sandoz and Eli Lilly. Gibofsky reports he is a consultant for AbbVie, Celgene, Eli Lilly, Novartis and Pfizer.