Rheumatologists Cautioned to Avoid Overtreatment

Janis C. Kelly

July 13, 2018

Rheumatologists should exercise "a little caution" in applying six common rheumatology treatment paradigms because each might lead to overdiagnosis and overtreatment, a rheumatologist warns in a commentary. Another expert disagrees about the frequency of overdiagnosis, however.

Robert Landewé, MD, PhD, a professor of rheumatology from Amsterdam Rheumatology and Clinical Immunology Center, the Netherlands, published the commentary online July 4 in the Annals of the Rheumatic Diseases.

Landewé discusses early diagnosis, intensive treatment, remission, prognosis and risk stratification, evidence-based rheumatology, and precision medicine as applied in diagnosis and treatment of inflammatory rheumatic diseases. He notes that overdiagnosis is often an adverse effect of screening and can lead to overtreatment, psychological distress, and "huge costs."

"After a long history of therapeutic nihilism and acquiescence, the focus in rheumatology has shifted from caring for the disabled to actively finding the unrecognized, from a wait-and-see policy to early intervention, and from careful step-up treatment with poorly effective but toxic drugs to immediate intervention with powerful new drug combinations," he explains.

"To date, no one seems to bother about overdiagnosis and overtreatment, since for the first time in history we have an opulence of effective drugs and more are coming."

Is Overdiagnosis Really a Problem in Rheumatology?

Vivian P. Bykerk, MD, associate attending rheumatologist at the Hospital for Special Surgery and Director of the Inflammatory Arthritis Center of Excellence, New York City, agreed with some but not all of Landewé's critique. In particular, Bykerk is not convinced that overdiagnosis occurs very often.

"In 25 years of treating early [rheumatoid arthritis (RA)] patients, I have observed that they usually present with very obvious RA (they have waited, their primary care physician waited to refer, etc). I can think of only a few that might have come too early," she told Medscape Medical News.

According to Bykerk, data from the IMPROVE study showed that only 4% of patients with early RA could permanently stop disease-modifying antirheumatic drugs (DMARDs). "I am speaking from a North American perspective, where waiting time to see a rheumatologist is 3 to 6 months. If an individual had a self-resolving synovitis, it would have resolved," Bykerk added.

Parallels With Oncology

The current situation in diagnosis and treatment of inflammatory rheumatic diseases such as RA, psoriatic arthritis (PsA), and ankylosing spondylitis (AS) roughly resembles certain developments in oncology care during the past decade. These include an initial push for very early detection, recognition that this can lead to unnecessary treatment of cancers that would have had little effect on the patient's life, and a shift from trying to achieve complete remission and remove all signs of the cancer to recognition that for many patients, achieving durable stable disease is nearly as good as cure, with generally lower rates of adverse effects. A key parallel for rheumatologists is to consider whether aiming for remission in the patient who already has durable low disease activity is worth the cost and potential hazards of additional treatment.

Landewé told Medscape Medical News that his concern about possible overdiagnosis and overtreatment in rheumatology was triggered by data on the harmful effects of prostate cancer screening. "I was intrigued by the fact that despite compelling evidence of the existence of overdiagnosis in cancer screening, health authorities in my country and beyond had decided to ignore these issues, likely due to fear of public beliefs and disdain," he said.

"This made me wonder if overdiagnosis (followed by overtreatment) could also play a role in modern rheumatology, especially because we pay so much attention to early diagnosis, intensive treatment, and treat-to-target."

The discussion of six popular treatment paradigms rests mainly on Landewé's expert opinion, in the absence of scientific data, and the analysis identifies areas where additional studies are needed.

Paradigm 1: Early Diagnosis

Landewé suggests that initiatives to encourage earlier diagnosis (such as revised diagnostic criteria) might be a double-edged sword. "Bad cases may have reached the office of the rheumatologist timelier, but in their wake, many milder cases with less clear symptoms and a better prognosis will have shown up too," he warns. This could create more "grey-zone medicine" for patients with milder disease subjected to more intensive treatments.

The push for early diagnosis has been succeeded by attempts to identify patients likely to develop true inflammatory disease, who then become candidates for preemptive treatment. "This is befuddling risk of disease with presence of disease, which may turn healthy people into patients and can cause collateral damage," Landewé says.

Advice for the Clinician on Early Diagnosis

Landewé told Medscape Medical News that clinicians should "be parsimonious" when considering a diagnosis that would have important treatment implications.

"The first step in the chain of 'unnecessary healthcare' is a careless diagnosis and prognosis," Landewé said. "After that diagnosis, any treatment, including an expensive biological treatment, becomes far easier to justify (psychologically). There is a lot of pressure nowadays (by professional organizations, guidelines, big pharma, patients, etc) to 'not waste time' and considering 'arthritis as an emergency,' " Landewé explained.

"But we are not talking about acute myocardial infarction or pneumonia. We do have some time to wait and see what happens. A 3-month delay followed by the right diagnosis/prognosis and the most appropriate treatment (sequence) is to be preferred over a quick but careless diagnosis of RA, PsA, or AS in a patient who had self-limiting arthritis but will likely be treated with [conventional synthetic/biologic] (cs/b) DMARDs for years, if not the rest of her life."

Paradigm 2: Intensive Treatment

Early diagnosis and intensive treatment have substantially improved outcomes in RA, PsA, and AS, along with the availability of more effective drugs and treatment strategies. However, Landewé warns that the outcomes in clinical trials are average group results, and that apparent clinical superiority may be a result of outcomes in a minority of patients, whereas the majority might have done just as well on usual therapy. "To date, we still extrapolate trial results obtained in severe patients to those with milder disease, convinced that we do a good job, but we likely overtreat some of them," he writes.

Paradigm 3: Remission

The paradigm that clinical remission is the best treatment goal is based on evidence that clinical disease activity leads to progression of structural damage. Landewé suggests that aiming for clinical remission may be appropriate for patients with severe disease and high disease activity, but that "pushing toward remission" for the patient with low disease activity might do more harm than good. He notes that the question of whether the difference between zero and minimal structural progression adds value from the patient's standpoint has not been definitively answered. "What we have learnt, though, is that drug use in patients who were intensively monitored and steered towards remission was markedly higher than in patients treated regularly — a subtle sign of overtreatment," Landewé writes.

Paradigm 4: Prognosis and Risk Stratification

Most of the potential imaging tests and biomarkers studied for risk stratification in rheumatic diseases have not been reproducible or have been only marginally more accurate than simple clinical measurements, according to Landewé, and researchers, clinicians, and the public often greatly overestimate the value of prognostic (risk) information.

Landewé told Medscape Medical News, "The main concerns are that our prediction rules are fallible: There is a lot of misclassification (false-positive cases). Data proving that treating patients with a certain risk of RA (so including the false positives) is better than waiting until they have actually developed RA are lacking. That is only a hypothesis with a lot of room for conflict of interest (researchers know that funding bodies are sympathetic to this hypothesis) and immense potential for unnecessary healthcare. And think about the psychological consequences: healthy people face a risk of getting a chronic disease that urges them to lifelong treatment. Given the lack of insight of most people in proper risk valuation, this is indeed making healthy people sick."

Paradigm 5: Evidence-Based Rheumatology

Evidence-based guidelines based on systematic literature review and expert judgement can help inform treatment decisions for individual patients. However, because they describe best practice approaches irrespective of the local situation, rigid adherence to guidelines without consideration for the individual patients may also contribute to overdiagnosis and overtreatment, Landewé says.

Paradigm 6: Precision Medicine

"Precision medicine" refers to the idea that patients can be separated into subgroups based on genetic, cellular, or molecular markers and that those markers can be used to improve treatment of individual patients. Landewé dismisses precision medicine as "a current hype" that is appealing to the public and to funding bodies, but is "unlikely to become the holy grail of rheumatology" as a result of the complex interplay of factors involved in rheumatic disease pathways.

Bykerk disagrees.

"There may be opportunities to find new treatment targets and/or repurpose older drugs using this approach. I do agree that any "biomarker" better have substantially improved predictive capacity over clinical measures (as we treat the patient not the test). Some tests, however, may guide which therapy could be of benefit over another," she said.

"Treat the Patient, Not the Test"

"Patients don't want to take medicine they don't have to, so if they feel their disease is well controlled (normal function, normal strength, no fatigue, no stiffness, and most importantly, no pain), they will not usually agree to step-up therapy, even if they have a couple of swollen tender joints ([proximal interphalangeal joints])," Bykerk told Medscape Medical News."

"We have tried to recruit patients at risk for RA (30% chance over 3 years), and have yet to find one who would agree to such a study. If they aren't in pain, they are less likely to agree to medication with a potential or experienced side effect list that is worse than their perceived problem. I have always said, 'Treat the patient, not the test,' so ultrasound-only synovitis may not be of sufficient rationale."

Important Questions Are Likely to Remain Unanswered

Landewé writes, "We may have to acknowledge in rheumatology that 'earlier' (diagnosis) is not always necessarily 'better', that 'more' (treatment) is not automatically 'more effective', and that 'increased sensitivity' (of diagnostic tests) is not synonymous to 'better yield'. We need to better distinguish 'disease' from the 'risk of disease'. We need trials investigating how to best manage all those milder cases that come to us as a result of our early diagnosis initiatives. We need to learn what does and does not really matter to patients with inflammatory rheumatic diseases. We need to pay more attention to possible overdiagnosis and overtreatment in our clinical research."

Landewé outlined the clinical trials he would like to see: "These trials should be explanatory (pragmatic) and test a strategy with cheap and widely available csDMARDs (and/or combinations) compared to a strategy with the presumably best possible standard treatment (likely [methotrexate] MTX plus a bDMARD)," he told Medscape Medical News.

"They should not have the usual inclusion criteria of high disease activity, but should allow every patient with a diagnosis of RA (PsA) in whom a DMARD is considered, except for patients with highly active disease, since it is not ethical to treat them with presumably less effective drugs."

"Such trials should test a noninferiority hypothesis with a noninferiority margin based on consensus of all stakeholders. Outcomes should include 1- and 2-year radiological progression...to not overlook unacceptable radiographic progression despite good clinical outcomes. Costs, treatment retention, and adverse events are important outcomes of such a trial. Blinding is not necessary, but the duration should be long enough (2 years?) to support proper conclusions," he explained.

However, Landewé doubts that such a trial will ever be performed, as it would be very expensive. "It is not in the interest of pharmaceutical industry, and public funding will fall short," he predicted.

Landewé and Bykerk have disclosed no relevant financial relationships.

Ann Rheum Dis. Published online July 4, 2018. Abstract

For more news, join us on Facebook and Twitter


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.
Post as: