Introduction
The last few decades have seen major therapeutic advancements in rheumatoid arthritis (RA) therapeutics. New disease-modifying antirheumatic drugs (DMARDs) have created more choices for people, but no therapeutic works for all patients. There has been a focus on personalized medicine initiatives that tailor therapeutic decisions to patients based on their unique characteristics or biomarkers. However, trade-offs still need to be made between treatments. This paper reviews the history of RA therapeutics and progress toward personalized risk predictive models for DMARDs, outlining where knowledge gaps still exist. It also discusses why patient preferences play a key role in a holistic view of personalized medicine and how this links with shared decision-making.
The evolution of rheumatoid arthritis therapeutics
Scientific investigation and medical advancements have allowed us to evolve past the ineffective rheumatoid arthritis (RA) therapy models that marked the early 20th century. A timeline summarizing the evolution of RA therapeutics is presented in Fig. 1.
Novel and unapproved therapies
Research is underway in novel RA therapeutic, with several notable drugs being ozoralizumab, obefazimod, and ianalumab. Ozoralizumaab (ATN-103) is a trivalent compound with two anti-TNF- antibodies and one antialbumin antibody that has exhibited efficacy in animal models and holds promise for human trials. Obefazimmod (ABX464) causes the upregulation of anti-inflammatory microRNA (miR-124), which has been implicated in several anti-inflammatory pathways, including the downregulation of TNF-α. Ianalumab, an anti-B cell monoclonal antibody, is currently being investigated in a phase 1 trial. Available RA therapeutics, while effective, can have unpleasant side effects. Research is looking at drug delivery mechanisms that can reduce unwanted side effects, such as extracellular vesicles released from mesenchymal stem cells (MSC). An MSC-based delivery system, a microdevice is developed using nanoparticles, human MSCs, and an anti-inflammatory drug, has shown the promising potential of such delivery systems. Several RA therapeutics were thoroughly investigated but ultimately shown to be ineffective for treating RA. Promising research suggested granulocyte colony-stimulating factor (GCF) as an alternative delivery system.
The need for personalized predictive models
Infliximab was the first anti-TNF therapeutic approved for use in RA, revolutionizing the way RA was treated and inspiring further advancements, but it has not had as much impact as prednisone, MTX, and infliximab. Systematic reviews have not found convincing differences in terms of benefits between biological or targeted synthetic therapeutic options in terms of treatment benefits. Personalized medicine is the ability to predict an individual patient’s risk or benefit from a treatment to allow more tailored therapeutic decisions.
Clinical prediction modeling for RA
Studies have shown that RA treatment benefits and risks vary by age, sex/gender, baseline disease activity, functional ability, seropositivity, smoking status, and other sociodemographic variables. A 2018 systematic review by Archer et al. identified 35 prognostic factors and 3 predictive factors, but noted that there were uncertainties around the significance of these factors. There were limitations identified in the available models, such as flawed variable selection techniques, inadequate internal validation, and the lack of external validation.
Modeling to distinguish between treatments
Prognostic factors can help guide the overall therapeutic strategy, but can be less helpful for choosing a specific treatment. To generate estimates that can help distinguish drugs in terms of their treatment benefits, we need to explore predictive factors that may serve as effect modifiers. Randomized controlled trials (RCTs) allow the estimation of effect modifiers, and several genetic and biological markers have been linked to treatment response in RA. Risk-specific biomarkers have been identified, such as mutations of the p53 gene and overexpression of dihydrofolate reductase (DHFR) and FPGS genes. Biomarker identification has the potential to guide therapeutic decisions to maximize treatment efficacy and limit risk on a personalized level.
Precision medicine RCTs
RCTs can be used to generate drug-specific predictive models by stratifying participants by their biomarker profile, genotype, or other biological characteristics. A recent randomized trial by Humby et al. found that patients with low CD20 B-cell signatures had a higher response rate to tocilizumab than rituximab. In a follow-up study, Rivellese et al. identified genes and pathways associated with response to each treatment. However, synovial biopsies may be impractical for routine care.
Moving from personalized outcome estimates to personalized treatment
The best predictive models for treatment benefits and harms can, at best, produce estimates of treatment response. In diseases, such as RA, where treatment response is heterogeneous and there is no distinctly superior treatment, predictive models cannot tell us which treatment is “best”. To a patient, personalized medicine provides information about the chance of a future outcome for the treatments available. Patients still need to choose between treatments. As biomarkers only shift our estimates of outcomes, a decision will still need to be made that weighs these outcomes against the other pros and cons. Understanding patients’ unique treatment-related preferences can facilitate informed decision-making between available treatments.
Patient preferences
A systematic review summarized patients’ preferences for RA therapeutics, highlighting the importance of treatment benefits. However, preference variability does exist, such as preference related to treatment cost or route of drug administration. A discrete choice experiment identified a subgroup of risk-averse patients among a broader group of patients who were more risk tolerant. To personalize RA therapy, decisions about risk and benefit trade-offs need to be made in the context of the patient’s unique preferences. For example, a patient who has failed first-line therapy with conventional synthetic DMARDs may have a choice between the JAK inhibitor tofacitinib or a similarly effective anti-TNF. However, recent concerns regarding elevated cancer and cardiovascular event rates in individuals treated with JAK inhibitors may impact decision-making. A fundamental principle of shared decision-making is that patients have inherent values and preferences that should guide treatment decisions. Preference misdiagnosis occurs when a clinician incorrectly assesses a patient’s treatment-related values and preferences, but recognition and corrective action is rare. Patients who have experienced a preference misdiagnosis may be at risk for nonadherence to their medication, which can be highly detrimental in RA, where uncontrolled disease can lead to irreversible bone erosion. The trade-off between risks and benefits is a highly personalized assessment guided by each patient’s unique values.
Conclusion
Medical advancements have provided a comprehensive array of RA treatments, but which treatment is best suited to each patient is still unclear. Patients and clinicians would benefit from a personalized approach to RA treatment, which involves developing and tailoring medication decisions to patients based on their unique characteristics or biomarkers. Researchers should utilize all available data to develop the best available treatment effect estimates for patients. Personalized medicine can expand beyond treatment effect estimates and incorporate patient preferences, particularly in cases where some aspects of medication are not acceptable for some patients. It is most optimized when it takes place in a shared decision-making context where the clinician and patient work together to develop a treatment plan.
References:
1. G.G. Hunder, E.L. Matteson Rheumatology practice at Mayo clinic: the first 40 years-1920 to 1960 Mayo Clin Proc, 85 (4) (2010), pp. e17-e30.
2. B.I. Comroe The use and abuse of gold therapy in rheumatoid arthritis J Am Med Assoc, 128 (12) (1945), pp. 848-851.
3. T.G. Benedek History of the development of corticosteroid therapy Clin Exp Rheumatol, 29 (5 Suppl 68) (2011), pp. S5-S12.
4. L.E. Ward, H.F. Polley, M.H. Power, et al. Prednisone in rheumatoid arthritis: metabolic and clinical effects Ann Rheum Dis, 17 (2) (1958), pp. 145-159.
5. T. Hedner, B. Everts The early clinical history of salicylates in rheumatology and pain Clin Rheumatol, 17 (1) (1998), pp. 17-25.
6. T.M. Chalmers Clinical experience with Ibuprofen in the treatment of rheumatoid arthritis Ann Rheum Dis, 28 (5) (1969), pp. 513-517.
7. B. McConkey istory of the development of Sulphasalazine1 in rheumatology Drugs, 32 (Supplement 1) (1986), pp. 12-17.
8. E. Schrezenmeier, T. Dörner Mechanisms of action of hydroxychloroquine and chloroquine: implications for rheumatology.
9. S. Shams, J.M. Martinez, J.R.D. Dawson, et al. The therapeutic Landscape of rheumatoid arthritis: current state and future Directions Front Pharmacol (2021), p. 12.
10. R. Gubner, S. August, V. Ginsberg Therapeutic suppression of tissue reactivity. II. Effect of aminopterin in rheumatoid arthritis and psoriasis Am J Med Sci, 221 (2) (1951), pp. 176-182.
11. B.N. Cronstein, T.M. Aune Methotrexate and its mechanisms of action in inflammatory arthritis Nat Rev Rheumatol, 16 (3) (2020), pp. 145-154.
12. J.A.M. Wessels, T.W.J. Huizinga, H.-J. Guchelaar Recent insights in the pharmacological actions of methotrexate in the treatment of rheumatoid arthritis Rheumatology, 47 (3) (2007), pp. 249-255.
13. B. Friedman, B. Cronstein Methotrexate mechanism in treatment of rheumatoid arthritis Joint Bone Spine, 86 (3) (2019), pp. 301-307.
14. M. Cutolo, A. Sulli, C. Pizzorni, et al. Anti-inflammatory mechanisms of methotrexate in rheumatoid arthritis Ann Rheum Dis, 60 (8) (2001), pp. 729-735.
15. E.S.L. Chan, B.N. Cronstein Molecular action of methotrexate in inflammatory diseases Arthritis Res Ther, 4 (4) (2002), p. 266.
16. M. Cutolo, A. Sulli, C. Pizzorni, et al. Anti-inflammatory mechanisms of methotrexate in rheumatoid arthritis Ann Rheum Dis, 60 (8) (2001), p. 729.
17. P. Jones Reappraisal of the clinical use of leflunomide in rheumatoid arthritis and psoriatic arthritis Open Access Rheumatol Res Rev, 53 (2010).
