Use of Disease-modifying Antirheumatic Drugs, Biologics, and Corticosteroids in Older Patients With Rheumatoid Arthritis Over 20 Years

DISCUSSION

The current study examines the change in medication utilization derived from prescribing patterns in health administrative data for older adults with RA between 1997 and 2017. During these 20 years of observation, there were substantive changes in the therapeutic approach to RA that included more aggressive use of conventional DMARDs and the introduction of biologics. The increased utilization of disease-modifying therapies is particularly encouraging. The results of our study in an older adult population with universal access to health care indicate that many of the important outcomes from RA clinical trials were translated into clinical practice, as well as implementation of practice guidelines developed by many professional rheumatology organizations. However, the use of CS was largely unchanged over time. Given the risks associated with chronic CS use, especially in older adults, renewed efforts are required to minimize their use in the long-term pharmacological management of RA.

A number of previous studies, cross-sectional^4,5,6,7,8 and longitudinal,^9,10,11 have used health administrative data to evaluate medication utilization in RA at a population level. To our knowledge, the current study in RA has the longest duration to date. Some of the changes in utilization are in line with expectations. For example, the temporal fall in coxib prescribing in both RA and control populations is in keeping with the effect of regulatory changes and prescribing patterns for this drug class. At study onset, the use of DMARDs was low but comparable to the findings in another Canadian jurisdiction by Lacaille, et al,¹ who reported that only 43% of patients with RA received a DMARD at least once over 5 years between 1996 and 2000. In our study over 20 years of observation, there was a progressive increase in DMARD use that had almost doubled by 2017. Once biologics became available to the insured population of the current study, there was also a steady increase in utilization over time. In light of these changes, in particular with the increased use of disease-modifying therapies, it was surprising to find that almost one-third of patients received some exposure to CS on an annual basis and that the prevalent use of CS did not change appreciably over the 20 years of observation. A comparable trend, albeit with slightly lower utilization, was seen when the analysis was confined to prednisone only.

The frequency of CS use in RA is variable and influenced by a number of factors. In a cross-sectional study of 10,262 patients with RA in the United States, 23% were receiving CS.⁶ In another US study,⁴ 10% of 8125 patients with RA received CS without DMARDs and were more likely to be older, had lower incomes, and less likely to have seen a rheumatologist compared to DMARD users. In a UK study of 38,884 patients with RA over 17 years, 47% of patients received oral CS for a median percentage of 26.3% of follow-up time and CS were more commonly prescribed in older patients.¹¹ In a Canadian study of 16,207 patients with RA over 65 years of age, 35% of patients received CS.⁸ Finally, in an incident RA cohort, 63% of 3992 patients received CS within 12 months following diagnosis.¹⁰ The current study confirms and expands these observations. In our prevalent RA cohort, approximately one-third of cases had an annual exposure to CS compared to up to 10% of controls, and the majority of prescribing was by family physicians. Although there was little change in the mean annual dose of prednisone over time, it was encouraging that the duration of exposure in any given year fell consistently over the latter 10 years of observation. For all physician groups, CS were more frequently prescribed close to the time of RA diagnosis, but less frequently by rheumatologists who also curtailed subsequent prescribing over a follow-up of 14 years. A similar trend was seen when the analysis was restricted to prednisone use and the annual dose of prednisone.

CS play an important role in gaining prompt control of acute polyarthritis in patients with RA. However, there are concerns surrounding their long-term use, even in low doses. Although clinical trials of up to 2 years in duration have suggested less radiographic joint destruction in patients with RA treated concurrently with low-dose CS,^12,13,14 this has not been replicated in a more recent observation study of 5 years’ duration.¹⁵ Further, as reflected in recent European League Against Rheumatism guidelines for management of RA,¹⁶ in the era of synthetic and biologic DMARDs, the potential benefit of long-term CS, even in low doses, is now challenged by expert opinion. The reasons responsible for the persistent use of CS in our study were not explored and are currently unknown. Possible explanations include reluctance by both patients and physicians to intensify treatment with DMARDs or to commence with biologics due to a fear of drug toxicity or some degree of CS dependency because of adrenal suppression from prolonged use.

What is perhaps less controversial is evidence of toxicity associated with long-term CS in patients with RA. This is particularly persuasive when examining the association with infection. Severe infections,¹⁷ some requiring hospitalization,^18,19 nonserious infections,⁸ and infection risk above that seen with methotrexate²⁰ are all associated with CS use. The frequency of use, even in low doses, is also an important factor. For example, Dixon, et al¹⁷ found that the use of 5 mg prednisolone for 3 months, 6 months, and 3 years was associated with a 30%, 46%, and 100% increased risk of serious infection compared to patients with RA who did not use CS. Concern about using CS is also shared by patients with RA, with 68% of 158 consecutive ambulatory patients with RA seen in 1 center declining to take them.²¹ Notwithstanding the possibility of confounding by indication due to disease activity and severity,²¹ there is a compelling need to minimize and preferably eliminate the use of low-dose CS in the long-term management of RA.

What steps can be taken to lower long-term CS exposure in patients with RA? As acknowledged by previous²² and more current¹⁶ treatment guidelines for RA, CS can play an important role in the management of patients with RA. Prompt control of active disease at presentation or with subsequent flares provides a bridging strategy to allow time for conventional or biologic DMARDs to take effect. Rapid tapering of CS is universally recommended, with total exposure of less than 3 months in the majority of cases.¹⁶ The use of long-term CS, even in doses of prednisone 5 mg/day, are not recommended.¹⁶ In the Canadian healthcare system, and in many others, the role of primary care physicians (PCP) is central to the treatment of patients with RA at presentation, during flares, and in the long term. The current study indicates that not only are the majority of CS for older patients with RA prescribed by PCPs but in contrast to rheumatologists, the dose is usually higher and is tapered less aggressively over time. This difference in practice patterns could be addressed though enhanced medical education for PCPs, more communication between rheumatologists and PCPs, and improved access to rheumatologists. The proportion of patients with RA who were seen by a rheumatologist in our study (48.8%) is similar to that reported in another Canadian population health study (48%).¹ In both studies, the remaining patients were seen by general internists and PCPs. Renewed efforts to educate patients on the risks associated with long-term use of CS may also be required because for some patients, the symptomatic benefit of low-dose CS may outweigh what appears to be a low risk of toxicity. Finally, in some situations, improved and timelier access to biologic and synthetic DMARDs may be beneficial, especially in view of the prompt onset of action of these agents that may diminish the need for bridging therapy with CS.

Strengths of the current study include prior validation of the case definition for RA against a clinical dataset,² a prolonged duration of observation, and the fact that the Nova Scotia population is stable with a mix of urban and rural communities and a range of socioeconomic groups representative of a general Canadian population. There are also some limitations to the study. First, due to the homogenous nature of the Nova Scotia population, it was not possible to examine the effect of race/ethnicity on medication utilization. Second, although the Canada Health Act provides universal coverage for inpatient health services, public payment for medication outside of hospital is restricted predominantly to individuals 65 years of age and older. The health administrative data for the current study were limited to this older adult population, many of whom likely had well-established RA. This was addressed in part by the identification of an incident RA cohort in whom the pattern of CS utilization was comparable to that seen in the prevalent RA cohort. Although the definition of incident cases was in agreement with traditional methodology in population health studies,¹ it would not have excluded patients with RA with longstanding disease who relocated to Nova Scotia during the period of study. Very likely, this represented a minority of RA cases. Third, the cases were not stratified for disease activity or severity, which may have influenced medication utilization patterns. Finally, the possibility of bias due to misclassification of CS exposure needs to be considered. In a UK study²³ that compared the “true” CS utilization determined by patient report with prescription data, ascertainment agreement was high with 86% accuracy but was still considered large enough to lead to important misclassification bias.

Despite these limitations, the current study provides a high-level overview of changes in medication utilization over a 20-year period during which the pharmacological approach to the treatment of RA underwent several paradigm shifts. Many of the transformative outcomes of clinical trials in RA and RA treatment guidelines from rheumatology organizations are reflected in this population health study. Despite the many advances in disease-modifying therapies, an important finding of the current study is the durable utilization of CS that needs to be addressed. The findings presented here provide a benchmark to determine the effect of future efforts to reduce the long-term use of low-dose CS in older adults with RA.