The Risk of Cardiovascular Events in Patients with RA & More Explored in 3rd Plenary Session

SAN DIEGO—The ACR Convergence 2023 Plenary Sessions highlight select abstracts of interest to the general ACR audience. In one presentation at this year’s Plenary Session 3, Beth Wallace, MD, MSc, a staff physician at the VA Ann Arbor Healthcare System, Michigan, and an assistant professor at the University of Michigan, Ann Arbor, shared important data on the relationship between time-dependent cumulative glucocorticoid exposure and major adverse cardiovascular events (MACE) in a cohort of veterans with rheumatoid arthritis (RA).

Background

Dr. Wallace commenced her talk with an unfortunate truth: “Up to 50% of RA patients in the United States use long-term glucocorticoids despite previous work suggesting they increase the risk of MACE in a dose-dependent way.”¹

She and her colleagues previously presented work demonstrating that fewer than 14 days of glucocorticoid use in a six-month period is associated with a two-thirds increase in the odds of MACE over the following  six months, and 90 days of use was associated with over a two-fold increase. However, studies to date have been limited by the fact that no adjustments were made for the timing of prior glucocorticoid use or dose per unit of time, both of which matter when discussing glucocorticoid risk profile.

Dr. Wallace

Study Design

Wallace et al. conducted a retrospective cohort study using national data from the Veterans Affairs (VA) Corporate Data Warehouse. The researchers included patients with RA who were between 40 and 90 years old and had an initial rheumatology visit between 2010 and 2018. They excluded patients with other rheumatic disorders and prior MACE or equivalent. The primary outcome was time to first MACE, defined as an acute myocardial infarction, stroke or transient ischemic attack, cardiac arrest, coronary revascularization or death from a cardiovascular (CV) cause.

They used pharmacy dispensing data to calculate a weighted cumulative dose (WCD) of oral glucocorticoids (see below), and medical claims data to identify incident MACE.² They adjusted the model for many baseline (i.e., age, sex, race, BMI, Elixhauser index) and time-varying covariates updated every six months (i.e., tobacco use, VARS-CVD, malignancy claim, lipid-lowering pharmacy claim, opioid claim, methotrexate use, biologic use, hydroxychloroquine use, claim for hospitalized infection, number of rheumatology clinic visits in the past six months), including the VA Risk Score–Cardiovascular Disease (VARS-CVD). The VARS-CVD is an estimate of five-year MACE risk specific to the veteran population that’s generated from multiple variables.³ It’s similar to the Atherosclerotic Cardiovascular Disease (ASCVD) Risk Estimator from the American Heart Association/American College of Cardiology.⁴ 

The Weighted Cumulative Dose Model

Dr. Wallace took time to explain exactly why they chose the WCD model to determine glucocorticoid exposure. “Most studies that evaluate glucocorticoid use over time take complex raw data and convert it to a simplified summary measure that can be fixed in time or time-varying,” she said. “But none account for the fact that use, dose and duration are all likely to affect real-world risk, and past exposure might affect risk differently than ongoing exposure.”

The WCD was calculated by taking each day’s actual glucocorticoid exposure and multiplying it by an empirically derived, weight-based coefficient, with more recent doses assigned higher weights. Then, the weighted doses were summed up over a two-year period. “The WCD model allows us to generate estimates that incorporate use, dose and duration of glucocorticoid exposure, and evaluate both remote and recent glucocorticoid exposure and how they contribute to the outcome of interest,” said Dr. Wallace.

Results

Wallace et al. analyzed almost 19,000 patients, the majority of whom were men in their 60s. About two-thirds of these patients received glucocorticoids during the study period, and about one in six had at least one six-month interval during which they received glucocorticoids for more than 90 days. Of note, this is the definition of long-term glucocorticoid use, according to ACR clinical practice guidelines.

“About 4% of patients overall had a MACE event, with a median time-to-event of three years,” said Dr. Wallace. “Although VARS-CVD risks were similar across all groups, the patients who received more than 1.2 mg of glucocorticoids per day had twice the risk of incident MACE than those who did not.”

Next, Dr. Wallace described the data for patients actively using glucocorticoids at the time of assessment and those who had stopped using glucocorticoids one year earlier. Relative to those without prednisone use, patients actively using 5 mg of prednisone for 30 and 90 days had a 5% and 10% increased hazard of MACE, respectively. Interestingly, patients who stopped using glucocorticoids one year earlier had similar increases in hazard of MACE: 3% for 30 days of use and 9% for 90 days of use.

Limitations

Dr. Wallace took care to highlight some limitations of their work. For example, claims data always come with the risk of residual confounding.  Because they didn’t use a causal inference model, the analysis may not adequately capture risk factors that are both confounders and part of the causal pathway for MACE (e.g., blood sugar and blood pressure). She also noted, “Some of the results presented aren’t statistically significant, although they approach statistical significance. We think this is, in part, due to inadequate bootstrapping of the sample, which we are working to correct.”

Conclusion

“Long-term glucocorticoid use is common in RA, and ongoing use is associated with a dose and duration-dependent increase in MACE risk, even at doses as [low] as 5 mg and durations as short as 15 to 30 days. And prior glucocorticoid use, even up to a year before risk assessment, may remain associated with increased risk of MACE,” concluded Dr. Wallace.

Moving forward, Dr. Wallace and her colleagues will be doing additional analyses to confirm results.  They are investigating whether associations persist after more robust adjustment for disease activity score and remote glucocorticoid exposure, and whether there’s an interaction between background cardiovascular risk and glucocorticoid exposure. We can look forward to more data in this regard.

Samantha C. Shapiro, MD, is the executive editor of Harrison’s Principles of Internal Medicine. As a clinician educator, she practices telerheumatology and writes for both medical and lay audiences.

References

1. Wallace B, Gao Y, Kim H, et al. Time-dependent evaluation of glucocorticoid exposure duration and major adverse cardiovascular events in a cohort of veterans with rheumatoid arthritis [abstract]. Arthritis Rheumatol. 2022;74(suppl 9).
2. Sylvestre MP, Abrahamowicz M. Flexible modeling of the cumulative effects of time-dependent exposures on the hazard. Stat Med. 2009;28(27).
3. Sussman J, Wiitala WL, Zawitowski M, et al. The Veterans Affairs cardiac risk score: Recalibrating the ASCVD score for applied use. J Gen Intern Med. 2016;(1).
4. Goff DC Jr., Lloyd-Jones DM, Bennett G, et al.; American College of Cardiology/American Heart Association Task Force on Practice Guidelines. 2013 ACC/AHA guideline on the assessment of cardiovascular risk: A report of the American College of Cardiology/American Heart Association Task Force on Practice Guidelines. Circulation. 2014 Jun 24;129(25 Suppl 2):S49–S73. Erratum in: Circulation. 2014 Jun 24;129(25 Suppl 2):S74–S75.