Digital Tool May Help Neurologists Assess Steroid Toxicity

SAVANNAH, Ga. — A digital tool to help neurologists assess steroid toxicity in patients with myasthenia gravis (MG) demonstrated sensitivity in distinguishing between different doses and durations of steroid exposure in a retrospective, real-world study.

The Glucocorticoid Toxicity Index-Metabolic Domains (GTI-MD), an abbreviated version of the GTI (Steritas), used weighted, standardized clinical outcome assessments to calculate steroid toxicity using a de-identified electronic health record (EHR) dataset.

“The results of our study indicate that patients with MG who initiated steroids demonstrated evidence of steroid toxicity in as little as 90 days after initial exposure, which was significant for patients with 20+ mg at index with repeated use,” study investigators, led by Neelam Goyal, MD, clinical professor of neurology and neurological sciences at Stanford University School of Medicine in Palo Alto, California, noted.

The findings were presented on October 15 at American Association of Neuromuscular & Electrodiagnostic Medicine (AANEM) 2024.

Rapid Evidence of Toxicity

The GTI uses nine health domains to calculate steroid toxicity scores, and the GTI-MD, which has been shown to be closely correlated, uses four domains collected routinely in clinical practice: Body mass index (BMI), blood pressure, glucose tolerance, and lipid metabolism.

The study used the Optum EHR dataset to identify 682 adult patients with MG, mean age of 70 years, 38% women, with at least two confirmed diagnoses of MG between 30 and 730 days apart and information on steroid utilization.

Patients were divided into two groups: Steroid initiators (SI; n = 377) were those whose steroid use was already in progress at the index date, whereas steroid-naive (SN) patients (n = 305) began their steroid use at the index date. Among the SI group, 30% were on doses > 20 mg/d and 22% were on lower doses. Among the SN group, 22% were on doses > 20 mg/d and 26% were on lower doses.

As expected, mean GTI-MD scores measured 90 days after the index date were higher in the SI group than in the SN group, indicating a higher level of steroid toxicity in the SI group. This was measured with two subscores of the GTI-MD: The Cumulative Worsening Score (22.6 vs 18.7; P = .007) and the Aggregate Improvement Score (4.9 vs 1.9; P = .27), the latter incorporating resolved toxicities resulting from the introduction of steroid-sparing agents.

The authors commented that scores were higher in the SN group than expected, “which could be explained by age, previous steroid exposure, comorbidities, and side effects from other medications.” However, they concluded that the findings suggest utility of the tool retrospectively, with EHR data.

Clinical Application

The GTI and related measurements are proprietary tools and therefore not readily available to all clinicians, noted Marie Beaudin, MD, another neurologist at Stanford University School of Medicine, who was not involved in the research.

In a separate, observational, ongoing study, Beaudin and Goyal’s team are examining the use of the tool prospectively for following the steroid toxicity burden in 50 patients with MG and correlating it with MG outcomes measured using the MG-Activities of Daily Living, MG Composite, and MG-Quality of Life 15R validated scales, as well as the adverse event unit.

“The objective of this study is to quantify the burden of toxicity that our patients are having from glucocorticoids, see how sensitive to change the scale is as their dosage of prednisone changes, and explore the correlations between the score and their disease outcome measures,” Beaudin told Medscape Medical News.

Unlike the abbreviated GTI-MD, the GTI measures nine domains: Bone mineral density, BMI, lipid metabolism, blood pressure, glucose tolerance, myopathy, skin toxicity, neuropsychiatric symptoms, and infections.

The score involves actively prompting and examining the patient, making it quite comprehensive. Beaudin said the study has revealed interesting insights into how patients report their side effects. When asked broadly about steroid-related side effects, many patients mention issues like weight or skin issues.

However, she noted, when prompted specifically about symptoms like insomnia, irritability, depression, or cognitive changes, there was an unexpected increase in positive responses, as patients are often unaware these could be side effects. This suggests the study may capture a greater burden than originally anticipated, said Beaudin.

She added that the long-term utility of the GTI score might be to help clinicians predict steroid toxicity and guide management.

“Then we would get more aggressive in trying to wean or taper patients. But these are often complicated cases because as soon as we taper, the disease flares. It’s a difficult decision whether to reduce the dosage of prednisone because toxicity burden is high, when disease burden is high too, and that’s where other medications can come into play.”

For example, she said, for insurance coverage, a high steroid toxicity score could justify the need to initiate more expensive steroid-sparing agents.

Both studies were funded by argenx. Goyal reported that she has consulted and received grant support from argenx, UCB, Alexion, and Janssen argenx. Beaudin is supported by a McLaughlin Scholarship from Laval University, Quebec, Quebec, Canada.