Children with lupus carry a heavy steroid-toxicity burden that cannot be captured by measuring dose. A real-world study from Boston Children's Hospital and the Children's Hospital of Philadelphia confirms the pediatric Glucocorticoid Toxicity Index (pGTI) is ready for clinical practice and trial design.
Steroid-sparing strategies in pediatric rheumatology have long rested on an uncomfortable and inaccurate assertion: that reducing dose is equivalent to reducing harm. A recent study published in Seminars in Arthritis and Rheumatism by Emily Zhang, Joyce C. Chang[1], and colleagues dismantles that assumption with data from 126 children with pediatric-onset systemic lupus erythematosus (pSLE) followed across two leading academic medical centers over more than two decades.
As the first systematic application of the pGTI in a real-world clinical setting, the study carries direct implications for drug developers pursuing steroid-sparing programs in pediatric autoimmune disease. The findings reframe the central question: moving from steroid-sparing, measured by dose reduction, to steroid-toxicity sparing, measured by the actual toxicity patients experience.
Children treated with steroids face risks that extend well beyond the known adult toxicity profile. Growth delay and disruption of normal psychosocial development constitute irreversible, time-sensitive harms. Yet until the pGTI was developed, no standardized, validated instrument existed to capture these pediatric-specific consequences alongside the metabolic and cardiovascular toxicities shared with adults.
The Zhang et al. study applied a modified pGTI across seven health domains, scoring change in toxicity at six-month intervals for each patient. The Cumulative Worsening Score (CWS), a continuous score of new or worsening toxicity, provides a longitudinal picture of accumulated harm that dose-reduction metrics cannot deliver.
The findings reveal a steroid-toxicity burden that is both substantial and clinically meaningful across multiple domains:
Critically, the correlation between cumulative steroid dose and cumulative toxicity as measured by the pGTI was modest (rho 0.30). This directly supports a core principle of the pGTI framework: dose is a weak proxy for individual toxicity burden. Younger age, elevated BMI at the time of steroid initiation, and the need for adjunctive rituximab were all independently associated with greater cumulative toxicity. Baseline steroid history does not reliably predict which patients will sustain the greatest harm.
One of the study's most important findings for drug developers is the persistence of toxicity accrual well beyond the initial treatment period. While the largest average increase in cumulative toxicity occurred within the first six months, 35% of patients developed new or worsening toxicity between 12 and 24 months, and 44% between 12 and 36 months.
This has a direct implication for trial design: steroid-sparing endpoints that measure early dose reduction may miss the window in which toxicity continues to accumulate. Longitudinal measurement using the pGTI captures ongoing harm that cross-sectional dose metrics will never detect.
The study also addressed a question that has been open since the pGTI was first described: "Is retrospective scoring feasible in real-world multi-center data?" The answer from Zhang et al. is yes, with qualifications that are useful for planning future applications.
Variability in blood pressure measurement in routine clinical practice, inconsistent documentation of mood disturbance, and the exclusion of three domains due to infrequent clinical assessment of relevant laboratory and imaging data all point to specific areas where standardized data collection protocols would strengthen future prospective applications. The authors note that EHR-enabled structured fields and routine mental health screening would materially improve pGTI scoring, and they support the instrument's use across pediatric conditions beyond lupus, including malignancy, asthma, and other autoimmune kidney diseases.
The implications of this study for pharmaceutical development programs are concrete. Pediatric populations with chronic inflammatory conditions represent an area of significant unmet need in which steroid-toxicity has been accepted rather than measured. The pGTI provides the validated, pediatric-specific instrument to change that.
For drug developers, the pGTI now has real-world construct validity established in pSLE, paving the way for its use as a primary or secondary endpoint in trials evaluating steroid-sparing biologics, targeted therapies, and combination regimens in pediatric autoimmune disease. For HEOR teams, the pGTI provides a validated score that can be connected to the costs and complications that drive healthcare utilization in a population that will carry the consequences of inadequate steroid stewardship for decades.
As Zhang et al. conclude, reducing steroid-related morbidity in children with chronic inflammatory conditions is of paramount importance with lifelong implications. The standardized, validated tools to pursue that objective are now available to the benefit of young patients.
Zhang E, Capponi S, Scobell R, Alonzi G, Hlobik M, Daga A, Meidan E, Wobma H, Kim L, Henderson LA, Case S, Nigrovic PA, Stone JH, Costenbader KH, Son MBF, Chang JC. Real-world application of the pediatric Glucocorticoid Toxicity Index in childhood-onset lupus. Seminars in Arthritis and Rheumatism. 2024;68:152516. https://doi.org/10.1016/j.semarthrit.2024.152516