Glucocorticoid Exposure and Infection in Children and Adults With Glomerular Disease: Findings From the Cure Glomerulonephropathy Study

To the Editor:

Infection-related acute care events contribute to increased morbidity and health care utilization among patients with glomerular disease (GD),¹ and are a major concern for patients, caregivers, and health professionals.² Novel therapeutic agents have reduced or eliminated glucocorticoids (GCs) from some treatment regimens for GD; however, GCs are still used frequently for remission induction or maintenance.³ Our previous analysis of infection risk in patients with GD found a 1.3 to 3 times higher risk of first infection in patients with GC exposure, but had limited clinical application owing to the absence of GC dosing information.^1,4

The aim of the present study was to describe longitudinal GC dose exposures among children and adults with biopsy-proven primary GD enrolled in the Cure Glomerulonephropathy (CureGN) Study⁵ and to estimate infection risk associated with low-, medium-, and high-dose GC exposures. Marginal structural models were used to account for time-dependent confounding by disease activity⁶ (Item S1). Models were adjusted for baseline characteristics at enrollment (age, time from kidney biopsy, sex, race, GD subtype, estimated glomerular filtration rate, and comorbid conditions) and weighted to account for time-varying markers of disease activity (urinary protein-creatinine ratio and serum albumin).

The cohort included 2,568 participants (43% female; 35% <18 years old; 582 [23%] with minimal change disease, 650 [25%] focal segmental glomerulosclerosis, 578 [23%] membranous nephropathy, and 758 [30%] immunoglobulin A nephropathy/vasculitis) (Tables S1 and S2). A total of 446 (17%) experienced at least 1 infection-related acute care event over a median follow-up of 58 months (interquartile range [IQR], 32-77). The median prescribed daily oral prednisone dose-equivalent was 23 mg (IQR, 13-36) and 28 mg (IQR, 15-40) in children and adults, respectively. Among those who received intravenous GCs, the median intravenous methylprednisolone dose-equivalent was 60 mg (IQR, 32-250) and 125 mg (IQR, 100-125), respectively (Table S3 and Fig S1).

In the multivariable marginal structural model with “no immunosuppression” as the reference group, low-, medium-, and high-dose GC exposures (inclusive of concurrent immunosuppression) were associated with a 2.26 (95% CI, 1.23-4.15), 2.70 (95% CI, 1.58-4.62), and 2.98 (95% CI, 1.65-5.36) times higher hazard of first infection, respectively (Figs 1 and S2, and Table S4). A model restricted to only hospitalized infections yielded similar effect estimates. Among those <18 years of age, the hazard ratios of first infection were 3.48 (95% CI, 1.42-8.56), 3.67 (95% CI, 1.65-8.16), and 4.83 (95% CI, 2.35-9.94) for low-, medium-, and high-dose GC exposure (Table S4). Similar hazards of first infection were observed in sensitivity analyses when assumptions related to GC tapering and overlapping GC exposure dates and doses were varied to minimize overall GC exposure estimates (Table S5 and Fig S3).

Figure 1.

Hazard of first infection-related acute care event among all participants (fully adjusted marginal structural model; adjusted for age, time from biopsy to study enrollment, sex, race, glomerular disease subtype, estimated glomerular filtration rate, comorbid conditions, and time-varying markers of disease activity [urinary protein-creatinine ratio and serum albumin]). GC, glucocorticoid; IS, immunosuppression. Low dose: ≤7.5 mg/d if ≥30 kg or <0.2 mg/kg/d if <30 kg. Medium dose: >7.5 mg/d and ≤30 mg/d if ≥30 kg or ≥0.2 mg/kg/d and ≤1 mg/kg/d if <30 kg. High dose: >30 mg/d if ≥30 kg or >1 mg/kg/d if <30 kg.

These analyses demonstrate the significant infection risk associated with contemporary GD treatment regimens, particularly among children. This may reflect greater high-dose GC exposure in children compared to adults (4.3% vs 1.8% of total study follow-up time, respectively) and greater exposure to multiple concurrent immunosuppressants in children (9.9% vs 4.8%).

There is conflicting literature regarding the safety profile of low-dose GCs.⁷ We found that low doses of GC were associated with an increased hazard of infection. This finding is consistent with observations in other patient populations, though there is variability in reported dosing thresholds for increased infection risk across different study populations, ranging from below 5 mg/d in patients with rheumatoid arthirits⁸ to >30 mg/d in patients with newly diagnosed autoimmune disease.⁹ We hypothesize that in our analysis the concomitant use of other immunosuppressants may be driving the infection risk among the low-dose GC group, as raw incidence rates of infection among those with low-dose GC alone are similar to those without immunosuppression exposure (Fig 2).

Figure 2.

Percent follow-up time and rate of infection-related acute care events per 100 person-years, stratified by glucocorticoid dose exposure level. Low dose: ≤7.5 mg/d if ≥30 kg or <0.2 mg/kg/d if <30 kg. Medium dose: >7.5 mg/d and ≤30 mg/d if ≥30 kg or ≥0.2 mg/kg/d and ≤1 mg/kg/d if <30 kg. High dose: >30 mg/d if ≥30 kg or >1 mg/kg/d if <30 kg. IS, immunosuppression.

A strength of our study includes its large, geographically diverse, and well-phenotyped population. Owing to sample size limitations, our models were unable to separate out the effect of GC from other immunosuppressant exposure, and do not differentiate every-other-day from daily or oral from intravenous GC exposures. To reduce the risk of outcome misclassification, we identified infections using a previously validated ICD-10-based computable phenotype.¹⁰ Dosing information for non-GC immunosuppressive medications, use of prophylactic antibiotics, and vaccination history were not available for analyses.

Our results highlight the importance of steroid-sparing medication regimens for patients with GD. Although trials support the efficacy of lower-dose GC treatment (TESTING, PEXIVAS, AURORA), clinicians should remain aware that immunosuppression regimens that include lower-dose GC exposure may carry excess infection risk. Additional studies are needed to ascertain the clinical efficacy of shorter-duration GC regimens for achieving and maintaining remission, particularly among patients with minimal change disease and focal segmental glomerulosclerosis, for whom overall GC exposure is high. Clinicians should continue to vigilantly monitor for infection and minimize infection risk through vaccination and patient education.

Data Sharing:

Deidentified participant data presented in the current report are available to investigators who submit and receive approval from the CureGN Data Coordinating Center and execute a data sharing agreement. More information can be found at the CureGN website: curegn-org.webflow.io/for-researchers.