Steroid-induced adrenal insufficiency in children with nephrotic syndrome: a systematic review.
Kalliopi Vardaki, Ilianna Maniadaki, Emmanouil Galanakis et al.
✦ AI-curated · Sources linked
In 30 seconds
This systematic review evaluated steroid-induced adrenal insufficiency (AI) in children with idiopathic nephrotic syndrome treated with corticosteroids. Thirteen studies involving 516 children were included, with findings synthesised narratively because of methodological heterogeneity. Reported AI prevalence ranged from 5.9 to 92.9%, varying substantially with the diagnostic test and timing of assessment. AI was more often reported in children with frequently relapsing or steroid-dependent disease and prolonged or repeated corticosteroid exposure. The review also identified potential clinical risks, including relapse during infections and adrenal crisis during physiological stress when stress-dose corticosteroids are not given.
Key findings
- Across 13 studies and 516 children, reported AI prevalence ranged from 5.9 to 92.9%.
- Seven adrenal-function testing methods were identified; the 2-h ACTH stimulation test had the highest diagnostic yield.
- AI was more commonly observed with frequently relapsing or steroid-dependent disease and prolonged or repeated corticosteroid exposure.
- AI was linked to increased relapse risk, particularly during infections; suppressed adrenal function may also leave children vulnerable to adrenal crisis during physiological stress.
Why it matters
The wide prevalence range reflects inconsistent diagnostic methods and criteria, so the true burden of AI in children receiving corticosteroids for nephrotic syndrome remains uncertain. Recognising possible adrenal suppression matters because physiological stress, including infection, may coincide with relapse or risk of adrenal crisis. The review highlights the need for more consistent assessment and better evidence on longer-term outcomes.
What to keep in mind
The included studies were mostly small and single-centre, with substantial methodological heterogeneity, inconsistent AI definitions and cortisol cutoffs, and limited long-term follow-up.