Summary: In adults with established adrenal insufficiency, once-daily low-dose prednisolone reduced osteocalcin levels compared with thrice-daily hydrocortisone in a double-blind crossover trial. Secondary analyses favoured prednisolone for weight and HbA1c, but these short-term biomarker findings do not establish fewer fractures, cardiovascular events or superior long-term safety.
PICO Summary
| Element | Detail |
|---|---|
| Population | 47 adults were randomised and 46 analysed, including 16 with primary and 30 with secondary adrenal insufficiency. Median age was 55 years. Participants had established disease and stable hormone replacement. Diabetes mellitus, glucocorticoid-induced adrenal insufficiency and congenital adrenal hyperplasia were excluded. |
| Intervention | Prednisolone 2-5 mg once each morning for 120 days, with matching placebo at noon and in the afternoon. The median study dose was 3.5 mg daily. Among the analysed participants, 24 received prednisolone first; participants crossed over to the other treatment. |
| Comparator | Hydrocortisone in three daily doses for 120 days; median total study dose 20 mg daily. Among the analysed participants, 22 received hydrocortisone first. Study regimens were individualised before randomisation; these findings do not provide a universal dose-conversion rule. |
| Outcomes | Primary: Day-120 differences in change from baseline, prednisolone minus hydrocortisone, were -1.22 ng/mL for carboxylated osteocalcin (95% CI -2.35 to -0.10; P=0.04) and -1.38 ng/mL for undercarboxylated osteocalcin (95% CI -2.32 to -0.44; P=0.005). Lower turnover markers are not evidence of fewer fractures. Selected secondary outcomes: Weight difference -1.87 kg (95% CI -3.02 to -0.72; P=0.002), waist circumference -2.26 cm (95% CI -3.97 to -0.56; P=0.01), and HbA1c -1.23 mmol/mol (95% CI -1.95 to -0.51; P=0.001). There was no significant difference in Addi-QoL or SF-36 domains. Secondary P values were not adjusted for multiple comparisons. ARR and NNT are not applicable to these continuous outcomes. Harms: There were 133 adverse events in 37 participants. Supplement 2 lists 67 events during prednisolone periods, 60 during hydrocortisone periods and six outside the blinded treatment periods. All three serious adverse events occurred during prednisolone periods: viral gastroenteritis, acute gastroenteritis and hyponatraemia. One participant was withdrawn following hyponatraemia during a prednisolone period. No adrenal crises were observed, but this small trial cannot establish equivalence for rare harms or prove a treatment-related difference in serious-event risk. |
| Study Design | Single-centre UK, double-blind, randomised, two-period crossover trial. Each treatment lasted 120 days, separated by at least two weeks on usual open-label replacement therapy. Analysis used a repeated-measures mixed model; no carryover effect was detected. The main limitation is reliance on short-term surrogate outcomes, not fractures, cardiovascular events or mortality. Trial registration: NCT03936517. |
Expert Commentary
I regard this as a useful replacement-therapy comparison, rather than evidence that one glucocorticoid is universally safer. The crossover design reduces between-person variability, and the double-blind assessment strengthens confidence in the measured biochemical differences. The central limitation, however, is the distance between surrogate markers and outcomes that matter to patients. Lower osteocalcin does not itself demonstrate stronger bone, and the weight and HbA1c findings were secondary analyses without multiplicity adjustment. Diabetes was excluded, so the glycaemic result should not be presented as treatment evidence for people with diabetes. Safety also deserves prominence: all three serious adverse events occurred during prednisolone periods, although the small numbers cannot establish causation or a comparative safety difference. Can I use this with my patients? I would use it to inform a specialist discussion with an adult who has stable adrenal insufficiency and finds multiple daily doses difficult, not to recommend a routine switch. Individualised replacement, sick-day education and emergency steroid arrangements remain essential. The study does not supply a universal conversion ratio. I would want longer follow-up with bone density, fractures, adrenal crises and patient-reported outcomes before preferring one regimen on safety grounds. An adequately powered comparison should establish whether these short-term signals translate into meaningful clinical benefit.
References
Choudhury S, Lazarus K, Sharma A, et al. Prednisolone once daily vs hydrocortisone thrice daily in hypoadrenalism: a randomized clinical trial. JAMA Netw Open. 2026;9(3):e262982. doi:10.1001/jamanetworkopen.2026.2982. PMID: 41874506. Safety counts checked against Supplement 2, eTable 5, available with the primary report.
