Evaluating the Low-dose ACTH Stimulation Test in Children

Ideal Times for Cortisol Measurement

Harpreet Gill; Nick Barrowman; Richard Webster; Alexandra Ahmet


J Clin Endocrinol Metab. 2019;104(10):4587-4593. 

In This Article

Abstract and Introduction


Context: Central adrenal insufficiency (AI) can be diagnosed with the low-dose ACTH stimulation test (LDST). Protocols determining timing of cortisol sampling vary, with 30 minutes after stimulation being most common.

Objectives: To determine optimal times to draw cortisol levels and factors predicting timing of peak cortisol levels in children undergoing LDST.

Design: Retrospective chart review of LDSTs between February 2014 and September 2017.

Setting: The Children's Hospital of Eastern Ontario.

Patients: Patients 3 months to 20 years who underwent LDSTs.

Intervention: LDSTs were performed with cortisol levels at 0, 15, 30, and 60 minutes after 1 & μu;g cosyntropin. Cortisol values <18 & μg/dL (500 nmol/L) determined AI.

Main Outcome Measures: The incremental value of testing cortisol at 15 or 60 minutes, in addition to the standard 30-minute sample, was estimated.

Results: A total of 221 patients met inclusion criteria. The mean age was 9.7 years, and 32% were female. Peak cortisol levels were 19%, 67%, and 14% at 15, 30, and 60 minutes, respectively. One false positive LDST result would be prevented for every 24 (95% CI, 13 to 46) or 55 (95% CI, 22 to 141) patients tested at 15 or 60 minutes in addition to the standard 30-minute test. Of the 122 patients who passed the LDST, discontinuing the 15- and 60-minute samples would have misdiagnosed 12 patients (9.8%). Glucocorticoid exposure, age, and body mass index z scores were independent predictors of peak cortisol timing.

Conclusion: Although the majority of patients peak 30 minutes after cosyntropin administration, testing cortisol levels at 15 and 60 minutes reduces the risk of false positive LDSTs.


The diagnosis of adrenal insufficiency (AI) presents a challenge for the clinician and often necessitates provocative testing. The insulin tolerance test (ITT) has been the gold standard provocative test in the diagnosis of AI since its introduction in the 1960s.[1] However, its use for children is limited by the risk of hypoglycemia.[2]

Given the limitations of the ITT, the 250 μg standard-dose ACTH stimulation test was reevaluated and validated as an alternative provocative test for the diagnosis of AI.[3] Given concerns that the 250-μg standard-dose ACTH stimulation test was not sufficiently sensitive for the diagnosis of central AI, several studies and a meta-analysis validated the 1 μg low-dose ACTH stimulation test (LDST) as a more sensitive diagnostic test for the evaluation of secondary and tertiary AI.[1,2,4–7] However, there continues to be debate and significant variability across centers regarding the optimal time to measure cortisol levels after the administration of ACTH in the LDST.

A pediatric study in 1998 showed that peak cortisol levels occurred 30 minutes after ACTH administration in the LDST.[2] Several adult studies showed similar results.[1,5,7] Therefore, many centers measure a cortisol level only 30 minutes after ACTH administration. This practice is reflected in a meta-analysis by Kazlauskaite et al.,[6] where 10 out of 11 studies evaluating the LDST sampled cortisol only at baseline and 30 minutes after stimulation. However, a study by Zarković et al.[8] showed that in adults, the cortisol increment between baseline and 15 minutes after ACTH stimulation was the most valuable parameter in diagnosing AI. Furthermore, a recent study by Cartaya and Misra[9] showed that when tested at 30 and 60 minutes after ACTH administration, the majority (54%) of their retrospective pediatric cohort reached peak cortisol levels at 60 minutes. Moreover, 13% of patients who did not pass the test at 30 minutes did so at 60 minutes. This finding suggests that measuring 60 minutes after ACTH is important to prevent overdiagnosis of AI. However, this study did not measure cortisol levels at 15 minutes.[9]

Overdiagnosing AI leads to unnecessary glucocorticoid exposure, the usage of health care professional time and resources, medication costs, the need for repeat testing, and patient and parental stress associated with the diagnosis. However, reducing the number of cortisol levels drawn decreases blood loss through phlebotomy and the cost of additional cortisol tests.

The aim of this study was to determine the benefit of testing cortisol levels at 15 and 60 minutes, in addition to the standard 30-minute sample in the LDST. We evaluated this benefit by assessing whether there were patients who would have failed the LDST if tested only at 30 minutes but who passed with the addition of the 15- or 60-minute samples. Secondarily, we evaluated predictors of the timing of peak cortisol levels. The Children's Hospital of Eastern Ontario is well positioned to assess these outcomes given the strictly protocoled LDSTs, with cortisol levels measured at baseline and at 15, 30, and 60 minutes after administration of 1 μg of ACTH (Cortrosyn™).