The insulin tolerance test (ITT) is the most reliable method for diagnosing growth hormone deficiency (GHD). Our study aimed to determine a new glycemic threshold that allows for an adequate growth hormone (GH) response without increasing the risk of deleterious outcomes.

This study was conducted retrospectively on children who were suspected of GHD. There were two groups of children: group 1 (G1) consisted of children who underwent an ITT and reached a hypoglycemic level of 2.2 mmol/l, and group 2 (G2) included those who did not reach biochemical hypoglycemia but exhibited clinical hypoglycemic symptoms, which led to cessation of the test.

Our population consisted of 129 patients. The mean nadir blood glucose level was 1.798 ±0.25 mmol/l in G1, significantly lower than in G2, at 2.57 ±0.28 mmol/l (p < 10–3). GH in G1 reached a mean peak of 4.71 ±3.20 ng/ml, significantly lower than in G2 (8.56 ±5.99 ng/ml) (p = 0.01). Among G1, 86% of the patients presented a GHD, compared to 55.6% in G2. The proportion of somatotrophic deficiency was significantly higher in G1 compared to G2 (p = 0.01). Receiver operating characteristic curve analysis revealed a threshold of 2.44 mmol/l with a sensitivity of 90% and a specificity of 60% (area under the curve = 0.915; confidence interval 95% (0.87–0.96).

Based on our findings, we recommend implementing this new threshold when hypoglycemic symptoms are present. This higher threshold offers several benefits, including reduced risk to patients, while maintaining accurate interpretation of the test with good sensitivity and specificity.