Hyperglucagonaemia analysed by glucagon sandwich ELISA: nonspecific interference or truly elevated levels?

Aim/hypothesis Hyperglucagonaemia is a characteristic of several clinical conditions (e.g. end-stage renal disease (ESRD), type 2 diabetes, obesity before and after Roux-en-Y gastric bypass (RYGB) and vagotomy with pyloroplasty), but the molecular nature of ‘immunoreactive’ glucagon is poorly characterised. The specific determination of fully processed, intact glucagon requires a ‘sandwich’ assay employing a combination of antibodies directed against both N- and C-termini. We compared a novel assay for intact glucagon with a highly sensitive C-terminal RIA (hitherto considered specific) to determine the extent to which the hyperglucagonaemia measured in clinical samples was caused by authentic glucagon. Methods We examined the performance of three commercial glucagon ‘sandwich’ ELISAs. The ELISA with the best overall performance was selected to compare glucagon measurements in clinical samples with an established glucagon RIA. Results The first assay performed poorly: there was high cross-reactivity with glicentin (22%) and a lack of sensitivity for glucagon. The second and third assays showed minor cross-reactivity (1–5%) with oxyntomodulin and glicentin; however, the second assay had insufficient sensitivity for glucagon in plasma (>10–20 pmol/l). Thus, only the third assay was suitable for measuring glucagon concentrations in clinical samples. The ELISA and RIA measured similar glucagon levels in healthy individuals. Measurements of samples from individuals with abnormally high (type 2 diabetes or obese) or very elevated (post vagotomy with pyloroplasty, post-RYGB) glucagon levels were also similar in both assays. However, glucagon levels in participants with ESRD were much lower when measured by ELISA than by RIA, indicating that the apparent hyperglucagonaemia is not caused by fully processed intact glucagon. Conclusions/interpretation For most purposes, sensitive C-terminal glucagon RIAs are accurate. However, measurements may be spuriously high, at least in patients with renal disease. Trial Registration Samples from type 2 diabetic and normoglucose-tolerant patients before and 1 year after RYGB were from a study by Bojsen-Møller et al (trial registration number NCT 01202526). Samples from vagotomised and control individuals were from a study by Plamboeck et al (NCT01176890). Samples from ESRD patients were from a study by Idorn et al (NCT01327378).