Assessment and Management of Anti-Insulin Autoantibodies in Varying Presentations of Insulin Autoimmune Syndrome

Abstract

Context: Insulin autoimmune syndrome (IAS), spontaneous hyperinsulinemic hypoglycemia due to insulin-binding autoantibodies, may be difficult to distinguish from tumoral or other forms of hyperinsulinemic hypoglycemia, including surreptitious insulin administration. No standardized treatment regimen exists.

Objectives: To evaluate an analytic approach to IAS and responses to different treatments.

Design and setting: Observational study in the UK Severe Insulin Resistance Service.

Patients: Six patients with hyperinsulinemic hypoglycemia and detectable circulating anti-insulin antibody (IA).

Main outcome measures: Glycemia, plasma insulin, and C-peptide concentrations by immunoassay or mass spectrometry (MS). Immunoreactive insulin was determined in the context of polyethylene glycol (PEG) precipitation and gel filtration chromatography (GFC). IA quantification using ELISA and RIA, and IA were further characterized using radioligand binding studies.

Results: All patients were diagnosed with IAS (five IgG, one IgA) based on a high insulin/C-peptide ratio, low insulin recovery after PEG precipitation, and GFC evidence of antibody-bound insulin. Neither ELISA nor RIA result proved diagnostic for every case. MS provided a more robust quantification of insulin in the context of IA. One patient was managed conservatively, four were treated with diazoxide without sustained benefit, and four were treated with immunosuppression with highly variable responses. IA affinity did not appear to influence presentation or prognosis.

Conclusions: IAS should be considered in patients with hyperinsulinemic hypoglycemia and a high insulin/C-peptide ratio. Low insulin recovery on PEG precipitation supports the presence of insulin-binding antibodies, with GFC providing definitive confirmation. Immunomodulatory therapy should be customized according to individual needs and clinical response.

Figure 1.

Variable patterns of dysglycemia of patients studied. (a) Venous plasma glucose concentrations during a 75-g OGTT at presentation of patient 1; ○ denotes glucose measurements following glucose rescue. The glucose nadir was 39 mg/dL (2.2 mmol/L). (b) Demonstration of labile glycemia in patient 1 at presentation by CGMS. (c) Demonstration of normoglycemia in patient 1 following immunomodulation therapy. (d) Demonstration of labile glycemia in patient 3 concomitant with glucocorticoid therapy. (e) Demonstration of reactive hypoglycemia in patient 4 at presentation by mixed-meal tolerance test. The peak glucose concentration was 232 mg/dL (12.9 mmol/L) with glucose nadir at 300 minutes of 29 mg/dL (1.6 mmol/L). (f) Demonstration of reactive and nocturnal hypoglycemia in patient 4 at presentation by CGMS. (g) Demonstration of reactive hypoglycemia in patient 6 at presentation by 75-g OGTT. The glucose nadir was 26 mg/dL (1.4 mmol/L). (h) Demonstration of labile glycemia in patient 6 at presentation by CGMS.

Figure 2.

Displacement curves for serum samples from patients 1 to 5 at various dilutions in antibody-negative serum following competitive displacement with unlabeled human insulin. Although identified as low affinity (4.1 × 10⁻⁷ mol/L), patient 6 plasma was considered unreliable because baseline levels of insulin binding were very low. Serum was diluted as follows: patient 1, 10-fold; patient 2, 50-fold; patient 3, 100-fold; patient 4, neat; patient 5, 10-fold.

Figure 3.

Demonstration of insulin-antibody complexes using GFC. Results of insulin assay after GFC of nonfasting plasma. Elution volumes of immunoglobulin (Ig), albumin (Alb), and monomeric insulin (mIns) are shown. Results are shown for patient 2 at (a) presentation (pretherapy) and (b) with and without preincubation of plasma with exogenous insulin posttherapy, as well as with and without preincubation of plasma with exogenous insulin at presentation for (c) patient 3, (d) patient 4, and (e) patient 6.

Figure 4.

Response of biochemical markers to therapy in patient 3. (a) Cumulative results for patient 3 over course of treatments, including MMF, azathioprine (AZA), prednisolone (Pr), dexamethasone (Dex), rituximab (R), and plasma exchange (PEx), showing anti-insulin IgG concentrations (in-house human insulin–specific ImmunoCAP) and immunoassay insulin recovery following PEG precipitation over time. (b) Effect of plasma exchange on insulin immunoassay linearity to dilution. Calculated insulin concentration plotted against plasma dilution of patient 3 plasma before plasma exchange and following cycle 1 and cycle 9.