ISSN 1470-3947 (print) | ISSN 1479-6848 (online)

Endocrine Abstracts (2019) 63 P587 | DOI: 10.1530/endoabs.63.P587

Prevalence and characteristics of Nivolumab-induced diabetes: results from a monocentric observational study

Hélène Leclerc, Delphine Drui, Pascale Mahot-Moreau, Matthieu Pichelin, Patrick Thomare, Ester Landau, Anne-Laure Fournier, Clémentine Fronteau, Eric Dresco, Bernadette Lucas-Pouliquen & Bertrand Cariou

L’institut du thorax, Department of Endocrinologie, CHU Nantes, Nantes, France.

Introduction: Nivolumab is an anti-PD1 immunotherapy that restores the immune response against cancer cells, but can induce fulminant diabetes. The prevalence of such anti-PD1 induced diabetes is not clearly determined, but it has been recently estimated between 0.4% and 0.9% in recent studies. Here, we performed a monocentric epidemiological study in order to assess the prevalence and provide a description of Nivolumab-induced diabetes cases.

Material and method: We identified all Nivolumab deliveries by the pharmacy of Nantes University Hospital from its marketing authorization in October 2014 until July 2017. Nivolumab-induced diabetes (history of diabetes or biologically determined) were found by a file review.

Results: We identified a total of 5009 deliveries of Nivolumab for 377 patients (on average 13 cures per patient). We recorded 2 Nivolumab-induced decompensations of diabetes (including one case of pre-existing diabetes that required definitive use of insulin therapy), i.e. a prevalence of 0.5%. Since July 2017, we collected 10 other cases: 7 women and 5 men concerned with anti-PD1-induced diabetes, average age of 69.7 years, and average BMI of 22.6 kg/m2. The time of onset of diabetes was between 3rd and 28th cures, with a median after the 5th cure. A fulminant pheotype with sudden onset of diabetes with frank hyperglycemia and ketoacidosis was noted in 7 patients. The average glycemia was 32.4 mmol/l, and the average HbA1c was 8.0% at diagnosis. All patients required insulin therapy and undetectable C peptide was found in 5 patients (5/9). Autoimmunity was mainly negative. Anti-GAD and anti-IA2 antibodies were absent in all patients tested (11/11 and 10/10 respectively). Anti-ZNT8 antibodies were positive in only 1 of the 6 patients tested. Increased lipase levels was found in 3 of the 6 patients tested. Unstable diabetes with glycemic variability and difficulty to achieve glycemic control was noted in 8 patients.

Discussion – conclusion: Our study found a prevalence of anti-PD1 induced diabetes of 0.5%. The clinical presentation was severe with a frequent occurrence of fulminant diabetes without markers of auto-immunity. We are currently analyzing the glycemic variability of these immunotherapy-induced diabetes, which seems to be difficult to manage for patients and clinicians. This new form of secondary diabetes requires close collaboration between oncologists and diabetologists.