Endocrine Abstracts

Background: Targeted therapy using phosphatidylinositol 3-kinase (PI3K) inhibitors has become widespread in the treatment of cancers. Although PI3K inhibitors raised concern about the alteration of glucose and lipid metabolism from animal studies, human clinical data have been rarely reported.

Aim: We investigated the clinical manifestations, incidence of diabetes and its predictive factors in cancer patients who treated with PI3K inhibitors.

Method: We analyzed 38 diabetes-free patients with advanced solid cancer who initiated PI3K inhibitor treatment at the university-affiliated hospital and followed up for average 239 days retrospectively. Cox regression analysis was performed to identify independent predictive parameters for the development of diabetes after initiation of PI3K inhibitors.

Results: Of 38 patients with PI3K inhibitors initiation (mean age 54.5 years, 23.7% female), 55.3% developed diabetes during the treatment within the mean period of 29 days. Of these, 28.6% presented the remission of diabetes, 72 mean days after the discontinuation of PI3K inhibitor use. Incident diabetes patients revealed higher proportions of use of hypertension medication, higher levels of HbA1c and fasting glucose at baseline, and longer duration of PI3K inhibitor use (Ps<0.005). Presence of prior use of steroid and higher baseline HbA1c levels were associated with the development of diabetes (HR=8.41,95%CI=1.89–37.33; HR=2.16,95%CI=1.09–4.25, respectively). Patients showing remission of diabetes after discontinuing PI3K inhibitor use were younger (P=0.035) and maintained lower fasting glucose levels during PI3K inhibitor use (P=0.001) compared to patients with persistent diabetes.

Discussion: In conclusion, history of steroid use and higher HbA1c levels at baseline may be important predictors for developing diabetes in patients with cancers who treated with PI3K inhibitors. Close observation and careful intervention are needed when treating with PI3K inhibitors in these patients.