Is Metformin Use Associated With Decreased Thyroid Cancer Risk in Patients With Acromegaly?

Abstract

Context: Acromegaly has long been blamed to portend an increased risk for benign and malignant thyroid neoplasia. Growth hormone (GH) and consequent insulin-like growth factor 1 (IGF-1) hypersecretion are implicated in cancer promotion. Metformin, a biguanide derived from the French lilac, is gaining considerable interest because of its plausible anti-tumor properties. Besides, metformin has been shown to inhibit somatotroph proliferation and decrease GH secretion in in vivo studies. Patients with acromegaly have high incidence of diabetes and were thereof treated with metformin. We hypothesized metformin use may be linked to decreased thyroid cancer incidence in patients with acromegaly. Study Design and Methods: The medical records of 508 patients with acromegaly followed at our tertiary referral center between 1969 and 2019 were retrospectively reviewed. The inclusion criteria were having a follow-up duration for at least 12 months and being regularly screened for nodular thyroid disease and thyroid cancer by ultrasonography as indicated in respective guidelines. Patients with acromegaly were evaluated based on ongoing or prior history of metformin use or thyroid cancer diagnosis. Metformin exposure was defined as use of metformin for at least 12 months on a regular basis between initial date of acromegaly and time prior to cancer diagnosis date. Considering the long patency period of cancer of interest, we excluded exposures in the year immediately prior to index cancer date. We evaluated the effect of metformin exposure on risk of thyroid cancer using Kaplan-Meier analysis. Results: Final analysis included 377 patients with acromegaly. Mean age at acromegaly diagnosis was 41.6 ± 11.7 and 60.5% of the patients were female. Three hundred twenty-two patients (85.4%) had undergone transsphenoidal surgery as primary therapy, 73 patients (19.4%) needed radiotherapy and 178 patients (46%) received post-operative medical therapy. Median follow-up duration was 73.5 months (IQR [31.0-137.7]). One hundred twenty patients (31.9%) had an ongoing or prior use of metformin, and total of 19 patients (5%) had thyroidcancer. Age at acromegaly diagnosis, gender distribution, baseline GH and IGF-1 levels, pituitary tumor size and invasiveness, biological aggressiveness, curative therapy options, treatment responses didn’t differ between metformin users and non-users, as well as between those having and not having thyroid cancer. Kaplan-Meier estimates for 1 year, 3 years and 5 years of metformin exposure showed decreased probability of thyroid cancer incidence (p<0.05 for all). Conclusion: Although our results imply decreased thyroid cancer risk upon metformin exposure, prospective study designs with larger cohorts are obliged in order to fully elucidate the effect of metformin use on thyroid cancer.