Metformin Use and the Risk of Myeloproliferative Neoplasms in a Danish Population Based Cohort

INTRODUCTION The peroral antidiabetic drug metformin (MT) is primarily prescribed for hyperglycaemia due to insulin-resistance in type-2 diabetes mellitus or metabolic syndrome and is a widely used drug worldwide. Use of MT has been associated with reduced cancer risk and mortality in meta-analyses (Gandini et al, 2014). Constitutive activation of the JAK-STAT pathway by the recurrent gain-of-function mutation JAK2^V617F is a hallmark of MPN and has been related to the formation of bone marrow fibrosis by altering the bone marrow microenvironment. Preclinical studies of MT have shown antileukemic activity in myeloid neoplasms by activating the AMPK-mTOR-pathway, downregulating JAK2/STAT signaling, and mitochondrial activity suggesting MT as a possible treatment option in MPN (Biondani & Peyron, 2018; MacHado-Neto et al, 2018; Kawashima & Kirito, 2016). Recently, the phase II study Fibromet failed to demonstrate effect of MT in reversing bone marrow fibrosis in myelofibrosis. However, significant down regulation of the JAK-STAT-pathway and cytokine production were demonstrated (Campos et al, 2021). Currently, the association between MT use and the risk of MPNs remains to be elucidated.

METHODS We conducted a nationwide population-based nested case-control study to compare MT use among patients diagnosed with MPN between 2010 and 2018 and a matched population without MPN in the Danish general population to estimate odds ratios (ORs) for MPN associated with MT use. We utilized five large nationwide Danish health registries to obtain data on exposure, outcome, and covariates for adjustment. The index date was defined as the date of MPN diagnosis and cases were identified using the National Chronic Myeloid Neoplasia Registry covering >90% of MPN cases in Denmark (Bak et al, 2016). Controls were selected by incidence density sampling (five to one) from the Danish general population matched on age, gender, and calendar year. Data on MT exposure were collected from the Danish National Prescription Registry. "Ever-use” was defined as a minimum of one prescription before the index date and "Long-term use” as >5 years of MT use (assuming an intake of the WHOs defined daily dose for MT at 2000 mg plus 25% to account for minor prescription variations and incompliances). Exposure to MT, comorbidity, and concomitant drug use were disregarded 12 months prior to the index date to avoid reverse causation. By conditional logistic regression analysis, we obtained ORs and 95% confidence intervals (95% CIs) adjusted for study design (age, gender, and index date) and fully adjusted ORs (aORs) (see footnote in Table 2 for adjusted variables).

RESULTS The study population included a total of 3,816 cases and 19,080 matched controls (Table 1). Exposure assessment categorized 7.0% of cases and 8.2% of controls as "Ever-users” of MT resulting in an aOR for MPN of 0.70 (95% CI, 0.61-0.81). Long-term users were rare and comprised 1.1% of cases and 2.0% of controls resulting in an aOR for MPN 0.45 (0.33-0.63). When assessing the duration of MT exposure, we found a trend towards a dose-response relationship (<1 year: aOR 0.79 [0.62-1.00]; 1-4.99 years: 0.78 [0.64-0.95]; 5-9.99 years: 0.42 [0.29-0.61]; >10 years: 0.56 [0.30-1.03]); however, this was not statistically significant for the >10 years category, most likely due to small numbers. In stratified analyses the protective effect of long-term MT use was consistent according to gender with aOR of 0.37 (0.23-0.60) and 0.56 (0.36-0.89) for males and females, respectively, compared to males and females without MT use. In a subgroup analysis the effect was consistent across all age groups and MPN-subtypes. However, aORs did not reach statistical significance for age <60 years, myelofibrosis, and MPN unclassifiable (Table 2).

CONCLUSION In this study we found an inverse association between MT use and risk of MPNs indicating a potential protective effect of MT on the development of MPNs. The protective effect was most pronounced for long-term use, older age (≥60 years) and for polycythemia vera and essential thrombocythemia. Whether this effect is causal should be determined interventional studies.

Jakobsen:Roche: Honoraria. El-Galaly:Roche: Ended employment in the past 24 months; Abbvie: Other: Teaching in 2021. Roug:Jazz Pharmaceuticals: Other: Travel grant.