Metformin Use Reduces the Risk of Olfactory Dysfunction in Diabetic Patients: A 22‐Year Follow‐Up Study

Olfaction is an important sense in the body and olfactory dysfunction may contribute to significant impact on various aspects of safety, nutrition, social interactions, quality of life, and overall mortality [1]. Diabetes mellitus (DM) has shown a greater risk of developing OD and been proposed as one of the factors associated with smell deterioration in various populations [2, 3]. Therefore, there is a need to explore a method to prevent the occurrence of OD in diabetic patients. Oral hypoglycemic drugs are commonly used to control blood sugar in diabetic patients; it is valuable to recommend whether a certain drug for treating diabetes also has a protective effect against OD. Metformin is a first-line oral hypoglycemic agent prescribed for diabetic patients and has been reported to associate to reduce the risk of cardiovascular disease, cognitive decline, dementia, stroke, and sudden deafness in patients with DM [4]. Recently, metformin had also been reported to protect olfactory function in patients with DM, but its underlying mechanisms and contributing factors remain unclear [5-7]. Notably, there is still a paucity of research examining the impact of DM on the olfactory function; we conducted this retrospective, nationwide, and long-term follow-up cohort study to determine whether metformin has its protective effect from OD in diabetic patients. This retrospective matched-cohort study used data from the National Health Insurance Research Database (NHIRD) in Taiwan, focusing on patients diagnosed with type 2 diabetes (DM) between January 1, 2000, and December 31, 2021. Patients over 18, diagnosed with DM at least three times within a year, were selected. The two cohorts comprised those who received metformin treatment and those who did not. Olfactory dysfunction (OD) was identified using ICD-9 codes, including V41.5 (“problems with smell and taste”), 781.1 (disturbances in sensor perception), 352.0 (olfactory nerve disorders), and 951.8 (traumatic anosmia). The code of traumatic anosmia was also used because an increased risk of trauma in type 2 DM patients was reported and traumatic brain injury is often associated with olfactory problem [8, 9]. Patients with prior OD diagnoses, those treated with metformin before 2000, or diagnosed with chronic kidney disease were excluded. Potential confounders, including age, sex, comorbidities, and sociodemographic variables, were considered. The study was approved by the Institutional Review Board of Tri-Service General Hospital (TSGHIRB No. E202516005), and written informed consent was waived. Data analysis was performed using IBM SPSS version 22.0. The chi-square and Fisher's exact tests compared categorical variables, while Student's t-test assessed continuous variables. Multivariate Cox proportional hazards regression was used to evaluate OD risk, with results presented as hazard ratios (HR) and 95% confidence intervals (CI). The Kaplan–Meier method with log-rank testing estimated OD risk differences between groups. A p-value < 0.05 was considered significant. Based on the data collected for this study from 2000 to 2021, a total of 45,201 DM patients who used metformin were included, along with a matched control group of 135,603 DM patients who did not use metformin. At the end of the follow-up period, 370 patients using metformin (0.82%, 370/45,201) and 1275 patients not using metformin (0.94%, 1275/135,603) developed OD, indicating a significantly lower incidence of OD among those using metformin (p = 0.019). The cumulative incidence curve for OD in the metformin cohort was significantly lower than that of the control cohort, after adjusting for age and other variables (Figure 1; log-rank test, p < 0.001). Even after adjusting for age and variables in the Cox proportional hazard regression analysis, metformin use remained significantly linked to a reduced risk of OD, with an adjusted HR of 0.635 (95% confidence interval: 0.417–0.892; p < 0.001; Table 1). In analyzing other risk factors for OD, older age, catastrophic illness, HTN, stroke, hyperlipidemia, IHD, depression, insomnia, thyroid dysfunction, residing in the winter season, and living in areas with high urbanization levels were associated with a significantly elevated risk of OD. This large-scale retrospective matched-cohort study found that metformin use is associated with a reduced incidence of OD in patients with DM. After adjusting for sociodemographic factors, comorbidities, and other variables, patients on metformin were 65% less likely to experience OD compared to non-users. Previous research by Kim et al. showed that DM significantly increases the risk of OD, with an incidence rate 1.58 times higher than in controls [2]. Sani et al. also linked advancing age and longer DM duration to reduced olfactory function [1]. A recent review noted various factors such as hypertension, obesity, thyroid dysfunction, and chronic diseases influencing olfactory function in DM patients, consistent with our study's findings (Table 1) [3]. Our results align with other studies, like Mozzanica et al., who found that metformin users had 74% lower odds of OD, Vohra et al., who observed reduced OD among metformin users in a nationally representative sample, and Assi et al., who conducted the first longitudinal study to exam a decreased OD among metformin user [5-7]. The potential mechanisms behind OD in DM include medication effects, metabolic changes, hypoxemia, oxidative stress, and vascular complications. However, the exact link between diabetic complications and OD remains debated and requires further research [3, 7]. The strengths of our study include a large sample size, nationwide scope, and long-term follow-up. Limitations include the lack of data on DM severity, metformin dosage, and olfactory severity assessments, and the inability to establish causal relationships due to the population-based design (Supporting information). Yu-Chun Chen analyzed the data and drafted the original manuscript. Chi-Hsiang Chung collected the data. Tsu-Hsuan Weng and Chun-Teng Tsai supervised the data and reviewed the manuscript. Li-Hsiang Cheng and Wu-Chien Chien designed the study and editing of the manuscript. We also appreciate the Health and Welfare Data Science Center, Ministry of Health and Welfare (HWDC, MOHW), Taiwan, for providing the National Health Insurance Research Database (NHIRD). This retrospective study obtained approval from the Institutional Review Board of Tri-Service General Hospital (IRB approval number: E202516005). The authors declare no conflicts of interest. Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.