Prediction of hospital-acquired influenza using machine learning algorithms: a comparative study.

BACKGROUND: Hospital-acquired influenza (HAI) is under-recognized despite its high morbidity and poor health outcomes. The early detection of HAI is crucial for curbing its transmission in hospital settings. AIM: This study aimed to investigate factors related to HAI, develop predictive models, and subsequently compare them to identify the best performing machine learning algorithm for predicting the occurrence of HAI. METHODS: This retrospective observational study was conducted in 2022 and included 111 HAI and 73,748 non-HAI patients from the 2011-2012 and 2019-2020 influenza seasons. General characteristics, comorbidities, vital signs, laboratory and chest X-ray results, and room information within the electronic medical record were analysed. Logistic Regression (LR), Random Forest (RF), Extreme Gradient Boosting (XGB), and Artificial Neural Network (ANN) techniques were used to construct the predictive models. Employing randomized allocation, 80% of the dataset constituted the training set, and the remaining 20% comprised the test set. The performance of the developed models was assessed using metrics such as the area under the receiver operating characteristic curve (AUC), the count of false negatives (FN), and the determination of feature importance. RESULTS: Patients with HAI demonstrated notable differences in general characteristics, comorbidities, vital signs, laboratory findings, chest X-ray result, and room status compared to non-HAI patients. Among the developed models, the RF model demonstrated the best performance taking into account both the AUC (83.3%) and the occurrence of FN (four). The most influential factors for prediction were staying in double rooms, followed by vital signs and laboratory results. CONCLUSION: This study revealed the characteristics of patients with HAI and emphasized the role of ventilation in reducing influenza incidence. These findings can aid hospitals in devising infection prevention strategies, and the application of machine learning-based predictive models especially RF can enable early intervention to mitigate the spread of influenza in healthcare settings.

Disparities in cause-specific mortality by health insurance type and premium: evidence from Korean NHIS-HEALS cohort study, 2002-2019.

BACKGROUND: Although one's socioeconomic status affects health outcomes, limited research explored how South Korea's National Health Insurance (NHI) system affects mortality rates. This study investigated whether health insurance type and insurance premiums are associated with mortality. METHODS: Based on the National Health Insurance Service-Health Screening cohort, 246,172 men and 206,534 women aged ≥ 40 years at baseline (2002-2003) were included and followed until 2019. Health insurance type was categorized as employee-insured (EI) or self-employed-insured (SI). To define low, medium, and high economic status groups, we used insurance premiums at baseline. Death was determined using the date and cause of death included in the cohort. Cox proportional hazard models were used to analyze the association between insurance factors and the overall and cause-specific mortality. RESULTS: The SI group had a significantly higher risk of overall death compared to the EI group (adjusted hazard ratio (HR) [95% confidence interval]: 1.13 [1.10-1.15] for men and 1.18 [1.15-1.22] for women), after adjusting for various factors. This trend extended to death from the five major causes of death in South Korea (cancer, cardiovascular disease, cerebrovascular disease, pneumonia, and intentional self-harm) and from external causes, with a higher risk of death in the SI group (vs. the EI group). Further analysis stratified by economic status revealed that individuals with lower economic status faced higher risk of overall death and cause-specific mortality in both sexes, compared to those with high economic status for both health insurance types. CONCLUSION: This nationwide study found that the SI group and those with lower economic status faced higher risk of overall mortality and death from the five major causes in South Korea. These findings highlight the potential disparities in health outcomes within the NHI system. To address these gaps, strategies should target risk factors for death at the individual level and governments should incorporate such strategies into public health policy development at the population level. TRIAL REGISTRATION: This study was approved by the Institutional Review Board of Chungbuk National University Hospital (CBNUH-202211-HR-0236) and adhered to the principles of the Declaration of Helsinki (1975).

Variable Selection in Bayesian Multiple Instance Regression using Shotgun Stochastic Search.

In multiple instance learning (MIL), a bag represents a sample that has a set of instances, each of which is described by a vector of explanatory variables, but the entire bag only has one label/response. Though many methods for MIL have been developed to date, few have paid attention to interpretability of models and results. The proposed Bayesian regression model stands on two levels of hierarchy, which transparently show how explanatory variables explain and instances contribute to bag responses. Moreover, two selection problems are simultaneously addressed; the instance selection to find out the instances in each bag responsible for the bag response, and the variable selection to search for the important covariates. To explore a joint discrete space of indicator variables created for selection of both explanatory variables and instances, the shotgun stochastic search algorithm is modified to fit in the MIL context. Also, the proposed model offers a natural and rigorous way to quantify uncertainty in coefficient estimation and outcome prediction, which many modern MIL applications call for. The simulation study shows the proposed regression model can select variables and instances with high performance (AUC greater than 0.86), thus predicting responses well. The proposed method is applied to the musk data for prediction of binding strengths (labels) between molecules (bags) with different conformations (instances) and target receptors. It outperforms all existing methods, and can identify variables relevant in modeling responses.

Positive association between the ratio of triglycerides to high-density lipoprotein cholesterol and diabetes incidence in Korean adults.

BACKGROUND: Insulin resistance is associated with the incidence of diabetes and cardiovascular diseases such as myocardial infarction. The ratio of triglycerides (TG) to high-density lipoprotein cholesterol (HDL-C) (TG/HDL-C ratio) is positively correlated with insulin resistance. This study aimed to investigate the relationship between the TG/HDL-C ratio and the incidence of diabetes in Korean adults. METHODS: This retrospective study used data from the National Health Insurance Service-National Health Screening Cohort. The TG/HDL-C ratio was divided into three tertiles, the T1, T2, and T3 groups, based on sex. We estimated the hazard ratios (HRs) and 95% confidence intervals (CIs) for diabetes using multivariate Cox proportional hazards regression analyses. RESULTS: A total of 80,693 subjects aged between 40 and 79 years were enrolled. The median follow-up period was 5.9 years. The estimated cumulative incidence of diabetes in the T1, T2, and T3 groups was 5.94%, 8.23%, and 13.50%, respectively, in men and 4.12%, 4.72%, and 6.85%, respectively, in women. Compared to T1, the fully adjusted HRs (95% CIs) of the T2 and T3 groups for new-onset diabetes were 1.17 (1.06-1.30) and 1.47 (1.34-1.62), respectively, in men and 1.20 (1.02-1.42) and 1.52 (1.30-1.78), respectively, in women. CONCLUSIONS: Increased TG/HDL-C ratio was significantly associated with a higher risk of new-onset diabetes in both sexes.

Statin Use and Incidence of Chronic Kidney Disease in Hypercholesterolemia Patients with Normal Renal Function.

INTRODUCTION: Dyslipidemia is a known risk factor for chronic kidney disease (CKD). The effects of statins on CKD have already been studied in patients with CKD; however, data on the general population are limited. This study aimed to determine the relationship between statin use and the incidence of CKD in patients with hypercholesterolemia having normal renal function. METHODS: A total of 7,856 participants aged 40-79 years at baseline (2009-2010) were included in the final analyses. The participants were divided into statin users (n = 4,168) and statin nonusers (n = 3,668), according to the statin usage. The Cox proportional hazard regression model was used to evaluate the adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for CKD. RESULTS: The median follow-up duration was 5.8 years. A total of 543 cases of CKD (285 cases in males and 258 cases in females) occurred during the study period. The estimated cumulative incidence of CKD was significantly different between male statin nonusers and users (p < 0.001), while it was not statistically significant between female statin nonusers and users (p = 0.126). Compared with statin nonusers, the fully adjusted HRs (95% CIs) for CKD in statin users were 1.014 (0.773-1.330) in males and 1.117 (0.843-1.481) in females. CONCLUSION: Dyslipidemia is an obvious risk factor for CKD; however, statin use in patients with hypercholesterolemia having normal renal function does not demonstrate a clear relationship with the incidence of CKD.

Metformin usage and the risk of colorectal cancer: a national cohort study.

PURPOSE: This study aimed to investigate the association between metformin usage and the risk of colorectal cancer (CRC) using data from the Korean National Health Insurance Service-National Health Screening Cohort database. METHODS: Data from the NHIS-HEALS cohort between 2002 and 2015 were longitudinally analyzed. Subjects were divided into three groups: metformin non-users with diabetes mellitus (DM), metformin users with DM, and no DM group. CRC was defined using the ICD-10 code (C18.0-C20.0) at the time of admission. Cox proportional hazard regression models were adopted after stepwise adjustment for confounders to investigate the association between metformin usage and colorectal cancer risk. RESULTS: During the follow-up period, of the total 323,430 participants, 2341 (1.33%) of the 175,495 males and 1204 (0.81%) of the 147,935 females were newly diagnosed with CRC. The estimated cumulative incidence of CRC was significantly different among the three groups based on Kaplan-Meier's survival curve (p values < 0.05 in both sexes). Compared with metformin non-users, hazard ratios (95% CIs) of metformin users and the no DM group were 0.66 (0.51-0.85) and 0.72 (0.61-0.85) in males and 0.59 (0.37-0.92) and 0.93 (0.66-1.29) in females, respectively, after being fully adjusted. CONCLUSIONS: Metformin users with diabetes appear to have a significantly lower risk of CRC compared with metformin non-users.

Metformin use reduced the risk of stomach cancer in diabetic patients in Korea: an analysis of Korean NHIS-HEALS database.

BACKGROUND: Diabetes mellitus (DM) increases atherosclerotic cardiovascular complications and cancer risks. Stomach cancer is the most common cancer in Korea. Although the survival rate of stomach cancer has improved, the disease burden is still high. METHODS: This retrospective study investigated the association between metformin use and stomach cancer incidence in a Korean population using the National Health Insurance Service-National Health Screening Cohort database. Participants aged 40-80 years old at the baseline period (2002-2003) were enrolled. The study population was categorized into three groups of metformin non-users with DM, metformin users with DM, and individuals without DM (No DM group). RESULTS: A total of 347,895 participants (14,922 metformin non-users, 9891 metformin users, and 323,082 individuals without DM) were included in the final analysis. The median follow-up duration was 12.70 years. The estimated cumulative incidence of stomach cancer was highest in metformin non-users and lowest in the No DM group (men vs. women: 3.75 vs. 1.97% in metformin non-users, 2.91 vs. 1.53% in metformin users, and 2.54 vs. 0.95% in the No DM group). Compared with metformin non-users, the hazard ratios (95% confidence intervals) for stomach cancer incidence of metformin users and the No DM group were 0.710 (0.579-0.870) and 0.879 (0.767-1.006) in men and 0.700 (0.499-0.981) and 0.701 (0.544-0.903) in women, respectively, after full adjustment. CONCLUSIONS: Metformin users with DM in the Korean population were at lower risk of stomach cancer incidence after controlling for potential confounding factors.

Metformin use reduced the overall risk of cancer in diabetic patients: A study based on the Korean NHIS-HEALS cohort.

BACKGROUND AND AIM: Several studies have reported the preventive effect of metformin on cancer development. This study aimed to investigate the relationship between use of metformin and risk of cancer in Koreans. METHODS AND RESULTS: This study was designed retrospectively using the National Health Insurance Service-National Health Screening Cohort conducted between 2002 and 2015. 40 to 69-year-old subjects who received a health screening examination from 2002 to 2003 were enrolled. Hazard ratios (HRs) and 95% confidence intervals (CIs) for cancer were estimated in a multivariate Cox proportional regression analysis. A total of 323,430 subjects was enrolled (301,905 individuals without diabetes [No DM], 8643 diabetic patients with metformin treatment [metformin users], and 12,882 diabetic patients without metformin treatment [metformin non-users]). The median follow-up period was 12.7 years. Cumulative incidence of overall cancer was 7.9% (7.7, 10.3, and 11.1% in No DM, metformin users and non-users, respectively). Compared to metformin non-users, the fully adjusted HRs (95% CIs) of metformin users and No DM for overall cancer incidence were 0.73 (0.66-0.81) and 0.75 (0.64-0.88), respectively, in men and 0.83 (0.78-0.89) and 0.81 (0.72-0.92) in women. CONCLUSIONS: Diabetic patients receiving metformin treatment, and individuals without diabetes were at lower risk for cancer incidence than diabetic patients without metformin treatment.

Association between statin use and all-cause mortality in cancer survivors, based on the Korean health insurance service between 2002 and 2015.

BACKGROUND AND AIMS: Cancer is the number one cause of death in Korea. This study aimed to investigate if statin use in cancer survivors was inversely associated with all-cause mortality. METHODS AND RESULTS: Data from the 2002 to 2015 National Health Insurance Service-National Health Screening Cohort (NHIS-HEALS) were used. The Kaplan-Meier estimator was used to estimate the survival function according to statin usage. Cox proportional hazards regression models were adopted after stepwise adjustment for potential confounders to calculate hazard ratios (HRs) and 95% confidence intervals (CIs) for all-cause mortality. The median follow-up duration was 10.0 years. Statin users had a higher percentage of diabetes and hypertension in both sexes. Survival rates of statin users were higher than non-users (p-values <0.001 in men and 0.021 in women). Compared to non-users, the HRs (95% CIs) of statin users for all-cause mortality were 0.327 (0.194-0.553) in men and 0.287 (0.148-0.560) in women after adjustment for potential confounding factors. CONCLUSIONS: Statin users in cancer survivors had higher survival rate than non-users in both sexes.

Statin use and site-specific risk of colorectal cancer in individuals with hypercholesterolemia from the National Health Insurance Service-National Health Screening Cohort (NHIS-HEALS).

BACKGROUND AND AIMS: We investigated the association between statin use and site-specific risk of colorectal cancer in individuals with hypercholesterolemia. METHODS AND RESULTS: This study is based on the National Health Insurance Service-National Health Screening Cohort, conducted during 2002-2015. Statin users were classified as high and low users according to medication possession ratio (MPR). Statin nonusers comprised participants who did not use statins during the entire follow-up period. In total, 17,737 statin users and 13,412 statin nonusers were included in the analysis, with a median follow-up period of 12.7 years. Cox proportional hazards regression models were adopted after stepwise adjustment for confounders to investigate prospective association between statin usage and colorectal cancer risk. In total, 378 (2.3%) of 16,588 male participants and 239 (1.6%) of 14,561 female participants had colorectal cancer during the follow-up period. Compared to nonusers, fully adjusted hazard ratios (HRs) (95% confidence intervals [95% CIs]) for colorectal cancer risk in high statin users were 0.56 (0.42-0.75) in men and 0.64 (0.46-0.90) in women. In men, the fully adjusted HRs for proximal and rectal cancer for high users were 0.29 (0.15-0.56) and 0.52 (0.35-0.78), respectively, compared to those for nonusers. In women, statistical significance was seen only in rectal cancer (HR 0.43 [0.25-0.72]) but not in proximal or distal colon cancer. CONCLUSIONS: High statin users with hypercholesterolemia were associated with lower risk of overall colorectal cancer, especially proximal colon cancer in men and rectal cancer in both sexes.

Detection of gene-environment interactions in a family-based population using SCAD.

Gene-environment interaction (GxE) is emphasized as one potential source of missing genetic variation on disease traits, and the ultimate goal of GxE research is prediction of individual risk and prevention of complex diseases. However, there are various challenges in statistical analysis of GxE. In this paper, we focus on the three methodological challenges: (i) the high dimensions of genes; (ii) the hierarchical structure between interaction effects and their corresponding main effects; and (iii) the correlation among subjects from family-based population studies. In this paper, we propose an algorithm that approaches all three challenges simultaneously. This is the first penalized method focusing on an interaction search based on a linear mixed effect model. For verification, we compare the empirical performance of our new method with other existing methods in simulation study. The results demonstrate the superiority of our method under overall simulation setup. In particular, the outperformance obviously becomes greater as the correlation among subjects increases. In addition, the new method provides a robust estimate for the correlation among subjects. We also apply the new method on Genetics of Lipid Lowering Drugs and Diet Network study data. Copyright © 2017 John Wiley & Sons, Ltd.

Identification of Imaging Predictors Discriminating Different Primary Liver Tumours in Patients with Chronic Liver Disease on Gadoxetic Acid-enhanced MRI: a Classification Tree Analysis.

OBJECTIVES: To identify predictors for the discrimination of intrahepatic cholangiocarcinoma (IMCC) and combined hepatocellular-cholangiocarcinoma (CHC) from hepatocellular carcinoma (HCC) for primary liver cancers on gadoxetic acid-enhanced MRI among high-risk chronic liver disease (CLD) patients using classification tree analysis (CTA). METHODS: A total of 152 patients with histopathologically proven IMCC (n = 40), CHC (n = 24) and HCC (n = 91) were enrolled. Tumour marker and MRI variables including morphologic features, signal intensity, and enhancement pattern were used to identify tumours suspicious for IMCC and CHC using CTA. RESULTS: On CTA, arterial rim enhancement (ARE) was the initial splitting predictor for assessing the probability of tumours being IMCC or CHC. Of 43 tumours that were classified in a subgroup on CTA based on the presence of ARE, non-intralesional fat, and non-globular shape, 41 (95.3 %) were IMCCs (n = 29) or CHCs (n = 12). All 24 tumours showing fat on MRI were HCCs. The CTA model demonstrated sensitivity of 84.4 %, specificity of 97.8 %, and accuracy of 92.3 % for discriminating IMCCs and CHCs from HCCs. CONCLUSIONS: We established a simple CTA model for classifying a high-risk group of CLD patients with IMCC and CHC. This model may be useful for guiding diagnosis for primary liver cancers in patients with CLD. KEY POINTS: • Arterial rim enhancement was the initial splitting predictor on CTA. • CTA model achieved high sensitivity, specificity, and accuracy for discrimination of tumours. • This model may be useful for guiding diagnosis of primary liver cancers.

Bayesian estimation of the phylogeography of African gorillas with genome-differentiated population trees.

Phylogeography investigates the historical process that is responsible for the contemporary geographic distributions of populations in a species. The inference is made on the basis of molecular sequence data sampled from modern-day populations. The estimates, however, may fluctuate depending on the relevant genomic regions, because the evolution mechanism of each genome is unique, even within the same individual. In this article, we propose a genome-differentiated population tree model that allows the existence of separate population trees for each homologous genome. In each population tree, the unique evolutionary characteristics account for each genome, along with their homologous relationship; therefore, the approach can distinguish the evolutionary history of one genome from that of another. In addition to the separate divergence times, the new model can estimate separate effective population sizes, gene-genealogies and other mutation parameters. For Bayesian inference, we developed a Markov chain Monte Carlo (MCMC) methodology with a novel MCMC algorithm which can mix over a complicated state space. The stability of the new estimator is demonstrated through comparison with the Monte Carlo samples and other methods, as well as MCMC convergence diagnostics. The analysis of African gorilla data from two homologous loci reveals discordant divergence times between loci, and this discrepancy is explained by male-mediated gene flows until the end of the last ice age.

Higher high-density lipoprotein cholesterol levels increased vertebral osteoporotic fracture, but reduced hip fracture in men based on the National Health Insurance Service-National Health Screening Cohort.

BACKGROUND CONTEXT: Lipids are currently known to play an important role in bone metabolism. PURPOSE: This study aimed to investigate the effect of high-density lipoprotein cholesterol (HDL-C) on osteoporotic fractures beyond its beneficial effects on the cardiovascular system. STUDY DESIGN/SETTING: This was a retrospective, observational study that used data from the National Health Insurance Service-Health Screening cohort database. PATIENT SAMPLE: This study included 318,237 participants who were 50 years or older and with HDL-C levels of 10 to 200 mg/dL. OUTCOMES MEASURES: Physiologic measure-Diagnosis of osteoporotic fracture during the follow-up period. METHODS: The study participants were categorized into four quartiles according to baseline HDL-C levels. The Cox proportional hazards model was used to assess osteoporotic fracture risk according to HDL-C levels. RESULTS: After full adjustment and with the Q1 group as the reference group, estimates of hazard ratios (HRs; 95% confidence intervals [CIs]) for any osteoporotic fracture in men were 1.03 (0.94-1.12), 1.02 (0.93-1.11), and 1.07 (0.98-1.18) for the Q2, Q3, and Q4 groups, respectively. After classifying osteoporotic fractures according to the body location, the fully adjusted HRs for vertebral and hip fractures in the men's Q4 groups were 1.16 (1.02-1.31) and 0.74 (0.57-0.96), respectively. In women, fully adjusted HRs (95% CIs) of the female Q4 group for any osteoporotic, vertebral, and hip fractures were 1.03 (0.95-1.11), 0.96 (0.86-1.07), and 1.06 (0.80-1.41), respectively. CONCLUSION: In this study, HDL-C levels were positively associated with vertebral fractures in both men and women but inversely related to hip fractures in men. Therefore, monitoring the lipid profiles of patients with osteoporosis may be beneficial for the prevention of osteoporotic fractures.

A prediction model for childhood obesity risk using the machine learning method: a panel study on Korean children.

Young children are increasingly exposed to an obesogenic environment through increased intake of processed food and decreased physical activity. Mothers' perceptions of obesity and parenting styles influence children's abilities to maintain a healthy weight. This study developed a prediction model for childhood obesity in 10-year-olds, and identify relevant risk factors using a machine learning method. Data on 1185 children and their mothers were obtained from the Korean National Panel Study. A prediction model for obesity was developed based on ten factors related to children (gender, eating habits, activity, and previous body mass index) and their mothers (education level, self-esteem, and body mass index). These factors were selected based on the least absolute shrinkage and selection operator. The prediction model was validated with an Area Under the Receiver Operator Characteristic Curve of 0.82 and an accuracy of 76%. Other than body mass index for both children and mothers, significant risk factors for childhood obesity were less physical activity among children and higher self-esteem among mothers. This study adds new evidence demonstrating that maternal self-esteem is related to children's body mass index. Future studies are needed to develop effective strategies for screening young children at risk for obesity, along with their mothers.

Association between polycythemia and risk of ischemic stroke in males based on the national health insurance service-health screening cohort.

BACKGROUND: Polycythemia, a state in which the hematocrit or hemoglobin (Hb) concentration in the peripheral blood increases, is associated with several thrombosis-related diseases, of which cerebral infarction is relatively common. This study aimed to investigate the association between ischemic stroke and polycythemia, as a potential risk factor. RESEARCH DESIGN AND METHODS: This study included men who had undergone national health checkups between 2002 and 2003; the data were extracted from the Korean National Health Insurance Service-Health Screening database. The primary outcome was the risk ischemic stroke; adjusted hazard ratios (HRs) with 95% confidence intervals (CIs) for ischemic stroke were calculated using Cox proportional hazards regression models. RESULTS: In total, 207,737 male participants aged 40-79 years were included in this study. At the baseline, 13972 (6.7%) participants met the polycythemia criteria (Hb >16.5 g/dL). During the study period, 897 and 12,440 cases of ischemic stroke occurred in the polycythemia and normocythemia (13.0 g/dL ≤ Hb ≤16.5 g/dL) groups, respectively. Compared with the normocythemia group, the polycythemia group showed an adjusted HR (95% CI) for ischemic stroke of 1.12 (1.04-1.20). CONCLUSIONS: The risk of ischemic stroke was higher in participants with polycythemia than in those with normocythemia.

Effect of metabolic health and obesity on all-cause death and CVD incidence in Korean adults: a retrospective cohort study.

This study aimed to investigate the risk of all-cause mortality and incidence of CVD according to metabolic health and body mass index (BMI) in Korean adults. This study was retrospectively designed using the National Health Insurance Service-National Health Screening Cohort data. Participants were divided into six groups according to two category of metabolic syndrome and three categories of BMI. Hazard ratios (HRs) and 95% confidence intervals (CIs) for the composite outcome (all-cause mortality and incidence of CVDs) were estimated using multivariable Cox proportional hazards regression models. 151,706 participants aged ≥ 40 years were enrolled; median follow-up period was 9.7 years in the study. Compared to metabolically healthy normal weight, the fully adjusted HRs (95% CIs) of metabolically healthy overweight, metabolically healthy obese, metabolically unhealthy normal weight, metabolically unhealthy overweight, and metabolically unhealthy obese for composite outcome were 1.07 (1.03-1.12), 1.12 (1.07-1.17), 1.33 (1.25-1.41), 1.28 (1.22-1.34), and 1.31 (1.26-1.37), respectively, in men, and 1.10 (1.05-1.16), 1.22 (1.16-1.29), 1.34 (1.26-1.43), 1.27 (1.19-1.34), and, 1.40 (1.34-1.47), respectively, in women. High BMI and metabolic unhealthiness were associated with an increased risk on the composite of all-cause mortality and incidence of CVD in both sexes.

Association Between Migraine and Ischemic Cardio-Cerebrovascular Disease (CCVD) and Effects of Triptans and Ergotamine on the Risk of Ischemic CCVD in Patients with Migraine in the Korean NHIS-HEALS Cohort.

BACKGROUND AND OBJECTIVES: Triptans and ergotamine are commonly used to treat migraine, a risk factor for ischemic stroke. This study aimed to investigate the association between migraine and ischemic cardio-cerebrovascular disease (CCVD). Further analyses were performed to examine whether symptom-relieving treatment of migraine with triptans and ergotamine reduces ischemic CCVD in migraineurs. METHODS: Participants from the Korean NHIS-HEALS cohort database were divided into patients reporting headache without migraine (HA), migraineurs who received at least one prescription for triptans or ergotamine (TE), and migraineurs who were prescribed neither triptans nor ergotamine (NTNE). Ischemic CCVDs comprised ischemic cerebrovascular diseases and cardiovascular diseases. Using cox proportional hazards regression models, primary and secondary analysis for risk of ischemic CCVDs was compared. RESULTS: Among 62,272 patients diagnosed with migraine or HA, men with migraine or HA numbered 14,747 and 8935, respectively, while the numbers of women were 27,836 and 10,754, respectively. The median follow-up was 6.65 years. The overall incidence rate of CCVDs was 4728/38,590 (12.25%) in females and 3158/23,682 (13.33%) in males. Compared with the HA group, the hazard ratios (HRs) (95% CIs) of the TE and NTNE groups for ischemic CCVDs were 1.18 (1.01-1.39) and 1.39 (1.28-1.50), respectively, in males, and 1.22 (1.09-1.37) and 1.53 (1.42-1.65), respectively, in females, after full adjustment for confounding variables. Compared with the NTNE group, the HRs (95% CIs) of the TE group for ischemic CCVDs were 0.86 (0.73-0.999) in males and 0.80 (0.72-0.88) in females. CONCLUSIONS: Migraine increased the risk of ischemic CCVDs in both sexes, and migraineurs treated with triptans and ergotamine were at lower risk of ischemic CCVDs than migraineurs who did not take those medications, especially in women.

Metformin Use May Increase Risk of Pancreatic Cancer in Diabetic Women: An Analysis of the Korean National Health Insurance Service-National Health Screening Cohort Database.

BACKGROUND: In addition to its antidiabetic effects, metformin has pleiotropic effects, such as the inhibition of carcinogenesis. This study aimed to investigate the association between metformin use and pancreatic cancer risk in the Korean National Health Insurance Service (NHIS)-National Health Screening Cohort (HEALS). METHODS: Of the individuals in the Korean NHIS-HEALS, 29,271 men and 19,091 women were included in the final analysis after propensity score matching based on age, body mass index, and smoking status. The study population was categorized into three groups: metformin non-users with diabetes mellitus (DM), metformin users with DM, and non-diabetic users. A Cox proportional hazards regression model was used to examine the association between metformin use and pancreatic cancer. RESULTS: The median follow-up period was 12.9 years. The estimated pancreatic cancer incidence was highest in metformin users with DM, regardless of sex (P<0.001), and lowest in non-diabetic men and female metformin non-users (P=0.053). The hazard ratios (95% confidence interval) for pancreatic cancer incidence in metformin users and non-diabetic individuals were 1.116 (0.648-1.923) and 0.447 (0.259-0.771) in men and 2.769 (1.003-7.642) and 1.451 (0.529-3.984) in women, respectively, after full adjustment. CONCLUSION: Women with diabetes using metformin are at a higher risk of pancreatic cancer than women with diabetes not using metformin. Meanwhile, men with DM using metformin have a similar risk of pancreatic cancer as men with DM not using metformin.