Impact of linkage level on inferences from big data analyses in health and medical research: an empirical study

BACKGROUND

Linkage errors that occur according to linkage levels can adversely affect the accuracy and reliability of analysis results. This study aimed to identify the differences in results according to personally identifiable information linkage level, sample size, and analysis methods through empirical analysis.

METHODS

The difference between the results of linkage in directly identifiable information (DII) and indirectly identifiable information (III) linkage levels was set as III linkage based on name, date of birth, and sex and DII linkage based on resident registration number. The datasets linked at each level were named as databaseIII (DBIII) and databaseDII (DBDII), respectively. Considering the analysis results of the DII-linked dataset as the gold standard, descriptive statistics, group comparison, incidence estimation, treatment effect, and moderation effect analysis results were assessed.

RESULTS

The linkage rates for DBDII and DBIII were 71.1% and 99.7%, respectively. Regarding descriptive statistics and group comparison analysis, the difference in effect in most cases was "none" to "very little." With respect to cervical cancer that had a relatively small sample size, analysis of DBIII resulted in an underestimation of the incidence in the control group and an overestimation of the incidence in the treatment group (hazard ratio [HR] = 2.62 [95% confidence interval (CI): 1.63-4.23] in DBIII vs. 1.80 [95% CI: 1.18-2.73] in DBDII). Regarding prostate cancer, there was a conflicting tendency with the treatment effect being over or underestimated according to the surveillance, epidemiology, and end results summary staging (HR = 2.27 [95% CI: 1.91-2.70] in DBIII vs. 1.92 [95% CI: 1.70-2.17] in DBDII for the localized stage; HR = 1.80 [95% CI: 1.37-2.36] in DBIII vs. 2.05 [95% CI: 1.67-2.52] in DBDII for the regional stage).

CONCLUSIONS

To prevent distortion of the analyses results in health and medical research, it is important to check that the patient population and sample size by each factor of interest (FOI) are sufficient when different data are linked using DBDII. In cases involving a rare disease or with a small sample size for FOI, there is a high likelihood that a DII linkage is unavoidable.

A Prediction Model for Osteoporosis Risk Using a Machine-Learning Approach and Its Validation in a Large Cohort

BACKGROUND

Osteoporosis develops in the elderly due to decreased bone mineral density (BMD), potentially increasing bone fracture risk. However, the BMD is not regularly measured in a clinical setting. This study aimed to develop a good prediction model for the osteoporosis risk using a machine learning (ML) approach in adults over 40 years in the Ansan/Anseong cohort and the association of predicted osteoporosis risk with a fracture in the Health Examinees (HEXA) cohort.

METHODS

The 109 demographic, anthropometric, biochemical, genetic, nutrient, and lifestyle variables of 8,842 participants were manually selected in an Ansan/Anseong cohort and included in the ML algorithm. The polygenic risk score (PRS) of osteoporosis was generated with a genome-wide association study and added for the genetic impact of osteoporosis. Osteoporosis was defined with < -2.5 T scores of the tibia or radius compared to people in their 20s-30s. They were divided randomly into the training (n = 7,074) and test (n = 1,768) sets-Pearson's correlation between the predicted osteoporosis risk and fracture in the HEXA cohort.

RESULTS

XGBoost, deep neural network, and random forest generated the prediction model with a high area under the curve (AUC, 0.86) of the receiver operating characteristic (ROC) with 10, 15, and 20 features; the prediction model by XGBoost had the highest AUC of ROC, high accuracy and k-fold values (> 0.85) in 15 features among seven ML approaches. The model included the genetic factor, genders, number of children and breastfed children, age, residence area, education, seasons to measure, height, smoking status, hormone replacement therapy, serum albumin, hip circumferences, vitamin B6 intake, and body weight. The prediction models for women alone were similar to those for both genders, with lower accuracy. When the prediction model was applied to the HEXA study, the correlation between the fracture incidence and predicted osteoporosis risk was significant but weak (r = 0.173, P < 0.001).

CONCLUSION

The prediction model for osteoporosis risk generated by XGBoost can be applied to estimate osteoporosis risk. The biomarkers can be considered for enhancing the prevention, detection, and early therapy of osteoporosis risk in Asians.

Vitamin D intake and bone mineral density in Korean adults: analysis of the 2009-2011 Korea National Health and Nutrition Examination Survey

BACKGROUND/OBJECTIVES

The prevalence of vitamin D deficiency in Koreans is quite high; however, until recently, Korean National Health and Nutrition Survey (KNHANES) had not analyzed the vitamin D intake among Koreans. Additionally, the Korean Dietary Reference Intake for vitamin D was established based on insufficient evidence. Therefore, we investigated vitamin D intake and its relationship with bone mineral density (BMD) in Korean adults using the combined data from the 2009-2011 KNHANES.

MATERIALS AND METHODS

This study was conducted in 11,949 healthy adults. Vitamin D intake was assessed using a 24-h recall method, and the BMD was measured using dual-energy X-ray absorptiometry.

RESULTS

The prevalence of vitamin D deficiency (< 20 ng/mL) was 64% in men and 77% in women. In women aged ≥ 50 yrs and men aged < 50 yrs, there was a significant positive correlation between vitamin D intake and serum 25-hydroxyvitamin D level after sun exposure adjustment. The BMD of postmenopausal women aged ≥ 50 yrs with a vitamin D intake of 5 µg/day or more was significantly higher than that of women with intake less than 5 µg/day. After adjusting for age, energy, and calcium intake, the vitamin D intake of the osteoporotic group was significantly lower than that of the osteopenia group in women.

CONCLUSIONS

Since the relationship between vitamin D intake and BMD was observed in women aged ≥ 50 yrs, further research is needed to clarify these findings using cohort or randomized controlled trials.

Musculoskeletal Health of the Adults Over 50 Years of Age in Relation to Antioxidant Vitamin Intakes

As the proportion of the elderly population increases rapidly, interest in musculoskeletal health is also emerging. Here, we investigated how antioxidant vitamin intake and musculoskeletal health are related. Adults aged 50 to 80 years with a body mass index (BMI) of 18.5 to 27.0 kg/m² were included. Bone mineral density (BMD), lean mass (LM), appendicular skeletal muscle mass index (ASMI) were measured using dual-energy X-ray absorptiometry (DXA), and the grip strength and knee extension using hand dynamometer. Nutrient intakes were measured using a 24-hour recall questionnaire. A total of 153 adults (44 men and 109 women) participated in this study. A partial correlation analysis showed a significant positive relationship between vitamin E and BMD and between vitamin C and LM/Height. Participants were classified into three groups according to whether their vitamin E and C intake met the recommended intake for Dietary Reference Intakes for Koreans (KDRIs). The prevalence of having low T-score (< −1.0) and low ASMI (< 7.0 for men and < 5.4 for women) was 51.3% and 15.4% in the group with vitamins C and E intakes below KDRIs. After adjusting for sex, smoking status and energy, protein, vitamin D, and calcium intake, the group with vitamins C and E both below the KDRIs displayed a significantly lower BMD at all test sites and LM/Height compared with vitamin C and/or E intake above the KDRIs groups. We conclude that sufficient intake of vitamin E and C is important for maintaining BMD and lean mass in Korean adults over 50 years of age.

The Trend in the Sales of Menopausal Hormone and Other Osteoporosis Medications in South Korea from 2016 to 2019

Background

Increasing geriatric population, osteoporosis prevalence, and interest in bone health are key contributing factors for the growth of osteoporosis medication markets. Thus, this study evaluated changes in the menopausal hormone and other osteoporosis medication markets from 2016 to 2019.

Methods

This study’s dataset was obtained from the International Marketing Services of IQVIA Inc. in South Korea. The sales of medications for osteoporosis treatment with menopausal hormones were evaluated for drug sales for osteoporosis treatment from 2016 to 2019.

Results

The results showed that the tissue-selective estrogen complex (TSEC) sales had increased annually while the estrogen-progesterone therapy (EPT) sales had decreased. Excluding menopausal hormones, bisphosphonates were the most widely sold medications for osteoporosis treatment. Among the bisphosphonate medications, sales of ibandronate and zoledronate increased annually, while alendronate and risedronate decreased. Teriparatide also showed increasing sales. A rapid rise was noted in the sales of denosumab.

Conclusions

While the sales of TSEC, injectable bisphosphonates, and denosumab have increased annually, the sales of EPT, estrogen therapy, oral bisphosphonates have not increased, as reflected in hormone therapy and osteoporosis medication market trends. This study showed the recent trends in hormone therapy and the osteoporosis medication market from 2016 to 2019 in South Korea.

Cardiorespiratory Fitness is Inversely Associated with Risk of Low Bone Mineral Density in Older Korean Men

Little is known regarding the association between physical fitness and bone health in older Korean men. This study investigated the relationship between estimated cardiorespiratory fitness (eCRF) and bone mineral density (BMD). This cross-sectional study included 2715 Korean men aged 50 years and older selected from those who participated in the 2008–2011 Korea National Health and Nutritional Examination and Survey. eCRF was obtained using a sex-specific algorithm developed on the basis of age, body mass index, resting heart rate, and physical activity and classified into low, middle, and high categories. Femoral neck BMD was assessed by dual X-ray absorptiometry. Odds ratios (OR) and 95% confidence intervals (CI) for osteopenia, osteoporosis, and low BMD were calculated for eCRF categories in models fully adjusted for age, waist circumference, education, income, smoking, heavy alcohol intake, serum vitamin D, serum parathyroid hormone, and dietary intake of energy, protein, calcium, and vitamins A and C. Overall, eCRF levels were positively associated with BMD and negatively with prevalence of osteopenia, osteoporosis, and low BMD. Logistic regression showed inverse trends in the risks of osteopenia (high vs. low: OR = 0.692; 95% CI, 0.328–0.517; p = 0.049) and low BMD (high vs. low: OR = 0.669; 95% CI, 0.497–0.966; p = 0.029) by eCRF category in models fully adjusted for all the measured covariates. The current findings suggest that maintaining high eCRF via regular physical activity may contribute to attenuation of age-related loss of BMD and decreased risk for low BMD in older Korean men.

Anti-fracture Efficacy of Monthly Risedronate Compared with That of Weekly Risedronate in Postmenopausal Korean Women with Osteoporosis: A Nationwide Cohort Study

Background

Intermittent dosing regimens for oral risedronate (once-monthly and once-weekly) were developed for patient convenience. While several studies have reported the anti-fracture efficacy of weekly dosing, few have assessed monthly dosing. The lower efficacy of monthly dosing has been previously suggested. The aim of this study was to compare the anti-fracture efficacy of monthly and weekly dosing.

Methods

We obtained information from the Korea National Health Insurance Service database from 2012 to 2017 of Korean women of ≥50 years of age who used weekly or monthly risedronate. We compared the time of occurrence of the first osteoporotic fracture after the first prescription of risedronate. Using a Cox proportional model, we assessed incidence rate ratios (IRRs) with 95% confidence intervals (CIs) for fractures at any site, and the hip, vertebral, and non-vertebral sites between both regimens. Propensity score weighting was used to balance the treatment groups.

Results

The study populations were distributed according to dosing frequency (monthly, 27,329; weekly, 47,652). There was no significant difference in the incidence rate of new fractures in any site (IRR, 1.008; 95% CI,0.963–1.055; P=0.737), hip (IRR, 0.999; 95% CI, 0.769–1.298; P=0.996), vertebral (IRR, 0.962; 95% CI, 0.890–1.040; P=0.330), or non-vertebral (1.022; 95% CI, 0.968–1.078; P=0.439) sites between monthly and weekly risedronate.

Conclusion

The anti-fracture efficacy at any site and the examined individual sites was similar for the monthly and weekly risedronate regimens. Large-scale randomized controlled trials are required for confirmation.

CXCL2 attenuates osteoblast differentiation by inhibiting the ERK1/2 signaling pathway

The C-X-C motif chemokine ligand 2 (CXCL2), a member of the CXC receptor ligand family, is involved in various immune and inflammatory processes, but its effect(s) on bone formation have not yet been reported. We report here that CXCL2 is enriched in bone marrow and show abundant expression of CXCL2 in osteoblasts of osteoporotic mice. CXCL2 neutralization within the bone marrow by using antibody alleviated bone loss in mice, indicating a negative role of CXCL2 in bone formation. In line with this, CXCL2 overexpression attenuated proliferation, as well as differentiation, of osteoblasts in vitro By contrast, CXCL2 downregulation promoted osteoblast expansion and differentiation. Mechanistically, CXCL2 inhibits the ERK1/2 (MAPK3/1) signaling pathway in osteoblasts. Activation of ERK1/2 abolishes the inhibitory effect of CXCL2 in osteoblasts, whereas inactivation of ERK1/2 reverses the osteogenic role of CXCL2 inhibition. These results show that CXCL2 attenuates osteoblast differentiation through inhibition of the ERK1/2 signaling pathway. We demonstrate here that CXCL2 is a negative regulator of bone formation and clarify the responsible mechanisms. Therefore, pharmaceutical coordination of CXCL2 and of the pathways through which it is regulated in osteoblasts might be beneficial regarding bone formation.

Hemoglobin levels and low bone mineral density in non-anemic older adults: Secondary analysis of the Korean National Health and Nutrition Examination Survey

BACKGROUND

Although, previous studies have reported a positive association between hemoglobin levels and bone mineral density (BMD), the majority of the studies were limited in patients with chronic hypoxemic conditions and findings concerning the association among non-anemic populations are inconclusive. We aimed to examine the association between hemoglobin levels and BMD in non-anemic healthy adults.

METHODS

This cross-sectional study included 3626 non-anemic men and women aged ≥ 60 years who participated in the Korean National Health and Nutrition Examination Survey (KNHANES). The BMD of the lumbar spine and both femurs was measured by dual-energy X-ray absorptiometry (DXA). Participants with T-score for BMD < -1.0 SD were defined as having low BMD. The odds ratios (ORs) and 95% confidence intervals (CIs) for low BMD were calculated using multiple logistic regression analyses across sex-specific hemoglobin quartiles.

RESULTS

The prevalence of low BMD gradually decreased in accordance with hemoglobin quartiles in both sexes. Compared with the group in the lowest quartile, the OR (95% CI) for low BMD in the lumbar spine was 0.78 (0.54-0.93) for men and 0.67 (0.50-0.93) for women after adjusting for age, BMI, tobacco smoking, alcohol intake, physical activity, walking difficulty, household income, total calorie intake, calcium intake, iron intake, 25(OH)D, alkaline phosphatase, and parathyroid hormone levels. However, these positive associations were not found in femur after adjusting for the same co-variables.

CONCLUSIONS

Hemoglobin levels were inversely associated with low BMD in lumbar spine among non-anemic adults.

Biomechanical role of osteoporosis affects the incidence of adjacent segment disease after percutaneous transforaminal endoscopic discectomy

Study design

Variation in the biomechanical characteristics of intervertebral discs adjacent to the segment disc after undergoing percutaneous transforaminal endoscopic discectomy (PTED) in models with normal and abnormal bone mineral density (BMD) was estimated using the finite element method.

Objective

The study investigated the change in the incidence of adjacent segment disease (ASD) after PTED in patients without and with osteoporosis.

Backgrounds

PTED has been widely used for treating lumbar disc herniation (LDH); changes in BMD will affect biomechanical characteristics, possibly leading to changes in the incidence of ASD after PTED. However, this issue remains largely unclear.

Methods

A non-linear, lumbosacral finite element model was reconstructed based on imaging data and validated using compared values computed by the current model from published and well-validated, in vitro biomechanical experiment studies. Corresponding PTED models with normal and abnormal BMDs were also reconstructed. Shear and von Mises stresses on the annulus fibrosis, the von Mises stress on the endplates in L5–S1 segment discs, and the total deformation of current lumbosacral models were computed in different body positions by changing loading conditions, including flexion, extension, left and right lateral bending, and axial rotation.

Results

In most loading conditions, biomechanical characteristics of the lumbosacral segment discs with normal BMDs after PTED slightly increased. However, in the PTED model with osteoporosis, most of the biomechanical characteristics dramatically increased.

Conclusion

Osteoporosis leads to the deterioration of biomechanical characteristics in the adjacent segment disc after PTED; this variation may also result in an increase in the incidence of ASD. However, further studies on the interactions between pathological changes are warranted.

Association between Inflammatory Potential of Diet and Bone-Mineral Density in Korean Postmenopausal Women: Data from Fourth and Fifth Korea National Health and Nutrition Examination Surveys

Post menopause is considered a critical period for bone-mass loss. Impaired bone metabolism during this phase can increase the risk of fractures in old age. Inflammation is a risk factor for bone health, and diet is a potential source of inflammation. However, few studies have examined the association between the dietary inflammatory index (DII®) and bone-mineral density (BMD) in postmenopausal women in Korea. The objective of this study was to determine, by means of a Korean cross-sectional investigation, whether higher DII scores are associated with decreased BMD in postmenopausal women. To that end, the raw data from the fourth and fifth Korea National Health and Nutrition Examination Surveys (KNHANES, 2009-2011) for 2778 postmenopausal women aged over 50 years were analyzed. The subjects' BMD were measured by dual-energy x-ray absorptiometry, and their DII® scores were calculated from a single 24-h dietary recall. Further, the participants were classified into three groups according to DII® score. Women with more pro-inflammatory diets (i.e., those in the highest tertile of DII®) had significantly lower BMD in the femoral neck as compared with women in the lowest tertile (p for trend <0.05) after adjustment for age, body-mass index (BMI), household income, education status, smoking habits, physical activity, total calcium intake, female-hormone use, age at menopause, and blood vitamin D levels. Multiple logistic regression analyses revealed that the odds ratio (OR) of total femur osteopenia/osteoporosis was higher in women in the highest tertile of DII® than in those in the lowest (OR 1.27, 95% CI 1.00-1.62, p for trend < 0.05). This study established that more pro-inflammatory diets might be associated with lower BMD in postmenopausal Korean women.

Age-related Reference Intervals for Total Collagen-I-N-terminal Propeptide in Healthy Korean Population

Background

Procollagen type I N-terminal propeptide (PINP) is one of the most clinically useful bone formation biomarkers. Therefore, the purpose of this study was to independently evaluate the performance of automated total PINP assay and established age- and gender-specific reference intervals for PINP in healthy Korean population.

Methods

The imprecision, linearity, and detection capability of Elecsys total PINP assay was determined and reference interval was established using 599 serums from Korean population with normal bone mineral densities based on bone densitometry. Age groups were divided into 20s, 30s, 40s, 50s, 60s and over.

Results

Elecsys total PINP had excellent performance in imprecision, linearity, and detection capability. When partitioning age groups in Korean male and female populations, there was significant difference in total PINP between different age groups. In male populations, PINP level was decreased with increasing age, then it remained steady after middle-age. In female populations, there was a decreasing tendency similar to that in the male population with a sharp increase in the 50 to 59 age group.

Conclusions

Elecsys total PINP assay showed precise and reliable performance in our study. We established age-related PINP reference intervals for Korean male and female population with normal bone mineral densities.

Prevalence and Fracture Risk of Osteoporosis in Patients with Rheumatoid Arthritis: A Multicenter Comparative Study of the FRAX and WHO Criteria

(1) Background: We evaluated the prevalence and fracture risk of osteoporosis in patients with rheumatoid arthritis (RA), and compared the fracture risk assessment tool (FRAX) criteria and bone mineral density (BMD) criteria established by the World Health Organization (WHO). (2) Methods: This retrospective cross-sectional study, which included 479 RA patients in 5 hospitals, was conducted between January 2012 and December 2016. The FRAX criteria for high-risk osteoporotic fractures were calculated including and excluding the BMD values, respectively. The definition of high risk for fracture by FRAX criteria and BMD criteria by WHO was 10-year probability of ≥ 20% for major osteoporotic fracture or ≥ 3% for hip fracture, and T score ≤ −2.5 or Z score ≤ −2.0, respectively. (3) Results: The mean age was 61.7 ± 11.9 years. The study included 426 female patients (88.9%), 353 (82.9%) of whom were postmenopausal. Osteoporotic fractures were detected in 81 (16.9%) patients. The numbers of candidates for pharmacological intervention using the FRAX criteria with and without BMD and the WHO criteria were 226 (47.2%), 292 (61%), and 160 (33.4%), respectively. Only 69.2%–77% of the patients in the high-risk group using the FRAX criteria were receiving osteoporosis treatments. The following were significant using the WHO criteria: female (OR 3.55, 95% CI 1.46–8.63), age (OR 1.1, 95% CI 1.08–1.13), and BMI (OR 0.8, 95% CI 0.75–0.87). Glucocorticoid dose (OR 1.09, 95% CI 1.01–1.17), age (OR 1.09, 95% CI 1.06–1.12), and disease duration (OR 1.01, 95% CI 1–1.01) were independent risk factors for fracture. (4) Conclusions: The proportion of RA patients with a high risk of osteoporotic fractures was 33.4%–61%. Only 69.2%–77% of candidate patients were receiving osteoporotic treatments while applying FRAX criteria. Independent risk factors for osteoporotic fractures in RA patients were age, the dose of glucocorticoid, and disease duration.

Effect of improved medication adherence on health care costs in osteoporosis patients

Osteoporosis is a chronic disease that requires continuous health care spending for pharmacotherapy and examinations. Osteoporotic fractures are a major economic burden. However, little is known about the economic effects of osteoporosis and osteoporotic fractures in Korea.The purpose of this study was to determine the predictors of osteoporosis-related health care costs and to evaluate the economic effects of fracture prevention through medication adherence among osteoporosis patients.Using the Korea National Health Insurance Claims Database (KNHICD), we identified osteoporosis patients aged 50 years and older from 2011 to 2012. Annual health care costs of osteoporosis were analyzed from the insurer's perspective and compared between patients with fractures and those without fractures. Adherents were defined as patients with a medication possession ratio of ≥80%. A generalized linear model (GLM) was used to estimate the predictors of osteoporosis-related health care costs.The major predictors of osteoporosis-related health care costs were age, medication adherence, and the occurrence of fractures (P < .001). The proportion of fractures among non-adherents was approximately 1.1 times the proportion among adherents. Health care costs per patient with fractures were 3.8 times the costs per patient without fractures. Patients with fractures had higher health care costs due to hospitalization and outpatient costs but lower pharmacy costs than non-adherents. We estimated that about $5 million of health insurance expenses could be saved annually if all non-adherents became adherents.Improved osteoporosis medication adherence can reduce osteoporosis-related health care costs by preventing fractures. Persistent pharmacotherapy for osteoporosis is necessary to prevent osteoporotic fractures and to reduce osteoporosis-related health care costs.