Search

AIM This study assessed the incidence of epilepsy in preterm infants and those small for gestational age (SGA) at term and identified risk factors associated with higher epilepsy incidence in these children. METHODS We enrolled children (from 2000 to 2010) who were premature (n = 21 474) or SGA (n = 2206); we then included a matched control cohort (n = 94 720). Cox regression was used to assess the epilepsy risk in preterm and SGA children. To determine the associated factors for epilepsy, the preterm and SGA infants were divided into six groups according to the common complications related to brain development and were separated into three subgroups based on birthweight (BW). RESULTS The cumulative incidence of epilepsy was significantly higher in preterm or SGA children than in the control group. The overall incidence densities (per 1000 person-years) of epilepsy were: 0.37 in the control, 2.96 in the preterm, 2.90 in the SGA, 15.9 in the preterm with cerebral haemorrhage, 14.6 in the SGA with cerebral haemorrhage, 6.92 in the preterm with asphyxia, 3.82 in the SGA with asphyxia, 14.3 in the preterm with congenital brain anomalies, and 25.4 in the SGA with congenital brain anomalies cohorts. Infants with BW

Benign Prostate Hyperplasia (BPH) has been associated with prostate cancer prevalent among men after 50 years of age, however, it is unclear whether the antidiabetic drug, metformin, can reduce prostate cancer for men with BPH. The insurance claims data of men aged 50 years or older, with both type 2 diabetes mellitus (T2DM) and BPH diagnosed from 1997 to 2007 were analyzed. Individuals were followed up for at least 5 years. We identified 2906 and 2906 patients as the metformin cohort and nonmetformin cohort, respectively. The Cox method analysis showed that the metformin cohort had an adjusted hazard ratio (aHR) of 0.69 (95% confidence interval [CI] = 0.49‐0.96, P = 0.0298) for prostate cancer, compared to the nonmetformin cohort after controlling for age, traditional Chinese medicine (TCM) use, prostate specific antigen, and Charlson comorbidity index. Patients using TCM for BPH (per 6 months) also had an aHR of 0.41 (95% CI = 0.24‐0.69; P = 0.0009). In conclusion, both metformin medication and TCM use could be associated with reduced risk of prostate cancer for men with BPH and diabetes.

AIM We used insurance claims data of Taiwan to compare the risk of non-traumatic lower extremity amputation between haemodialysis (HD) and peritoneal dialysis (PD) patients. METHODS We identified 77 669 HD patients and 10 035 PD patients without prior amputation from 2000 to 2010. Incidence rates and hazard ratios (HRs) of lower extremity amputation, and subsequent 30-day mortality after amputation were evaluated up to 31 December 2011. RESULTS There were 2427 and 216 patients undergoing lower extremity amputation during follow-up in the HD and PD groups with incidence rates of 8.35 and 5.79 per 1000 person-years, respectively. Compared with the HD group, the overall adjusted HR of lower extremity amputation for the PD group was 1.27 (95% CI = 1.10-1.46). The impact of diabetes status on the risk of lower extremity amputation interacted with dialysis modality significantly (P

Background: Studies have reported an association between asthma and oral diseases, including periodontal diseases. The aim of this retrospective study is to investigate risk of periodontal diseases for patients with asthma.Methods: Using the claims data of National Health Insurance of Taiwan and patients without a history of periodontal diseases, 19,206 asthmatic patients, who were newly diagnosed from 2000 through 2010, were identified. For each case, four comparison individuals without history of asthma and periodontal disease were randomly selected from the general population and frequency matched (categorical matched) by sex, age, and year of diagnosis (n = 76,824). Both cohorts were followed to the end of 2011 to monitor occurrence of periodontal diseases. Adjusted hazard ratios (aHRs) of periodontal disease were estimated using Cox proportional hazards regression analysis.Results: Overall incidence of periodontal diseases was 1.18-fold greater in the asthma cohort than in the comparison cohort (P

Aim The study evaluated the risk of sudden sensorineural hearing loss (SSHR) in patients with end-stage renal disease (ESRD) undergoing dialysis METHODS: We identified 18 905 patients newly diagnosed with ESRD, comprising 6382 patients on peritoneal dialysis (PD) and 12 523 patients on hemodialysis (HD) using Taiwan's National Health Insurance claims data from 2000 to 2010. A total of 76 520 controls without kidney disease were randomly selected with frequency matched with the ESRD group by age and sex. We further established additional cohorts with 6375 patients on PD and 6375 patients on HD matched by their propensity scores. Cox proportional hazards models were used to examine the effect of ESRD and dialysis modality on the risk of sudden sensorineural hearing loss (SSHL) by the end of 2011. Results The incidences for SSHR were 0.49, 2.37, 2.95, and 2.12 per 1000 person-years in the control, ESRD (PD and HD), PD, and HD cohorts, respectively. The adjusted HRs of SSHR was 2.71, 3.60, and 2.33 for ESRD, PD, and HD groups (95% CI = 2.17-3.38, 2.71-4.78 and 1.81-2.99, respectively) compared with the control cohort. Patients on PD exhibited 1.53-fold (95% CI = 1.14-2.04) and 1.64-fold (95% CI = 1.19-2.27) higher SSHL risk than their age- and sex-matched and propensity score-matched patients on HD, respectively. Conclusion Patients with ESRD exhibit a higher risk of SSHL than people without kidney disease. The SSHL risk is greater for patients on PD than on HD.

Background Increasing evidence suggests that Helicobacter pylori infection (HPI) may have extragastric manifestations, including the respiratory system. This study investigated the role of HPI in increasing the subsequent risk of chronic obstructive pulmonary disease (COPD) in a nationwide population. Methods We conducted this retrospective cohort study using data from the Longitudinal Health Insurance Database, which is derived from the Taiwanese National Health Insurance Research Database. A total of 5941 adults who were newly diagnosed with HPI between 2005 and 2006 were selected. Healthy patients without HPI were selected from the general population and frequency matched as a ratio of 4:1, according to age, sex, and index years. Both cohorts were followed up from the index date to the end of 2011 to measure the incidence of COPD. Cox proportional hazard regression analysis was used to assess the hazard ratio (HR) of COPD between the HPI cohort and non-HPI cohorts. Results The overall HR of COPD was 1.84 (95% confidence intervals = 1.57–2.17) for the HPI cohort, compared with the non-HPI cohort, after adjusting for age, sex, and comorbidities. Although the incidence of COPD was substantially higher in the elderly participants (age, ≥ 65 years) than that in younger participants, the highest HR (4.05, 95% confidence intervals = 1.39–11.8) of COPD was observed in the youngest (age, 20–49 years) participants. Conclusion In this study, the patients with HPI exhibited a significantly higher risk of COPD than those without HPI did.

Aims Cholangiocarcinoma (CCA) is the second most common primary liver cancer in the world. Due to the lack of effective treatments, the survival rate of CCA is low and it is usually considered difficult to diagnose early. To date, no effective strategies for the prevention of CCA have been developed. Statins are cholesterol-lowering agents which possess pleiotropic properties and the use of statins may reduce cancer risk. The aim of the study was to investigate the effect of statin use on the risk of CCA. Methods We used nationwide insurance data to perform a case–control study including 3174 CCA patients diagnosed in 2002–2011 and 3174 propensity score matched controls. Odds ratios (ORs) and 95% confidence intervals (CI) were calculated to assess the association between CCA risk and statin use by type of statin and dose. Results Patients with CCA were slightly younger than controls with mean ages of 67.4 (SD 12.3) and 68.5 (SD 13.2) years (P = 0.001), respectively, and had less users of statins (22.7 vs. 26.5%, P

Aim The prevalence of end-stage renal disease in Taiwan is among the highest in the world. Treatment reimbursement for haemodialysis was capped in 1996 in order to contain costs. This study evaluated temporal changes in the costs and utilization of medical care and mortality in patients receiving haemodialysis following capped reimbursement. Methods Using insurance claims data in Taiwan between 1998 to 2009, we established eight annual subcohorts of patients with incident haemodialysis, increasing from 6099 in 1998 to 7745 in 2005. With a 4-year follow-up paradigm for each subcohort, we evaluated resources use and costs of medical services, as well as mortality trends. Results The annual mean cost for each haemodialysis patient increased from US $431 to $737 for emergency visits, US $9007 to $13 280 for hospitalizations and US $79 141 to $92 416 (16.8% increase) for total costs, from the initial to final subcohorts, respectively. Compared to the 1998 subcohort, the adjusted hazard ratio of deaths declined from 0.97 (95% CI 0.91 to 1.02) for the 1999 subcohort to 0.86 (95% CI 0.82 to 0.91) for the 2005 subcohort (P for trend

Background Chronic rhinosinusitis (CRS) is a chronic illness with a high prevalence worldwide. The purpose of this study was to investigate the characteristics of adjunctive traditional Chinese medicine (TCM) use in patients with CRS. Methods Data were retrieved from a dataset made available by the Taiwan National Health Insurance Research Database (NHIRD) containing the medical records of 1,000,000 randomly sampled beneficiaries insured in the year 2000. Patients newly diagnosed with CRS in the year 2000 were enrolled, and their medical records during the period 2000 to 2011 were collected and analyzed. Results A total of 14,806 CRS subjects were enrolled in this study: 10,512 were non-TCM users and 4294 were TCM users. TCM group had a significantly higher proportion of females, were younger, and a greater proportion lived in urban areas compared with the non-TCM group (all p < 0.0001). Ninety-seven percent (97%) of the TCM users received herbal remedies. The most common Chinese herbal formula used was Xin-Yi-Qing-Fei-Tang and the most commonly used single herb was Baizhi. The hazard ratio (HR) for functional endoscopic sinus surgery (FESS) was 0.17 for TCM users (95% confidence interval [CI], 0.15 to 0.20) compared with non-TCM users after adjusting for age, sex, urbanization, and comorbidities. Conclusion Our investigation found that 29% of CRS patients used TCM in addition to Western medical treatment. A lower proportion of patients in the TCM group underwent ESS compared with that in the non-TCM group. These findings may be of value in further studies evaluating the efficacy and safety of TCM use in CRS patients.

Background Studies have associated diabetes mellitus (DM) with the reduced risk of abdominal aortic aneurysm and thoracic aortic aneurysm and dissection. We used the national insurance data of Taiwan to examine these correlations for an Asian population. The association was also evaluated by DM severity. Methods We identified 160 391 patients with type 2 DM diagnosed from 1998 to 2008 and 646 710 comparison subjects without DM, frequency matched by diagnosis date, sex and age (mainly the elderly). The DM severity was partitioned into advanced and uncomplicated status according to DM-related comorbidities. Results By the end of 2010, the overall pooled incidence rate of thoracic aortic aneurysm and abdominal aortic aneurysm was 15% lower in the type 2 DM cohort than in non-DM cohort, with an adjusted hazard ratio of 0.64 [95% confidence interval (CI) 0.56–0.74] in the multivariable Cox model. Patients with advanced type 2 DM were significantly associated with reduced thoracic aortic aneurysm rupture and abdominal aortic aneurysm without rupture, with adjusted hazard ratios of 0.50 (95% CI 0.35–0.71) and 0.53 (95% CI 0.40–0.69), respectively. Uncomplicated type 2 DM was also associated with reduced abdominal aortic aneurysm without rapture (aHR = 0.58, 95% CI 0.45–0.74). Conclusions Our results demonstrate that patients with diabetes in this Asian population have reduced prevalence of thoracic and abdominal aortic aneurysms. The observed paradoxical inverse relationship between severity of DM and aortic aneurysms is clear. Further research is required to investigate the underlying mechanisms for the reduced risk of aortic aneurysms associated with diabetes. Copyright © 2014 John Wiley & Sons, Ltd.

Objectives/Hypothesis Tonsillectomy is related to increased risk of certain types of cancer. This study evaluates the possible association between tonsillectomy and cancer in Taiwan. Methods Data from the National Health Insurance system of Taiwan was used. A cohort study consisting of 997 patients with a new diagnosis of tonsillectomy from 2000 to 2005 and a comparison cohort of 3,988 subjects without tonsillectomy were used in the Poisson regression analysis to estimate the incidence rate ratios (IRRs) and 95% confidence intervals (CIs) of cancers. Results The incidence rate of all cancer types was higher in the tonsillectomy group than in the control group (4.28 vs. 2.97 per 1,000 person-years, respectively), with an IRR = 1.54 and 95% CI of 1.05 to 2.25, but the significant difference was limited to patients with more than 3 years follow-up. Site-specific analysis found no significant association between tonsillectomy and any individual cancer after adjusting for age, sex, hyperlipidemia, hypertension, and diabetes; however, tonsillectomy patients with more than 3 years follow-up had a marginally significantly higher risk for breast cancer development (adjusted IRR = 2.62; CI = 0.97–7.03). Conclusion Our study found that Taiwanese people with tonsillectomy have a significantly higher overall risk of developing cancer and a marginally higher risk of developing breast cancer when follow-up is longer than 3 years. Level of Evidence 2b. Laryngoscope, 125:134–139, 2015

It is well known that patients with diabetes have higher extent and severity of periodontitis, but the backward relationship is little investigated. The relationship between periodontitis needing dental surgery and subsequent type 2 diabetes mellitus (DMT2) in those individuals without diabetes was assessed.This is a retrospective cohort study using data from the national health insurance system of Taiwan. The periodontitis cohort involved 22,299 patients, excluding those with diabetes already or those diagnosed with diabetes within 1 year from baseline. Each study participant was randomly frequency matched by age, sex, and index year with one individual from the general population without periodontitis. Cox proportional hazards regression analysis was used to estimate the influence of periodontitis on the risk of diabetes.The mean follow-up period is 5.47 ± 3.54 years. Overall, the subsequent incidence of DMT2 was 1.24-fold higher in the periodontitis cohort than in the control cohort, with an adjusted hazard ratio of 1.19 (95% confidence interval = 1.10 to 1.29) after controlling for sex, age, and comorbidities.This is the largest nation-based study examining the risk of diabetes in Asian patients with periodontitis. Those patients with periodontitis needing dental surgery have increased risk of future diabetes within 2 years compared with those participants with periodontitis not requiring dental surgery.

Aim Attention-deficit-hyperactivity disorder (ADHD) is a disorder that is associated with accidental injuries. The aim of this study was to evaluate the relationship between ADHD and bone fracture in children. Method The study cohort comprised 3640 children (2874 males, 766 females; mean age 8y 5mo, SD 3y) with ADHD (International Classification of Diseases, Ninth Revision) who were matched to children without ADHD at a ratio of 1:4 (n=14 560; 11 496 males, 3064 females; mean age 8y 5mo, SD 3y). A Cox proportional hazard regression analysis was conducted to estimate how ADHD affected the risk of bone fracture. Results The incidence of fracture among the ADHD cohort was 197.67 per 10 000 person-years, and was 1.3-fold greater than in the comparison cohort (147.54 per 10 000 person-years). The risk in children with ADHD was higher than that in children without ADHD (p value for log-rank test < 0.001). After adjusting for potential confounding factors, the ADHD cohort was 1.32 times more likely to have bone fracture accidents than the comparison cohort (hazard ratio, 1.32; 95% confidence interval 1.17–1.49). Interpretation Children with ADHD have a higher risk of experiencing bone fracture accidents than do children without ADHD.

BACKGROUND: Few studies have investigated the relationship between injury location, mechanism and their association with complex regional pain syndrome (CRPS). We conducted a nationwide database survey to explore this issue. METHODS: This was a population-based case-control study. Five hundred and eighty-nine patients with at least one ambulatory visit or admission with a principal diagnosis of CRPS from 2004 to 2009 were selected. For each CRPS patient, ten age- and sex-matched non-CRPS subjects were randomly selected. The odds ratios (PLoS One. 2013;8:e57205) and 95% confidence intervals (95% CIs) of risk factors for CRPS were derived from multivariate logistic regression models. RESULTS: Injury was a risk factor for CRPS (OR, 2.96; 95% CI, 2.18 to 4.02) independent of age and sex. In adjusted models, open wound on the upper limbs (OR 1.25, 95% CI 1.02 to 1.54) conferred higher CRPS risk. Injury mechanisms including nerve and spinal cord injury (OR 2.42, 95% CI 1.44 to 4.08), muscle and joint sprain and strain (OR 1.69, 95% CI 1.40 to 2.03), superficial injury (OR 1.23, 95% CI 1.00 to 1.51), and contusion (OR 1.44, 95% CI 1.20 to 1.74), but not fracture, increased the risk of CRPS. CONCLUSION: Injury in the extremities rather than the trunk is an important risk factor for CRPS. Certain injury mechanisms confer higher risk of CRPS. This nationwide study demonstrated that injury increased CRPS nearly threefold. Open wound, sprain and strain, superficial injury, contusion, and nerve and spinal cord injury are main injury mechanisms. Injury in the extremities confers a higher risk of CRPS. Language: en

Objectives This study investigates the relationship between current morphine use and the risk of pulmonary embolism (PE) development in deep vein thrombosis (DVT) patients. Methods We conducted a population-based nested case-control retrospective analysis using the Longitudinal Health Insurance Database 2000 of Taiwan. A DVT cohort of 3668 patients with no history of PE from 1998 to 2010 and the other cohort of 174 patients who subsequently developed PE were evaluated. Morphine use was designated as ‘current’ if the prescription duration covered the index date or ended within 30 days before the index date. Logistic regression was used to estimate the odds ratios and 95% confidence intervals (CI), and the multivariable model was applied to control for age. Results Compared with non-morphine users, DVT patients who received morphine within 30 days of the index date had a 4.54-fold (95% CI = 2.30–8.97) chance of developing PE. The risk of PE development increased with an increase in cumulative dosage and in the average dosage of morphine. Conclusion The incidence of PE in DVT patients in Taiwan is associated with current morphine treatment (≤30 days) and is dependent on dosage.

Inflammatory processes, which kindle endothelial dysfunction and atherosclerosis, may facilitate the development of cardiovascular disease, including ischaemic stroke. Evident stroke risk factors may not be identified in up to 40% of stroke patients, especially in the younger population. Inflammation remains to be established as a stroke risk factor. In this study, it was assessed whether chronic osteomyelitis (COM), an infectious disease with chronic inflammation, increases stroke risk.A national insurance claim data set of 22 million enrollees in Taiwan was used to identify 18 509 patients with COM and 74 034 randomly selected age- and gender-matched controls for a follow-up period of 11 years starting 1 January 2000 and ending 31 December 2010. Stroke risk was analyzed using the Cox proportional hazards regression model.Comorbidities known to increase stroke risk, including hypertension, diabetes, hyperlipidemia, coronary heart disease and peripheral arterial disease, were more frequently noted in the COM group who had significantly greater stroke risk than the control cohort. Comparing only those without comorbidities, COM carried greater stroke risk than the control group [hazard ratio (HR) = 1.40, 95% confidence interval (CI) 1.22-1.62, P 0.001]. The younger population carried even greater risk (age45, HR = 2.73, 95% CI 1.71-4.35; age65, HR = 1.16, 95% CI 1.02-1.31).This is the first report linking COM to an increased risk of developing stroke. Results suggest that COM is a significant stroke risk factor and call for closer attention to this group of patients for more rigorous stroke prevention, especially in the younger age group.

Background and purpose The aim of this study was to explore whether hearing loss is associated with the risk of Parkinson's disease in the elderly in Taiwan. Methods Using claims data of the Taiwan National Health Insurance Program, 4976 patients (aged 65 years or older) with newly diagnosed hearing loss from 2000 to 2010 were identified and 19 904 subjects without hearing loss were randomly selected as comparisons, frequency matched by sex, age and index year of diagnosing hearing loss. The incidence of Parkinson's disease by the end of 2010 and the associated risk factors were investigated. Results The incidence of Parkinson's disease in the hearing loss group was 1.77-fold higher than that in the non-hearing-loss group (3.11 vs. 1.76 per 1000 person-years). After controlling for confounding factors, the adjusted hazard ratio (HR) of Parkinson's disease was 1.53 (95% CI 1.17, 1.99) for the hearing loss group compared with the non-hearing-loss group. Male sex (HR = 1.33, 95% CI 1.02, 1.74), age (for each year, HR = 1.06, 95% CI 1.04, 1.09), hypertension (HR = 1.70, 95% CI 1.26, 2.30) and cerebrovascular disease (HR = 1.78, 95% CI 1.37, 2.32) were also significantly associated with the risk of Parkinson's disease. Conclusions Hearing loss correlates with an increased risk of Parkinson's disease in the elderly. Further studies are needed to confirm whether hearing loss could be a non-motor feature of Parkinson's disease.

The association between herpes zoster and cardiovascular complications remains vague with limited study on the association between these two disorders. This study evaluated the risk of cardiovascular diseases in patients with herpes zoster. From insurance claims data of Taiwan, 19,483 patients with herpes zoster diagnosed in 1998–2008 and 77,932 subjects without herpes zoster were identified in this study. Both cohorts were followed up until the end of 2010 to measure the incidence of arrhythmia and coronary artery disease. The incidence rate ratio and adjusted hazard ratio (HR) of the cardiovascular complications with 95% confidence interval (CI) were estimated. The incidence of arrhythmia was 1.17-fold greater in the herpes zoster cohort than in the non-herpes zoster cohort (13.2 vs. 11.3 per 1,000 person-years), with an adjusted HR of 1.16 (P < 0.01). The coronary artery disease incidence in the herpes zoster cohort was 1.16-fold higher than that in the non-herpes zoster cohort (9.02 vs. 7.83 per 1,000 person-years), with an adjusted HR of 1.11 (P < 0.01). Over the stratified follow-up years, adjusted HRs were 1.22 (95% CI ¼ 1.12– 1.34) for arrhythmia and 1.14 (95% CI ¼ 1.02– 1.28) for coronary artery disease within 2 years after herpes zoster diagnosis. The risk measured for these disorders declined over time. Comorbidities of hypertension, diabetes, and hyperlipidemia also contributed to these cardiovascular disorders with greater extent. It is concluded that the contribution of herpes zoster to the risk of arrhythmia and cardiovascular diseases is less strong than that of hypertension, diabetes, and hyperlipidemia. J. Med. Virol. 86:772–777, 2014.

The association between periodontal disease treatment and end-stage renal disease (ESRD) remains unclear. This study aims to determine whether surgical periodontal treatment reduces ESRD risk.From the insurance claims data of patients with periodontal disease who were free of ESRD from 1997 to 2009, 35,496 patients were identified who underwent surgery for subgingival curettage and/or periodontal flap and are considered the treatment cohort. For comparison, 141,824 patients who did not undergo these treatments were considered the no-treatment cohort. Follow-ups were performed until the end of 2009 to estimate the incidence and risk of ESRD in these two cohorts. Cox proportional hazard regression was used to estimate the related hazard ratio (HR) and 95% confidence interval (CI) of ESRD.The incidence of ESRD was lower in the treatment cohort than in the no-treatment cohort (4.66 versus 7.38 per 10,000 person-years), with an adjusted HR of 0.59 (95% CI = 0.46 to 0.75). Sex- and age-specific analysis showed that the incidence rate ratio of the treatment cohort to the no-treatment cohort was higher for women than for men and declined with age. The risks of ESRD were consistently lower in the treatment cohort even when compared by comorbidity.Patients with periodontal disease who undergo procedures for subgingival curettage and/or periodontal flap have a remarkably decreased risk of ESRD.

Aim The purpose of this study was to explore the possible association between subsequent acute coronary syndrome (ACS) risk and depressive disorder. Methods We used data from the National Health Insurance system of Taiwan to address the research topic. The exposure cohort contained 10 871 patients with new diagnoses of depressive disorders. Each patient was randomly frequency-matched for sex and age with four participants from the general population who did not have any ACS history before the index date (control group). Cox's proportion hazard regression analyses were conducted to estimate the relation between depressive disorders and subsequent ACS risk. Results Among patients with depressive disorders, the overall risk for developing subsequent ACS was significantly higher than that of the control group (adjusted hazard ratio: 1.88, 95% confidence interval: 1.63–2.17). Further analysis revealed that the higher risk was observed in patients who were male, were of older age, or whose diagnosis was combined with other comorbidities. Conclusions The findings from this population-based retrospective cohort study suggest that depressive disorder is associated with an increased subsequent ACS risk.

Objective To evaluate the association between urinary tract stone (UTS) and urinary tract cancer (UTC) in Taiwanese patients, as the results of epidemiological studies about the relationship between UTS and the development of UTC remain inconclusive. Patients/Subjects and Methods We conducted a population-based cohort study using data from the Taiwanese National Health Insurance system. The UTS cohort included 21 862 patients, and each patient was randomly frequency-matched for age, sex, and index year with two insured members of the general population who did not have UTS. Cox proportional hazard regression analysis was performed to estimate the effect of UTS on the risk of UTC. Results Patients with UTS were at a significantly higher risk of developing UTC compared with the comparison group (adjusted hazard ratio 4.66; 95% confidence interval 2.97–7.30). Women were at higher risk than men. Further analyses showed that the level of UTC was associated with that of UTS, and that the risk for UTC became more divergent for the two groups over time. Conclusions Taiwanese patients with UTS, particularly women, had a higher risk for developing UTC than patients without UTS. The risk became more marked over time for this group.

Aims This study aimed to assess the risk of non-fatal cardiovascular events among patients with type 2 diabetes mellitus (T2DM) who are taking metformin, glimepiride or glyburide. Materials and methods Using the National Health Insurance Research database in Taiwan, this retrospective cohort study identified 1159 patients with newly diagnosed T2DM from 1998 to 2007, 30 years and older and without a history of cardiovascular disease at baseline. Patients with cancer, liver cirrhosis or chronic kidney disease were excluded. On the basis of prescription, patients were grouped into three medication subcohorts: metformin (N = 595), glimepiride (N = 234) or glyburide (N = 330) monotherapy for 100% of the follow-up period without any oral anti-diabetic agents added or changed, by the end of 2009. Incidence and hazard ratios of non-fatal cardiovascular events including coronary artery disease, peripheral artery disease, stroke and heart failure among these three subcohorts were compared. Results The overall incidence of non-fatal cardiovascular events was the highest for patients taking glyburide (169.1 per 1000 person-years), followed by for those taking glimepiride and metformin (95.2 and 49.1 per 1000 person-years, respectively). Compared with the adjusted hazard ratio for patients taking glyburide, the adjusted hazard ratio for those taking glimepiride was 0.52 (95% CI 0.40–0.69) and for those taking metformin was 0.31 (95% CI 0.24–0.40). Conclusions T2DM patients taking metformin and glimepiride are at lower risk of non-fatal cardiovascular events than those taking glyburide. Copyright © 2013 John Wiley & Sons, Ltd.

Purpose We conducted a nested case-control study to evaluate the association between risk of cancer and mirtazapine use in depression patients in Taiwan. Methods We obtained data from the Taiwan National Health Insurance Research Database to conduct a population-based nested case-control study. The study cohort included 16 897 patients diagnosed with depression between January 1, 2000 and December 31, 2008. We identified 530 cancer patients as the study group and matched 4 non-cancer subjects with each cancer patient by incident density, age, and sex. Odds ratios and 95% confidence intervals were estimated using multivariate conditional logistic regression analysis. Results Use of mirtazapine for depression did not have significant effect on overall cancer incidence (odds ratio: 1.03, 95% confidence interval: 0.72–1.48). Further analysis of annual mirtazapine dosages and the duration of mirtazapine use revealed no significant effect on cancer risk. Conclusion The findings of this population-based nested case-control study suggest that mirtazapine use may not provide a tumor suppression effect in humans such as that seen in the animal model. Future large-scale and in-depth investigations in this area are warranted. Copyright © 2013 John Wiley & Sons, Ltd.

Background and purpose Studies have suggested that multiple sclerosis (MS) might be linked to an overall reduced cancer rate, but a positive relationship is also found for several types of cancer. This study determines whether MS is associated with cancer risk in Taiwan. Methods Data from the National Health Insurance System of Taiwan were used to assess this issue. The MS cohort included 1292 patients, and each patient was randomly frequency-matched with four participants without MS, based on their age, sex and index year (control cohort). Cox proportional hazards regression analysis was conducted to estimate the influence of MS on cancer risk. Results A significantly higher risk of developing overall cancer was found amongst the MS cohort compared with the control cohort [adjusted hazard ratio (HR) 1.85, 95% confidence interval (CI) 1.26–2.74], as well as breast cancer (adjusted HR 2.23, 95% CI 1.11–4.46). Conclusions The nationwide population-based cohort study revealed that Taiwanese patients with MS have a higher risk of developing overall cancer types and breast cancer in particular.

1.97), breast cancer (HR=1.45, 95% CI=1.00, 2.09), and non-cervical gynecological cancer (HR=4.03, 95% CI=1.77, 9.17), compared with the general population. Conclusions: Physicians are at lower overall risk of cancer than the general population, apart from cancer of the thyroid, prostate, breast, and non-cervical gyne- cological cancer. (J Occup Health 2013; 55: 158−166) Abstract: A Comparison of Cancer Incidence among Physician Specialists and the General Population: A Taiwanese Cohort Study: Shih-Yi LIN, et al. Grad- uate Institute of Clinical Medical Science, College of Medicine, China Medical University, Taiwan— Background: Physicians are frequently studied as a population that experiences extremely high stress, burnout, and prolonged working hours that might harm one's health. However, they have sound medical knowl- edge and have easy access to medical resources. We studied the incidence of cancer among Taiwanese physi- cians using a nationwide cohort study design. Methods: Data were obtained from the National Health Insurance (NHI) system in Taiwan. The physician cohort contained 22,309 physicians, and each physician was randomly frequency-matched according to age and sex with 4 people from the general population. Results: The overall incidence ratio of cancer was 27% lower in the physician cohort than in the nonphysician comparison cohort (33.9 vs. 46.5 people per 10,000 person-years, crude hazard ratio (HR)=0.73, 95% CI=0.70, 0.76). The adjusted HR was 0.78 (95% CI=0.72, 0.84). Female physicians experienced a higher incidence rate ratio of overall cancer, compared to male physicians (crude HR=1.17, 95% CI=1.03, 1.33 vs. crude HR=0.70, 95% CI=0.67, 0.74, respectively). Physicians were at a significantly higher risk of thyroid cancer (HR 1.75, 95% CI=1.14, 2.68), prostate cancer (HR=1.54, 95% CI=1.21

This study examines the possible association between urinary tract infection (UTI) and urinary tract cancer (UTC). Data from the National Health Insurance system of Taiwan were used for the analysis. The UTI cohort included 70 116 patients who were diagnosed and recruited between 1997 and 2010. Each patient was randomly frequency-matched with two people without UTI from the general population based on their age, sex, and month of UTI diagnosis. Cox's proportional hazard regression analysis was used to estimate the effects of UTI on UTC risk until the end of follow-up on December 31, 2010. Patients with UTI had a significantly higher risk of developing UTC than healthy people (adjusted hazard ratio, 4.66; 95% confidence interval, 3.55-6.10). Further analyses indicated that risks are potentially related to the level of the lesion site. This study indicated that patients with UTI had a higher risk of developing UTC.

Diabet. Med. 30, 318–325 (2013) Abstract Aims To evaluate whether homeostasis model assessment and high-sensitivity C-reactive protein improve the prediction of isolated post-load hyperglycaemia. Methods The subjects were 1458 adults without self-reported diabetes recruited between 2006 and 2010. Isolated post-load hyperglycaemia was defined as fasting plasma glucose

Clin. Otolaryngol. 2012, 37, 369–375 Objectives: To explore the possible association between the risk of coronary artery disease and head or neck cancer based on some possible shared risk factors and/or treatment-related complications. Design and setting: A population-based retrospective cohort study. Participants: All new 729 patients with head and neck cancer diagnosed between 2000 and 2003 were followed up till 2009, and the risk of subsequent coronary artery disease was calculated. For each patient, the risk was calculated in 4 age-and sex-matched population controls. Matching was not possible for factors including socio-economic group and smoking. Main outcome measures: Cox’s proportional hazard regression analysis was conducted to estimate the relationship between head or neck cancer and risk of coronary artery disease. Results: For patients with head or neck cancer, the overall risk for developing coronary artery disease was almost the same as that of the control group [adjusted hazard ratio (but without control for some lifestyle factors): 0.95; 95% confidence interval: 0.65–1.35]. Overall, we found no increased risk of coronary artery disease for patients with head or neck cancer when the data were categorised either by treatment methods or by cancer subsites. Conclusions: This population-based study indicated that patients with head or neck cancer were at no higher risk of developing coronary artery disease than was the general population over a 6- to 9-year period; however, we cannot exclude the risk over a longer period of time.

Aim: The aim of the present study was to investigate whether polypharmacy correlates with the risk of dementia in older people. Methods: From representative claims data established from the National Health Insurance with a population coverage rate of 99% in Taiwan, we identified 7135 newly diagnosed patients with dementia in 2000–2008 and 2 8540 randomly selected controls without dementia, both aged ≥65 years. The daily use of prescribed drugs in the past 2 years was compared between cases and controls, controlling for demographic characters and comorbidities. Results: The incidence of dementia increased with the number of medications used and age. Cases were older than controls, predominant with women and more likely to use five or more drugs daily (44.0% vs 32.0%, P

SUMMARY Purpose: Schizophrenia and epilepsy may share a mutual susceptibility. This study examined the bidirectional relation between the two disorders. Methods: We used claims data obtained from the TaiwanNationalHealthInsurancedatabasetoconductretrospective cohort analyses. Analysis 1 compared 5,195 patients with incident schizophrenia diagnosed in 1999‐ 2008 with 20,776 controls without the disease randomly selected during the same period, frequency matched with sex and age. Analysis 2 comprised a similar method to compare 11,527 patients with newly diagnosed epilepsy with 46,032 randomly selected sex- and age-matched controls. At the end of 2008, analysis 1 measured the incidence and risk of developing epilepsy and analysis 2 measuredtheincidenceandriskofdevelopingschizophrenia. Key Findings: In analysis 1, the incidence of epilepsy was higher in the schizophrenia cohort than in the nonschizophrenia cohort (6.99 vs. 1.19 per 1,000 person-years) with an adjusted hazard ratio (aHR) of 5.88 [95% confidence interval (CI) 4.71‐7.36] for schizophrenia patients. In analysis 2, the incidence of schizophrenia was higher in the epilepsy cohort than in the nonepilepsy comparison cohort (3.53 vs. 0.46 per 1,000 person-years) with an aHR of 7.65 (95% CI 6.04‐9.69) for epilepsy patients. The effect of schizophrenia on subsequent epilepsy was greater for women, but the association between epilepsy and elevated incidence of schizophrenia was more pronounced in men. Significance: We found a strong bidirectional relation between schizophrenia and epilepsy. These two conditions may share common causes. Further studies on the mechanism are required.

Background: Little is known about the relationship between polypharmacy and Parkinson's disease (PD). In the present study, we tested the hypothesis that polypharmacy increases the risk of PD in older people in Taiwan. Methods: From a randomly sampled cohort of one million health insurance enrollees, we identified 2827 new cases of PD over the period 2000–2008 aged ≥65 years and selected 11 308 age-matched controls without PD. Medication history and baseline comorbidities were compared between the two groups. We defined ‘polypharmacy’ as an average daily use of five or more prescribed drugs. Results: Compared with patients using between none and one drug, the odds ratios (ORs) for PD increased to 1.53, 2.08, 2.64, and 2.95 for patients using two to four, five to seven, eight to nine, and ≥10 drugs, respectively. The other conditions associated with PD were dementia (OR 3.43), stroke (OR 2.30), depression (OR 2.15), and alcoholism (OR 2.11). Hyperlipidemia (OR 0.90) was inversely associated with PD. Conclusions: Polypharmacy was shown to increase the risk of PD in older people in Taiwan, with risk increasing with the number of medications taken daily. Dementia, stroke, depression, and alcoholism were also associated with an increased risk of PD, whereas there was an inverse correlation between PD and hyperlipidemia.

Background: Carpal tunnel syndrome (CTS), with unclear etiology, is the most common entrapment neuropathy. Its occurrence is related to lots of medical and non-medical conditions with uncertain causality. With a large population, we characterized selected demographical and clinical factors to add more information on CTS-correlated factors and new insight into future CTS prevention. Methods: A national insurance claim dataset of one million enrollees in Taiwan was used to identify 15 802 patients with CTS and 31 604 randomly selected controls, during a period of 7 years starting 1 January 2000. Statistical association with CTS was determined for five sociodemographic and nine medical factors. Results: Patients were predominantly women (65.6% vs. 47.7% in the control group) and older (40 and above, 62.6% vs. 36.2%). Rheumatoid arthritis was found to be the most significant comorbidity associated with CTS, followed by gout, hypertension, diabetes, obesity, uremia, and acromegaly. For younger group age ≤39, the association of these comorbidities was stronger, and hypothyroidism and vitamin B6 deficiency were additional comorbidities. Aging appears to reduce the relative impact of the diseases commonly associated with CTS as the possible risk factors. Conclusions: Identification of the CTS correlates in younger group would be of greater value in timely detection and treatment for these diseases. Correcting these disorders may aid in removing possible causes of CTS. This is the first report on the effect of aging on probable CTS risk factors. How factors associated with aging contribute to the development of CTS remains to be determined.

Liao C-H, Chang C-S, Chang S-N, Lane H-Y, Lyu S-Y, Morisky DE, Sung F-C. The risk of temporomandibular disorder in patients with depression: a population-based cohort study. Community Dent Oral Epidemiol 2011; 39: 525-531. � 2011 John Wiley & Sons A ⁄ S Abstract - Objectives: This study used a population-based retrospective cohort design to examine whether depression is a risk factor of temporomandibular disorder (TMD). Methods: From a universal insurance database, we identified 7587 patients who are newly diagnosed individuals with depression in 2000 and 2001. A total of 30 197 comparison subjects were randomly selected from a nondepression cohort. Both groups were followed until the end of 2008 to measure the incidence of TMD. Results: The incidence of TMD was 2.65 times higher in the depression cohort than in the nondepression cohort (6.16 versus 2.32 per 1000 person-years). The hazard ratio (HR) measured by multivariate Cox's proportional hazard regression analysis of TMD for the depression cohort was 2.21 (95% confidence interval (CI) 1.83-2.66), after controlling for socio- demographic factors and other psychiatric comorbidities. Women had higher risk to develop TMD than men (HR 1.61, 95% CI 1.36-1.92 for women without depression; HR 3.54, 95% CI 2.81-4.45 for women with depression). Conclusions: This study demonstrates that patients with depression are at an elevated risk of developing TMD.

Aliment Pharmacol Ther 2011; 33: 89–98 Summary Background Gastro-oesophageal reflux disease (GERD) has been associated with reflux laryngitis. Aims To investigate the risk factors and the predictors of pharyngeal acid reflux (PAR) in Taiwanese patients with suspected reflux laryngitis. Methods With referral from ENT physicians, 104 patients with symptoms and signs suggestive of reflux laryngitis completed a validated symptom questionnaire, an upper endoscopy exam and ambulatory 24-h pH tests with three sensors located at the hypopharynx, proximal and distal oesophagus. Patients with one or more episodes of PAR were considered abnormal. Results Pharyngeal acid reflux was identified in 17% (18/104) of patients. In multivariate logistic regression analysis, PAR was independently associated with classical reflux symptoms [adjusted odds ratio (aOR) = 3.5, 95% confidence interval (CI): 1.0–12.8], hiatus hernia (aOR = 6.7, 95% CI: 1.5–30.2) and overweight (aOR = 3.4, 95% CI: 1.0–11.0). In predicting PAR, classical reflux symptoms had a sensitivity of 78% and hiatus hernia had a specificity of 95%. With all three factors, the positive predictive value for PAR was 80%. Classical reflux symptoms included heartburn, chest pain, dyspepsia and acid regurgitation. Conclusions Classical reflux symptoms, hiatus hernia and overweight are independent risk factors that may predict pharyngeal acid reflux in patients with suspected reflux laryngitis.

Aims. To investigate the incidence of ambulatory care visits among female nursing staff in Taiwan. Background. Literature frequently indicates increased risks of needlestick injury and musculoskeletal disorders among practicing nurses. With increased workload in the healthcare system, nursing staff could also be vulnerable to other diseases. Design. A retrospective cohort study design including 27,624 female nursing staff with a contract with Taiwan’s National Health Insurance (NHI) programme between 2001–2004. Methods. Information on ambulatory care visits was retrieved from the NHI claim data. Comparison groups included other female medical personnel and active non-medical working women. Incidence rates of ambulatory care visits were calculated using person-year approach. Rate ratio adjusted for potential confounders was estimated from the Poisson regression model. Results. The nurse cohort developed a total of 781,352 ambulatory care visits, representing an incidence rate of 85,285/104 person-years. Compared with the other female medical personnel, female nurses had significantly elevated incidence of genitourinary [adjusted rate ratio (ARR) = 1·20], circulatory (ARR = 1·14), mental (ARR = 1·12), infectious/parasitic diseases (ARR = 1·11) and pregnancy/birth complications (ARR = 1·07). Significantly increased incidence sustained for both infectious/parasitic diseases and pregnancy/birth complications as the nurse cohort was compared with the active non-medical working women. The nurses, on the contrary, had significantly lower ARRs than the comparison groups of neoplasm, injury, metabolic, respiratory, digestive and musculoskeletal diseases. Conclusions. Female nurses had increased incidences of certain types of systemic illness when compared with other medical personnel and with women working outside of the health industry. Whether the findings found in this study were region specific or could be applicable to other nations, requires further investigations. Relevance to clinical practice. Policy makers and hospital administrators must not overlook nurses’ potentially unseen health problems. A mandatory periodical physical examination for nursing staff must be considered.

Background Chronic exposure to carbon disulfide (CS2) has been associated with hypertension, elevated cholesterol, arteriosclerosis, ischemic heart disease and related mortality, adverse nervous and reproductive health effects, and hearing impairment. This study assessed the hypertension risk for male rayon workers exposed to CS2. Methods A total of 251 rayon workers and 226 administrative clerks at two plants in Taiwan received health examination and interviews, and were compared for hypertension risk. On-site CS2 levels were measured. Results The hypertension was more prevalent in the exposure group (43.4%) than in controls (7.1%) with greater impact on systolic blood pressure than diastolic. Multivariate logistic regression analysis showed a significant dose–response relationship between hypertensive risk and cumulative exposure index (CEI) with an odds ratio of 15.1 for workers exposed to 343–468 year-ppm of CS2. The overall risk was elevated to 7.6 times higher for rayon workers. The risk increased significantly after more than 10 years of employment. Conclusions This study suggests that it takes a long exposure period to develop hypertension for rayon workers with CS2 exposure. Am. J. Ind. Med. 50:22–27, 2007. © 2006 Wiley-Liss, Inc.

A retrospective cohort morbidity study based on standardized incidence ratios (SIRs) was conducted to investigate the possible association between exposure to chlorinated organic solvents and various types of cancers in an electronics factory. The cohort of the exposed group was retrieved from the Bureau of Labor Insurance (BLI) computer database records dating from 1978 through December 31, 1997. Person-year accumulation began on the date of entry to the cohort, or January 1, 1979 (whichever came later), and ended on the closing date of the study (December 31, 1997), if alive without contracting any type of cancers, or the date of death, or the date of the cancer diagnosis. Vital status and cases of cancer of study subjects were determined from January 1, 1979 to December 31, 1997 by linking cohort data with the National Cancer Registry Database. The cancer incidence of the general population was used for comparison. After adjustment for age and calendar year, only SIR for breast cancer in the exposed female employees were significantly elevated when compared with the Taiwanese general population, based on the entire cohort without exclusion. The SIR of female breast cancer also showed a significant trend of period effect, but no significant dose-response relationship on duration of employment. Although the total cancer as well as the cancer for trachea, bronchus and lung for the entire female cohort was not significantly elevated, trend analysis by calendar-year interval suggested an upward trend. However, when duration of employment or latency was taken into consideration, no significantly elevated SIR was found for any type of cancer in either male or female exposed workers. In particular, the risk of female breast cancer was not indicated to be increased. No significant dose-response relationship on duration of employment and secular trend was found for the above-mentioned cancers. This study provides no evidence that exposure to chlorinated organic solvents at the electronics factory was associated with elevated human cancers. Dominant short-term employees may bias the cancer risk toward false positive.

Purpose: This study investigated the outcome of childhood hypertension among children urine screened positive for proteinuria, glucosuria and/or hematuria. Methods: Between 1992 and 2000, an islandwide mass urine screening program for proteinuria, glucosuria, and hematuria was conducted annually for all students aged 6–18 years in Taiwan Province. Among 732 students also found to be hypertensive in Taipei County, 164 were invited for a follow-up check in 2006 for their present health status. They were compared with 97 good-friend controls (invited by cases) for blood pressure, anthropometric and lipid measures, etc. Results: Compared with controls, cases were younger but had greater body mass index, waist circumference, waist-hip ratio, blood pressure, and lipidemia prevalence. The stepwise logistic regression analysis showed that cases were at greater risk than controls for chronic disease (odds ratio [OR]=3.27 (95% confidence interval [CI]=1.17–9.12), low high-density lipoprotein (OR=1.76, 95% CI=1.01–3.09), pre-hypertension (OR=2.97, 95% CI=1.43–6.15), and greater waist to hip ratio (OR=6.46, 95% CI=1.87–22.3). Conclusions: Our findings suggest that children screened with hypertension remain at elevated risk of having chronic disease and risk factors associated with hypertension.