Gen-Min Lin

Serum uric acid as an independent predictor of mortality in high-risk patients with obstructive coronary artery disease: A prospective observational cohort study from the ET-CHD registry, 1997–2003

BACKGROUND Serum uric acid (SUA) has been observed to be highly associated with the development of cardiovascular disease for more than 50 years. Several studies have reported elevated SUA as an independent predictor of mortality in patients with coronary artery disease (CAD) after adjustment for classic risk factors but some studies did not find similar results. METHODS Between January 1997 and December 2003, a prospective cohort study was performed in 1054 patients with angiographically defined CAD, and their classic risk factors and SUA levels were determined at enrollment. The study cohort was followed for an average of 3.2 years, with a median of 3.1 years. The main outcome measure was death from cardiac disease and any cause. RESULTS Of all study patients, 789 (74.9%) were men and 265 (25.1%) were women. The mean age of the male and female patients was 64.8 and 66.9 years, respectively. The mean SUA level of all patients was 410.4 μmol/L. There were grading effects of SUA quartiles on cardiac and all-cause mortality in univariate and multivariate Cox regression analyses. After adjustment, the multivariate analyses revealed that patients in the highest SUA quartile (>487 μmol/L) had 2.08 (95% CI=1.19-3.62, p=0.01) fold increased risk of cardiac death, and 1.68 (95% CI=1.10-2.57, p=0.017) fold increase risk of overall mortality compared with the lowest quartile (<315 μmol/L). CONCLUSIONS SUA may be a significant predictor of cardiac and overall mortality, independent of classic risk factors in high-risk patients with obstructive CAD.

Relation of body mass index to mortality among patients with percutaneous coronary intervention longer than 5 years follow-up: A meta-analysis

percutaneous coronary intervention longer than 5 years follow-up: A meta-analysis Yi-Hwei Li ⁎, Gen-Min Lin ⁎, Chin-Lon Lin , Ji-Hung Wang , Chih-Lu Han d a Department of Public Health, Tzu-Chi University, Hualien, Taiwan b Department of Medicine, Hualien Armed Forces General Hospital, Hualien, Taiwan c Division of Cardiology, Buddhist Tzu-Chi General Hospital, Hualien, Taiwan d Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan

Relation of serum uric acid and body mass index to mortality in high-risk patients with established coronary artery disease: a report from the ET-CHD registry, 1997-2006.

BACKGROUND Hyperuricemia is associated with a higher risk of death in patients with coronary artery disease (CAD). In contrast, overweight or obesity was associated with survival benefits in this population. However, the relation of body mass index (BMI) and serum uric acid (SUA) to mortality has not been clarified in this population. METHODS We studied a cohort of 1202 patients with angiographic CAD from the ET-CHD registry during 1997-2003 in Taiwan. To evaluate the relation of BMI and SUA on mortality, the subjects were categorized into 4 groups by BMI >/= 25 kg/m(2) (overweight or obesity) or BMI<25 kg/m(2) (normal- or under-weight), and SUA levels higher or lower than the median of 6.6 mg/dl. At a median follow-up of 5.4 years, cardiac and all-cause deaths were the primary end points. RESULTS Multivariate analyses demonstrated that high SUA group had a significantly higher cardiac mortality [hazard ratio (HR): 1.79, 95% confidence interval (CI): 1.14-2.82, p=0.023] and overall mortality (HR: 1.68, 95% CI: 1.19-2.36, p=0.003) than low SUA group only in overweight or obese patients. Additionally, high BMI was associated with a significantly lower cardiac mortality (HR: 0.58, 95% CI: 0.38-0.99, p=0.023) and overall mortality (HR: 0.62, 95% CI: 0.41-0.82, p=0.003) than low BMI in patients with low SUA levels. Furthermore, normal-low weight and underweight patients (BMI<21 kg/m(2)) were found to have a higher risk of mortality regardless of SUA levels. CONCLUSIONS Among patients with established CAD, SUA may be a potent predictor to mortality in overweight or obese patients. Moreover, the obesity-mortality paradox phenomenon was mainly driven by higher mortality risk in underweight patients and lower mortality risk in overweight and obese patients with low SUA.

Gender differences in the impact of diabetes on mortality in patients with established coronary artery disease: A report from the Eastern Taiwan integrated health care delivery system of Coronary Heart Disease (ET-CHD) registry, 1997–2006

OBJECTIVES The effect of type 2 diabetes mellitus (DM) on mortality was more pronounced in women than men with coronary artery disease (CAD) in the pre-stent era before 1996. However this relationship is controversial in the post-stent era. METHODS We studied a cohort of 1073 patients with angiographically defined CAD from the Eastern Taiwan integrated health care delivery system of Coronary Heart Disease (ET-CHD) registry during 1997-2003 in Tzu-Chi General Hospital, Hualien, Taiwan. To evaluate gender-specific DM effect on mortality, the subjects were divided into 4 groups: diabetic women (n=147), non-diabetic women (n=127), diabetic men (n=239), and non-diabetic men (n=560). At a mean follow-up of 5.4 years, cardiac and all-cause mortality were the primary end points. RESULTS Annual total mortality rates were 10.2%, 5.1%, 7.2%, and 4.8%; annual cardiac mortality rates were 8.2%, 3.0%, 4.3%, and 2.6% for diabetic women, non-diabetic women, diabetic men, and non-diabetic men, respectively. Multivariate Cox regression models, adjusted for possible confounders showed that gender-specific hazard ratios (HRs) of DM for total mortality were 2.02 (95% CI: 1.32-3.09), and 1.72 (95% CI: 1.32-2.25) for women and men, respectively. The HRs for total mortality associated with diabetes were not different between women and men (p=0.53). Similarly, adjusted gender-specific HRs of DM for cardiac mortality were 2.46 (95% CI: 1.45-4.19) for women, and 1.83 (95% CI: 1.28-2.62) for men, which were also not significantly different (p=0.36). CONCLUSIONS Among patients with CAD, the impact of DM on mortality was consistently higher in women than in men, but the differences across sexes were not statistically significant after 1996 in Taiwan.

The obesity-mortality paradox in elderly patients with angiographic coronary artery disease: a report from the ET-CHD registry.

Background The body mass index (BMI)-mortality paradox has been well known in patients with obstructive coronary artery disease (CAD). However, this phenomenon has rarely been described among elderly patients over a 5-year follow-up. Methods We studied a cohort of 722 elderly patients (age ≥ 65 years) with angiographic CAD from the ET-CHD registry during 1997-2003 in eastern Taiwan. To evaluate the BMI eff ect on mortality, the elderly subjects were categorized into 5 groups by BMI (kg/m2): underweight and normal-low weight (< 21), normal-high weight (21-23.9) overweight (24-26.9), mild obesity (27-29.9) and moderate/severe obesity (≥ 30). At a maximal 10-year follow-up, cardiac and all-cause deaths were the primary end points. Results After multivariate analysis, patients from the category of underweight and normal-low weight in reference to those from the normal-high weight category had a signifi cantly higher risk of cardiac and all-cause mortality (hazard ratio (HR): 1.68 (95% CI: 1.04-2.70) and 2.02 (95% CI: 1.42-2.87), respectively) following a median of 5.4 years. Obese elderly patients tended to have the lowest risk of all-cause death across all the study BMI categories in the early stage. However, after 5 years, mortality increased in the obese patients surviving beyond 5 years, and was higher than that in overweight patients. Conclusions The obesity-mortality paradox was present in elderly patients with angiographic CAD in Taiwan and the risk of death was signifi cantly higher in those with a BMI < 21 kg/m2. However, a J-shaped relationship between mortality and BMI developed after 5 years of follow-up.

Relation of body mass index to mortality among patients with percutaneous coronary intervention in the drug-eluting stent era: A systematic review and meta-analysis

with percutaneous coronary intervention in the drug-eluting stent era: A systematic review and meta-analysis Gen-Min Lin ⁎, Yi-Hwei Li ⁎, Chin-Lon Lin , Ji-Hung Wang , Chih-Lu Han d a Department of Public Health, Tzu-Chi University, Hualien, Taiwan b Department of Medicine, Hualien Armed Forces General Hospital, Hualien, Taiwan c Division of Cardiology, Buddhist Tzu-Chi General Hospital, Hualien, Taiwan d Department of Medicine, Taipei Veterans General Hospital, Taipei, Taiwan

The "smoker's paradox" in Asian versus non-Asian patients with percutaneous coronary intervention longer than 6 months follow-up: a collaborative meta-analysis with the ET-CHD registry.

Cigarettesmokingis recognizedasanimportantriskfactorfor car-diovascular disease for decades [1]. Among patients with acute coro-nary events, current smoker has been observed to be paradoxicallyassociated with a survival benefit. This is called as the “smokers para-dox” [2,3]. On average, symptomatic smokers were found with youngeragesandlowerratesofcomorbiditythannon-smokersatentryinprevi-ous studies [4]. Therefore, thephenomenon may be mostly reasoned bythe discrepancies in baseline characteristics. In addition, some reportsuncovered that the “smokers paradox” was only present in short-termbutitwasalmostabsentwithfollow-uplongerthan6 months[5].Nota-bly, most of these studies were carried out in the fibrinolysis era. Hencetherewerefewerreportswithregardtotheissuespeci ficallyforpatientsundergoing percutaneous coronary intervention (PCI). Furthermore, arecent large-sized report from the combined Korean Acute MyocardialInfarction Registry (KAMIR) and the Korean Myocardial Infarction(KorMI) registry revealed that current smokers were associated with a48% decrease in the risk of total death in reference to non-smokers at1 year after primary PCI for acute MI [6]. This study implied that anethnic differencemightbe presentin the “smokers paradox”.Thereforewe conducted a meta-analysis to compare the impact of smoking statusat baseline on the subsequent coronary events between Asian andnon-Asian patients with PCI and follow-up longer than 6 months.Detailedsearchstrategiesweredesignedtoidentifystudiesevaluatingtheeffectofbaselinesmokingstatusonthemortalityandcardiaceventsfollowing PCI for at least 6 months. One author (Lin GM) pre-screenedthe search results from MEDLINE and Cochrane Central Register ofControlled Trials (CENTRAL) from 1970 to 2013 and removed citationsthat were not relevant. Two reviewers (Lin GM and Li YH) indepen-dently reviewed the remaining citations and performed data extrac-tion. Disagreements were resolved through consensus. Study qualitywas assessed using the New Castle–Ottawa quality scale (NOS) for ob-servational studies [7]. Studies were required to separately report theadjusted risk ratios of total or cardiac deaths/events in patients by thesmokingstatusatentry.Theprimaryoutcomewastotalorcardiacmor-tality and the secondary outcomes were target vessel revascularization(TVR) and recurrent MI. Finally, there were ten articles (four for Asiansand six for non-Asians) extracted to synthesize their results and thestudy characteristics were displayed in Table 1 [6,8–16].Ofnote,thefollow-upperiodsinmostoftheseselectedPCIregistrieswere reported between 6 months and 1 year(three for Asians and fivefor non-Asians) and longer than 3 years (one for Asians and one fornon-Asians). Additionally, we analyzed the adjusted hazard ratios(HRs) of all-cause death at 1 year and 3 years, respectively betweensmokers and non-smokers with PCI in the Eastern Taiwan integratedhealth care delivery system of Coronary Heart Disease (ET-CHD) regis-try in Table 2 [17–19].The baseline profiles of patients undergoing PCI werepresented aspercentages or as mean ± SD. Chi-square tests were used to testbaseline differences across the study groups for categorical variables.Since most continuous variables in this study did not follow a normaldistribution, Wilcoxon rank sum test, a nonparametric method with-out the requirement of normal assumption, was used to determinethe difference between smokers and non-smokers. Cox proportionalhazard regression models were used to determine the adjusted HRsfor total death associated with the baseline smoking status. All con-founders listed in Table 2 were entered into the model except totalcholesterol, triglycerides, and glucose levels which were removedfrom thefinal model due to their high association with the other co-variates. All statistical analyses were performed with SAS version 9.2(SAS Institute, Cary, NC). A P b 0.05 was considered statisticallysignificant.

Cancer Risk Before Schizophrenia Diagnosis in Taiwan, 1995–2009

The debate on whether schizophrenia precipitates or prevents cancers has lasted for over 1 century.1 Earlier researches almost focused upon the crude cancer risk after schizophrenia diagnosis. Inconsistent and even conflicting cancer incidence ratios were published. The inconsistency could be due to various sources of confounding. Until 1994, Mortensen was the first to take into account different smoking rates between schizophrenia patients and general population to attain smoking-adjusted cancer risks.2 Other lifestyle factors and psychotropic medications were also controlled for in the subsequent studies.3,4 Lichtermann even took a step further to conduct age, gender, and age-of-onset stratified analyses in both schizophrenia patients and their first-degree relatives.5 Despite all these efforts, the adjusted risk ratios were still at odds with one another and no conclusions could be made concerning how schizophrenia alters one’s cancer risks. On this very topic, Ji et al provided 3 distinct indices using the Swedish schizophrenia registry: (1) smoking-adjusted cancer risk in schizophrenia patients, (2) cancer risk in their first-degree relatives, and (3) before-diagnosis cancer risk in schizophrenia patients.6 Among them, the first 2 indices were frequently reported in the literature. Because most of the factors related to schizophrenia process are impossible to be fully adjusted in registry data, the first-degree relatives of schizophrenia patients turn up as perfect proxies for investigation. The cancer risks for the disease-free relatives could be attributed to either shared genetic background or shared environmental background, but the risks are hardly related with the schizophrenia disease process itself. According to Ji et al, male schizophrenia patients have lower cancer risk than the general population (standardized incidence ratio [SIR]: 0.79, 95% CI: 0.75–0.82), but female patients have higher risk (SIR: 1.20, 95% CI: 1.15–1.24). Moreover, relative to the general Swedish population, the first-degree relatives are generally protected from cancer occurrence (Female SIR: 0.96, 95% CI: 0.94–0.98; Male SIR: 0.92, 95% CI: 0.89–0.96). In a recent meta-analysis, the pooled overall SIRs come fairly close (SIR for siblings: 0.89, 95% CI: 0.84–0.94; SIR for parents: 0.90, 95% CI: 0.88–0.93).7 The work of Ji et al indeed shed new light on “differential cancer risks” throughout schizophrenia development. They constructed not only “cancer risk after schizophrenia diagnosis” but also counter-intuitively, “cancer risk before schizophrenia diagnosis.” There are excellent reasons to do so. First, if schizophrenia and cancers are exclusively genetically linked, whichever condition is first identified should not alter the risk (as long as the observation period is adequately long). Second, cancer risk may distribute unevenly during different stages in schizophrenia. According to their results, the latter scenario is supported. The cancer-protection advantage disappears as patients go from psychiatrically undiagnosed state (SIR: 0.40, 95% CI: 0.38–0.43) to diagnosed state (SIR: 1.00, 95% CI: 0.97–1.03). In women, cancer protection before schizophrenia diagnosis (overall SIR before and after schizophrenia diagnosis: 0.94, 95% CI: 0.91–0.97) shifts into cancer proneness after schizophrenia diagnosis (SIR: 1.20, 95% CI: 1.15–1.24). In men, as schizophrenia manifests itself, the impact of cancer-protection also lessens (overall SIR before and after schizophrenia diagnosis: 0.63, 95% CI: 0.61–0.66; SIR after schizophrenia diagnosis: 0.79, 95% CI: 0.75–0.82). To our best knowledge, Ji was one of the very few to estimate before-schizophrenia-diagnosis SIRs for cancers.6 Because evidence cumulated concerning cancer risks after schizophrenia diagnosis and those in first-degree relatives, it will be reassuring to establish results on cancer risks before schizophrenia diagnosis too. To obtain “cancer risk before schizophrenia diagnosis,” we used the National Health Insurance Research Database in Taiwan to identify 74 448 schizophrenia patients (39 669 men and 34 779 women) diagnosed between 1995 and 2009.8 The follow-up time for cancers was defined either by person-years until cancer identification before schizophrenia diagnosis or by person-years until schizophrenia diagnosis. The mean ages in 1995 were 30.92 for females and 27.54 for males; the mean follow-up time was 8.13 years for women and 7.95 years for men. A total of 471 patients with schizophrenia (337 women and 134 men) were identified with cancers by the end of 2009. The mean age for cancer identification was 45.6, and schizophrenia was diagnosed averagely 5.3 years later. We calculated expected number of cancers by multiplying the sum of at-risk person-years by the overall cancer rate in the corresponding population (from the cancer registry database associated with “Taiwan Cancer Control Act”). The standardized incidence ratios were calculated as the observed/expected number of cancer cases, with 95% confidence intervals assuming a Poisson distribution for the observed number of cancers. In summary, the before-schizophrenia SIR for cancers was 0.58 (95% CI: 0.52–0.65). The gender-specific SIR was 0.77 (95% CI: 0.67–0.88) in females and 0.36 (95% CI: 0.30–0.44) in males. As such, our results reinforce the findings by Ji et al that cancer risk is generally reduced before schizophrenia manifestation.2,6,8 Based on our previous work concerning “cancer risk after schizophrenia diagnosis,” the overall SIR for cancers was 1.17 (95% CI: 1.08–1.28), female SIR 1.31 (95% CI: 1.17–1.48) and male SIR was 1.02 (95% CI: 0.90–1.16).8 Taken together with the before-schizophrenia cancer risks, the trend of “loss of cancer protection (as schizophrenia evolves)” was replicated. Furthermore, from our previous study, we also found the cancer risks appeared significantly lower than those of the general population after 5 years of schizophrenia diagnosis. That is, schizophrenia is generally cancer-protective in both the prodromal and chronic stages. In other words, schizophrenia solely increases cancer risk during the early deteriorating phase. The concurrent cancer risk is likely to be caused by the immunity-inflammation process that also contributes to schizophrenia itself.9,10 For completion, our overall SIR before and after schizophrenia diagnosis was 1.14 (95% CI: 1.07–1.21): separately, 1.30 (95% CI: 1.19–1.41) in women, and 0.94 (95% CI: 0.85–1.04) in men. As far as “trend” is concerned, our findings are consistent with those from Ji et al although both the overall before-and-after-schizophrenia-diagnosis SIRs and the overall after-schizophrenia-diagnosis SIRs are consistently higher in our study. In Swedish registry, schizophrenia men are protected from cancers, but schizophrenia women do not have the same benefit. In Taiwanese registry, schizophrenia women are at higher risk, whereas schizophrenia men are not apt for cancers. Interestingly, our findings are in line with the after-schizophrenia-diagnosis SIRs from some other Asian-based studies.11 Given no other plausible explanations, racial disparities are likely to account for the systemic difference. In short, Asian schizophrenia patients tend to possess relatively higher cancer risk profiles than Caucasians irrespective of gender. Catts reasoned that the racial disparity in cancer SIRs among schizophrenia patients might result from the disparity in background cancer incidence rates for general population.7 For instance, assuming cancer risk in schizophrenia patients is constant across countries, the cancer SIRs would be exaggerated in Asian countries wherein the background cancer incidence rates are usually lower than those of European countries. In accordance with the after-schizophrenia-diagnosis SIR, Swedish female patients also have an increased cancer risk. Just as we did, Ji et al found female schizophrenia patients have increased risks specifically for female- specific cancers such as breast and endometrium cancers. Knowing the “cancer risk before schizophrenia diagnosis” appears to be lower in both Swedish and Taiwanese women, we are assured that some factors associated with the disease process of schizophrenia (for instance, antipsychotic-induced hyperprolactinemia) may have a universally promoting effect on female-specific cancers. According to the 2 so far largest population-based studies, cancer risk in schizophrenia is not constant across the life span. To get a renewed in-depth outlook on the relationship between schizophrenia and cancers, we propose that cancer risk in schizophrenia should be framed as stage-specific risk rather than lifetime risk. In the meanwhile, more parallel comparisons should be made between international groups to elucidate risk transition trends and ethnic discrepancies. Finally, the before-after (schizophrenia diagnosis) comparisons may testify the environmental effects on specific cancers.

Relation of estimated glomerular filtration rate and body mass index to mortality in non-dialysis patients with coronary artery disease: A report from the ET-CHD registry, 1997–2003

BACKGROUND Low estimated glomerular filtration rate (eGFR) and body mass index (BMI) have been known as poor prognostic factors in patients with coronary artery disease (CAD). Besides, high BMI was associated with higher survival rate in patients on dialysis as well. However, the relation of eGFR and BMI to mortality has not been clarified in non-dialysis CAD patients. METHODS We studied a cohort of 1243 non-dialysis patients with angiographic CAD from the ET-CHD registry during 1997-2003 in Taiwan. To evaluate the relation of eGFR and BMI to mortality, the subjects were categorized into 4 groups by BMI higher or lower than 25 kg/m2 and eGFR higher or lower than 60 mL/min/1.73 m2, a cut-off value for chronic kidney disease (CKD). At a mean follow-up of 5.4 years, cardiac and all-cause deaths were the primary end points. RESULTS In the high BMI group, CKD was a strong predictor of cardiac and overall mortality [hazard ratio (HR): 1.63 (95% CI: 1.05-2.53) and 2.17 (95% CI: 1.54-3.07), respectively]. Besides, CKD marginally elevated the mortality risk in the low BMI group. Among CKD patients, BMI was not associated with mortality except for a high death rate in patients with BMI lower than 21 kg/m2. In contrast, in non-CKD patients, high BMI group had a lower cardiac and overall mortality than the low BMI group [HR: 0.75 (95% CI: 0.48-1.16) and 0.55 (95% CI: 0.39-0.77), respectively]. CONCLUSIONS Among non-dialysis CAD patients in eastern Taiwan, lower eGFR was associated with worse prognosis regardless of BMI levels. Notably, in patients with high BMI, the relationship between CKD and mortality was extremely remarkable. Moreover, the obesity-mortality paradox phenomenon was not present in non-dialysis CKD patients.

Role of Yin and Yang in Diurnal Blood Pressure Variation for Cardiovascular Disease

To the Editor: We read with great interest the work by Verdecchia et al,1 which reported that, in initially untreated subjects with hypertension, a blunted day-night blood pressure (BP) dip was associated with a blunted morning BP surge and vice versa. In these subjects, a blunted morning BP surge was an independent predictor of cardiovascular events, whereas an excessive BP surge did not portend an increased risk of events.1 As compared with previous reports, there are conflicting results regarding the association between morning BP surges and following cardiovascular events, and these results should be interpreted carefully. Here, we use the concept of yin and yang representing …