Person:

Glynn, Robert

Objectives: To evaluate the association between restless legs syndrome (RLS) and incident cardiovascular disease (CVD). Design: Prospective cohort study. Setting: Women's Health Study (WHS) and Physicians' Health Study (PHS), USA. Participants: 29,756 female health professionals aged ( \geq 45) years and 19,182 male physicians aged ( \geq 40) years at baseline. Main outcome measures: Main outcome was incidence of major CVD; secondary outcomes were first incidence of myocardial infarction, stroke, death due to CVD or coronary revascularisation. Results: 3487 (11.7%) women and 1373 (7.2%) men met International Restless Legs Study Group criteria for RLS. In the WHS 450 major CVD events occurred and 1064 major CVD events were confirmed in the PHS. In both cohorts, RLS was not associated with increased risk of major CVD, stroke, myocardial infarction, CVD death or coronary revascularisation. After adjustment for major vascular risk factors, the HRs (95% CI) for major CVD were 1.15 (0.88 to 1.50) in women and 1.01 (0.81 to 1.25) in men. Highest multivariable-adjusted HRs were 1.29 (0.91 to 1.82) for total stroke in women and 1.22 (0.87 to 1.70) for CVD death in men. Excluding participants with comorbidities potentially leading to RLS did not substantially change the effect estimates. Conclusions: In these large prospective studies of female and male health professionals, RLS was not associated with an increased risk of any incident CVD event. The data do not support the hypothesis that RLS is a marker of increased risk of vascular disease.

Objectives: Research about work-related stressors and cardiovascular disease (CVD) has produced mixed findings. Moreover, a paucity of data exists regarding the long-term associations between job strain and job insecurity and CVD among women. Methods We used Cox proportional hazard models to examine the relationship between job strain, job insecurity, and incident CVD over 10 years of follow-up among 22,086 participants in the Women’s Health Study (mean age 57±5 years). Results: During 10 years of follow-up there were 170 myocardial infarctions (MI), 163 ischemic strokes, 440 coronary revascularizations, and 52 CVD deaths. In models adjusted for age, race, education, and income, women with high job strain (high demand, low control) were 38% more likely to experience a CVD event than their counterparts who reported low job strain (low demand, high control; Rate Ratio (RR) = 1.38, 95% Confidence Interval (CI) = 1.08–1.77), and women with active jobs (high demand, high control) were 38% more likely to experience a CVD event relative to women who reported low job strain (95% CI = 1.07–1.77). Outcome-specific analyses revealed that high job strain predicted non-fatal myocardial infarction (RR = 1.67, CI = 1.04–2.70), and coronary revascularization (RR = 1.41, CI = 1.05–1.90). No evidence of an association between job insecurity and long-term CVD risk was observed. Conclusion: High strain and active jobs, but not job insecurity, were related to increased CVD risk among women. Both job strain and job insecurity were significantly related to CVD risk factors. With the increase of women in the workforce, these data emphasize the importance of addressing job strain in CVD prevention efforts among working women.

In the Justification for Use of statins in Prevention: an Intervention Trial Evaluating Rosuvastatin (JUPITER), random allocation of rosuvastatin compared to placebo among primary prevention patients with a low-density lipoprotein cholesterol level of <130 mg/dl and a high-sensitivity C-reactive protein (hs-CRP) level of ≥2 mg/L resulted in a highly significant 44% reduction in major vascular events. However, the relation of baseline hs-CRP levels to risk within JUPITER has not previously been described and has been an area of controversy for study interpretation. As reported in the present study for the first time, despite enrolling patients with a constrained range of values, increasing baseline hs-CRP levels within JUPITER were nonetheless associated with increasing vascular risk in analyses treating hs-CRP as a continuous variable, as an ordinal variable, and as a threshold variable. As anticipated, the relative risk reduction associated with rosuvastatin was similar in magnitude across the tertile and threshold levels of entry hs-CRP. In conclusion, as the absolute risk increased with increasing hs-CRP, the absolute risk reduction associated with rosuvastatin within JUPITER was also greatest among those with the greatest entry hs-CRP levels.

Background: Older‐generation anticonvulsants that highly induce cytochrome P450 enzyme system activity produce metabolic abnormalities that may increase cardiovascular risk. The objective of this study was to evaluate the risk of ischemic cerebrovascular and coronary events in adult new users of anticonvulsants that highly induce cytochrome P450 activity compared with other anticonvulsant agents, as observed in a routine care setting. Methods and Results: This was a cohort study of patients 40 to 64 years old from the HealthCore Integrated Research Database who had initiated an anticonvulsant medication between 2001 and 2006 and had no recorded major coronary or cerebrovascular condition in the 6 months before treatment initiation. Propensity score (PS) matching was used to evaluate ischemic cerebrovascular and coronary risk among anticonvulsant new users. High‐dimensional propensity score (hdPS)–matched analyses were used to confirm adjusted findings. The study identified 913 events in 166 031 unmatched new treatment episodes with anticonvulsant drugs. In a PS‐matched population of 22 864 treatment episodes, the rate ratio (RR) for ischemic coronary or cerebrovascular events associated with highly inducing agents versus other agents was 1.22 (95% CI, 0.90‐1.65). The RR moved to 0.99 (95% CI, 0.73‐1.33) with adjustment for hdPS matching (RR, 1.47; 95% CI, 0.95‐2.28 for cerebrovascular events; RR, 0.70; 95% CI, 0.47‐1.05 for coronary events). Conclusions: In this exploratory analysis, there was no evidence of a consistent and statistically significant effect of initiating anticonvulsants that highly induce cytochrome P450 activity on ischemic coronary or cerebrovascular outcomes compared with other agents, given routine care utilization patterns.

Limited data are available on the association between adiponectin and incident heart failure. In the current ancillary study to the Physicians' Health Study, we used a prospective nested-case control design to examine whether plasma adiponectin concentration was related to the risk of heart failure. We selected 787 incident heart failure cases and 787 matched controls for the current analysis. Each control was selected using a risk set sampling technique at the time of the occurrence of the index case and matched on year of birth, age at blood collection, and race. Adiponectin was measured using ELISA. Heart failure occurrence was self-reported in annual follow-up questionnaire. Validation of self-reported heart failure in this cohort has been published. The mean age was 58.7 years. In a conditional logistic regression adjusting for age, race, time of blood collection, year of birth, hypertension, atrial fibrillation, smoking, alcohol intake, and exercise, estimates of the relative risk (95% confidence interval) were 1.0 (ref), 0.74 (0.53–1.04), 0.67 (0.48–0.94), 0.70 (0.50–0.99), and 0.92 (0.65–1.30) from the lowest to the highest quintile of adiponectin, respectively, p for quadratic trend 0.004. Additional adjustment for potential mediating factors including diabetes, C-reactive protein, and body mass index led to the attenuation of the estimate of effect [1.0 (ref), 0.81 (0.57–1.15), 0.75 (0.53–1.06), 0.83 (0.58–1.18), and 1.26 (0.87–1.81) across consecutive quintiles of adiponectin]. Our data are consistent with a J-shaped association between total adiponectin and the risk of heart failure among US male physicians.

Background: Once atrial fibrillation (AF) progresses to sustained forms, adverse outcomes increase and treatment success rates decrease. Therefore, identification of risk factors predisposing to persistence of AF may have a significant impact on AF morbidity. Methods and Results: We prospectively examined the differential associations between traditional, lifestyle, and biomarker AF risk factors and development of paroxysmal versus nonparoxysmal AF (persistent/permanent) among 34 720 women enrolled in the Women's Health Study who were free of cardiovascular disease and AF at baseline. AF patterns were defined based on current guidelines and classified according to the most sustained form of AF within 2 years of diagnosis. During a median follow‐up of 16.4 years, 690 women developed paroxysmal AF and 349 women developed nonparoxysmal AF. In multivariable time‐varying competing risk models, increasing age (hazard ratio [HR] 1.11, 95% CI 1.10 to 1.13, versus HR 1.08, 1.07 to 1.09, per year), body mass index (HR 1.07, 1.05 to 1.09, versus HR 1.03, 1.02 to 1.05, per kg/m2), and weight (HR 1.30, 1.22 to 1.39, versus HR 1.14, 1.08 to 1.20, per 10 kg) were more strongly associated with the development of nonparoxysmal AF compared with paroxysmal AF. Hemoglobin A1c levels at baseline were directly related to the development of nonparoxysmal AF but inversely associated with paroxysmal AF in multivariable competing risk models (P for nonequal association=0.01). Conclusions: In women without AF or CVD at baseline, increasing age, adiposity, and higher hemoglobin A1c levels were preferentially associated with the early development of nonparoxysmal AF. These data raise the hypothesis that efforts aimed at weight reduction or glycemic control may affect the proportion of the population with sustained AF.

We developed a semi-automated active monitoring system that uses sequential matched-cohort analyses to assess drug safety across a distributed network of longitudinal electronic healthcare data. In a retrospective analysis, we showed that the system would have identified cerivastatin-induced rhabdomyolysis. In this study, we evaluated whether the system would generate alerts for three drug-outcome pairs: rosuvastatin and rhabdomyolysis (known null association), rosuvastatin and diabetes mellitus, and telithromycin and hepatotoxicity (two examples for which alerting would be questionable). During >5 years of monitoring, rate differences (RDs) comparing rosuvastatin to atorvastatin were -0.1 cases of rhabdomyolysis per 1,000 person-years (95% CI, -0.4, 0.1) and -2.2 diabetes cases per 1,000 person-years (95% CI, -6.0, 1.6). The RD for hepatotoxicity comparing telithromycin to azithromycin was 0.3 cases per 1,000 person-years (95% CI, -0.5, 1.0). In a setting in which false positivity is a major concern, the system did not generate alerts for three drug-outcome pairs.

Women are advised to be vaccinated for influenza during pregnancy and may receive vaccine at any time during their pregnancy. In observational studies evaluating vaccine safety in pregnancy, to account for such time-varying vaccine exposure, a time-dependent predictor can be used in a proportional hazards model setting for outcomes such as spontaneous abortion or preterm delivery. Also, due to the observational nature of pregnancy exposure cohort studies and relatively low event rates, propensity score (PS) methods are often used to adjust for potential confounders. Using Monte Carlo simulation experiments, we compare two different ways to model the PS for vaccine exposure: (1) logistic regression treating the exposure status as binary yes or no; (2) Cox regression treating time to exposure as time-to-event. Coverage probability of the nominal 95% confidence interval for the exposure effect is used as the main measure of performance. The performance of the logistic regression PS depends largely on how the exposure data is generated. In contrast, the Cox regression PS consistently performs well across the different data generating mechanisms that we have considered. In addition, the Cox regression PS allows adjusting for potential time-varying confounders such as season of the year or exposure to additional vaccines. The application of the Cox regression PS is illustrated using data from a recent study of the safety of pandemic H1N1 influenza vaccine during pregnancy.

Background: Multivariable confounder adjustment in comparative studies of newly marketed drugs can be limited by small numbers of exposed patients and even fewer outcomes. Disease risk scores (DRSs) developed in historical comparator drug users before the new drug entered the market may improve adjustment. However, in a high dimensional data setting, empirical selection of hundreds of potential confounders and modeling of DRS even in the historical cohort can lead to over-fitting and reduced predictive performance in the study cohort. We propose the use of combinations of dimension reduction and shrinkage methods to overcome this problem, and compared the performances of these modeling strategies for implementing high dimensional (hd) DRSs from historical data in two empirical study examples of newly marketed drugs versus comparator drugs after the new drugs’ market entry—dabigatran versus warfarin for the outcome of major hemorrhagic events and cyclooxygenase-2 inhibitor (coxibs) versus nonselective non-steroidal anti-inflammatory drugs (nsNSAIDs) for gastrointestinal bleeds. Results: Historical hdDRSs that included predefined and empirical outcome predictors with dimension reduction (principal component analysis; PCA) and shrinkage (lasso and ridge regression) approaches had higher c-statistics (0.66 for the PCA model, 0.64 for the PCA + ridge and 0.65 for the PCA + lasso models in the warfarin users) than an unreduced model (c-statistic, 0.54) in the dabigatran example. The odds ratio (OR) from PCA + lasso hdDRS-stratification [OR, 0.64; 95 % confidence interval (CI) 0.46–0.90] was closer to the benchmark estimate (0.93) from a randomized trial than the model without empirical predictors (OR, 0.58; 95 % CI 0.41–0.81). In the coxibs example, c-statistics of the hdDRSs in the nsNSAID initiators were 0.66 for the PCA model, 0.67 for the PCA + ridge model, and 0.67 for the PCA + lasso model; these were higher than for the unreduced model (c-statistic, 0.45), and comparable to the demographics + risk score model (c-statistic, 0.67). Conclusions: hdDRSs using historical data with dimension reduction and shrinkage was feasible, and improved confounding adjustment in two studies of newly marketed medications. Electronic supplementary material The online version of this article (doi:10.1186/s12982-016-0047-x) contains supplementary material, which is available to authorized users.

Background/Objectives: Data from previous studies consistently suggest that maternal smoking is positively associated with obesity later in life. Whether this association persists across generations is unknown. We examined whether grand-parental smoking was positively associated with overweight status in adolescence. Subject/Methods: Participants were grandmother-mother-child triads in the Nurses’ Health Study II (NHS II), the Nurses Mothers’ Cohort Study, and the Growing up Today Study (GUTS). Grandmothers provided information on their and their partner’s smoking during pregnancy with the child’s mother. Information on child's weight and height at ages 12 (N = 3094) and 17 (N = 3433) was obtained from annual or biennial GUTS questionnaires. We used logistic regression to estimate ORs of being overweight or obese, relative to normal weight. Results: Grand-maternal smoking during pregnancy was not associated with overweight status in adolescence. After adjusting for covariates, the OR of being overweight or obese relative to normal weight at age 12 years in girls whose grandmothers smoked 15+ cigarettes daily during pregnancy was 1.21 (95% CI 0.74-1.98; ptrend = 0.31) and 1.07 (0.65-1.77; ptrend = 0.41) in boys. Grand-paternal smoking during pregnancy was associated with being overweight or obese at age 12 in girls only, but not at age 17 for either sex: the OR for being overweight or obese at age 12 was 1.38 (95% CI 1.01-1.89; ptrend = 0.03) in girls, and 1.31 (95% CI 0.97-1.76; ptrend = 0.07) in boys. Among children of non-smoking mothers, the OR for granddaughter obesity for grand-paternal smoking was attenuated and no longer significant [OR 1.28 (95% CI 0.87-1.89; ptrend= 0.18)]. Conclusions: Our findings suggest that the association between maternal smoking and offspring obesity may not persist beyond the first generation. However, grand-paternal smoking may affect overweight status of the granddaughter, likely through the association between grand- paternal smoking and maternal smoking.