A comparative study of different methods for automatic identification of clopidogrel-induced bleedings in electronic health records.

Electronic health records (EHRs) linked with biobanks have been recognized as valuable data sources for pharmacogenomic studies, which require identification of patients with certain adverse drug reactions (ADRs) from a large population. Since manual chart review is costly and time-consuming, automatic methods to accurately identify patients with ADRs have been called for. In this study, we developed and compared different informatics approaches to identify ADRs from EHRs, using clopidogrel-induced bleeding as our case study. Three different types of methods were investigated: 1) rule-based methods; 2) machine learning-based methods; and 3) scoring function-based methods. Our results show that both machine learning and scoring methods are effective and the scoring method can achieve a high precision with a reasonable recall. We also analyzed the contributions of different types of features and found that the temporality information between clopidogrel and bleeding events, as well as textual evidence from physicians' assertion of the adverse events are helpful. We believe that our findings are valuable in advancing EHR-based pharmacogenomic studies.

Plasma hepatocyte growth factor is a novel marker of AL cardiac amyloidosis.

BACKGROUND: Cardiac amyloidosis is an infiltrative cardiomyopathy that is challenging to diagnose. We hypothesized that the novel biomarkers hepatocyte growth factor (HGF), galectin-3 (GAL-3), interleukin-6 (IL-6), and vascular endothelial growth factor (VEGF) would be elevated in cardiac amyloidosis and may be able to discriminate from non-cardiac systemic amyloidosis or other cardiomyopathies with similar clinical or morphologic characteristics. METHODS: Patients were selected from the Vanderbilt Main Heart Registry according to the following groups: (1) amyloid light-chain (AL) cardiac amyloidosis (n = 26); (2) transthyretin (ATTR) cardiac amyloidosis (n = 7); (3) left ventricular hypertrophy (LVH) (n = 45); (4) systolic heart failure (n = 42); and (5) non-cardiac systemic amyloidosis (n = 7). Biomarkers were measured in stored plasma samples. Biomarkers' discrimination performance in predicting AL cardiac amyloidosis (i.e., Concordance index) was reported. A survival analysis was used to explore the relationship between HGF levels and mortality among AL cardiac amyloidosis patients. RESULTS: HGF levels were markedly elevated in patients with AL cardiac amyloidosis (median = 622, interquartile range (IQR): 299-1228 pg/mL) compared with the other groups, including those with non-cardiac systemic amyloidosis (median = 134, IQR: 94-163 pg/mL, p < 0.001). HGF was not a specific marker for ATTR amyloidosis. Gal-3 was elevated in all groups with amyloidosis but could not differentiate between those with and without cardiac involvement. There was no difference in IL-6 or VEGF between those with AL cardiac amyloidosis compared to other groups (p = 0.13 and 0.057, respectively). CONCLUSIONS: HGF may be a specific marker that distinguishes AL cardiac amyloidosis from other cardiomyopathies with similar clinical or morphologic characteristics. Further studies are necessary to determine whether HGF levels predict the likelihood of survival.

Evaluation of the F2R IVS-14A/T PAR1 polymorphism with subsequent cardiovascular events and bleeding in patients who have undergone percutaneous coronary intervention.

Abnormal platelet reactivity is associated with recurrent ischemia and bleeding following percutaneous coronary intervention (PCI). Protease-activated receptor-1 (PAR1), encoded by F2R, is a high affinity thrombin receptor on platelets and the target of the antiplatelet drug vorapaxar. The intronic single nucleotide polymorphism F2R IVS-14 A/T affects PAR1 receptor density and function. We hypothesized that carriers of the T allele, who have been shown to have decreased platelet reactivity, would be at lower risk for thrombotic events, but higher risk for bleeding following PCI. Using BioVU, the Vanderbilt DNA repository linked to the electronic medical record, we studied 660 patients who underwent PCI for unstable or stable coronary artery disease. Primary outcome measures were major adverse cardiovascular events (MACE, composite of revascularization, MI, stroke, death) and bleeding (assessed by Bleeding Academic Research Consortium scale) over 24 months. The minor allele (T) frequency was 14.8 %. There were no genotypic differences in the frequency of MACE (33.7, 28.8, and 31.6 % for A/A, A/T, and T/T respectively, P = 0.50) or bleeding (15.7, 14.7, and 18.8 % for A/A, A/T, and T/T respectively, P = 0.90). In a Cox regression model, fully adjusted for age, race, sex, BMI, and smoking status, carrying a T allele was not associated with MACE (HR 1.19, 95 % CI 0.89-1.59, P = 0.23) or bleeding (HR 0.73, 95 % CI 0.37-1.4, P = 0.34). In conclusion, in our population, F2R IVS-14 PAR1 variability does not affect risk of MACE or bleeding following PCI.

Understanding the role of prostaglandin E2 in regulating human platelet activity in health and disease.

The platelet thrombus is the major pathologic entity in acute coronary syndromes, and antiplatelet agents are a mainstay of therapy. However, individual patient responsiveness to current antiplatelet drugs is variable, and all drugs carry a risk of bleeding. An understanding of the complex role of Prostaglandin E2 (PGE2) in regulating thrombosis offers opportunities for the development of novel individualized antiplatelet treatment. However, deciphering the platelet regulatory function of PGE2 has long been confounded by non-standardized experimental conditions, extrapolation of murine data to humans, and phenotypic differences in PGE2 response. This review synthesizes past and current knowledge about PGE2 effects on platelet biology, presents a rationale for standardization of experimental protocols, and provides insight into a molecular mechanism by which PGE2-activated pathways could be targeted for new personalized antiplatelet therapy to inhibit pathologic thrombosis without affecting hemostasis.

Catecholamine pathway gene variation is associated with norepinephrine and epinephrine concentrations at rest and after exercise.

OBJECTIVE: To examine the hypothesis that genetic variation in enzymes and transporters associated with synthesis, storage, release, and metabolism of catecholamines contributes to the interindividual variability in plasma catecholamine concentrations at rest and after exercise. METHODS: We measured plasma norepinephrine (NE) and epinephrine concentrations at rest and after a standardized exercise protocol in 165 healthy individuals (60% White, 40% African-American) and examined 29 functional or common variants in 14 genes involved in synthesis, transport, or metabolism of catecholamines. We examined the relationship between genotypes and NE concentrations at rest and the increase after exercise (ΔNE) by multiple linear regression with adjustment for covariates [age, race, sex, BMI, fitness, and resting NE (for ΔNE)]. As a secondary outcome, we carried out similar analyses for epinephrine concentrations. RESULTS: There was large interindividual variability in resting NE (mean, 204±102 pg/ml; range, 39-616 pg/ml) and ΔNE (mean, 256±206 pg/ml; range, -97 to 953 pg/ml). Resting NE was significantly associated with variants of four genes: CYB561 (P<0.001), VMAT2 (P=0.016), CHGA (P=0.039), and PNMT (P=0.038). ΔNE after exercise was associated with three variants of PNMT (P=0.041) and COMT (P=0.033 and 0.035), and resting and exercise epinephrine concentrations were associated with two variants each. CONCLUSION: The findings of this exploratory study suggest that variation in catecholamine pathway genes contributes to the interindividual variability in plasma NE and epinephrine concentrations at rest and after exercise.

CYP2A6 genetic variation and dexmedetomidine disposition.

PURPOSE: There is a large interindividual variability in dexmedetomidine dose requirements for sedation of patients in intensive care units (ICU). Cytochrome P450 2A6 (CYP2A6) mediates an important route of dexmedetomidine metabolism, and genetic variation in CYP2A6 affects the clearance of other substrate drugs. We examined whether CYP2A6 genotypes affect dexmedetomidine disposition. METHODS: In 43 critically ill ICU patients receiving dexmedetomidine infusions adjusted to achieve the desired level of sedation, we determined a median of five plasma dexmedetomidine concentrations each. Forty subjects were genotyped for five common CYP2A6 alleles and grouped into normal (n = 33), intermediate (n = 5), and slow metabolizers (n = 2). RESULTS: Using a Bayesian hierarchical nonlinear mixture model, estimated dexmedetomidine clearance was 49.1 L/h (posterior mean; 95% credible interval 41.4-57.6 L/h). There were no significant differences in dexmedetomidine clearance among normal, intermediate, and slow CYP2A6 metabolizer groups. CONCLUSION: Genetic variation in CYP2A6 does not appear to be an important determinant of dexmedetomidine clearance in ICU patients.

Genetic variations in the α(2A)-adrenoreceptor are associated with blood pressure response to the agonist dexmedetomidine.

BACKGROUND: α(2A)-Adrenoceptors (α(2A)-ARs) have important roles in sympathetic cardiovascular regulation. Variants of ADRA2A affect gene transcription and expression and are associated with insulin release and risk for type 2 diabetes. We examined whether ADRA2A variants are also associated with cardiovascular responses to the selective α(2)-AR-agonist dexmedetomidine. METHODS AND RESULTS: Seventy-three healthy subjects participated in a placebo-controlled, single-blind study. After 3 infusions of placebo, subjects received 3 incremental infusions of dexmedetomidine (cumulative dose, 0.4 µg/kg). Primary outcomes were changes in systolic blood pressure (SBP) and plasma norepinephrine concentrations, measured as difference of the area-under-the-curve during placebo and dexmedetomidine infusions (ΔAUC). We used multiple linear regression analysis to examine the associations between 9 ADRA2A tagging variants and 5 inferred haplotypes and ΔAUC after adjustment for covariates. Homozygous carriers of rs553668 and the corresponding haplotype 4, previously associated with increased α(2A)-AR expression, had a 2.2-fold greater decrease in AUC(SBP) after dexmedetomidine (adjusted P=0.006); similarly, the maximum decrease in SBP was 24.7±8.1 mm Hg compared with 13.6±5.9 mm Hg in carriers of the wild-type allele (P=0.007). Carriers of haplotype 3, previously associated with reduced α(2A)-AR expression, had a 44% smaller decrease in AUC(SBP) (P=0.013). Haplotype information significantly improved the model predicting the decrease in SBP (P<0.001). There were similar but nonsignificant trends for diastolic blood pressure and heart rate. Genotypes were not significantly associated with norepinephrine responses. CONCLUSIONS: Common ADRA2A variants are associated with the hypotensive response to dexmedetomidine. Effects of specific variants/haplotypes in vivo are compatible with their known effects on gene expression in vitro.

Effects of variation in the human alpha2A- and alpha2C-adrenoceptor genes on cognitive tasks and pain perception.

BACKGROUND: The mechanisms underlying interindividual variability in pain perception and cognitive responses are undefined but highly heritable. alpha(2C)- and alpha(2A)-adrenergic receptors regulate noradrenergic activity and are important mediators of pain perception and analgesia. We hypothesized that common genetic variants in these genes, particularly the ADRA2C 322-325 deletion variant, affect pain perception or cognitive responses. METHODS: We studied 73 healthy subjects (37 Caucasians and 36 African-Americans) aged 25.4+/-4.6years. Pain response to a cold pressor test was measured using a 10cm visual analog scale and again on the next day, after three infusions of the selective alpha(2)-agonist dexmedetomidine. Standardized cognitive tests were administered at baseline and after each infusion. The contribution of ADRA2C deletion genotype, dexmedetomidine concentration, and other covariates to pain perception and cognitive responses was determined using multiple linear regression models. Secondary analysis examined the effects of ADRA2A and other ADRA2C variants on pain perception. RESULTS: ADRA2C Del homozygotes had higher pain scores in response to cold at baseline (6.3+/-1.8cm) and after dexmedetomidine (5.6+/-2.2cm) than insertion allele carriers (4.6+/-2.1cm [baseline] and 3.8+/-1.9cm [after dexmedetomidine]; adjusted P-values=0.019 and 0.004, respectively). Cognitive responses were unrelated to ADRA2C Ins/Del genotype. None of the other ADRA2A and ADRA2C variants was significantly related to cold pain sensitivity before dexmedetomidine; after dexmedetomidine, ADRA2A rs1800038 was marginally associated (P=0.03). CONCLUSION: The common ADRA2C del322-325 variant affected pain perception before and after dexmedetomidine but did not affect other cognitive responses, suggesting that it contributes to interindividual variability in pain perception.

GRK5 Gln41Leu polymorphism is not associated with sensitivity to beta(1)-adrenergic blockade in humans.

AIMS: A common, functionally significant polymorphism in GRK5 (Gln41Leu) encodes a gain-of-function enzyme that enhances desensitization of the beta(1)-adrenergic receptor. GRK5 Leu41 has been postulated to confer endogenous 'genetic beta-blockade' and contribute to an attenuated response to beta-blockers in black subjects. The effects of this GRK5 variant on sensitivity to a beta-blocker have not been studied in humans. We hypothesized that the GRK5 Gln41Leu variant contributes to interindividual variability in response to beta-blockade and to the ethnic difference in sensitivity between black and Caucasian individuals. MATERIALS & METHODS: We measured the heart rate at rest and during a graded incremental exercise in 154 healthy subjects (85 white and 69 black) before and after an oral administration of 25 mg atenolol. We determined the genotypes of GRK5 (Gln41Leu), beta(1)-adrenergic receptor (ADRB1 Ser49Gly and Arg389Gly) genotypes and plasma atenolol concentrations. The effects of genotype and covariates on sensitivity to atenolol, measured as the reduction in exercise-induced tachycardia, were determined using multiple regression analyses. RESULTS: The minor allele frequency of GRK5 Leu41 was 32.6% in blacks and 0% in whites. Black individuals were less sensitive to atenolol than white individuals (p < or = 0.011) but this was not explained by the GRK5 genotype. The GRK5 genotype had no effect on resting heart rate before (p = 0.61) and after adjustment for age, sex, ethnicity, atenolol concentrations, BMI and ADRB1 genotypes (p = 0.81). The decrease in heart rate after atenolol administration did not differ significantly according to the GRK5 genotype at rest or after exercise, before (all p > 0.14) and after statistical adjustment for covariates (all p > 0.17). CONCLUSION: The GRK5 Gln41Leu polymorphism does not affect sensitivity to the beta(1)-adrenergic blocker, atenolol, during acute physiological adrenergic stimulation, nor does it contribute to the ethnic differences in sensitivity to atenolol among black and Caucasian individuals.

The alpha2C-adrenoceptor deletion322-325 variant and cold-induced vasoconstriction.

OBJECTIVES: Cold-induced vasoconstriction is mediated in part by selective enhancement of local alpha(2C)-adrenoceptor (alpha(2C)-AR) activity. A common insertion-deletion variant in the alpha(2C)-AR gene (ADRA2C del322-325) results in an approximately 85% reduction of agonist-mediated function in vitro. We tested the hypothesis that individuals with the ADRA2C del322-325 variant have attenuated vasoconstriction in response to cold. METHODS: Cutaneous digital blood flow (flux) was measured by laser Doppler flowmetry in a controlled environment at room temperature and during two cycles of graduated local heat and cold exposure in 31 subjects. Temperature-response curves were analyzed to estimate the following measures: E(min) (minimal flux during cooling), and ET(50) and ET(90) (the local temperature at which flux decreased by 50 and 90%, respectively). RESULTS: We found no significant genotypic differences in E(min) (24.3 +/- 19.5, 30.0 +/- 20.5, and 21.5 +/- 25.9 AU for ins/ins, ins/del, and del/del genotypes, respectively; P = 0.48), ET(50) (25.5 +/- 6.0, 25.1 +/- 6.7, and 25.1 +/- 7.1 degrees C; P = 0.99), or ET(90) (20.5 +/- 4.7, 22.1 +/- 4.0, and 20.8 +/- 6.7 degrees C; P = 0.77) in either the first or second heating and cooling cycle (cycle 1 values presented). INTERPRETATION: The ADRA2C del322-325 variant did not affect vascular sensitivity to local cold exposure.

Beta-1-adrenoceptor genetic variants and ethnicity independently affect response to beta-blockade.

OBJECTIVES: Black patients may be less responsive to beta-blockers than whites. Genetic variants in the beta1-adrenergic receptor (beta1-AR) associated with lesser response to beta-blockers are more common in blacks than in whites. The purpose of this study was to determine whether ethnic differences in response to beta-blockade can be explained by differing distributions of functional genetic variants in the beta1-AR. METHODS: We measured sensitivity to beta-blockade by the attenuation of exercise-induced tachycardia in 165 patients (92 whites), who performed a graded bicycle exercise test before and 2.5 h after oral atenolol (25 mg). We determined heart rate at rest and at three exercise levels from continuous ECG recordings and calculated the area under the curve. We also measured plasma atenolol concentrations and determined genotypes for variants of the beta1-AR (Ser49Gly, Arg389Gly) and alpha2C-AR (del322-325). The effects of ethnicity, genotype, and other covariates on the heart rate reduction after atenolol were estimated in multiple regression analyses. RESULTS: Atenolol resulted in a greater reduction in exercise heart rate in whites than in blacks (P=0.006). beta1-AR Arg389 (P=0.003), but not the alpha2C-AR 322-325 insertion allele (P=0.31), was independently associated with a greater reduction in heart rate area under the curve. Ethnic differences in sensitivity to atenolol remained significant (P=0.006) after adjustment for beta1-AR and alpha2C-AR genotypes. CONCLUSION: Ethnic differences in sensitivity to the beta1-blocker atenolol persist even after accounting for different distributions of functional genetic beta1-AR variants, suggesting that additional, as yet unidentified factors contribute to such ethnic differences.

Genetic variants in the alpha2C-adrenoceptor and G-protein contribute to ethnic differences in cardiovascular stress responses.

OBJECTIVES: Cardiovascular responses to stressors are regulated by sympathetic activity, increased in black Americans, and associated with future cardiovascular morbidity. Our aim was to determine whether two functional variants in genes regulating sympathetic activity, a deletion in the alpha2C-adrenergic receptor (ADRA2C del322-325) and a G-protein beta3-subunit variant (GNB3 G825T), affect cardiovascular responses to physiologic stressors and contribute to their ethnic differences. METHODS: We measured heart rate and blood pressure responses to a cold pressor test (CPT) in 79 healthy participants (40 blacks, 39 whites), aged 25.7+/-5.3 years, and determined genotypes for the ADRA2C and GNB3 variants. We examined the response variables (increase in heart rate and blood pressure) in multiple linear regression analyses adjusting first for baseline measures, ethnicity, and other covariates, and then additionally for genotypes. RESULTS: Black participants had a greater heart rate response to CPT than whites [mean difference, 9.9 bpm; 95% confidence interval (CI), 4.1 to 15.6; P=0.001]. Both the ADRA2C del/del (15.8 bpm; 95% CI, 8.0-23.7; P<0.001) and GNB3 T/T genotypes (6.8 bpm; 95% CI, 0.9-12.7; P=0.026) were associated with greater heart rate response. After adjusting for genotypes, the ethnic difference was abrogated (1.3 bpm; 95% CI, -5.4-8.0; P=0.70), suggesting that the genetic variants contributed substantially to ethnic differences. CONCLUSION: Variation in genes that regulate sympathetic activity affects hemodynamic stress responses and contributes to their ethnic differences. This study elucidates how genetic factors may in part explain ethnic differences in cardiovascular regulation.

Ethnic and genetic determinants of cardiovascular response to the selective alpha 2-adrenoceptor agonist dexmedetomidine.

The alpha(2)-adrenoceptor agonist clonidine reduces blood pressure more effectively in White than Black Americans despite similar degrees of sympatholysis. Functional genetic variation in receptor signaling mechanisms, for example in the beta 3 G-protein subunit (GNB3 C825T) and in the alpha(2C)-adrenoceptor subtype (ADRA2C del322-325), may affect drug responses. We examined the hypothesis that there are ethnic differences in the responses to the highly selective alpha(2)-agonist, dexmedetomidine, and that these genetic variants contribute to interindividual variability in drug responses. In a placebo-controlled, single-masked study, 73 healthy subjects (37 whites and 36 blacks) received 3 placebo infusions and then 3 incremental doses of dexmedetomidine (cumulative dose, 0.4 microg/kg), each separated by 30 minutes. Blood pressure, heart rate, and plasma catecholamine concentrations were determined after each infusion. We measured dexmedetomidine concentrations after the last infusion and determined ADRA2C del322-325 and GNB3 C825T genotypes. Dexmedetomidine lowered blood pressure and plasma catecholamine concentrations significantly (all P<0.001). There was substantial interindividual variability in the reduction of systolic blood pressure (range, 1 to 34 mm Hg) and plasma norepinephrine concentrations (range, 24 to 424 pg/mL). However, there were no differences between black and white subjects in dexmedetomidine responses (P>0.16 for all outcomes) before or after adjustment for covariates. Neither ADRA2C del322-325 nor GNB3 C825T genotypes affected the responses to dexmedetomidine (all P>0.66). There is large interindividual variability in response to the selective alpha(2)-AR agonist dexmedetomidine, and neither ethnicity nor ADRA2C and GNB3 genotypes contribute to it. Further studies to identify determinants of alpha(2)-AR-mediated responses will be of interest.

Effect of the alpha2C-adrenoreceptor deletion322-325 variant on sympathetic activity and cardiovascular measures in healthy subjects.

BACKGROUND: The alpha2C-adrenergic receptor plays an important role in the regulation of the sympathetic nervous system and, therefore, blood pressure and heart rate. A deletion polymorphism in its gene (ADRA2C del322-325), ten times more common in black than white Americans, has been associated with a loss of function in vitro and, under controlled study conditions, raised blood pressure and catecholamine secretion. We therefore examined the hypothesis that the ADRA2C deletion variant would alter sympathetic activity and contribute to ethnic differences in blood pressure. METHODS: We measured resting plasma norepinephrine and epinephrine concentrations, blood pressure and heart rate in 224 healthy subjects (127 whites), and determined their ADRA2C del322-325 genotype. Additionally, we analyzed heart rate variability (HRV) in a subgroup of 50 black subjects. RESULTS: Systolic (SBP) and diastolic blood pressure (DBP) were higher in blacks than whites [difference (95% confidence interval), 4.4 (1.5-7.4) mmHg, P = 0.003; and 2.7 (0.7-4.6) mmHg, P = 0.01, respectively]. Norepinephrine concentrations did not differ among subjects with 0, 1 and 2 copies of the deletion variant [median (interquartile range), 185.0 (147.5-269.8), 200.0 (154.9-257.0) and 173.8 (158.5-235.8) pg/ml, respectively; P = 0.54]. Similarly, none of the HRV parameters differed among the genotype groups. In multiple linear regression analyses adjusting for multiple covariates, the deletion genotype was not associated with SBP or DBP. In contrast, black ethnicity was associated with higher SBP (P = 0.001) and DBP (P = 0.005). CONCLUSION: The ADRA2C deletion polymorphism had no effect on markers of resting sympathetic activity and cardiovascular measures, and did not account for ethnic differences in blood pressure.