Brian R. Overholser

Development and Validation of a Risk Score to Predict QT Interval Prolongation in Hospitalized Patients

Background—Identifying hospitalized patients at risk for QT interval prolongation could lead to interventions to reduce the risk of torsades de pointes. Our objective was to develop and validate a risk score for QT prolongation in hospitalized patients. Methods and Results—In this study, in a single tertiary care institution, consecutive patients (n=900) admitted to cardiac care units comprised the risk score development group. The score was then applied to 300 additional patients in a validation group. Corrected QT (QTc) interval prolongation (defined as QTc>500 ms or an increase of >60 ms from baseline) occurred in 274 (30.4%) and 90 (30.0%) patients in the development group and validation group, respectively. Independent predictors of QTc prolongation included the following: female (odds ratio, 1.5; 95% confidence interval, 1.1–2.0), diagnosis of myocardial infarction (2.4 [1.6–3.9]), sepsis (2.7 [1.5–4.8]), left ventricular dysfunction (2.7 [1.6–5.0]), administration of a QT-prolonging drug (2.8 [2.0–4.0]), ≥2 QT-prolonging drugs (2.6 [1.9–5.6]), or loop diuretic (1.4 [1.0–2.0]), age >68 years (1.3 [1.0–1.9]), serum K+ <3.5 mEq/L (2.1 [1.5–2.9]), and admitting QTc >450 ms (2.3; confidence interval [1.6–3.2]). Risk scores were developed by assigning points based on log odds ratios. Low-, moderate-, and high-risk ranges of 0 to 6, 7 to 10, and 11 to 21 points, respectively, best predicted QTc prolongation (C statistic=0.823). A high-risk score ≥11 was associated with sensitivity=0.74, specificity=0.77, positive predictive value=0.79, and negative predictive value=0.76. In the validation group, the incidences of QTc prolongation were 15% (low risk); 37% (moderate risk); and 73% (high risk). Conclusions—A risk score using easily obtainable clinical variables predicts patients at highest risk for QTc interval prolongation and may be useful in guiding monitoring and treatment decisions.

Prevalence of QT Interval Prolongation in Patients Admitted to Cardiac Care Units and Frequency of Subsequent Administration of QT Interval-Prolonging Drugs

BackgroundCardiac arrest due to torsades de pointes (TdP) is a rare but catastrophic event in hospitals. Patients admitted to cardiac units are at higher risk of drug-induced QT interval prolongation and TdP, due to a preponderance of risk factors. Few data exist regarding the prevalence of QT interval prolongation in patients admitted to cardiac units or the frequency of administering QT interval-prolonging drugs to patients presenting with QT interval prolongation.ObjectiveThe aim of this study was to determine the prevalence of Bazett’s-corrected QT (QTc) interval prolongation upon admission to cardiac units and the proportion of patients presenting with QTc interval prolongation who are subsequently administered QT interval-prolonging drugs during hospitalization.MethodsThis was a prospective, observational study conducted over a 1-year period (October 2008–October 2009) in 1159 consecutive patients admitted to two cardiac units in a large urban academic medical centre located in Indianapolis, IN, USA. Patients were enrolled into the study at the time of admission to the hospital and were followed daily during hospitalization. Exclusion criteria were age <18 years, ECG rhythm of complete ventricular pacing, and patient designation as ‘outpatient’ in a bed and/or duration of stay <24 hours. Data collected included demographic information, past medical history, daily progress notes, medication administration records, laboratory data, ECGs, telemetry monitoring strips and diagnostic reports. All patients underwent continuous cardiac telemetry monitoring and/or had a baseline 12-lead ECG obtained within 4 hours of admission. QT intervals were determined manually from lead II of 12-lead ECGs or from continuous lead II telemetry monitoring strips. QTc interval prolongation was defined as ≥470 ms for males and ≥480 ms for females. In both males and females, QTc interval >500 ms was considered abnormally high. A medication was classified as QT interval-prolonging if there were published data indicating that the drug causes QT interval prolongation and/or TdP. Study endpoints were (i) prevalence of QTc interval prolongation upon admission to the Cardiac Medical Critical Care Unit (CMCCU) or an Advanced Heart Care Unit (AHCU); (ii) proportion of patients admitted to the CMCCU/AHCU with QTc interval prolongation who subsequently were administered QT interval-prolonging drugs during hospitalization; (iii) the proportion of these higher-risk patients in whom TdP risk factor monitoring was performed; (iv) proportion of patients with QTc interval prolongation who subsequently received QT-prolonging drugs and who experienced further QTc interval prolongation.ResultsOf 1159 patients enrolled, 259 patients met exclusion criteria, resulting in a final sample size of 900 patients. Patient characteristics: mean (± SD) age, 65 ±15 years; female, 47%; Caucasian, 70%. Admitting diagnoses: heart failure (22%), myocardial infarction (16%), atrial fibrillation (9%), sudden cardiac arrest (3%). QTc interval prolongation was present in 27.9% of patients on admission; 18.2% had QTc interval >500ms. Of 251 patients admitted with QTc interval prolongation, 87 (34.7%) were subsequently administered QT interval-prolonging drugs. Of 166 patients admitted with QTc interval >500ms, 70 (42.2%) were subsequently administered QT interval-prolonging drugs; additional QTc interval prolongation ≥60 ms occurred in 57.1% of these patients.ConclusionsQTc interval prolongation is common among patients admitted to cardiac units. QT interval-prolonging drugs are commonly prescribed to patients presenting with QTc interval prolongation.

Effectiveness of a Clinical Decision Support System for Reducing the Risk of QT Interval Prolongation in Hospitalized Patients

Background—We evaluated the effectiveness of a computer clinical decision support system (CDSS) for reducing the risk of QT interval prolongation in hospitalized patients. Methods and Results—We evaluated 2400 patients admitted to cardiac care units at an urban academic medical center. A CDSS incorporating a validated risk score for QTc prolongation was developed and implemented using information extracted from patients’ electronic medical records. When a drug associated with torsades de pointes was prescribed to a patient at moderate or high risk for QTc interval prolongation, a computer alert appeared on the screen to the pharmacist entering the order, who could then consult the prescriber on alternative therapies and implement more intensive monitoring. QTc interval prolongation was defined as QTc interval >500 ms or increase in QTc of ≥60 ms from baseline; for patients who presented with QTc >500 ms, QTc prolongation was defined solely as increase in QTc ≥60 ms from baseline. End points were assessed before (n=1200) and after (n=1200) implementation of the CDSS. CDSS implementation was independently associated with a reduced risk of QTc prolongation (adjusted odds ratio, 0.65; 95% confidence interval, 0.56–0.89; P<0.0001). Furthermore, CDSS implementation reduced the prescribing of noncardiac medications known to cause torsades de pointes, including fluoroquinolones and intravenous haloperidol (adjusted odds ratio, 0.79; 95% confidence interval, 0.63–0.91; P=0.03). Conclusions—A computer CDSS incorporating a validated risk score for QTc prolongation influences the prescribing of QT-prolonging drugs and reduces the risk of QTc interval prolongation in hospitalized patients with torsades de pointes risk factors.

Sex‐Related Differences in the Pharmacokinetics of Oral Ciprofloxacin

The oral pharmacokinetics of ciprofloxacin were studied in healthy volunteers to assess the influence of sex on its disposition. Subjects (8 males, 7 females) received a single oral dose of ciprofloxacin 750 mg, blood and urine samples were collected, and ciprofloxacin concentrations were determined. A two‐compartment open‐model with two or three absorption phases, each one having a fitted independent lag time, best fit the data using a weighted least squares estimator. Univariate and multivariate regression analyses were performed to determine the influence of renal function, weight, and subject sex on the oral clearance (CLS/F) and apparent steady‐state volume of distribution (Vss/F) of ciprofloxacin. Females had a median Cmax of ciprofloxacin that was 30% greater than males and a significantly smaller median (range) Vss/F: 81.1 (44.8–111.6) versus 170.9 (140.9–213.4), respectively (p < 0.01). In addition, females had increased exposure to ciprofloxacin, with a slower median (range) CLS/F of 28.3 L/h (24.5–33.4) compared to 44.4 L/h (41.4–53.7) for males (p < 0.01). Regression analyses revealed that subject sex was the only significant predictor of CLS/F (p < 0.001), but both body weight (p= 0.04) and subject sex (p< 0.005) were significant predictors of Vss/F. Fixed oral doses of ciprofloxacin will lead to higher maximum concentration and total drug exposure in females compared to males and do not appear to be solely related to weight‐based differences.

Biostatistics Primer: Part I

Biostatistics is the application of statistics to biologic data. The field of statistics can be broken down into 2 fundamental parts: descriptive and inferential. Descriptive statistics are commonly used to categorize, display, and summarize data. Inferential statistics can be used to make predictions based on a sample obtained from a population or some large body of information. It is these inferences that are used to test specific research hypotheses. This 2-part review will outline important features of descriptive and inferential statistics as they apply to commonly conducted research studies in the biomedical literature. Part 1 in this issue will discuss fundamental topics of statistics and data analysis. Additionally, some of the most commonly used statistical tests found in the biomedical literature will be reviewed in Part 2 in the February 2008 issue.

Pharmacokinetics of Intravenously Administered Levofloxacin in Men and Women

Study Objective. To characterize and compare the pharmacokinetics of levofloxacin in men and women after systemic administration.

Efavirenz Inhibits the Human Ether-A-Go-Go Related Current (hERG) and Induces QT Interval Prolongation in CYP2B6*6*6 Allele Carriers

Efavirenz (EFV) has been associated with torsade de pointes despite marginal QT interval lengthening. Since EFV is metabolized by the cytochrome P450 (CYP) 2B6 enzyme, we hypothesized that EFV would lengthen the rate‐corrected QT (QTcF) interval in carriers of the CYP2B6*6 decreased functional allele.

Limited sampling strategies to estimate exposure to the green tea polyphenol, epigallocatechin gallate, in fasting and fed conditions

The objective of this study was to develop an efficient sampling strategy to predict epigallocatechin gallate (EGCG) pharmacokinetics after green tea administration. Ten healthy subjects received a single 800-mg oral dose of EGCG administered as Polyphenon E under both fasting and fed conditions. Plasma samples were serially collected over 24 hours and EGCG concentrations were determined. A one-compartment model with a lag time for absorption best fit the concentration-time data. Maximum A Posteriori Bayesian (MAPB) priors were developed by simultaneously fitting pharmacokinetic parameters from both study phases. The D-optimal sampling designs were determined and Monte Carlo simulations were performed. The original model with the estimators was used to fit the simulated data with the optimized sampling schemes. Two and three optimal sampling strategies (OSS-2 and OSS-3, respectively) were developed. The median two sampling times for OSS-2 were 1.3 and 6.9 hours (fasting conditions) and 3.4 and 8.7 hours (fed conditions). The median three sampling times for OSS-3 were 0.7, 1.4, and 7.0 hours (fasting conditions) and 1.4, 3.6, and 8.7 hours (fed conditions). The predictive power of OSS-3 was greater than that of OSS-2, under both fasted and fed conditions, and both strategies had greater predictive performance under fasting conditions. The sampling schemes were accurate and precise in predicting EGCG oral clearance (or area under the curve with known doses), and hence exposure, under both fasting and fed conditions. The increased predictive performance for estimating pharmacokinetic parameters under fasting conditions appeared to be the result of a decreased variability in absorption.

Population pharmacogenetic‐based pharmacokinetic modeling of efavirenz, 7‐hydroxy‐ and 8‐hydroxyefavirenz

The purpose of this study was to determine the demographic and pharmacogenetic covariates that influence the disposition of efavirenz (EFV) and its major metabolites. A population pharmacokinetic (PK) model was developed from a randomized, cross‐over, drug‐interaction study in healthy male Korean subjects (n = 17). Plasma concentrations of EFV and its hydroxy‐metabolites (0–120 hours) were measured by LC/MS/MS. Genomic DNA was genotyped for variants in the cytochrome P450 (CYP) 2A6, 2B6, 3A5, and MDR1 genes. A PK model was built in a stepwise procedure using nonlinear mixed effect modeling in NONMEM 7. The covariate model was built using the generalized additive modeling and forward selection‐backward elimination. Model‐based simulations were performed to predict EFV steady‐state concentrations following 200, 400, and 600 mg daily oral dose among different CYP2B6 genotypes. The final model included only CYP2B6 genotype as a covariate that predicts EFV clearance through the formation of 8‐OH EFV that represented 65% to 80% of EFV clearance. The total clearance of EFV in CYP2B6*6/*6 genotype was ∼30% lower than CYP2B6*1/*1 or CYP2B6*1/*6 alleles (P < .001). Clopidogrel reduced both formation and elimination clearances of 8‐OH EFV by 22% and 19%, respectively (P = .033 and .041). Other demographics and genotype of accessory CYP pathways did not predict EFV or metabolites PK. CYP2B6 genotype was the only significant predictor of EFV disposition. The developed model may serve as the foundation for further exploration of pharmacogenetic‐based dosing of EFV.

Development of an efficient sampling strategy to predict enoxaparin pharmacokinetics in stage 5 chronic kidney disease.

Patients with renal dysfunction are at greater risk for hemorrhagic events than patients with normal renal function. The objective of this study was to develop an efficient sampling strategy that optimally predicts anti-Xa activity exposure after enoxaparin administration in patients with chronic kidney disease to optimize enoxaparin therapy. The antifactor Xa activity data of 8 anuric patients who were administered 1 mg/kg enoxaparin immediately after hemodialysis were used for the development of the optimal sampling strategies. Maximum A Posteriori Bayesian (MAPB) priors were developed by pharmacokinetic analysis. The 2, 3, and 4 D-optimal sampling designs were determined and Monte Carlo simulations using the MAPB estimator were performed. The original model with the estimator was used to determine the predictive power of the sampling schemes. The most efficient sampling scheme (OSS-2) was accurate and precise in predicting apparent enoxaparin clearance (−3.2% bias, 6.5% precision). The addition of a third sampling time at 30 minutes (OSS-3) was more accurate (P < 0.01) and precise (P < 0.01) than OSS-2 at predicting the rate of absorption (ka) but did not improve the accuracy (P > 0.05) or precision (P = 0.20) of the CLs/F estimate. The determination of antifactor Xa activity at 5 and 24 hours, along with the Bayesian estimators generated from this study, accurately and precisely predict the apparent clearance of antifactor Xa activity. This sampling strategy may be used for therapeutic drug monitoring of enoxaparin and for future clinical trials in patients with chronic kidney disease.