Search PubMed⌕ Search

Biomedical subjects

L B Sheiner

Publications and source records attributed to L B Sheiner.

At least 37 records · Page 2Linked to original sources
PubMed1996

A population pharmacokinetic model for docetaxel (Taxotere): model building and validation.

A sparse sampling strategy (3 samples per patient, 521 patients) was implemented in 22 Phase 2 studies of docetaxel (Taxotere) at the first treatment cycle for a prospective population pharmacokinetic evaluation. In addition to the 521 Phase 2 patients, 26 (data rich) patients from Phase I studies were included in the analysis. NONMEM analysis of an index set of 280 patients demonstrated that docetaxel clearance (CL) is related to alpha 1-acid glycoprotein (AAG) level, hepatic function (HEP), age (AGE), and body surface area (BSA). The index set population model prediction of CL was compared to that of a naive predictor (NP) using a validation set of 267 patients. Qualitatively, the dependence of CL on AAG, AGE, BSA, and HEP seen in the index set population model was supported in the validation set. Quantitatively, for the validation set patients overall, the performance (bias, precision) of the model was good (7 and 21%, respectively), although not better than that of the NP. However, in all the subpopulations with decreased CL, the model performed better than the NP; the more the CL differed from the population average, the better the performance. For example, in the subpopulation of patients with AAG levels > 2.27 g/L (n = 26), bias and precision of model predictions were 24 and 32% vs. 53 and 53%, respectively, for the NP. The prediction of CL using the model was better (than that of the NP) in 73% of the patients. The population model was redetermined using the whole population of 547 patients and a new covariate, albumin plasma level, was found to be a significant predictor in addition to those found previously. In the final model, HEP, AAG, and BSA are the main predictors of docetaxel CL.

Antineoplastic Agents, Phytogenic↗
PubMed1996

Do we need full compliance data for population pharmacokinetic analysis?

For population pharmacokinetic analysis of multiple oral doses one of the key issues is knowing as precisely as possible the dose inputs in order to fit a model to the input-output (dose-concentration) relationship. Recently developed electronic monitoring devices, placed on pill containers, permit precise records to be obtained over months, of the time/date opening of the container. Such records are reported to be the most reliable measurement of drug taking behavior for ambulatory patients. To investigate strategies for using and summarizing this new abundant information, a Markov chain process model was developed, that simulates compliance data from real data from electronically monitored patients, and data simulations and analyses were conducted. Results indicate that traditional population pharmacokinetic analysis methods that ignore actual dosing information tend to estimate biased clearance and volume and markedly overestimate random interindividual variability. The best dosing information summarization strategies consist of initially estimating population pharmacokinetic parameters, using no covariates and only a limited number of dose records, the latter chosen based on an a priori estimate of the half-life of the drug in the compartment of interest; then resummarizing the dose records using either population or individual posterior Bayes parameter estimates from the first population fit; and finally reestimating the population parameters using the newly summarized dose records. Such summarization strategies yield the same parameter estimates as using full dosing information records while reducing by at least 75% the CPU time needed for a population pharmacokinetic analysis.

Administration, Oral↗
PubMed1996

Modeling a bivariate control system: LH and testosterone response to the GnRH antagonist antide.

A pharmacodynamic analysis of the input-response relationship between the gonadotropin-releasing hormone antagonist antide and luteinizing hormone (LH) and testosterone concentrations is presented. A control compartmental model is developed using pharmacokinetic and pharmacodynamic data from experiments in which different short intravenous antide infusions were given to healthy male volunteers. Because of the control interdependence between serum LH and testosterone a separation principle similar to one we have used previously to analyze physiological pharmacokinetic data is used for model exploration: testosterone and LH are first modeled separately, conditioning on the other observed response. This reveals that the LH effect on testosterone depends on previous LH exposure and that LH depends not on current but on previous testosterone exposure, resulting in an LH overshoot after antide-induced suppression. Both submodels are combined into one global model, which in addition includes a model for testosterone circadian variation. This model describes the data well and can be used to predict responses for some nonstudied antide dosages. However, the sensitivity of predictions to model assumptions limits the range of valid extrapolation, and this, too, is illustrated.

Humans↗
PubMed1995

Three new residual error models for population PK/PD analyses.

Residual error models, traditionally used in population pharmacokinetic analyses, have been developed as if all sources of error have properties similar to those of assay error. Since assay error often is only a minor part of the difference between predicted and observed concentrations, other sources, with potentially other properties, should be considered. We have simulated three complex error structures. The first model acknowledges two separate sources of residual error, replication error plus pure residual (assay) error. Simulation results for this case suggest that ignoring these separate sources of error does not adversely affect parameter estimates. The second model allows serially correlated errors, as may occur with structural model misspecification. Ignoring this error structure leads to biased random-effect parameter estimates. A simple autocorrelation model, where the correlation between two errors is assumed to decrease exponentially with the time between them, provides more accurate estimates of the variability parameters in this case. The third model allows time-dependent error magnitude. This may be caused, for example, by inaccurate sample timing. A time-constant error model fit to time-varying error data can lead to bias in all population parameter estimates. A simple two-step time-dependent error model is sufficient to improve parameter estimates, even when the true time dependence is more complex. Using a real data set, we also illustrate the use of the different error models to facilitate the model building process, to provide information about error sources, and to provide more accurate parameter estimates.

Models, Theoretical↗
PubMed1995

A nonlinear mixed-effects pharmacokinetic model comparing two formulations of cyclosporine in stable renal transplant patients.

A nonlinear mixed-effects model simultaneously modeled two pharmacokinetic (PK) variables in patients administered cyclosporine twice daily: (i) concentration of drug in blood at the end of the 12-hr dosing interval (C12) and (ii) area under the concentration-time curve within the dosing interval (AUC). For two formulations (Neoral and Sandimmune), the model assessed the following: nonlinearity with respect to dose, interoccasion (intraindividual) variability, interindividual variability, and within- and across-individual correlation between C12 and AUC. Data were pooled from six clinical studies in stable renal transplant patients administered each formulation. PK samples on two occasions were taken usually for each formulation. Each individual's random effect was eight-dimensional consisting of two PK variables for each formulation on two occasions. An ANOVA-like partitioning worked well and reduced the variance matrix for the random effect to a known function of 13 parameters to be estimated, thereby making a numerically intensive computation feasible. Simulations were used to check the model fit, to compute standard errors, and to account for peculiarities in the residual analysis. Outcomes of tests comparing formulations, most of which were statistically significant, favored Neoral (dose proportional, lower interoccasion variability, lower interindividual variability, and higher correlation between C12 and AUC).

Computer Simulation↗
PubMed1995

A population model for the leukopenic effect of etoposide.

We present a new model-dependent approach to quantify hematologic toxicity in a patient population after anticancer therapy. The population model consists of three submodels that are simultaneously fit to the data: (1) a cubic spline function describing the average response of the population versus time ("structural model"), (2) a covariate model, which relates parameters of the structural model to measured demographic or therapeutic variables that are found to be of predictive value (in this study: white blood cell (WBC) count baseline, drug concentration, serum albumin, and serum bilirubin concentration), and (3) a variance model, which estimates the contribution to the response from random variability between patients and from variability within patients, both between courses and within courses, between days. To demonstrate the approach, previously reported data from 118 courses of etoposide therapy in 71 patients with cancer were used to model the decrease in WBC count after 3-day continuous infusions of drug. The estimated typical response profile is characterized by (1) a lag-time of 4 1/2 days before any WBC count decline is observed, (2) a duration of time below baseline of 22 days, and (3) half-maximal effect (i.e., decrease to 50% of baseline WBC count) after exposure to C50 = 3 mg/L etoposide (mean) over 3 days. Lower serum albumin concentrations, higher bilirubin concentrations, or both are associated with greater effects at a given etoposide exposure. Large variability in the estimated response was found between individuals and within individuals, between courses. The total variabilities (SD) in lag-time, duration of the decrease, and C50 were 1 day, 6 days, and 1.8 mg/L, respectively. The population model can also be used to predict the consequence of as-yet untested therapy and sampling strategies, as well as to relate acceptable risks of toxicity to target drug exposure.

Adult↗
PubMed1995

A new method to explore the distribution of interindividual random effects in non-linear mixed effects models.

This article presents a new approach for exploring the distribution of interindividual random effects in nonlinear mixed effect models. The approach introduces a spline function, which transforms an assumed normally distributed interindividual random effect to an arbitrary distribution approximating that of the data. The performance of this tool is illustrated using simulated pharmacokinetic data with non-normally distributed random effects. Results of the analyses of two real kinetic data sets are also presented.

Biometry↗
PubMed1994

Comparison of the Akaike Information Criterion, the Schwarz criterion and the F test as guides to model selection.

In pharmacokinetic data analysis, it is frequently necessary to select the number of exponential terms in a polyexponential expression used to describe the concentration-time relationship. The performance characteristics of several selection criteria, the Akaike Information Criterion (AIC), and the Schwarz Criterion (SC), and the F test (alpha = 0.05), were examined using Monte Carlo simulations. In particular, the ability of these criteria to select the correct model, to select a model allowing estimation of pharmacokinetic parameters with small bias and good precision, and to select a model allowing precise predictions of concentration was evaluated. To some extent interrelationships among these procedures is explainable. Results indicate that the F test tends to choose the simpler model more often than does either the AIC or SC, even when the more complex model is correct. Also, the F test is more sensitive to deficient sampling designs. Clearance estimates are generally very robust to the choice of the wrong model. Other pharmacokinetic parameters are more sensitive to model choice, particularly the apparent elimination rate constant. Prediction of concentrations is generally more precise when the correct model is chosen. The tendency for the F test (alpha = 0.05) to choose the simpler model must be considered relative to the objectives of the study.

Mathematics↗
PubMed1994

A new approach to the analysis of analgesic drug trials, illustrated with bromfenac data.

A clinical trial of an analgesic agent compares pain relief scores (ordered categorical responses) over time among groups of patients, each subject to a painful procedure and given various doses of active agent (including zero, i.e., placebo) on demand. Patients may elect to remedicate with an active agent if their pain relief is insufficient, so the sample of patients at any given time is biased toward those with better relief. Standard analyses usually (1) fill in the missing data but make no correction for so doing and (2) treat the ordered categorical variable as continuous. Both of these create problems in interpretation and inference, but the former is more serious than the latter. An alternative analysis has been recently proposed that deals with these problems. This article presents that method for a nonstatistical audience and illustrates its use on some data from the analgesic bromfenac.

Analgesics↗
PubMed1994

Estimating bioavailability when clearance varies with time.

The influence of interoccasion variability in clearance on bioavailability estimates from a traditional two-period crossover design is reported for five methods of analysis: (1) the standard crossover analysis, (2) a groupwise, parallel, analysis, (3) and (4) two correction procedures suggested by J.G. Wagner and by P.S. Collier and S. Riegelman, and (5) a pharmacokinetic nonlinear mixed-effects model analysis. Three bioavailability parameters are considered the population mean bioavailability (F), the interindividual variance of bioavailability (omega 2F) and the correlation of bioavailability with clearance [cor (CL,F)]. Data are simulated with different degrees of interoccasion variability and/or non-zero cor(CL,F). With the standard crossover analysis of these data, estimates of F, omega F, and cor(CL,F) are all biased in the presence of interoccasion variability in clearance. Estimates of F and omega 2F obtained from the parallel-group analysis are not reliable because the approach relies on the assumption that cor(CL,F) is zero. The two correction procedures are very sensitive to random error in the estimates of terminal half-life. The mixed-effect model approach produces unbiased estimates of all three bioavailability parameters. These results from simulations are supported by a real data example.

Administration, Oral↗
PubMed1994

Comparison of twitch depression of the adductor pollicis and the respiratory muscles. Pharmacodynamic modeling without plasma concentrations.

BACKGROUND: Although the respiratory muscles (the diaphragm and the laryngeal adductors) recover from paralysis more rapidly than does the adductor pollicis, patients can develop complete paralysis of the respiratory muscles, but not of the adductor pollicis, after bolus administration of vecuronium. The authors used a pharmacodynamic model not requiring muscle relaxant plasma concentrations to reconcile these findings. METHODS: The pharmacodynamic model is based on the traditional model, in which: (1) vecuronium concentration at the neuromuscular junction (C(effect)) is a function of the plasma concentration versus time curve and a rate constant for equilibration between plasma and the neuromuscular junction (k(eo)); and (2) effect is a function of C(effect), the steady-state plasma concentration that produces 50% effect (C50), and a factor to explain the sigmoid relationship between concentration and effect. In the absence of vecuronium plasma concentrations, an empiric model (rather than the usual effect compartment model) can be used to mimic the time delay (proportional, but not identical, to 1/k(eo)) between dose and effect. The model can be used to estimate the steady-state infusion rate that produces 50% effect (IR50), equal to the product of C50 and vecuronium plasma clearance; IR50 for different muscle groups then can be compared to assess relative sensitivity. The authors applied this model to published effect data for subjects given 40-70 micrograms/kg vecuronium in whom paralysis of three muscle groups was measured during opioid/propofol anesthesia. RESULTS: For IR50, the ratio of values for the larynx:diaphragm:adductor pollicis was 1.4:1.2:1; for the equilibration constant (inversely proportional to the time delay), the ratio for the respiratory muscles to the adductor pollicis was 2.5:1. CONCLUSIONS: Vecuronium concentrations peak earlier at the respiratory muscles than at the adductor pollicis, possibly the result of greater perfusion to these organs, leading to earlier onset of paralysis. The observation that bolus injection of vecuronium produces greater paralysis of the respiratory muscles than of the adductor pollicis, despite greater resistance of the respiratory muscles, can be explained by differential rates of equilibration between plasma and various muscles.

Humans↗
PubMed1994

Semiparametric models for antagonistic drug interactions.

A new class of models to describe antagonistic drug interactions are presented. They are semiparametric in that they use nonparametric functions (splines) but are forced to obey certain constraints corresponding to reasonable assumptions. We propose the models primarily for exploratory data analysis, but they may also be definitive models for such purposes as predicting future responses. Certain problems that arise in semiparametric modeling, such as model selection, are addressed so that we can propose a relatively automatic and objective approach to model determination. We demonstrate the applicability of the class of models we propose to two real data set examples involving pain relief response to opioid agonists/antagonists. The results suggest that the semiparametric approach is particularly useful when unusual shapes link dose to response.

Animals↗
PubMed1994

Case-mix adjustment using objective measures of severity: the case for laboratory data.

OBJECTIVE: We evaluate the use of routinely gathered laboratory data to subclassify surgical and nonsurgical major diagnostic categories into groups homogeneous with respect to length of stay (LOS). DATA SOURCES AND STUDY SETTING: The source of data is the Combined Patient Experience database (COPE), created by merging data from computerized sources at the University of California San Francisco (UCSF) Medical Center and Stanford University Medical Center for a total sample size of 73,117 patient admissions. STUDY DESIGN: The study is cross-sectional and retrospective. All data were extracted from COPE consecutive admissions; the unit of analysis is an admission. The outcome variable LOS proxies hospital resource utilization for an inpatient stay. Nine (candidate) predictor variables were derived from seven lab tests (WBC, Na, K, C02, BUN, ALB, HCT) by recording the whole-stay minimum or maximum test result. DATA COLLECTION/EXTRACTION METHODS: Patient groups were formed by first assigning to major diagnostic categories (MDCs) all 73,117 admissions. Each MDC was then partitioned into medical and surgical subgroups (sub-MDCs). The 13 sub-MDCs selected for study define a study population of 32,599 patients that represents approximately 45 percent of inpatients. Within each of the 13 sub-MDCs, patients were randomly assigned to one of two data sets in a ratio of 2:1. The first set was used to create, the second to validate, three different LOS predictors. Predictive accuracies of individual DRG classes were compared with those of two alternative classification schemes, one formed by recursive partitioning (the sub-MDC) using only lab test results, the other by partitioning with both lab test results and individual DRGs. PRINCIPAL FINDINGS: For the eight largest sub-MDCs (81 percent of study population), individual DRGs explained 23 percent of the within sub-MDC variance in LOS, laboratory data classes explained 31 percent, and classes derived by considering individual DRGs and laboratory data explained 37 percent. (Each result is a weighted average R2. The average number of LOS classes into which the eight largest sub-MDCs were partitioned were 20, 10, and 10, respectively. Within six of the eight, partitioning on the basis of laboratory data alone explained more within sub-MDC variance than did partitioning into individual DRGs. CONCLUSIONS: Routine lab test data improve the accuracy of LOS prediction over that possible using DRG classes. We note that the improvements do not result from overfitting the data, since the numbers of LOS classes we use to predict LOS are considerably fewer than the numbers of individual DRGs.

California↗
PubMed1993

The importance of modeling interoccasion variability in population pharmacokinetic analyses.

Individual pharmacokinetic parameters may change randomly between study occasions. Analysis of simulated data with NONMEM shows that ignoring such interoccasion variability (IOV) may result in biased population parameter estimates. Particular parameters affected and the extent to which they are biased depend on study design and the magnitude of IOV and interindividual variability. Neglecting IOV also results in a high incidence of statistically significant spurious period effects. Perhaps most important, ignoring IOV can lead to a falsely optimistic impression of the potential value of therapeutic drug monitoring. A model incorporating IOV was developed and its performance in the presence and absence of IOV was evaluated. The IOV model performs well with respect to both model selection and population parameter estimation in all circumstances studied. Analysis of two real data examples using this model reveals significant IOV in all parameters for both drugs and supports the simulation findings for the case that IOV is ignored: predictable biases occur in parameter estimates and previously nonexistent period effects are found.

Computer Simulation↗