The lasso--a novel method for predictive covariate model building in nonlinear mixed effects models

Covariate models for population pharmacokinetics and pharmacodynamics are often built with a stepwise covariate modelling procedure (SCM). When analysing a small dataset this method may produce a covariate model that suffers from selection bias and poor predictive performance. The lasso is a method suggested to remedy these problems. It may also be faster than SCM and provide a validation of the covariate model. The aim of this study was to implement the lasso for covariate selection within NONMEM and to compare this method to SCM. In the lasso all covariates must be standardised to have zero mean and standard deviation one. Subsequently, the model containing all potential covariate-parameter relations is fitted with a restriction: the sum of the absolute covariate coefficients must be smaller than a value, t. The restriction will force some coefficients towards zero while the others are estimated with shrinkage. This means in practice that when fitting the model the covariate relations are tested for inclusion at the same time as the included relations are estimated. For a given SCM analysis, the model size depends on the P-value required for selection. In the lasso the model size instead depends on the value of t which can be estimated using cross-validation. The lasso was implemented as an automated tool using PsN. The method was compared to SCM in 16 scenarios with different dataset sizes, number of investigated covariates and starting models for the covariate analysis. Hundred replicate datasets were created by resampling from a PK-dataset consisting of 721 stroke patients. The two methods were compared primarily on the ability to predict external data, estimate their own predictive performance (external validation), and on the computer run-time. In all 16 scenarios the lasso predicted external data better than SCM with any of the studied P-values (5%, 1% and 0.1%), but the benefit was negligible for large datasets. The lasso cross-validation provided a precise and nearly unbiased estimate of the actual prediction error. On a single processor, the lasso was faster than SCM. Further, the lasso could run completely in parallel whereas SCM must run in steps. In conclusion, the lasso is superior to SCM in obtaining a predictive covariate model on a small dataset or on small subgroups (e.g. rare genotype). Run in parallel the lasso could be much faster than SCM. Using cross-validation, the lasso provides a validation of the covariate model and does not require the user to specify a P-value for selection.

Reduced baroreflex sensitivity in acute alcohol withdrawal syndrome and in abstained alcoholics

Acute alcohol withdrawal is often associated with increased sympathetic activity, and a decreased baroreflex sensitivity (BRS) can be assumed. Parameters of heart rate variability (HRV), blood pressure variability (BPV), BRS as well as cardiac index (CI), left ventricular work index (LVWI) and total peripheral resistance (TPR) were investigated in 20 patients undergoing acute alcohol withdrawal and matched controls. Measures were obtained during the peak of withdrawal symptomatology prior to treatment as well as 2 and 24h under continuous clomethiazole treatment. Alcohol withdrawal scores were obtained and correlated with autonomic measures. In addition, parameters were assessed in 15 subjects who abstained from alcohol after long-term intake. We found a severe down-regulation of BRS during acute alcohol withdrawal and to a milder extent in abstained alcoholics. Furthermore, HRV and BPV did not unequivocally reveal signs of elevated sympathetic activity. Non-linear parameters of HRV and parameters of BRS correlated with the severity of AWS. The distinct decrease of BRS in AWS and in long-term abstained subjects described here is of importance since similar alterations have been identified as independent prognostic factors for cardiac mortality in other diseases.

Modeling drug- and system-related changes in body temperature: application to clomethiazole-induced hypothermia, long-lasting tolerance development, and circadian rhythm in rats

The aim of the present investigation was to develop a pharmacokinetic-pharmacodynamic model for the characterization of clomethiazole (CMZ)-induced hypothermia and the rapid development of long-lasting tolerance in rats while taking into account circadian rhythm in baseline and the influence of handling. CMZ-induced hypothermia and tolerance was measured using body temperature telemetry in male Sprague-Dawley rats, which were given s.c. bolus injections of 0, 15, 150, 300, and 600 micromol kg(-1) and 24-h s.c. continuous infusions of 0, 20, and 40 micromol kg(-1) h(-1) using osmotic pumps. The duration of tolerance was studied by repeated injections of 300 micromol kg(-1) at 3- to 32-day intervals. Plasma exposure to CMZ was obtained in satellite groups of catheterized rats. Fitted population concentration-time profiles served as input for the pharmacodynamic analysis. The asymmetric circadian rhythm in baseline body temperature was successfully described by a novel negative feedback model incorporating external light-dark conditions. An empirical function characterized the transient increase in temperature upon handling of the animal. A feedback model for temperature regulation and tolerance development allowed estimation of CMZ potency at 30 +/- 1 microM. The delay in onset of tolerance was estimated via a series of four transit compartments at 7.6 +/- 2 h. The long-lasting tolerance was assumed to be caused by inactivation of a mediator with an estimated turnover time of 46 +/- 3 days. This multicomponent turnover model was able to quantify the CMZ-induced hypothermia, circadian rhythm in baseline, and rapid onset of a long-lasting tolerance to CMZ in rats.

Rapid and long-lasting tolerance to clomethiazole-induced hypothermia in the rat

Mechanism, onset and duration of tolerance development to clomethiazole-induced hypothermia were investigated in rats using telemetry. The hypothermic effect of clomethiazole was completely abolished for 10 days after an s.c. injection of 300 micromol/kg and the effect returned to approximately 50% in 32 days. The gamma-aminobutyric acidA (GABA(A)) receptor agonist muscimol induced hypothermia at 88 micromol/kg without any (cross-) tolerance. GABA(A) receptor antagonists, bicuculline (5.4 micromol/kg) and picrotoxin (3.3 micromol/kg), did not inhibit clomethiazole-induced hypothermia nor the tolerance. The noncompetitive N-methyl-d-aspartate (NMDA) receptor antagonist, dizocilpine, counteracted clomethiazole-induced hypothermia at 3 micromol/kg but not the tolerance. Tolerance to the 5-hydroxytryptamine(1A) (5-HT(1A)) receptor agonist R-(+)-8-hydroxy-2-(di-n-propylamino)tetralin (R-8-OH-DPAT)-induced hypothermia was blocked by dizocilpine and clomethiazole but not vice versa. No pharmacokinetic interaction was observed. In conclusion, long-lasting tolerance to clomethiazole-induced hypothermia does not involve GABA(A) or 5-HT(1A) receptor functions. Glutamate via NMDA receptors may be involved in the hypothermic response but not in the tolerance.

Population pharmacokinetics of clomethiazole and its effect on the natural course of sedation in acute stroke patients

AIMS

This analysis was performed to investigate the population pharmacokinetics of clomethiazole and its effect on the natural course of sedation in acute stroke patients using a nonlinear mixed effects modelling approach.

METHODS

One thousand five hundred and forty-six acute stroke patients (774 on active treatment) from 166 centres were included in three randomized, double-blind, placebo-controlled phase III efficacy and safety studies. A total dose of 68 mg kg(-1) clomethiazole edisilate was given as a three-phase i.v.-infusion over 24 h. Three blood samples were drawn from all patients to characterize the pharmacokinetics. Sedation was monitored throughout the entire treatment period and the degree of sedation was measured on a discrete ordinal scale with six levels. Models were fitted to the data using the software NONMEM.

RESULTS

Clomethiazole was characterized by a two-compartment pharmacokinetic model with interindividual variability in all structural parameters. For a patient weighing 75 kg, the average CL, V1, Q, and V2 was estimated to be 52.7 l h(-1), 82.5 l, 167 l h(-1) and 335 l, respectively. The interindividual variability in CL, V1, Q and V2 was estimated to be 48%, 53%, 42% and 54%, respectively. Increasing body weight and concomitant administration of liver enzyme inducing drugs were found to increase clearance (by 0.5 l h(-1) kg(-1) and 40%, respectively). Increasing weight also increased the volume of distribution (1.1 l kg(-1) for V1 and 4.7 l kg(-1) for V2). A six-category proportional odds model with a component including the natural course of sedation following placebo administration, a drug component (present or absent) and an interindividual variability component described the degree of sedation. Stroke severity as measured on the NIH-stroke scale on admission and drug treatment were the most important predictors of sedation, but a nonlinear increase in sedation with increasing age was also found. Increasing body weight increased the sedative drug effect.

CONCLUSIONS

The pharmacokinetics of clomethiazole were characterized in acute stroke patients and the analysis excluded several possible covariates of interest in drug development. The time course of sedation could be quantitatively described during the first 24 h following an acute stroke in the presence or absence of clomethiazole treatment.

Pharmacokinetics and sedative effects in healthy subjects and subjects with impaired liver function after continuous infusion of clomethiazole

OBJECTIVE

Clomethiazole is virtually completely eliminated by hepatic metabolism. This study was designed to assess the impact of liver impairment on its elimination and sedative effects.

METHODS

Eight patients with mild liver impairment (Child-Pugh grade A), eight patients with moderate/severe liver impairment (Child-Pugh grade B/C) and eight healthy subjects of similar age were given 68 mg/kg clomethiazole edisilate according to a 24-h infusion scheme aimed at producing minimum sedation as it was intended for clinical use in patients with stroke. Concentrations of clomethiazole and its active alpha-carbon hydroxylated metabolite NLA-715 were followed in plasma and urine for 96 h and 24 h, respectively. Sedation was monitored using a scale from 1 to 6.

RESULTS

The fraction excreted unchanged in urine was less than 0.2% for clomethiazole and less than 0.4% for NLA-715. Urine concentrations of clomethiazole were strongly correlated (r(2)=0.60) to plasma concentrations and approximately equal to unbound plasma concentrations. Plasma levels of NLA-715 increased steadily during the infusion, eventually reaching mean levels exceeding those of clomethiazole in all groups. Plasma clearance of clomethiazole in subjects with mildly impaired liver function was not statistically different from that of healthy controls (40 l/h vs 44 l/h). In subjects with moderate/severe liver impairment, there was a 50% reduction in clearance. Sedation was not observed except in two subjects in the Child-Pugh A group showing mild sedation.

CONCLUSION

The reduced clomethiazole clearance in patients with moderate/severe liver impairment seems to call for a reduction of clomethiazole dosage. However, sedation was not observed in this group at the investigated dose level.

Serum concentration of chlormethiazole and therapeutic effect in acute alcohol withdrawal syndrome: an open clinical trial

It was the aim of this study to find a relationship between the serum concentration of chlormethiazole and its therapeutic effect in acute alcohol withdrawal syndrome. As a secondary subject, the concentration of chlormethiazole was investigated in relation to variables of treatment and variables of physical status of patients. In an open clinical trial, the clinical status of patients was rated by the Mainz Alcohol Withdrawal Scale (MAWS) and the Delirium Rating Scale (DRS). Chlormethiazole concentration was measured by gas-liquid chromatography. Patients were dichotomized according to minimum values of MAWS and DRS after 2 days of treatment (good response and retarded or no response). Chlormethiazole concentration and dose per body weight and MAWS and DRS scores before treatment were compared by the Student t test and the Mann-Whitney test. The two groups were also analyzed by logistic regression with chlormethiazole concentration, MAWS and DRS score before treatment, age, gender, body weight, years of alcoholism, and dose per body weight as independent variables. Chlormethiazole concentration was analyzed by multiple regression with dose, age, gender, smoking, initial alcohol, body weight, and liver dysfunction as independent variables. Forty-three patients were included in the study. Twenty-four patients reached a minimum time of investigation of 2 days. The chlormethiazole concentration was in the range of 0.3 to 5.4 microg/mL at doses of 10 to 24 capsules/d (1 capsule = 192 mg chlormethiazole). As the main result, significantly increased chlormethiazole concentrations were found in patients with retarded or no response; however, in addition the DRS score before treatment and dose per body weight were increased. In addition, the final models of logistic regression contained only DRS score before treatment. As a secondary result, the final model of multiple regression revealed an increased chlormethiazole concentration with dose of chlormethiazole and concentration of alcohol in blood and a decreased chlormethiazole concentration with body weight. This was the first study to investigate the relationship between the chlormethiazole concentration and therapeutic effect in alcohol withdrawal. No robust relationship could be detected that could be separated from the control of treatment by clinical variables. Rather, a poor therapeutic outcome is mainly predicted by an increased initial severity of symptoms, and higher doses are applied in more severely ill patients. Thus, pharmacokinetic control of treatment is not recommended.

The metabolism of clomethiazole in gerbils and the neuroprotective and sedative activity of the metabolites

A single dose of clomethiazole (600 micromol kg(-1) i.p.) has previously been shown to be neuroprotective in the gerbil model of global ischaemia. In gerbils, clomethiazole (600 micromol kg(-1)) injection produced a rapid appearance (peak within 5 min) of drug in plasma and brain and similar clearance (plasma t(1/2): 40 min) from both tissues. The peak brain concentration (226+/-56 nmol g(-1)) was 40% higher than plasma. One major metabolite, 5-(1-hydroxyethyl-2-chloro)-4-methylthiazole (NLA-715) and two minor metabolites 5-(1-hydroxyethyl)-4-methylthiazole (NLA-272) and 5-acetyl-4-methylthiazole (NLA-511) were detected in plasma and brain. Evidence suggested that clomethiazole is metabolized directly to both NLA-715 and NLA-272. Injection of NLA-715, NLA-272 or NLA-511 (each at 600 micromol kg(-1)) produced brain concentrations respectively 2.2, 38 and 92 times greater than seen after clomethiazole (600 micromol kg(-1)). Clomethiazole (600 micromol kg(-1)) injected 60 min after a 5 min bilateral carotid artery occlusion in gerbils attenuated the ischaemia-induced degeneration of the hippocampus by approximately 70%. The metabolites were not neuroprotective at this dose. In mice, clomethiazole (600 micromol kg(-1)) produced peak plasma and brain concentrations approximately 100% higher than in gerbils, drug concentrations in several brain regions were similar but 35% higher than plasma. Clomethiazole (ED(50): 180 micromol kg(-1)) and NLA-715 (ED(50): 240 micromol kg(-1)) inhibited spontaneous locomotor activity. The other metabolites were not sedative (ED(50) >600 micromol kg(-1)). These data suggest that the neuroprotective action of clomethiazole results from an action of the parent compound and that NLA-715 contributes to the sedative activity of the drug. British Journal of Pharmacology (2000) 129, 95 - 100

Cubic phases for studies of drug partition into lipid bilayers

Drug partition into lipid bilayers in a cubic liquid-crystalline phase was investigated. Glyceryl monooleate was used to form the lipid bilayer in a reversed bicontinuous cubic liquid-crystalline phase. The reason for using the cubic phase is that it may coexist with an external aqueous phase, and that the phase boundary (cubic phase/aqueous bulk) is well-defined due to the stiffness of the cubic phase. This makes the cubic phase a potential candidate for high throughput screening (HTS) of the lipophilicity and the dissociation constant (if any) of drug compounds. Clomethiazole (CMZ), lidocaine, prilocaine and 4-phenylbutylamine (4-PBA) were chosen as model drug compounds. It was shown that it is possible to determine a pH-dependent apparent partition coefficient, Kbl/w, of a drug compound using a lipid bilayer expressed as a cubic liquid-crystalline structure. Good agreement was found when the resulting Kbl/w vs. pH curves for CMZ, lidocaine and prilocaine were fitted to a mathematical expression. This included the bilayer/water partition coefficient for the unionised and ionised drug respectively and the pKa of the drug. The effect of different experimental conditions; such as amount of cubic phase, temperature, agitation, sample preparation and interfacial area between the cubic phase and the aqueous bulk on the partition kinetics were investigated as well. The studies reveal that the time needed to reach partition equilibrium was, as expected, substantially reduced (from days to hours) by decreasing the amount of cubic phase, increasing the interfacial area between the cubic phase and the aqueous phase, and increasing the temperature and the agitation of the sample. It was also shown that the bilayer affinity of 4-PBA was increased when a zwitterionic lipid (i.e. dioleoyl phosphatidylcholine, DOPC) was incorporated in the bilayer.

Pharmacokinetics and pharmacodynamics of clomethiazole after oral and rectal administration in healthy subjects

Clomethiazole, a sedative-hypnotic and anticonvulsant drug, has been successfully administered orally and intravenously, but in cases where either of these methods presents complications, rectal administration may represent a practical alternative. We sought to compare the single-dose pharmacokinetics and pharmacodynamics of clomethiazole after oral and rectal administration. Ten healthy adult volunteers were given 600 mg clomethiazole edisylate (corresponding to 390 mg clomethiazole base) in 2 capsules as a single oral or rectal dose in a double-masked, double-dummy, crossover fashion. Serum concentrations were measured up to 10 hours after administration using a specific high-performance liquid chromatography method. Computerized reaction-time measurement and visual analogue scales (VAS) were used to assess drug effects. Peak serum concentrations were significantly higher after oral administration (mean +/- SEM, oral 1.76 +/- 0.47 microg/mL vs rectal 0.48 +/- 0.14 microg/mL; P = 0.03) and appeared earlier (55 +/- 12 vs 89 +/- 11 min; P = 0.04). Area under the concentration-time curve values were similar after administration by both routes (oral 116 +/- 20.6 vs rectal 105 +/- 36.0 microg x min/mL), with a relative rectal bioavailability of 90% compared with oral administration. The objective pharmacodynamic effects on reaction time (increase of 104 +/- 26 vs 66 +/- 22 ms, oral vs rectal) and working speed (decrease of 132 +/- 38 vs 97 +/- 32 ms, oral vs rectal) were not significantly different. Subjective pharmacodynamic effects, as measured on the VAS, were comparable with both routes of administration. Clomethiazole was well tolerated, with a similar adverse effect profile for both routes of administration. The effects of rectal dosing of clomethiazole were similar to those of oral dosing but appeared to occur later. Our results suggest that rectal administration of a single 600-mg clomethiazole edisylate dose bears no safety risk. Therefore, rectal administration could be considered when neither oral nor parenteral administration is possible and a later onset of effect is not critical.

Reticuloendothelial stimulation: levamisole compared

PURPOSE

Levamisole in combination with 5-fluorouracil is an effective adjuvant for the treatment of resected Dukes stage C colon cancer. Since the mechanism of action of levamisole is not known, we have investigated its effects on hepatic and splenic reticuloendothelial system (RES) activity in the rat and compared the effect of levamisole with other known RES stimulators.

METHODS

The hepatic and splenic uptake of an intravenous dose of technetium-99m-sulfur colloid has been used to measure RES activity in rats treated with levamisole, glucan, zymosan, chlormethiazole, octreotide, and saline.

RESULTS

Levamisole significantly increased the hepatic uptake of technetium-99m-sulfur colloid and is comparable in its effect to the other RES stimulators. In contrast, levamisole has no effect on splenic RES activity.

CONCLUSION

RES function is considered to be a potentially important factor in the development of liver metastases, and the stimulatory effect of levamisole on the hepatic RES may partly explain its efficacy as an adjuvant treatment in colon cancer.

Intravenous chlormethiazole

Intravenous chlormethiazole is widely used for its sedative and anticonvulsant action in patients with acute alcohol withdrawal, status epilepticus, pre-eclampsia and eclampsia. Concern remains over its safety if it is given carelessly and without careful monitoring.

A single-dose study of the pharmacodynamic effects of chlormethiazole, temazepam and placebo in elderly parkinsonian patients

Nine elderly parkinsonian volunteers took single doses of 384 mg of chlormethiazole, 10 mg of temazepam and placebo capsules in a double-blind three-way cross-over study on separate visits at least one week apart. In the 6 hours following the dose, the level of drowsiness, performance on a series of psychomotor tests, effects on parkinsonian symptoms and signs, and standing and lying blood pressure were recorded. Chlormethiazole produced drowsiness on all tests and impaired psychomotor performance, as compared with placebo, without affecting parkinsonian symptoms and signs, or postural blood pressure. Temazepam was consistently less potent than chlormethiazole on tests of drowsiness and psychomotor performance. Both treatments were well tolerated. It is suggested that chlormethiazole is safe to use as a hypnotic at this dosage in this group of patients with Parkinson's disease, while temazepam did not appear to be effective as a hypnotic at this dosage.

Uptake and elution of chlormethiazole, meperidine, and minaxolone in the hindquarters of sheep: implications for clearance calculations

Mass balance principles were used to describe the uptake and elution of chlormethiazole, meperidine, and minaxolone in the hindquarters of sheep. Sheep received a right atrial infusion of either chlormethiazole (3.71 mg/min) or meperidine (2.70 mg/min) for 180 min, or minaxolone (0.37 mg/min) for 120 min. Paired arterial and inferior vena cava (draining the hindquarters) blood samples were taken during and after the infusion. The mean and SD (n = 4) of the time-averaged extraction ratios across the hindquarters (determined from the relevant arterio-venous area under blood concentration--time curves) were 0.12 (0.10), 0.36 (0.13), and 0.27 (0.05) for chlormethiazole, meperidine, and minaxolone, respectively. The rank order of the rate of uptake of the drugs into the hindquarters was the same as the rank order of their lipophilicity, and uptake still continued when both the arterial and inferior vena cava drug concentrations were essentially constant. For chlormethiazole, meperidine, and minaxolone, 48% (44), 4% (6), and 35% (17), respectively, of the drug taken into the hindquarters eluted from the hindquarters after the infusion. Drug uptake and retention in extravisceral tissues, represented here by the hindquarters, can result in the mean total body drug clearance being overestimated when determined by traditional systemic pharmacokinetic methods.

A study of the psychometric effects of chlormethiazole in healthy young and elderly subjects

The pharmacokinetics and effects of chlormethiazole (91 mg intravenously together with 356 mg orally) on psychomotor performance were studied in 10 young (mean age 29 years) and 10 older (mean age 66.2 years) volunteers using an open design. Chlormethiazole affected psychomotor function and decreased subjective arousal in both age groups. The peak effect was found at approximately 30 min, i.e. at the end of the infusion, and performance returned to normal by 3 h. There was no evidence of increased sensitivity of the older subjects to the psychomotor or subjective effects of chlormethiazole. The incidence and type of reported symptoms was also similar in the two age groups. The volume of distribution of chlormethiazole was greater in the old than in the young subjects as was the elimination half-life after intravenous administration. Other pharmacokinetic variables showed no significant differences between young and old subjects.