Search PubMed⌕ Search

SEARCH · Search PubMed

Results for “linear mixed model”

Search indexed PubMed citations on genomics, clinical trials, systematic reviews and public health. Explore titles, authors and supplied subject terms, then open the PubMed record.

Quote a phrase for an exact phrase match. Source license links do not imply unrestricted reuse.

At least 1,441 records · Page 80Linked to original sources
PubMed1994

Evaluation of serum prostate-specific antigen velocity after radical prostatectomy to distinguish local recurrence from distant metastases.

OBJECTIVE: Serum prostate-specific antigen (PSA) values are most useful for prediction of disease recurrence after surgery. It is unknown whether a detectable PSA level after surgery indicates a local recurrence potentially benefiting from pelvic irradiation or distant metastases requiring hormonal treatment. METHODS: We analyzed postoperative rate of change of serum PSA levels as a predictor of local versus distant disease recurrence after radical prostatectomy. Between 1982 and 1991, 1,058 men underwent radical prostatectomy for localized prostate cancer and follow-up consisted of determining serum PSA levels and digital rectal examinations. Clinical follow-up of 542 men for four or more years and 78 men for eight or more years yielded ten-year actuarial disease recurrence rates of 4 percent for local recurrence, 8 percent for distant metastases, and 23 percent for an isolated elevation of serum PSA level only. Fifty-one patients with isolated elevations of PSA levels only were followed expectantly until they were diagnosed with either local or distant metastases. RESULTS: A linear mixed effects regression analysis was used to model these data. Using these models, the time to a serum PSA level of 0.5 ng/mL, the PSA level one year following surgery, pathologic stage, Gleason sum, and the rate of change of PSA (PSA velocity [PSAV]) were tested as predictors of local versus distant metastases. A combination of PSAV, pathologic stage, and Gleason grade best distinguished local from distant metastases. CONCLUSIONS: These data suggest that PSAV in men with an isolated elevation of PSA levels following radical prostatectomy might aid in clinical decision making.

Actuarial Analysis↗
PubMed2004

Associations of sole ulcer at claw trimming with reproductive performance, udder health, milk yield, and culling in Swedish dairy cattle.

The association between sole ulcer found at routine claw trimming 2-6 months after calving and reproductive, health and productive traits in the same lactation, was investigated in an observational study of 2368 dairy cows in 102 Swedish herds. The data were analysed by mixed multivariable linear-, logistic- and Poisson-regression modelling at the cow level, accounting for clustering within herd. Associations were found between sole ulcer and first-service conception risk in the first study year (OR 0.59), calving interval (2% longer), treatment for anoestrus (OR 1.61), and 305-day milk yield (479 kg ECM higher). No associations could be shown between sole ulcer and first-service conception risk in the second study year, the number of services per conception, clinical mastitis, high milk somatic-cell counts or culling.

Animal Husbandry↗
PubMed2001

Costs and savings associated with infection control measures that reduced transmission of vancomycin-resistant enterococci in an endemic setting.

OBJECTIVE: To determine the costs and savings of a 15-component infection control program that reduced transmission of vancomycin-resistant enterococci (VRE) in an endemic setting. DESIGN: Evaluation of costs and savings, using historical control data. SETTING: Adult oncology unit of a 650-bed hospital. PARTICIPANTS: Patients with leukemia, lymphoma, and solid tumors, excluding bone marrow transplant recipients. METHODS: Costs and savings with estimated ranges were calculated. Excess length of stay (LOS) associated with VRE bloodstream infection (BSI) was determined by matching VRE BSI patients with VRE-negative patients by oncology diagnosis. Differences in LOS between the matched groups were evaluated using a mixed-effect analysis of variance linear-regression model. RESULTS: The cost of enhanced infection control strategies for 1 year was $116,515. VRE BSI was associated with an increased LOS of 13.7 days. The savings associated with fewer VRE BSI ($123,081), fewer patients with VRE colonization ($2,755), and reductions in antimicrobial use ($179,997) totaled $305,833. Estimated ranges of costs and savings for enhanced infection control strategies were $97,939 to $148,883 for costs and $271,531 to $421,461 for savings. CONCLUSION: The net savings due to enhanced infection control strategies for 1 year was $189,318. Estimates suggest that these strategies would be cost-beneficial for hospital units where the number of patients with VRE BSI is at least six to nine patients per year or if the savings from fewer VRE BSI patients in combination with decreased antimicrobial use equalled $100,000 to $150,000 per year.

Adult↗
PubMed2006

Fast multifield optimization of the biological effect in ion therapy.

In this paper, we present a new technique for simultaneous multifield optimization of the biological effect (i.e. relative biological effectiveness times dose) for intensity modulated radiotherapy with ion beams. It offers complete inverse treatment planning by taking into account planning constraints for the target volume as well as for organs at risk. The approach is based on the mixed irradiation formalism of the linear-quadratic model from radiobiology. We employ a novel objective function to directly optimize the biological effect rather than the physical dose. The required biological input data are reduced to a minimum and are completely independent from the optimization itself. They can be derived from any radiobiological model or even from directly measured data. The new optimization method was fully integrated into our inverse treatment planning tool KonRad. Comparisons with the TRiP98 treatment planning code were done for simple spread-out Bragg peaks as well as for three-dimensional treatment plans, where all fields were optimized separately. While the agreement between both planning systems was very good, the calculation time was substantially reduced in KonRad. By enabling the multifield optimization, the quality of the treatment plans and the sparing of healthy tissues can be clearly improved.

Algorithms↗
PubMed2006

Mixed dentition space analysis in a Jordanian population: comparison of two methods.

AIMS: To test the reliability of the Moyers charts and the Tanaka and Johnston equations in Jordanians and to derive coefficients of correlation between the combined mesiodistal widths of the four permanent mandibular incisors and the combined widths of the canine and premolars of the maxillary and mandibular quadrants. METHODS: The dental models of the dentition of 130 male and 96 female Jordanian subjects (age range is 14-16 years) with complete permanent dentition were randomly selected. The mesiodistal width of the four permanent mandibular incisors, maxillary and mandibular canines, and premolar teeth were measured using a vernier gauge caliper to the nearest 0.1 mm. The actual teeth measurements were then compared with the predicted values derived from the Tanaka and Johnston equations and Moyers probability tables. Correlation and linear regression analyses were performed between the predicted and actual tooth size for Jordanian children and standard regression equations were developed. RESULTS: Except for the maxillary arch in male subjects, Tanaka and Johnston regression equations underestimated the mesiodistal widths of permanent canines and premolars. On the other hand, there were no statistically significant differences between actual mesiodistal widths of canines and premolars and the predicted widths from Moyers charts at the 65% and 75% level for the lower and upper arches in male subjects and at the 85% level for the upper and lower arches in female subjects. For the newly developed regression equations, the correlation coefficients between the sum of the mandibular four incisors and the sum of the canine and premolars were 0.60 (for lower) and 0.51 (for upper) in male subjects and 0.59 and 0.64, respectively, in female subjects. CONCLUSIONS: (1) Tanaka and Johnston prediction method was not accurate when applied to a Jordanian population. (2) Moyers method for prediction can be used for Jordanian children at different probability levels for male and female subjects.

Adolescent↗
PubMed2003

Hypothesis testing in semiparametric additive mixed models.

We consider testing whether the nonparametric function in a semiparametric additive mixed model is a simple fixed degree polynomial, for example, a simple linear function. This test provides a goodness-of-fit test for checking parametric models against nonparametric models. It is based on the mixed-model representation of the smoothing spline estimator of the nonparametric function and the variance component score test by treating the inverse of the smoothing parameter as an extra variance component. We also consider testing the equivalence of two nonparametric functions in semiparametric additive mixed models for two groups, such as treatment and placebo groups. The proposed tests are applied to data from an epidemiological study and a clinical trial and their performance is evaluated through simulations.

Anticonvulsants↗
PubMed2004

Comparison of mathematical model predictions to experimental data of fatigue and performance.

As part of the "Fatigue and Performance Modeling Workshop," six modeling teams made predictions for temporal profiles of fatigue and performance in five different scenarios. One scenario was based on a laboratory study of fatigue and performance during 88 h of extended wakefulness with or without nap opportunities. Another scenario was based on a field study of alertness in freight locomotive engineers. Two scenarios were based on laboratory studies with various conditions of chronic sleep restriction and recovery. There was a theoretical scenario for future ultra-long-range flight operations as well. Experimental data were available for all scenarios except the latter. The model predictions were compared with the experimental data; after linear scaling using mixed-effects regression, mean square errors were computed to quantify goodness-of-fit. The six models were also compared among each other on the basis of these mean square errors. The present paper provides detailed information about the results of these comparisons. The models were capable of predicting the data for some scenarios fairly well. However, predicting the data for the two scenarios involving chronic sleep restriction was more problematic. Differences among the predictions from the six models were relatively small, suggesting that these models have a broad common basis. More experimental research is needed to yield new insights for the further development of fatigue and performance models.

Adult↗
PubMed1996

Latent class analysis in medical research.

In the introduction we give a brief characterization of the usual measures for indicating the quality of diagnostic procedures (sensitivity, specificity and predictive value) and we refer to their relationship to parameters of the latent class model. Different variants of latent class analysis (LCA) for dichotomous data are described in the following: the basic (unconstrained) model, models with parameters fixed to given values and with equality constraints on parameters, multigroup LCA including mixed-group validation, and linear logistic LCA including its relationship to the Rasch model and to the measurement of change in latent subgroups. The problem with the identifiability of latent class models and the possibilities for statistically testing their fit are outlined. The second part refers to latent class models for polytomous data. Special attention is paid to simple variants having fixed and/or equated parameters and to log-linear extension of LCA with its possibility for including on the latent level. Several examples are presented to illustrate typical applications of the model. The paper ends with some warnings that should be taken into consideration by potential users of LCA.

Clinical Trials as Topic↗
PubMed2006

What is syncrystallization? States of the pH indicator methyl red in crystals of phthalic acid.

The concept of syncrystallization was reinvestigated by focusing on phthalic acid (PA) grown with methyl red (MR). Crystals are alternately red and yellow in adjacent growth sectors. X-ray structures of MR and its cocrystals, revealing MR in the neutral, zwitterionic, and protonated states, as well as measurements of linear birefringence and linear dichroism of mixed crystals, were used to investigate mechanisms of PA coloring. These experiments were complemented by force field calculations of the lowest energy stable surfaces of expressed facets and energies of MR on and in crystals, as well as molecular orbital calculations of MR. Two MR species were detected in PA having distinct energies, polarizations, and face selectivities. Assignments of structures to these MRs, previously thought to be neutral and protonated, required a nuanced analysis of hydrogen bonds. The essential difference between yellow and red species is whether the MR carboxylic acid proton is inter- or intramolecularly hydrogen bound. Inferences about mixed crystal structure drawn from an examination of cocrystals of PA and MR are inconsistent with polarization spectroscopy signaling caution when using stoichiometric compounds as models of dilute solid solutions. Upon heating mixed crystals, linear dichroism diminishes and oriented, elongated pools of MR separate and pass through the bulk in directions perpendicular to the direction of elongation. These bâtonnets subsequently crystallize leaving macroscopic oriented crystals of a MR-rich phase within PA. No evidence was found for the simultaneous crystallization of MR and PA; however, the MR reorientation on heating as well as the separation and recrystallization of a MR-rich phase are distinct processes that could be embraced by the literal meaning of syncrystallization.

Journal Article↗
PubMed2003

Bias in assessment of health-related quality of life in a hemodialysis population: a comparison of self-administered and interviewer-administered surveys in the HEMO study.

ABSTRACT. Examined is the relationship of patient-reported health-related quality of life (HRQOL) to the mode of survey administration in the Hemodialysis Study. In addition to self-administered surveys to assess HRQOL, interviewer-administered surveys were made available to include patients with poor vision, decreased manual dexterity, or strong preference. For examining the predictors of participation by self-administration of the survey, multiple logistic regression was performed. For examining the relationship of HRQOL results to mode of survey administration, adjusted differences between the self-administered and interviewer-administered groups were obtained from multiple linear regression models accounting for sociodemographic and case-mix factors. A total of 978 of the first 1000 subjects in the Hemodialysis Study completed the survey by interview (n = 427) or by self-administration (n = 551). The interviewer-administered group was older, was more likely black, had longer duration of ESRD, had a higher prevalence of diabetes, and had more severe comorbidity (all P < 0.01). After adjustment for these differences, patients in the interviewer-administered group had higher scores on scales that measured Role-Physical, Role-Emotional, and Effects of Kidney Disease (all P < 0.001). Dialysis studies that restrict HRQOL measurement to patients who are able to complete surveys without assistance will not accurately represent the health of the overall hemodialysis population. Clinical studies and clinical practices using HRQOL as an outcome should include interviewer administration or risk a selection bias against subjects with older age, minority status, and higher level of comorbidity. Future investigation should include research of survey modalities with a low response burden such as telephone interview, computer-assisted interview, and proxy administration.

Adolescent↗
PubMed2002

Fitting mixed-effects models for repeated ordinal outcomes with the NLMIXED procedure.

This paper presents an analysis of repeated ordinal outcomes arising from two psychological studies. The first case is a repeated measures analysis of variance; the second is a mixed-effects regression in a longitudinal design. In both, the subject-specific variation is modeled by including random effects in the linear predictor (inside a link function) of a generalized linear model. The NLMIXED procedure in SAS is used to fit the mixed-effects models for the categorical response data. The presentation emphasizes the parallel between the model specifications and the SAS statements. The purpose of this paper is to facilitate the use of mixed-effects models in the analysis of repeated ordinal outcomes.

Algorithms↗
PubMed1994

A comparison of population and standard two-stage pharmacokinetic analyses of vigabatrin data.

Vigabatrin (VGB), an irreversible inhibitor of GABA, is being developed as an add-on therapy for uncontrolled complex partial seizure. A single-dose study was conducted in three groups of subjects with normal, mild-to-moderate, and moderate-to-severe renal impairment to examine the effect of renal function on the pharmacokinetics of VGB. Serial blood samples were collected up to 60 h following a single 750 mg oral dose of VGB for the quantitation of drug concentrations. The plasma VGB concentration-time data were analyzed by mixed-effects modeling to estimate population pharmacokinetic parameters and to identify any significant demographic covariates. The parameters of VGB were also calculated by standard two-stage techniques and then compared to the results obtained using the mixed-effects analysis. Population VGB plasma concentration-time profiles were best described by a two-compartment model with zero-order absorption. Creatinine clearance was observed to significantly affect the oral clearance of VGB (p < 0.05), i.e. a linear increasing relationship existed between the two variables. Other demographic factors had no influence on VGB pharmacokinetics. There were agreements in the oral clearance, apparent volume of distribution during elimination, and half-life estimates calculated by both methods. In addition, the conventional technique identified a linear relationship between oral and creatinine clearances. In summary, mixed-effects modeling of serial vigabatrin data validated results determined by the standard two-stage technique.

4-Aminobutyrate Transaminase↗
PubMed1994

The survival of asynchronous V79 cells at low radiation doses: modeling the response of mixed cell populations.

We have observed that when a single linear-quadratic (LQ) function is used to fit the radiation survival response of an asynchronously dividing population of V79 cells, a consistent misfit occurs at low doses. The data can be better described by fitting the low-dose and high-dose ranges separately, and there is evidence of a two-component response. The most obvious explanation is that we may simply be seeing the response of subpopulations of cells of different radiosensitivity: sensitive G1-, G2- and M-phase cells and resistant S-phase cells. The cell sorting assay for cell survival which we have used in these studies may thus be providing sufficient accuracy to resolve these subpopulations, not previously seen in conventional survival measurements. An alternative explanation is that the linear-quadratic function may be inappropriate for accurate description of the radiation survival response at low dose, at least for these cells. To test this hypothesis we have used three other models to fit the data: the single-hit plus multi-target (SHMT) model and the two-parameter repair-misrepair (RMR) model both yielded inferior fits to the asynchronous survival data; the three-parameter RMR model provided an improved fit to the data. The best fit, however, was obtained using a two-population LQ model, which suggested approximately equal numbers of sensitive and resistant cells. When the survival response of tightly synchronized G1/S-phase cells was measured using the cell sorting assay, no substructure was observed. This offers strong support to the hypothesis that the substructure observed in the asynchronous survival response is due to subpopulations of cells of different, cycle-dependent radiosensitivity.

Animals↗
PubMed2006

A warfarin-dosing model in Asians that uses single-nucleotide polymorphisms in vitamin K epoxide reductase complex and cytochrome P450 2C9.

INTRODUCTION: Because of the unique lack of genetic diversity despite the multiethnicity in the Asian population, we hypothesize that single-nucleotide polymorphisms in cytochrome P450 (CYP) 2C9 (CYP2C9*3) and vitamin K epoxide reductase complex subunit 1 (VKORC1) at position 381, used to infer VKORC1haplotype in combination with demographic factors, can accurately predict warfarin doses. The aims of this study were to derive a pharmacogenetics-based dosing algorithm by use of retrospective information and to validate it through a data-splitting method in a separate cohort of equal size. METHODS: We used 215 records of warfarin patients recruited into a CYP2C9/VKORC1 genotyping study to perform this analysis. Univariate analyses for individual predictors, including age, weight, gender, serum albumin concentration, ethnic group, international normalized ratio, and CYP2C9 and VKORC1 381 genotypes, were conducted to select variables with P < .1 for further inclusion into the multivariate regression analysis. In the final model only predictors reaching a statistical significance of P < .05 were retained. RESULTS: Data from 107 subjects undergoing maintenance warfarin therapy with an international normalized ratio stabilized between 2 and 3 were used to derive the final model, as an exponential function of age, weight, CYP2C9*3 allele, and VKORC1 381 CC and TC genotypes, and this model accounted for 60.2% of the variability in daily warfarin dose requirement. The model was validated in a separate cohort of 108 subjects and showed a mean underestimation of 0.23 +/- 1.21 mg/d. CONCLUSION: Warfarin dose requirements in Asians can be accurately predicted by use of a combination of patient demographics and a simplified genotyping approach for single variants in CYP2C9 and VKORC1.

Adult↗
PubMed1999

In vivo comparison of putative probes of CYP3A4/5 activity: erythromycin, dextromethorphan, and verapamil.

OBJECTIVES: Multiple in vivo CYP3A4/5 probes have been proposed. We compared verapamil clearance measures (CYP3A4/5 substrate) to the erythromycin breath test (ERBT) and the cumulative urinary dextromethorphan/3-methoxymorphinan test. METHODS: Clearance of intravenous and oral racemic verapamil and the area under the plasma concentration versus time curve (AUC) ratio of norverapamil (N-demethylated metabolite) to verapamil after oral verapamil dosing, the ERBT, and the dextromethorphan urinary metabolite ratios were measured in 84 healthy nonsmoking subjects (42 men and 42 women; age, 47 +/- 23 (mean +/- SD) years; weight, 69 +/- 11 kg). Relationships between putative CYP3A4/5 probes were assessed by linear regression. RESULTS: The strongest correlation was between intravenous and oral verapamil clearance (r2 = 0.26; P = .0001). Relationships between cumulative urinary dextromethorphan/3-methoxymorphinan and (1) intravenous verapamil clearance (r2 = 0.073; P = .024), (2) oral verapamil clearance (r2 = 0.144; P = .001), and (3) plasma AUC(norverapamil)/AUC(verapamil) after oral verapamil (r2 = 0.10; P = .01) were also detected. The ERBT and intravenous verapamil clearance were weakly related (r2 = 0.04; P = .067). No relationship was detected between ERBT and dextromethorphan/3-methoxymorphinan ratios (r2 = 0.00006; P = .945), oral verapamil clearance (r2 = 0.00006; P = .94), or plasma AUC(norverapamil)/AUC(verapamil) after oral verapamil (r2 = 0.0002; P = .9). CONCLUSIONS: Intravenous and oral verapamil clearance values were significantly correlated, and cumulative dextromethorphan/3-methoxymorphinan urinary ratios correlated with both plasma AUC(norverapamil)/AUC(verapamil) after oral verapamil dosing and with oral and intravenous verapamil clearance. The ERBT correlated only weakly with intravenous verapamil clearance. Results with verapamil are comparable to results with other intravenous and oral CYP3A4/5 probes. Lack of correlation between putative CYP3A4/5 probe results may be attributable to the route of administration; probe characteristics; and intersubject, intrasubject, between-day, and testing measurement variability.

Administration, Oral↗
PubMed2003

Mixed dentition analysis in a Senegalese population: elaboration of prediction tables.

Estimating the size of unerupted teeth is an essential aspect of orthodontic diagnosis and treatment planning in the mixed dentition. Three main methods have been used to estimate the mesiodistal crown widths of unerupted canines and premolars: direct measurements on radiographs; calculations from prediction equations and tables; and a combination of radiographic measurements and prediction tables. Because there are clear racial differences in tooth sizes, the objectives of this study were to produce odontometric data, correlation coefficients between the combined mesiodistal widths of the permanent mandibular incisors and the canine and premolars for each quadrant, and prediction tables with the regression equations, specifically for Senegalese children. Fifty black Senegalese students (25 women, 25 men, mean age 23.50 years) were selected from the University Cheikh Anta Diop in Dakar, Senegal. The mesiodistal crown diameters of the permanent teeth were measured with calipers. Significant sexual dimorphism was found in tooth sizes. The correlation coefficients between the total mesiodistal width of the mandibular permanent incisors and that of the maxillary and mandibular canines and premolars were found to be 0.53 and 0.70, respectively. The standard error of the estimate was better (0.66) for women in the maxilla, and the r(2) values ranged from 0.46 to 0.57 for both sexes. Prediction tables were prepared. The accuracy of the prediction tables should be tested in a larger sample including more ethnic groups.

Adult↗
PubMed2003

Limits for the scaled average bioequivalence of highly variable drugs and drug products.

PURPOSE: To provide a rational procedure for establishing regulatory bioequivalence (BE) limits that can be applied in determinations of scaled average BE for highly-variable (HV) drugs and drug products. METHODS: Two-period crossover BE investigations with either 24 or 36 subjects were simulated with assumptions of a coefficient of variation of 10, 20, 30, or 40%. The decline in the fraction of accepted studies was recorded as the ratio of geometric means (GMR) for the two formulations was raised from 1.00 to 1.45. Acceptance of BE was evaluated by scaled average BE, assuming various BE limits, and, for comparison, by unscaled average BE. A procedure for calculating exact confidence limits in two-period studies is presented, and an approximate method, based on the linearization of the regulatory model, is applied. RESULTS: A mixed model is proposed for average BE. Accordingly, at low variabilities, the BE limit is constant, +/-BELo, generally log(1.25). Beyond a logarithmic, limiting, "switching" variability (sigma(o)), in the region of HV drugs, the approach of scaled average BE is applied with limits of +/-(BEL(o)/sigma(o)). It is demonstrated that the performance of the mixed model corresponds to these expectations. The effect of sigma(o), and of the resulting BE limits is also demonstrated. Scaled average BE, with all reasonable limits for HV drugs, requires fewer subjects than an unscaled average BE. In two-period studies, the exact and approximate methods calculating confidence limits yield very comparable inferences. CONCLUSIONS: Scaled average BE can be effectively applied, with the recommended limits, for determining the BE of HV drugs and drug products. The limiting, "switching" variability (sigma(o)) will have to be established by regulatory authorities.

Computer Simulation↗
PubMed2000

Single and multiple dose pharmacokinetics of nelfinavir and CYP2C19 activity in human immunodeficiency virus-infected patients with chronic liver disease.

AIMS: To evaluate the single-dose and multiple-dose pharmacokinetics of nelfinavir and its active M8 metabolite in eight HIV-seropositive patients with liver disease, and to examine the relationship between CYP2C19 activity (genotype and plasma M8/nelfinavir metabolic ratio) and the severity of liver disease in these patients. METHODS: Nelfinavir was given as a single dose (500 or 750 mg) to patients beginning therapy and twice (500, 750 or 1000 mg) or three times (250 or 750 mg) daily during chronic therapy. Single-dose pharmacokinetic values were used to predict multiple-dose regimens. Peak and total plasma exposures between 2-4 microg ml-1 and 45-75 microg ml-1 h, respectively, and predose levels > 0.7 microg ml-1 were targeted for multidose nelfinavir. Genotype was determined by analysis for CYP2C19*1, CYP2C19*2, and CYP2C19*3. Individuals were grouped according to their genotype, molar M8/nelfinavir AUC ratio (low: < 0.1, intermediate: 0.1-0.3, high > 0.3), and Child-Pugh classification for severity of liver disease. RESULTS: Nelfinavir pharmacokinetics were characterized by wide interindividual variability, low clearance (181-496 ml min-1 70 kg-1, n = 7), and prolonged half-life (5-20 h, n = 7). M8/nelfinavir AUC ratio increased 58% (n = 4) and alpha 1-acid glycoprotein levels decreased up to 39% (n = 5) from single to multiple dosing. CYP2C19 activity was low (metabolic AUC ratio < 0.1) in four patients with moderate to severe liver disease even though they were genetically extensive CYP2C19 metabolizers (*1/*1 or *1/*2). Three patients required lower daily doses than the standard regimen of 750 mg every 8 h to achieve target concentrations and maintain virologic suppression at < 50 RNA copies ml-1 (up to 20 months). CONCLUSIONS: Acquired CYP2C19 deficiency from moderate or severe liver disease resulted in decreased M8 formation. Long-term HIV suppression is possible using low nelfinavir doses in patients with liver disease.

Adult↗