| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
BIOLOGICS |
From Centocor Research and Development, Inc., Malvern, Pennsylvania.
Address for reprints: Honghui Zhou, PhD, FCP, Pharmacokinetics, Modeling & Simulation, Clinical Pharmacology Sciences, Centocor Research and Development, Inc., 200 Great Valley Parkway, Malvern, PA 19355; e-mail: hzhou2{at}cntus.jnj.com.
 |
ABSTRACT |
|---|
 TOP ABSTRACT METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Key Words: Human anti-TNF-
monoclonal antibody • infliximab • population pharmacokinetics • ankylosing spondylitis
) has been found in the sacroiliac joints of patients with AS,5,6 suggesting that the inflammation appears to be at least partly mediated by TNF-.
Infliximab is a recombinant immunoglobulin (Ig) G1
, human-murine chimeric anti-TNF- monoclonal antibody that specifically and potently binds and neutralizes the soluble TNF- homotrimer and its membrane-bound precursor. The safety and efficacy of infliximab was evaluated in patients with AS in the Ankylosing Spondylitis Study for the Evaluation of Recombinant Infliximab Therapy (ASSERT).7
In this study, we report the results of a population pharmacokinetic (PK) analysis of the ASSERT study. The objectives of this population PK analysis were to assess the overall PK characteristics of infliximab in patients with active AS who received repeated intravenous (IV) infusions of infliximab 5 mg/kg or 7.5 mg/kg and also to identify and quantify covariates that may potentially affect infliximab pharmacokinetics in this patient population. The population PK model developed in the current analysis may be useful for the future clinical development of infliximab.
|
METHODS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
3 for 2 consecutive visits, his or her dose was then increased to 7.5 mg/kg every 6 weeks. Patients who did not meet the prespecified criterion for dose escalation continued to receive infliximab 5 mg/kg every 6 weeks through week 96.
|
Serum samples analyzed for antibodies to infliximab were collected prior to the first infusion at week 0 and prior to infusions at weeks 24, 54, 78, 96, and 102. Antibodies to infliximab were determined using an antigen bridging enzyme immunoassay.9 Patients were classified as antibody positive if antibodies to infliximab were detected at any visit. If antibodies were not detected, patients were characterized as either inconclusive or negative for antibodies to infliximab depending on whether a measurable infliximab concentration was present in the serum samples evaluated for anti-infliximab antibodies.
Of the 279 patients who were enrolled in the study, 274 received at least 1 infliximab infusion and had serum infliximab concentration data. The baseline demographic and clinical characteristics of the population included in the PK analysis are summarized in Table I. A total of 8554 serum concentrations, including 8137 measurable (ie, LLOQ) concentrations and 417 preinfusion samples with concentration values <LLOQ, were available for this population PK analysis. Considering that the samples with concentrations <LLOQ accounted for only a very small percentage (ie, approximately 5%) of the total samples, this population PK analysis was primarily performed with concentration values <LLOQ excluded from the data set. However, the impact of concentrations <LLOQ on the PK parameter estimates was evaluated by imputing these concentrations as 
, 
, or 
of the LLOQ value.10
|
Population Pharmacokinetic Model Development
The hierarchical population PK model was built using NONMEM (Double Precision, Version V, Level 1.1). The first-order conditional estimation (FOCE) method with 
-
interaction was used throughout the model-building procedure. One- and 2-compartment models with zero-order infusion and first-order elimination were explored during the structural model development. Interindividual variability (IIV) was described using the exponential error model: Pj = TVP·ej. In the IIV model, Pj is the individual value for the PK parameter in the jth patient, TVP is the typical value of the parameter in the population, and j (ie, eta) is a random effect with a mean of zero and variance of 
2. The residual variability was modeled as a combination of proportional and additive components: Csij = 
sij· (1+ij1) + ij2. In the residual variability model, Csij is the ith concentration measured in the jth patient, sij is the model-predicted concentration, and ij1 and ij2 are error terms with means of zero and variances 
2q and 22, respectively. To examine the IIV in PK parameters, as well as the potential correlations among these parameters, a full omega (
) matrix was initially attempted. During the modeling process, however, the matrix was simplified by reducing the number of correlations when the covariance step failed.
From the output of the base model-fitting procedures, individual estimates of random effects () and individual PK parameter values were obtained using a post hoc empirical Bayesian estimation based on the population parameters (as prior information) and the patient's observed concentrations. Covariate effects on the PK parameters were first explored graphically and then tested computationally through estimation in NONMEM. A "forward-addition/backward-elimination" strategy was used to obtain the full covariate model.
Continuous covariates (eg, age, weight, and serum albumin) on clearance (CL) or volume of distribution (V) were assessed using a multilinear function as expressed in the following equation:
 |
i represents a scaling factor for the effect of that covariate. Other functions such as the power function (eg, allometric equations) or an "Emax-type" function were also examined when appropriate. For example, allometry was used to model the effect of weight or body surface area (BSA) on both CL and V, as shown in the following equation:
 |
i represents the exponent of the power function.
Categorical covariates (eg, sex and binary indicators for concomitant medications) were modeled using one of the following equations:
 |
i represents a scaling factor for the effect of that covariate. The difference in the scaling factor i between the 2 equations is either the covariate effect modeled as an absolute change or as a fractional change. Covariates that individually influenced the PK parameters (ie, as determined by the single covariate model assessments) were added in a descending order of magnitude with regard to reduction of the objective function values (ie, forward-addition method). The likelihood ratio test was the primary criterion used to determine the appropriateness of covariate selection. The covariate was retained in the model if it reduced the objective function value (OFV) by at least 10 points (note: for P < .001 with 1 degree of freedom, the change in OFV is 10.83) compared with the previous model using the log-likelihood ratio test. Because of multiple comparisons inherent to the forward selection procedure, a stringent (ie, P < .001) level of statistical significance was used to make decisions on covariate inclusion. This adjustment abrogated the need to do a Bonferroni correction for multiple comparisons of nested models.11
Alternate models and the relative impact of these covariates on each PK parameter were evaluated by decreasing OFVs (10 points), reducing IIV, using goodness-of-fit plots such as predicted (PRED) versus observed (DV) and weighted residuals (WRES) versus PRED, and reducing the standard error of estimates.
The percentage of IIV (% variance) explained by the covariate(s) for a given PK parameter (eg, CL) was computed as follows:
 |
Population Pharmacokinetic Model Qualification
Model qualification and internal validation procedures were conducted, including a posterior predictive check and a bootstrap analysis.12 A posterior predictive check was used to evaluate the model's predictability by comparing observed concentration data with simulated concentration profiles based on the final PK parameter estimates. Concentration-time profiles of at least 1000 patients were simulated for each of the 2 infliximab dosing regimens (the 5-mg/kg regimen without dose escalation and the 5- to 7.5-mg/kg regimen) using the population PK parameter estimates derived from the final covariate model. The observed concentration versus time data were graphically overlaid with the median values along with the 5th and 95th percentiles from the simulated concentration-time profiles. The predictability of the model was deemed adequate if the observed concentration data were appropriately scattered within the 5th and 95th percentiles of the simulated data.
The bootstrapping method was also used to evaluate the stability and performance of the developed model. With this approach, the population PK parameters were repeatedly estimated by fitting the final population model to a sufficient number of bootstrap replicate data sets (eg, 1000 bootstrap replicates). The median values and 95% confidence intervals (CIs) of the population PK parameter estimates from these 1000 bootstrap data sets were compared with the point estimates (ie, typical PK parameter values) from the original data set used to develop the final model. The model was considered stable if the typical population PK parameter values from the final model fell within the 95% CI of the PK parameter estimates obtained from the bootstrap runs.
|
RESULTS |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
Covariates to Infliximab Clearance
The demographic and clinical characteristics of the patients were examined as potential covariates for infliximab CL (Table I). In addition, the effects of selected concomitant medications on infliximab CL were also evaluated. For a given concomitant medication to be considered as a covariate, this medication was required to be taken by at least 25 patients. Table II summarizes the concomitant medications that were assessed as potential covariates for CL. The potential effect of ethnic differences (ie, race) was not evaluated because 97.8% of patients were Caucasian.
|
According to the specified OFV change criterion (ie, 10.83 at P 
.001, df = 1), antibody-to-infliximab status (IR) (
OFV = –20.1) and white blood cell (WBC) count (OFV = –14.6) independently influenced CL. However, CL was not significantly affected by other covariates such as body weight, BSA, sex, age, baseline disease activity (C-reactive protein [CRP] and BASDAI), liver function enzymes, creatinine clearance, and concomitant use of nonsteroidal anti-inflammatory drugs (NSAIDs), prednisolone, or omeprazole.
Body surface area and body weight are of particular interest because they have been reported as having a significant effect on the PK of other monoclonal antibodies.14,15 Although a weak correlation was noted between infliximab CL and either body weight or BSA, incorporating these 2 covariates into the model using different functions (eg, linear and power functions) did not produce a significant decrease in OFV (OFV = –1 for body weight and OFV = 0 for BSA with allometric functions).
Of the 274 patients included in the analysis, 23 (8.4%) tested positive for antibodies to infliximab, 22 patients were classified as negative for antibodies to infliximab, and 194 patients were classified as inconclusive (ie, antibodies were not found, but the presence of detectable infliximab in the serum may have caused interference in the assay). Thirty-five patients did not have appropriate samples for antibody testing and were classified as having an unknown antibody-to-infliximab status. In the NONMEM data file, the variable IR was coded as a binary variable: positive was set to 1, and all other classifications (negative, inconclusive, and unknown) were set to 0. Incorporating IR as a covariate for CL decreased the OFV by 20.1 points. As shown in Figure 3A,B, the dependency of CL on IR for the base model was adequately accounted for by the final model.
|
The potential effect of WBC on infliximab CL was also assessed. Figure 3C,D shows the random effect (eta) for infliximab CL by WBC before and after incorporating WBC as a covariate. Adding WBC as a covariate resulted in a significant decrease in OFV (OFV = –14.6).
Covariates to Volume of Distribution of Infliximab in the Central Compartment
When appropriate, the effects of certain potential covariates (eg, age, weight, BSA, sex, WBC, serum albumin, total serum protein, and CRP) for the volume of distribution in the central compartment (V1) were also assessed. In search of potentially important covariates to V1, the addition of BSA, weight, sex, and serum albumin to the base model resulted in OFV decreases of 62.0, 39.1, 32.2, and 12.1 points, respectively. Because weight and BSA were highly correlated, and because including BSA led to a greater decrease in OFV than weight, only BSA was selected for inclusion in the final covariate model. However, during the forward selection of covariates to V1, the influence of serum albumin became insignificant after adding BSA and sex on V1.
Figure 4A,B displays the correlation of the random effect (eta) for V1 with BSA for the base and final models, respectively. The addition of BSA led to the greatest decrease in OFV (OFV = –62.0). As shown in Figure 4A,B, the final model with an allometric function for the effect of BSA on V1 adequately accounted for the dependency of V1 on BSA.
|
Final Covariate Population Pharmacokinetic Model
The following set of equations represents the full covariate model, including IR and WBC as covariates for CL and including BSA and sex as covariates for V1.
 |
 |
In the above equations, both WBC and BSA were centered by their median values in the study population, 7.9 x 109/L and 1.94 m2, respectively. In the NONMEM data file, IR was coded as a binary variable with 1 for positive and 0 for all others (ie, negative, inconclusive, and unknown). Sex was coded as 0 for male and 1 for female.
During the model refinement process, covariates that already existed in the full covariate model described above were removed one at a time by fixing the covariate parameter to a null value (ie, 5, 6, 7, or 8 was fixed to 0, one at a time), and the resultant OFV was compared with the previous model. Removal of each of these covariates resulted in significant increases in OFV (OFV = +20.1, +14.0, +40.0, and +11.7 for removal of IR on CL, WBC on CL, BSA on V1, and sex on V1, respectively). Therefore, this full covariate model was considered to be the final covariate model. In all, incorporating the covariates into the model resulted in a total OFV decrease of 107.7 points as compared with the base model.
The parameter estimates and IIV values from the final covariate model are shown in Table III. Compared with the base model, the covariate effects in the final model explained approximately 12.8% of the IIV in CL and 31.9% of the IIV in V1.
|
Goodness-of-fit plots for the final covariate PK model are provided in Figure 5. Figure 5A shows the scatterplot of typical predicted (PRED) versus observed (DV) serum infliximab concentrations, whereas Figure 5B shows the plot of the Bayesian-predicted (IPRED) versus observed (DV) serum infliximab concentrations. Figure 5C shows the weighted residuals (WRES) versus population-predicted (PRED) serum infliximab concentrations. Figure 6 shows the observed and model-predicted concentration-time profiles in a typical patient who received 5-mg/kg infliximab infusions throughout the trial and who was not positive for antibodies to infliximab. These diagnostic plots generally suggested that the final model adequately described the observed concentration-time data, except for the data points with high infliximab concentrations of >300 µg/mL.
|
|
The condition number of 30 (calculated as the ratio of the largest eigenvalue to the smallest eigenvalue) for the final covariate model indicated that the final model was also well conditioned in terms of the matrix and that the final model had adequate stability.16
Qualification of the Final Population Pharmacokinetic Model
Posterior Predictive Check
Simulations were performed with the final covariate model using corresponding covariate values for the 274 patients with AS who were included in the population PK analysis. Concentration-time profiles were simulated for each of the 2 general infliximab dosing regimens (ie, 5 mg/kg without dose escalation and 5 mg/kg with dose escalation to 7.5 mg/kg). As such, the observed concentration data in the placebo to 5-mg/kg infliximab group and the 5-mg/kg infliximab group were combined together and compared with the simulated data for infusions of 5 mg/kg infliximab. It should be noted that because dose escalation from 5 to 7.5 mg/kg could have occurred at any visit between weeks 36 and 96, the simulations for the 5- to 7.5-mg/kg infliximab group were not expected to exactly reflect the clinical scenario of dose escalation in the study.
Concentration data for a total of 1096 patients (4 times 274 patients) were simulated. Most of the observed concentrations fell within the range of values between the 5th and the 95th percentiles of the simulated concentrations (data not shown). There was a slight bias toward underprediction for a small percentage of the observed serum infliximab concentrations, mainly in postinfusion samples with concentration values >300 µg/mL. The reasons for these incongruous concentrations could not be identified. Overall, the final model was able to predict the observed concentration data reasonably well.
Bootstrap Analysis
The final population PK model for infliximab in patients with AS was fitted to 1000 bootstrapped samples to evaluate its stability and performance.17 The 95% confidence intervals (ie, the 2.5th and 97.5th percentiles) for all parameter estimates obtained by bootstrapping are summarized in Table III, along with the parameter estimates of the final model. The typical value of the PK parameter estimates from the final model fell within the 2.5th and 97.5th percentiles of the respective bootstrapping values, indicating that the performance and stability of the final population PK model of infliximab in patients with AS was acceptable.
Comparison of Post Hoc Infliximab Pharmacokinetic Parameters Between Patients With and Without Dose Escalation
Of the 201 patients who were initially randomized to receive treatment with 5 mg/kg infliximab, 106 met the prespecified criterion for dose escalation to 7.5 mg/kg every 6 weeks. Both CL and Vss were similar between the 5-mg/kg group without dose escalation and the 5-mg/kg group with dose escalation (Figure 7A,B).
|
, , or of the LLOQ (ie, 0.05, 0.025, or 0.0125 µg/mL, respectively).
Table IV shows the PK parameter estimates that were derived using the data sets with and without the values <LLOQ (values <LLOQ were imputed as of LLOQ). In general, no appreciable differences were observed in the structural parameter values between the 2 data sets, except for the impact of antibodies to infliximab on the CL. When the values <LLOQ were included and imputed as of the LLOQ (ie, 0.05 µg/mL), the CL for patients who tested positive for antibodies to infliximab was 76.5% higher than for the remaining patients (ie, those who were negative, inconclusive, or unknown for antibodies to infliximab). When all values <LLOQ were removed from the data set, however, the CL for the antibody-positive patients was 41.9% higher than for the remaining patients. Of note, when values <LLOQ were imputed as either or of the LLOQ value, the magnitude of the impact of antibody status on CL was essentially similar (76.6% and 76.7%, respectively) to that when the values <LLOQ were imputed as of the LLOQ value.
|
|
DISCUSSION |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
monoclonal antibodies such as adalimumab18 and golimumab.19 Serum infliximab concentration data in this study were collected prior to the next infusion (trough levels) and 1-hour postinfusion (peak levels), but the infliximab dosing schedules consisting of an induction phase at weeks 0, 2, and 6 followed by a maintenance dosing regimen every 6 weeks enabled us to collect data points 2 and 4 weeks after the previous infusions. In addition, patients randomized to the 5-mg/kg infliximab group received a placebo infusion at week 26, and serum samples were also collected at this visit. These data points collected at weeks 2, 6, and 26, which were not regular trough and peak concentrations, have helped in the construction of a 2-compartment structural model in this study. The final population PK model described the observed concentration-time profiles of infliximab reasonably well. The plot of individual Bayesian-predicted versus observed concentration values for the final model showed a slight bias toward underprediction for a small percentage of the observed serum infliximab concentrations, mainly postinfusion samples with incongruous concentrations >300 µg/mL. Exact causes for these incongruous concentration values could not be identified. Nevertheless, results from the posterior predictive check showed that the final model adequately and accurately estimated the PK parameters and had reasonable predictability. This population PK model was used to characterize the disposition and PK variability of infliximab in patients with AS. Results from the final model showed that the typical systemic CL of infliximab was 0.273 L/day with an IIV of 34.1%. Furthermore, in this study of patients with AS, the systemic CL of infliximab was similar to that (0.27 L/day) reported for patients with rheumatoid arthritis20 but lower than that (0.36 L/day) for patients with Crohn's disease.21 The CL of infliximab in the 2 rheumatologic populations (rheumatoid arthritis and AS) was also similar to the CL (0.24-0.31 L/day) of adalimumab in patients with rheumatoid arthritis.18 The typical volume of distribution for the central compartment (V1) was 3.06 L, which is approximately equal to the human plasma volume, and the IIV for V1 was 17.5%. The typical volume for the peripheral compartment (V2) was 2.94 L. The results for V1 and V2 indicate that infliximab is located primarily in the circulatory system with some distribution to extravascular tissues. The estimate for the volume of distribution of infliximab is also in agreement with those for other IgG-based monoclonal antibodies.22 Attempts to include IIV terms for the intercompartment clearance (Q) and/or V2 were not successful in terms of the minimization step and/or the covariance step. This was not unexpected, however, because of the limited sampling scheme that was used during routine clinical care in this phase III efficacy and safety study. As such, the PK sampling scheme was insufficient to support a full characterization of the distribution phase of infliximab following an IV infusion. Therefore, the final model only included IIV terms for the 2 PK parameters (CL and V1) for which the assessment of IIV is of primary interest.
This study was also conducted to identify and quantify the effects of clinically important covariates on infliximab PK. In all, the significant covariates included in the final model explained 12.8% of the IIV for CL and 31.9% of IIV for V1. This relatively small percentage of IIV in CL that was accounted for by the covariates was not surprising because little is known about the various intrinsic or extrinsic factors that regulate the disposition of monoclonal antibodies.
Of all the covariates evaluated, antibody-to-infliximab status had the most significant influence on infliximab CL. For patients who had detectable antibodies to infliximab, CL was estimated to be 41.9% to 76.7% higher than that for those with negative, inconclusive, or unknown antibody-to-infliximab status. It is of interest that, when concentration values <LLOQ were handled differently (ie, imputed or completely removed), results could vary quite significantly. For example, when values <LLOQ were imputed as the LLOQ (0.05 µg/mL), patients with positive antibody status had 76.5% higher CL than the remaining patients; when all the values <LLOQ were removed from the data set, results showed a 41.9% higher CL for patients with positive antibody status. In other words, complete removal of values <LLLQ from the NONMEM data set would underestimate the impact of antibodies to infliximab on the CL because a greater proportion of trough concentrations was reported as <LLOQ in patients who tested positive for antibodies to infliximab than in patients who did not. These results indicate that the inclusion of observations <LLOQ may be important for estimating the impact of antibodies on CL of the study agent. When LLOQ is low like in this study, inclusion of observations <LLOQ could minimize the bias for the estimate of the impact of antibodies on CL. The accelerated CL of infliximab in patients who tested positive for antibodies to infliximab may explain why a lower proportion (8/16, 50%) of these patients achieved a 20% improvement response according to the Assessment in Ankylosing Spondylitis criteria (ASAS20) at week 102, compared with patients in the inconclusive category (145/183, 79%) (data on file). This was not unexpected because a higher drug exposure was indeed maintained in patients with inconclusive antibody status.
Baseline WBC count was also shown to significantly affect the CL of infliximab. A higher baseline WBC count tended to be associated with an increase in CL. However, because the coefficient (0.0106) was small and the range of WBC counts was generally narrow, this effect is not expected to be clinically relevant. Sun et al15 have recently reported that lymphocyte count independently influenced the CL of efalizumab, an anti-CD11a monoclonal antibody. Because catabolism of IgG antibodies targeted against soluble antigens is generally believed to be mediated by the reticulo-endothelial system,23 elevated lymphocyte levels could presumably be associated with a more rapid catabolism of antibodies.
Body surface area and body weight have frequently been reported as significant covariates for CL of other IgG-based monoclonal antibodies.14,15 However, in this study of patients with AS, the effect of either body weight or BSA on infliximab CL was not significant, albeit a weak correlation was observed.
The laboratory test covariates, such as liver function enzymes and creatinine clearance, did not have statistically significant effects on infliximab CL in patients with AS. A population-based approach24 was also used to assess the impact of the concomitant use of NSAIDs, prednisolone, or omeprazole on infliximab CL. No drug-drug interactions were observed. These findings are not unanticipated. In contrast to many small-molecule drugs, cytochrome P450 enzymes and/or transporters do not play a role in IgG-based monoclonal antibody disposition. Therefore, in most cases, a lack of drug-drug interaction propensity with cytochrome P450-metabolized medications can be reasonably expected for IgG-based monoclonal antibodies such as infliximab.25
When covariate effects on V1 were assessed, both BSA and sex were found to be significant covariates. Body surface area was a major determinant for V1. In addition, women appeared to have a slightly smaller volume of distribution (10%) when compared with men after the correction of BSA. Due to the limited PK data and potential correlations among covariates in the study, sex could be coincidentally identified as a significant covariate to V1. Nevertheless, this marginal impact of sex on V1 should not be of any clinical relevance. Because the V1 for infliximab is dependent on BSA or body weight, an infliximab dosing regimen that is based on body weight may help to reduce variability in peak exposure.
In addition, the correlations between baseline disease activity measures (CRP and BASDAI) and infliximab PK parameters (CL and V1) were assessed. The CRP represents acute inflammatory status, whereas the BASDAI score defines the overall disease activity.8 Nevertheless, neither BASDAI nor CRP showed significant correlation with infliximab PK.
The PK characteristics of infliximab in patients who required dose escalation were similar to the PK characteristics in patients who maintained treatment with 5 mg/kg infliximab every 6 weeks throughout the study. Consequently, the observed interindividual difference in clinical response to infliximab therapy is more likely the result of a pharmacodynamic rather than a PK difference. For example, the extent to which TNF contributes to the pathophysiology of AS may vary significantly among individual patients.
In summary, this study is the first to characterize the population PK of infliximab in patients with AS. Results showed that both the PK characteristics and the observed variability of infliximab in this patient population were typical of intravenous immunoglobulin G1 (IgG1).22,26 Antibody-to-infliximab status was the most significant covariate for systemic CL of infliximab, whereas BSA was the major predictor for the central volume of distribution of infliximab. This population PK analysis also indicates that the development of antibodies to infliximab is associated with accelerated infliximab CL. This may represent an underlying mechanism for an inadequate response or loss of response to infliximab treatment.
The authors thank the patients, investigators, and study personnel who made the ASSERT study possible. They also thank Scott Newcomer, MS, and Rebecca Clemente, PhD, of Centocor, Inc. for their assistance in preparing the manuscript.
Financial disclosure: This study was funded by Centocor Research and Development, Inc. All authors are employees of Centocor Research and Development, Inc and own stock in Johnson & Johnson.
|
REFERENCES |
|---|
|
TOPABSTRACTMETHODSRESULTSDISCUSSIONREFERENCES |
|---|
1. Braun J, van der Heijde D. Novel approaches in the treatment of ankylosing spondylitis and other spondyloarthritides. Expert Opin Investig Drugs. 2003;12: 1097-1109.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
2. Boonen A, van der Heijde D, Landewe R, et al. Work status and productivity costs due to ankylosing spondylitis: comparison of three European countries. Ann Rheum Dis. 2002;61: 429-437.
3. Braun J, Pincus T. Mortality, course of disease and prognosis of patients with ankylosing spondylitis. Clin Exp Rheumatol. 2002;20(suppl 28): S16-S22.[Web of Science][Medline] [Order article via Infotrieve]
4. Lehtinen K. Mortality and causes of death in 398 patients admitted to hospital with ankylosing spondylitis. Ann Rheum Dis. 1993;52: 174-176.
5. Braun J, Bollow M, Neure L, et al. Use of immunohistologic and in situ hybridization techniques in the examination of sacroiliac joint biopsy specimens from patients with ankylosing spondylitis. Arthritis Rheum. 1995;38: 499-505.[Web of Science][Medline] [Order article via Infotrieve]
6. Braun J, Sieper J. Anti-TNF: a new dimension in the pharmacotherapy of the spondyloarthropathies!? Ann Rheum Dis. 2000;59: 404-407.
7. van der Heijde D, Dijkmans B, Geusens P, et al. Efficacy and safety of infliximab in patients with ankylosing spondylitis: results of a randomized, placebo-controlled trial (ASSERT). Arthritis Rheum. 2005;52: 582-591.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
8. Garrett S, Jenkinson T, Kennedy LG, Whitelock H, Gaisford P, Calin A. A new approach to defining disease status in ankylosing spondylitis: the Bath Ankylosing Spondylitis Disease Activity Index. J Rheumatol. 1994;21: 2286-2291.[Web of Science][Medline] [Order article via Infotrieve]
9. Hanauer SB, Wagner CL, Bala M, et al. Incidence and importance of antibody responses to infliximab after maintenance or episodic treatment in Crohn's disease. Clin Gastroenterol Hepatol. 2004;2: 542-553.[CrossRef][Medline] [Order article via Infotrieve]
10. Beal SL. Ways to fit a PK model with some data below the quantification limit. J Pharmacokinet Pharmacodyn. 2001;28: 481-504.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
11. Bland JM, Altman DG. Multiple significance tests: the Bonferroni method. BMJ. 1995;310: 170.
12. Guidance for Industry: Population Pharmacokinetics. Rockville, MD: US Department of Health and Human Services; 1999.
13. Rajagopalan P, Gastonguay MR. Population pharmacokinetics of ciprofloxacin in pediatric patients. J Clin Pharmacol. 2003;43: 698-710.
14. Ng CM, Bruno R, Combs D, Davies B. Population pharmacokinetics of rituximab (anti-CD20 monoclonal antibody) in rheumatoid arthritis patients during a phase II clinical trial. J Clin Pharmacol. 2005;45: 792-801.
15. Sun YN, Lu JF, Joshi A, Compton P, Kwon P, Bruno RA. Population pharmacokinetics of efalizumab (humanized monoclonal anti-CD11a antibody) following long-term subcutaneous weekly dosing in psoriasis subjects. J Clin Pharmacol. 2005;45: 468-476.
16. Montgomery DC, Peck EA. Introduction to Linear Regression Analysis. New York: John Wiley; 1982: 301-302.
17. Parke J, Holford NHG, Charles BG. A procedure for generating bootstrap samples for the validation of nonlinear mixed-effects population models. Comput Methods Programs Biomed. 1999;59: 19-29.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
18. Weisman MH, Moreland LW, Furst DE, et al. Efficacy, pharmacokinetic, and safety assessment of adalimumab, a fully human anti-tumor necrosis factor-alpha monoclonal antibody, in adults with rheumatoid arthritis receiving concomitant methotrexate: a pilot study. Clin Ther. 2003;25: 1700-1721.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
19. Zhou H, Jang H, Fleischmann RM, et al. Pharmacokinetics and safety of golimumab, a fully human anti-TNF-alpha monoclonal antibody, in subjects with rheumatoid arthritis. J Clin Pharmacol. 2007;47: 383-396.
20. Kavanaugh A, St Clair EW, McCune WJ, Braakman T, Lipsky P. Chimeric anti-tumor necrosis factor- monoclonal antibody treatment of patients with rheumatoid arthritis receiving methotrexate therapy. J Rheumatol. 2000;27: 841-850.[Web of Science][Medline]
[Order article via Infotrieve]
21. Boyle A, Tawadros R, Zhu Y, et al. Comparative pharmacokinetics of single and multiple-dose infliximab in Crohn's disease patients. Gastroenterology. 2002;122: A614-A615.[CrossRef]
22. Mascelli MA, Zhou H, Sweet R, et al. Molecular, biologic, and pharmacokinetic properties of monoclonal antibodies: impact of these parameters on early clinical development. J Clin Pharmacol. 2007;47: 553-565.
23. Tabrizi MA, Tseng CML, Roskos LK. Elimination mechanisms of therapeutic monoclonal antibodies. Drug Discov Today. 2006;11: 81-88.[CrossRef][Web of Science][Medline] [Order article via Infotrieve]
24. Zhou H. Population-based assessments of clinical drug-drug interactions: qualitative indices or quantitative measures? J Clin Pharmacol. 2006;46: 1268-1289.
25. Seitz K, Zhou H. Pharmacokinetic drug-drug interaction potentials for therapeutic monoclonal antibodies: reality check. J Clin Pharmacol. 2007;47: 1104-1118.
26. Koleba T, Ensom MHH. Pharmacokinetics of intravenous immunoglobulin: a systematic review. Pharmacotherapy. 2006;26: 813-827.[CrossRef][Web of Science][Medline]
[Order article via Infotrieve]
CiteULike 
Connotea 
Del.icio.us 
Digg 
Reddit 
Technorati What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
