Search

Population pharmacokinetic (PK) modeling serves as the cornerstone for understanding drug behavior within a specific population. It integrates subject covariates to elucidate the variability in PK parameters, thus enhancing predictive accuracy. However, covariate modeling within this framework can be intricate and time-consuming due to the often obscure structural relationship between covariates and PK parameters. Previous attempts, such as deep compartment modeling (DCM), aimed to streamline this process using machine learning techniques. Nonetheless, DCM fell short in assessing residual errors and interindividual variability (IIV), potentially leading to model misspecification and overfitting. Furthermore, DCM lacked the ability to quantify model uncertainty. To address these limitations, we introduce hierarchical deep compartment modeling (HDCM) as an advancement of DCM. HDCM harnesses machine learning to discern the interplay between covariates and PK parameters while simultaneously evaluating diverse levels of random effects and quantifying uncertainty through Bayesian inference. This tutorial provides a comprehensive application of the HDCM workflow using open-source Julia tools., (© 2024 The Author(s). Clinical and Translational Science published by Wiley Periodicals LLC on behalf of American Society for Clinical Pharmacology and Therapeutics.)

Drug side effects that impair cognition can lead to diminished quality of life and discontinuation of therapy. Topiramate is an antiepileptic drug that elicits cognitive deficits more frequently than other antiepileptic drugs, impairing multiple cognitive domains including language, attention, and memory. Although up to 40% of individuals taking topiramate may experience cognitive deficits, we are currently unable to predict which individuals will be most severely affected before administration. The objective of this study was to show the contributions of plasma concentration and working memory capacity in determining the severity of an individual's topiramate-related cognitive impairment. Subjects were enrolled in a double-blind, placebo-controlled crossover study during which they received a single dose of either 100, 150, or 200 mg topiramate. Working memory function was assessed using a modified Sternberg working memory task with 3 memory loads administered 4 hours after dosing. After adjustment for differences in working memory capacity, each 1 μg/mL of topiramate plasma concentration was associated with a 3.6% decrease in accuracy for all memory loads. Placebo effects occurred as a function of working memory capacity, with individuals with high working memory capacity experiencing less severe placebo-related impairment compared with those with low working memory capacity. Our results demonstrate that severity of topiramate-related cognitive deficits occurs as a function of both drug exposure and baseline cognitive function. By identifying patient- and exposure-related characteristics that modulate the severity of cognitive side effects, topiramate dosing strategies may be individually tailored in the future to prevent unwanted cognitive impairment., (© 2020, The American College of Clinical Pharmacology.)

Objective: To determine how early lamotrigine clearance (LTG-CL/F) increases during early pregnancy in women with epilepsy and to quantify the relationship of LTG-CL/F to estradiol concentrations and gestational week., Methods: This was a multicenter, observational study of pregnant women with epilepsy on lamotrigine and no interacting concomitant medications, employing frequent blood sampling prior to and early in pregnancy. A population mixed-effects modeling approach was used to describe the relationship between LTG-CL/F and gestational week and between LTG-CL/F and estradiol. Akaike information criterion (AIC) compared goodness of fit between final models and a generalized estimating equation to compare differences between low and high percentage LTG-CL/F change groups (p < 0.05)., Results: Twenty-five pregnancies (22 participants) were available. Increases in LTG-CL/F were present at 5 weeks gestational age. Both estradiol and gestational week were highly correlated with LTG-CL/F changes; LTG-CL/F increased at the rate of 0.115l/h for every gestational week and 0.844l/h for every 1ng/ml of estradiol, with women in the high LTG-CL/F percentage change group changing at a faster rate (p < 0.001). Models using gestational week performed better than models using estradiol., Interpretation: Gestational week was a better predictor of changes in LTG-CL/F than estradiol concentration and may reflect additional factors, although neither was robust enough to use clinically due to substantial interpatient variability. Changes in LTG-CL/F begin as early as the 5th gestational week, often before women know they are pregnant, emphasizing the importance of planning and early detection of pregnancy and consideration of early implementation of therapeutic drug monitoring. Ann Neurol 2018;84:556-563., (© 2018 American Neurological Association.)

Vaccine responses vary by geographic location. We have previously described how HIV-associated inflammation leads to fibrosis of secondary lymph nodes (LNs) and T cell depletion. We hypothesized that other infections may cause LN inflammation and fibrosis, in a process similar to that seen in HIV infection, which may lead to T cell depletion and affect vaccine responses. We studied LNs of individuals from Kampala, Uganda, before and after yellow fever vaccination (YFV) and found fibrosis in LNs that was similar to that seen in HIV infection. We found blunted antibody responses to YFV that correlated to the amount of LN fibrosis and loss of T cells, including T follicular helper cells. These data suggest that LN fibrosis is not limited to HIV infection and may be associated with impaired immunologic responses to vaccines. This may have an impact on vaccine development, especially for infectious diseases prevalent in the developing world.

In the quest for a functional cure or the eradication of HIV infection, it is necessary to know the sizes of the reservoirs from which infection rebounds after treatment interruption. Thus, we quantified SIV and HIV tissue burdens in tissues of infected nonhuman primates and lymphoid tissue (LT) biopsies from infected humans. Before antiretroviral therapy (ART), LTs contained >98% of the SIV RNA + and DNA + cells. With ART, the numbers of virus (v) RNA+ cells substantially decreased but remained detectable, and their persistence was associated with relatively lower drug concentrations in LT than in peripheral blood. Prolonged ART also decreased the levels of SIV- and HIV-DNA + cells, but the estimated size of the residual tissue burden of 10 8 vDNA + cells potentially containing replication-competent proviruses, along with evidence of continuing virus production in LT despite ART, indicated two important sources for rebound following treatment interruption. The large sizes of these tissue reservoirs underscore challenges in developing 'HIV cure' strategies targeting multiple sources of virus production.

Antiretroviral therapy (ART) suppresses HIV replication in most individuals but cannot eradicate latently infected cells established before ART was initiated. Thus, infection rebounds when treatment is interrupted by reactivation of virus production from this reservoir. Currently, one or a few latently infected resting memory CD4 T cells are thought be the principal source of recrudescent infection, but this estimate is based on peripheral blood rather than lymphoid tissues (LTs), the principal sites of virus production and persistence before initiating ART. We, therefore, examined lymph node (LN) and gut-associated lymphoid tissue (GALT) biopsies from fully suppressed subjects, interrupted therapy, monitored plasma viral load (pVL), and repeated biopsies on 12 individuals as soon as pVL became detectable. Isolated HIV RNA-positive (vRNA+) cells were detected by in situ hybridization in LTs obtained before interruption in several patients. After interruption, multiple foci of vRNA+ cells were detected in 6 of 12 individuals as soon as pVL was measureable and in some subjects, in more than one anatomic site. Minimal estimates of the number of rebounding/founder (R/F) variants were determined by single-gene amplification and sequencing of viral RNA or DNA from peripheral blood mononuclear cells and plasma obtained at or just before viral recrudescence. Sequence analysis revealed a large number of R/F viruses representing recrudescent viremia from multiple sources. Together, these findings are consistent with the origins of recrudescent infection by reactivation from many latently infected cells at multiple sites. The inferred large pool of cells and sites to rekindle recrudescent infection highlights the challenges in eradicating HIV.