A hierarchical Bayesian approach for combining pharmacokinetic/pharmacodynamic modeling and Phase IIa trial design in orphan drugs: Treating adrenoleukodystrophy with Lorenzo’s oil

Cynthia Basu, Mariam A. Ahmed, Reena V. Kartha, Richard C. Brundage, Gerald V. Raymond, James C. Cloyd, Bradley P. Carlin

Research output: Contribution to journalArticle

5 Citations (Scopus)

Abstract

X-linked adrenoleukodystrophy (X-ALD) is a rare, progressive, and typically fatal neurodegenerative disease. Lorenzo’s oil (LO) is one of the few X-ALD treatments available, but little has been done to establish its clinical efficacy or indications for its use. In this article, we analyze data on 116 male asymptomatic pediatric patients who were administered LO. We offer a hierarchical Bayesian statistical approach to understand LO pharmacokinetics (PK) and pharmacodynamics (PD) resulting from an accumulation of very long-chain fatty acids. We experiment with individual- and observational-level errors and various choices of prior distributions and deal with the limitation of having just one observation per administration of the drug, as opposed to the more usual multiple observations per administration. We link LO dose to the plasma erucic acid concentrations by PK modeling, and then link this concentration to a biomarker (C26, a very long-chain fatty acid) by PD modeling. Next, we design a Bayesian Phase IIa study to estimate precisely what improvements in the biomarker can arise from various LO doses while simultaneously modeling a binary toxicity endpoint. Our Bayesian adaptive algorithm emerges as reasonably robust and efficient while still retaining good classical (frequentist) operating characteristics. Future work looks toward using the results of this trial to design a Phase III study linking LO dose to actual improvements in health status, as measured by the appearance of brain lesions observed via magnetic resonance imaging.

Original languageEnglish (US)
Pages (from-to)1025-1039
Number of pages15
JournalJournal of Biopharmaceutical Statistics
Volume26
Issue number6
DOIs
StatePublished - Nov 1 2016

Fingerprint

Orphan Drug Production
Adrenoleukodystrophy
Pharmacodynamics
Bayes Theorem
Pharmacokinetics
Bayesian Approach
Drugs
Oils
Modeling
Dose
Fatty Acids
Biomarkers
Pediatrics
Operating Characteristics
Magnetic Resonance Imaging
Toxicity
Prior distribution
Adaptive Algorithm
Neurodegenerative Diseases
Health Status

All Science Journal Classification (ASJC) codes

  • Statistics and Probability
  • Pharmacology
  • Pharmacology (medical)

Cite this

Basu, Cynthia ; Ahmed, Mariam A. ; Kartha, Reena V. ; Brundage, Richard C. ; Raymond, Gerald V. ; Cloyd, James C. ; Carlin, Bradley P. / A hierarchical Bayesian approach for combining pharmacokinetic/pharmacodynamic modeling and Phase IIa trial design in orphan drugs : Treating adrenoleukodystrophy with Lorenzo’s oil. In: Journal of Biopharmaceutical Statistics. 2016 ; Vol. 26, No. 6. pp. 1025-1039.
@article{2d9d5e24f25a43a88363267d7684cc14,
title = "A hierarchical Bayesian approach for combining pharmacokinetic/pharmacodynamic modeling and Phase IIa trial design in orphan drugs: Treating adrenoleukodystrophy with Lorenzo’s oil",
abstract = "X-linked adrenoleukodystrophy (X-ALD) is a rare, progressive, and typically fatal neurodegenerative disease. Lorenzo’s oil (LO) is one of the few X-ALD treatments available, but little has been done to establish its clinical efficacy or indications for its use. In this article, we analyze data on 116 male asymptomatic pediatric patients who were administered LO. We offer a hierarchical Bayesian statistical approach to understand LO pharmacokinetics (PK) and pharmacodynamics (PD) resulting from an accumulation of very long-chain fatty acids. We experiment with individual- and observational-level errors and various choices of prior distributions and deal with the limitation of having just one observation per administration of the drug, as opposed to the more usual multiple observations per administration. We link LO dose to the plasma erucic acid concentrations by PK modeling, and then link this concentration to a biomarker (C26, a very long-chain fatty acid) by PD modeling. Next, we design a Bayesian Phase IIa study to estimate precisely what improvements in the biomarker can arise from various LO doses while simultaneously modeling a binary toxicity endpoint. Our Bayesian adaptive algorithm emerges as reasonably robust and efficient while still retaining good classical (frequentist) operating characteristics. Future work looks toward using the results of this trial to design a Phase III study linking LO dose to actual improvements in health status, as measured by the appearance of brain lesions observed via magnetic resonance imaging.",
author = "Cynthia Basu and Ahmed, {Mariam A.} and Kartha, {Reena V.} and Brundage, {Richard C.} and Raymond, {Gerald V.} and Cloyd, {James C.} and Carlin, {Bradley P.}",
year = "2016",
month = "11",
day = "1",
doi = "10.1080/10543406.2016.1226326",
language = "English (US)",
volume = "26",
pages = "1025--1039",
journal = "Journal of Biopharmaceutical Statistics",
issn = "1054-3406",
publisher = "Taylor and Francis Ltd.",
number = "6",

}

A hierarchical Bayesian approach for combining pharmacokinetic/pharmacodynamic modeling and Phase IIa trial design in orphan drugs : Treating adrenoleukodystrophy with Lorenzo’s oil. / Basu, Cynthia; Ahmed, Mariam A.; Kartha, Reena V.; Brundage, Richard C.; Raymond, Gerald V.; Cloyd, James C.; Carlin, Bradley P.

In: Journal of Biopharmaceutical Statistics, Vol. 26, No. 6, 01.11.2016, p. 1025-1039.

Research output: Contribution to journalArticle

TY - JOUR

T1 - A hierarchical Bayesian approach for combining pharmacokinetic/pharmacodynamic modeling and Phase IIa trial design in orphan drugs

T2 - Treating adrenoleukodystrophy with Lorenzo’s oil

AU - Basu, Cynthia

AU - Ahmed, Mariam A.

AU - Kartha, Reena V.

AU - Brundage, Richard C.

AU - Raymond, Gerald V.

AU - Cloyd, James C.

AU - Carlin, Bradley P.

PY - 2016/11/1

Y1 - 2016/11/1

N2 - X-linked adrenoleukodystrophy (X-ALD) is a rare, progressive, and typically fatal neurodegenerative disease. Lorenzo’s oil (LO) is one of the few X-ALD treatments available, but little has been done to establish its clinical efficacy or indications for its use. In this article, we analyze data on 116 male asymptomatic pediatric patients who were administered LO. We offer a hierarchical Bayesian statistical approach to understand LO pharmacokinetics (PK) and pharmacodynamics (PD) resulting from an accumulation of very long-chain fatty acids. We experiment with individual- and observational-level errors and various choices of prior distributions and deal with the limitation of having just one observation per administration of the drug, as opposed to the more usual multiple observations per administration. We link LO dose to the plasma erucic acid concentrations by PK modeling, and then link this concentration to a biomarker (C26, a very long-chain fatty acid) by PD modeling. Next, we design a Bayesian Phase IIa study to estimate precisely what improvements in the biomarker can arise from various LO doses while simultaneously modeling a binary toxicity endpoint. Our Bayesian adaptive algorithm emerges as reasonably robust and efficient while still retaining good classical (frequentist) operating characteristics. Future work looks toward using the results of this trial to design a Phase III study linking LO dose to actual improvements in health status, as measured by the appearance of brain lesions observed via magnetic resonance imaging.

AB - X-linked adrenoleukodystrophy (X-ALD) is a rare, progressive, and typically fatal neurodegenerative disease. Lorenzo’s oil (LO) is one of the few X-ALD treatments available, but little has been done to establish its clinical efficacy or indications for its use. In this article, we analyze data on 116 male asymptomatic pediatric patients who were administered LO. We offer a hierarchical Bayesian statistical approach to understand LO pharmacokinetics (PK) and pharmacodynamics (PD) resulting from an accumulation of very long-chain fatty acids. We experiment with individual- and observational-level errors and various choices of prior distributions and deal with the limitation of having just one observation per administration of the drug, as opposed to the more usual multiple observations per administration. We link LO dose to the plasma erucic acid concentrations by PK modeling, and then link this concentration to a biomarker (C26, a very long-chain fatty acid) by PD modeling. Next, we design a Bayesian Phase IIa study to estimate precisely what improvements in the biomarker can arise from various LO doses while simultaneously modeling a binary toxicity endpoint. Our Bayesian adaptive algorithm emerges as reasonably robust and efficient while still retaining good classical (frequentist) operating characteristics. Future work looks toward using the results of this trial to design a Phase III study linking LO dose to actual improvements in health status, as measured by the appearance of brain lesions observed via magnetic resonance imaging.

UR - http://www.scopus.com/inward/record.url?scp=84989809572&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84989809572&partnerID=8YFLogxK

U2 - 10.1080/10543406.2016.1226326

DO - 10.1080/10543406.2016.1226326

M3 - Article

C2 - 27547896

AN - SCOPUS:84989809572

VL - 26

SP - 1025

EP - 1039

JO - Journal of Biopharmaceutical Statistics

JF - Journal of Biopharmaceutical Statistics

SN - 1054-3406

IS - 6

ER -