The Novel Use of Structural Equation Modeling to Investigate Oxidative Stress, its Determinants, and its Association with Colorectal Adenoma Pubblico

Abstract

Abstract

The Novel Use of Structural Equation Modeling to Investigate Oxidative Stress, its Determinants, and its Association with Colorectal Adenomas
By Ronald Eldridge

Despite strong basic science evidence demonstrating the role of oxidative stress in carcinogenesis, the results of epidemiologic studies addressing this issue are unconvincing. Oxidative stress is a complex, multifaceted, incompletely understood process that is unobservable in vivo. Numerous biomarkers of in vivo oxidation have been used in research, but none of these serve as a comprehensive measure. Structural equation modeling (SEM) offers the possibility of measuring oxidative stress through a latent (unobserved) variable derived from the shared covariance of multiple imperfect biomarkers, modeled in a system of a priori specified structural (causal) equations.

The primary objective of this dissertation was to investigate the validity and utility of the SEM method to study oxidative stress, its determinants, and its health effects.

In the first study, using three different datasets, I investigated whether a SEM would suitably identify and characterize oxidative stress from five a priori selected biomarkers: F2-isoprostanes, fluorescent oxidation products, mitochondrial DNA copy number, gamma-tocopherol, and C-reactive protein. From the resulting characterization of the latent variable, and its associations with pro- and antioxidant exposures, I determined that the latent variable could be justifiably called "oxidative stress".

In the second study, I investigated the association between the latent oxidative stress variable and newly diagnosed colorectal adenoma. Based on the data from two colonoscopy-based cross-sectional studies, oxidative stress was strongly associated with colorectal adenoma with an odds ratio of 2.61 and a 95% confidence interval of 1.25-5.46 per standard deviation change in oxidative stress.

In the third study, I critically evaluated the causal assumptions of SEM when applied to studies of biologic pathways and constructs. Through multiple Monte Carlo simulations, I examined measurement error, selection bias, and unmeasured confounding using the previously examined models as case studies.

Compared to previous studies that relied on more traditional analytic techniques, the SEM method allows for a better measure of oxidative stress, and yielded a stronger association with colorectal adenoma. The methodology can be applied to other oxidative stress-related health outcomes, or possibly extended to other areas of research where it is necessary to combine different, but imperfect measurements to describe a complex biologic phenomenon.

Table of Contents

Table of Contents
LIST OF FIGURES
LIST OF TABLES
CHAPTER 1. BACKGROUND AND RESEARCH PLAN...1

Section 1. Background...1

Oxidative Stress...1
Structural Equation Modeling...4
Oxidative Stress Biomarkers...7
Relation between Oxidative Stress and Inflammation...10

Section 2. Research Plan...11

Objectives...11
Specific Aims...11
Data Sources...12
Research Plan...14

Section 3. Novelty, Significance, and Impact...25

CHAPTER 2. USING MULTIPLE DETERMINANTS AND BIOMARKERS TO OBTAIN A BETTER MEASUREMENT OF OXIDATIVE STRESS: A LATENT VARIABLE STRUCTURAL EQUATION MODEL APPROACH...27

Abstract...27

Introduction...28

Materials and Methods...30

Results...37

Discussion...41

CHAPTER 3. A NEW LOOK AT OXIDATIVE STRESS AND ITS ASSOCIATION WITH COLORECTAL ADENOMA: A LATENT VARIABLE, STRUCTURAL EQUATION MODELING APPROACH...52

Abstract...52
Introduction...54
Materials and Methods...56
Results...61
Discussion...63

CHAPTER 4. ASSESSING THE CAUSAL ASSUMPTIONS AND METHODOLOGICAL CHALLENGES OF STRUCTURAL EQUATION MODELS IN EPIDEMIOLOGIC STUDIES OF BIOLOGIC CONSTRUCTS...73

Abstract...73
Introduction...74
SEM Assumptions...75
Temporality and Information Bias...77
Selection Bias...83
Confounding...85
Discussion...87

CHAPTER 5. DISCUSSION AND FUTURE DIRECTIONS...98

Overview of Findings...98
Implications and Future Directions...100

REFERENCES...103

About this Dissertation

Rights statement

• Permission granted by the author to include this thesis or dissertation in this repository. All rights reserved by the author. Please contact the author for information regarding the reproduction and use of this thesis or dissertation.

School	Laney Graduate School
Department	Epidemiology
Degree	PhD
Submission	Dissertation
Language	English
Research Field	Health Sciences, Epidemiology
Parola chiave	Oxidative stress Structural Equation Model Biomarker Colorectal adenoma Epidemiology
Committee Chair / Thesis Advisor	Goodman, Michael, Emory University Flanders, W Dana, Emory University
Committee Members	Bostick, Roberd M, Emory University Fedirko, Veronika , Emory University

Ultima modifica

Primary PDF

Thumbnail	Title	Date Uploaded	Actions
	The Novel Use of Structural Equation Modeling to Investigate Oxidative Stress, its Determinants, and its Association with Colorectal Adenoma ()	2018-08-28 10:38:44 -0400	Press to Select an action Download

Supplemental Files

Thumbnail	Title	Date Uploaded	Actions