The Design and Analysis of Calcium-41 Clinical Trials to Study Treatments for Bone Loss

Courtney N Henry, Purdue University

Abstract

Calcium-41 (41Ca) is a biological tracer used to rapidly and sensitively monitor skeletal calcium metabolism. Measurements of the 41Ca:Ca ratio in urine quantify changes in net bone turnover (i.e. differences in bone formation and resorption). A biomedical application of 41Ca technology clinically assesses the efficacy of treatments for preventing bone loss, primarily targeting susceptible populations such as post-menopausal women. This research initially describes a 41Ca simulation study to determine efficient 41Ca experimental designs that preserve the accuracy and reliability of estimation and the statistical power of our conventional design, whilst substantially lowering costs by reducing the number of samples and/or study time. The simulation study investigates how sample size, number of study periods, number of samples per period, and number of days between sample collection influence results. Specifically, we quantify their effect on (1) the accuracy (standard error) of reported mean estimates of treatment, (2) the power of the study to detect given effect sizes, and (3) the cost of the study. We conclude with a recommendation of a new experimental design which maintains the statistical integrity of our conventional design, and significantly reduces the analytical (and logistical) expenditures. Ultimately, we anticipate these new designs to increase the financial feasibility and practicality of 41Ca clinical experimentation. The final research component details a (step-by-step) modeling procedure for characterizing the 41Ca urinary excretion curve (in the absence of treatment). This modeled curve provides a fixed index from which perturbations due to treatment can be directly quantified (once steady-state kinetics have been established). Moreover, this model may be used to assess and/or determine the clinical study protocol for the 41Ca equilibration period (i.e. time at steady-state), at which time the 41Ca:Ca signal is indicative of net bone turnover. We propose a 5-parameter, non-linear spline model formulation for this curve, where model parameters describe individual subject kinetics. We intend for this modeling procedure to serve as a standard procedure for future 41Ca data handling and analysis of the dynamic behavior of 41Ca urinary excretion in humans. We anticipate that the compilation of both research products (i.e. simulation study and modeling procedure) will further promote the widespread access and utility of 41Ca technology and clinical experimentation for advancing the discovery of novel therapies for osteoporosis prevention.

Degree

Ph.D.

Advisors

McCabe, Purdue University.

Subject Area

Statistics|Nutrition

Off-Campus Purdue Users:
To access this dissertation, please log in to our
proxy server
.

Share

COinS