Contamination Levels in Blood Samples Drawn from the Injection Intravenous Line

Abstract

Purpose: Securing two intravenous lines, one for injection and one for blood sampling, can be nearly impossible in compromised patients, therefore, a need exists to quantify the potential error when simplified techniques are employed.

Method: Two venous catheters were placed. 2-deoxy-2-[¹⁸F]fluoro-glucose (FDG) was infused through one of the catheters. Venous blood samples were drawn from each line. Triplicate aliquots of plasma were analyzed in duplicate.

Results: Concentrations from the infusion line were 2.0% higher than the concentrations from the noninfusion line. The average error was 3.3%, 2.0%, and 0.7% higher for the first, second, and third samples, respectively.

Conclusions: Blood sampling through the infusion catheter is a viable alternative to the placement of separate venous catheters. Sampling from the injection catheter, even with tubing flush and replacement, will potentially incur small (generally < 10%) overestimations in concentration in initial samples. Subsequent sampling reduces the error to essentially zero by the third sample.

Introduction

Positron emission tomography (PET) with 2-deoxy-2-[¹⁸F]fluoro-glucose (FDG) is used for the diagnosis, staging, and treatment monitoring of heart, brain, and oncologic disease. Functional tissue information is available from both the images produced, as well as the pharmacokinetic behavior of the tracer. In PET imaging, through the implementation of tracer kinetic models, a number of physiological parameters may be calculated such as tissue blood flow, receptor density or in the case of the use of FDG, glucose metabolic rate. Frequently, the implementation of these models requires some type of blood sampling for the determination of arterial input functions, even when population-based curves are employed. Patients undergoing these studies are frequently compromised to such an extent that securing two separate intravenous lines, one for drug injection and one for venous blood sampling, can be extremely difficult if not impossible. Yet, just as with therapeutic drug monitoring, it is frequently these very same patients that the need for the information derived from blood sampling may be the most urgent. Therefore, a need exists to develop standard techniques for the minimization of contamination and for estimates of the degree of contamination when a single intravenous catheter is utilized, and to quantify the potential sampling error when simplified sampling and assay techniques are employed.

It was the purpose of the present investigation to examine the potential errors associated with simplified blood sampling techniques amenable to clinical situations.