By Coleen Curran

For most of the population, the words “matrix effect” conjures up images of Keanu Reeves flying though steel walls in a floor-length leather coat, but for laboratorians it brings to mind something far less entertaining – bad results.

NCCLS defines matrix effect
As defined by the NCCLS, matrix effect is the influence of a sample property, other than the analyte, on the measurement, and thereby on the value of the measurand. In other words, matrix effects in analytical systems are a source of error because they mask the true results.

Streck Laboratories’ Sugar-Chex II glucose control

Although matrix effect can occur across the spectrum of clinical laboratory tests, it tends to become a pronounced problem in hospitals that use a number of different types and brands of point-of-care glucose meters. To control each individual glucose meter correctly, the user commonly employs a brand-specific control. When one considers the number of glucose meters, test strips and control strips residing in the average hospital, it’s easy to see how things can quickly get out of hand.

Verifying the linearity of glucose meters is a fact of life for point-of-care departments. CLIA mandates it, and the CAP survey requires that linearity checks be performed on each new meter before it is used on a patient and then twice a year on working glucometers. In determining a department’s capability of capturing a “true glucose,” it is necessary to verify the linearity of the entire glucose monitoring system — operator technique, reagent strip lot variation and instrument performance.

Unlike glucose controls for wet chemistry analyzers, the suspending medium in controls for point-of-care glucose meters plays a major role in determining the final glucose concentration. Thus, the preparation of glucose controls for POC devices requires that the characteristics of whole blood be present in order to obtain accurate values.

Early glucose controls
Thus, the earliest glucose controls relied on the addition of metabolic inhibitors to fresh blood. Unfortunately, the useful life of such controls was only a few days. Others attempted to make commercial glucose controls with serum and aldehyde-stabilized red blood cells (RBC). While these controls usually gave stable reproducible values with any one strip, between strips and different instruments, the values differed greatly. Further, the values did not agree with the central laboratories analyses.

Building a blood-like linearity check
The goal at Streck Labs was to provide a stable, reproducible glucose control that provides values within a few percentage points of the true value. To do this, the company knew it had to provide a close simulation of the RBC and serum. If a RBC can pass through a two-micron filter in spite of its much larger size, an appropriate surrogate must produce the same effect as the elongated RBC. Movement of control through the test strip must mimic whole blood because the diffusion of glucose through the strip membrane is a rate-limiting step for the enzymatic reaction. And that’s exactly what Sugar-Chex II and Sugar-Chex Linearity accomplishes.

For example, glucose in H2O on YSI = 60 mg/dL. On a test strip, the same sample measures 175 mg/dL. Similar discrepancies are seen with serum. A Sugar-Chex II/Sugar-Chex Linearity glucose control gives good agreement with YSI and other instruments. (Yellow Springs Inc., is the Gold Standard glucose reference analyzer.)

Glucose agreement between POC and the central lab
To achieve glucose agreement with the central laboratory, it is necessary to identify the key elements in a control that help it achieve accurate whole blood simulation. The failure to effectively replicate and combine these key elements is largely responsible for the introduction of matrix effects in both instrument and strip technologies.

Streck Labs has identified these key elements as: a cellular element, viscosity and ionic strength. The cellular element in whole blood contributes viscosity, which in turn affects the regulation of sample flow through the various strip layers. For example, red blood cells physically fill pores on the reagent membrane, thus regulating flow to the subsequent layer. Ionic strength impacts reaction kinetics that in turn shortens or extends the reaction phase. In the absence of either the cellular element or ionic strength, sample migration through the layers is markedly different. Rate of absorption, sample delivery through the reagent membrane and rate/speed of the enzymatic reaction is changed. Ionic strength, combined with close simulation of the cellular elements functional characteristics (size, density, charge and concentration) permit the control solution to function universally and maintain good agreement with the central laboratory.

Bradford A. Hunsley, research and development manager at Streck Labs describes matrix effect as inaccurate measurement of the specified parameter. “If you have a true glucose concentration of 100 mg/dL; a matrix effect would be the recovery of an excessive amount of or an insufficient amount of glucose. The true glucose concentration is being masked in one way or another,” he said.

Precision cannot replace accuracy
Routinely, point-of-care facilities verify the performance characteristics of blood glucose meters according to CLIA requirements. Different glucose meter technologies require different controls. For example, two of the major manufacturers support multiple glucose control formulations for their wide array of instrument technologies. Some would say that the reason for the multiple controls is the failure of any one of them to perform in a manner similar to whole blood. Control solutions that lack the properties of blood (viscosity, ionic strength, rate of absorption, sample delivery through the reagent membrane or the rate of the enzymatic reaction) effectively fail to test the strip’s ability to deliver an accurate and precise glucose measurement. In the absence of accuracy, precision alone (i.e. run-to-run reproducibility) fails to assure the clinician that patient results are correct.

Through research, Streck has identified the essential elements in blood that are relative to point-of-care glucose meter performance and constructed a matrix that closely mimics whole blood. To achieve this mimicry across multiple strip and meter platforms, it recognized the essential features associated with flow rate/viscosity, wetting properties and ionic strength. In combination, these elements produce a glucose composition that closely replicates the behavioral characteristics of fresh whole blood on point-of-care glucose monitoring technologies.

“Many commercially available controls do not contain the cellular element,” Hunsley said. “Given that these strips are designed to handle, manage and separate cellular materials, you can see that the flow characteristics and the reaction kinetics are altered when you apply a control material that the strip is not designed for.”

Baylor Medical’s experience
Baylor University Medical Center of Dallas recently switched its linearity control to Sugar-Chex to help them accurately capture the very high and very low ends of the glucose spectrum.

“With the Level 5, you’re trying to get a higher value and [Sugar-Chex Linearity] was able to produce a higher value,” said Claudia Hill, point-of-care coordinator for Baylor University Medical Center at Waxahachie. “The other product we were using didn’t quite get it as high as we needed it to.”

Kathy Belote, point-of care coordinator at Baylor University Medical Center, Garland reported the same problem, “We had trouble with Level 5 dropping off,” she said. “It was coming out lower than what we like. It was in the range, but when you graphed it, it looked like it was way off. So we and some of the other Baylor affiliate hospitals switched to Streck’s product because we were all having the same problems.”

Like many busy point-of-care departments that strive for optimal patient care, Baylor realized that it’s not enough just to have a reproducible number. Reproducibility and precision must be coupled with accuracy. With Sugar-Chex Linearity, its point-of-care departments are confident that the glucose numbers recovered are meaningful, even in the very high and low ranges.

Coleen Curran is a freelance writer and healthcare consultant based in Charlotte, N.C.