Vol. 19 • Issue 20 • Page 28
With the emergence of evidence-based medicine and practice guidelines based on specific lab tests and defined cutoffs, the need for lab method standardization is increasingly important. However, healthcare providers are often unaware of the significant differences in results observed when different measurement methods are employed. The potential impact to the healthcare system is enormous; accordingly, both labs and manufacturers of analytical systems and reagents must make every effort to standardize their methods.
While some progress in standardization is being made regarding well-defined analytes such as cholesterol, the list of concerns in the standardization of immunoassays is a complex one lacking clear solutions. Issues involve everything from the analytes and antibodies to reporting of end results. Although there have been -ongoing -conversations throughout the industry regarding these problems and the need for standardization, success has been limited.
Assessing the Issues
To start, immunoassays are often not a direct measurement of the analyte. Unlike many chemistry assays, immunoassays are an indirect measurement-we’re actually assessing what’s left after the patient’s antigen has competed with the label antigen. In competitive assays, for example, the reagent might be a chemiluminescent or enzyme label antigen competing with the patient’s antigen for antibody binding sites. Thus, it’s the patient’s antigen that indirectly determines how much of the label antigen will actually bind to the antibody. The measurement is then related back indirectly to the amount of analyte that must have been in the patient sample.
The laboratorian operates under the assumption that the antibody reacts with the label antigen in the same manner as the unlabeled patient antigen, i.e., that the competition for antigen binding is equivalent. However, the label may slow down the reaction or impair it in some way so that the laboratorian does not see a true recovery of the patient antigen. Also, large labels can even obscure antigen binding sites to a degree. Size of the label selected does matter, especially in competitive asays. The efficiency of separation of antibody-bound antigen from free is also a difficult aspect to control in competitive immunoassays. Non-competitive sandwich assays for larger antigens have similar complexity in terms of antibody-antigen issues influencing recovery and making standardization difficult; these methods also involve multiple wash steps.
Another challenge lies in the differences in antibody development by vendors. Even among the same vendor, each batch of antibodies might measure or bind to antigens differently enough to show significant bias with earlier lots of reagents. Some vendors have taken impressive steps to overcome this hurdle; we’ll discuss that below.
The list of problems continues. The analytes a laboratorian is trying to measure in immunoassays most often do not exist in a purified, weighed-out form. It’s very difficult to get a chemically well-characterized peptide like human chorionic gonadotropin (HCG) or troponin into an isolated, purified material and determine its exact weight. Even if proteins for immunoassays could be purified to the degree required, the purification technique could change the structure, raising a whole new set of issues. Many of these proteins and peptides are not particularly stable in the first place, thus making storage after ultrapurification a risk.
An approach that has worked reasonably well is to have a group such as the World Health Organization (WHO) define a material and assign a value to it, allowing manufacturers and labs to buy that material to test their calibration to see if their methods recover the assigned value for that reference preparation. Even so, subtle differences between methods and labs result in different normal ranges. The ultimate goal, of course, is to be able to get the same results and ranges from the same tests in different labs to eliminate physician confusion.
Establishing a Reference Point
Agreement on the point of reference is necessary for standardization. For analytes with no suitable WHO standard or International Reference Preparation available, the common point of reference could be based on a commercially available method used for the original clinical studies for that analyte.
Let’s examine prostate specific antigen (PSA) testing as an example. The first commercial PSA immunoassay included many studies that assessed men over the age of 40 and looked at their digital rectal exam, biopsy results, signs and symptoms. The manufacturer showed that PSA, at a certain cutoff point, is useful for diagnosing and managing prostate cancer. As other vendors offered PSA assays, they were required to correlate almost exactly with the original method on which all of the original studies were performed. After all, if a new commercial kit is to be used in place of the original, it should correlate with the clinically proven assay and give essentially the same results.
Thus, performing a correlation study by testing aliquots of fresh sample of the new commercial method against the original “gold standard,” a regression line slope of 1.0 combined with a y-axis intercept close to zero versus the original method is at least statistically indicative that the methods are essentially identical. Of course, additional studies proving accuracy and precision of the new method are required before approval and adoption of the method for clinical use.
The FDA already requires such correlation studies for new methods prior to approval of a new commercial test kit. Presently, some degree of bias between commercial tests is allowed by the FDA; however, going forward, evidence-based protocols and harmonization requirements for specific cut-offs for immunoassays may require that new methods compare both qualitatively and quantitatively to the original.
As a step toward success, some vendors have achieved standardizing their own operations to create a factory calibration between reagent lots. They have very tight controls on new lots of reagents and can standardize the production to supply a factory calibration on a -two-dimensional bar code. Factory-based instruments must be performing analysis in close agreement with those in the field. The system runs calibrators included in the kit called adjusters, reads the factory calibration parameters from the bar code and is able to make the lab’s calibration curve match the one at the factory. So, within the manufacturer, all of the variables surrounding their antigen, antibody and buffers can be controlled enough to consistently make their instruments work like those in remote client labs.
This is an impressive improvement, but other companies aren’t able to reach that point of standardizing their instruments-and that serves as a poignant illustration of how complicated the process really is. This is a red flag for the industry as we hope for achieving inter-comparability between vendors for the same analytes. If a factory calibration is such a challenge, coming up with a universal calibrator seems even more out of reach. However, this within-vendor factory calibration presents the industry with an interesting option for making methods compare more favorably between vendors. The idea of “universal adjusters” supplied with each new lot of immunoassay reagent comes to mind.
Other Harmonization Efforts
Organizations such as the AACC are sponsoring forums with a focus on test harmonization as a means of improving test performance. This global effort faces challenges because of the small number of analytes with well-established standards and suitable reference methods for measurement. Future global consensus standardization efforts and agreements must focus on those analytes with no reference methods available and no suitable reference preparation available. Harmonization of analytes that play an essential role in evidence-based protocols should be a priority. Effective protocols for achieving global standardization efforts must be established, as well as a rigorous outcomes assessment program for evaluating the overal adoption and success.
An idea I have developed is a computerized method of reducing immunoassays and virtually all other medical and laboratory results to a single, dimensionless universal scale. Regardless of the method employed for analysis, harmonized/normalized results are mapped together and presented numerically and graphically with ease so that physicians can access them even on wireless devices, smart phones, etc. This system of reporting integrates and simplifies things for the physician and allows patients to be more involved in their healthcare. Patients, and even some physicians, have limited interest and knowledge about the actual value of units and numbers being reported. Patients simply want to know if their result is normal and how they compare to others of their same age and sex, and how their results are trending.
Such normalized and harmonized data could more conveniently be integrated into evidence-based cut-offs, regardless of the analyte. When viewing the issues relative to immunoassays and harmonization, a solution involving data reduction may, in the long run, be more feasible than analytic harmonization.
Dr. Blick is professor, Department of Pathology, University of -Oklahoma Health Sciences Center and OU Medical Center.