The Critical Components of Successful Assay Design

By April 17, 2018 November 14th, 2018 Articles
rapid contaminant identification, lateral flow, lateral flow test, lateral flow assay

Written by: George Parsons, Ph.D.

A rapid diagnostic test is a tool that, like any other tool, must be designed and built for its intended purpose. Before going into the lab to develop an assay, one must start at the end and correctly identify the goals and use of the product. Only then can point of care diagnostic products be designed correctly. This is the critical first step in successful assay design.

For successful assay design, the following items should be clearly articulated prior to development:

  • A written description of the assay’s intended use
  • What the analyte will measure
  • What sample—that may or may not contain the analyte—will be collected
  • How much sample is required
  • How much experience the operator must have and what support resources, such as instrumentation for quantitation, are available

The availability of critical raw materials, including antibodies and calibration materials, is another important consideration of successful assay design. Since developed rapid diagnostic tests may have useful lives measured in decades, reliable sources of these critical raw materials must be assured. In addition, development resources, such as experienced lateral flow assay developers and adequate laboratory facilities are critical. Finally, a manufacturing infrastructure with the requisite personnel, documentation and management to reproducibly produce this assay for years to come must also be available or built. All of these requirements are assembled into a multi-page specification document that defines the desired parameters of the diagnostic product. Contact us for more information about filling out a product requirements document for your diagnostic test.

Diagnostic assays require resource teams with different perspectives for successful development and transfer to manufacturing.  Technical resources can address questions, such assay feasibility and performance characteristics.  Marketing represents the end user and provides inputs to the team as to what the end user needs the assay to do.  Manufacturing suggests how best to make the assay in a consistent, reliable and cost-efficient manner.  Quality defines how best to test the assay to ensure it is meeting specifications.

During the development process, R&D will challenge the rapid diagnostic test with various conditions known to interfere with similar assays.  One classic parameter that can affect urinalysis assays is pH.  In humans, urine samples can have pH values ranging from 4.5 to 8.[1]  Antibodies work best around a pH of 7.4 and can usually only tolerate deviations of plus or minus one pH unit.  A sample with a pH of 5 may give erroneous or misleading results in an assay not designed to handle these extremes.  Temperature also affects antibody reactions.  An assay that works well at room temperature or body temperature may not give the same results in colder temperatures.  Other substances in the sample can also interfere with the assays.  Assays that use blood or blood fractions, such as serum or plasma, must be tested to ensure that they will perform adequately in the presence of hemoglobin released from disrupted red blood cells, or lipids from dietary sources, or bilirubin from liver abnormalities.  Exogeneous substances, such as biotin and fluorescein, can also interfere with some assays.[2]

Lateral flow assays can be hardened to address some, but not all of these issues.  For example, the pH parameter can be dealt with by ensuring that sufficient buffering capacity is present either on the device or in the solutions intended for use with the device.  Written warnings in the instructions for use can help address temperature fluctuation and naturally occurring substances in blood.  Intelligent assay design can help manage biotin interference.

One critical aspect of assay design is the provision of adequate controls.  Controls are often assay specimens that have been engineered to mimic actual samples in composition and concentration of the analyte under investigation while maintaining stability and ease of use.  In single use assays, including many lateral flow assays, this type of control is clearly impractical and other measures must be taken to ensure the assay is run appropriately.  Over the counter pregnancy tests routinely include a control line that ONLY appears if there has been an adequate amount of sample added to the device and that it is not of a pH that would interfere with an immunoassay.  This is often accomplished by adding a second line of reagent in the reaction area of the device.  The first line is the capture antibody for the analyte of interest in this case human chorionic gonadotropin (hCG).  The second line is usually an anti-species IgG antibody that can capture the labeled detector antibody.  For example, most pregnancy assays use mouse monoclonal antibodies for both capture and detector antibodies.  A second line that contains goat anti-mouse antibody will capture any excess detector antibody.  What the user sees is that a single line on the device indicates that the test has been run properly with enough sample to carry the reagents to their intended capture lines, but the analyte is not present in detectable amounts.  If the analyte were present, then the user would see two lines, one from the capture antibody line and one from the anti-mouse line.  If no lines are visible the test is invalid and no conclusion can be drawn.  An inadequate amount of sample is frequently the cause of a failed assay.

A recently published paper provides a real-world example of how an assay was developed without consideration of how it was to be used.[3]  An alert and well-trained operator noticed that a certain urine specimen failed to give either a result or control line on a Point of Care (POC) pregnancy test from one manufacturer.  The urine specimen was tested on six other commercially available POC pregnancy tests and they all gave appropriate test and control results.  Usually when a POC pregnancy test is run the colors of the test and control lines are the same—but not in this case.  The test line was red and the control line was blue.  The authors speculated that the usual anti-IgG capture line had been replaced by an immobilized streptavidin line and the capture antibody reagent had been supplemented by that addition of a biotin labeled blue dye.  The suspect sample was thought to contain excess biotin ingested by the patient from an over the counter preparation that are supposed to promote good growth of hair and nails.  Excess biotin in the sample overwhelmed the streptavidin in the control line, and it could not bind the biotin labeled blue dye.  The other POC pregnancy tests did not use the biotin-streptavidin chemistry and were not affected by the presence of biotin.

The above represents a failure in assay design, but it was a soft failure.  No patients were harmed in the incident.  It did, however, cause financial harm in that it triggered an investigation requiring people and resources to unravel.

What did we learn from the failure?  Obviously, that incorporating a biotin-avidin linkage in a test that uses human urine as a sample is inherently a risky decision.  Six other tests were unaffected by the excess biotin which is good news because it points a way forward to a more robust design. The biotin-avidin linkage was probably introduced into the failed test to provide a different color for the control strip to better differentiate it from the test strip.  Another way to accomplish this would be use a different species antibody for the control strip chemistry.  Let’s assume the test strip is a mouse antibody as is the detector labeled antibody.  If the control material were a blue dyed rabbit antibody and the control strip were a goat anti-rabbit antibody, that should provide a robust chemistry for the control strip while preserving the different colors for test and control results.

This incident and the excellent scientific detective work that went into determining root causes provide a good lesson in what can happen if good assays design is not implemented and rigorously tested.

If you are not sure whether your test is sufficient for its intended purpose, contact us.  Our team of engineers, scientists and marketers will help you identify the goals, use and consumers for your test.

[1] Moeller, KE; Kissack, JC ; Atayee, RS and Lee, KC, ”Clinical Interpretation of Urine Drug Tests: What Clinicians Need to Know About Urine Drug Screens” Mayo Clin Proc., 92(5):774-796 (2017)

[2] Li D; Radulescu A, Shrestha R, Root M, Karger A, Killeen A, Hodges J et al., ”Association of Biotin Ingestion With Performance of Hormone and Nonhormone Assays in Healthy Adults” JAMA 318(12):1150-1160 (2017)

[3] Williams G, Cervinski M , Nerenza, R “Assessment of biotin interference with qualitative point-of-care hCG test devices” Clin Biochem (2018).