Edition 1, 2021
IS INFECTIOUS DISEASE SEROLOGY TESTING DIFFERENT TO OTHER MEDICAL MEASUREMENT SYSTEMS?
In response to an article published by us in 2019, a correspondent suggested that infectious disease serology is “just another assay which follows the same laws as any other”. This is an important concept to debate, as it impacts on the way we understand and manage serology testing; in particular, how we approach standardization and control of infectious disease serology. Can we assume that the laws accepted as truth and applied to clinical chemistry can be universally applied? Do they apply to infectious disease serology? First, we can review the differences between testing for an inert chemical such as glucose and testing for antibodies.
Glucose is generally a homogeneous molecule of a known structure (C6H12O) and molecular weight of 180.16 g/mol, and can be found in a pure state. Accurate measurement of glucose using a certified reference method such as mass spectrometry can be utilized to create a higher-order certified reference material (CRM). This CRM can be used to calibrate all measurement systems to standardize reporting. If a measurement system demonstrates bias, it can be recalibrated, and therefore adjust for reagent lot changes and other sources of variation. When testing a population for glucose, the results form a normal or Gaussian distribution, where the majority of the population have results within a normal range, but some individuals will be hypo- and other hyper-glycemic. That is, there are multiple clinical decision points.
When looking at antibodies to infectious diseases, irrespective of the organism, there are many differences to clinical chemistry testing. Antibody responses are a polyclonal in nature and assays may have multiple isotypes of IgG, IgM and/or IgA. Early in infection, the immune response is immature, where the antibodies have low avidity and affinity. IgM is usually first to be detected and gradually decreases over time. Antibodies to some antigens are detected earlier than others and are present for variable lengths of time. Over time, the response matures, becoming more specific and therefore having stronger capability to bind to antigens. Each individual produces a different immune response, although each infected person will be detected as antibody positive. Also, different genotypes or antigenic mutations can elicit variable antibody responses. So, the immune response to an infection is variable over time and is never homogeneous. When we review a population for antibodies to infection, we do not have a Gaussian distribution, we have two populations – a negative and a positive population. In well-designed assays, these two populations are well differentiated. There is a single decision point; that is the presence or absence of antibodies.
Assay design for infectious disease serology varies from manufacturer to manufacturer. Each assay uses different source of antigen to detect antibodies, have different chemistries to detect signals and, in some cases, detect multiple antibody isotypes. Most serology assays are highly regulated, which means that the user is not able to adjust for bias, especially the bias introduced by new reagent lot numbers. It is important to note that there are no certified reference methods for infectious disease serology, and therefore any standards that are created are not traceable to SI units. These analytes are referred to as “Type B” quantities and measure functional or biological activity.
When we test for glucose we are measuring “how much” glucose is in the sample. When we measure antibodies, we are measuring “how well” the antibodies bind to the solid phase and how well our detection system can produce a signal. We do not measure the numbers of antibodies to a particular antigen. We cannot assume that the numbers of antibodies are proportional to the signal.
In summary, there are major differences between testing for clinical chemistry analytes and antibody testing. Why is this important? It is because of statements like serology testing is “just another assay which follows the same laws as any other”. Plainly this is not the case. We cannot apply the same methods to standardize and control clinical chemistry to infectious disease serology. When we do, we encounter issues. NRL has been at the forefront of demonstrating that the application of Westgard rules for monitoring QC results of infectious disease testing is not appropriate (Clin Chem Lab Med 2018; 56(11): 1970–1978). We have also demonstrated the lack of standardization when serology assays have results reported in International Units per Milliliter, as experienced in anti-rubella IgG testing. We championed a change to the Intended Use of the WHO International Standard anti-Rubella Immunoglobulin standard (NIBSC RUB-1-94), resulting in the statement “IVD manufacturers, regulators and assay users should be made aware of RUBI-1-94 potential lack of commutability when used as a calibrant. This was highlighted by the WHO Expert Committee on Biological Standardization in TRS 68th Report (Section 3.3.4: 2018).” This change was in response to the fact that, even after almost 20 years of calibration of anti-rubella IgG assays using this standard, there continues to be a lack of standardization of quantitative results.
It is of great interest to note that a WHO International Standard First WHO International Standard for anti-SARS-CoV-2 immunoglobulin (human) NIBSC code: 20/136 has been released, with a stated intended use of “the International Standard is for the calibration and harmonisation of serological assays detecting anti-SARS-CoV-2 neutralising antibodies. The preparation can also be used as an internal reference reagent for the harmonisation of binding antibody assays”. It is not clear exactly if the neutralizing antibody assays referred to are viral plaque neutralizing assays used in vaccine development and potency testing, or it included immunoassays. However, manufacturers of immunoassays are releasing “quantitative” anti-SARS-CoV-2 assays.
Are we to experience the same issues faced with anti-rubella IgG? I predict so.
Executive Manager, Scientific and Business Services