Sample non-linearity can be defined as a sample that when measured at different dilutions gives a substantially different result. Two examples are shown in Table 7.4. Sample A is an example of a non-linear sample, while sample B shows a linear response (i.e. further dilution [1/100] gives a value close to that obtained at the initial dilution [1/10]).
Table 7.4. Examples of non-linear and linear sample behaviour.
Sample non-linearity should not be confused with assay linearity; if an assay is non-linear all samples will give substantially different results when measured at different dilutions. Assessment of Freelite assay linearity forms an important part of immunoassay development and validation (Section 5.6).
As with other immunoassays, FLC assays are potentially prone to sample non-linearity. Pretorius et al.  concluded that non-linearity of FLC assays was a property of the individual sample, and not a method-specific phenomenon. Sample specific non-linearity may occur in some samples when measured with Freelite FLC assays due to: 1) non-specific interference (matrix effects); 2) the inherent structural diversity of monoclonal FLCs (Chapter 3); 3) FLC polymerisation (Section 7.6); or 4) other unknown factors.
Hitchcock et al.  investigated the incidence of sample non-linearity in 1452 serum samples sent for routine analysis. Non-linearity was defined as samples that gave a > value at a 1/10 sample dilution, but a result within the 1/10 measuring range when assayed at a 1/100 dilution. Overall, 5% samples exhibited non-linear behaviour, although the authors comment that only 2/1452 (0.14%) had noticeable non-linearity (defined as a 30% difference between the two dilutions) .
7.4.1. Managing non-linearity
Binding Site recommend that customers follow the sample dilution protocol shown in each product insert and report the first plausible (non-flagged) result returned by the analyser. Laboratories should also follow their own in-house protocols to ensure that non-linear samples are managed consistently. The use of non-standard sample dilutions, or skipping dilutions should be avoided. Examples of how to report non-linear samples are shown in Tables 7.5 and 7.6.
Table 7.5. Examples of non-linear κ Freelite results using a Siemens BN™II analyser. In this example, the first result was reported (values in bold).
Table 7.6. Example of non-linear λ Freelite results using a Binding Site SPAPLUS analyser. In this example, as the analyser has generated flags against the results at 1/10 and 1/100 dilutions the first valid result was reported (value in bold).
It is important to note that Freelite results should not be interpreted in isolation and other laboratory findings and clinical symptoms should be considered when evaluating the status of the patient. In addition, IMWG guidelines state that documentation of response requires two consecutive sFLC assessments made at any time, but to confirm response or progressive disease, two discrete samples are required .