Lateral-flow assay reproducibility is not only influenced by design and manufacturing, but also by the components used in the assembly of the test, the most critical of which is the reaction membrane.
By Dr Klaus Hochleitner and Brendan O’Farrell
Diagnostic testing has become an essential aspect in a wide variety of settings. Testing assists in the identification of risk factors such as cholesterol levels for coronary disease risk. Establishing the presence of disease or condition (for example, respiratory infection, hepatitis, and myocardial infarction) would be difficult if not impossible without the aid of a diagnostic test. Diagnostic testing also can be used to differentiate between acute and previous infection. For example, differential analysis of an immunological response can distinguish an active infection from a latent one.
The results of a diagnostic test have a large influence on clinical management. Differential diagnosis such as viral versus bacterial infection can reduce the incidence of overprescribing antibiotics. Together with clinical symptoms, the detection of biomarkers such as troponin I or C-reactive protein (CRP) indicates the need for prompt medical intervention.
Disease management activities can also be influenced by diagnostic test results. Monitoring of cholesterol levels allows for assessing the efficacy of cholesteremia modifying therapeutics, diet, and exercise. Assessment of thrombolytic agents plays an important role in the management of coronary artery disease (CAD).
Diagnostic tests are not limited to human healthcare. Diagnostic tests are also utilized in other areas where analysis is important, ranging from agriculture to environmental testing to food safety to veterinary applications. Irrespective of the area of use and application, diagnostic tests should demonstrate cost effectiveness and clinical utility. A test that fails to demonstrate these aspects is apt to be replaced by one that does. However simple and cost effective, a test must also demonstrate adequate reproducibility to distinguish clinically relevant results from assay variability. There is little value and utility in a test if the lack of reproducibility casts doubt on the results.
|Figure 1 (click to enlarge). High reproducibility (green) versus low reproducibility (red). Negative sample (triangle) and positive sample (circle). High reproducibility tests (1 and 2) show clear differentiation between negative and positive results, whereas tests 3 and 4 (low reproducibility) show less differentiation between negative and positive results.|
Consider the scenario where a single, positive sample and a negative sample containing no analyte are analyzed multiple times in the same assay. An assay with high reproducibility will report values in a smaller distribution than will an assay with low reproducibility. Figure 1 shows the results when testing a positive and negative sample in an assay with high and low reproducibility.
Referring to Figure 1, assume the assay has a defined cut-off limit of 7.5 units. The assay with high reproducibility allows for clear differentiation between positive and negative results. This contrasts with results obtained in the assay with low reproducibility. Because of the large variation, there is reduced confidence distinguishing clinically relevant results (positive sample) from the assay variability (negative sample).
High assay reproducibility not only improves the distinction between clinically relevant results and assay variability, but also has an effect on an assay detection limit. A common practice is to establish the limit of detection at a point that is three standard deviations above normal assay variability. Figure 2 shows the effect of assay reproducibility on assay detection limit.
There are many factors that can influence the performance of a test. Temperature differences can alter the assay kinetics; procedural errors, such as those associated with timing and dispensing, contribute to assay variability. Other factors including assay components, manufacturing processes, and reagent variation also contribute to assay precision. It is important to note that assay reproducibility is the cumulative effect of individual sources of variability.
|Figure 2 (click to enlarge). Effect of reproducibility on assay detection limit. The high reproducibility assay (left) has a lower detection limit than a low reproducibility assay (right).|
Technological advances have enabled manufacturers to develop assays with improved reproducibility and simplicity since the early days of diagnostic testing. Tests that once required hours or days to complete can now generate results in minutes. Today, automated assay platforms have replaced repetitive manual operations and eliminated many causes of procedural errors. Engineering advances, combined with components developed specifically for diagnostic applications, have not only contributed to increased assay performance, but have also enabled tests to be performed in a variety of settings ranging from central laboratory environments to in-home testing to remote field locations.
Under ideal situations, a developer would assess test reproducibility under the actual conditions of final use. However, in many instances, assessing test reproducibility under conditions of actual use is neither practical nor possible. This is especially true for tests intended for in-home and field use.
Instead, manufacturers rely on a combination of robust assay design, manufacturing methods, and multiple in-process and final quality control procedures to assess and ensure reproducibility.
Assays intended for decentralized analysis, such as in-home and remote test settings, have an additional simplicity requirement as they are often performed by individuals with little or no technical training. One test platform especially suited for use by nonprofessionals is the lateral-flow assay (LFA).
LFA reproducibility is not only influenced by design and manufacturing, but also by the components used in the assembly of the test. The most critical component in these types of assays is the reaction membrane, which is typically a nitrocellulose membrane.
|Figure 3 (click to enlarge). Average peak heights of FF120HP (roll 7A/Q).|
With the need for more-sensitive assays and the reporting of quantitative results, the consistency of the LFA reaction membrane has a significant impact on the performance of the final test.
Nitrocellulose membranes are manufactured by dissolving the raw materials in a mixture of organic solvents and water, pouring this casting mix onto a solid belt-like support, and evaporating the solvents under controlled conditions of temperature, humidity, and belt speed within the membrane manufacturing machine.
As the organic solvents evaporate, the water concentration in the casting mix increases to the point that the nitrocellulose fibers precipitate and form a random mesh- or sponge-like structure. The consistency and reproducibility of this structure is key to the reproducibility of the LFA devices manufactured using this type of membrane.
Using new membrane formulations and improved manufacturing conditions, GE Healthcare Life Sciences has developed a new family of highly consistent membranes. One of these membranes is Whatman FF120HP.
Manufacture of the test components, assembly of tests, performance studies, including all data and conclusions, were provided by DCN Diagnostics (Carlsbad, CA).
A colloidal gold myoglobin assay was previously developed at DCN Diagnostics for internal use and demonstration purposes. The assay, utilizing the same conjugate pads, antibodies, colloidal gold conjugate, and processes, was transferred to the FF120HP membrane.
|Figure 4 (click to enlarge). Intra-lot variability of FF120HP based on testing of three rolls from the same casting lot.|
All tests were performed in DCN Diagnostics’ laboratories using standards with various concentrations of myoglobin prepared in myoglobin-free serum. Tests were done by pipetting 100 µl of the myoglobin standard into a test tube, then adding a test strip and waiting for 15 minutes. After 15 minutes, the strip was removed and peak heights were determined using an ESEQuant Lateral Flow Immunoassay Reader (Qiagen).
Results and Discussion
The best line appearance on the FF120HP membrane was obtained at a dispensing rate of 0.6 µl/cm. The antibody mass was 0.099 µg per test strip.
Reproducibility of the FF120HP membrane was tested using the conditions described above. Concentrations of myoglobin in myoglobin-free serum were 0, 0.25, 2.5 and 10 ng/ml. Twenty strips were prepared for each concentration of myoglobin. Test strips were read using a reflectance reader (Table I), and peak heights were obtained. The % CV was determined by dividing the standard deviation (SD) by the mean times 100%.
|Table I. ESEQuant reader results of colloidal gold myoglobin assay using a constant amount of anti-human myoglobin antibody per test.|
The results of this limited testing of a single roll of membrane showed that the peak heights (Figure 3) had low variability.
Intra-lot Variability of the Whatman FF120HP Membrane
The intra-lot variability of the FF120HP membrane was evaluated by striping the membranes from three different rolls (7A/Q, 7A/R, 7A/S) from a single casting lot (G1471608). A test line of 0.2 mg/ml of anti-human myoglobin antibody was dispensed at 0.6 µl/cm and 50 mm/sec. Twenty tests were performed for each roll using 0, 0.25, 2.5, and 10 ng/ml of myoglobin in myoglobin-free serum.
The results show that the three rolls within the same casting lot produced similar peak heights when striped with the same antibody solutions (Figure 4).
Inter-lot Variability of Whatman FF120HP Membrane
The inter-lot variability of the FF120HP membrane was evaluated by striping membranes from three different casting lots (D013024, roll 15A/T; G1471608, roll 7A/Q; G1471610, roll1A/T) with a test line of 0.2 mg/ml of anti-human myoglobin antibody dispensed at 0.6 µl/cm and 50 mm/s. Twenty tests were performed for each roll using 0, 0.25, 2.5 and 10 ng/ml of myoglobin in myoglobin-free serum.
The results showed that the three rolls of FF120HP membrane produced very similar peak heights when striped with the same antibody solution (Figure 5).
|Figure 5 (click to enlarge). Inter-lot variability of FF120HP based on the testing of three rolls from three different casting lots.|
The FF120HP membrane provides a very high level of consistency. Additionally, the FF120HP membrane demonstrated remarkable intra- and inter-lot reproducibility with very low CV values that provide assay developers better control of manufacturing processes, reduced QC burden, and lower scrap rates in assay manufacturing.
Dr Klaus Hochleitner is global lead technical product specialist,
diagnostics, at GE Healthcare Life Sciences. He can be reached via e-mail at email@example.com.
Brendan O’Farrell is president and founding partner at DCN
Diagnostics. He can be reached via e-mail at firstname.lastname@example.org.
Also contributed to this article:
Vincent Sy is global lead applications specialist, diagnostics,
at GE Healthcare Life Sciences.
Wei Sun is research scientist, polymer and analytical chemistry, at GE Healthcare Life Sciences.