Analytical sensitivity limits for lateral flow immunoassays.

A plethora of literature describes the use of lateral flow immunoassays (LFIA). Only a fraction of that contains information about possible analytical sensitivity or limit of detection. Accordingly, we have undertaken a comprehensive survey of the literature in PubMed and present here that fraction that has information on limits of detection or analytical sensitivity. Further detail of the search strategy is at http://www.finddiagnostics.org/ news/resources/gordon_michael_ latflow_immunoassays.pdf. In general, there is a need for rapid tests, such as LFIA, capable of being performed by unskilled operators yet providing rapid and reliable results in areas such as diagnosis of diseases in developing countries (1, 2 ), emergency room use (3 ), and biodefense (4 ). The utility of such tests is enhanced by the ability to determine analytes at very low concentrations. In LFIA, sample is added, the analyte and label are subjected to chromatography-like migration through a membrane, and a result is read at the site of an immobilized capture reagent. The term LFIA has not been universally used and so does not capture all of the relevant publications in a key word search. A formalized structure (http://www.finddiagnostics.org/ news/resources/gordon_michael_ latflow_immunoassays.pdf) was used as a device to create a terminology-independent tool for a search of publications in PubMed in July 2007. The number of publications using LFIA was 287. The most common “standard” technology type used a combination of colloidal gold label dried in a conjugate pad and nitrocellulose membrane lateral flow medium. All the publications were categorized by having either standard technology or the following significant variations: alternative label, no conjugate pad (“mix and run”), or alternative membrane (http://www.finddiagnostics.org/ news/resources/gordon_michael_ latflow_immunoassays.pdf). Of the 287 publications, the 17 containing information on limits of detection are summarized in Supplemental Table 1 that accompanies the online version of this letter at http://www.clinchem.org/content/ vol54/issue7. Supplemental Table 1 alsodescribes thenatureof theanalyte, thematrixinwhichtheanalytewasapplied, the technology, the method of determining detection limit, and molarity at the detection limit. We converted the units used in each publication to molarity to facilitate comparisons across different technologies and analytes. The lowest detection limit using the standard format, 3 10 11 mol/L, was that of FernandezSanchez et al. (see ref. 7 in Supplemental Table 1). Improvement appears to result from mix-and-run format, although direct comparison is not possible since the only reports listed used an additional alternative label. Possible reasons for using this format might be (a) to avoid the difficulty of stabilizing dried reagents, (b) to optimize sensitivity by prereacting first components in a liquid phase, or (c) to provide the option of combining with a microplate format. In that case, dried reagents in the wells can be preincubated for a specified time. Higher throughput might then be achieved by running multiple patient samples in parallel. Alternative labels described were colloidal carbon, europium III and time-resolved fluorescence, upconverting phosphors, and dyeloaded liposomes. The lowest limit of detection of all was 10 mol/L for the detection of cholera toxin with mix-and-run and dye-loaded liposomes. This is the record for this type of technology. A more direct comparison with other methods is possible for europium III and upconverting phosphors. Limits of detection reported for microparticle-based immunoassay systems by Ukonaho et al. (5 ) were approximately 100-fold lower than for LFIA in Supplemental Table 1. It is therefore possible that these 2 labels have potential for lower limits of detection, but the fact that special optics or illumination are required might be incompatible with their use in a resource-limited environment. There appeared to be no advantage in the use of alternative membranes, the only example in Supplemental Table 1 being nylon.