Abstract
Many tests in clinical laboratories use fully automated ‘closed systems’ but others use manual techniques or generic ‘open systems’ such as enzyme immunoassay (EIA) plate processors. These latter techniques rely on accurate manufacturers' instruction sheets. Errors in these have implications on patients and unambiguous instructions are important.
Kit instructions often use the terms 1:x and 1/x interchangeably in describing dilution steps despite these having distinct meanings. Instruction to dilute a sample 1:10 indicates addition of one part of sample to 10 parts of diluent, whereas 1/10 indicates addition of one part to nine parts of diluent. Discussion with a small number of scientists in our laboratory demonstrated that some considered the terms interchangeable.
Current kit inserts from tests within our immunology department were reviewed to determine the extent of this problem in an area that uses many manual assays. Selected commercial kits that required a manual dilution step or set-up on an ‘open system’ EIA processor were examined. The instruction and intended meaning were determined by careful reading of all parts of the kit insert.
Seventeen kits from seven manufacturers were examined: The Binding Site, Birmingham, UK (anti-gliadin EIA); Technochrome, Vienna, Austria (functional C1 inhibitor kinetic assay); Euroimmun, Lübeck, Germany (anti-intrinsic factor EIA, MPO-ANCA EIA, PR3-ANCA EIA, anti-liver antibody line immunoblot, anti-extractable nuclear antigen [ENA] line immunoblot); SeraQuest, Miami, FL, USA (ENA Plus screen EIA); Aesku Diagnostics, Wendelsheim, Germany (IgA anti-tissue transglutaminase [tTG] EIA). Eight assays from two fully automated systems that provided instructions for off-line dilution of reagents or samples with high results were included: Architect, Abbott Diagnostics, Maidenhead, UK (total prostate-specific antigen, folate, prolactin, vitamin B12, total β-human chorionic gonadotropin [hCG]); DELFIA Xpress, Perkin-Elmer, Turku, Finland (alpha-fetoprotein, unconjugated estriol, free-β-hCG).
Of the kits examined only two used the ratio term correctly (The Binding Site anti-gliadin and SeraQuest anti-ENA screen). The other 15 used the ratio as though it was a fraction to describe the dilution of samples or reagents. In all cases the correct meaning could be discerned by careful reading of the kit insert, e.g. Aesku anti-tTG instructs ‘Dilute the concentrated sample buffer 1:5 with distilled water (e.g. 20 mL plus 80 mL)’. The clear intention here is a 1/5 dilution, not a 1:5 (1/6). All errors were of this nature, i.e. use of a ratio as equivalent to the fraction. The nature of the error often implied that this was a clear misunderstanding and not an oversight by the manufacturer, e.g. describing a 10 μL plus 1000 μL dilution as 1:101 when it should be 1:100.
When standards, controls and samples are treated identically, even a 1:10 compared with a 1/10 dilution is unlikely to have a significant effect on the results. However, some kits provide prediluted standards and controls but samples require dilution.
It may seem pedantic to insist upon the correct use of these mathematical terms, especially as in the examples seen a careful reading of the kit insert prevented an error. However, the potential for error exists and use of fractions and ratios as equivalent fosters misunderstanding about standard mathematical nomenclature that should be unambiguous and may have consequences elsewhere when correct use of these terms may be critical.