Inter-laboratory Comparisons

An alternative method for performance evaluation is to participate in inter-laboratory comparisons, often known as round-robins . Usually this involves sending sub-samples of a selection of appropriate samples to a number of independent laboratories, to be analysed either by a fully specified procedure and technique, or to be analysed by whatever method each laboratory choses. Obviously, these two approaches test different things. The former indicates the precision attainable using a specified procedure, and tests both the adequacy of the specification and the competence of participating laboratories. Only the 

It should be stressed that inter-laboratory comparisons are not a substitute for regular, in house quality control procedures. Their main value is in novel procedure evaluations or, especially in the present context, when several different laboratories are collaborating in an integrated environmental study. 

---

Reference methods are generally arrived at by consensus and fairly extensive testing by a number of laboratories. For example, the flame atomic absorption method for Ca in serum developed under the leadership of the agency fondly remembered as NBS, now NIST (Cali et al. 1972), was established after several inter-laboratory comparison exercises. The results were evaluated after each exercise and the procedure was changed as necessary. After five exercises, it was felt that the state-of-the-art had been reached, with the reference method being capable of measuring Ca in serum with an accuracy of 2% of the true value determined by IDMS (note that attainment of high accuracy and precision is not only a matter of the method, but is a function of both the method and analyst expertise). 

NIST has also used results obtained from inter-laboratory studies as an additional set of results in the two or more methods approach (mode 2 in Table 3.13). For example for the recent value assignment for PCBs and pesticides in SRM 1944, the mean of results from 19 laboratories participating in an inter-laboratory comparison exercise was used as an additional set of data in the determination of the certified values. Similar inter-laboratory study results were also included in the value assignment of PAHs, PCBs, and pesticides for two recently issued mussel tissue materials, SRM 2977 and SRM 2978. 

A number of scientific journals regularly publish papers reviewing the state of the art in the RM business as well as original contributions on certification, inter-laboratory comparisons, and on RM/CRM applications. Table 8.2 lists the most popular and widely cited. The most prolific journals are Fresenius Journal, JAAS, Science of the Total Environment and Water, Air, Soil Pollution, all with around or even more than 50 papers mentioning reference materials from 1998 and 1999. Of these, Fresenius Journal led in 1998, with more than 80 papers. This was partly because it traditionally publishes, so far in six special issues, a series of papers presented at the International Biological and Environmental Reference Materials Symposia (BERM) series. The role and contribution of the BERM series of meetings is reviewed below. 

Dearman, R.J., Skinner, R.A., Herouet, C., Labay, K., Debruyne, E. and Kimber, I., Induction of IgE antibody responses by protein allergens Inter-laboratory comparisons. Food Chem. Toxic., 41, 1509, 2003. 

Recent results demonstrated the importance of an inter-laboratory comparison in order to ensure a correct method transfer. Earlier inter-company collaborations were performed for small molecules and biomolecules. ... 

The laboratory is also involved in quahty assessment, such as participation in an external quahty- control scheme (inter-laboratory comparison). Here, a variety of analytes in matrices such as hard water, soft water, and sludge are circulated to laboratories and the returned results are processed in a codified format so that each participating laboratory can only identify its own values. 

A further advancement comes from inter-laboratory comparison of two standards having different isotopic composition that can be used for a normalization procedure correcting for all proportional errors due to mass spechomehy and to sample preparation. Ideally, the two standard samples should have isotope raUos as different as possible, but still within the range of natural variations. There are, however, some problems connected with data normalization, which are still under debate. For example, the CO2 equilibration of waters and the acid extraction of CO2 from carbonates are indirect analytical procedures, involving temperature-dependent fractionation factors (whose values are not beyond experimental uncertainties) with respect to the original samples and which might be re-evaluated on the normalized scale. 

Stevens, J.L., Green, N.J.L., Bowater, R.J., Jones, K.C. (2001). Inter-laboratory comparison exercise for the analysis of PCDD/Fs in samples of digested sewage slrrdge. Chemosphere 45 1139-1150. 

Stronkhorst J., Ciarelli, S., Schipper, C.A., Postma, J.R, Dubbeldam, M., Vangheluwe, M., Brils, J.M. and Hooftman, R. (2004). Inter-laboratory comparison of five marine bioassays for evaluating the toxicity of dredged material. Aquatic Ecosystem Health Management, 7, 147-159. 

Valkirs, A.O., Seligman, PF., Olson, G.J., Brinckman, F.E., Matthias, C.L. and Bellama, J.M. (1987) Di- and tributyltin species in marine waters. Inter-laboratory comparison of two ultratrace analytical methods employing hydride generation and atomic absorption or flame photometric detection. Analyst, 112, 17-21. 

Whereas type A and B uncertainty components [6] are almost always added by root-mean-square additions, there are circumstances under which direct additions of uncertainties from dependent error sources are preferable. When protocols use such direct additions of uncertainties, these conditions should be clearly explained in the protocol. That information is needed for a fair inter-laboratory comparison of values and uncertainties of measurements on the same material. 

In their regular day to day practice, field laboratories use commercial reagents or prepare in-house solutions for the calibration of instruments, and they rely on purity assessment of producers. For method validation and even measurement uncertainty, field labs regularly participate in proficiency testing schemes. In such inter-laboratory comparisons, the reference value is usually obtained as the arithmetic mean of results of participants. 

Forty six field laboratories (mainly from France) have participated in the inter-laboratory comparison. They have analysed the water sample using their regular calibration solution and then have repeated this analysis using the certified standards. They also have analysed the matrix CRM. Laboratories have produced two results (duplicate) per sample. Techniques used were mainly atomic absorption spectroscopy with furnace but also ICP-OES and ICP-MS. 

Analysis of a matrix CRM has shown a noteworthy bias in lead analysis of about 10%. This bias was observed for all routine methods used by laboratories. Complementary experiments have indicated that a strong matrix effect was responsible for this bias. This underlined bias has demonstrated that for this matrix water, only a primary method can provide the true value of the sample of an inter-laboratory comparison. The mean value of laboratory results cannot be used as a reference value. 

M. Ratio, F. International Inter-Laboratory Comparison (Round-Robin) Test for the Verification of Chemical Disarmament. F.l Testing of Existing Procedures, 1990 F.2. Testing of Procedures on Simulated Industry Samples, 1991 F.3. Testing of Procedures on Simulated Military Facility... 

M. Rautio (Ed.), H. Inter-laboratory Comparison Test Coordinated by the Provisional Technical Secretariat for the Preparatory Commission for the Organization for the Prohibition of Chemical Weapons. H.l First Inter-laboratory Comparison Test. Methodology and Instrumentation for Sampling and Analysis in the Verification of Chemical Disarmament, The Ministry for Foreign Affairs of Finland, Helsinki, 1994. 

On the other hand, PAMPA is a purely artificial method and PAMPA membranes do not reassemble real lipid bilayer structures as barriers for permeation but much thicker barriers. The thickness and material of the supporting PVDF filters also influences artificially the permeation of compounds depending on the lipophilicity of the compounds more than the thin polycarbonate filter does in CACo2 experiments. Also the best choice of membrane constituents for PAMPA experiments is still under investigation and it seems that it will depend a lot on the goal of the PAMPA experiment which membrane is used (e.g. blood brain barrier permeation or intestinal absorption). One has to take into account that PAMPA today is a summary term on a lot of different methods applied in different laboratories using different membrane constituents, sink conditions, permeation times etc., which makes inter laboratory comparison difficult. 

Some reagents react with the initiating radical to give unreactive substances, a process known as inhibition. A common inhibitor for vinyl polymerisations is hydroquinone, which reacts by the transfer of two hydrogen radicals to the initiator radicals (Fig. 2.4). This gives quinone and unreactive initiator and has the net effect of causing a lag time in the polymerisation and a decrease in the initiator concentration. Monomers are often stored in the presence of inhibitor in order to prevent polymerisation. The amount and type of inhibitor may vary depending on the monomer batch and the manufacturer. For inter-laboratory comparisons of materials to be possible, it is therefore important to remove the inhibitor and purify the monomers prior to use [13]. 

The ISO 17025 standard [3] describes monitoring the quality assurance of test and calibration results by, amongst other means, the regular use of CRMs and/or internal quality control using secondary reference materials and by participation in inter-laboratory comparisons or proficiency testing programmes. 

---