The never-ending search of an acceptable compromise for pancreatic lipase standardisation.

Human pancreatic lipase (HPL) is a single-chain glycoprotein with a molecular weight of 48,000 Da (1) . Its concentration in the pancreas is approximately 9000-fold greater than in other tissues, and the concentration gradient between pancreas and serum is approximately 20,000-fold (2) . For full catalytic activity and greatest specifi city, the presence of bile salts (such as sodium deoxycholate) and a cofactor called colipase , a small protein secreted by the pancreas, are required. HPL activity depends on the substrate, which in vivo conditions are glycerol esters of long-chain fatty acids (triglycerides), being present as an emulsion. Thus, the biological reaction does not obey classical Michaelis kinetics and the enzyme activity depends not on substrate concentrations but rather on interface area. Colipase, aided by the presence of bile salts to stabilise the triglyceride emulsion, binds to HPL to form a complex. This association produces a conformational change for HPL, such that the latter can more effi ciently bind to the substrate to produce lipolytic products (3) . HPL measurement on serum has been used for several years to diagnose acute pancreatitis and its elevation is considered a more specifi c diagnostic fi nding than increases in serum amylase catalytic activity (4) . A number of substrates and complex auxiliary and indicator systems are currently used in commercial HPL methods (1, 3) . In general, longchain triglyceride substrates have demonstrated a correlation of results with the biological and clinical state that is superior to that seen with methods using other substrates (5) . In 1994, Lessinger and F é rard fi rst pointed out considerable inter-assay discrepancies in the plasma HPL catalytic activities according to different measurement procedures, possibly leading to incorrect interpretation of test results (6) . In order to improve comparability of HPL measurements and in agreement with a similar approach already recommended for other clinically important enzymes, they proposed the development of a reference measurement system for HPL catalytic activity, including both a reference measurement procedure and suitable reference materials, based on the concept of metrological traceability (7) . In applying the reference system theory, enzymes represent a special class of analytes (8, 9) . As they are defi ned in terms of the “ catalytic amount ” , their numerical results depend entirely on the experimental conditions under which measurements are made. Therefore, a reference measurement procedure, which defi nes the conditions under which the catalytic activity of a given enzyme is measured, occupies the highest level of the metrological traceability chain (10) . On the basis of the seminal work performed by Tietz et al. (11) , the kinetic titrimetric method employing an automated potentiometric titrator (an instrument commonly referred to as a “ pH-stat ” ) has long been considered as a candidate reference method for HPL. Lessinger and coworkers developed an optimised titrimetric procedure at constant pH for HPL using triolein-based emulsion as substrate, a standardised mode of substrate emulsifi cation, and optimisation of HPL effectors, such as sodium deoxycholate, calcium ions and colipase, and proposed its use for certifi cation of reference materials (12, 13) . Problems affecting the type of emulsifi cation procedure and the reproducibility and stability of the emulsion were examined thoroughly and the feasibility of interlaboratory transferability of the method under controlled conditions was demonstrated by an international exercise (14) . The production of two reference materials for HPL was also described (15, 16) . These enzyme preparations were characterised in terms of catalytic properties, homogeneity, stability, and commutability (16) . Particularly, the reference material containing HPL purifi ed from human pancreatic juice (coded BCR 693) was found fully commutable between the mentioned standardised titrimetric procedure and a widely used colorimetric commercial method, thus potentially serving to transfer trueness from titrimetric procedure (taken as reference) to fi eld methods, through the assignment of traceable values to commercial calibrators, and improve comparability of HPL results (Figure 1 ). With this titrimetry-based reference system in place from 2004, why we are still discussing about the best practical way for standardising HPL measurements ? The concept of a reference measurement system is valid only if the reference procedure and corresponding lower-order routine methods have similar analytical specifi cities toward the specifi c enzyme (7, 17) . In the case of HPL, commercial methods employ different measurement principles that may refl ect differences in analytical specifi city (5) . Furthermore, a number of suboptimal routine methods leading to interference by esterases or nonpancreatic lipases are still available in the market (18 – 20) , even if no signifi cant differences were found in their diagnostic effi ciency when compared to more specifi c HPL assays (21) . Consequently, the idea of relating commercial assays to each other through the use of the highly, but differently specifi c titrimetry-based reference system could create some insurmountable problems in the establishment of traceability of commercial methods. Theoretically, these assays