Search

Guaiac-based faecal occult blood tests (gFOBT) have some limited applications in colorectal cancer screening programmes, but are not recommended for assessment of patients with lower abdominal symptoms and are badly used in clinical practice. Faecal immunochemical tests for haemoglobin (FIT) are replacing gFOBT in screening programmes because they have many advantages, including that faecal haemoglobin concentration can be quantitated using automated immunoturbidimetry. In countries with limited colonoscopy resources, the services are now becoming overwhelmed with referrals from primary care as well as from screening and surveillance programmes. Options to deal with this include reinforcing current objective-based guidelines, expanding colonoscopy resources, or using simple faecal tests to assess patients presenting with suspected colorectal disease. There are many data on use of FIT in screening for detection of colorectal cancer and advanced adenoma: newer data on the use of FIT in primary care show that, although faecal haemoglobin is not a good diagnostic rule-in test for colorectal disease, it has very high negative predictive value, making it a very good rule-out test. Along with use in surveillance programmes, this application of faecal haemoglobin assays would improve use of colonoscopy services and facilitate patient care. Faecal haemoglobin analyses should become an essential examination.

Since 1999, the ‘Consensus Agreement’ from the Stockholm Conference has provided professionally accepted guidance as to a hierarchical approach to setting analytical performance goals. However, many aspects of laboratory medicine have changed since then. In consequence, the 1 st EFLM Strategic Conference on ‘Defining analytical performance goals 15 years after the Stockholm Conference on Quality Specifications in Laboratory Medicine’ was held on 24–25 November, 2014. The conference developed a proposal for a modified and updated consensus statement. There was general agreement that three different models, again set in a hierarchy, should be used to set analytical performance goals. The proposal is simple. Model 1 is based on the effect of analytical performance on clinical consequences, which can be assessed using different approaches including outcome studies, simulations of uses of test results, and surveys of clinicians’ and/or experts’ opinions investigating clinical settings. Model 2 is the widely used strategy based on components of biological variation. Model 3 is founded on state of the art, but that should be based on the highest level technically achievable or, alternatively, could be defined as the analytical quality achieved by a certain percentage of laboratories. Whether this new approach becomes widely adopted remains to be assessed.

Generation of numerical estimates of the within- and between-subject components of random biological variation has been studied for over four decades. These estimates are applied to set analytical performance goals for imprecision, bias and total analytical error, examine the utility of conventional population-based reference values through investigation of individuality, and assess the significance of changes in serial results from an individual using reference change values. The number of publications still appearing on newly- and well-studied analytes shows the ongoing interest in biological variation. Moreover, the classic works continue to be well cited. New concepts appear regularly. Applications of numerical data are facilitated by the availability of a database ( www.westgard.com/biodatabase-2014-update.htm ) that is updated every two years. However, there have been criticisms over recent times regarding the validity of these data, particularly the heterogeneity of estimates of the components of variation published for a single analyte. This has led to recent publication of the methodology used to generate the database and a plea for harmonisation and standardisation of terms and symbols. A Working Group of EFLM aims to improve the performance of studies by dissemination of a check-list that will stimulate better execution and documentation. Biological variation does remain relevant.

Calprotectin is a heterodimer of calcium-binding proteins released from polymorphic nucleated cells and monocytes upon cellular activation (or death) at sites of active inflammation. It is present in plasma and faeces and said to be highly stable in faeces after collection. The symptoms of functional disorders such as irritable bowel syndrome (IBS) and organic inflammatory bowel disease (IBD) can be very similar, but these are two very different clinical conditions. Recent guidelines from NICE state that faecal calprotectin testing is recommended as an option to support clinicians with the differential diagnosis of IBD or IBS in adults with recent onset lower gastrointestinal symptoms for whom specialist assessment is being considered, if cancer is not suspected, having considered the risk factors (for example, age), and appropriate quality assurance processes and locally agreed care pathways are in place for the testing. This guidance has led to an explosion in requests for calprotectin. Calprotectin is also becoming widely used in monitoring treatment of IBD. A number of methods for measurement of calprotectin are available, including point-of-care tests. However, there are still many problems associated with the assays, particularly the pre-analytical aspects. Moreover, the setting of appropriate cut-off concentrations is still much debated.

An inter-laboratory survey of qualitative urinalysis was carried out in Australasia during 1981. Eighty-one laboratories analysed 6 samples of urine distributed in 3 batches of 2 at regular bimonthly intervals, mostly with commercially available reagent strips. Fewer than 30% of laboratories performed any form of quality control for analytes other than pH. Results indicated that improvement in the analysis of urine bilirubin, protein, glucose, ketones and blood is required, and recommendations to improve standards are made.

Summary A regional survey of the performance of quantitative analysis of ten commonly requested urinary analytes was carried out in South Australia. Using the results achieved by the better laboratories as a basis, an empirical set of analytical goals for total laboratory error was derived; this is the first definition of such goals for quantitative urinary analysis. The goals are particularly for use in the assessment of laboratory performance in inter-laboratory surveys.

Analytical performance standards for the imprecision of commonly requested clinical biochemistry tests are proposed in the form of goal graphs. The goals delineated in graphic form are a consensus of published performance standards. The graphs define acceptable and unacceptable performance standards for levels of analyte commonly encountered in clinical laboratory practice. The graphic approach allows a flexible presentation of goals that is easy to interpret. The disadvantage of goal graphs is their necessarily empirical foundation.

Summary Eighty-six Australasian laboratories participated in an inter-iaboratory quality assurance programme for 10 urine analytes. Twelve liquid samples were prepared from commercial lyophilized urine control material and distributed in 3 batches of 4. Use of pre-set acceptability limits for total laboratory error and target values facilitated timely feedback in graphic form. The samples had concentrations which were linearly related; this allowed simple calculation of overall imprecision and bias, graphic feedback of all submitted results, and comparison of performance between laboratories. A number of unsuitable and poor methods were identified. Particular attention must be paid, in future, to more widespread use of appropriate calibration and quality control materials, to avoidance of transcription and calculation errors, and to analysis of urine samples with elevated levels of analyte. Current laboratory performance can meet analytical goals for analyses of urine creatinine, phosphate, urate, and glucose but analysis of urine sodium, potassium, urea, calcium, osmolality, and proteins require significant improvement.

An evaluation of the Beckman Creatinine Analyser 2 has been carried out. The assay is based upon a kinetic alkaline picrate method. Good day-to-day and within-day precision (CV = 3.4%; upper limit of reference range) was obtained. Analyses of quality control materials and materials with consensus values, and comparison of the results obtained for samples of plasma and urine from patients with the results obtained by an alkaline picrate continuous-flow method, evidenced good accuracy. The method was linear up to 3.00 mmol/l. Glucose, lipaemia, haemolysis and icterus caused negligible interference in analysis of plasma creatinine. Albumin, ascorbate, glucose and acetoacetate caused negligible interference in analysis of urine creatinine. Acetoacetate in serum, however, caused significant positive interference. The desirable performance characteristics and practicability make this analyser attractive to the clinical chemist, particularly for paediatric and emergency laboratory use.

The precise and accurate analysis ofiron binding capacity still poses a number of problems. A commercial kit (RESO-MAT, Mallinckrodt Australia Pty Ltd) for the determination of total and unsaturated iron binding capacities based on the use of Fe 59 as a transferrin saturant has heen evaluated and compared with a mechanized colorimetric method. The method was found to have acceptable precision and accuracy. The results obtained by the radiometric method compared well with those obtained using a colorimetric method. The radiometric test kit has a number of advantages over colorimetric methods in that haemolysis, lipaemia and icterus do not interfere, preparation of iron standards and calibration curves is not required, and minimal preparation and/or addition of reagents is necessary.

A favoured concept is that all analyses should be standardized with the purest materials available. This ideal is unattainable as a significant number of problems remain unsolved and, to eliminate certain of these problems, secondary standards are often used. An alternative approach is to relate concentration directly to absorbance by a calibration relationship dependent on the molar extinction coefficient of a reaction product or reactant. The present study examined the dogma that a standard must be analysed in each batch of tests by comparing the analytical performance achieved in 4 manual colorimetric tests urea, glucose, phosphate and albumin—using daily primary or serum standards (the variable calibration mode) with that obtained using constant calibration relationships (the constant calibration mode). It was shown that the mode of standardization selected for each and every method must be chosen as an objective consequence of experimental study.