Abstract
Lack of reference materials and standard procedures, on faecal tests leads to major problems in harmonisation of methods and do not allow the comparison of outcome data.
In particular the absence of standardisation of pre-analytical characteristic was noted for faecal test methods for haemoglobin since different manufacturers have developed different sampling procedures and report units.
Moreover the physical characteristics of the faecal specimen and the designs of specimen collection devices do not allow analysis of samples on different systems in consequence, faecal tests cannot be compared using standard evaluation protocols.
To improve the harmonization of results generated using different analytical systems and the overall performances of test on faecal materials we propose the introduction of standard procedures for sampling and pre-analytical phase and the adoption of specific procedures based on the use of artificial biological samples for comparison of methods.
Harmonization of sampling devices with the use of a standard design for pickers and a standard ratio between analyte and buffer for different manufacturers represent a mandatory step in the roadmap for harmonization of clinical laboratory measurement on faecal materials and can allow a significant standardisation of results generated by different devices.
The creation of specific protocols for the evaluation and comparison of analytical methods for analyse of faeces could lead to a significant improvement in the performance of methods and systems.
Clinical protocols for the use of the fecal immunochemical test (FIT) for hemoglobin and the resulting economic implications have become of paramount importance. Recent EU recommendations (1) suggest that all member countries adopt best practice in population-screening programs for early colorectal cancer (CRC) detection and follow detailed EU guidelines (2) when using existing screening strategies, including FIT.
Currently, the results obtained from the comparisons of FIT methods are flawed due to the lack of standardization of the pre-analytical phase and due to the wide range of designs of the studies undertaken worldwide to assess the performance characteristics.
In 2012, discussion documents on the use of FIT for screening from the Expert Working Group (EWG), established by the Colorectal Cancer Screening Committee of the Word Endoscopy Organization (WEO) (3, 4) have urged manufacturers and other authorities to improve a number of aspects of FIT standardization, and indicated that clinical research must follow the recommendations in performance and data reporting.
EWG recommends that quantitative data on fecal hemoglobin concentrations should be reported as µg hemoglobin/g feces rather than as ng hemoglobin/mL buffer, since the latter unit of measure is specific for each specimen collection device used, making transferability of data across studies impossible (5). Furthermore, the mass of fecal sample used for hemoglobin measurements and the volume of buffer in the collection device are also inconsistent among manufacturers (5).
These issues could be resolved through the standardization of the procedures used in the pre-analytical phase, thereby making FIT results more consistent across time, geography, and methodology. However, several aspects of the pre-analytical and analytical phases on fecal tests need to be examined in detail.
Standardization of the Sampling and Pre-Analytical Phase
Specimen collection devices consist of different components, all of which have an effect on the fecal hemoglobin concentration measured and, consequently, on the screening result and outcome. The following aspects can all influence the pre-analytical phase:
specimen collection probes/pickers that collect different fecal masses;
collars in the buffer tubes that remove excess feces from the probe/pickers and prevent leakage;
different buffer volumes in the tubes;
buffers with different preservatives and hemoglobin stabilizers.
The physical characteristics of the fecal specimen, such as texture and hardness, and the probes/picker design determine the amount collected from the feces and has an impact on the mass collected (examples are shown in Fig. 1). The collar regulates the mass of feces passed into the collection tube, but it does not guarantee that the exact quantity of feces claimed by the manufacturer has been collected.
Specimen collection devices and probes/pickes used by different manufacturers for fecal immunochemical test for hemoglobin analytical systems.
Buffer composition as well as preservatives and stabilizers used by different manufacturers play an important role in hemoglobin stability at different temperatures, as demonstrated in a study on FIT variability in different seasons reporting a significant relationship between the clinical efficacy of FIT and temperature (6). Despite the current improvements in stability of hemoglobin in the recently introduced buffers, this issue still poses a problem in the interpretation of existing data, since clinical research publications often do not report details of the methodology used, and manufacturers may modify the reagent composition without informing the users of their possible effects on analytical performance and clinical outcomes.
Furthermore, the collection devices and buffer compositions, crucial to the determination of fecal hemoglobin concentrations, are often covered by patents, a fact that renders it extremely difficult to reach a consensus for the standardization of design aspects relevant to the pre-analytical phase.
We propose that only through probe/picker design standardization, which might be achieved without significant investment by manufacturers, the collected fecal mass and buffer volume will be standardized, leading to a significant reduction in pre-analytical variability and enhanced comparability across data.
The status quo so far has been that manufacturers present a specific design and declare the fecal mass collected by their own collection device. However, considering that fecal characteristics vary considerably and cannot be standardized, the mass declared by the manufacturers is unlikely to apply to all fecal samples and will vary across different collection devices.
Diet and fecal consistency can affect the fecal mass collected by the different probes/pickers and this should also be taken into account in the assessment of pre-analytical sources of variation. Considering that CRC screening is practiced on a global basis, the habitual diets in different countries could determine a systematic bias in the fecal mass collected by different devices, though it should also be taken into account that the dietary effects for a population screened in a specific country are not necessarily the same (7).
In our opinion, the best strategy to harmonize fecal hemoglobin concentrations generated by different methods is a complete standardization of the design of probes/pickers, collars, buffer volume, and buffer composition, with these being produced by all the different manufacturers.
As a consequence of this, a marked improvement would be gained in terms of standardization of fecal mass collected in different specimen collection devices to be analyzed by different clinical researchers, without excessive investment in research or in the production of new specimen collection devices. Figure 2 shows some examples of a possible standardized design for probes/pickers as to regulate the relationship between fecal mass collected and buffer volume, in order to reduce the overall variability.
Example of a standardized design for probe/pickers for fecal specimen collection and a suggested relationship between sample size and buffer volume to reduce the overall specimen collection variability.
The introduction of a major standardization of sample collection devices would reduce the variability due to fecal characteristics, even though it is impossible to completely ignore this aspect. It is also possible that the nature of the plastic (in particular the porosity) of the collection devices could play a role in the final estimation of fecal hemoglobin concentration, but it is assumed that this plays only a minor role in the pre-analytical phase.
We recognize that manufacturers are not keen on changing their own collection devices, and are actually very proud of the benefits of their individual products. Possibly, only with the full involvement of the European Regulatory Authorities this issue may be resolved; however, we are hopeful that existing scientific data may support such a change. The standardization of the pre-analytical phases has played a crucial role in many fields of laboratory medicine. In particular, the use of a standard blood/citrate ratio has significantly improved the comparison of coagulation parameters generated by different analytical systems (8).
The standardization of the design and the characteristics of probe/pickers used for fecal sampling could play a significant role not only for FIT, but also for other fecal tests. Starting from the design of a single specimen collection device, it is possible to adjust the amount of feces needed for the analysis through manipulation of the ratio between the mass collected by the probe/picker and the buffer volume, as to obtain the analyte concentration in the buffer required for its analysis (Fig. 2).
It must also be considered that sampling problems related to the possibly intermittent nature of bleeding into the colorectum, and the lack of homogeneity of any blood present in a single fecal sample will not be resolved by our suggested sampling strategy; however, our strategy will have a beneficial effect on the comparison of clinical data obtained by different systems (9, 10) and in different screening programs (11, 12).
Revision of the Procedures for Comparison of Methods
The non-homogeneous hemoglobin distribution in feces and differences in the fecal characteristics, which lead to considerable inter-sample variation, as described in the first section, do not allow the collection of identical specimens from the same fecal sample. Moreover, the lack of standardized specimen collection devices does not allow direct sampling from the same tube on different analytical systems without significant sources of error. Therefore, comparison of data obtained from different analytical systems for fecal hemoglobin cannot be performed without considering and assessing the pre-analytical and analytical phases separately.
Traditional protocols for the assessment of analytical performance characteristics involve the comparison of the system under investigation with a well-characterized reference (predicate) system.
All analytical systems for fecal tests are designed to work with the related sampling devices and, inter alia, ionic strength and pH of the material sampled for analysis, as well as analytical working ranges, are closely related to the specific system.
The introduction of specific protocols for comparison of methods of fecal analysis and the separate investigation of the pre-analytical and analytical phases could be useful to address these issues.
We suggest the introduction of 2 procedures that somewhat vary from the standard protocols normally used for comparison of methods.
Both procedures adopt the use of artificial biological samples (ABS) obtained by adding different amounts of analyte to native fecal specimens with no detectable analyte, often termed spiking.
Suggested Procedures for Comparison of fit for Hemoglobin
Details of the suggested procedures may be fully addressed only by a specific working group or committee for fecal test standardization, and the final draft should be an issue of the regulatory authority. Our attempt was to submit a new strategy to approach the fecal test investigations.
ABS for the investigation of FIT for hemoglobin are prepared adding different amounts of a human blood lysate to freshly collected human fecal samples with no detectable hemoglobin content. Final concentrations of hemoglobin in ABS are determined considering the analytical working range and cutoff concentration suggested by the manufacturer. The samples are kept at 2°C-4°C (for no more than 48 hours) or at -20°C (for more than 48 hours) until they are used. A 156 mmol/L aqueous sodium chloride solution is uses as diluent.
The aim of this procedure is the generation of data on the concordance between 2 methods, the evaluation of the comparative bias in the analyte measurements from different systems, and the investigation of the effect of specimen collection devices on the measurements.
A 2 x 2 contingency table can be built using the same data to simulate the diagnostic assessments on the material analyzed. The presence or absence of analyte in the ABS is used as reference, and the cutoff for fecal hemoglobin concentrations suggested by the manufacturers are used in the assessment of diagnostic accuracy. Preliminary information on sensitivity and specificity of methods can be obtained from these data.
Conclusions
Currently, there are no specifically designed protocols for the assessment of the pre-analytical and analytical performance characteristics to facilitate the standardization for defining and comparing data on fecal testing. Manufacturers have developed different specimen collection strategies for different tests and methods, and any attempt to compare method performance characteristics can be based only on clinical epidemiological investigations with a consequent increase in the costs and the overall working time.
In our opinion, the creation of specific protocols for the evaluation and comparison of analytical methods for the analysis of feces could lead to a significant improvement in the methods and systems available in the EU market, and to a significant standardization of the overall results generated using different devices. In particular, the harmonization of the sampling and pre-analytical phase represents a mandatory step in the roadmap for the harmonization of clinical laboratory measurement on fecal materials (14).