Introduction: Concepts in Clinical Laboratory Testing

The method used to perform a measurement of a compound in a biological sample is extremely important to know so that it is possible to determine if the result reflects the true value in the body of the patient. Knowledge of methodologic details is also required to understand the many variables that can make the result incorrect and clinically misleading.

Laboratory testing was performed for decades with no instrumentation in the laboratory. All of the tests were individually developed, often locally by the laboratory director, and performed manually. Prior to the 1970s, it was not uncommon to create a standard curve for glucose measurement by pipetting increasing amounts of glucose into a series of tubes with reagents, plotting the optical density to create a standard curve on graph paper, and then measuring the amount of glucose in the sample by obtaining the optical density of the patient specimen and using the standard curve. Automation of test performance began in the 1960s and 1970s, and then has continued to increase. Tests which are not fully automated are often partially automated. The methods which are fully manual at this point are far fewer than those which are fully or partially automated. The methods described in this text are largely automated to at least some degree. The details of instrument performance and the chemistry of the reactions are depicted in the figures. The methodology, the reagents used, and the instrument all have an impact on the outcome of the test. It is for this reason that healthcare providers and especially experts in laboratory medicine must be fully aware of the methodology implemented for testing and the impact of the methodology on the test result.

When manual methods dominated in most clinical laboratories, the clinical laboratory scientist was almost completely focused on the analytical performance of the test. When automated instruments arrived in the laboratory, the laboratory scientist at the bench had to develop new skills that focused on demonstrating that the instrument was generating accurate test results. The need for a high level of technical competence became required because so much of the testing in the clinical laboratory was being performed on increasingly sophisticated instruments. At this time, the clinical laboratory scientist not only has to assure functional performance of testing instruments, but because there are so many tests which are highly complex, consultation by clinical laboratory scientists with patient facing healthcare providers on test selection and result interpretation is increasingly necessary. If a sample for glucose measurement was collected in a vacuum tube without a glycolysis inhibitor (a gray top tube), and the sample was transported over an 8-hour route before it arrived in the laboratory, the glucose is likely to be artefactually low because of the metabolism of the glucose by the platelets and the white blood cells prior to analysis. A practicing doctor is unlikely to know that the low glucose value is not reflective of hypoglycemia in the patient, but instead represents a preanalytical error. Communication between the clinical laboratory scientist and the healthcare provider, in this example to avoid a false diagnosis of hypoglycemia, is essential to establishing a correct diagnosis.

When a genetic cause for cystic fibrosis was first described in the late 1980s, one mutation in one individual gene was associated with cystic fibrosis. Now more than 2000 different mutations have been found in that one gene, and the genetic classification of cystic fibrosis has produced six major types of the disease, which are treated with different highly expensive medications. What was once a disease thought to be diagnosable with a test for sweat chloride is now a disorder which requires complex genetic analysis. “Molecular” methods (this generally refers to molecular genetic methods, when, in fact, most assays measure the amount of “molecules”) are highly complex and usually quite expensive. Now, for many diseases, it is possible to do one expensive genetic test to make a definitive diagnosis that stands for a lifetime, as opposed to performing many simpler tests that suggest a diagnosis, but never actually establish it. Experts in laboratory medicine who can communicate the results of genetic tests to a well-intended patient facing provider with only a modest understanding of genetic terminology produces an extraordinary benefit for the patient.

The use of point-of-care testing for many compounds other than glucose has often been a matter of some controversy. It is challenging for individuals working in a clinical laboratory to understand that a person with poor eyesight and limited dexterity, and possibly even having difficulty understanding simple instructions, can effectively use a small plastic kit. It must be recognized that performing a test and making a mistake that can produce the wrong result can have lethal consequences. The methodologies used for point-of-care testing are by necessity simpler than the methodologies used in the clinical laboratory itself. For example, the test for the international normalized ratio (INR) for a patient monitoring the effects of the drug warfarin on a point-of-care device involves the use of very different reagents from those used within the clinical laboratory. For the point-of-care test, a capillary blood sample from a finger stick is used to determine the INR, and in the clinical laboratory, venous blood collected in a vacuum tube containing citrate as an anticoagulant is used. It is not uncommon to have a correction factor modify the result from a point-of-care instrument to make it more similar to the result generated in a clinical laboratory.

It is also a matter of controversy if a test that can be performed more often, even if it is less likely to be correct, may be of more benefit than a clinical laboratory test which requires much longer to generate a result. One example is the use of a test to detect antigens for SARSCoV2 performed by the patient at low cost multiple times in a week. This is in comparison to a laboratory-based polymerase chain reaction (PCR) test which may require as many as 7–10 days to provide a result, though with a higher sensitivity to detect a SARS-CoV2 infection. Even if the PCR result is more accurate, some argue that it is clinically more beneficial to perform a less sensitive test every day. This is because during the week of waiting for the PCR test result, a patient infected with SARSCoV2 may experience at least one positive test result using the less sensitive antigen test.

In general, the cost per test is much greater for reagents in point-of-care testing. A test that costs $50 in a drugstore to be performed by the patient as a point-of-care test can easily cost less than $10 in reagents for the same test performed in a clinical laboratory.

An excellent example of variability in the “same” test methodology is the PCR test for SARSCoV2 RNA. There are literally dozens of available PCR tests, but they can vary substantially in methodology. This is because of differences in the primers and probes used as reagents, or in the individual steps in the test for SARS-CoV2 RNA. Some of the available PCR tests have the ability to detect as little as 600 nucleic acid units of the RNA per mL, and other assays are not able to detect SARS-CoV2 RNA unless there are at least 300,000 nucleic acid units per mL. This 5000-fold difference in sensitivity is clearly influential in deciding whether a patient is infected, especially for those who have a low-tomoderate amount of virus in the sample. One of the more insensitive tests for SARS-CoV2 RNA is genetic but has a methodological difference from classical PCR that allows it to have a rapid turnaround time and be used at the point of care. In general, optimizations for speed of test performance to shorten turnaround time increase simplicity of performance or reduce invasiveness of sample collection (sample collection from the nasopharynx is more invasive than collection of saliva). Such optimizations have all been incorporated into methods to detect SARS-CoV2 RNA rapidly at the expense of reduced sensitivity. Very few practicing doctors are made aware of the methods performed in the laboratories which they use, and even if they are aware, the clinical implications of using a particular method are usually not made known to them.

For many compounds measured in a clinical laboratory, it is possible to obtain biological samples from 20 or more individuals without disease, on no medications with no other variables that would influence the result of the compound being quantitated. Statistical assessment of the middle 95% of a Gaussian distributed range can easily be performed to establish a reference range. This within laboratory generation of a reference range removes some of the variables which might be present for a reference range established from individuals who might be different in some way. For example, dietary preferences that influence laboratory tests may be significantly different in one population versus another. In the 1980s, it was realized that the middle 95% of an Asian population demonstrated a much lower total cholesterol value than the middle 95% from a population in the United States. This was quite certainly a result of differences in intake of dietary fat.

With smaller clinical laboratories, and certainly for doctor’s office laboratories and pointof- care testing, it is too difficult to create a locally validated reference range. For that reason, use of the reference range provided by the manufacturer is a common practice.

For some tests, there is a threshold value rather than a reference range. In some tests, the high value is abnormal, while in others the low value is abnormal. The threshold value itself varies by methodology. For some time, healthcare providers were aware of a certain threshold below which a D-dimer level ruled out a deep vein thrombosis or pulmonary embolism. As new methods appeared to measure D-dimer, the threshold values of the new methods were different from the one first introduced and remembered by clinicians evaluating patients who may have a thrombotic event. Because the practitioners had become used to a specific threshold value, the methods introduced shortly after the original D-dimer assays incorporated an artificial multiplication factor to create a result that would be less likely to be misinterpreted.

To perform proficiency testing, a clinical laboratory purchases samples with known amounts of specific compounds through a program that supplies them to the laboratory on a regular basis. Proficiency testing is absolutely necessary to prove that the analytical phase of testing in a given laboratory is at least approximately correct. Proficiency testing does not establish a reference range. Proficiency testing establishes whether a laboratory is providing correct measurements for the samples sent to the laboratory for testing. The results of the proficiency test measurements are sent to a team which performs a complex statistical analysis of the data. The goal of the analysis is to compare test results from laboratories in which the reagents and instruments are the same. In this way, the proficiency testing process avoids making inappropriate comparisons between laboratories that use reagents or instruments which are not the same. The licensing of laboratories to perform tests and bill for the results is highly dependent upon success with proficiency testing, as it is a major portion of the laboratory inspection. Maintaining the integrity of samples in shipping and handling is a major task for those who manage and sell proficiency tests samples.

Companies that sell proficiency test samples do not sell samples for every possible measurement, often because they do not have a central group which can define what the acceptable range or threshold should be. To know with confidence that all test results generated in a clinical laboratory are correct, it is necessary to document an analysis of some kind which demonstrates accuracy for whatever is being measured.

When a lab test result is unexpected by healthcare provider, a common interpretation is that the laboratory made some kind of error in the analysis of the specimen. In fact, that is not true for the vast majority of such cases. The assays now are so reproducible, especially the automated ones, that analytical errors are much less common than errors that occur before the sample is ever tested. These are called preanalytical errors. For example, for some tests the patients must be fasting; for other tests the samples must be collected in a specific vacuum tube with an anticoagulant; and for some tests the sample must be kept on ice during transportation. If there is a mistake in any of these steps, the error occurs before the test is ever performed, and that is why it is known as preanalytical. To the practicing doctor, the failure to instruct the patient to fast for at least 12 hours before a sample is collected, which produces a result that makes no medical sense for the patient’s condition, is most often perceived as an analytical failure in the clinical laboratory. Therefore, it is important to know that selection of an optimal methodology, even with a flawless measurement, does not guarantee that a patient’s measured result will represent what is actually present in the patient’s body because preanalytical errors are extremely common.

Some patients will need a test that is too complex to be performed in a nearby laboratory to establish or to rule out a diagnosis. This is true for large and small laboratories because no laboratory performs every available laboratory test. Selecting a single commercial laboratory for most tests that must be sent out is a major decision which involves, among other variables, the methods used to perform measurements. It is important to know which methods are used by the commercial laboratory, so that the results are understood by the laboratory scientists at the laboratory from which the sample was sent. Issues like potential interferences that affect a particular laboratory test can then be communicated to the ordering provider. The preanalytical, analytical, and postanalytical phases are emphasized to different extents by the major commercial laboratories. The choice of a lowcost provider can have negative clinical impact if the commercial laboratory is not willing to pay attention to all the details that will affect the test result. For example, a vacuum tube filled with blood anticoagulated with citrate (a blue top tube), submitted for performance of coagulation testing, must have the tests run within a limited number of hours from the time of collection. If the commercial laboratory allows such samples to sit overnight, it will provide inaccurate results that could alter, for example, the dose of an anticoagulant, and permit a patient to suffer a bleeding or thrombotic disorder. Another critical point about tests that a clinical laboratory may perform in low volume is that for some of these, the results are needed as soon as possible for both quality of care and cost considerations. An example of this is a test for heparin- induced thrombocytopenia. If the patient has this condition, the treatment involves substitution of heparin with a very expensive alternative anticoagulant. It can cost hundreds of thousands of dollars for this alternative anticoagulant to be used while waiting for a test result from a commercial laboratory. For that reason, maintaining a test which is not often performed in a clinical laboratory for a typically urgent decision makes both clinical and financial sense.

The methods described in this book are predominantly the common ones found in clinical laboratories. Each method description provides an overview of the basic concept of the assay, minimizing the details, while including clinically important information and a comment on the expense of the test and the complexity of the assay in the laboratory.

Each method also has a descriptor to reflect whether it is manual, semiautomated, or highly automated. A comment has been added if microscopy is involved, as this makes any technique highly manual. It should be noted that for some methods, there is an option for manual performance or for using some level of automation. There is usually greater automation in the larger clinical laboratories because larger laboratories are more likely to have the test volume and the financial resources to justify the automated option. The purchase of test reagents in large quantities also results in a lower purchase price. The term “semiautomated” indicates that there is a manual component associated with the use of an instrument that performs some steps of the analysis.

The expense assessment attached to each assay described in this chapter is an approximation, listed as low, moderate, or high. The charge for the test set by the institution operating the laboratory is usually proportional to the expense of the reagents, supplies, and labor required to perform the test. On occasion, however, there is a great disparity between the actual expense to perform the test and the amount charged by the institution for the assay. With this in mind, the expense estimation provided for each method in this chapter is more closely related to the actual cost of reagents, supplies, and labor in the laboratory, with the understanding that the amount charged for the test should be in the same range of low, moderate, or high—but it is not always the case.

The turnaround time for an assay is not provided because it is impossible to know all of the elements associated with the turnaround time for a test within an individual institution. Broadly speaking, the turnaround time is shorter for assays that are highly automated and less expensive, and longer for assays that are manual and highly expensive. It is important to understand that the turnaround time for an assay can be calculated using different starting points. For example, one starting point is the time a sample is collected. Another starting point is the time that a sample enters the laboratory. However, the most relevant starting time clinically, which predates the previous two starting times, is the time at which the physician actually orders the test. Similarly, there are different end points in the assessment of turnaround time. Most commonly, the end point is the time at which the result is reported by the laboratory into the laboratory information system. However, it is most important to know when the physician becomes aware of the result, because it is at that point that treatment may commence.

Finally, it should be noted what methods are not presented in this chapter. There are a number of methods that have been used progressively less over time, and in many institutions these assays are no longer performed at all in the clinical laboratory. These are numerous and include, among many, the radioimmunoassay (RIA), immunoelectrophoresis, lipoprotein electrophoresis, and the bleeding time. A major effort was made to greatly expand the number of molecular genetic methods. These molecular genetic methods are described in this book. These are often highly complex, and the intention was to provide enough detail for the methods to be understood without making them confusing at the same time.