In health care, the analysis of body fluids plays an important role in the diagnosis and management of a wide variety of conditions. Traditionally, clinical laboratorians have provided analysis of body fluid specimens without question while also recognizing the sometimes difficult work that goes into their collection. As standards for method validation and laboratory developed tests (LDTs) evolve, the analysis of body fluids is receiving increased attention from both laboratories and regulatory bodies.
The clinical laboratory’s overarching goal is to ensure accurate test results from all specimens. Therefore, it is the responsibility of every laboratory to investigate the analytical performance of the tests performed on the various fluid types accepted and to provide a context for result interpretation.
The Importance of Body Fluid Chemistries
A review of the variety of patient specimens that are submitted to the clinical laboratory illustrates that several major fluid types do not fall into conventional specimen categories. However, proper analysis of these fluid types—including peritoneal, pleural, synovial, amniotic, pericardial, and cerebrospinal fluid (CSF)—is integral to the diagnosis and management of a multitude of disease states. Among the major applications of body fluid analysis is the differentiation of transudates from exudates for fluids originating from pleural effusions.1 Given the commonality of pleural effusions presented in the lab, clinical distinction between transudate and exudate pleural effusions aids in the diagnosis of the origin of the pleural effusions, which in turn affects the differential diagnosis and treatment. Further analysis may include albumin, pH, lactate dehydrogenase (LDH), and total protein, whereas additional applications include identifying bacterial infection, diagnosing chylous ascites, and distinguishing the origins of peritonitis.
Emerging applications of fluid analysis include the measurement of antibiotic concentrations in CSF to achieve sterilization in the setting of bacterial meningitis.2 Ultimately, regardless of the fluid type or application, its thorough analysis in the laboratory plays an important role in patient care.
Off-Label Use of FDA-Approved Assays
For most chemistry assays, manufacturers have defined acceptable specimen types as serum, plasma, whole blood, and urine. With the exception of glucose, lactate, immunoglobulin G (IgG), and total protein for CSF, alternative fluid specimens likely are not evaluated or included in FDA-510k submissions. Laboratory use of a specimen type that is not covered by an assay’s intended use statement constitutes off-label use of the reagent; a process that begets several consequences. Using reagents outside their intended use defaults the test classification to high complexity. The assay then becomes an LDT, which requires validation. As the impact of a test being classified as an LDT continues to evolve, it appears clear that the FDA is increasing its oversight of assays falling into this category.
CAP Guidelines and Regulatory Requirements
Since 2013, the CAP All Common Checklist has contained one item dedicated to body fluid testing. Within the most recent checklist iteration, this one item (COM.40620; Body Fluid Validation) requires that the testing of body fluid specimens using methods intended for other specimen types (eg, blood or other fluid) have been validated by the laboratory for accuracy, precision, analytic sensitivity, analytic specificity, interferences, and reportable range.3 However, this item applies only to fluids used in tests that the laboratory considers routinely ordered. Other specimens that the laboratory may receive outside of routine orders can be considered clinically unique and should require director or designee approval prior to analysis. The validation or verification of these tests must meet all of the criteria listed elsewhere in the checklist for method performance specifications, as well as those pertinent to LDTs. This includes adding these fluid tests to the lab’s running list of LDTs, listing appropriate clinical claims, and appending to the results a statement that the laboratory developed the assay.
Strategies for Validation
The approach to validation of body fluid specimen testing for chemistry analysis is similar to that of other validations performed in the laboratory. FIGURE 1 outlines an example of a typical workflow required to validate a series of body fluid tests. The strategies described are meant to serve as a guide and should not replace the laboratory’s specific accrediting agency’s regulations. It is the responsibility of the laboratory director to determine the extent of validation or verification required.
The first step in adopting the CAP body fluid checklist item is to decide which tests require validating. CAP requires validation only of body fluid tests that are routinely orderable at a given institution. A retrospective analysis of data from the laboratory information system (LIS) can identify frequently ordered tests involving body fluids. At our institution, we identified five major fluid types and a total of 28 analytes as routinely ordered (see TABLE). In most cases, the patient population served by each laboratory will influence which (if any) body fluid tests require validating.
The second step is to acquire actual specimens. As body fluid samples or appropriately matched matrices are not commercially available, the laboratory must collect its own specimens for testing validation. This process may take several months and may require collaboration with other laboratories in order to obtain the right type and amount of fluid. Collaborating not only increases sample numbers, but establishes an outside contact for method comparison studies. During the collection phase, the laboratory should continue to plan the validation phase by identifying relevant medical decision points (MDP) and researching relevant reference intervals for the tests in question. Once a sufficient number of specimens are collected for each fluid type, the recommendation is to analyze all fluid specimens for every test being validated. Doing so accounts for any degradation of target analyte during the collection phase and provides the laboratory with an estimation of what can be used for precision, spiking, and mixing studies.
If the laboratory performs its own assay validation without relying on method performance specifications from the manufacturer or published literature, it should include accuracy, precision, method comparison, limit of detection (sensitivity), interfering substances (specificity), and reference interval studies per the CAP checklist.
Precision and Sensitivity
During the specimen collection phase, the laboratory should research how the specific analytes measured are used clinically. This exercise sets the required analytical measuring ranges and any relevant MDPs, and understanding these two components helps guide precision and sensitivity studies. For instance, glucose values less than 50 mg/dL may be used to differentiate secondary bacterial peritonitis from spontaneous bacterial peritonitis. If all of the peritoneal fluids collected by the lab have values similar to serum in the 100-150 mg/dL range, the laboratory should be mindful that the medical decision point is actually 50 mg/dL. Appropriate dilutions should be made to achieve a specimen around 50 mg/dL to demonstrate acceptable precision at the clinical cutoff and also near the lower limit of the measuring interval for the glucose assay (possibly as low as 10 mg/dL). Several CLSI documents (eg, EP15-A2 and EP5-A2) detail approaches to precision studies that laboratories may choose to follow.4,5 Conventional approaches to establish precision performance specifications for FDA-cleared assays usually involve two levels of material above or below a relevant MDP or simply with low and high values that are run consecutively 20 times. Standard deviation (SD) and coefficient of variation (CV) are then calculated to assess intra-run precision. A CV below 15% often is used as a general guide for acceptable precision.
In addition to precision, the laboratory needs to demonstrate and define the limit-of-detection (sensitivity) using the fluid type being validated. Using the values generated from the initial analysis of all specimens, the laboratory can choose a single specimen at or near the lower limit of the reference interval to validate the sensitivity. If a specimen is unavailable at the cutoff, two fluid specimens can be mixed to achieve a suitable concentration. Once an appropriate fluid specimen has been identified, a precision study can be performed to calculate both inter– and intra-run precision. The generally accepted CV criterion (at the lower limit of the reference interval) is suggested to be less than 20%.
Methods to Determine Accuracy
Laboratories can employ several distinct strategies to determine the accuracy of testing using different fluid types. The simplest strategy involves taking an approved specimen type with a known high value of the target analyte and creating a series of mixed samples containing increasing concentrations of fluid. The samples are analyzed and the measured values are compared to the expected values for each mixture. The difference between the expected and measured values can be used to assess or exclude the existence of fluid matrix effects.
Another approach uses information from the initial measurement of all fluids for mixing studies. For each analyte validation, the specimen with the lowest analyte concentration and the specimen with the highest analyte concentration are used to create mixtures across the analytical measuring range. Results ratios of 0:100, 25:75, 50:50, 75:25, and 100:0 mixtures can be used to generate a five-point curve for every fluid analyte. Expected and measured concentrations then can be compared to evaluate the linearity of the assay.
A third approach allows for interrogation of accuracy when available fluid specimens do not approach the upper limit of the analytical measuring range, and also provides an objective evaluation of matrix effects for that fluid type. The highest calibrator or synthetic calibration verification material is used for spiking studies of a fluid specimen with a low target analyte value. For these mixtures, as the fluid specimen concentration increases, so does the percent matrix; a spectrum ranging from 0% fluid to 100% fluid will help indicate the extent of potential matrix inferences present and the deviation from the expected target value. For fluid types that vary significantly in protein content and viscosity, multiple fluid specimens should be measured using this calibrator spiking method.
For most test validation studies, laboratories perform some form of method comparison. This can be a comparison of old versus new instrumentation or a comparison of results with an outside facility (a process that can provide an accuracy assessment independent of the laboratory’s current platform). The addition of fluid validation requirements in the CAP checklist likely is indicative that many laboratories do not retain data from previous instrumentation and that outside facilities remain the best option for correlation studies in this instance. If laboratories work together during the specimen collection phase, they also can serve as collaborators for method comparison studies. Measurement on different instruments from different manufacturers provides additional insight regarding platform-specific biases. Comparisons between wet chemistry analysis and dry slide technology adds an additional method to evaluate matrix biases.
Specificity and Interfering Substances
The possibility for unexpected interferences may increase when unconventional fluids are tested. The approach for evaluating assay specificity and identifying interfering substances depends upon the results of the accuracy study. Current CAP guidelines state that performance specifications provided by the manufacturer can be used if the laboratory can reasonably exclude the existence of matrix effects. If accuracy studies find no significant difference between a certain body fluid and other approved specimens, then referencing the package insert or published literature may be sufficient. If the fluid type shows unique properties, such as high viscosity or large degrees of lipemia or hemolysis, the laboratory may choose to perform interference studies that are unique to that fluid type. These may include deproteinization or enzymatic treatment to reduce viscosity.
Establishment of reference intervals for body fluids may seem to be a daunting task given their clinical use. If conventional approaches to establishing reference intervals were used for body fluids, many labs would ask: How do we collect fluids from normal individuals? Furthermore, what is considered a normal fluid? For many body fluid chemistries, the results are used in comparison with the same analyte measured in blood (see FIGURE 2A). The comparison can either be a gradient difference between specimen types or a ratio of blood to fluid. Current CAP guidelines state that a reference interval must be reported with the fluid results unless the value is compared to its concentration in blood. The intent of any reference interval is to provide context for clinical decision-making. This can be achieved through a traditionally established normal range or through an interpretive comment describing how the result should be used. Published literature can be helpful in crafting comments that describe how to interpret the fluid analyte relative to blood and any relevant clinical cutoffs or MDPs based on those relationships. Such interpretive comments provide guidance for proper calculation with blood concentrations and are therefore clinically relevant.
LIS and Other Considerations
The functionality of a laboratory’s LIS can affect how ordering and reference intervals are handled. If the LIS does not capture specific fluid types, then it will be challenging to identify the specific fluids received and determine which fluid types need validating. The laboratory may choose to build order sets containing both blood and fluid analytes, which would allow for automated calculations of relevant gradient or ratios (see FIGURE 2A). If paired ordered sets are not built, then expanded choices of specimen types (eg, pericardial, peritoneal, synovial, cyst, CSF, etc) can be added to allow for interpretative comments. In this case, the LIS performs a lookup calculation using the fluid type and analyte to determine the correct comment to append to the test result (see FIGURE 2B). Each laboratory should decide what approach works best for its LIS, workflow, and test menu.
In addition to the CAP checklist item specific to body fluids, the laboratory should be aware of checklist items related to LDTs. These items detail that LDT results should include a statement that it was developed by the laboratory and meets certain performance characteristics.3 Any clinical claims based on the use of the LDT should be referenced and these assays also should be added to the laboratory’s active list of LDTs with an established method for proficiency testing (PT). If commercial PT material is unavailable for the target fluids, then a system of alternative proficiency verification should be developed using blinded samples. The laboratory director should establish acceptable criteria for alternative PT relevant to the clinical utility and any relevant MDPs.
Body fluid testing provides important information for clinical diagnosis of a variety of medical conditions. The lack of commercially approved clinical diagnostics for fluid chemistries means facilities often must rely on LDTs to meet these unique clinical needs. At first glance, the validation process may seem extensive and intimidating. However, method performance specifications for fluids on several instrumentation platforms have been published that can aid the laboratory in its verification studies.6,7 Planning the validation strategy during the specimen collection phase can greatly streamline the work required. Furthermore, collaboration with other laboratories can strengthen validation sample sizes and provide a means for intra-platform method comparisons.
Steven W. Cotten, PhD, DABCC, FACB, is an assistant clinical professor and Director of Chemistry, Immunology, and Toxicology at the Ohio State University Wexner Medical Center. He received his PhD in pharmaceutical sciences from the University of North Carolina at Chapel Hill and completed a fellowship in clinical chemistry in the department of pathology and laboratory medicine at UNC Chapel Hill. Steven is a diplomat of the American Board of Clinical Chemistry and a fellow of the National Academy of Clinical Biochemistry.