Reference Intervals in Clinical Chemistry: Establishment and Verification

Clinical laboratories must provide reference intervals that accurately reflect the local population and the specific analytical method in use. A reference interval is the central 95 percent range of test results observed in a healthy reference population, typically defined by the 2.5th and 97.5th percentiles. The Clinical and Laboratory Standards Institute guideline EP28-A3c provides the framework for establishing new reference intervals through direct sampling of healthy individuals or for verifying intervals adopted from external sources such as manufacturers, published literature, or harmonized multicenter studies. This article explains the practical steps for both processes, including sample size planning, reference individual selection, partitioning decisions, transference procedures, and documentation requirements, with attention to the limitations that laboratories face in routine practice.

At a Glance

The decision to establish or verify a reference interval depends on the laboratory's resources, the analyte in question, and the availability of a suitable source interval. The table below summarizes the main pathways and their requirements.

The choice between these approaches is not fixed. A laboratory may verify a manufacturer interval initially and later establish its own interval when resources permit. The CLSI EP28-A3c guideline remains the common standard referenced across published studies in clinical chemistry, including work on calculated globulin, tumor markers, pediatric analytes, and thyroid hormones in pregnancy.

Scope and Purpose of Reference Intervals

Reference intervals serve as the baseline for interpreting individual patient results. Without an appropriate interval, a laboratory result has limited clinical meaning because the clinician cannot determine whether the value falls within the expected range for a healthy person of similar age, sex, and physiological status. The World Health Organization Laboratory Quality Management System Handbook addresses the broader quality framework in which reference intervals must be managed, including the need for documented procedures and regular review.

Laboratory investigations provide objective data that aid in disease diagnosis, clinical decision making, and patient follow up. Clinical interpretation of laboratory test results relies heavily on the availability of appropriate population-based reference intervals or decision limits developed through clinical outcome studies. Although reference intervals are fundamental to accurate laboratory test interpretation, the need for sound evidence-based reference intervals has been largely overlooked, particularly in the pediatric population. In the field of pediatric laboratory medicine, accurate age and sex specific reference intervals established using samples from healthy children and adolescents have not been readily available, forcing many clinical laboratories to report adult reference intervals with pediatric test results.

Reference intervals differ from decision limits. A reference interval describes what is observed in a healthy population, while a decision limit is derived from clinical outcome studies and indicates a value at which medical action is recommended. For example, the 99th percentile upper reference limit for high sensitivity cardiac troponin is used as a decision limit for myocardial injury, not simply as a description of healthy individuals. Laboratories must distinguish between these concepts when selecting or establishing intervals.

Core Principles of CLSI EP28-A3c

The CLSI EP28-A3c guideline defines the procedures for both establishing and verifying reference intervals. The guideline emphasizes that reference intervals are method specific and population specific. An interval established on one analytical platform cannot be assumed to apply to another platform without verification, because different methods may produce systematically different results for the same analyte.

Reference Individual Selection

The quality of a reference interval depends entirely on the quality of the reference sample. Reference individuals must be selected to represent the healthy population for which the interval will be used. The selection process involves both a priori and a posteriori exclusion criteria.

A priori exclusion criteria are applied before sample collection. These include questionnaires and health screens that exclude individuals with known diseases, recent acute illness, pregnancy, use of medications that affect the analyte, and other factors that could influence test results. A posteriori exclusion criteria are applied after results are obtained, such as removing values that are clearly abnormal based on other laboratory findings or clinical assessment.

The CALIPER initiative demonstrated the importance of rigorous reference individual selection in pediatric reference interval studies. The Canadian Laboratory Initiative on Pediatric Reference Intervals recruited healthy children and adolescents through an outreach campaign and applied strict exclusion criteria to ensure that the reference sample represented the healthy pediatric population. This approach addressed the unacceptable limitations of previous pediatric reference intervals that were often established with small sample sizes, inpatient or outpatient samples, outdated methodologies, or inappropriate statistical procedures.

Sample Size Requirements

For de novo establishment, CLSI EP28-A3c recommends a minimum of 120 reference individuals per partition. This sample size allows the nonparametric estimation of the 2.5th and 97.5th percentiles with acceptable statistical precision. A partition is a subgroup defined by age, sex, or other relevant characteristics.

For verification of an existing interval, the guideline recommends a minimum of 20 reference individuals per partition. The verification procedure is straightforward. The laboratory collects samples from 20 healthy individuals who meet the same exclusion criteria used in the original study. If no more than two of the 20 results fall outside the candidate reference interval, the interval is considered verified. This rule is based on the binomial distribution and provides 90 percent confidence that the interval is acceptable.

A study on reference interval verification in routine clinical laboratories noted that the standard approach recommended by CLSI EP28-A3c is to collect and analyze a minimum of 20 samples from healthy subjects from the local population. The same review acknowledged that pediatric and geriatric age groups pose additional challenges in acquiring and verifying reference intervals because of the difficulty of recruiting healthy volunteers in these populations.

Partitioning Decisions

Partitioning refers to the decision to establish separate reference intervals for different subgroups, such as males and females or different age bands. Partitioning should only be performed when the differences between subgroups are clinically significant and statistically demonstrable.

The Harris and Boyd test is commonly used to assess whether partitioning is necessary. This test compares the means and standard deviations of candidate subgroups and determines whether the differences are large enough to warrant separate intervals. A study establishing reference intervals for carbohydrate antigen 72-4 in healthy adults in Shenzhen, China used the Harris and Boyd Z test to assess the need for partitioning by sex or age. The study found no significant sex or age related differences and adopted a combined reference interval.

Age partitioning is particularly important in pediatrics. A study of high sensitivity cardiac troponin T in healthy children aged 0 to 14 years divided the reference sample into nine age groups, from umbilical cord blood to 6 to 14 years. The study found significant differences in troponin concentrations across age groups, with the highest levels in neonates and a gradual decrease to adult reference ranges by one year of age. This finding demonstrates that a single pediatric reference interval would be inappropriate for this analyte.

Sex partitioning is also common. A study of neutrophil to lymphocyte ratio, lymphocyte to monocyte ratio, and platelet to lymphocyte ratio in Chinese healthy adults found significant differences between males and females for all three ratios. The study established separate reference intervals by sex and age group, noting that the reference upper limits changed with age in both sexes.

Establishing Reference Intervals

De novo establishment is the most rigorous approach but also the most resource intensive. The process requires careful planning, recruitment of a sufficient number of healthy reference individuals, standardized preanalytical conditions, and appropriate statistical analysis.

Planning the Reference Interval Study

The first step is to define the purpose of the study and the population to which the interval will apply. The laboratory must decide whether the interval will be used for a specific age group, sex, or clinical setting. The study protocol must specify the inclusion and exclusion criteria, the number of reference individuals needed, the sample collection procedures, and the statistical methods to be used.

Sample size calculation should follow the CLSI EP28-A3c recommendation of at least 120 individuals per partition. A study establishing age specific reference intervals for amino acids and acylcarnitines in dried blood spots by tandem mass spectrometry enrolled 480 apparently healthy children and subdivided them into four age groups of 120 participants each. The sample size was calculated according to CLSI approved guidelines.

Recruitment of Reference Individuals

Recruitment is often the most difficult part of a reference interval study. Healthy volunteers may be recruited from the community, from blood donors, or from staff and students. The recruitment process must include a health questionnaire to apply a priori exclusion criteria.

The CALIPER initiative provides a model for large scale recruitment. CALIPER launched an outreach campaign in 2008 to recruit healthy children and adolescents from the community. The campaign involved schools, community centers, and other public venues, and it applied rigorous exclusion criteria to ensure the health status of the reference sample.

For smaller studies, recruitment may be limited to available volunteers. A study establishing a reference interval for calculated globulin on the Roche platform used 310 highly selected adults from primary care, ranging in age from 16 to 89 years. The study strictly followed the CLSI EP28-A3c guideline for direct reference interval studies.

Sample Collection and Handling

Preanalytical conditions must be standardized to minimize variation. This includes the time of day for sample collection, the fasting status of the reference individuals, the type of collection tube, and the time between collection and analysis. The World Health Organization Laboratory Quality Management System Handbook emphasizes the importance of standardized preanalytical procedures in producing reliable laboratory results.

For some analytes, posture, tourniquet time, and recent exercise can affect results. The study protocol should specify these conditions and document any deviations. The Croatian newborn study used the direct a posteriori sampling method and analyzed samples on the Beckman Coulter AU680 biochemical analyzer under standardized conditions.

Statistical Analysis

The statistical analysis of reference interval data involves several steps. First, the distribution of the data must be examined. Many clinical chemistry analytes have skewed distributions, particularly tumor markers and enzymes. The CA72-4 study found that values were markedly right skewed, which is typical for tumor markers.

Outlier detection is the next step. The CA72-4 study used Tukey's 1.5 times interquartile range rule to remove high outliers, excluding 258 outliers or 9.9 percent of the initial sample. Other methods include the Dixon test and visual inspection of histograms or box plots.

The reference interval is then calculated using nonparametric methods. The 2.5th and 97.5th percentiles are estimated from the ranked data. Confidence intervals for the reference limits can be obtained by bootstrap resampling. The CA72-4 study used 2000 bootstrap resamples to obtain 90 percent confidence intervals for the upper reference limit.

A study establishing trimester specific reference intervals for TSH and thyroid hormones in pregnant women in Oran, Western Algeria followed the CLSI EP28-A3c guideline and included 401 apparently healthy pregnant women. The reference intervals were derived from the 2.5th and 97.5th percentiles for each trimester.

Validation of the Established Interval

After establishing a reference interval, the laboratory should validate its performance in an independent sample. The CA72-4 study evaluated the established interval in an independent set of 206 adults by calculating the exceedance rate, with less than 5 percent of results falling outside the interval considered acceptable. The validation set showed an exceedance rate of 4.85 percent, meeting the preset criterion.

Verifying Reference Intervals

Most clinical laboratories do not have the resources to establish their own reference intervals for every analyte. The alternative is to verify intervals established by external sources, such as manufacturers, published literature, or harmonized multicenter studies. Verification is less resource intensive but still requires careful attention to the comparability of the source population and the local population.

The 20 Sample Verification Procedure

The CLSI EP28-A3c verification procedure requires a minimum of 20 samples from healthy individuals from the local population. The samples are analyzed using the laboratory's routine methods, and the results are compared with the candidate reference interval. If no more than two results fall outside the interval, the interval is considered verified.

A study on verification of harmonized reference intervals in Croatia included 100 apparently healthy adults selected using both a priori and a posteriori exclusion criteria following CLSI EP28-A3c guidelines. Reference intervals were considered verified if at least 90 percent of results, or 18 of 20, fell within the predefined intervals. All tested hematology and coagulation reference intervals were successfully verified, and 25 of 27 biochemical analytes were verified in the first sample set. Total calcium and alkaline phosphatase required additional verification.

Transference of Reference Intervals

Transference is the process of adopting a reference interval established for one method or population and applying it to another method or population. Transference may be appropriate when the analytical methods are comparable and the populations are similar. However, transference carries risks that must be evaluated.

A study on reference interval transference via linear regression examined the conditions under which transference is appropriate. The study established reference intervals for 27 analytes on Roche and Beckman systems and converted Roche intervals to Beckman intervals using linear regression. The concordance rates between transferred and measured reference intervals varied depending on the method used and the number of test samples. The study concluded that transferability is affected by many factors, including correlation, test number, regression equation type, and quality requirements.

The CALIPER reference intervals have been transferred to multiple analytical platforms. A study on CLSI based transference of CALIPER pediatric reference intervals to Beckman Coulter AU biochemical assays examined whether the CALIPER intervals could be applied to this platform. A Croatian study verified CALIPER reference intervals for 19 biochemical assays in newborns and found that 14 of 19 intervals were adopted for use after the first set of measurements. Additional samples were tested for five analytes, and new reference intervals were determined for potassium, magnesium, and direct bilirubin because verification remained unsatisfactory.

Verification in Special Populations

Pediatric and geriatric populations present particular challenges for reference interval verification. The difficulty of recruiting healthy children and older adults often makes the 20 sample verification procedure impractical. Indirect methods using routine patient data may be more feasible in these populations.

A study on verification of reference intervals in routine clinical laboratories noted that pediatric and geriatric age groups continue to pose additional challenges in respect of acquiring and verifying reference intervals. The study recommended practical approaches for routine implementation, including the use of data mining techniques when direct sampling is not feasible.

The CALIPER initiative addressed the pediatric reference interval gap by establishing comprehensive age and sex specific intervals for a wide range of analytes. The initiative has made significant strides towards improving pediatric healthcare in Canada and globally, providing a resource for clinical laboratory specialists, clinicians, and other healthcare workers.

Indirect Methods for Reference Interval Estimation

Indirect methods use large volumes of routine patient data to estimate reference intervals. These methods are based on the assumption that the majority of patient results come from individuals without the disease or condition of interest, and that the central portion of the distribution approximates the healthy reference distribution.

Data Mining Approaches

Data mining techniques can be used to verify reference intervals established by direct methods. A study on verification of reference intervals in routine clinical laboratories described data mining techniques using large amounts of patient test results to verify reference intervals, considering both the laboratory method and local population.

The calculated globulin study compared direct and indirect approaches. The reference interval established using direct sampling techniques was 23 to 35 g/L, while the interval established using indirect sampling techniques in a much larger unselected reference group of 8466 patients was 22 to 37 g/L. The study noted that the range was a little broader but not materially altered when strict exclusion criteria were removed and when using a data mining approach that resulted in a 27 fold larger reference group.

Software Tools for Indirect Methods

Several software tools have been developed to facilitate indirect reference interval estimation. The ReferenceRangeR tool supports five indirect methods for reference interval estimation, including refineR, TMC, TML, kosmic, and reflimR. The tool can include up to 200,000 laboratory test results through a copy and paste input and provides recommendations for sex based stratification by performing statistical analysis. A drift detection algorithm analyzes whether age based stratification is necessary.

The VeRUS method offers an alternative approach to indirect verification. VeRUS compares the candidate reference interval to an interval estimated from local routine data, with acceptable differences based on the sampling uncertainty intrinsic to the nonparametric method for establishing reference intervals with 120 samples. A simulation study comparing VeRUS with the binomial test and equivalence limits found that the binomial test was inherently unable to reject intervals that were too wide, while VeRUS demonstrated robust performance without the need for sample collection.

Limitations of Indirect Methods

Indirect methods have limitations that laboratories must understand. The patient population may include a substantial proportion of individuals with the disease or condition of interest, which can bias the estimated reference interval. The methods also require statistical expertise and careful interpretation of the underlying data distribution.

The calculated globulin study noted that the indirect approach used a much larger reference group but produced a slightly broader interval. The study concluded that the direct method, following CLSI EP28-A3c, remains the standard for establishing reference intervals, while indirect methods may be useful for verification or when direct sampling is impractical.

Partitioning by Age and Sex

Partitioning decisions have a direct impact on the clinical utility of reference intervals. Separate intervals for different age and sex groups can improve diagnostic accuracy but also increase the complexity of reporting and the risk of errors.

Age Partitioning

Age partitioning is essential for analytes that change with age, particularly in pediatrics and geriatrics. The high sensitivity cardiac troponin T study in children found significant differences across age groups, with the highest levels in neonates and a gradual decrease to adult reference ranges by one year of age. The upper limits of the reference intervals ranged from 96.6 ng/L in the 2 to 28 day group to 7.9 ng/L in the 6 to 14 year group.

The amino acid and acylcarnitine study in dried blood spots established age specific reference intervals for four age groups from birth to 12 years. The study emphasized that interpretation of extended newborn screening results should be based on age specific cutoffs established by the laboratory for primary analyte concentration and secondary analyte concentration ratios.

Sex Partitioning

Sex partitioning is required for analytes that differ between males and females. The neutrophil to lymphocyte ratio, lymphocyte to monocyte ratio, and platelet to lymphocyte ratio study found significant differences between sexes for all three ratios. The study established separate reference intervals by sex and age group, noting that the reference upper limits changed with age in both sexes.

The CA72-4 study found no significant sex or age related differences and adopted a combined reference interval. This finding illustrates that partitioning decisions must be based on data from the local population instead of assumptions from other studies.

Statistical Tests for Partitioning

The Harris and Boyd test is the standard method for assessing the need for partitioning. The test compares the means and standard deviations of candidate subgroups and determines whether the differences are large enough to warrant separate intervals. The CA72-4 study used the Harris and Boyd Z test and found no significant differences, supporting the use of a combined interval.

The ReferenceRangeR tool provides recommendations for sex based stratification by performing statistical analysis. The tool also includes a drift detection algorithm to analyze whether age based stratification is necessary.

Method Specificity and Platform Differences

Reference intervals are method specific. Different analytical platforms may produce systematically different results for the same analyte, even when the same principle of measurement is used. This method specificity has important implications for transference and verification.

Platform Differences in Free Light Chains

A study on serum free light chain immunoglobulins found that reference intervals and diagnostic ranges vary by instrument platform. The study investigated the transference of manufacturer reported reference intervals for kappa and lambda free light chains and established de novo intervals on four instruments at three academic medical centers. Three of four instrument platforms did not exhibit acceptable transference of the manufacturer reported kappa free light chain reference interval. The manufacturer reported diagnostic range did not encompass all values observed in reference sera for any of the four platforms evaluated.

The study concluded that transference of manufacturer reported free light chain reference intervals may be inappropriate for select instrument platforms. De novo establishment of free light chain reference intervals specific to instrument platform is highly recommended to assure correct patient result classification.

Electrode Type Differences in Electroretinography

Method specificity extends beyond clinical chemistry to other diagnostic fields. A study on ISCEV standard full field electroretinogram reference limits found clinically significant amplitude differences between electrode types. Silver thread electroretinograms were 55 to 65 percent of the amplitude of gold foil electroretinograms, and skin electroretinograms were 35 to 38 percent of the amplitude of silver thread electroretinograms. The study used a linear model to transform gold foil reference data for inclusion in the silver thread reference sample.

Implications for Transference

The method specificity of reference intervals means that transference requires careful evaluation. The linear regression transference study found that concordance rates between transferred and measured reference intervals varied depending on the method used. For most analytes, accurate results could be obtained when the correlation coefficient was greater than 0.800 and the test number was sufficient, regardless of the regression equation type.

The study concluded that to reduce the risk of transference, it is very important to select the right method with reasonable conditions. Laboratories should evaluate the correlation between their method and the source method before transferring a reference interval.

Quality Control and Documentation

Reference intervals must be managed within a broader quality management system. The World Health Organization Laboratory Quality Management System Handbook provides guidance on the quality framework for laboratory operations, including the management of reference intervals.

Ongoing Verification

Reference intervals should be reviewed periodically to ensure they remain appropriate for the local population and the current analytical method. Changes in analytical methods, reagents, or calibrators may require reverification of reference intervals. Changes in the population, such as demographic shifts or changes in health status, may also affect the validity of existing intervals.

A study on verification of harmonized reference intervals in Croatia noted that notable methodology and population characteristics changes have occurred since the original harmonized reference intervals were published in 2004. The study concluded that periodic re evaluation is needed due to changes in analytical methods and population characteristics.

Documentation Requirements

Laboratories must document the source of each reference interval, the method of establishment or verification, the date of verification, and the individuals responsible. This documentation supports accreditation requirements and provides a basis for troubleshooting when discrepancies arise.

The nationwide survey on knowledge, attitudes, and practices regarding reference interval utilization in clinical laboratories in Nepal found that most laboratories relied on manufacturer provided reference intervals or published literature. Accredited laboratories demonstrated better knowledge of reference intervals and higher confidence in using current intervals. The survey highlighted the need for training and standardization in reference interval management.

Autoverification Systems

Autoverification systems can incorporate reference intervals into automated result validation. A study on an autoverification system for thyroid function profiles used quality control checks, instrument error flags, limit range rules, delta check rules, and logical rules. The system achieved an overall autoverification pass rate of 75.2 percent, which increased to 77.8 percent after optimization. The median laboratory turnaround time decreased from 122.1 minutes to 88.6 minutes.

The study noted that reference change value based delta checks require prior results and are therefore not applicable to new patients. Other autoverification rules remain active, and only results that cannot be autoverified are routed to manual review.

Common Failure Patterns

Laboratories encounter several recurring problems in reference interval establishment and verification. Understanding these failure patterns can help laboratories avoid common errors.

Inadequate Sample Size

The most common failure in reference interval establishment is the use of too few reference individuals. A sample size below 120 per partition produces imprecise estimates of the reference limits, particularly for analytes with skewed distributions. The CALIPER white paper noted that pediatric reference intervals have often been established with a small sample size, inpatient or outpatient samples, outdated methodologies, or inappropriate statistical procedures.

Inappropriate Reference Individual Selection

The selection of reference individuals is critical to the validity of the reference interval. Including individuals with subclinical disease, medication use, or other factors that affect the analyte can bias the interval. The calculated globulin study used highly selected adults from primary care and applied strict exclusion criteria to ensure the health status of the reference sample.

Failure to Partition Appropriately

The failure to partition by age or sex when significant differences exist can produce reference intervals that are too wide for some subgroups and too narrow for others. The high sensitivity cardiac troponin T study in children demonstrated substantial age related differences that would be obscured by a single pediatric interval.

Uncritical Transference

The uncritical adoption of manufacturer or published reference intervals without verification is a common failure. The free light chain study found that transference of manufacturer reported reference intervals was inappropriate for select instrument platforms. The nationwide survey in Nepal found that most laboratories relied on manufacturer provided reference intervals without local validation.

Ignoring Method Changes

Changes in analytical methods, reagents, or calibrators can shift results and invalidate existing reference intervals. Laboratories must reverify reference intervals after any significant method change. The Croatian harmonized reference interval study emphasized the need for periodic re evaluation due to changes in analytical methods.

Limitations and Professional Escalation

Reference intervals have inherent limitations that laboratories and clinicians must understand. A reference interval describes the central 95 percent of results in a healthy population, which means that 5 percent of healthy individuals will have results outside the interval. This is a statistical property, not an indication of disease.

Limitations of Reference Intervals

Reference intervals do not define health or disease. A result outside the reference interval does not necessarily indicate disease, and a result within the interval does not exclude disease. The clinical context, including symptoms, signs, and other laboratory findings, must be considered in interpretation.

The CALIPER white paper noted that clinical interpretation of laboratory test results relies heavily on the availability of appropriate population based reference intervals or decision limits developed through clinical outcome studies. The distinction between reference intervals and decision limits is important for analytes such as cardiac troponin, where the 99th percentile upper reference limit is used as a decision limit.

Professional Escalation Criteria

Laboratories should have procedures for escalating concerns about reference intervals. These include situations where verification fails repeatedly, where results suggest a method problem, or where clinical feedback indicates that an interval is inappropriate.

The Croatian newborn study provides an example of escalation in practice. After the first set of measurements, 14 of 19 tested reference intervals were adopted for use. A second set of samples was tested for five analytes, and new reference intervals were determined for potassium, magnesium, and direct bilirubin because verification remained unsatisfactory.

Laboratories should also escalate concerns when reference interval verification reveals systematic differences between the local population and the source population. This may indicate that the source interval is not transferable and that de novo establishment is required.

Safety and Regulatory Context

Reference interval management operates within a broader regulatory and safety framework. The World Health Organization Laboratory Quality Management System Handbook addresses the quality requirements for medical laboratories, including the management of reference intervals. The World Health Organization Laboratory Biosafety Manual provides guidance on the safe handling of biological samples, which is relevant to reference interval studies that involve collection and analysis of human samples.

Biosafety Considerations

Reference interval studies involve the collection and handling of human blood or other biological samples. Laboratories must follow biosafety procedures to protect staff and prevent contamination. The World Health Organization Laboratory Biosafety Manual provides guidance on safe handling practices, personal protective equipment, and waste management.

Regulatory Requirements

Accreditation standards, such as ISO 15189, require laboratories to have documented procedures for reference interval management. The nationwide survey in Nepal found that accredited laboratories demonstrated better knowledge of reference intervals and higher confidence in using current intervals. The survey highlighted the need for training and standardization in reference interval management.

The U.S. Food and Drug Administration Bioanalytical Method Validation Guidance addresses the validation of analytical methods used in clinical studies. While this guidance is primarily focused on drug development, it emphasizes the importance of method validation and the need for documented procedures.

Practical Implementation Steps

The following steps provide a practical framework for implementing reference interval establishment or verification in a clinical laboratory.

Step 1: Define the Scope

Determine which analytes require reference interval establishment or verification. Prioritize analytes with new methods, changed methods, or no existing interval. Consider the clinical impact of incorrect intervals for each analyte.

Step 2: Select the Approach

Decide whether to establish a new interval or verify an existing interval. Consider the availability of a suitable source interval, the resources for recruiting reference individuals, and the characteristics of the local population.

Step 3: Plan the Study

Develop a study protocol that specifies the inclusion and exclusion criteria, the sample size, the sample collection procedures, and the statistical methods. For verification studies, confirm that the source interval was established using appropriate methods and that the source population is comparable to the local population.

Step 4: Recruit Reference Individuals

Recruit healthy volunteers who meet the inclusion criteria and do not meet any exclusion criteria. Document the recruitment process and the health screening results. For pediatric studies, consider using established pediatric reference intervals such as CALIPER when local recruitment is not feasible.

Step 5: Collect and Analyze Samples

Collect samples under standardized preanalytical conditions. Analyze the samples using the laboratory's routine methods. Document any deviations from the protocol.

Step 6: Analyze the Data

For establishment studies, examine the data distribution, detect and remove outliers, and calculate the reference interval using nonparametric methods. For verification studies, compare the results with the candidate interval and apply the 20 sample rule.

Step 7: Document and Review

Document the source of the interval, the method of establishment or verification, the date, and the responsible individuals. Review the interval periodically and after any significant method change.

Records and Measurements

Laboratories must maintain records of reference interval activities. These records support accreditation, troubleshooting, and continuous improvement.

Essential Records

The following records should be maintained for each reference interval:

Measurement of Verification Success

The success of a verification study is measured by the proportion of reference individual results that fall within the candidate interval. The CLSI EP28-A3c criterion is that no more than 2 of 20 results may fall outside the interval. The Croatian harmonized reference interval study used the criterion that at least 90 percent of results, or 18 of 20, must fall within the predefined intervals.

Frequently Asked Questions

What is the difference between a reference interval and a reference range?

The terms reference interval and reference range are often used interchangeably, but reference interval is the preferred term in CLSI EP28-A3c. A reference interval is the central 95 percent range of results observed in a healthy reference population, typically defined by the 2.5th and 97.5th percentiles. The term reference range can be misleading because it suggests a range of acceptable values, while a reference interval is a statistical description of healthy population results.

How many samples are needed to establish a reference interval?

CLSI EP28-A3c recommends a minimum of 120 reference individuals per partition for de novo establishment. This sample size allows nonparametric estimation of the 2.5th and 97.5th percentiles with acceptable precision. For verification of an existing interval, a minimum of 20 samples per partition is recommended. The amino acid and acylcarnitine study in dried blood spots used 120 participants per age group, calculated according to CLSI approved guidelines.

When should a laboratory verify instead of establish a reference interval?

A laboratory should verify an existing reference interval when a suitable source interval is available and the local population and analytical method are comparable to the source. Verification requires only 20 samples per partition and is less resource intensive than establishment. A laboratory should establish its own reference interval when no suitable source interval exists, when the analytical method differs substantially from the source method, or when verification fails repeatedly.

What is reference interval transference?

Reference interval transference is the process of adopting a reference interval established for one method or population and applying it to another method or population. Transference may be appropriate when the analytical methods are comparable and the populations are similar. However, transference carries risks that must be evaluated. The linear regression transference study found that transferability is affected by many factors, including correlation, test number, regression equation type, and quality requirements.

Why do reference intervals differ between analytical platforms?

Reference intervals are method specific because different analytical platforms may produce systematically different results for the same analyte. The free light chain study found that reference intervals and diagnostic ranges vary by instrument platform, and transference of manufacturer reported intervals was inappropriate for select platforms. Laboratories must verify intervals on their own platform before adoption.

How are pediatric reference intervals established?

Pediatric reference intervals require age and sex specific partitioning because many analytes change with age during childhood. The CALIPER initiative established comprehensive pediatric reference intervals using samples from healthy children and adolescents. When local establishment is not feasible, laboratories may verify CALIPER intervals for their platform. The Croatian newborn study verified CALIPER intervals for 19 biochemical assays and adopted 14 of 19 intervals after the first set of measurements.

What are indirect methods for reference interval estimation?

Indirect methods use large volumes of routine patient data to estimate reference intervals. These methods assume that the central portion of the patient distribution approximates the healthy reference distribution. Software tools such as ReferenceRangeR support multiple indirect methods and can include up to 200,000 test results. The VeRUS method compares a candidate interval to an interval estimated from local routine data, with acceptable differences based on sampling uncertainty.

How often should reference intervals be reviewed?

Reference intervals should be reviewed periodically and after any significant method change. Changes in analytical methods, reagents, or calibrators may require reverification. Changes in the population, such as demographic shifts or changes in health status, may also affect the validity of existing intervals. The Croatian harmonized reference interval study concluded that periodic re evaluation is needed due to changes in analytical methods and population characteristics.

Related Diagnostic Guides

• Establishing Reference Intervals in Veterinary Clinical Chemistry
• Procedure for Quality Control: Step-by-Step Implementation in a Molecular Lab
• Spread Plate Method: Advantages, Disadvantages, and Step-by-Step Protocol
• BCA Assay Protocol: Principles and Step-by-Step Instructions
• Bradford Assay Protocol: Step-by-Step for Protein Quantification

References and Further Reading

• Laboratory Quality Management System Handbook. World Health Organization.
• Laboratory Biosafety Manual. World Health Organization.
• Assay Guidance Manual. National Center for Advancing Translational Sciences.
• Bioanalytical Method Validation Guidance. U.S. Food and Drug Administration.
• NCBI Literature Resources. National Center for Biotechnology Information.
• The Canadian laboratory initiative on pediatric reference intervals: A CALIPER white paper.. Critical reviews in clinical laboratory sciences, 2017.
• Establishment of a reference interval for calculated globulin on the Roche platform.. Annals of clinical biochemistry, 2026.
• Verification of reference intervals in routine clinical laboratories: practical challenges and recommendations.. Clinical chemistry and laboratory medicine, 2018.
• Establishment of the reference interval for high-sensitivity cardiac troponin T in healthy children of Chongqing Nan'an district.. Scandinavian journal of clinical and laboratory investigation, 2021.
• Distribution and reference interval establishment of neutral-to-lymphocyte ratio (NLR), lymphocyte-to-monocyte ratio (LMR), and platelet-to-lymphocyte ratio (PLR) in Chinese healthy adults.. Journal of clinical laboratory analysis, 2021.
• Establishment of Reference Interval for CA72-4 in Healthy Adults in Shenzhen, China.. Journal of clinical laboratory analysis, 2026.
• CLSI-based verification and de novo establishment of reference intervals for common biochemical assays in Croatian newborns.. Biochemia medica, 2024.
• Establishment of Age Specific Reference Interval for Aminoacids and Acylcarnitine in Dried Blood Spot by Tandem Mass Spectrometry.. Indian journal of clinical biochemistry : IJCB, 2024.
• Verification of harmonized reference intervals in Croatia: is it time for a change?. 2026.
• ReferenceRangeR: a novel tool designed to facilitate reference interval estimation and verification.. 2026.
• VeRUS: verification of reference intervals based on the uncertainty of sampling.. 2026.
• Designing and Evaluating an Autoverification RCV-Based System for Thyroid Function Profiles.. 2026.
• Nationwide Survey on Knowledge, Attitudes, and Practices regarding Reference Interval Utilization in Clinical Laboratories in Nepal.. 2026.
• Establishment of Trimester-Specific Reference Intervals for TSH and Thyroid Hormones in Pregnant Women living in Oran, Western Algeria.. 2026.
• ISCEV standard full-field ERG reference limits from 407 healthy subjects, derived from transference and validation of reference data between electrode types and centres. Documenta Ophthalmologica, 2025.
• A study on reference interval transference via linear regression. Clinical Chemistry and Laboratory Medicine, 2019.
• Alkaline phosphatase : reference interval transference from CALIPER to a pediatric Brazilian population. 2018.
• Shifting-reference concentration cells to refine composition-dependent transport characterization of binary lithium-ion electrolytes. 2020.
• Reference intervals and diagnostic ranges for serum free κ and free λ immunoglobulin light chains vary by instrument platform: Implications for classification of patient results in a multi-center study.. Clinical Biochemistry, 2018.
• A long-range generalized predictive control algorithm for a DFIG based wind energy system. IEEE/CAA Journal of Automatica Sinica, 2019.
• Esprit de corps, Work Transference and Dissolution: Lacan as an Organisational Theorist. 2016.
• Characterizing Thermodynamic Properties in Concentrated Binary Electrolytes: Combined Differential Reference Concentration Cell and Hittorf Methods. ECS Meeting Abstracts, 2019.
• CLSI-based transference of CALIPER pediatric reference intervals to Beckman Coulter AU biochemical assays. Clinical Biochemistry, 2015.
• Establishment of SARS-CoV-2 Immunoglobulins (IgM, IgG) Reference Intervals for Elder Population in China based on 3,733 Samples. Clinical Laboratory, 2022.
• Establishment of DBS-based reference intervals for vitamins, essential elements, and potentially toxic elements in Chinese adults from Hubei Province. Clinical Chemistry and Laboratory Medicine, 2026.

This article is educational and does not replace validated laboratory procedures, institutional biosafety review, manufacturer instructions, or professional interpretation.