Reference Intervals in Veterinary Clinical Pathology: Establishment and Use

By Dr. Zubair Khalid, DVM, MS, PhD · 2026-08-11

Key Takeaways

• Reference intervals represent the central 95% of values from a defined healthy population, meaning 5% of healthy individuals will fall outside the range for any single analyte, necessitating cautious interpretation of isolated out-of-range results.
• Establishing de novo nonparametric reference intervals requires a minimum of 120 carefully selected healthy individuals, with explicit inclusion/exclusion criteria and rigorous quality control, as per ASVCP and CLSI guidelines.
• Partitioning reference intervals by biologically relevant subgroups such as age, sex, breed, or species is critical when significant variation exists, as demonstrated by age-related changes in Holstein calves or sex-specific differences in Sprague-Dawley rats.
• Transference of existing reference intervals from other laboratories or published sources is permissible but mandates formal validation using a smaller local population (typically 20-60 animals) to confirm applicability to the local analyzer, reagents, and patient demographic.
• Decision limits, which distinguish diseased from non-diseased animals or trigger interventions, are distinct from reference intervals and are derived from outcome data or receiver operating characteristic analysis, not healthy population statistics.
• Subject-based reference intervals, comparing a patient's current results to their own historical values, are valuable for monitoring chronic conditions or therapeutic responses when intra-individual variation is low relative to inter-individual variation.

Reference intervals are the interpretive backbone of clinical pathology. Every complete blood count, serum biochemistry panel, and endocrine assay is judged against a set of values derived from a defined population of healthy animals. When those values are flawed, the diagnostic conclusions drawn from them are flawed as well. This article explains what reference intervals are, how they are constructed, and how the practicing veterinarian should interpret patient results in relation to them. It serves the clinician who must decide whether a laboratory value represents health, disease, or an artifact of the reference population itself.

The central question this article addresses is deceptively simple: what does it mean when a patient's result falls outside the reported range? The answer depends on how that range was established, whether it applies to the patient in front of you, and what clinical consequence should follow from the finding. A reference interval is not a biological truth. It is a statistical description of a particular group of animals measured under particular conditions at a particular time. Understanding its limitations is as important as understanding its construction.

At a Glance

The Conceptual Foundation of Reference Intervals

A reference interval describes the dispersion of a variable in a healthy population. The standard approach defines the interval as the range containing 95% of reference values, leaving 2.5% of healthy individuals above and 2.5% below the limits. This means that by definition, one in twenty healthy animals will have a result outside the interval for any single analyte. When multiple analytes are measured on a single panel, the probability that a healthy animal has at least one abnormal result rises substantially. This statistical reality underpins the clinical rule that an isolated out-of-range value, particularly one just outside the limit, warrants cautious interpretation instead of reflexive action.

The international recommendations for establishing reference intervals are published by the Clinical and Laboratory Standards Institute (CLSI) and adapted for veterinary species by the American Society for Veterinary Clinical Pathology (ASVCP). The preferred method is a priori nonparametric determination from at least 120 carefully selected reference individuals. This approach requires that the reference population be defined by explicit inclusion and exclusion criteria before sampling begins, and that all analytical procedures be subject to quality control. When fewer than 120 animals are available, robust statistical methods can estimate intervals, but the resulting limits may be highly imprecise and should be interpreted with corresponding caution ASVCP reference interval guidelines.

The selection of reference individuals is the most critical step in the entire process. Animals must be healthy by history, physical examination, and ideally by ancillary testing. Exclusion criteria should address recent medication, subclinical disease, pregnancy, lactation, and any factor known to influence the analyte of interest. The reference population must also match the patients to whom the interval will be applied. A reference interval derived from adult working dogs cannot be applied to puppies, and an interval derived from one breed may not fit another Reference values: a review.

Partitioning by Biologically Relevant Subgroups

Reference values are not uniform across all members of a species. Age, sex, reproductive status, and breed can each exert substantial effects on hematologic and biochemical analytes. The decision to partition a reference interval into subgroups should be based on demonstrated biological differences instead of convenience.

Age is a particularly important partitioning variable in growing animals. Studies in Holstein dairy calves have shown significant age-related changes in hemoglobin, mean corpuscular volume, serum total protein, globulin, aspartate aminotransferase, and alkaline phosphatase during the first three months of life. Neutrophil numbers and glucose concentrations differ markedly within the first 24 to 48 hours after birth. Applying adult reference intervals to neonatal or juvenile animals produces systematic misinterpretation of laboratory results age related changes in Holstein dairy calves.

Sex-based partitioning is equally important in many species. Work in Sprague-Dawley rats demonstrated that most hematologic and biochemical analytes are significantly influenced by sex, with males showing higher hemoglobin, hematocrit, and red blood cell counts than females. The ASVCP guidelines recommend that partitioning be considered whenever the proportion of reference values falling outside a combined interval exceeds a defined threshold, typically 1% or more of the reference population sex-specific reference intervals in Sprague-Dawley rats.

Strain and breed differences further complicate the picture. Studies in laboratory mice have documented inter-strain variation in glucose, total bilirubin, potassium, calcium, and phosphate across C57BL/6J, 129SV/EV, and C3H/HeJ strains. These differences are large enough that therapeutic monitoring in one strain cannot rely on reference values derived from another age-related reference intervals in mouse strains. The same principle applies in veterinary practice: breed-specific intervals exist for many canine analytes, and their use is strongly encouraged when available.

Transference and Validation of Existing Intervals

Collecting 120 healthy animals for every analyte, species, and subgroup is often impractical in clinical practice. The ASVCP guidelines therefore describe alternative approaches. Transference involves adopting a reference interval established elsewhere, such as from a published study, a commercial laboratory, or an instrument manufacturer. Validation requires that the transferred interval be checked against a smaller number of reference individuals from the local population, typically 20 to 60 animals, to confirm that the interval remains appropriate for the new setting ASVCP reference interval guidelines.

Validation is not optional when an interval is transferred between laboratories. Differences in analyzer platforms, reagents, calibration, and sample handling can shift results systematically. A reference interval that fits one laboratory perfectly may misclassify patients at another. The validation process should include both a comparison of reference values against the transferred limits and an assessment of whether the proportion of out-of-range results in the local reference population is acceptable. If more than a small percentage of healthy local animals fall outside the transferred interval, the interval should be recalculated or a different source should be sought Reference values: a review.

Reference Intervals Versus Decision Limits

A reference interval answers a descriptive question: what values occur in healthy animals? A decision limit answers a clinical question: what value distinguishes diseased from non-diseased animals, or what value triggers a specific intervention? These are fundamentally different constructs. Decision limits are derived from outcome data, receiver operating characteriztic analysis, or expert consensus, and they may fall inside, outside, or at the boundary of the reference interval. The ASVCP guidelines explicitly distinguish decision limits from reference intervals and recommend that clinicians use decision limits when they exist for a particular analyte and clinical context ASVCP reference interval guidelines.

Subject-Based Reference Intervals

Population-based reference intervals have a well-documented limitation: they may be insensitive to meaningful changes within an individual animal. When intra-individual variation is small relative to inter-individual variation, a patient can move from its own homeostatic set point to a clearly abnormal value while remaining within the population interval. Subject-based reference intervals, sometimes called individual reference intervals, address this problem by comparing a patient's current result to its own previous values. This approach requires serial data from the individual and is most useful for monitoring chronic disease or therapeutic response. The ASVCP guidelines include recommendations for the use and interpretation of subject-based intervals as an alternative to population-based intervals when sample size is limited or inter-individual variation is high ASVCP reference interval guidelines.

Species-Specific Considerations

Reference intervals are species-specific by necessity, but the degree of specificity required extends beyond the species level. Domestic species present particular challenges because of the wide range of breeds, body sizes, and management systems encountered in practice. Production animals are further complicated by the influence of production stage, nutrition, and environmental conditions on many analytes. Laboratory animals used in toxicology require strain-specific and sex-specific intervals to detect treatment effects reliably sex-specific reference intervals in Sprague-Dawley rats. The same logic applies to exotic, avian, and wildlife species, where reference data are often sparse and the clinician must rely on published intervals from closely related species with explicit acknowledgment of the associated uncertainty.

Alternative sample matrices introduce additional interpretive challenges. Salivary hormone analysis, for example, offers a noninvasive and stress-free alternative to plasma and serum, but the establishment of defined reference intervals for salivary analytes remains incomplete. The mode of hormone entry into saliva, the collection method, and the analytical platform all influence results, and standardized reference intervals are needed before salivary testing can be interpreted with the same confidence as blood-based testing current status of salivary hormone analysis.

Verifying a Reference Interval Before Clinical Use

Before a laboratory result is interpreted against a published or transferred interval, the interval must be verified for the local setting. Verification confirms that the preanalytical methods, analytical platform, and patient population of the receiving laboratory are sufficiently similar to those of the source laboratory. The ASVCP reference interval guidelines recommend a structured verification process that begins with a documented comparison of the two laboratories' methods, including sample type, anticoagulant, analyzer, reagent lot, and calibration protocols. If any of these differ materially, the interval should not be adopted without a full de novo study.

The practical verification procedure uses a small number of reference individuals, typically 20, sampled from the local healthy population. Each value is compared against the proposed interval. If no more than 2 of 20 values fall outside the interval, the interval is accepted for local use. If 3 or more values fall outside, the interval is rejected and the laboratory must either establish its own interval or investigate whether the discrepancy reflects a true population difference, an analytical shift, or a preanalytical variation. This rule is derived from the binomial distribution and is stated in the ASVCP consensus guidelines. The verification sample must be drawn from healthy individuals meeting the same inclusion and exclusion criteria used by the source laboratory, and the samples must be handled identically to routine patient samples.

A common failure mode in verification is the use of convenience samples, such as blood donors or presurgical screening patients, without confirming health status. Subclinical disease, recent medication, or physiologic stress can shift values and produce false verification failures. Conversely, a narrowly selected verification group, such as young adult males only, can mask a partition problem that will surface when the interval is applied to the full hospital population. The verification group should reflect the demographic range of the patients for whom the interval will be used.

Factors That Alter Reference Values

Reference values are not fixed biological constants. They shift with intrinsic and extrinsic factors, and the clinician must know which factors apply to the patient at hand. The table below summarizes the major categories and their typical effects.

Age is among the most consequential factors in growing animals. Studies in Holstein dairy calves demonstrate that hemoglobin, MCV, MCH, MCHC, inorganic phosphorus, total protein, globulin, AST, and ALP change significantly during the first three months of life, and neutrophil numbers and glucose differ within the first 24 to 48 hours after birth. The authors of that work concluded that age-specific reference values are required for precise interpretation of laboratory results in calves. Similar age dependence is documented in laboratory mice, where biochemical and hematologic parameters vary by strain and by age range, with differences observed between 1 to 2 months, 3 to 8 months, and 9 to 12 months of age. The mouse strain reference interval study also reported sex differences for glucose, LDH, cholesterol, and BUN, reinforcing that both age and sex partitions may be needed within a single species.

Sex partitioning is not limited to laboratory rodents. In Sprague-Dawley rats used in toxicology studies, most hematologic and biochemical analytes were significantly influenced by sex, with males showing higher hemoglobin, hematocrit, and red blood cell counts. The rat reference interval study applied the CLSI C28-A3 and ASVCP guidelines to establish sex-specific intervals from 500 healthy animals. For the practitioner, the lesson is that a single unpartitioned interval can misclassify a substantial proportion of healthy patients when sex or age effects are large.

Interpreting a Single Patient Result

When a patient result falls outside the reported interval, the first question is not "what disease causes this" but "is this result truly abnormal for this patient." The interval describes the central 95% of a healthy reference population, so by definition 5% of healthy individuals will have a value outside the interval for any single analyte. The review of reference value methodology emphasizes that reference intervals are estimates of the dispersion of variables in healthy individuals, not diagnostic thresholds. A single marginally elevated value in an otherwise healthy animal may represent the tail of the healthy distribution instead of disease.

The second question concerns the magnitude of the deviation. A value that exceeds the upper limit by 1% is biologically different from one that exceeds it by 300%. The interval itself does not convey the clinical significance of the deviation. The practitioner must integrate the magnitude of change, the direction of change, the presence of concurrent abnormalities, and the pretest probability of disease in that patient.

The third question is whether the change is real. Analytical imprecision, sample hemolysis, lipemia, or delayed separation can produce spurious abnormalities. If the result does not fit the clinical picture, repeat testing on a fresh sample is appropriate before pursuing an extensive diagnostic workup.

Monitoring Trends Versus Comparing to an Interval

For serial monitoring, the population-based interval is a blunt instrument. A patient can move from the middle of the interval to the upper edge and still remain "within normal limits," yet that trajectory may be clinically important. Conversely, a patient can oscillate around a single limit without true change. The ASVCP guidelines describe subject-based reference intervals as an alternative when inter-individual variation is high or when the analyte is monitored longitudinally in the same individual. In practice, the clinician should compare serial results to the patient's own baseline and to the magnitude of change expected from biological and analytical variation, instead of relying solely on the population interval.

Documentation of reference interval use should include the source of the interval, the verification date and method, the analyzer and sample type, and any partitions applied. This record supports consistent interpretation across clinicians and provides a basis for periodic re-verification.

Recognized Complications and Failure Modes

Reference intervals fail in predictable ways, and most failures are detectable before they cause clinical harm. The most common failure mode is an interval that does not match the patient population. This occurs when the reference population differed from the patient population in age, breed, sex, or physiologic state. The second most common failure is analytical drift, where the instrument or reagent lot changes shift results systematically while the interval remains fixed. A third failure mode is preanalytical variation, including prolonged sample storage, hemolysis, lipemia, or inconsistent collection technique.

Each failure mode has a characteriztic signature. Population mismatch produces results that cluster near one limit of the interval across many analytes. Analytical drift produces gradual, parallel shifts in control values that may be visible on quality control charts before patient results appear abnormal. Preanalytical error typically affects a single sample or a small batch and often involves analytes known to be unstable, such as potassium, glucose, or enzymes.

Early detection depends on routine quality control and on periodic verification of the interval itself. The ASVCP guidelines recommend that laboratories monitor quality control data continuously and revalidate intervals when instrument platforms change or when the patient population served by the laboratory shifts ASVCP reference interval guidelines. For in-clinic analyzers, the same discipline applies: run controls with each batch, track control values over time, and compare patient results against intervals supplied by the manufacturer only after verifying those intervals against your own patient population.

Common Errors in Interpretation

Less experienced clinicians often treat the reference interval as a binary boundary. A value at 0.1 units above the upper limit is interpreted as disease, while a value at 0.1 units below the limit is interpreted as health. This ignores the statistical construction of the interval. By design, 95% of healthy individuals fall within the interval, which means 5% of healthy individuals fall outside it. A single marginal abnormality, especially in a low-prevalence disease context, is more likely to represent a healthy outlier than disease.

A second common error is interpreting a change within the interval as insignificant. A patient whose creatinine rises from 0.8 to 1.7 mg/dL may remain within the interval yet have lost substantial renal function. Serial monitoring detects trends that single comparisons cannot, and this principle applies across species and analytes Reference values: a review. The corrective action is to compare current results to the patient's own baseline when prior values exist, and to interpret the magnitude and direction of change alongside the interval.

A third error is applying an interval to a patient for whom it was not validated. Age-specific intervals are required for growing animals. Holstein calves show significant age-related changes in hemoglobin, MCV, phosphorus, total protein, globulin, AST, and ALP during the first three months of life, and adult intervals misclassify these values Hematology and serum biochemistry of Holstein dairy calves. Similarly, sex-specific intervals matter in species where partitioning is statistically justified, as demonstrated in Sprague-Dawley rats where most hematologic and biochemical analytes differ by sex Sex-specific reference intervals in Sprague-Dawley rats. The corrective action is to confirm that the interval was derived from or validated for a population matching the patient.

Limitations of the Evidence and Areas of Expert Disagreement

The evidence base for veterinary reference intervals is uneven across species. Companion animal intervals are comparatively well developed, while intervals for exotic species, wildlife, and many production settings rely on small sample sizes. The ASVCP guidelines acknowledge that collecting sufficient reference samples is challenging and provide methods for intervals derived from small numbers of animals, but the resulting limits may be highly imprecise ASVCP reference interval guidelines. Expert opinion differs on how small a sample can be before an interval becomes clinically misleading, and on whether transference from another laboratory is preferable to a locally derived interval with limited sample size.

Salivary hormone analysis illustrates a further limitation. Saliva offers a noninvasive alternative to plasma and serum, but the field lacks standardized analytical tools and defined reference intervals, and round-robin trials are needed before the approach gains broad clinical acceptance Current status of salivary hormone analysis. Practitioners should treat published salivary intervals with caution until local validation is performed.

When to Escalate

Referral or specialist consultation is warranted when a patient's results are persistently outside the interval, when the pattern of abnormalities suggests a disease process that exceeds local diagnostic capacity, or when serial monitoring shows a progressive trend that crosses into abnormal territory. Laboratory involvement is appropriate when quality control data suggest drift, when an interval appears mismatched to the patient population, or when a new analyzer or reagent lot is introduced and interval verification is required. The ASVCP guidelines provide a framework for transference and validation that laboratories can apply in these circumstances ASVCP reference interval guidelines.

Regulatory reporting applies in specific circumstances, including notifiable disease detection and food animal residue concerns. The World Organization for Animal Health maintains international standards for disease surveillance and trade-related reporting, and practitioners should consult those standards when a laboratory result raises a reportable disease question WOAH terrestrial animal health standards. Regional requirements differ, and the responsible approach is to confirm local reporting obligations before acting.

Frequently Asked Questions

How Many Animals Do I Really Need to Establish a Reference Interval?

The internationally preferred standard is at least 120 healthy reference individuals for nonparametric determination of a 95% reference interval. This number allows the 2.5th and 97.5th percentiles to be estimated with acceptable precision. When that many animals are impractical, the ASVCP reference interval guidelines describe robust statistical methods for smaller sample sizes, but reference limits derived from small groups may be highly imprecise. A more practical approach for most practices is transference and validation of an interval from a commercial laboratory or published source, which requires far fewer animals, typically 20 to 40, to confirm that the existing interval fits your local population and analytical system.

Can I Use a Reference Interval From Another Laboratory or Another Instrument?

Yes, but only after formal validation. Transference is acceptable when the analytical method, preanalytical handling, and patient population are comparable to those of the source laboratory. The ASVCP guidelines on transference and validation recommend collecting samples from a small number of healthy animals, typically 20 per subgroup, and comparing their results to the proposed interval. If fewer than 10% of results fall outside the interval, the transferred interval is generally acceptable. If more results fall outside, you must investigate preanalytical differences, instrument calibration, or population characteriztics before rejecting the interval. Document the validation date, the source of the interval, and the results of your comparison in the laboratory quality records.

What Should I Do When Only a Few Healthy Animals Are Available?

Small sample sizes are common in exotic species, wildlife, and rare breeds. When you cannot collect 120 animals, the reference value review by Geffré and colleagues notes that robust estimators and bootstrap methods can generate intervals from as few as 20 to 40 values, though the limits will be imprecise. You should report such intervals with their 90% confidence intervals so users understand the uncertainty. An alternative is to rely on subject-based reference intervals, where serial samples from an individual establish that animal's own baseline, which is particularly useful when inter-individual variation is high. For production species, consider pooling data across cooperating herds or using historical quality control data to supplement the reference sample.

How Do Age and Sex Partitioning Affect My Daily Interpretation?

Partitioning matters most when biological variation is large. For example, age-related changes in Holstein dairy calves show that hemoglobin, MCV, alkaline phosphatase, and globulins change substantially during the first three months of life, so adult intervals are inappropriate for calves. Similarly, sex-specific intervals in Sprague-Dawley rats demonstrate that many hematologic and biochemical analytes differ significantly between males and females. In practice, check whether the laboratory report already applies age and sex partitions. If it does not, interpret results against the partition that best matches your patient, and be cautious when a result falls near a boundary between partitions. For growing animals, serial monitoring is often more informative than a single comparison to an adult interval.

How Should I Document Reference Interval Verification in My Practice Records?

Record the source of the interval, whether it was established de novo, transferred, or validated, and the date of verification. For validated intervals, document the number of reference animals used, the percentage of results falling within the interval, and any outliers detected. Note the analytical platform and reagent lot if relevant. The ASVCP quality assurance and laboratory standards guidance recommends periodic re-verification, particularly after instrument maintenance, reagent changes, or software updates. Keep this documentation accessible to all clinicians who interpret results from your in-house laboratory. If you use a commercial laboratory, retain their current reference interval documentation and note the date you confirmed it matches your report format.

How Do I Explain an Abnormal Result to an Owner When the Interval Seems Unreliable?

Be transparent about what a reference interval does and does not mean. Explain that the interval represents the central 95% of healthy animals, so 1 in 20 healthy animals will fall outside it by chance alone. If the interval was derived from a small or different population, state that limitation directly. Frame the result in context: a single mild elevation in a clinically normal animal may warrant rechecking instead of immediate intervention. The MSD Veterinary Manual emphasizes that laboratory results must be integrated with history, physical examination, and other diagnostics. For clients, use language such as "this value is outside our expected range, but that does not confirm disease" and outline the next diagnostic step. Document your interpretation and the reasoning behind your recommendation in the medical record.

Related Clinical & Scientific Guides

• Peripheral Blood Smear Evaluation: A Step-by-Step Guide
• Reticulocyte Counts in Veterinary Medicine: Clinical Utility and Interpretation
• Cerebrospinal Fluid Analysis in Veterinary Neurology: Collection and Interpretation

References and Further Reading

• Reference values: a review.. 2009.
• Sex-specific reference intervals of hematologic and biochemical analytes in Sprague-Dawley rats using the nonparametric rank percentile method.. 2017.
• ASVCP reference interval guidelines: determination of de novo reference intervals in veterinary species and other related topics.. 2012.
• Current status of salivary hormone analysis.. 2008.
• Age-Related Reference Intervals of the Main Biochemical and Hematological Parameters in C57BL/6J, 129SV/EV and C3H/HeJ Mouse Strains.. 2008.
• Hematology and serum biochemistry of Holstein dairy calves: age related changes and comparison with blood composition in adults.. 2007.
• American Society for Veterinary Clinical Pathology Guidelines. American Society for Veterinary Clinical Pathology.
• MSD Veterinary Manual, Professional Edition. MSD Veterinary Manual.
• American Veterinary Medical Association Practice Resources. American Veterinary Medical Association.

Related Articles

• Reticulocyte Counts in Veterinary Medicine: Clinical Utility and Interpretation
• Platelet Function Testing in Veterinary Patients: Clinical Applications
• Proteinuria in Dogs and Cats: Diagnostic Approach and Clinical Significance
• Serial Monitoring of Canine and Feline Proteinuria: Clinical Decision-Making
• Monitoring Hepatic Enzyme Trends in Dogs and Cats: Clinical Significance

This article is educational professional reference material for veterinary audiences. It is not a substitute for veterinary diagnosis, individual clinical judgment, current product labeling, or applicable regulatory requirements.