LOINC codes — Logical Observation Identifiers Names and Codes — are universal identifiers for laboratory tests, clinical measurements, vital signs, survey instruments, and healthcare documents.
LOINC answers one question across every healthcare system that uses it — what was measured or observed?
A blood glucose test ordered at one hospital and performed at another carries the same LOINC code regardless of the local test name, the lab information system (LIS), or the EHR platform.
LOINC Version 2.82, released February 2026, contains 109,325 terms maintained by the Regenstrief Institute, and the database is available as a free download with a free login.
In this guide, we’ll be exploring:
- How to find the right LOINC code using SearchLOINC
- How LOINC codes are structured (the six-part naming model)
- What LOINC codes are used for across clinical, lab, and public health settings
- How LOINC differs from CPT, ICD-10-CM, and SNOMED CT
- Common mapping mistakes and how to avoid them
How are LOINC codes structured?
Every LOINC term is built from up to six semantic parts. The structure is why two lab tests with similar names can have different LOINC codes — and why choosing a code based on the test name alone is a common mapping error.
| LOINC part | What it represents | Example |
| Component | The substance or analyte being measured | Glucose, hemoglobin A1c, sodium |
| Property | The characteristic measured | Mass concentration, substance concentration, presence |
| Time aspect | When or over what period the measurement occurs | Point in time, 24-hour collection |
| System | The specimen or body site | Serum/plasma, urine, whole blood, CSF |
| Scale | The type of result value | Quantitative, ordinal, nominal, narrative |
| Method | The measurement technique (included only when clinically significant) | Immunoassay, PCR, automated count |
A single analyte like glucose can have dozens of LOINC codes depending on the specimen (serum vs. urine vs. CSF), the property (mass concentration vs. substance concentration), the timing (point-in-time vs. 24-hour), and the scale (quantitative vs. ordinal).
The six-part structure is what makes LOINC specific enough to support interoperability — and what makes code selection harder than it looks.
Anatomy of a LOINC Code
Each LOINC term uses up to six parts to define exactly what was measured, from what specimen, using what method.
Component
What is measured (Glucose)
Property
What attribute (Mass conc.)
Time
When measured (Point in time)
System
What specimen (Serum/Plasma)
Scale
Result type (Quantitative)
Method
How measured (if clinically relevant)
What are LOINC codes used for?
LOINC is not a billing code, not a diagnosis code, and not primarily a procedure code.
It identifies the observation — the question that was asked — while other terminologies identify the answer (SNOMED CT for clinical findings), the condition (ICD-10-CM for diagnoses), and the service (CPT for procedures).
| Use case | How LOINC helps |
| Lab result exchange | Identifies the exact test or measurement being reported across systems |
| EHR and LIS interoperability | Helps receiving systems file incoming results into the correct field |
| Health information exchange | Standardizes observations so records shared between organizations retain clinical meaning |
| Quality reporting (HEDIS, eCQMs) | Defines the observation codes that quality measures reference |
| Public health reporting (ELR, eCR) | Standardizes reportable lab results for electronic lab reporting to agencies like the CDC |
| Research and analytics | Makes data from different sites and EHRs comparable in common data models like OMOP |
| Patient apps and FHIR APIs | Gives third-party applications a standard way to identify observations |
LOINC’s official description says it is used globally to identify data and move it between systems — by reference labs, healthcare providers, government agencies, insurance companies, software manufacturers, and researchers.
The Version 2.81 release added more than 4,000 new and updated concepts, including nearly 2,000 lab ordering concepts, 350+ PhenX Toolkit concepts, and nearly 400 food and environmental allergen concepts.
How do you find the right LOINC code?
Finding a LOINC code is not a keyword search — it is a six-part matching exercise. SearchLOINC is the official web-based tool, and RELMA (Regenstrief LOINC Mapping Assistant) is the downloadable mapping application.
Step-by-step process:
Identify the Analyte
Determine exactly what laboratory analyte or clinical observation is being measured.
Confirm the Specimen
Identify the specimen type or body system associated with the observation.
Determine Result Type
Classify the result as quantitative, qualitative, ordinal, or narrative.
Review Timing
Determine whether the observation is point-in-time or collected over a defined interval.
Review the Method
Determine whether the testing method is clinically significant for code selection.
Search SearchLOINC
Use SearchLOINC or RELMA to identify candidate LOINC terms.
Compare Candidates
Review every candidate across all six LOINC attributes before selecting the best match.
Validate
Review the proposed mapping with laboratory, clinical, and informatics experts.
Document the Mapping
Record the local-code-to-LOINC mapping and maintain version history.
Maintain & Update
Re-check mappings after every LOINC release to keep terminology current.
LOINC recommends sending the sender’s local code and local name alongside the LOINC code and name to support debugging and mis-mapping detection — and updating to the current LOINC version within 90 days of publication.
In practice, the search process is where most mapping errors originate. Two tests with nearly identical names — “Glucose, serum” and “Glucose, plasma” — may share a LOINC code if the terminology treats serum/plasma as equivalent, or they may not.
Confirming the actual specimen used by the lab is a step many teams skip, and the result is a mapping that looks correct but misrepresents the data.
How do LOINC codes differ from CPT codes?
AMA defines CPT codes as describing medical services and procedures performed by physicians and other qualified healthcare professionals.
LOINC identifies what was measured. CPT identifies what service was performed. The two systems answer different questions about the same clinical event.
| Question | LOINC | CPT |
| What does it identify? | The observation, test, measurement, or document | The service or procedure performed |
| Primary purpose | Clinical data exchange and interoperability | Billing, claims, and procedure reporting |
| Example use | What lab result was reported | What lab service was billed |
| Level of detail | Specific to specimen, scale, property, and method | Often less granular for lab result meaning |
| Maintained by | Regenstrief Institute | American Medical Association |
| Best question it answers | “What was measured?” | “What service was provided?” |
A single CPT code can map to multiple LOINC codes because CPT is less specific about specimen type, measurement property, and result scale.
When a payer receives a claim with a CPT code for a metabolic panel, the CPT says “a metabolic panel was performed.”
When an EHR receives LOINC-coded results from that panel, each individual observation (sodium, potassium, glucose, creatinine) carries its own LOINC code specifying exactly what was measured, from what specimen, using what scale.
LOINC’s FAQ notes that CPT may be used for ordering and billing purposes and is often less specific than LOINC for lab tests.
How do LOINC codes differ from ICD-10?
CDC states that ICD-10-CM is used to code and classify medical diagnoses. LOINC identifies measurements and observations. The two systems describe different parts of the clinical picture.
| Question | LOINC | ICD-10-CM |
| What does it identify? | Tests, measurements, observations, documents | Diagnoses, diseases, conditions, injuries |
| Clinical role | What was measured or observed | What condition the patient has |
| Example relationship | HbA1c lab observation | E11.65 — Type 2 diabetes with hyperglycemia |
| Used for billing? | Not primarily | Required on claims to support medical necessity |
| Best question it answers | “What result is this?” | “What condition does the patient have?” |
LOINC and ICD-10-CM work together in clinical workflows. The LOINC code identifies the HbA1c test. The ICD-10-CM code identifies the diabetes diagnosis being monitored.
The CPT code identifies the billable lab service. SNOMED CT may describe the clinical finding. Each terminology has a defined lane.
How does LOINC fit with SNOMED CT, RxNorm, and other code systems?
SNOMED International’s 2025 LOINC implementation guide describes use cases including lab ordering, standardizing results, cross-terminology retrieval, and clinical decision support.
| Code system | What it typically represents |
| LOINC | Observations, measurements, lab tests, survey questions, documents |
| SNOMED CT | Clinical findings, problems, procedures, organisms, body structures |
| ICD-10-CM | Diagnoses and conditions for morbidity classification and billing |
| CPT | Services and procedures for billing and claims |
| RxNorm | Medications and normalized drug names |
| UCUM | Units of measure for quantitative results |
In many implementations, LOINC identifies the observation and SNOMED CT codes the result value — the two terminologies complement each other rather than compete.
Where do LOINC codes appear in healthcare data exchange?
LOINC codes travel inside messages and resources, not as standalone identifiers. Clinicians rarely see a raw LOINC code — it sits inside the data infrastructure that powers result delivery, quality reporting, and interoperability.
- EHR lab result sections
- LIS test catalogs and interface builds
- HIE result exchange and document sharing
- HEDIS and digital quality measure reporting
- Patient access APIs delivering lab data to third-party apps
- Electronic laboratory reporting (ELR) to public health agencies
- FHIR Observation resources (the code element references LOINC)
- HL7 v2 lab result messages (OBX segments carry LOINC codes for each observation)
Mayo Clinic Laboratories states that it systematically assigns LOINC values to its assays and provides a LOINC values spreadsheet for LIS integration.
The TriNetX Global Collaborative Network, with EHR data for about 179 million patients as of April 2025, structures laboratory terms by LOINC code — though the network also warns that tests with the same LOINC code may still differ by laboratory method.
What are the most common LOINC mapping mistakes?
A 2026 JMIR Medical Informatics study on a real-world oncology dataset of approximately 10 million cancer patients found that correct LOINC code and unit conformance improved from 73.1% to 99.7% after applying a harmonization framework — evidence that raw LOINC mappings in production data carry significant error rates before cleanup.
Previous studies cited in the same paper reported that 6% to 19% of laboratory tests cannot be accurately mapped to LOINC. A 2025 study of university hospital research mappings found a LOINC mapping inconsistency rate above 15%.
| Mistake | Why it weakens the mapping | Better approach |
| Choosing by test name only | Similar names can have different specimens or scales | Compare all six LOINC parts |
| Ignoring specimen type | Serum, plasma, urine, blood, and CSF may differ | Confirm actual specimen used by the lab |
| Ignoring result scale | Quantitative and qualitative results need different codes | Match scale to what the lab actually reports |
| Mixing order and result codes | Orders and observations have different use cases | Confirm whether mapping an order or a result |
| Ignoring method when relevant | Method may change clinical meaning for some assays | Include method if clinically significant |
| Using deprecated codes | Retired LOINC terms create downstream data issues | Check term status with each release |
| Not involving lab experts | Informatics-only mapping misses clinical detail | Use multidisciplinary review for complex tests |
| Not updating after releases | LOINC adds thousands of terms per release | Review and update mappings within 90 days of each version |
The strongest takeaway from recent evidence — LOINC codes support interoperability, but they do not guarantee it alone.
Accurate mapping requires local-code provenance, unit standardization, version control, expert review, and ongoing governance.
Is LOINC free to use?
Regenstrief distributes LOINC and RELMA free of charge under its license terms.
The database download requires a free account. Organizations should use the current production version and maintain update cycles aligned with LOINC release schedules — currently twice per year.
The free-of-charge model is a deliberate strategy to maximize adoption. License terms still apply, particularly around redistribution and derivative works.
Organizations should review the license before incorporating LOINC data into commercial products.
LOINC codes identify the test — MedHeave identifies the revenue attached to it.
Standardized observation data powers interoperability.
But when lab results, diagnostic observations, and clinical measurements do not connect to properly coded and submitted claims, revenue stays on the table — especially in multi-specialty practices where coding complexity varies by payer and procedure type.
- MedHeave manages coding, claim submission, and follow-up for every billable encounter
- Charge capture reviews catch missed or undercoded services before claims go out
- Payer-specific billing guideline sheets map coding requirements by specialty and payer
- Denial rework happens the same day the ERA arrives
Talk to MedHeave about connecting your clinical data quality to consistent revenue performance.
Frequently asked questions
Here are some commonly asked questions about LOINC codes:
LOINC codes identify laboratory tests, clinical observations, vital signs, survey instruments, and healthcare documents. A LOINC code tells any system receiving the data exactly what was measured or observed — the analyte, the specimen, the measurement property, and the result type. LOINC is used in lab result exchange, EHR interoperability, health information exchange, quality measure reporting, public health surveillance, and research data harmonization. It is not a billing code or a diagnosis code.
CPT codes identify the service or procedure performed (what was billed). LOINC codes identify what was measured or observed (the clinical data). CPT is maintained by the AMA for billing and claims reporting. LOINC is maintained by the Regenstrief Institute for clinical data exchange and interoperability. A single CPT code for a lab service can correspond to multiple LOINC codes because LOINC is more specific about specimen, property, scale, and method than CPT needs to be for billing purposes.
Start by identifying the exact analyte, specimen, and result type. Then search in SearchLOINC (the official web application at loinc.org) or RELMA (the downloadable mapping tool). Compare candidate terms across all six LOINC parts — component, property, time, system, scale, and method. Validate your selection with lab and informatics experts, and document the mapping with version information. Re-check mappings after each LOINC release to catch deprecated or updated terms.
ICD-10-CM classifies diagnoses, diseases, and conditions. LOINC identifies observations, tests, and clinical measurements. ICD-10 answers “what condition does the patient have?” — LOINC answers “what was measured or observed?” The two systems work together in clinical workflows — LOINC identifies the HbA1c observation while ICD-10-CM codes the diabetes diagnosis being monitored. ICD-10 is required on claims for medical necessity. LOINC is required for interoperable clinical data exchange.
No. LOINC covers laboratory observations, clinical measurements (vital signs, functional assessments), survey instruments (PHQ-9, social needs screening tools), clinical document types (discharge summaries, progress notes, radiology reports), and panels (CBC, metabolic panel). The Clinical LOINC domain includes vitals, surveys, and document ontology codes. Laboratory LOINC covers specimen-based tests. The terminology spans far beyond the lab.
