IVD Analytical Performance Studies under IVDR
An IVD analytical performance study demonstrates that an in vitro diagnostic device correctly detects or measures a specific analyte. Under IVDR (EU) 2017/746 it is one of the three pillars of performance evaluation, with requirements set out in Annex I, Section 9.1(a) and Annex XIII. We help IVD manufacturers plan, run and document these studies, from protocol design to a notified-body-ready Analytical Performance Report.
Three Pillars of IVD Performance Evidence
Every IVD placed on the EU market must show that it is safe and performs as claimed. Under IVDR (EU) 2017/746, that proof comes from a performance evaluation built on three pillars.
Scientific validity
The association between the analyte and a clinical condition or physiological state.
Analytical performance
The ability of the device to correctly detect or measure a specific analyte.
Clinical performance
The ability of the device to give results that correlate with the clinical condition in the target population.
Why IVD Analytical Performance Studies Matter
These laboratory studies are the technical foundation of your performance evaluation. Results are documented in an Analytical Performance Report (APR), which feeds directly into the Performance Evaluation Report (PER) reviewed by your notified body.
- Proves accuracy, precision, sensitivity and specificity for the target analyte
- Shows reliable, reproducible results across claimed specimen types and conditions
- Supports the GSPRs in Annex I and your labelling claims
Incomplete analytical data is a common source of notified body findings. Getting the study design right first time avoids costly delays.
Are you looking for an email with full information about IVDR CE Marking?
How to Conduct IVD Analytical Performance Studies
Analytical performance is purely technical: no correlation with the target disease is required. Studies are run in a laboratory by testing the device against samples with known concentrations or characteristics of the analyte.
Define claims and scope
Link each claim to the intended purpose, assay type and applicable GSPRs.
Write study protocols
Set design, sample size, statistics and acceptance criteria using CLSI guidelines.
Select and qualify samples
Use well-characterised samples for every claimed specimen type.
Run, analyse and report
Test against acceptance criteria and compile the Analytical Performance Report.
Sample types used in analytical studies
Analytical data can be generated from a combination of:
- Well-characterised clinical samples with documented collection, handling and storage
- Well-documented panels of contrived or spiked samples
- External quality assessment (EQA) panels and certified reference materials
One dataset per specimen type
Each specimen type you claim, such as whole blood, plasma or serum, needs its own supporting analytical data. The results are summarised in your performance evaluation and IVDR technical documentation.
IVDR Analytical Performance Characteristics
The IVDR lists the analytical performance characteristics a device must achieve, where applicable. Which ones apply depends on the assay technology (qualitative, quantitative or semi-quantitative) and the intended use environment. Any omission must be justified.
| Characteristic | What it demonstrates | Typical guideline |
|---|---|---|
| Analytical sensitivity | Limit of blank, limit of detection (LoD) and limit of quantitation (LoQ) | CLSI EP17 |
| Analytical specificity | Cross-reactivity and endogenous or exogenous interference | CLSI EP07 |
| Trueness (bias) | Closeness of the mean result to the reference value | CLSI EP09 |
| Precision | Repeatability and reproducibility of results | CLSI EP05 |
| Accuracy | Combined effect of trueness and precision | CLSI EP21 (total analytical error) |
| Measuring range and linearity | Interval over which results remain proportional and reliable | CLSI EP06 |
| Cut-off | Decision point, with specimen collection and handling criteria | CLSI EP12 (qualitative tests) |
| Carry-over and stability | No sample-to-sample effect; specimen and reagent validity under storage | CLSI EP25 (reagent stability) |
Which characteristics apply depends on the assay type. CLSI guidelines are widely recognised references, not mandatory requirements.
Guidelines and standards
Studies are typically designed using Clinical and Laboratory Standards Institute (CLSI) guidelines, for example:
The evidence must address every applicable parameter in Annex I, Section 9.1(a) of Regulation (EU) 2017/746, and support the related GSPRs with existing or newly generated data.
Using data from earlier device versions
Data generated on a previous version of the device, including under the IVDD, can be reused where it remains valid.
You must justify why the device changes do not affect the data's validity and show that it still reflects the current state of the art.
When is analytical performance evidence enough?
This evidence is assessed together with scientific validity to decide whether clinical performance data is also needed, as described in MDCG 2022-2.
Established, standardised tests
Analytical data plus scientific validity is often enough to demonstrate conformity.
Established, non-standardised tests
Additional clinical performance data is usually required.
Novel tests
Dedicated clinical performance studies will most likely be needed.
Designing Robust IVD Analytical Performance Studies
Meeting each Annex I characteristic is only part of the job. Notified bodies also check whether the study design, statistics and links to risk management make the results credible for the way the device will actually be used.
Robustness testing
Robustness shows that results stay reliable when operating conditions vary within the limits a user or laboratory could reasonably meet. Typical factors include incubation time and temperature, reagent and calibrator lots, instruments, operators and specimen handling.
A PCR assay, for example, may be challenged with small changes in pipetting volume, cycling temperature or extraction input to confirm that detection near the limit of detection does not drift.
Robustness data also supports the operating limits in your instructions for use (IFU), so every tolerance stated in the IFU should trace back to a tested condition.
Statistics and sample size
The IVDR does not set fixed sample numbers, so the justification has to come from the study design. Recognised guidance gives a starting point: the CLSI EP05 single-site precision design, for example, uses 20 days, two runs per day and two replicates per run.
- Report point estimates with 95% confidence intervals, not single values
- Pre-define acceptance criteria in the protocol, linked to the state of the art and intended use
- Cover the full measuring range, including the cut-off and medical decision points
- Use several reagent lots and instruments where they could affect results
- Present results separately for each claimed specimen type
For linearity, a glucose assay claiming 20 to 600 mg/dL would test samples spread across, and slightly beyond, that interval to show results stay proportional at both ends.
Linking results to risk management
Analytical studies and risk management under EN ISO 14971 inform each other. Interference, cross-reactivity, carry-over and reagent instability are all potential causes of false-positive or false-negative results, so each identified hazard should map to a study that measures it, and each study result should feed back into the risk management file.
Where a limitation remains, such as a known interfering substance, it must be stated in the IFU so users can control the residual risk. Your notified body will check that the technical documentation, risk file and IFU tell the same story.
Common notified body findings on analytical studies
Most deficiencies come from gaps in planning rather than poor laboratory work. Typical findings include:
- Acceptance criteria set after results were known, or not linked to the state of the art
- Studies run on a single reagent lot or instrument
- Specimen types claimed in the IFU but not tested
- Interfering substances chosen without a rationale, such as ignoring medicines common in the target population
- Limit of detection claimed from too few replicates or without confidence intervals
- Reports that do not trace back to the protocol, or excluded data and deviations left unexplained
- Data from an earlier device version reused without a change impact assessment
A protocol review before testing starts is the cheapest way to avoid these findings.
Analytical Performance for Novel IVD Technologies
Newer test types need validation beyond the classic parameters, and they are reviewed closely because there is less established state of the art to compare against.
Next-generation sequencing (NGS) tests
Analytical validation must cover the whole workflow, from extraction and library preparation to sequencing and the bioinformatics pipeline. Typical evidence includes:
- Detection accuracy for each claimed variant type, such as SNVs, indels, copy-number changes and fusions
- Coverage depth and uniformity across target regions
- Limit of detection expressed as variant allele fraction
- Verification of the bioinformatics pipeline as IVD software, repeated after each update
Companion diagnostics
A companion diagnostic identifies patients who are suitable for a specific medicinal product and is classified as Class C under IVDR. Analytical data must show the test reliably detects the biomarker at the clinical cut-off used in the medicine's trials.
Before issuing a certificate, the notified body must consult the EMA or a national medicines authority on the device's suitability for that medicine. Aligning analytical and clinical study plans with the drug developer early avoids delays at this stage.
Self-testing and near-patient tests
For self-tests and near-patient tests, Annex I Section 19 requires the device to perform as intended in the hands of its intended users, taking their skills and environment into account. Analytical studies should therefore include conditions such as capillary specimens, untrained operators and variable room temperature, and the results should support the usability evidence and the simplified instructions for lay users.
Analytical Performance for EU and US Submissions
Many manufacturers place the same IVD on both the EU and US markets. Much of the laboratory work can be shared, because notified bodies and the FDA both accept studies designed to recognised CLSI guidelines, but the way evidence is packaged and reviewed differs.
| Aspect | EU (IVDR) | US (FDA) |
|---|---|---|
| Legal basis | Regulation (EU) 2017/746, Annex I Section 9.1(a) and Annex XIII | 21 CFR Part 809 plus 510(k), De Novo or PMA requirements |
| Where results go | Analytical Performance Report, feeding the Performance Evaluation Report | Analytical studies section of the premarket submission |
| Who reviews | Notified body for Class A sterile, B, C and D devices | FDA Center for Devices and Radiological Health (CDRH) |
| Reference standards | CLSI guidelines, harmonised standards and common specifications where they exist | FDA-recognised consensus standards, including many CLSI guidelines |
| After approval | Post-market performance follow-up (PMPF) under Annex XIII Part B | Device changes assessed for a new 510(k) or PMA supplement |
Plan one study programme against both sets of expectations: agree acceptance criteria, specimen types and reagent lots before testing starts, then present the same results in each format. See our guides to FDA IVD clinical evidence and the 510(k) IVD submission workflow.
IVDR Analytical Performance Support from I3CGLOBAL
Our IVDR consultants support IVD manufacturers worldwide across all risk classes, from Class A to Class D.
- Study protocols with defensible acceptance criteria
- Interpretation of sensitivity, specificity, trueness, precision and linearity results
- Analytical Performance Report aligned with your Performance Evaluation Plan
- Gap assessment of existing IVDD analytical data
Start your IVDR project
Share your device details and get a tailored proposal.
IVDR Quote Request Contact UsIVD Analytical Performance: Frequently Asked Questions
What is an IVD analytical performance study?
It is a laboratory study that shows how well an in vitro diagnostic detects or measures a specific analyte. Under IVDR it is one of three components of the performance evaluation, alongside scientific validity and clinical performance.
Which parameters demonstrate analytical performance under IVDR?
Annex I, Section 9.1(a) lists analytical sensitivity, analytical specificity, trueness (bias), precision (repeatability and reproducibility), accuracy, limits of detection and quantitation, measuring range, linearity, cut-off (including specimen collection and handling criteria and control of known interference) and cross-reactions.
What is the difference between analytical and clinical performance?
Analytical performance shows the test correctly detects or measures the analyte. Clinical performance shows the result correlates with a particular clinical condition or physiological state in the target population, for example through diagnostic sensitivity and specificity.
Can I reuse analytical data generated under the IVDD?
Yes, where it remains valid. You must justify that any device changes since the data was generated do not affect its validity and that it still reflects the current state of the art.
How many samples does an analytical performance study need?
The IVDR does not set fixed sample numbers. Sample size, replicates, runs and lots should follow recognised guidance for each characteristic, such as the relevant CLSI protocols or common specifications, and be statistically justified in the study protocol.
Where is analytical performance documented in the technical file?
Study protocols, data and reports are summarised in the analytical performance report and referenced in Annex II, Section 6.1 of the IVDR technical documentation. The conclusions then feed into your performance evaluation report.
Do analytical studies need repeating after a design or supplier change?
Not always. Assess each change, such as a new reagent supplier, component or software version, for its impact on performance. Where a characteristic could be affected, run targeted verification or bridging studies and update the performance evaluation.
Why are robust analytical performance studies needed for IVDR CE marking?
Notified bodies assess analytical performance as part of the IVDR CE marking process. Robust, well-documented studies support your performance claims, reduce review findings and form the basis of your Summary of Safety and Performance for Class C and D devices.
What is robustness testing for an IVD?
Robustness testing shows that an IVD gives reliable results when operating conditions vary within realistic limits, such as temperature, incubation time, reagent lots, instruments or operators. The results support the operating limits stated in the instructions for use.
How is analytical performance shown for NGS tests?
For NGS tests, analytical performance covers the whole workflow, including extraction, library preparation, sequencing and the bioinformatics pipeline. Evidence typically includes accuracy for each claimed variant type, coverage depth and uniformity, and limit of detection expressed as variant allele fraction.
Can the same analytical studies support both EU and US submissions?
Often, yes. Notified bodies and US regulators both accept studies designed to recognised CLSI guidelines. Agreeing acceptance criteria, specimen types and reagent lots before testing lets one study programme support both, with results presented in each submission’s format.
What is an Analytical Performance Report?
An Analytical Performance Report (APR) summarises the protocols, results and conclusions of all analytical performance studies for a device. It shows how each applicable characteristic in Annex I, Section 9.1(a) has been demonstrated and feeds into the Performance Evaluation Report, together with scientific validity and clinical performance evidence.
Which IVDR device classes need notified body review of analytical performance?
Notified bodies review analytical performance evidence for Class A sterile, Class B, Class C and Class D devices. For Class D devices, an EU reference laboratory, where designated, also verifies the performance claimed and compliance with any applicable common specifications. Non-sterile Class A devices are self-declared, but the same evidence must still be in the technical documentation.
Updated | Author: Sobha Augustine, Founder, I3CGLOBAL | Reviewed by: Swathi Tummala, Senior Regulatory Consultant, I3CGLOBAL