Lot release testing is where manufacturing process control becomes a regulatory obligation. Before any diagnostic consumable ships, ISO 13485:2016 Clause 8.2.6 requires documented acceptance criteria, evidence of conformity against those criteria, and authorization by a designated individual. Under the FDA's Quality Management System Regulation (QMSR), which took effect February 2, 2026 and incorporates ISO 13485:2016 by reference into 21 CFR Part 820, that requirement now governs every FDA-regulated IVD manufacturer.
The challenge at scale isn't knowing the requirement exists — it's translating it into a test menu that catches real failure modes without creating hold-up cycles that stall shipment.
What Tests Belong in a Lot Release Package
There's no universal test menu mandated for IVD consumables. What belongs in your lot release package depends on the product, its intended use, and the failure modes your process can introduce. The following categories cover the ground most diagnostic consumable manufacturers need.
Dimensional and Physical Checks
For lateral-flow strips, microfluidic cartridges, and multi-well plates, dimensional conformance directly affects assay performance. Membrane width, flow-control zone dimensions, reagent-pad position, well volume, and seal geometry are all attributes where a manufacturing shift produces a functional failure — often without any visible defect.
Where production-scale data are available, tolerances grounded in actual process capability — rather than engineering drawings alone — are more likely to reflect the real distribution of manufactured units and to remain defensible if an inspector asks why a given limit was chosen. If your process runs at ±0.3 mm but your labeled specification is ±1.0 mm, a lot can pass release while running at the edge of what the assay actually tolerates. Tighter internal release limits — derived from process capability and, where applicable, stability degradation over shelf life — create a buffer that keeps released product conformant through expiry. ASTM E2709-23 provides a general statistical methodology for demonstrating capability to comply with an acceptance procedure — computing a confidence-bounded probability of passing defined acceptance limits from process distribution data. The standard's own scope notes that the methodology was originally developed for pharmaceutical drug-product quality characteristics but is applicable to acceptance procedures across industries; applied to IVD dimensional and reagent-concentration attributes, the approach requires adaptation — grounding the degradation models and acceptance pathways in ISO 13485:2016 Clause 8.2.6 rather than drug GMP frameworks — but the underlying statistical logic for capability-adjusted release intervals translates directly.
Any measurement tool used in dimensional acceptance — calipers, vision systems, coordinate measuring machines — must be part of a calibration program with documented intervals and acceptance criteria, per ISO 13485:2016 Clause 7.6. A lot release record that references measurements from an out-of-calibration instrument is not defensible.
Functional Assays
Functional testing is the core of IVD lot release. The specific assay depends on the product, but the question is always the same: does this lot perform within the limits defined in the cleared or approved product specification?
For lateral-flow devices, that typically means panel testing against positive controls at the stated sensitivity threshold and negative controls at the stated specificity threshold. A WHO/FIND lot-testing program that evaluated 6,056 malaria rapid diagnostic test lots between 2007 and 2017 found that failure modes included non-detection of target parasite antigens at the stated sensitivity threshold (the dominant finding among lots that failed at initial receipt) and insufficient buffer volumes in single-use vials (accounting for the majority of failures detected after long-term storage) — both of which a manufacturer-side functional assay and dimensional/volume acceptance program would be positioned to catch before shipment (PMC7359453). The 1.1% lot failure rate observed in that program represents failures detected by the WHO/FIND lot-testing program itself — 26 at receipt of the RDT lot in the testing laboratory (intercepted before distribution) and 43 after long-term storage — not failures that reached end-users (PMC7359453). Robust manufacturer-side functional assay and dimensional/volume acceptance testing is what prevents non-conforming lots from ever reaching an external lot-testing program or the field in the first place.
For reagent-based consumables — enzyme-linked immunoassay plates, PCR cartridges, clinical chemistry substrates — a different set of failure modes applies than for lateral-flow RDTs, and the functional assay design considerations differ accordingly. A widely recognized assay-design concern in these categories is the use of acceptance criteria that are anchored entirely to a single lot-specific control prepared and run at the time of release: because control material itself can vary between preparations, lot-to-lot variability in the control propagates directly into pass/fail outcomes, in effect moving the acceptance gate rather than holding it fixed. Where technically feasible, criteria expressed in absolute terms — for example, signal-to-noise ratio ≥ X at a defined analyte concentration, or Ct value within a defined range — provide a stable acceptance gate that does not shift with control-material variation. The appropriate form of acceptance criteria for a given product should be justified in the design history file based on the assay's performance characteristics and the sources of variability identified during development and scale-up.
Bioburden and Sterility
Not all diagnostic consumables require sterility. Many IVD consumables are non-sterile by design and labeling. But where a bioburden limit is part of the product specification — either because the consumable contacts clinical samples with a defined contamination-risk requirement, or because bioburden control is a precondition for a downstream sterilization step — the test method must be ISO 11737-1:2018 compliant, with a validated sampling plan and calculation method.
For consumables that are labeled sterile, the primary lot release basis is the validated sterilization process itself — ISO 13485:2016 Clause 7.5.6 requires that any special process whose output cannot be fully verified by subsequent testing be validated before use, and sterilization is the textbook example. Where a confirmatory sterility test is performed at lot release, the applicable compendial method is USP <71> Sterility Tests, not ISO 11737-2:2019; ISO 11737-2:2019 is scoped exclusively to sterility tests performed during sterilization process definition, validation, and maintenance — its scope explicitly excludes routine product release testing. Sterility testing alone is not a substitute for a validated sterilization process — ISO 13485:2016 Clause 7.5.6 requires that any process whose output cannot be fully verified by subsequent testing must be validated before use. Sterilization is a textbook special process under that clause: the validated cycle is the release basis, and the lot release sterility test is a confirmatory check, not the primary assurance mechanism.
Writing Release Specifications That Hold Up
A lot release specification is a controlled document. Under ISO 13485:2016 Clause 4.2.3, it belongs in the Medical Device File — the post-QMSR equivalent of the Device Master Record — and must be version-controlled and formally approved. That governance structure matters: a release spec that exists in a spreadsheet someone emails around will not survive an FDA audit.
Beyond document control, defensible release specifications have three properties.
Criteria are measurable and unambiguous. "Adequate color development" is not an acceptance criterion. "Line intensity ≥ 2.0 on a calibrated reader per SOP-QC-047" is. Every attribute tested at release needs a numeric pass/fail threshold or an explicit reference to the test method that defines it.
Criteria are traceable to product performance. The link between a dimensional tolerance and the assay performance data that justified it should be documented — in a design history file, a validation report, or a design of experiments record. If an inspector asks why the membrane-positioning tolerance is ±0.2 mm and not ±0.5 mm, the answer should be in the file.
The release record captures everything ISO 13485:2016 Clause 8.2.6 requires. That means: the acceptance criteria used (not just a pass/fail notation), the identity of the person authorizing release, and evidence that all planned acceptance activities were completed. A release record missing the authorizing signature or referencing an outdated version of the acceptance criteria creates a nonconformance even if the product itself is conformant.
Common Gaps That Cause Lot Failures and Hold-Ups at Scale
These are the patterns that appear consistently as diagnostic consumable manufacturers move from low-volume R&D production to commercial volumes.
Specifications written for development lots, not production lots. Acceptance criteria set during development often reflect a process that looked stable at 500 units. At 50,000 units with different raw material lots, a new equipment qualification, and three operators instead of one, the process has more sources of variation. Specifications that weren't set with production-scale process capability data will produce unexpected failures — not because the product changed, but because the specification never accounted for the real production distribution.
No internal release limits tighter than the labeled spec. Releasing to the full width of the labeled specification assumes the product will never degrade. For reagent-containing consumables with a defined shelf life, a lot released at the edge of specification may fall out of spec before expiry. Process capability-derived internal limits — set inside the labeled specification to account for analytical measurement uncertainty and stability degradation — prevent this without requiring a labeled specification change.
Functional assay controls that shift between lots. Inter-lot variability in control materials propagates directly into functional assay results. If the positive control used in lot release testing comes from a different preparation than the one used to establish acceptance criteria, the test is no longer measuring the same thing. Lock in reference material preparation, storage, and lot qualification as part of the release procedure — not as an informal practice.
Calibration gaps in measurement systems. ISO 13485:2016 Clause 7.6 requires calibration records for all measurement and test equipment used in acceptance activities. At scale, this means not just the primary instruments but the reference standards used to calibrate them. A missing calibration cycle on a plate reader used for functional assays puts every lot released with that instrument at risk of a retroactive nonconformance.
Release records that can't survive an audit. Incomplete records — missing authorization signatures, test results recorded after the release date, references to superseded SOPs — are recurring audit vulnerabilities, because each one directly contradicts what ISO 13485:2016 Clause 8.2.6 requires a lot release record to demonstrate: that all planned acceptance activities were completed, that acceptance criteria are identified, and that a designated individual authorized release. At low volumes, these gaps are often caught informally. At commercial volumes, they accumulate and surface during audits as systemic quality system failures, not isolated events.
Building a Release Package That Scales
The mechanics of lot release testing don't change as volume increases, but the failure modes do. At scale, the risks are consistency failures: instrument calibration gaps, operator-to-operator variation in assay execution, raw material variability that wasn't visible at low volumes, and release record integrity under throughput pressure.
Building a release package designed for commercial volumes means setting specifications against production-scale capability data, maintaining strict control of reference materials, running a disciplined calibration program, and treating the lot release record with the same rigor as any other controlled QMS record — because under the QMSR and ISO 13485:2016 Clause 8.2.6, it is a required, auditable record that must identify acceptance criteria, document evidence of conformity, and bear the authorization of a designated individual before product ships.