PMA: FDA Premarket Approval for Class III Devices
Follow a US device application from scientific evidence through FDA review to controlled changes and lifecycle records.

On this page · 12 sections
- 1What PMA actually means: application, review and approval
- 2Which devices require assessment for PMA?
- 3Clinical investigation before PMA: permissions and evidence planning
- 4PMA application contents: scientific evidence and a coherent device baseline
- 5FDA review: filing, the stated interval and possible actions
- 6PMA supplements: define the change and support the decision
- 7Post-approval lifecycle: conditions, labeling and assigned actions
- 8Manufacturing and QMSR: product scope before site assumptions
- 9Evidence model, system handoffs and exception recovery
- 10Application weaknesses and practical prevention controls
- 11Metrics that reveal unresolved evidence and lifecycle work
- 12Implementation checklist: review the decision, not just the document
How does PMA apply to your shop floor?
Pick your industry and scale — Ask V5 rewrites the definition in your context, gives a worked example, and shows what V5 does on day one.
01What PMA actually means: application, review and approval
PMA means any premarket approval application for a US Class III medical device, including all information submitted with it or incorporated by reference, as defined in 21 CFR 814.3(e)).
FDA's own premarket approval page puts the marketing consequence plainly for devices found not substantially equivalent to a Class I, II or Class III 510(k) device: "the device must have an approved PMA before marketing in the U.S." (FDA, Premarket Approval). Submission and filing are steps toward that decision; they are not permission to market.
FDA also uses premarket approval to describe its scientific and regulatory review of the safety and effectiveness of US Class III devices in its PMA overview. The same abbreviation can therefore identify the application or the broader review process. For operational clarity, distinguish these from the agency’s approval decision:
| Term | Meaning in a working record |
|---|---|
| Application | The information submitted or incorporated by reference. |
| Review | FDA’s evaluation of the application. |
| Approval | A recorded agency action, not another name for submission. |
For US medical-device PMAs, Part 814 is the governing regulation. Keep its binding provisions separate from FDA explanatory material and from recommended operating practices. An internal checklist can help organize work, but it is not itself a regulatory provision.
As a practical recordkeeping convention, label application versions, review correspondence and the approval decision separately. This helps a reader understand whether a document describes what was requested, what FDA evaluated or what the agency decided. Treat software support as an information-management design choice, not an approval authority; human review and approval remain human responsibilities.
For device-specific route questions, turn to the scope discussion below, supported by medical-device classification and the comparison of 510(k) and PMA pathways, rather than deciding from the abbreviation alone.
02Which devices require assessment for PMA?
Start with the product, intended use and jurisdiction—not the submission acronym. This section principally addresses US Class III medical devices. As a practical planning control, document the classification basis and submission-selection rationale separately; a class designation alone is not a complete pathway decision.
For US Class III devices, FDA’s PMA overview notes that some preamendment devices may require a Class III 510(k). This qualification matters when assessing the route for an established device type.
A narrow legal alternative also exists: for a US Class III device, 21 CFR 814.19 treats a product development protocol declared completed by FDA as an approved PMA.
Use the following as a planning comparison, not as eligibility tests or a required record format. Each row identifies a different decision to resolve.
| Term | Question to resolve | Decision record | Scope boundary |
|---|---|---|---|
| PMA | Does this device’s classification and regulatory history point to PMA? | Classification and route rationale | Do not infer the route from class alone. |
| 510(k) | Is premarket notification the appropriate submission? | Submission-selection rationale | Assess separately from PMA. |
| De Novo | Is De Novo appropriate to the classification question? | Classification-request rationale | Do not import PMA approval terminology. |
| IDE | What investigation permissions need assessment? | Investigation-permission assessment | Keep study planning distinct from marketing. |
| QMSR | What manufacturing-system scope applies to this product and activity? | Manufacturing applicability assessment | Keep separate from submission selection. |
For US medical devices associated with blood collection, processing and cellular therapies, FDA’s Devices Used in Blood Establishments describes CBER review of marketing and investigational submissions. Include center assignment in the planning record rather than assuming every device submission has the same destination.
Device-regulated human cells, tissues, and cellular and tissue-based products (HCT/Ps) are addressed in the manufacturing section. Non-US markets and non-device products are outside this article’s authorization analysis. For investigation planning, continue to clinical evidence before the application.
03Clinical investigation before PMA: permissions and evidence planning
For US medical-device investigations, check the current IDE and human-subject protection provisions for study-specific scope and permissions. Do not infer an investigation authorization, exemption or review timetable from the device’s PMA planning status. This section addresses evidence organization, not study eligibility or permission to begin enrolling participants.
As a practical approach across medical-device development phases, keep separate, linked records for study permissions, protocol versions, clinical device configurations and marketing-application status. This separation helps reviewers distinguish what was authorized, what was investigated and what the team proposes to market.
- Which intended use and patient population does the study address?
- Which device configuration was investigated, including relevant hardware, software and accessories?
- Which endpoints were evaluated, and which analysis version supports the reported results?
- Which differences from the proposed commercial configuration need assessment before the evidence is used?
Build an evidence applicability map linking each protocol version to the investigated configuration, study report, analysis version and intended application section. Record unresolved differences alongside the proposed assessment, supporting rationale and review status. Preserve those links when reports or analyses are revised rather than silently replacing the earlier baseline.
Identify scientific, clinical and regulatory review owners for the map. These are suggested coordination roles, not statutory assignments to job titles. Carry the resulting map into PMA application contents, where the configuration-mismatch worked case shows how to assess evidence applicability.
04PMA application contents: scientific evidence and a coherent device baseline
For US Class III device PMA applications, 21 CFR 814.20(b)(6)(6)) requires technical sections containing enough data and information for FDA to determine whether to approve or deny the application. The practical preparation task is to connect each scientific conclusion to the device configuration and intended use it actually supports.
For US Class III device applications, FDA’s technical-sections discussion describes a usual division between non-clinical laboratory studies and clinical investigations. For these US applications, FDA’s non-clinical discussion lists microbiology, toxicology, immunology, biocompatibility, stress, wear, shelf life and other laboratory or animal tests. These are examples, not a universal test battery.
Use a working index, not a substitute legal checklist
As an organizational aid, use the following working index. It is not an exhaustive list of mandatory PMA elements. For each entry, identify its owner, controlled revision, supporting records and unresolved questions:
- Administrative information and application identifiers.
- Intended use and device description.
- Configuration baseline identifying the proposed commercial design and relevant variants.
- Scientific reports, with links to protocols, analyses and the configurations evaluated.
- Manufacturing narrative tied to the proposed device.
- Labeling aligned with the intended use and supporting conclusions.
- Referenced records, including their identifiers, revisions and locations.
Where teams also maintain a technical file, use the index to identify reusable records without assuming that the dossier structures are interchangeable.
Worked case: clinical and commercial configurations differ
Assume the clinical study used a configuration different from the proposed commercial configuration. The unit of analysis is the study-to-configuration relationship, not the number of reports collected. No test outcome or bridging conclusion is assumed.
- Identify the difference. Compare the study configuration against the commercial baseline and describe the changed feature precisely, using controlled identifiers rather than informal product names.
- Assess affected conclusions. Identify which safety or effectiveness conclusions might depend on that feature; explain the possible connection rather than declaring the change immaterial.
- Locate supporting evidence. Link relevant laboratory reports, design analyses or clinical information to the difference. Record what each item addresses and what it leaves unanswered.
- Record unresolved questions. Assign an owner to each uncertainty, such as whether the available study conditions represent the proposed configuration. Keep missing support visible.
- Obtain a human scientific assessment. Have appropriate scientific reviewers assess applicability, document the reasoning and determine whether further evidence is needed. Do not infer acceptability merely because both configurations share a product name.
Carry these relationships into the evidence and integration model so reviewers can follow a conclusion back to its supporting configuration and records.
For US Class III device PMA applications, FDA recommends a quality-control audit before submission. A practical audit question is whether the scientific narrative, configuration references and unresolved-question log tell the same story.
05FDA review: filing, the stated interval and possible actions
For US Class III device PMA applications, 21 CFR 814.44(c)) states FDA’s action interval as the later of 180 days from filing or the number of days after filing determined under 814.37(c). For these US applications, the listed actions are an approval order, an approvable letter, a not approvable letter, or an order denying approval. This is an interval for agency action, not a promised approval date.
Keep the date basis visible
As an application-management practice, maintain a correspondence register that separates transmission, filing status, response work and the actual agency action. Recommended fields are:
- Submission transmission date and the identifier of the transmitted package.
- Filing status, with the agency communication supporting that status.
- FDA request reference and the question or issue requiring a response.
- Response owner and response version, distinguishing working drafts from submitted material.
- Stated date basis, identifying whether a date comes from agency correspondence or internal planning.
- Actual agency action and a link to the corresponding letter or order.
Use internal dates to coordinate resources, not to represent FDA commitments. If teams use backward scheduling for response preparation, label the planning assumptions and retain the agency communication separately. Avoid allowing a dashboard’s target date to replace the recorded filing status or action.
Recover exceptions before closing the review record
A practical exception-recovery procedure assigns unanswered questions to named owners, reconciles conflicting response drafts and preserves the version actually submitted. Escalate scientific disagreements to named reviewers rather than silently selecting one team’s conclusion. Record the disposition and supporting rationale so the final response remains traceable to the issue it addresses.
After recording the agency action, route unresolved response work to its owners. Where approval establishes associated conditions, connect the action and those conditions to post-approval lifecycle records, with assigned owners and linked source correspondence. This downstream workflow should follow the recorded decision, rather than begin from an assumed approval milestone.
- Define scopeRecord jurisdiction, intended use and pathway assessment.
- Baseline EvidenceLink studies to their tested device configurations.
- Assemble applicationReconcile technical sections, manufacturing information and labeling.
- Review internallyResolve scientific, regulatory and quality questions.
- Track FDA interactionRecord filing status, correspondence and actual agency action.
- Assess changesConnect proposals to baselines and documented decisions.
- Maintain lifecycleAssign applicable conditions and unresolved follow-up actions.
06PMA supplements: define the change and support the decision
For US PMA-approved devices, 21 CFR 814.39(c)(1)(1)) limits supplement information to that needed to support the change. For US device PMA supplements, that provision also requires a separate section identifying each change for which approval is requested and explaining its reason.
The decision rule comes first. Under 21 CFR 814.39(a)), after approval "an applicant shall submit a PMA supplement for review and approval by FDA before making a change affecting the safety or effectiveness of the device", unless FDA has advised that an alternate submission is permitted under paragraph (e) or the change falls under paragraph (f). The same paragraph adds that "the burden for determining whether a supplement is required is primarily on the PMA holder", so the change-control record should show who made that determination and on what evidence.
For US device changes after PMA approval, check the current text of 814.39 for the applicable submission route and implementation conditions. Keep that route assessment separate from the task of assembling supporting information.
Worked case: component substitution
Assume the responsible team has determined that a proposed component substitution will be handled through a PMA supplement. The unit of analysis is the proposed device change, not the supplier or purchasing transaction. The following is a suggested evidence structure, not a route-selection rule.
| Trace element | Suggested entry |
|---|---|
| Approved baseline | Reference the approved device configuration and existing component specification. |
| Changed component | Identify the replacement and differences in material, dimensions, interfaces or performance. |
| Reason for substitution | Explain the assumed business reason: the existing component is being discontinued. |
| Affected evidence | Map those differences to potentially affected performance, compatibility and manufacturing evidence. |
| Supporting assessment | Link comparison results, testing and the rationale for retaining any existing evidence. |
| Requested change | Describe precisely which component substitution and device configurations the request covers. |
To keep the submission focused, organize supporting material around the differences and their consequences. Do not assume that an unchanged report remains applicable: compare its test article, conditions and conclusions with the proposed configuration. Where applicability is uncertain, record the gap and proposed resolution rather than treating the report title as sufficient support.
Keep the assessment actionable
A practical change-control record can hold the baseline reference, affected device configuration, rationale, evidence links, regulatory assessment owner and implementation decision. Keep the proposed change distinct from the authorized implementation state, with unresolved evidence questions assigned to named reviewers. This structure helps reviewers assess the requested difference without reconstructing the entire device history.
For manufacturing-related changes, connect this assessment to QMSR applicability and site responsibilities.
07Post-approval lifecycle: conditions, labeling and assigned actions
For US PMA-approved devices, check the applicable approval order and current 21 CFR Part 814 for post-approval conditions and reporting scope. Build the lifecycle worklist from those applicable obligations, rather than assigning blanket reporting intervals, universal post-approval studies or generic retention periods.
For US devices whose PMA approval is conditioned on specified labeling corrections, 21 CFR 814.44(d)(1)(1)) requires incorporating those changes exactly as directed and submitting a copy of the final printed labeling to FDA before marketing.
As a practical management control, maintain an obligation register with the following fields:
- Originating approval condition or applicable provision, with a reference to its controlling text.
- Affected device and labeling version, where relevant.
- Responsible person and the specific deliverable.
- Due-date basis, including the event that triggers the action.
- Supporting records and documented closure approval.
For labeling corrections, a useful closure package links each directed correction to the final artwork, printed labeling and submission record. Have the assigned reviewer compare the actual labeling against the direction, rather than close the action solely because artwork was revised.
Escalate unassigned actions and unresolved labeling discrepancies to the responsible regulatory and quality leads. An exception-based review queue can help focus attention on missing owners, disputed applicability and incomplete closure records. Use lifecycle metrics to expose unresolved work and prompt decisions—not to imply continued regulatory acceptability. Keep closure approval with an accountable human reviewer.
08Manufacturing and QMSR: product scope before site assumptions
For US medical-device manufacturers, check the current 21 CFR Part 820 for QMSR scope and incorporated references. Keep this assessment separate from the PMA marketing-review decision; use the FDA QMSR overview for manufacturing-system context, not an ISO clause guide or a legacy quality-system checklist.
For US medical-device manufacturers engaged in only some covered operations, 21 CFR 820.1(a)) limits applicable requirements to the operations they perform. For US manufacturers of HCT/Ps that are devices, 21 CFR 820.1(a)(4)(4)) expressly applies Part 820 to those manufacturers.
Map products, activities and accountable reviewers
As an operational practice, build a product-and-activity matrix connecting the application owner, manufacturing sites, outsourced operations, controlled device configuration and responsible reviewers. Give each handoff an identified record owner and receiving reviewer. Keep these assignments separate from documented conclusions about legal applicability: assigning a supplier task does not itself resolve the regulatory assessment.
- Identify the product configuration and activity covered by each row.
- Record which site or supplier performs the activity and who reviews its output.
- Link unresolved scope questions to a responsible regulatory reviewer before treating the matrix as settled.
For software support, set design objectives around controlled records, review assignments and traceable decisions, including the electronic release record. Keep regulatory assessment, approvals and device-release decisions with responsible people. Software should support their work, not be treated as automatic compliance or release. Carry site and supplier handoffs into the implementation and review checklist.
09Evidence model, system handoffs and exception recovery
The following entity model and workflow are suggested operating designs for PMA coordination, not FDA-prescribed systems. Assign an authoritative record for each field before designing synchronization.
| Authoritative record | Fields it owns |
|---|---|
| Document control | Report identifier, controlled revision, document approval status |
| Clinical evidence repository | Study identifier, dataset version, analysis provenance |
| Product configuration record | Configuration identifier, component and specification revisions |
| Manufacturing record | Build identifier, configuration used, execution results |
| Submission tracker | Submission identifier, evidence manifest, lifecycle action owner and status |
Suggested entity diagram: evidence → supports an identified configuration; submission → references controlled evidence revisions; change assessment → identifies the affected baseline; lifecycle action → traces to an originating condition or decision. Keep these relationships explicit rather than inferring them from filenames.
Proposed interfaces exchange these identifiers and links: clinical repositories link analyses to controlled reports; configuration records connect evidence to manufacturing builds; submission tracking captures revision-specific references. Apply MES–QMS integration principles without transferring field ownership.
Worked case: assume an interface sends a revised report identifier while submission tracking still references the previous revision. Unit of analysis: controlled record revision. Preserve the submission baseline; isolate the synchronization exception; reconcile identifiers and versions against authoritative records; obtain owner review; record the corrected linkage through an audit-trail review workflow. Do not silently overwrite the submitted record.
Route duplicate identifiers, broken references, missing ownership and inconsistent status mappings to exception handling. Technical synchronization success establishes neither scientific acceptance nor approval; those decisions remain with responsible human reviewers.
Device Configuration
Device baseline under evaluation.
- device_id
- intended_use
- configuration_id
- baseline_revision
Evidence Item
Scientific report and tested configuration.
- evidence_id
- study_type
- report_revision
- tested_configuration_id
Submission
Application history and agency action.
- submission_id
- submission_type
- filing_date
- fda_action
Change Assessment
Proposed change and human decision.
- change_id
- baseline_id
- affected_evidence_ids
- decision_owner
Lifecycle Action
Assigned follow-up and originating condition.
- action_id
- origin_reference
- owner
- due_basis
- closure_approval
10Application weaknesses and practical prevention controls
For US Class III device PMA applications, FDA’s data-requirements discussion associates incomplete, inaccurate, inconsistent or poorly organized applications, including those omitting critical information, with approval delays or denial.
The table is a recommended review aid, not a ranked list of FDA findings. Suggested owners should be adapted to the organization’s responsibilities.
| Weakness | Observable symptom | Suggested control | Decision owner |
|---|---|---|---|
| Unsupported scientific conclusions | Cited report does not substantiate the conclusion | Trace each conclusion to results and limitations | Scientific lead |
| Configuration mismatches | Tested and submitted versions differ | Reconcile versions using the earlier applicability case | Design lead |
| Inconsistent labeling references | Sections cite different labeling revisions | Check references against the submission baseline | Labeling lead |
| Unresolved change questions | Change impact or submission route remains undecided | Document the assessment through change control, as illustrated earlier | Regulatory lead |
| Unowned lifecycle actions | Action lacks an accountable person or closure evidence | Assign ownership and define closure criteria | Lifecycle lead |
Separate document presence from evidentiary adequacy: a missing report calls for retrieval or generation; an available report that does not support its attributed conclusion calls for scientific reassessment. Use a data integrity risk assessment where reliability or traceability is uncertain, rather than treating attachment completeness as proof of support.
11Metrics that reveal unresolved evidence and lifecycle work
Suggested internal measures: count open items per PMA/device configuration at a dated snapshot; track completed responses in elapsed days within a defined reporting period.
| Measure | Source record | Owner | Escalation purpose |
|---|---|---|---|
| Unresolved evidence-to-configuration links | Evidence matrix | Engineering | Resolve applicability |
| Unanswered scientific questions | Question log | Scientific lead | Secure expertise |
| Blocked internal review actions | Review workflow | Review lead | Remove dependencies |
| Unassigned lifecycle actions | Lifecycle register | Regulatory lead | Assign accountability |
| Unreconciled interface exceptions | Reconciliation log | System owner | Restore traceability |
| Response elapsed days | Correspondence register | Submission lead | Address handling delays |
Response elapsed time = response submission date minus request receipt date, in calendar days. This measures internal handling, not an FDA review clock. Use management review to resolve escalations; these measures are neither regulatory thresholds nor predictors of FDA approval.
12Implementation checklist: review the decision, not just the document
Use this recommended cross-functional review to expose unresolved decisions, not to certify compliance or predict approval.
- Regulatory: confirm documented jurisdiction and pathway assessments, coherent application references, current correspondence status and identified lifecycle conditions.
- Scientific: assess whether evidence applies to the controlled device baseline; frame unresolved scientific adequacy as a review question.
- Clinical: assess clinical evidence applicability and identify questions needing clinical judgment rather than administrative closure.
- Manufacturing: confirm the controlled device baseline and recorded manufacturing-change decisions through change control.
- Quality: check reconciled system exceptions and reference consistency through document control; ask who owns each unresolved decision instead of marking it complete.
For unresolved issues, return to the relevant earlier section: scope for pathway; clinical investigation or application contents for evidence; review process for correspondence; supplements for changes; post-approval lifecycle for conditions; manufacturing and QMSR for manufacturing scope; evidence integrations for system exceptions and handoff ownership.
Frequently asked questions
Q.Does PMA terminology refer to the application, FDA's review or the approval decision?+
Context matters. For US Class III medical devices, FDA’s PMA overview uses PMA for the scientific and regulatory review process and PMA application for the submission. FDA describes approval as a determination that sufficient valid scientific evidence assures safety and effectiveness for the intended uses. In project records, distinguish application preparation, review status and approval decision explicitly. An internal technical file index can help organize supporting material, but its title should not imply an FDA decision.
Q.Does every [Class III device](/glossary/medical-device-classification) follow the same submission route?+
No. For US Class III medical devices, FDA’s PMA overview notes that some preamendment devices may require a Class III 510(k). Consequently, a practical route assessment should identify the particular device type and regulatory history rather than treating the class label as a complete submission instruction. Record the basis for the selected route before committing the evidence plan, and use the 510(k) reference when investigating that exception. This is a planning practice, not a substitute for the device-specific assessment.
Q.What can happen at the end of the PMA review interval, other than approval?+
For US Class III device PMAs, FDA’s data requirements discussion identifies denial and describes application weaknesses that have caused delays. It also describes refusal to file for missing administrative elements; that occurs before in-depth scientific and clinical review, not at its conclusion. Operationally, distinguish an intake problem, an unresolved scientific question and an adverse decision instead of labeling everything pending approval. Track the specific issue, response owner and next action; consider the FDA Q-Submission reference when planning questions.
Q.What should a contract manufacturer hand over to the application owner for a useful evidence review?+
As a practical handover package, provide an indexed set connecting the device configuration to manufacturing instructions, process validation, acceptance methods, relevant production results, deviations and unresolved investigations. Include document revisions, approval status, record locations and access arrangements so reviewers can distinguish current evidence from superseded material. A batch execution history is useful when it connects test results to the actual manufacturing conditions. Identify gaps explicitly and assign technical contacts; transferring files alone does not explain their relevance or limitations.
Q.How can a team keep a PMA supplement focused on the [proposed change](/glossary/change-control) without overlooking affected evidence?+
For US devices with approved PMAs, 21 CFR 814.39(c)(1) limits supplement information to what supports the change and requires a separate section identifying each requested change and its reason. As a practical method, build an impact map from the changed configuration specification to affected performance claims, tests, manufacturing controls and labeling. For each connection, identify updated evidence or explain why existing evidence remains applicable. Keep unrelated history out of the narrative without omitting dependencies that affect the change assessment.
Q.Does a PMA approval decision settle the manufacturer's QMSR applicability assessment?+
No; assess manufacturing activities separately. For US finished devices intended for human use, 21 CFR 820.1(a) requires manufacturers engaged in the listed activities to establish and maintain a device-appropriate quality management system. That same US device provision limits a manufacturer performing only some covered operations to requirements applicable to those operations. A practical assessment therefore maps each organization’s actual activities rather than using PMA ownership as the deciding criterion. Keep that activity map with the QMSR assessment.
Q.Which application weaknesses should trigger an internal [scientific review](/glossary/quality-assurance-process) rather than a document correction?+
For US Class III device PMAs, FDA’s data requirements discussion warns that missing valid clinical information and sound scientific analysis can affect review and approval. As a practical triage rule, escalate unexplained contradictory results, unsupported extrapolation, questionable test representativeness or missing links between the tested and proposed device. A formatting correction cannot resolve those uncertainties. Use a data integrity risk assessment where reliability is in question, and have appropriate scientific specialists decide whether reanalysis, additional testing or narrower conclusions are warranted.
Q.Who should own unresolved lifecycle actions when responsibility or [records move between organizations](/glossary/document-control)?+
As a practical governance arrangement, name an accountable owner for each open action and obtain explicit acceptance from the receiving organization. The transfer register should identify the issue, affected device configuration, supporting records, next decision, dependencies and escalation contact. Separate custody of records from authority to decide whether an action is complete. Bring disputed ownership or inaccessible evidence to management review, and retain a visible handover trail until the receiving team confirms access and understands the unresolved work.
Primary sources
Explore this topic
PMA sits inside this topic cluster in our glossary. Every neighbour is one click away.
Device-specific rules, submissions and the standards that bind them.
Want to see how PMA could fit into your own records and workflows? Explore the related V5 pages or talk to our team about what applies to your operation.
