FDA Expectations for Biocompatibility and Ethylene-Oxide Sterilization Residuals in Premarket Submissions
Biocompatibility testing and ethylene-oxide (EO) sterilization residuals are among the most common sources of deficiency letters in device premarket review. Because FDA's expectations are anchored in recognized consensus standards — ISO 10993-1 for biological evaluation and ISO 10993-7 for EO residuals — submission teams that misjudge the required scope of testing, or the residual limits that apply to their device's contact category, routinely lose review cycles to additional-information requests.
The analysis below walks through the framework FDA reviewers apply: how ISO 10993-1 situates biological evaluation within risk management, how body-contact nature and duration drive the endpoint selection, and what cleared submissions show about the data FDA expects for EO-sterilized devices, including residual testing against ISO 10993-7 limits.
Want to ask Rhizome your own regulatory questions? Try it for free.
FDA expectations for biocompatibility and ethylene-oxide sterilization residuals in premarket submissions
Biocompatibility and sterilization-residual data are two of the most frequently deficient areas in device premarket submissions, and both are governed by a well-defined set of FDA-recognized consensus standards. FDA does not expect a fixed battery of tests. It expects a risk-based biological evaluation keyed to how, and for how long, the device contacts the body, plus evidence that any ethylene-oxide (EO) sterilization process leaves residuals below the limits in ISO 10993-7. This overview lays out the framework a reviewer applies and the specific data expectations seen in cleared submissions.
The governing framework: ISO 10993-1 within a risk-management process
FDA's biological-evaluation expectations are built on ISO 10993-1, which FDA recognizes as recognition number 2-258 (fifth edition, 2018-08, partial recognition, entered 2019-01-14) 52. The standard, and FDA's recognition of it, frame biological evaluation as a process inside the device's overall risk management rather than a checklist of animal tests 5211.
In practice that process requires the submitter to 5211:
- apply the general principles of biological evaluation;
- categorize the device by the nature and duration of its body contact;
- gather and evaluate existing relevant data from all sources;
- identify gaps in the data set through a risk analysis;
- determine what additional data are needed; and
- assess the overall biological safety of the finished, sterilized device.
The evaluation is expected to consider biological risks arising from device constituents and from tissue-device interactions, including physical effects, across the device life cycle 11.
The FDA guidance document most often cited in cleared 510(k) and De Novo decision summaries is "Use of International Standard ISO 10993-1, 'Biological evaluation of medical devices - Part 1: Evaluation and testing within a risk management process'" 252253255264270. Older submissions reference the predecessor FDA Blue Book Memorandum #G95-1, which carried similar ISO 10993-1 language 261263271275. Reviewers also expect sample preparation and extraction to follow ISO 10993-12 where relevant 255256262275.
Categorization drives the endpoints
The single most important determinant of which endpoints FDA expects is the device's contact category, defined by both the type of tissue contacted and the duration of contact 5211125127140.
Contact type, as reflected in decision summaries, falls into the familiar ISO 10993-1 groupings 125127140:
- Surface-contacting devices (intact skin, mucosal membrane, breached/compromised surfaces).
- External communicating devices (contact with the bloodstream, tissue/bone/dentin, or an indirect blood path).
- Implant devices (contact with tissue/bone or with circulating blood).
Contact duration is classified as 125127140:
- Limited — up to 24 hours;
- Prolonged — greater than 24 hours to 30 days; and
- Permanent — greater than 30 days.
If a device has no patient contact, biocompatibility testing is stated as not applicable and the evaluation ends there 125. Otherwise, the contact category selects the relevant endpoints from the ISO 10993-1 matrix.
Endpoints FDA reviewers expect by category
Across cleared submissions, reviewers expect the endpoint set to scale with the invasiveness and contact duration of the device 123125127140:
- Surface contact, prolonged (intact skin): cytotoxicity, sensitization, and irritation 125.
- External communicating, blood-contacting, limited (<24 h): cytotoxicity, sensitization, irritation/intracutaneous reactivity, and acute systemic toxicity 127.
- Permanent implant, blood-contacting: cytotoxicity, sensitization, irritation/intracutaneous reactivity, acute systemic toxicity, hemocompatibility, genotoxicity, and implantation, with complement activation, thromboresistance, and pyrogenicity added depending on the device 140142.
- Blood-path or catheter devices generally: hemocompatibility (hemolysis, complement activation, thrombogenicity, coagulation and platelet/leukocyte measures) and material-mediated pyrogenicity are commonly included 132142146.
The endpoints that recur most often in decision summaries are cytotoxicity, sensitization, and irritation/intracutaneous reactivity (nearly universal for patient-contacting devices), followed by acute systemic toxicity, material-mediated pyrogenicity, genotoxicity, implantation, and hemocompatibility as the contact category escalates 122126128135138140142144146. Reviewers expect the submission to present this as a table listing the device/contact category, contact duration, each endpoint, and the test conclusion or the justification for not testing it 122125127140146.
Chemical characterization and toxicological risk assessment as an alternative to testing
FDA increasingly accepts chemical characterization under ISO 10993-18 paired with toxicological risk assessment under ISO 10993-17 to address, and in some cases replace, direct biological testing, particularly for the harder-to-test systemic endpoints 123135201212. In cleared submissions, acute systemic toxicity, genotoxicity, subacute/subchronic toxicity, chronic toxicity, and carcinogenicity are frequently "evaluated through" or "supported by" chemical characterization plus toxicological risk assessment rather than separate animal studies for each endpoint 198201207212214217.
Reviewer expectations for this pathway include 195201208209218:
- Identify and, where necessary, quantify extractables and leachables from the device materials under ISO 10993-18, including manufacturing- and sterilization-introduced substances 1618201.
- Use extraction conditions representative of clinical use or a conservative worst case (for example, aggressive conditions with both polar and non-polar solvents) 195.
- Analyze extracts with appropriate methods such as GC-MS, GC-FID, LC-HRMS, HS-GC-MS, and ICP-MS 196201.
- Subject compounds detected above the reporting/quantification threshold to a toxicological risk assessment under ISO 10993-17 to show that exposure is without appreciable harm 5208.
FDA also recognizes ISO/TS 21726 on applying the Threshold of Toxicological Concern (TTC) to biocompatibility (recognition 2-268, complete) 57, which can support a risk assessment when a substance's exposure is below a qualified threshold. Note that FDA has not published a single universal analytical evaluation threshold; the threshold applied is case-specific and must be justified in the submission 208195.
Ethylene-oxide sterilization residuals: ISO 10993-7 and ISO 11135
For EO-sterilized devices, FDA expects two linked things: a validated sterilization process and confirmation that post-aeration residuals fall below the ISO 10993-7 limits 160162166168.
Process validation. EO sterilization is developed, validated, and routinely controlled under ISO 11135, which FDA recognizes for both industrial and health-care-facility settings 229. Decision summaries pair residual data with sterilization validation demonstrating a sterility assurance level (SAL) of 10⁻⁶ or equivalent 160162166168171177181.
Residual limits. ISO 10993-7 (FDA recognition 2-275) specifies allowable limits for residual EO and ethylene chlorohydrin (ECH) in EO-sterilized devices, the procedures to measure them, and the methods to determine compliance for product release 159230231. The standard applies to devices with patient/body contact and excludes devices with no patient contact, such as in vitro diagnostics; the current revision also excludes components with neither direct nor indirect body contact and devices that do not absorb or retain EO/ECH 159230231. Limits are keyed to the same contact-duration categories used for biocompatibility. The values cited in FDA decision summaries are:
| Contact category | Ethylene oxide (EO) | Ethylene chlorohydrin (ECH) |
|---|---|---|
| Limited (≤24 h) | 4 mg per device | 9 mg per device 162168170171177181 |
| Prolonged (>24 h–30 d) | Average ≤ 0.1 mg/day; not to exceed 20 mg in 24 h and 60 mg in 30 days | Average ≤ 2 mg/day; not to exceed 12 mg in 24 h and 60 mg in 30 days 246286 |
| Permanent (>30 d) | Average ≤ 0.1 mg/day; not to exceed 20 mg in 24 h, 60 mg in 30 days, and 2.5 g lifetime | Average ≤ 2 mg/day; not to exceed 12 mg in 24 h, 60 mg in 30 days, and 50 g lifetime 246 |
Reviewers apply these limits to the specific device, and device-specific presentations appear as well. Some summaries express limits on a surface-area basis as a tolerable contact limit (for example, EO TCL and ECH TCL per cm²) 172, and pediatric/neonatal devices carry lower daily allowances derived under the ISO 10993-7 2019 amendment (for example, EO ≤ 0.21 mg/day or 0.30 mg/kg/day and ECH ≤ 0.19 mg/day or 0.27 mg/kg/day for neonates/infants) 164. Because EO and ECH carry genotoxic and carcinogenic concern, the residual assessment is treated as part of the overall biological safety picture rather than a stand-alone box to check.
Testing and aeration. Decision summaries show residuals verified against ISO 10993-7 limits after a defined aeration period (for example, 3, 9, or 10 days), either as part of EO cycle and aeration validation or through residual data confirming the limits will not be exceeded under worst-case conditions 161162163167169179. The evaluation is expected to demonstrate that the device meets its EO/ECH limit at the point of patient use.
Recognized standards commonly used
Beyond ISO 10993-1, -7, -17, and -18, FDA's biocompatibility recognition list (specialty task group "Biocompatibility") includes a broad set of test-method and characterization standards that submitters draw on to satisfy individual endpoints 565758. Frequently used entries include ISO/TS 10993-19 on physico-chemical characterization (recognition 2-281) 58, and a range of complete-recognition ASTM methods for specific endpoints, such as F756 hemolysis, F763 intramuscular implantation screening, F749 intracutaneous injection, F750 acute systemic toxicity, F813 direct-contact cytotoxicity, and F895 agar-diffusion cytotoxicity 7071726867. USP biological reactivity chapters <87> and <88> and the USP <151> pyrogen test are also FDA-recognized 63646261. Declaring conformity to the recognized version of the applicable standard is the cleanest way to align a submission with reviewer expectations, but conformity to a standard does not remove the obligation to document the risk-based rationale under ISO 10993-1.
What this means for a submission
A reviewer reading a biocompatibility and sterilization package expects to see, at minimum: a clear statement of the device's contact type and duration category; a biological-evaluation plan and endpoint matrix tied to that category under ISO 10993-1; test reports or a chemical-characterization-plus-toxicological-risk-assessment justification for each endpoint; and, for EO-sterilized devices, a validated sterilization process to SAL 10⁻⁶ under ISO 11135 with post-aeration EO and ECH residual data below the ISO 10993-7 limits for the applicable category 5211140229246. The most common weaknesses are an endpoint set that does not match the stated contact category, testing performed on a material coupon rather than the finished sterilized device, and residual data that are not tied to the worst-case aeration and use conditions 125140162. Aligning the evaluation to the contact category up front, and documenting the rationale for every endpoint addressed or omitted, is what turns a data package into a defensible biological-safety conclusion.