Engineering That Starts Where Other Projects Finish
An implantable device has no service port, no user-triggered firmware rollback, no battery door, and no acceptable failure mode. It sits in a 98.6 F (37 C) saline environment with an aggressive immune response for five, ten, or twenty years, and the only repair procedure is surgery. Every margin a consumer product treats as negotiable becomes fixed.
The regulatory consequence follows directly. Most implants are Class III and generally require premarket approval rather than clearance. PMA is a different proceeding, not a heavier 510(k): FDA evaluates whether there is reasonable assurance of safety and effectiveness on the strength of your own clinical evidence, with no predicate to lean on. The pathways are compared in 510(k) vs PMA; the practical difference is that a PMA program is measured in years and eight-figure budgets.
Materials That Survive Decades in Tissue
The implant materials palette is narrow because everything else has already failed somewhere. Titanium alloys for structural and hermetic housings. Platinum-iridium for electrodes. MP35N for lead conductors. Implant-grade silicone and polyurethane for flexible bodies. PEEK for load-bearing structures. UHMWPE for bearing surfaces. Parylene C as a conformal barrier.
Buy from a supplier who will provide a master file reference or full material traceability. A resin chemically identical to an implant grade but sold for industrial use is not a substitute, because you cannot document its additive package or lot history.
Biological testing scales with contact duration. Permanent contact with tissue or blood pulls in nearly the full ISO 10993 biocompatibility battery: cytotoxicity, sensitization, irritation, systemic toxicity, genotoxicity, implantation, and often carcinogenicity studies that run a year or longer. Chemical characterization and extractables work is the front door to that program, and it runs on finished, sterilized product, not raw resin. Budget 250,000 to 700,000 dollars and twelve to twenty-four months.
Hermeticity and Power for Life
Electronics inside the body need a barrier that keeps water vapor out for the device's entire life. Polymer encapsulation slows moisture ingress; it does not stop it. Active implants therefore use a welded titanium or ceramic enclosure with glass-to-metal or ceramic feedthroughs.
The acceptance criterion is a helium leak rate, commonly 1 times 10 to the minus 9 atm-cc per second or tighter, verified on every unit. That single requirement pulls in laser welding qualification, feedthrough vendor selection, internal getters to absorb residual moisture, and a leak test station on the line. It also constrains the industrial design, because the housing has to be weldable and testable.
Power is the other permanent constraint. Three approaches dominate: a primary lithium cell, usually Li-CFx or Li-I2, sized so the device reaches end of service before the cell does and paired with an elective replacement indicator; a secondary cell recharged inductively through tissue, which adds a coil, transcutaneous heating limits, and charge management that cannot run away; or no internal source at all, with power delivered inductively during use. Whichever you choose, the power budget drives the firmware architecture. Duty cycling, nanoamp sleep current, and telemetry scheduling are the design, not a late optimization.
Reliability When Repair Is Not an Option
Consumer reliability engineering asks how many units fail in warranty. Implant reliability engineering asks what the probability is that this specific unit survives a defined service life, and it must answer with data.
That means accelerated life testing at elevated temperature with a defensible model, real-time aging in parallel to validate it, mechanical fatigue testing to tens or hundreds of millions of cycles for anything that flexes with a heartbeat or a joint, and corrosion testing in accelerated saline. Leads and flexible interconnects are where most implant failures originate, and they get the harshest testing.
Redundancy and safe states matter more than performance. A hardware watchdog that forces a known-safe output, independent monitoring of any energy delivery path, and firmware that fails toward inactivity are all standard. This flows out of a formal ISO 14971 risk management process, which for an implant is the controlling document rather than a compliance artifact.
Sterilization, Packaging, and Manufacturing
Implants ship sterile, and the sterilization method has to be compatible with sealed electronics, batteries, and polymers simultaneously. Ethylene oxide cannot penetrate a hermetic housing, so it sterilizes the exterior only. Gamma and e-beam damage semiconductors and degrade some polymers. Steam is off the table for most active devices. Many implants end up using ethylene oxide with a validated exterior process, and the tradeoffs are compared in medical device sterilization methods.
The sterile barrier then has to hold for the full shelf life, validated by both accelerated and real-time aging with seal testing on aged samples. Any molded polymer component made in a controlled environment brings in cleanroom injection molding and its own qualification burden.
Manufacturing for a PMA product is inspected before approval. FDA conducts a preapproval inspection of the site, and the quality system must be operating with real records, not a binder written the month before. Your titanium supplier, feedthrough vendor, contract sterilizer, and molder all become auditable parts of your file, with change notification agreements in place.
Clinical Evidence and the Regulatory Clock
A PMA needs clinical data from a study of your device. That study is a significant risk investigation, so it requires an Investigational Device Exemption approved by FDA and by each site's institutional review board before the first patient is enrolled. The design, endpoints, sample size, and follow-up duration are negotiated with the agency, ideally through a pre-submission meeting long before the protocol is written. The mechanics are covered in IDE studies and clinical trials for devices.
Typical shape: a feasibility study of ten to thirty patients, then a pivotal study of one hundred to several hundred with follow-up of one to five years. FDA's review of the resulting PMA runs 180 days on the clock and much longer in practice, and may include an advisory panel. Approval carries conditions, usually a post-approval study and annual reporting.
Whole-program numbers are sobering. Seven to twelve years from concept to approval and 30 to 150 million dollars is the honest range for a novel active implant. Even at the low end it is an order of magnitude beyond the figures in typical medical device development costs.
How Small Teams Actually Make Progress
Nobody funds that in one round. The workable strategy stages it around evidence that changes the risk profile: bench proof of the core mechanism, a large-animal study demonstrating the effect and the surgical approach, a pre-submission meeting producing written FDA feedback on the evidence plan, then a first-in-human or early feasibility study. Each milestone is a financing event and a genuine off-ramp.
Design freeze discipline matters more here than anywhere else. Every material change after biocompatibility testing repeats the testing, and every geometry change after animal work weakens the data.
Get an Honest Read on an Implant Program
Projects House works with teams developing implantable and other high-risk devices on the engineering that decides feasibility: material and hermeticity strategy, power architecture, reliability evidence, and a staged plan that lines up technical milestones with financing. Send your device concept and intended indication through our contact form.