Why shared auto-injectors keep forcing rebuilt device files

Fubin Wu arrived at a Parenteral Drug Association (PDA) conference with a presentation to give the next day and a cost he had watched repeat across pharmaceutical companies, device platform vendors and consulting firms. One auto-injector serves several drug products, and each new pairing rebuilds the device file almost from scratch. His talk, on Module 3 submissions with low-variability design control and risk management, was aimed at the people who pay for that rebuild.

Wu, co-founder of Vertex Knight, a digital solution and consulting company, was speaking in a MarketScale Studio interview at the conference. Module 3 is the quality section of a marketing application, the chapters that document how a drug product is designed, manufactured and controlled, and for a combination product, a drug sold together with the device that delivers it, those chapters cover the device as well. By Wu's account, that section is where the repeated work shows up when a device is shared.

The complaint is old. The timing is new. In a November/December 2025 article in Pharmaceutical Engineering, the magazine of the International Society for Pharmaceutical Engineering (ISPE), Alan Stevens, Kent Abrahamson, Mahendran Ravichandran and Amy Wise wrote that platform devices have been discussed for years as a route to cost and time savings and have still not won widespread adoption from industry or regulators.

No regulatory pathway currently exists to review or approve a platform device, the ISPE authors wrote. Their article instead sets out how a platform device could support approval of a combination product, and it names FDA's May 2024 draft guidance on platform technology designation among three regulatory sources on platform use. That guidance, as the agency titled it, is for drug development.

Why a shared device still means a rebuilt design file

Wu's own example sets the scale of the waste, and it is the line his talk was built to attack.

ISPE's authors describe the same repetition as duplicative but necessary for generating safety and performance data on each combination product. The reason sits in design control, the documented chain a device maker keeps from design inputs, meaning what the device must do, through design outputs and verification, meaning tests showing the built device meets those inputs, to validation, meaning evidence that it works for the intended user. That chain lives in a medical device file, the term the ISPE article uses in place of design history file because of harmonization with ISO 13485:2016, the standard that sets quality management system requirements for medical device makers.

A new application wants that evidence for its own drug and device pairing, and much of the repetition is self-inflicted, by the same article's account. Manufacturers require redundant design control activities such as design inputs and design verification, which weakens the business case for a platform before a regulator ever sees it. Repurposing an older device without a plan to reuse its data carries its own risk, the authors add, because the original design can be stressed beyond its performance range and the current engineering team may not understand earlier design decisions.

The pressure is building from the product side as well. In a June 2022 editorial in Therapeutic Innovation & Regulatory Science, Robin Kumoluyi and Amit Khanolkar of Janssen, the pharmaceutical companies of Johnson & Johnson, wrote that combination products have matured into a market category and traced expanding adoption of medical devices for drug delivery and increasing complexity among the resulting trends. For a regulatory affairs lead at a pharmaceutical company or a contract development and manufacturing organization (CDMO), that means more device-bearing applications, each with a design control and risk file to defend.

Digitalization means structured data, not electronic documents

Wu calls the first trend a digital revolution and defines it narrowly.

The distinction matters in practice. Electronic documentation means specifications, test reports and submission chapters stored as files in a document system, where a related change is found only by someone reading for it. A digital system, as Wu describes it, holds product information as connected data, so a requirement, the test that verifies it and the submission paragraph that reports it point to one another.

Vertex Knight's own product follows that definition. Wu says in the interview that the company's digital solution connects product information so that it enables reuse, platform management and lifecycle change impact analysis, and his presentation applied the linkage to Module 3, with the submission narrative tied directly to the engineering data behind it.

That link is the mechanism behind lifecycle change impact analysis: when a part, supplier or specification changes after approval, the system identifies the tests, risk entries and submission sections the change touches. Wu does not name the database, software stack or data standard behind the Vertex Knight system in the interview, and he does not compare it with product lifecycle management (PLM) software, which engineering teams use to track part and document revisions, or with the electronic common technical document (eCTD), the file format in which regulators receive applications.

Digitalization carries a wider meaning across the device market. In its 2025 Global Drug Delivery System Trends Report, Alira Health describes delivery systems evolving from standalone mechanical tools into connected products with reminders, real-time data capture and remote care links, and it reports Riccardo Butta, president of health solutions at Flex, saying that fitting sensors into a device requires designing a complete connected system across the product lifecycle, from manufacturing and distribution through patient use and end of life. Wu's version sits upstream of all that, in the data a company holds about a device before a single unit ships.

Syntegon applies a related logic on the equipment side. On its own site, the company says it helps customers structure user requirement specifications, batch ranges and product families so that formulation room concepts scale from clinical to commercial batches, which it says reduces redesign effort in later project phases.

FDA's platform designation covers the drug while the device waits

FDA's May 2024 draft guidance sets out a platform technology designation under section 506K of the Federal Food, Drug, and Cosmetic Act, the provision that lets the agency designate a technology for reuse across drug products. Issued by the Center for Drug Evaluation and Research and the Center for Biologics Evaluation and Research, the draft lays out eligibility factors, potential benefits, how a sponsor can draw on data from a designated platform, what a designation request should contain, how to raise a planned request at a milestone meeting and how long FDA takes to evaluate one. It also covers revocation of a designation and postapproval changes to a designated platform technology, and its stated purpose is efficiency in development, manufacturing and review for drug product applications that incorporate a designated platform.

FDA adds a caution that cuts both ways for device teams. In the same draft, the agency acknowledges that both industry and FDA have used the term platform technology more loosely than the statute defines it, so some technologies long called platforms may not meet the statutory definition or eligibility factors. The draft is non-binding, was opened for a 60-day comment period, and will represent the agency's current thinking only once finalized.

ISPE's authors list that draft and the European Medicines Agency's May 2019 draft guideline on quality requirements for drug-device combinations as two of the three regulatory sources on platform use. The obstacles they name are investment cost, the risk that a regulator declines platform data in a clinical or marketing application, and confusion over what the word platform means, between industry and regulators and between industry partners. That confusion is the hardest part of Wu's thesis to argue past, since reuse a company can prove to itself still has to be read the same way by a reviewer.

Wu's digitalization argument leaves the regulatory obstacle in place and targets the duplication the ISPE article calls self-imposed, the repeated design inputs and verification a company runs because its own records make reuse hard to prove. The ISPE authors set out what closing the regulatory gap takes: agreement on the benefits of a platform approach, confidence in device performance, less workload for reviewers and manufacturers, and faster review of platform data.

Large-volume delivery already accounts for 15% of injectable biologics

Wu's second trend has a count behind it, and he put it last when he summed up his list.

ONdrugDelivery counted 182 large-volume subcutaneous biologics in January 2025, about 15% of IV and SC agents. Writing in the publication's issue 168, Andreas Schneider and Jakob Lange defined large volume as a maintenance dose above 2 mL, reviewed 1,338 approved and clinical-stage agents, and found the 182 were predominantly monoclonal or bispecific antibodies, antibody drugs engineered to bind one target or two, aimed at cancer and at chronic autoimmune, neurological and cardiovascular conditions. Wu says large-volume delivery seems preferred; the review puts a number on the preference.

Subcutaneous, or SC, means injected into the fatty layer under the skin rather than into a vein. Intravenous (IV) infusions, the dominant route for biologics, are usually given in acute-care settings because of safety concerns and doses too large for an SC injection, and they can run for hours, according to the ONdrugDelivery review, which is why moving a dose under the skin shifts care toward outpatient clinics and homes. Handheld auto-injectors today handle up to 2 mL in under 15 seconds, the review states, and doses from 2 mL to more than 30 mL move to on-body injectors, wearable pump devices that adhere to the skin, or to tethered syringe pump drivers.

Getting a large dose into that tissue relies on permeation enhancers such as hyaluronidase, an enzyme that temporarily opens the tissue matrix so fluid can spread, and on high-concentration formulations, the review states. The payoff the authors point to is DARZALEX FASPRO, Janssen's subcutaneous daratumumab, which replaced an intravenous formulation needing hours-long infusions with an injection completed in under five minutes.

A July 2026 review in Advanced Drug Delivery Reviews, published by Elsevier, shows how fast the ceiling moved. Andrea Allmendinger, Hanns-Christian Mahler and Beate Bittner write that fast bolus SC injections, meaning a dose pushed in at once rather than infused, were traditionally limited to 1 mL or at most 2 mL, and that volumes well above 2 mL, up to 10 mL and even 25 mL, are now in development, in clinical trials or on the market. Their own shorthand names the two routes a developer can take, a high-concentration, low-volume product or a low-to-mid-concentration, high-volume one, and they treat that choice as bound up with device selection across the whole product lifecycle.

The hypodermis itself is the obstacle, by that review's description: a layer of fat and collagen matrix that acts as the primary barrier to diffusion and carries a net negative charge that affects how therapeutic proteins move, with small molecules taken up directly by blood capillaries and molecules above 16 kilodaltons, a unit of molecular mass, absorbed differently. The review catalogs the current technical challenges of high-volume SC products rather than a settled method. That is what separates a large-volume device from a standard auto-injector: the device becomes a pump or wearable, and the formulation has to be concentrated or helped through the tissue.

Personalization is the trend with the thinnest numbers

Wu added his third trend after naming two, calling it procedure medicine or individualized treatment delivery and saying it is gaining market interest. He does not describe how an individualized treatment changes device requirements beyond the dose itself, and PDA's October 2025 conference program does not list personalization among its key topics.

Alira Health's 2025 report ties personalization to connectivity instead. Digital delivery systems capture adherence patterns, symptom evolution, injection timing and dose accuracy, the report says, and that data supports treatment personalization, post-market surveillance and real-world evidence, while electronic dose control and automatic logging cut the risk of under- or overdosing. Adoption is concentrated in diabetes, GLP-1 therapies, pediatric growth disorders and chronic injectables for rare diseases, according to the report.

PDA's own framing of the field matches the rest of Wu's list. The program for its Universe of Pre-Filled Syringes and Injection Devices Conference, held 21 to 22 October 2025 in Vienna and co-chaired by Theresa E. Bankston and Ypsomed's Jakob Lange, named digital health for usability and adherence, large-volume injections, wearable devices and new platform technologies, AI-driven testing, and sustainability through device reuse and lifecycle management as its key topics. Each of those arrives as a variant or a post-approval change on a shared platform, which is the lifecycle change impact analysis and management Wu says his company's system handles.

What a combination product team can do before a pathway exists

ISPE's prescription is to plan the platform from the start rather than repurpose a device later. The authors separate prospective platform development, where device data is generated once with reuse in mind and later products reference the platform medical device file, from the long-standing practice of reusing data from one approved combination product in the next, which they call not novel and leave out of scope. The benefit, they write, arrives with each future product that references the platform file instead of building its own.

Put that advice next to Wu's and a device team needs two things in order: a platform file designed for reference, and a data structure that can prove which parts of it a new product touches.

Two signals will tell a regulatory team whether that combination is arriving. One is FDA finalizing the May 2024 draft, which the agency says will then represent its current thinking on platform designation. The other is the one the ISPE authors set, a regulator accepting platform device data in a clinical or marketing application, and until then Wu's benchmark stands: change 1% of the file when 1% of the product changes, and reuse the knowledge base the organization has already built.