Abstract: When one high-concentration biologic enters ten companies' pipelines, it creates demand for ten delivery devices, ten sets of registration data and ten supply chains. The autoinjector export business is shifting from "quote, sample, ship" to "verify, document, stay locked into the medicine's entire product life."
Ⅰ. In the last two years, the opening line of an overseas enquiry has changed
More and more of the high-viscosity enquiries we receive now open like this:
"We need a 2.25 mL autoinjector. Drug viscosity is roughly 35-40 cP at 25 C (low shear), target injection time below 15 seconds, target markets the EU and the United States. Please quote together with the following: design verification reports to ISO 11608, injection performance data at end of shelf life (36 months), the device technical documentation package needed to support MDR Article 117, and your change control procedure."
Five years ago, the first email from the same kind of customer usually carried just three lines: do you have existing tooling? What is the minimum order quantity? What is the unit price?The difference is not rhetorical. It decides how you price the job, how many rounds of samples you build, and which clauses have to be written into the contract.What pushed enquiries into this shape is something quite specific happening on the formulation side.
Ⅱ.On the demand side: viscosity was not designed in - it was squeezed in by competition
The most direct force behind demand for high-viscosity delivery devices is not a technology. It is the wave of biosimilar launches.
- Adalimumab (Humira): after loss of exclusivity in the United States in 2023, more than ten biosimilars have entered the market.
- Ustekinumab (Stelara): peak sales of around USD 10.9 billion in 2023, loss of exclusivity in the US and EU in 2025, and more than six biosimilars approved to date.
- Pembrolizumab and nivolumab: expiry around 2028.
- Dupilumab and daratumumab: around 2030.
(The figures above come from third-party industry reports and market reviews; the estimates vary considerably between sources, so use them as a trend indication rather than a precise forecast.)
One inference matters enormously for the device industry: when a single molecule enters ten pipelines, it takes ten delivery devices, ten sets of registration data and ten supply chains to support it.More importantly, once the molecule itself can no longer differentiate, competition moves down to the injection experience: less frequent dosing, a finer needle, a shorter injection time, better stability. All four land squarely on the combination of the autoinjector and the prefilled syringe (PFS).On the device side, a few hard constraints cannot be engineered around. Put them on the table in the very first email of the project.

1. Viscosity
The working rule of thumb is that the comfort zone of a standard autoinjector sits below 20 cP. Once you pass 50 cP, it is usually no longer a question of "fitting a stronger spring"; you are looking at reformulating the drug product or moving to a different container system. Ardena sets out a comparable range in its review of PFS and autoinjector compatibility.
2. Time
As a rule, patients do not want to hold a device against the skin for more than 15 seconds, and that window has to cover needle insertion as well as the injection itself.
3. The needle
Under the Hagen-Poiseuille relationship, plunger force scales with the fourth power of the needle's internal diameter. Swap a 27G needle (bore approximately 0.191 mm) for a 30G (approximately 0.140 mm) and you need roughly 350% more force to deliver within the same injection time.
That is why "can you make it thinner" and "can you make it deliver" have always been two ends of the same equation, never two requirements that can be conceded separately.
4. Volume
Conventional PFS-based autoinjectors top out at a practical fill volume of about 2.25 mL. Going beyond that calls for a different primary container and a different drive concept - a limit Gerresheimer states explicitly in its public white paper on the Gx Inbeneo platform.Put simply: formulations are getting harder to deliver, so the device has no choice but to get harder to engineer.
Ⅲ.The business model has changed: you are not selling a pen, you are selling a fixed parameter inside a registration dossier
This is the point most easily buried under technical discussion, and the one that matters most to anyone in export.
In the EU, for integral drug-device combinations (IDDCs) such as prefilled autoinjectors, the device constituent must demonstrate conformity with the General Safety and Performance Requirements (GSPRs) of MDR Annex I. Where the device constituent itself requires notified body involvement - typically the case for sterile PFS presentations, dose-measuring functions or higher risk classes - the route is the Notified Body Opinion (NBOp) under Article 117.

The real impact arrives after approval. EMA's Questions and Answers on medicinal products used with medical devices (EMA/37991/2019 Rev.6, updated December 2025) makes one thing clear: any change to the device must be reassessed to determine whether a new NBOp or GSPR declaration of conformity is required, and submitted through the corresponding variation procedure. The revised Variations Guidelines apply from 15 January 2026 and introduce a dedicated Q.IV category for device-related changes.
In other words, once your pen is written into a customer's marketing authorisation dossier, you are no longer just a supplier - you become a parameter in their registration file that cannot be moved freely. The barrier to entry rises steeply: what customers want is not samples, but a complete technical documentation package ready for submission. The flip side is that this creates very high switching costs. A high-viscosity project running in steady supply tends to last for the whole life of that medicine - three years, five years, sometimes eight.
Ⅳ.The three checklists every overseas buyer eventually asks for
Across the projects we have handled over the past two years - European, Indian or Brazilian customers alike - the technical conversation always converges on three sets of questions. Checklist A: performance and verification - will it deliver the full dose?
Provide data against the design verification matrix of ISO 11608-1 and ISO 11608-5: dose accuracy, activation force, needle-shield lockout, injection time distribution. Life-cycle verification is normally run at 1.5 times the maximum intended number of actuations. Then the complete pre-conditioning and stress matrix - ambient, high and low temperature, damp heat (for example 40 C / 93% RH for 96 hours), cold storage, temperature cycling, vibration, 1 m free fall - with a repeat test after each condition. Increasingly, what customers actually want is the raw data together with the statistical treatment applied to it.
Performance at end of shelf life (say 36 months): injection time, residual volume and activation force still within specification. This is the requirement that separates high-viscosity projects from ordinary ones, and the hardest to backfill later. Then the cold-start case: injecting straight out of the refrigerator versus injecting after equilibrating to room temperature, and how much difference there is. This is routinely left out of a customer's Target Product Profile (TPP), yet it happens every day in real use.

Checklist B: primary container compatibility - what is left once the dose has been delivered?
PFS selection: flange design (round versus cut flange), barrel material, strength margin. Plunger stopper: consistency of gliding behaviour for low-friction or coated stoppers. Batch-to-batch variation shows up directly in the spread of injection force, so "same catalogue number" is not good enough. Siliconisation: silicone oil level, uniformity of application, adhesion stability. Containers using cross-linked (baked-on or otherwise cured) siliconisation carry significantly lower risk of silicone oil droplet migration and the protein aggregation it triggers - a point repeated throughout the published technical literature of several primary packaging suppliers.
Needle specification: for high-viscosity products the aim is to maximise the internal bore while preserving acceptable insertion feel, so 27G thin-wall or ultra-thin-wall is the usual landing point rather than the standard 29G. But it also moves the pain assessment, the needle description in the registration dossier and the clinical instructions for use, so it has to be locked down at RFQ stage and must never be left open until design verification. Compatibility and integrity: extractables assessment, sub-visible particles, container closure integrity (CCI). These are hard gates at the registration stage, and the later you close them, the more expensive they become.
Checklist C: quality system and supply chain - can you ship consistently for five years?
ISO 13485 is the floor here, not a differentiator. On the US side, one change landed in 2026: FDA's Quality Management System Regulation (QMSR) took effect on 2 February 2026, incorporating ISO 13485:2016 into 21 CFR Part 820 and retiring the QSIT inspection technique in favour of Compliance Program Manual CP 7382.850. FDA's inspection logic for device quality systems has moved onto ISO 13485. For Chinese suppliers serving both the EU and the US this is good news - one documentation set serves both markets - but it also means anything drafted in the idiom of the old QSR has to be re-mapped onto the ISO 13485 clause structure.
Article 117 requires a notified body, and notified body capacity is routinely cited by practitioners as the real bottleneck: the constraint is not how difficult the assessment is, but how long it takes to schedule. For projects involving an NBOp, the lead-time commitment has to leave room for that interval, which lies outside your control, and the contract has to state where responsibility sits. Capacity and tooling: tooling ownership, capacity commitments, a second source, and the notice period for supply interruption or product discontinuation. European customers take discontinuation clauses very seriously; a requirement of 12 to 24 months' prior notice is not unusual.
Ⅴ.Four traps that most often derail a quotation
Trap one: a viscosity value without a viscosity curve.
The same formulation can differ by a factor of two or more between 5 C and 25 C. You have to pin down the measurement temperature, the shear rate, and whether it has been retested under accelerated conditions or at end of shelf life. This one item decides whether the drive parameters you quote are real or guesswork.
Trap two: "we already have a PFS."
There is far more to check than fill volume: flange design, barrel material and supplier, stopper type and lot number, staked or removable needle, dimensional tolerances in relation to the plunger stroke. Quote on the strength of "1 mL or 2.25 mL" alone and rework later is close to guaranteed.
Trap three: "build us 30 prototypes first."
Prototypes for a high-viscosity project usually mean reworking the drive system, possibly new tooling or re-machined components, plus a representative verification build. It is not a catalogue pen with a different label. Spell out the workload, timeline and cost of the prototype phase up front: being explicit wins orders more reliably than vague agreement, and it keeps the relationship from breaking down further down the line.
Trap four: verbally promising that one device covers both low and high viscosity.
It can usually be done technically, but only if you first define the verification matrix at both ends of the range, how the SKUs are distinguished, and the differences in labelling and instructions for use. Otherwise you either sacrifice performance at one end or end up in an endless loop of rework.
Ⅵ. For Chinese suppliers: where the window is
In all honesty, the barriers to high-viscosity projects at large Western innovator companies are severe: full NBOp support, multi-year quality agreements, mandated imported primary packaging. Getting in on the first attempt is very hard.
There are three more realistic openings. First, biosimilar developers. Companies in India, Korea and China itself run several pipelines in parallel, respond better to a package of "device plus local technical service plus reasonable MOQ," and have far shorter decision chains than multinationals. Second, localised manufacturing in emerging markets. Brazil, the Middle East and South-East Asia are pushing local fill-finish and local delivery; regulatory pressure there is comparatively manageable, while the demands on delivery cadence and field support are higher - precisely where Chinese suppliers tend to be strongest. Third, delivery upgrades on products already on the market. Upgrading the delivery system of an existing product to extend its life cycle delivers results faster than a brand-new programme, with clearer objectives and an identifiable budget.
Three things are worth doing first, and none of them are glamorous. First, build verification capability in-house: mechanical testing, dose accuracy, injection time, ageing chambers. Being able to hand over raw reports within two days beats any brochure. Second, turn your technical parameters into a data package you can hand out, stating what viscosity you have qualified, what fill volume and which needle, what range the injection time falls in, and how end-of-shelf-life data were generated. Only verifiable parameters carry weight. Third, write change control into the contract: change notification lead time, the definition of a significant change, the conditions for a second source, and discontinuation terms. It reads like a legal matter, but it is what protects a long-term order.
High-viscosity formulations have not made the autoinjector business more sophisticated. They have made it more specific - specific down to one cP value, one needle bore, one 36-month ageing data set, one Q.IV variation category, one sentence of discontinuation notice in a contract. The dividing line was never who uses the loudest adjectives. It is who can pin every one of those variables down in the first week after the enquiry arrives, price every one of them into the quotation, and reproduce each of them without deviation on the production line.
