In ophthalmology, the molecule is no longer the hard part. The delivery system now decides whether a drug reaches the back of a patient’s eye and stays there, and the programs that reach patients treat sterile formulation and sustained-release manufacturing as first-order CMC work instead of a fill-finish afterthought. The headlines belong to gene therapy and bispecific antibodies. Approvals, prescriptions, and growth belong to how a drug is delivered and how cleanly it is made.

Spending backs this up. The global ophthalmic drug delivery systems market reached about $17.0 billion in 2025 and is forecast to hit $23.4 billion by 2030 at a 6.6% annual rate, with intravitreal delivery the single fastest-moving piece at 8.4% (Grand View Research, 2025). That layer is growing faster than the drugs it carries, because the growth is coming from a shift in format rather than a wave of new active ingredients.

The pipeline that gets written about is not the pipeline that reaches eyes

Gene and cell therapy is the fastest-growing ophthalmic segment, projected at 18.6% a year through 2033 (Nova One Advisor, 2025). It also compounds on under 5% of the market. A 20% growth rate on a sliver is still a sliver. Small-molecule delivery grows more slowly in percentage terms and off a base that is orders of magnitude larger, which is where the actual volume of made-and-shipped product sits.

Adherence is a formulation problem wearing a patient’s clothes

The strongest argument for sustained-release ophthalmics is not convenience. It is that eye-drop efficacy data is contaminated by whether patients take the drops at all. A 2025 meta-analysis of 47 studies covering roughly 961,000 patients found prostaglandin persistence falling from 75% at six months to 56% at one year and 31% by year three. Fewer than a third of glaucoma patients are still on their starting therapy three years in, and visual-field progression accelerates once adherence drops below about 80%.

Read that as a manufacturing brief. A depot that doses for months takes the patient’s behavior out of the efficacy equation, which is exactly why sustained delivery keeps clearing regulatory review while me-too drops stall. When a sponsor asks us to increase the solubility of an ocular drug so it can be dropped more concentrated, the better question is often whether the molecule belongs in a drop at all.

The implants that reach patients are small molecules in polymer

Here is the fact that reorganizes the whole conversation. Of the sustained-release ocular implants the FDA has approved, five of six are small molecules held in a biodegradable or non-erodible polymer matrix. Iluvien and Yutiq release fluocinolone for up to 36 months. Retisert runs about 30. Ozurdex delivers dexamethasone for three to six months, and Durysta places bimatoprost for a single multi-month course. The one biologic in the group, ranibizumab, couldn’t be molded into a months-long depot, so it rides a surgically placed refillable port instead.

A polymer implant that doses for three years is a manufacturing artifact before it is a clinical one. Getting there is hot-melt extrusion of a drug implant or pharmaceutical injection molding, with the release rate set by polymer selection, drug load, and the geometry of the extrudate, not by the API alone. That is the core of drug implant manufacturing: the same fluocinolone in two different matrices gives you a six-month product or a three-year product. The molecule is a starting material. The dosing duration is process control.

We have written before about how ocular implants improve patient outcomes, and the position holds and hardens here: the growth in ophthalmic delivery is a small-molecule polymer-processing story, and sponsors who file it under “biologics” misplan their CMC.

Sterility is where ophthalmic programs actually fail

Every dosage form that reaches the posterior segment is a sterile product injected into an immune-privileged compartment with no clearance mechanism for contaminants. A single endotoxin excursion or a sub-visible particle count over USP <789> limits does not read as a minor deviation in an ophthalmic; it reads as a safety signal in a tissue that cannot flush it. This is the failure mode that catches sponsors who treated the injectable step as commodity sterile injectable manufacturing and discovered late that ocular specifications are tighter than general parenteral ones.

Two mechanisms deserve naming, because incumbents assert “high quality standards” and stop. First, particle control in ophthalmic suspensions: an injectable suspension has to hold a tightly specified particle size distribution through terminal handling, because oversized particles both irritate the eye and change dissolution, and with it the release profile you validated. Second, container-closure and filter compatibility: preservative-free ophthalmics, now the default for chronic dosing, remove the antimicrobial safety net, which pushes the entire sterility assurance burden onto aseptic process design and sterile fill-finish. The analytical work behind those specifications, the method validation that proves particle size and sterility are in control, is not paperwork after the fact; it is what makes the batch releasable. That is why we run analytical services under the same quality system as the fill.

A poorly soluble ocular molecule needs particle engineering, not a bigger dose

Most small molecules aimed at the eye are poorly water-soluble, and the eye clears about 95% of a topical dose before absorption. The reflex is to push concentration. The better lever is particle size. Reducing an API to a controlled micron or sub-micron distribution raises the dissolution rate and the effective ocular exposure without touching the molecule, which is the whole point of micronization and nanomilling for insoluble compounds. For a suspension or an implant, that particle distribution is also a release-rate parameter, so it belongs in the complex formulation development work from the first feasibility batch rather than bolted on before scale-up.

Do not gold-plate a Phase I eye program, and do not split it across vendors

We will hold two positions against a sponsor’s short-term instinct. First, match the build to the phase. A preclinical or Phase I ophthalmic does not need a commercial-scale aseptic line or a three-year-release implant locked in; it needs a clean, well-characterized batch and a formulation path that will not have to be reinvented at registration. Over-engineering early burns a biotech’s runway on controls the program may never reach. Under-building the commercial process fails validation when the stakes are highest.

Second, the most dangerous moment in an ophthalmic program is a tech transfer between vendors. API in one place, sterile drug product in another, analytical in a third, and the seams are where the particle-size method drifts, the impurity gets re-qualified, and the timeline slips a quarter nobody budgeted. Keeping API, drug product, and analytical under one quality system removes those seams. It is cheaper on paper to shop each step; it is more expensive in practice when the receiving site has to redevelop a method that was already validated once.

Plan the program around the pipeline, not the forecast

One caution on the market numbers, since every competing article opens with one. Published forecasts for the 2035 global ophthalmology drug market run from about $73 billion to $112 billion, a $40 billion spread driven mostly by differing definitions and one 9.3% outlier the approval pipeline does not support. Three independent analysts cluster near $73 to $77 billion, so that is the range we plan against, with the caveat that it is a forecast rather than a measurement.

The useful signal in that spread is not the endpoint. The signal is that the fastest-growing modality in every one of those forecasts is small-molecule, and the fastest-growing route is delivery. A CMC lead does not need to know whether the market is $73 or $112 billion in 2035 to make the right call now: lock the formulation and the particle specification before you scale, decide early whether the molecule belongs in a drop, a suspension, or an implant, and put the sterile step under a quality system that treats ocular specifications as tighter than general parenteral ones. Our recent read on FDA ophthalmic approvals shows the products clearing review are disproportionately delivery and reformulation plays, which is the same conclusion from the other direction.

What this means for your program

The future of ophthalmic drug development is not waiting on a molecule. It is waiting on whoever can hold a particle-size distribution through a sterile fill, extrude a polymer that releases for three years, and keep the whole chain under one quality system so nothing gets lost in a handoff. That is a manufacturing thesis, and it is where sponsors underinvest and then lose the timeline they thought they were protecting. If you are developing a small-molecule eye drug, treat delivery and sterility as the program, not the packaging, and talk to an ophthalmic CDMO about the format decision before the molecule is locked, not after.