Intraocular Solutions · Science Series
Cohesive or dispersive?
The answer is
two numbers.
Every viscoelastic decision in cataract surgery comes down to zero-shear viscosity and the cohesion-dispersion index. Get those two right and the chamber stays deep, the endothelium stays covered, and the OVD comes out clean. Here is the rheology, the evidence, and the manufacturing limits nobody puts on a brochure.
index scale
with a soft shell
an OVD must clear
Section 01 — The Two-Axis Truth
Viscosity tells you
almost nothing.
Cohesion tells you the rest.
For two decades OVDs were sorted on a single line: thick ones on the right, thin ones on the left, with the assumption that thick meant cohesive and thin meant dispersive. That assumption broke, and the surgeon who still carries it in their head will pick the wrong product for a dense nucleus.
Ophthalmic viscosurgical devices are non-Newtonian fluids. Their viscosity is not a fixed property but a function of how hard you shear them. At rest inside the anterior chamber, a high-molecular-weight sodium hyaluronate is enormously viscous. Push it through a 27-gauge cannula and it thins by orders of magnitude, then recovers the instant the shear stops. That behaviour is pseudoplasticity, and it is why a material can be simultaneously easy to inject and stubborn about holding a chamber open.
The number that captures the resting state is zero-shear viscosity (V0) — viscosity measured as applied shear stress approaches zero. It predicts how well an OVD will maintain space against a collapsing chamber, deepen the anterior segment, and flatten a convex anterior capsule for capsulorhexis.
The number that captures what happens at the end of the case is the cohesion-dispersion index (CDI): the percentage of OVD aspirated per 100 mm Hg of applied vacuum, reported on a scale of 0 to 100. A CDI near 100 means the material behaves as one connected mass and evacuates in a single bolus. A CDI near 0 means it clings, resists aspiration, and stays where it was placed.
Cohesion and viscosity are independent properties. A cohesive OVD is made of long, entangled, high-molecular-weight chains that move as one body: it creates and holds space, flattens the anterior capsule, and leaves the eye cleanly at the end. A dispersive OVD has shorter chains and lower molecular weight, so it does not act as a single mass: it coats the corneal endothelium, resists being flushed out by irrigation, and partitions tissue planes. Cohesives make space. Dispersives protect surfaces. Because a material can be viscous and dispersive, the classification published by Steve A. Arshinoff and colleagues in 2005 abandoned the old single-axis table and plotted OVDs on two axes — zero-shear viscosity against cohesion-dispersion index. That two-dimensional map, not the word on the carton, is what predicts intraoperative behaviour.
The 2005 reclassification was forced by a real product. DisCoVisc, a 1.6% hyaluronic acid with 4% chondroitin sulfate, was viscous by every measurement yet behaved dispersively under aspiration. One axis could not describe it. Steve A. Arshinoff, Professor in the Department of Ophthalmology and Vision Sciences at the University of Toronto and the surgeon who introduced the CDI concept in the late 1990s, published the revised two-dimensional classification in the Journal of Cataract & Refractive Surgery in 2005, adding the "viscous dispersive" class that DisCoVisc had created.
Section 02 — Interactive
Watch cohesion decide
what stays
and what leaves.
Same anterior chamber, same vacuum, two different materials. Drag the vacuum up and watch what each one does. This is the CDI in motion — the property that determines whether your endothelial coat survives the case or disappears with the first burst of irrigation.
Section 03 — The Working Map
Four classes,
four jobs.
Once you plot OVDs on both axes, the clinical logic falls out. Each quadrant does one thing well and one thing badly, and the badly matters as much as the well.
| Class | Typical chemistry | What it does well | Where it fails you |
|---|---|---|---|
| Higher-viscosity cohesive | High molecular weight sodium hyaluronate, 1.0–1.4% | Creates and holds deep chamber; flattens convex anterior capsule for rhexis; tamponades vitreous; removes in one clean bolus | Swept out early by irrigation; leaves the endothelium bare during the phaco itself |
| Viscoadaptive | Very high MW hyaluronate, 2.3% | Extreme space maintenance; fractures into fragments under high flow and behaves dispersively at low flow | Retained fragments drive early IOP spikes unless removal is deliberate and complete |
| Viscous dispersive | Hyaluronic acid with chondroitin sulfate | Holds usable space and coats endothelium — a single-syringe compromise for routine cases | Neither the best chamber nor the best coating; slower to aspirate than a pure cohesive |
| Lower-viscosity dispersive | Chondroitin sulfate blends; hydroxypropyl methylcellulose (HPMC) 2% | Adheres to and shields the corneal endothelium through ultrasound and turbulence; partitions tissue; inexpensive enough to use generously | Will not hold a deep chamber under pressure; must be actively aspirated or it stays behind |
Section 04 — The Soft Shell
The technique that
made the choice
obsolete.
In 1999 Arshinoff published a deceptively simple idea in the Journal of Cataract & Refractive Surgery: stop picking one OVD and use both, in sequence, in a specific order. The soft-shell technique is now standard practice, and the mechanism is worth understanding rather than memorising.
Inject the dispersive first. It spreads across the corneal endothelium and stays there. Then inject the cohesive centrally, beneath it. The cohesive expands, flattens the anterior capsule and deepens the chamber, and in doing so it presses the dispersive upward and outward — pinning it against the cornea as a retained protective shell. You now have a deep, stable, manipulable working space with a coated roof. Neither material alone gives you both.
first published
endothelial loss reduction
at most
of the technique
The evidence is not subtle. In comparative studies of a Viscoat-and-Healon soft shell against a cohesive used alone, endothelial cell loss measured at two to three months postoperatively was reported to be 40 to 60 percent lower in the soft-shell groups. For a surgeon operating on dense brown cataracts in a high-volume list, that is the difference between a cornea that clears by day three and one that does not clear at all.
Variants worth knowing
The ultimate soft-shell technique substitutes balanced salt solution for the cohesive in the second step, using the dispersive shell plus fluid to control chamber depth precisely — useful where a viscoadaptive would risk an IOP spike. The tri-soft shell technique, described in the same journal in 2013, adds a third layer for specific challenging anatomies. All three are the same insight applied at different pressures: cohesion and dispersion are tools, not identities.
Relative endothelial protection under phacoemulsification, ranked from the published comparative literature. Directional, not a pooled meta-analysis — absolute cell loss depends on nuclear density, phaco time and technique.
Section 05 — The HPMC Question
The cheapest dispersive
on the shelf is also
a good one.
Hydroxypropyl methylcellulose is treated in some markets as the budget option you tolerate rather than the material you choose. The comparative data does not support that. It is a genuine low-viscosity dispersive, and for the job dispersives are hired to do it performs.
The strongest single result comes from a randomised controlled trial by N. Ray-Chaudhuri, G. M. Voros, S. Sutherland and F. C. Figueiredo, published in the European Journal of Ophthalmology in 2006. Comparing 2% HPMC against 1% sodium hyaluronate in phacoemulsification, mean endothelial cell density fell 4.27% with HPMC versus 11.76% with the hyaluronate at twelve weeks. The authors called HPMC an effective and cheaper alternative for routine small-incision cataract surgery. A later randomised trial pointed the same way: less endothelial loss with dispersive HPMC than with cohesive hyaluronate.
None of this makes HPMC a cohesive. It will not hold a chamber against positive vitreous pressure, it will not flatten a convex capsule, and left behind it is as capable of an IOP spike as anything else. What it does is coat, and it coats cheaply enough that a surgeon can use as much as the case actually needs rather than rationing a syringe. In a high-volume list, an OVD you can afford to use generously outperforms a premium one you use sparingly.
Section 06 — Case Studies
When the rheology is right
and the residue
is not.
An OVD can have flawless rheology and still blind people. The three cases below are published, investigated and attributable — and in every one, the product was sterile. What failed was residue control during manufacturing.
Talal Abdulrahman Althomali investigated two consecutive clusters of toxic anterior segment syndrome affecting 15 of 24 cataract procedures. Sterilisation failure was ruled out, as were instrument reprocessing, irrigating solutions and every other candidate on the standard TASS checklist. The one variable that had changed was a newly introduced 1.4% sodium hyaluronate supplied in a prefilled syringe.
The product was withdrawn and replaced with the sodium hyaluronate formulation previously in use. No further TASS occurred in subsequent operating sessions. The author concluded that endotoxin contamination introduced during manufacturing was the probable mechanism, and called explicitly for stricter endotoxin limits on ophthalmic viscosurgical devices.
Between July and November 2005, an outbreak of toxic anterior segment syndrome was investigated across seven surgical centres in the United States. Preeta K. Kutty and colleagues — a team that included Henry F. Edelhauser, Nick Mamalis and Arjun Srinivasan — identified 112 case patients with a median age of 74.
Eighty-nine percent had been exposed to a single brand of balanced salt solution. Laboratory investigation traced the cause to intrinsic endotoxin contamination of the product itself. It remains the first documented outbreak of TASS caused by intrinsic endotoxin contamination of a manufactured intraocular product, and it reset the industry's understanding of what sterility testing does and does not catch.
In October 2007, Advanced Medical Optics voluntarily recalled an entire lot of Healon D — 4,439 units — after receiving complaints of postoperative intraocular inflammation, including toxic anterior segment syndrome. Testing of the lot revealed elevated endotoxin levels.
The relevant detail for anyone specifying an OVD is that this was a market-leading product from a major manufacturer with a mature quality system. Endotoxin is a residue of the bacterial fermentation used to produce hyaluronic acid. It survives sterilisation. Lot-level control is the only thing standing between a fermentation batch and an inflamed anterior chamber.
Section 07 — Specification
What to demand
on the certificate
of analysis.
Rheology is what an OVD does. Purity is whether it should be in an eye at all. A viscoelastic can pass every sterility test and still cause a TASS cluster, because sterility and endotoxin are different questions. Sterility asks whether organisms are alive. Endotoxin asks what the dead ones left behind.
The threshold is 0.2 endotoxin units per millilitre. ISO 15798:2013 originally set the OVD limit at 0.5 EU/mL; Amendment 1, published in 2017, tightened it to 0.2 EU/mL. The US FDA does not recognise the older ISO figure and independently recommends 0.2 EU/mL for single-use intraocular ophthalmic devices. The reason for the severity is dose: quantities as small as 0.02 EU delivered into the anterior chamber have been shown to provoke inflammation. There is very little room between a passing lot and a clinical event.
Section 08 — The Agaaz Range
Both sides of
the two-axis map.
Agaaz Ophthalmics manufactures the intraocular solutions side of cataract surgery as a system rather than a catalogue — a cohesive for space, cellulose dispersives for protection, and the staining and prophylaxis products that sit between them in the same case. Manufacturing is GMP-based, and the specifications below are what the products are, not what they are positioned as.
Intraocular solutions are the larger share of what Agaaz makes; the OP-VIEW AS and TRICENTRA lens platforms are built to sit inside the same case. View the complete portfolio →
Section 09 — FAQ
Frequently asked questions
about OVD selection.
A cohesive ophthalmic viscosurgical device is built from long, high-molecular-weight chains that entangle into a single mass. It creates and holds space — deepening the anterior chamber, flattening the anterior capsule for capsulorhexis — and it leaves the eye in one clean bolus. A dispersive has shorter chains and lower molecular weight, so it does not act as one mass: it coats and clings to the corneal endothelium, stays put against irrigation, and partitions tissue. Cohesives make space; dispersives protect surfaces. The formal distinction is not viscosity but the cohesion-dispersion index.
A laboratory number between 0 and 100 describing the percentage of an OVD aspirated per 100 mm Hg of applied vacuum. Near 100 means the material leaves as one connected mass — highly cohesive. Near 0 means it resists aspiration and stays where it was placed. Steve A. Arshinoff and colleagues introduced the CDI in the late 1990s and made it, with zero-shear viscosity, the basis of the two-dimensional OVD classification published in 2005. It matters because cohesion, not viscosity, predicts whether an OVD will be swept out during phacoemulsification.
Described by Arshinoff in 1999, it uses both classes in sequence instead of choosing between them. The dispersive goes in first and coats the endothelium; the cohesive is then injected centrally, pushing the dispersive up against the cornea as a retained shell while holding a deep chamber below. In comparative studies of a Viscoat-and-Healon soft shell against a cohesive alone, endothelial cell loss at two to three months was reported to be 40 to 60 percent lower in the soft-shell groups.
The trials disagree, and the disagreement is informative. Ray-Chaudhuri and colleagues (Eur J Ophthalmol, 2006) found endothelial cell loss of 4.27% with 2% HPMC against 11.76% with 1% sodium hyaluronate at twelve weeks — a clear win for HPMC. Storr-Paulsen and colleagues (Acta Ophthalmol Scand, 2007) found 6.97% loss with a dispersive hyaluronate against 18.03% with HPMC and 18.46% with a cohesive hyaluronate — HPMC no better than the cohesive. The defensible reading: HPMC is a competent, inexpensive dispersive that beats a cohesive used alone in most routine work, but a dispersive hyaluronate may protect better where the endothelium is already compromised. What HPMC never does is hold a deep chamber under pressure, which is why dense nuclei, small pupils and shallow chambers still call for a high-molecular-weight cohesive or a soft shell.
0.2 endotoxin units per millilitre. ISO 15798:2013 originally set 0.5 EU/mL; Amendment 1 in 2017 tightened it to 0.2 EU/mL, and the US FDA independently recommends the same ceiling for single-use intraocular ophthalmic devices. Quantities as small as 0.02 EU delivered into the anterior chamber have been shown to provoke inflammation. Endotoxin is a residue of the bacterial fermentation used to make hyaluronic acid, so it is a manufacturing-control question — a product can be perfectly sterile and still cause TASS.
Yes. Althomali reported two consecutive TASS clusters affecting 15 of 24 cataract surgeries that stopped only when a newly introduced 1.4% sodium hyaluronate prefilled syringe was withdrawn. In 2007 Advanced Medical Optics recalled a 4,439-unit lot of Healon D after inflammation complaints, with testing showing elevated endotoxin. The 2005 multistate outbreak investigated by Kutty and colleagues found 112 patients across seven centres and traced it to intrinsic endotoxin contamination. In each case the products were sterile; the failure was residue control.
Any OVD left in the anterior chamber obstructs aqueous outflow through the trabecular meshwork until it clears, producing an IOP spike typically in the first 24 hours. The risk profile follows cohesion: a cohesive removes in one bolus and is easy to clear completely, while a dispersive resists aspiration by design and needs deliberate removal. Viscoadaptive materials sit in between — they fracture into fragments under high flow, and it is the fragments left behind that spike the pressure. Complete removal, not material choice, is the control.
References & Evidence Base
Peer-reviewed
citations.
Continue Reading
Related guides
from Beyond Vision.
Start writing here...
Cohesive vs Dispersive OVDs: The Rheology Behind Every Viscoelastic Decision in Cataract Surgery 2026