Vitrectomy for floaters: should the surgeon induce a posterior vitreous detachment?Floater surgery: should a PVD be induced?

What is a posterior vitreous detachment, and why does it matter in floater surgery?

The vitreous is the clear gel that fills the eye between the lens and the retina. With age the gel partly liquefies, and its back surface, a thin membrane of collagen, separates from the retina. That separation is a posterior vitreous detachment. It is a normal ageing change and the commonest cause of floaters, because the detached membrane and often a ring of tissue from around the optic nerve (a Weiss ring) drift in the fluid. What floaters are made of describes these in more detail.

A PVD does not detach the whole gel. The gel stays firmly attached at the outer edge of the retina, in a band called the vitreous base, and a PVD never separates that band.

This matters in surgery because a vitrectomy removes gel, and gel that has already separated is easy to cut away. Gel that is still attached has to be left in place or separated from the retina, and separating it is what surgeons mean by inducing a PVD. This also explains the risks of each approach. When the gel separates from the retina naturally, traction is applied to the retina, which can potentially pull on it and cause a retinal tear, as flashes, floaters and a curtain explains. When a surgeon separates the gel from the retina, traction is likewise applied to the retina, which can potentially pull on it and cause a tear.

Does every eye need a PVD induced?

No. Many eyes already have a complete PVD, and the proportion rises steeply with age. In a Japanese study that scanned about 50 eyes in each age group with OCT, complete PVD was found in none of the eyes of people in their thirties who were not highly short-sighted, in 14% in their fifties, 44% in their sixties and 70% in their seventies. It comes earlier in highly short-sighted eyes (longer than 26 mm): 14% in the thirties, 54% in the fifties and 94% in the seventies.

+

Age alone does not settle it. Even in the seventies, three in ten eyes without high short-sightedness had no complete PVD.

Floater surgery is also done mostly on eyes that already have a PVD, because a PVD is what causes many floaters. In a United States series of 413 eyes, 88% had a PVD before surgery, and surgeons induced one in only 8% of the 362 eyes with operation notes. In a California series of 286 eyes, 63% had a complete PVD, 27% had none and 9% had a partial one.

Cataract surgery can bring a PVD forward. In a Japanese study, 27% of eyes developed a complete PVD in the year after cataract surgery, compared with 5% of matched eyes that had no surgery, and new floaters after cataract surgery looks at what that means in practice. In another study of people aged 50 to 60, PVD was present in 51% of eyes with a lens implant and 21% of the fellow eye that still had its natural lens. Not every study agrees: a 2026 study from one Argentinian clinic found no acceleration in men and an excess only in women aged 70 to 79. So older eyes and eyes with a lens implant often have a PVD already, but often does not mean always.

Finding out beforehand is imperfect. Compared with what surgeons saw in theatre after staining the gel, a clinical examination, ultrasound and standard OCT each missed roughly four in ten to half of complete PVDs in one study of macular operations. Widefield swept-source OCT detected 99% in another study of 145 eyes, though it wrongly called about one in twelve attached gels detached. Some surgeons therefore confirm the PVD status at the start of the operation.

How is a PVD induced during surgery?

A PVD is induced during the vitrectomy, once the central gel has been removed. The surgeon uses the vitrectomy probe itself, switched from cutting to suction only, to lift the back surface of the gel away from the retina. The steps below describe the usual sequence.

Vitrectomy probe →1Floaters →2Light pipe →3← Retina4+

1. Vitrectomy probe2. Floaters3. Light pipe4. Retina

1. The surgeon first removes the central gel with the vitrectomy probe, working through three tiny ports in the white of the eye.
2. If the gel is still attached, the probe is switched from cutting to suction only. Its tip is placed just above the retina at the edge of the optic nerve, the usual starting point, and strong suction (reported at 500 to 650 mmHg) draws the thin back surface of the gel onto the opening.
3. The surgeon lifts the gel gently off the retina and sweeps it outward towards the periphery, watching its leading edge for tears. It is usually enough to lift it as far as the eye's mid-line, the equator, because the rest often follows as gel is cut away.
4. Because the gel is clear, some surgeons stain it with a few specks of triamcinolone, a steroid, so they can see whether it has lifted.
5. The gel is then cut away. The vitreous base stays attached in every case. Some surgeons also shave it, and others say shaving is generally unnecessary.
6. The surgeon checks the whole edge of the retina, often pressing on the outside of the eye to bring it into view, and treats any tear with laser or freezing.

Break rates when a PVD is induced vary widely, because the figures come from different operations, mostly macular surgery in older people. A 2026 Australian series summarised a published range of about 3% to 32%. Its own 185 eyes had a break in 15%, far more with 25-gauge instruments (21%) than with 27-gauge ones (3%), though most operations were done by surgeons in training and the gauge effect is an association, not proof of cause. The floater studies conflict. One series of 116 operations found that induction increased the risk of breaks, from under 12% to over 30% in the figures its later commentators quote. Another, of 168 eyes, found no difference, though only 12 eyes had a PVD induced. Breaks found during surgery are treated at the time.

Serious injuries from induction have been reported only in small numbers. A macular hole developed after induction in one 44-year-old man having floater surgery. A 2026 preprint that has not been peer reviewed, of 89 young eyes that had a complete vitrectomy with a PVD induced, reported permanent field defects in 2 eyes, attributed to touching the optic nerve, as well as one macular hole and one infection.

Vitrectomy for floaters

Vitrectomy for floaters

Dr Chen removing the vitreous gel and its floaters during vitrectomy. It shows real eye surgery.

Watch on YouTube →

Why is a PVD harder to induce in younger eyes?

Younger eyes hold the gel more tightly. In donor eyes, the thin inner layer of the retina remained attached to the gel's surface in 6 of 15 eyes from donors aged 20 or younger, but in none of the eyes from donors aged 21 or older. A later laboratory study of 17 human eyes found that the force needed to separate the gel fell significantly after age 60. A web textbook edited by a Sydney retinal surgeon lists young age among the situations in which inducing a PVD is more difficult, and adds that elderly people with a lens implant who need surgery for a membrane on the macula are more likely to have one already.

Fine collagen frameworkBundled collagen fibres

+

The gel is also usually less liquefied in young eyes, so it behaves as one firm body rather than a loose membrane around a watery core. Lifting it can then apply traction to the retina where the attachment is strongest.

This is reasoning supported by laboratory work and by expert experience, not a measured result in floater surgery. Clinical data on age and breaks conflict: a 2026 scoping review of ten studies found that younger age was linked with more breaks during induction, but a recent cohort of 185 eyes aged 20 to 94 found no age effect.

What is limited, or lens-sparing, vitrectomy?

Some surgeons, when the gel is still attached, do not induce a PVD. They remove only the central gel, often called core vitrectomy, and in an eye with its natural lens they leave 3 to 4 mm of gel behind the lens. This is usually called limited or lens-sparing vitrectomy. It was developed by a group in California, which publishes most of the data on it.

The aim is fewer retinal breaks, because the gel is not separated from the retina, and a later cataract, because less gel is removed. The cost is that some gel stays attached to the retina, and it may separate on its own years later.

The two approaches are points on a range rather than two boxes, as the sections below show. Vitrectomy describes the operation itself.

What are the pros and cons of each approach?

The table sets out what the studies report. Every row comes from studies without a control group that chose the operation at random, so none of it is proof that one approach causes a difference. Recovery after vitrectomy for floaters covers what follows either operation.

Do retinal tears happen more often when a PVD is induced?

They may, in some eyes, but the studies point in different directions and none isolates induction from the rest of the operation. The largest, an international survey by the European VitreoRetinal Society of 581 eyes operated on by 48 surgeons in 16 countries, sorted the operations into three kinds: complete (the gel removed out to the periphery, with a PVD induced unless one was already present), subtotal (a PVD induced but the periphery not cleared) and core (central gel only, no PVD induced).

+

The survey found breaks in 25 of 165 eyes (15%) after complete vitrectomy, 3 of 288 (1.0%) after subtotal and 1 of 128 (0.8%) after core. So the excess seems to be tied to clearing the gel out to the periphery, and to slow cutting speeds, which accounted for almost nine in ten of the breaks, rather than to lifting the gel alone, because the subtotal group also had a PVD induced. Most of the breaks (62%) also occurred with older 20-gauge instruments. The authors point out that they did not know how many of the complete group had a PVD to begin with.

The earlier floater series conflict. One found that inducing a PVD raised the rate of breaks, and another found no effect. A 2016 series from Amsterdam analysed 1,409 mixed elective operations, 153 of them for floaters, and found breaks in 2.8% of eyes when a PVD was induced and 0.6% when the gel had already separated. Its abstract does not report floaters separately.

Two things follow. The risk of a break when inducing a PVD is real. But how much of it comes from separating the gel itself, how much from the wider removal of gel, and how much from age and instruments is not known.

Will floaters come back if a PVD is not induced?

They can. If the gel is left attached, it may separate later, and the separation releases new floaters. In a series of 286 eyes, none of the 179 eyes that already had a complete PVD developed new floaters. Among the 99 eyes with no complete PVD, 14 (14%) did, on average 2 years after surgery and as early as 6 days and as late as 5.5 years. Eleven of the 99 eyes (11%) had a second vitrectomy. The eyes with new floaters were mostly young, short-sighted, male and had their own lens.

+

This group's own later report of 651 eyes counted 15 recurrences in 289 operations (5%). Both papers come from the same California group, whose authors now write that inducing a PVD should be considered in young, short-sighted eyes with their own lens and no complete PVD.

No study compares this rate with eyes that had a PVD induced, so the size of the benefit of inducing one is unmeasured. Whether not inducing a PVD leaves more floaters behind at the first operation has not been tested. What the series report is not floaters missed at the first operation but new floaters from a later PVD. Some surgeons add that gel left behind the lens can itself cause symptoms.

What happens to the retina later in life?

Nobody knows how much a limited vitrectomy changes the long-term risk of tears when a PVD develops naturally years later, and no study is long enough to say.

The concern is plausible. If the gel is left attached, it may detach later, and separation of any gel can apply traction to the retina where the attachment is firm, such as at areas of thin retina called lattice. But the vitreous base stays attached whichever operation is done, so inducing a PVD does not remove every attachment, and no study has measured how the risk of a later tear compares.

What the data show is short-term. Retinal detachment followed 7 of 289 limited vitrectomies (2.4%), at 2 weeks (two eyes), 7, 11, 15, 42 and 48 months, after an average follow-up of about 3 years, and up to 16 years in a few eyes. Two of the 7 eyes had no PVD before surgery, and the authors presumed a new PVD caused the detachment. In the same survey as before, detachment was 3.6% after complete vitrectomy, 2.4% after subtotal and 0.8% after core, a difference that was not statistically significant, with a median follow-up of 20 months. An older series of 110 eyes reported detachment in 10.9% of eyes, 4.5% within three months and 6.4% later, and another, described by others as mostly core vitrectomy, reported late detachment in 5.5% at 2 to 4 years. Those were different operations in a different era, and they cannot be compared with the newer series.

+

The honest summary is that it is reasonable to expect leaving the gel attached to raise the chance of a later PVD-related problem to some degree, that no study has measured by how much, and that the studies reporting low detachment rates for limited vitrectomy mostly follow eyes for two to three years. In a trade interview, the surgeon who developed limited vitrectomy said that about one in five of the eyes had laser or freezing to weak areas of retina before surgery, which he credits for the low detachment rate.

Does leaving the gel delay a cataract?

It seems to postpone cataract in many eyes, but the evidence does not show that it prevents it, and it does not show that not inducing the PVD is what does the postponing.

With vitreous gelLater: fluid and a cataract

+

Removing the gel exposes the natural lens to more oxygen, which is thought to be the main reason a cataract follows. In 69 eyes measured during vitrectomy, oxygen beside the lens rose from about 9 mmHg with the gel present to about 70 mmHg immediately after it was removed. In a group of 179 eyes with their own lens after limited vitrectomy, 64 (36%) had cataract surgery after an average of 2 years. The proportion rose with age.

In one comparison, 17 of 48 eyes (35%) had cataract surgery after limited vitrectomy and 20 of 23 (87%) after more extensive vitrectomy, both followed for at least 2 years, but the two groups differed in age and came from different surgeons, and only the abstract is available. That comparison supports the idea that leaving gel behind postpones cataract. It cannot tell whether the benefit comes from not lifting the gel, from the 3 to 4 mm left behind the lens, or from removing less gel altogether.

The idea that it only delays is supported for eyes over about 50. Even after limited surgery, about 4 in 10 had cataract surgery, on average 1.5 to 2.3 years after the floater operation. Under 50, the picture is less clear: 3 of 37 eyes had cataract surgery, but at 12, 56 and 102 months, so young eyes are not immune. Only a few eyes have been followed for more than 9 years, and the series report how many eyes had cataract surgery, not how the risk builds over time. If you have a lens implant, the question does not arise. Floater surgery and cataract and why vitrectomy causes cataract cover this in detail.

+

What do surgeons say when the evidence runs out?

Surgeons who operate on floaters fall into three groups. This section reports what they say, in journals and in the trade press. It is opinion and experience, not evidence, and it is labelled as such.

• Do not induce a PVD. The California group that developed limited vitrectomy avoids induction to limit retinal breaks and oxygen exposure of the lens. A US retinal surgeon advises the same for younger, often short-sighted people with no PVD, and warns them that new floaters may appear when a PVD develops on its own. A London surgeon leaves the peripheral and front gel as well, reasoning that the more gel left, the lower the chance of a later PVD and of cataract. That is a reasoned view, not a measured result.
• Induce a PVD. A surgeon at the University of Illinois in Chicago does a complete vitrectomy and induces a PVD if there is none, accepts a higher tear risk at surgery, especially in highly short-sighted eyes with thin retina, and expects the eye to get a PVD later anyway, with new floaters and a risk of tear. A surgeon in St Louis always creates a complete PVD because of a worry that residual gel may contract later. A Brussels-based group reported a macular hole after inducing a PVD in a young man, yet still prefer to induce one in every patient, reasoning that leaving the gel attached may lead to later breaks or detachment.
• Operate only if a PVD is already present. Several US surgeons require or strongly prefer a confirmed PVD, and often a lens implant too, so the question does not arise. A UK survey of surgeons who applied criteria found 38% listed an existing PVD as one.

Two points weaken these views as evidence. Most of the figures quoted for not inducing a PVD trace to one group's own series. And even the California group now discusses inducing a PVD with young, short-sighted patients who have no PVD, so the disagreement is narrower than it looks.

Case reports are anecdote, not evidence, but they show the risks are real on both sides. They include a macular hole after induction, a tear during induction with a small local detachment followed by a rapid cataract and a membrane on the macula, and a detachment 8.5 months after limited surgery, a few days after new floaters began. They cannot give rates.

Which approach suits which eye?

No approach suits everyone, and the decision depends on the eye. These are the factors that change it, not a recommendation for you. The first question is whether a PVD is already present, and the usual tests are imperfect.

+

• A PVD is already present. Nothing needs to be induced. The choice is then how much gel to remove and whether to clear the periphery. Most older eyes and many eyes with a lens implant are in this group.
• An older eye with no PVD. The gel is usually less firmly attached, so inducing a PVD may be easier. Some surgeons induce one. Others decline to operate until a PVD develops.
• A young, short-sighted eye with its own lens and no PVD. This is the hardest case. The gel is most firmly attached and the risk of new floaters later is highest, but cataract after limited surgery is less common under 50, and inducing a PVD may carry the most tear risk. The options are to wait, because a PVD is likely to come earlier in a short-sighted eye; limited vitrectomy, accepting a risk of new floaters and possibly a second operation; or inducing a PVD, accepting more risk of a break and possibly an earlier cataract. YAG laser or vitrectomy for floaters and treating floaters without surgery cover the alternatives.
• Wanting to avoid a cataract at all costs. No technique prevents it. A lens implant already in place removes the concern altogether.

Ask whether you already have a PVD, how that was checked, what your surgeon would do if you did not, and how much gel would be left.

Other questions about inducing a PVD

Where can I read more?

When are new floaters urgent?

A sudden shower of new floaters, new flashes of light, a dark shadow or curtain moving across your vision, or a sudden drop in vision can mean a retinal tear or detachment. This applies to anyone, whether or not they have had floater surgery. Contact your ophthalmologist urgently, or attend an emergency department.

After floater surgery, contact your surgeon if you develop pain, increasing redness or worsening vision, which can signal infection or raised pressure. Retinal detachment describes the treatment of a detachment.