Endophthalmitis after EVO ICL — the number, and the flight home
Published evidence researched September 2026 · Prices researched July 2026
An implanted lens is an operation inside the eye, and the defining complication of any operation inside the eye is infection inside it. The published figure for ICL is about 0.0167% — roughly one case per 6,000 implantations, and it is lower than the rate reported after cataract surgery, which is the number most people find first and then apply to the wrong operation. The part that is missing everywhere, and that this page exists to state: the published median interval from surgery to presentation is about eight days, and the ICL itinerary puts a fly-in patient on a plane home on day six or seven. The median moment of presentation therefore falls at or just after the flight. That makes after-hours access and the handover to an ophthalmologist at home a travel-planning question with a date on it.
The number, with the design it came from
The figure in circulation among people who go looking is 0.0167%, with a 95% confidence interval of 0 to 0.036%. It comes from an anonymous survey of ICL surgeons covering implantations between January 1998 and December 2006. Two hundred and thirty-four surgeons in twenty-one countries were approached; ninety-five responded; between them they reported 17,954 lenses implanted and three cases of endophthalmitis, one each from three different surgeons.
That is a real measurement and it is the best one there is, which is precisely why its shape should be stated rather than hidden behind the percentage. It is surgeon-reported and voluntary, with a response rate under half. Surveys of that design under-report; a surgeon who has had a case is not more likely to fill in a form about it. The period ends two decades ago, on earlier lens models and earlier antisepsis practice. And three events is a small numerator, which is what the wide confidence interval is telling you.
So the honest reading is not one in six thousand exactly. It is that the order of magnitude is thousands of operations per case, not hundreds — and that this is a risk you plan around rather than a risk you weigh against the decision to have surgery at all. What it is not is a number about you. Nobody can give you that one, and a centre that offers to should be treated with more suspicion, not less.
What this number is not: the figure you will find first
Search for infection after eye surgery and you will land on cataract data, because cataract surgery is the most-performed operation in medicine and its endophthalmitis literature is enormous. Those numbers describe a different operation on a different population, and they are higher.
| Operation and setting | Reported endophthalmitis rate | Roughly |
|---|---|---|
| EVO ICL — surgeon survey, 17,954 implantations, 1998–2006 | 0.0167% (95% CI 0–0.036%) | ~1 in 6,000 |
| Cataract surgery — ESCRS multicentre study, arms receiving an intracameral antibiotic | ~0.06% presumed · ~0.04% proven | ~1 in 1,700 – 1 in 2,500 |
| Cataract surgery — same study, arms not receiving an intracameral antibiotic | ~0.28% presumed · ~0.15% proven | ~1 in 360 – 1 in 670 |
Figures as published by their respective studies. These are not like-for-like trials: one is a voluntary surgeon survey, the other a multicentre prophylaxis study, and the populations differ sharply in age and comorbidity. The table is for order of magnitude, not for a ratio.
The useful conclusion is narrow and worth stating plainly: the ICL figure is the lowest line in that table, and the authors of the ICL survey make that comparison themselves. A reader who has been frightened by a cataract statistic has been frightened by a statistic about their grandmother's operation.
The arithmetic nobody publishes: your itinerary against the presentation window
Acute postoperative endophthalmitis does not announce itself on the table. It declares itself days later, and the published timing is consistent across series: a median of about eight days from surgery to presentation, with an interquartile range of roughly three to fourteen days, most cases arriving between the third and tenth postoperative day, and the large majority of all cases within six weeks.
Now set that against the trip. The ICL itinerary published on this site is six to seven days, and it is that length for clinical reasons — departure waits on the pressure check and on the lens-position check, not on the calendar. Neither of those gates was chosen with infection timing in mind, and neither should be. But the consequence is arithmetic:
| Day after surgery | Where a fly-in ICL patient is | Where the presentation window is |
|---|---|---|
| 0–2 | In the city, day-one and early checks | Below the interquartile range; early cases possible but uncommon |
| 3–5 | Still in the city; pressure and vault checks | Lower quartile onward — the window has opened |
| 6–7 | Cleared to fly; travelling home | Just below the median |
| 8 | Home, usually within a day of landing | The published median |
| 9–14 | Home, drops continuing, no scheduled contact | Upper half of the interquartile range |
| 15–42 | Home; most people have stopped thinking about it | The tail — the large majority of remaining cases |
Itinerary as published on this site, researched July 2026; presentation timings as published in the postoperative endophthalmitis literature. Individual schedules are set by the operating surgeon and vary.
Read the two right-hand columns together and the planning problem is obvious. Roughly the first half of the window is covered by a stay that a centre supervises directly. The median and everything after it happen at home, in a country where nobody has examined the eye, often in a different language, and frequently at a distance from the nearest ophthalmic emergency service.
The wrong response to that is to extend the trip. Waiting to day fourteen would not cover the tail either, and the gates that set the itinerary are clinical. The right response is to treat the second half of the window as something you arrange before you leave — which, unlike almost everything else on a medical trip, is entirely within the patient's control and costs nothing.
What ordinary recovery looks like, and what it looks like when it reverses
The reason a symptom list alone is not much use is that the first week after any eye operation involves discomfort, light sensitivity and fluctuating vision, and a reader told to watch for "pain and blurred vision" will either panic on day two or learn to discount both. The discriminator is not the symptom. It is the direction.
- Ordinary recovery improves. Scratchy and variable at first, then steadily better day on day. Small setbacks after a long screen session or a dusty commute settle by the next morning.
- A reversal does not. Vision that had been improving and then goes backwards over a day or two. Pain that increases rather than settles. Redness that deepens after the first few days rather than fading. Any of those inside the first six weeks is a same-day examination.
For scale rather than for self-diagnosis: in a large published series of postoperative endophthalmitis the presenting features were blurred or dropping vision in about 93% of cases, redness in about 81%, pain in about 75%, and eyelid swelling in about a third. Vision is the most consistent of them — which is worth knowing, because a patient waiting for severe pain before acting is waiting for the feature that is least reliably present.
One more figure, and it is the one a patient can act on. In that same series the mean duration of symptoms before patients presented was about 2.1 days, ranging from same-day to six. A meaningful part of the total delay was patients deciding whether it was serious enough to bother anyone. That is the interval you own, and the way to shorten it is to decide in advance — before you fly — where you would go and when it is open.
Four things to have in writing before you leave the country
These are ordinary requests. A centre that handles international patients routinely will not find any of them unusual, and the way it responds is itself informative.
- An operative record in English naming the procedure, the lens model and power, the date, and the operating surgeon. An ophthalmologist at home cannot manage a problem in an eye whose contents are undocumented, and "an ICL, I think" is not a lens specification.
- The drop regimen with generic names and end dates, in English. Brand names do not travel; molecules do.
- A named after-hours contact route for the first three weeks, with the hours it actually covers and the time zone it is in. A coordinator's daytime inbox is a business channel, not an emergency one, and the distinction only matters once.
- Written confirmation of the handover expectation — what the centre expects a local ophthalmologist to do if you present at home, and whether it will speak to that ophthalmologist directly. This is the item that converts a handover from an assumption into an arrangement, and it is the one most often missing.
Alongside those, do the two things that depend on nobody else: find the nearest ophthalmic emergency service to where you live and note its opening hours, and keep the operative record somewhere you can open it from your phone at three in the morning.
Where this sits among the other ICL questions
Infection is a rare and early risk. It is not the same conversation as the long-term corneal cell count, which is a slow measurement question and has its own page on this site; nor as lens vault, which is a positioning question and is covered separately; nor as the ordinary arc of getting used to an implanted lens, which is set out in the first three months after EVO ICL. The reason refractive centres discuss operating the two eyes on separate days at all is largely this risk, and that argument is made in full on the page about having both eyes done in one trip rather than repeated here.
Nor is this an argument against an implanted lens. It is the risk that comes with the category, it is smaller than the equivalent figure for the most common intraocular operation in the world, and it is manageable in exactly one way: by being examined quickly if the recovery reverses. Everything on this page is in service of making that faster.
Questions people actually ask
How common is endophthalmitis after ICL surgery?
The most-cited figure is approximately 0.0167% — roughly one case per 6,000 implantations, with a 95% confidence interval running from 0 to 0.036%. It comes from an anonymous survey of ICL surgeons covering implantations between January 1998 and December 2006: 234 surgeons in 21 countries were approached, 95 responded, and between them they reported 17,954 lenses implanted and three cases of endophthalmitis, one each from three different surgeons. Read the design before you read the number. It is surgeon-reported and voluntary, the response rate was under half, and surveys of this shape under-report rather than over-report, so the true figure is more plausibly at or above that estimate than below it. It also describes a period ending twenty years ago, on earlier lens models and earlier prophylaxis practice. What it is good for is establishing the order of magnitude, which is thousands of operations per case rather than hundreds. It is not a number about you, and no centre can honestly give you one that is.
Is the infection risk higher for ICL than for LASIK or SMILE?
The comparison is real and it is the honest reason this page exists rather than a laser equivalent. LASIK, SMILE and TransPRK are corneal operations: they work in the outer layer of the eye and do not open the interior. EVO ICL is an intraocular operation — a lens is placed inside the eye through an incision — and any procedure that enters the eye carries a risk of infection inside it that a corneal procedure essentially does not. Infection after laser surgery does occur and is treated as an emergency, but it is infection of the cornea rather than of the vitreous cavity, and the two are different in mechanism, in treatment and in what they threaten. So yes: this is a category of risk that ICL introduces and laser does not. It is also the single strongest argument for treating the choice between them as a clinical decision rather than a preference, and the surgeon who examines you is the person qualified to make it.
Is ICL more dangerous than cataract surgery in this respect?
On the published numbers, no — the direction runs the other way, and this matters because cataract figures are what you will find first if you search for infection after eye surgery. The ESCRS multicentre prophylaxis study found roughly 0.28% presumed and 0.15% proven endophthalmitis in the arm that did not receive an intracameral antibiotic, and roughly six per 10,000 presumed and four per 10,000 proven in the arms that did. Against about 1.67 per 10,000 for ICL, the ICL figure is the lower one on every comparison in that set, and the authors of the ICL survey say so explicitly. Two cautions before you lean on that. The populations are not alike: cataract surgery is performed overwhelmingly on older eyes, ICL on young healthy ones, and age and comorbidity are themselves risk factors. And the studies differ in design, one being a randomised prophylaxis trial and the other a voluntary survey. The comparison is useful for scale, not for a ratio.
When would an infection actually show up — before or after I fly home?
This is the part that is specific to travelling for surgery and that no consumer page states. Acute postoperative endophthalmitis is not an on-the-table event; it declares itself days later. Published series put the median interval from surgery to presentation at about eight days, with an interquartile range of roughly three to fourteen days, and most cases presenting between the third and tenth postoperative day. This site's own published ICL itinerary is six to seven days, because departure waits on the pressure and lens-position checks rather than on how you feel. Put those two facts side by side and the arithmetic is uncomfortable but useful: the median presentation moment falls at or just after the flight home. A meaningful share of the window is covered by your stay, and a meaningful share is not. The conclusion is not that the itinerary is wrong — it is set by clinical gates and extending it would not cover the tail either — but that arrangements for the second half of that window have to be made before you leave, not improvised from home.
What symptoms should make me seek care urgently?
Three, and the published frequencies are worth carrying because they tell you what the presentation actually looks like rather than what a leaflet says. In a large case series the presenting features were blurred or dropping vision in about 93% of cases, redness in about 81%, pain in about 75%, and eyelid swelling in about a third. The pattern that matters is not any one of them in isolation but the direction of travel: vision that was improving and then goes backwards, pain that is increasing rather than settling, or redness that deepens after the first few days. Ordinary early recovery moves the other way — scratchy, light-sensitive and variable at first, then steadily better. Anything that reverses that trend in the first two weeks is a same-day examination by an ophthalmologist, not a message to a coordinator and not a wait-and-see. If you are reading this because you have those symptoms now, stop reading and get examined today.
How fast does it have to be treated?
Fast enough that the delay you control is worth taking seriously. In the same series the mean duration of symptoms before patients presented was about 2.1 days, with a range from same-day to six days — which means a substantial part of the total time from onset to treatment was spent by patients deciding whether it was serious. That is the one interval a patient can shorten, and it is the reason the practical advice is to have the route to an eye casualty department identified before you need it rather than researched at two in the morning in your own country. Treatment itself is an ophthalmic emergency and is decided by the treating ophthalmologist on examination; this page does not describe management because management is not a decision a patient makes. What the patient decides is how quickly they are in front of someone who can look.
What should I ask the hospital in China to put in writing before I fly home?
Four things, and they are ordinary requests that a centre used to international patients will not find unusual. First, an operative record in English naming the procedure, the lens implanted with its model and power, the date, and the surgeon — a receiving ophthalmologist at home cannot manage a complication in an eye whose contents are undocumented. Second, the drop regimen with names and end dates, again in English, since brand names differ between countries. Third, a named after-hours contact route for the first three weeks, with the hours it actually covers and the time zone — a daytime coordinator's inbox is not an after-hours route. Fourth, written confirmation of what the centre expects a local ophthalmologist to do if you present at home, which is the question that turns a handover from an assumption into an arrangement. A centre that supplies these has thought about the second half of the window; one that cannot is telling you something about its model.
Should I arrange an eye examination at home even if nothing is wrong?
Identifying where you would go is worth doing regardless, and it costs nothing. Whether to book a routine review is a question for the operating surgeon, because it depends on the follow-up schedule they set and on what they want measured — and an ICL has checks that are specific to it, of which lens position is the obvious one. The general principle a fly-in patient should hold is that a follow-up schedule designed around a patient who lives nearby has to be explicitly re-planned for one who does not, and that re-planning is the surgeon's to do and yours to ask for before you book flights. What does not depend on anyone's schedule is the emergency route: know the name and the opening hours of the nearest ophthalmic emergency service to where you live before you travel, and keep the operative record where you can reach it from your phone.
Does a hospital's own infection rate tell me anything?
Very little that you can verify, and this site deliberately publishes no centre-level figure — not for its partner hospitals and not for anyone else's. There are three reasons. A rate of this order needs tens of thousands of operations before the confidence interval is narrow enough to distinguish one centre from another, so almost every published single-centre figure is dominated by chance. Self-reported rates are unauditable by a patient and are produced by the party with an interest in them. And a centre that quotes a very low or a zero rate is usually quoting a small denominator, which is the opposite of reassuring. What is worth asking about instead is process, because process is observable: whether the theatre is a dedicated intraocular one, what surface antisepsis is used, and how an after-hours presentation is handled. Those questions have answers a patient can hear and judge.