The number nobody agrees on
Written August 2026 · Informational — not medical advice, and not a substitute for the follow-up your operating surgeon arranges
Ask how often vision drifts back after refractive surgery and you will get answers between 1% and 40% — from serious published work, not from anyone being careless. That spread is the actual finding. It exists because "regression" is not one event with one definition: some papers count any measurable shift, some count only a shift big enough that a second operation was offered, and the threshold for offering one moves with the surgeon, the era and the laser. Two survival analyses of the same two procedures put the proportion of eyes free of regression at 88.1% and at 43.8%. Both are honest. They are counting different things. This page explains what is being counted, what the mechanisms are, what genuinely predicts a drift, and — the part that matters if you were operated on abroad — how to tell real regression from a measurement artefact before anyone reaches for a second procedure.
Why this page exists. There is no shortage of material on this topic; there is a shortage of material that survives contact with a second source. The clinical literature is detailed and consistent about mechanism but reports incidence across a range so wide it looks like a typo. The patient-facing material resolves the range by picking a number from it — you will read "fewer than 5% in the first year", "5–10% over ten years", "usually mild, 0.50 to 1.00 dioptre" — each defensible, none carrying the definition that makes it true, so a reader who checks two pages concludes the field is confused. It is not confused. It is measuring several different things and labelling them with one word.
Three different things get called "my vision went back"
Before any number is useful, the event has to be separated from two things that feel identical from the inside and have entirely different answers.
| What happened | What is changing | What it means for you |
|---|---|---|
| True regression | The cornea's shape, or the eye's length, has moved back toward its pre-operative refractive state. Measurable, repeatable, and present on a stable tear film. | Real, and the subject of this page. May or may not be worth treating depending on size, stability and what remains of your corneal thickness. |
| Age-related change | Nothing has undone. The near-focusing range that everybody loses from their forties is being lost on schedule, and surgery neither caused nor prevented it. | Not regression, however much it feels like it. It shows up as reading difficulty rather than distance blur, and it would have arrived anyway. See our page on refractive surgery in your forties. |
| A measurement artefact | The cornea has not changed. An unstable tear film is scattering light and inducing irregular astigmatism, so the refraction being measured is partly the surface film rather than the eye underneath it. | The most under-recognised of the three, and the one most likely to hit a fly-in patient. Treated as regression it leads to an unnecessary second procedure. See below. |
Distinguishing these three is a clinical judgement made on examination, not something a web page can do for you. What a web page can do is make sure the question gets asked.
The numbers, and why they disagree
Here is the honest state of the published record, with what each figure was actually counting. Read the middle column before the right one — it is the reason the right one moves.
| Reported figure | What was being counted | Source type |
|---|---|---|
| 1% – 40% (SMILE) 5% – 28% (LASIK) |
Enhancement rates across reviewed published series — i.e. how often a second procedure was performed. The width comes from era, nomogram, patient mix and each surgeon's threshold for offering one. | Range across series |
| 2.2% – 2.9% | Enhancement percentages in recent SMILE series, associated with older age and higher myopia. | Recent series |
| 1.3% SMILE · 3.0% PRK · 3.8% LASIK | Regression rates across all subjects in one multi-procedure comparison. | Single comparative study |
| 1.32% SMILE · 4.98% FS-LASIK | Enhancement rate in one direct comparative study. | Single comparative study |
| 13.1% FS-LASIK · 17.4% SMILE | Incidence of myopic regression itself over one year — not of re-operation. Note that the procedure ordering flips versus the enhancement figures, because these are different events. | One-year comparative study |
| 83.7% FS-LASIK · 88.1% SMILE free of regression | Survival analysis where the event was regression requiring enhancement. | Survival analysis |
| 42.1% FS-LASIK · 43.8% SMILE free of regression at 18 months | Survival analysis where the event was any regression meeting the study's refractive threshold. Same design family, roughly 40 percentage points apart from the row above — because "regression" and "regression that got operated on again" are not the same event. | Survival analysis |
| 23.6% of eyes | Myopic regression observed over a 5–16 year window after corneal refractive surgery — a long-horizon figure, and long horizons include ordinary ageing of the eye. | Long-term series |
Published research results, researched August 2026. These are group statistics from study populations with different inclusion criteria, different follow-up lengths and different definitions of the endpoint; none of them is a prediction for an individual eye, and none is an eligibility criterion.
The one thing to take from that table. When somebody quotes you a regression rate — a surgeon, a forum, a clinic page, this page — the first question is "counting what?" A 1.3% and a 43% can describe the same cohort. If the number arrives without its definition, it cannot be compared with any other number, and stacking two undefined numbers is how forum threads end in a fight.
Why it happens at all
Three mechanisms are described, and they are not alternatives — they can run together in the same eye.
1. The epithelium grows back thicker than it was
The surface layer of the cornea is not passive. After a myopic treatment it tends to thicken centrally, partially filling in the shape that was created, and the degree of that thickening is reported to depend strongly on the size of the correction — bigger correction, more remodelling. It happens after surface ablation, after LASIK and after lenticule extraction, to different extents. This is the mechanism that makes higher corrections a genuine risk factor rather than a superstition.
2. The stroma remodels, and the cornea's mechanics change
Removing tissue from the front of the cornea reduces its rigidity, and a less rigid cornea can bulge slightly at the back over time, altering the eye's total refractive power. Alongside that, the wound-healing signalling that follows any corneal procedure drives the cornea's resident cells into a repair state that itself remodels the tissue. This mechanism is why how much tissue is left is a load-bearing pre-operative number — the same number the ectasia screening is built on. What each figure on your workup report is, and what it is weighed against, is set out on our page decoding the workup report.
3. The eye keeps getting longer — and no corneal procedure touches that
The third mechanism is not a corneal one at all, and it is the reason a young, highly myopic patient can be told their surgery was technically excellent and still find themselves short-sighted again years later. Refractive surgery reshapes the cornea; it does not shorten the eye. If axial elongation is still running, the eye continues to become more myopic underneath a cornea that was set for the prescription it had on the day of surgery. That distinction matters far beyond spectacles — the risks that track axial length are not corrected by correcting the refraction, and we cover them separately in what surgery does and does not do to your myopia.
Get my early read →Send your rough prescription — a candid reply on fit and price within one business day.What predicts it — including one thing you can still change
Across the studies the predictors are consistent, and they split usefully into facts about your eye and one parameter that is chosen for you rather than found in you.
| Predictor | Direction | Can you do anything about it? |
|---|---|---|
| Size of the correction (pre-operative spherical equivalent) |
Higher myopia, more regression. A significant predictor in Cox proportional-hazard modelling, and consistent with the epithelial mechanism. | No — but it changes which procedure is the sensible one, which is a real decision. |
| Age | Older age is associated with higher enhancement rates in the recent SMILE series. | No. Worth knowing so the result is not misread as a technical failure. |
| Designed optical zone | A smaller optical zone contributes significantly to regression after both FS-LASIK and SMILE; it appears alongside pre-operative refraction as a significant predictor in the same modelling. One high-myopia SMILE protocol specifies zones of 6.2–6.7 mm. | Yes — this is a planning parameter, not a property of your eye. It is chosen when the treatment is designed, and it trades against how much tissue must be removed. You can ask. |
| Corneal and biometric measures (higher-order aberrations, anterior chamber depth, corneal diameter, predicted ablation depth) |
Reported as predictors in low-to-moderate myopia cohorts. | No, but they are all on the report you were given — see the workup page. |
| Which procedure | Contested. Some comparisons favour SMILE; one survival analysis found the procedure had no significant effect once the other factors were accounted for. | Partly — but corneal thickness and shape usually decide procedure long before this does. |
Published guidance and study findings researched August 2026, given so you can ask better questions at your consultation. None of it determines candidacy — candidacy and the choice of procedure are determined by the operating surgeon after examination.
The two questions worth asking before surgery, not after. "What optical zone are you planning for my treatment, and what made you choose that size?" and "Given my prescription and my corneal thickness, what does that zone cost me in tissue?" Neither is a challenge to the surgeon; both are answerable in a sentence by anyone who planned your treatment, and both are unanswerable afterwards, because once the treatment is delivered the trade-off has already been made.
When it happens
The clustering is early. Most of the shift that is going to happen tends to happen in the first year — the survival curves fall most steeply over the first twelve to eighteen months and then flatten — which is why "stable at a year" is a meaningful phrase and "stable at three weeks" is not. Long-horizon series that follow eyes for five to sixteen years report higher cumulative proportions, but by that horizon ordinary ageing of the eye is inside the number too, which is precisely the confusion this page opened with.
For a fly-in patient this timing has an awkward consequence: the whole of the period in which regression usually declares itself is a period you spend at home, measured by optometrists and ophthalmologists who never saw your eye before surgery and have no pre-operative record of it. That is not a reason to avoid travelling. It is a reason to leave China with the documents that make you measurable — and to understand the artefact below, because at home it is the failure mode nobody is watching for.
The artefact: a refraction taken on a dry eye is not a measurement of your cornea
This is the section this page was built for, and it comes directly out of a pattern we keep meeting: a patient a few weeks or months post-operative, vision noticeably worse than it was at their best, convinced the surgery has undone itself, sometimes already being offered a touch-up.
Refractive surgery reduces corneal sensation and disturbs the tear film for months — that is the subject of our page on corneal sensation after refractive surgery — and an unstable tear film has direct optical consequences. Tear-film instability increases optical aberrations and light scattering; if the eye is held open longer than the tear film stays intact, visual function deteriorates through induced irregular astigmatism and higher-order aberration. In other words the front surface of the eye stops being a clean optical surface between blinks, and whatever instrument is pointed at it measures that.
The practical signature is recognisable. Vision that fluctuates — worse at the end of the day, worse on a screen, worse in air conditioning, better immediately after a blink or a lubricating drop — is behaving like a surface problem. Regression does not come and go with a blink. A refractive error that has genuinely returned is stable: the same at nine in the morning as at nine at night, the same before and after a drop, the same at two visits a month apart.
The honest complication, stated rather than suppressed. The relationship between dryness and regression is not settled in the direction people assume. Several sources describe post-operative dry eye as contributing to later refractive regression by delaying corneal healing. But at least one study of tear-film stability after SMILE found that instability impaired visual-quality recovery and prolonged healing while not causing short-term refractive regression. Both can be true — the tear film reliably degrades the measurement and the experience, while its role in causing genuine long-run regression remains contested. The action is the same either way: do not accept a refraction taken on an unsettled surface as the basis for a second operation.
What to ask for, if a drift is found: that the surface be treated and the refraction repeated rather than acted on immediately; that the measurement be taken at a comparable time of day to the reading it is being compared with; and that the comparison be against your own documented post-operative best, not against the number you were promised. If dryness is active, our dry-eye page covers what settling it actually involves.
What you can leave China holding
Regression is one of the few post-operative questions that is answered almost entirely by paperwork, because it is a question about change, and change cannot be established from a single reading taken by someone with no baseline. The set below costs one conversation before you fly home and is difficult to obtain afterwards.
- Your pre-operative refraction, in writing. Surgery makes the pre-operative refraction permanently unmeasurable — it exists only in the operating hospital's file. Without it there is no anchor for the word "back".
- The treatment parameters actually delivered, including the planned optical zone and the attempted correction. These are what make later readings interpretable rather than merely alarming.
- Your post-operative refractions at every follow-up you attended, with dates — not just the last one. A curve is evidence; a point is not.
- A corneal topography or tomography print-out from after the procedure. If a shape question arises later, a local specialist comparing against nothing is guessing.
- The written enhancement terms — the eligibility window, what is included and what a second procedure would cost. That is a contract question rather than a clinical one, and it lives on our page about enhancement policy and aftercare.
One habit is worth adding, and it costs nothing: from the point your surgeon declares you settled, have your refraction recorded once a year at home, at roughly the same time of day. Three dated readings turn "I think it has got worse" into something a clinician can act on, and they are the difference between a stability question that can be answered and one that can only be argued about.
Where the boundary sits
Two things this page deliberately does not do. It does not tell you whether your own drift warrants a second procedure — that depends on the size and stability of the change, on how much corneal tissue you have left, and on which procedure you had, and it is a decision for the operating surgeon after examination. And it does not price anything: what a re-treatment costs, whether it is covered and how long the window lasts are contractual questions, answered on the aftercare page rather than here.
The third boundary is the one worth repeating because it is the commonest confusion in the threads that produced this page: age-related change is not regression. If reading has become harder while distance vision has held, nothing has undone — the eye has simply reached the age everybody's eye reaches, and it would have reached it whether or not you were ever operated on.
Get my early read →Send your rough prescription — a candid reply on fit and price within one business day.Questions people actually ask
What percentage of people get regression after LASIK or SMILE?
There is no single defensible percentage, and anyone who gives you one without saying what they counted has skipped the hard part. Across reviewed published series, enhancement rates — how often a second procedure was actually performed — run from about 1% to 40% after SMILE and about 5% to 28% after LASIK. Recent series report much tighter figures: enhancement percentages of 2.2% to 2.9% after SMILE, and one direct comparison found 1.32% for SMILE against 4.98% for FS-LASIK. But studies that count regression itself rather than re-operation report far higher numbers from the same kinds of population: 13.1% after FS-LASIK and 17.4% after SMILE within one year, and 23.6% of eyes over a five-to-sixteen-year window. The spread is not sloppiness. It reflects the era, the treatment nomogram, the mix of prescriptions in the study, and above all each surgeon's threshold for offering a second procedure. When you are quoted a rate, ask what event was being counted.
Why do two studies of the same procedures give completely different regression rates?
Because they defined the event differently, and the definition moves the answer more than the surgery does. One survival analysis of FS-LASIK and SMILE reported 83.7% and 88.1% of eyes free of regression, where the event counted was regression that required an enhancement. Another survival analysis of the same two procedures reported cumulative survival of 42.1% and 43.8% at eighteen months, where the event was any regression crossing the study's refractive threshold. Those two results are roughly forty percentage points apart and neither is wrong: most refractive drift is small, and most small drift is never operated on, so a study counting operations will always find far fewer events than a study counting millimetres of dioptre. This is the single most useful thing to understand about the literature on this topic, because it dissolves most of the apparent contradictions you will find in forum threads.
Is my vision getting worse regression, or is my eye just dry?
The distinguishing feature is stability rather than severity. A dry-eye artefact fluctuates: worse at the end of the day, worse on screens, worse in dry or air-conditioned air, and noticeably better immediately after a blink or a lubricating drop. Genuine regression does not behave that way — a refractive error that has really returned reads the same in the morning and the evening, the same before and after a drop, and the same at two visits a month apart. The mechanism behind the artefact is well described: an unstable tear film increases optical aberrations and light scattering and induces irregular astigmatism, so the instrument measuring your eye is partly measuring the film on its surface. There is an honest complication worth knowing — some sources describe post-operative dry eye as contributing to genuine later regression, while at least one study of tear-film stability after SMILE found it impaired visual-quality recovery without causing short-term refractive regression. Either way the practical rule holds: an unsettled ocular surface should be treated and the refraction repeated before anybody acts on the reading.
Why does refractive regression happen?
Three mechanisms are described and they can run together in the same eye. First, epithelial remodelling: the corneal surface layer tends to thicken centrally after a myopic treatment, partially filling in the shape that was created, and the extent of that thickening depends strongly on how large the correction was. Second, stromal and biomechanical change: removing tissue reduces the cornea's rigidity, which can allow slight posterior steepening that alters the eye's refractive power, while wound-healing signalling drives the cornea's resident cells into a remodelling state. Third, and quite separately, axial elongation — the eye continuing to grow longer. That third one is not a corneal process at all, which is why it can affect an eye whose surgery was technically excellent: reshaping the cornea does not shorten the eye, so if elongation is still running the eye becomes more myopic underneath a cornea that was set for an older prescription.
Does a bigger prescription make regression more likely?
Yes, and this is one of the more consistent findings in the literature rather than a rule of thumb. Higher pre-operative myopia contributes significantly to the development of regression after both FS-LASIK and SMILE, and in Cox proportional-hazard modelling pre-operative spherical equivalent appears as a significant predictor alongside the designed optical zone. It fits the epithelial mechanism directly: a larger correction means more tissue removed and more shape change for the surface layer to remodel against. Older age is also associated with higher enhancement rates in recent series. None of that rules anybody out of anything — it shifts which procedure is the sensible one and how carefully the treatment should be planned, and both of those are determined by the operating surgeon after full examination.
Can anything be done before surgery to reduce the chance of regression?
Most of the predictors are facts about your eye that cannot be changed — your prescription, your age, your corneal measurements. One is different, and it is worth knowing about because it is decided rather than discovered: the designed optical zone. A smaller optical zone contributes significantly to regression after both FS-LASIK and SMILE and appears as a significant predictor in the same modelling as pre-operative refraction; one high-myopia SMILE protocol specifies zones in the 6.2 to 6.7 mm range. The zone is chosen when your treatment is planned, and it trades against tissue depth, so it is a genuine question to raise at consultation rather than a criticism to make afterwards: what optical zone is planned, what made that size the right one for this eye, and what it costs in tissue. Beyond that, the honest answer is that treatment planning is the surgeon's domain and the decisions belong there.
Does regression mean I need a second operation?
Not automatically, and the sequence matters more than the answer. Most refractive drift is small, and the gap between the studies that count regression and the studies that count enhancements exists precisely because a large majority of drift is never operated on. Whether a second procedure is appropriate depends on how big the change is, whether it has been stable across more than one measurement, how much corneal tissue remains, which procedure you originally had, and whether the ocular surface was settled when the measurement was taken — all of which are assessed by an examining surgeon rather than deduced from a page. What you can do is make sure the question is asked in the right order: settle the surface, repeat the refraction, compare against your own documented post-operative best rather than against the result you were hoping for, and only then discuss treatment. The eligibility window, the terms and the cost of a re-treatment are contractual questions covered on our enhancement and aftercare page.
If I had surgery abroad, does regression become harder to deal with?
Harder to establish rather than harder to treat, and the difference is entirely about records. Regression is a claim about change, and change cannot be shown from a single reading taken by somebody who never saw your eye before. The period in which regression usually declares itself — most of the first twelve to eighteen months — is the period you spend back home, and surgery makes your pre-operative refraction permanently unmeasurable, so it exists only in the operating hospital's file. Leave China holding that pre-operative refraction in writing, the treatment parameters actually delivered including the planned optical zone, every post-operative refraction with its date, and a post-operative topography print-out. Then have your refraction recorded once a year at home at roughly the same time of day. Three dated readings turn an impression into evidence a local ophthalmologist can act on; one reading and a memory turn it into an argument.