"Never has there been such a vast amount of published science and outrage in the history of medicine regarding a procedure that was destined for the garbage heap from the beginning....the incised cornea never heals to it's former strength or regains it's pure optical perfection."


"There remains a great deal of skepticism about radial keratotomy in terms of its appropriateness and safety."

Dr Ivan Schwab, Refractive Keratoplasty - 1987


Eye (1988) 2, 478-483

Radial Keratotomy' Where did it go wrong?




A prospective study was set up to determine the effectiveness, predictability, stability, safety and feasibility within the National Health Service of radial keratotomy. The methods adopted include a new simplified guide to surgery with a predicted accuracy in 84 of 100 operations. Ninety-six percent of eyes with myopia of -6.0D or less preoperatively were seeing 6/12 or bet­ ter six months after surgery. Refraction remained stable within a range 0.5D spherical equiva­lent between the first and sixth postoperative months in 96%, between the first and twelfth postoperative months in 90% and between the first and second years in 100% of eyes examined. There were no cases of progressive hypermetropia. Sixty-four percent of the 61 patients admit­ted to postoperative symptoms. Three percent of eyes lost best corrected vision. The disappoint­ments were few but were analysed in detail. The least reliable were found to be associated with the higher myopes requiring a 3.0 mm clear optical zone.

Despite a rapid expansion of the practice of radial keratotomy (RK) in Europe, it is esti­mated that less than 5% of ophthalmic sur­geons in the United Kingdom practise the technique. Many are unwilling to adopt a new pattern of work into an already over­ loaded schedule and there is a need to assess the efficacy of this relatively new technique particularly in relation to the National Health Service (NHS).

In 1985 a prospective study was set up at Scarborough to answer three questions: 

(1) Keeping techniques as simple as possi­ble, is it safe and possible under the NHS? 

(2) How accurate or predictable is it? 

(3) What is the effect in the long term?

Material and Methods

One hundred consecutive eyes from 61 patients underwent RK between November 1985 and November 1987. Thirty­ three of the patients were males and 28 were females. The age range was 21 to 58 years. Criteria for entry into the study were myopia or myopic astigmatism with a spherical equi­valent of between -2.0D and -1O.OD. cor­rectable visual acuity (VA) to 6/12 or better, a stable refraction over a three year period, no other ocular pathology, failure of contact lens wearing and informed consent.

Of the 100 eyes entered, all have been fol­lowed for a minimum of six months, 60 have been followed for one year and 20 have been followed for two years. The protocol adopted was as follows:

(1) Informed consent. Some 80% of persons making enquiries were eliminated before consultaion by the following procedure. First a letter was sent to the patient stat­ing the necessary criteria for surgery, outlining the risks of surgery and asking for a doctor's referral letter. Specifically the letter stated that the operation worked by weakening the structure of healthy tissue, that there would be increased vulnerability to blunt trauma, that complications such as infection could lead to a worsening of vision, that problems of glare and fluctuating vision after surgery were common, that irregu­lar healing could lead to astigmatism and that there was no guarantee of success. On receipt of the referral letter an appointment was sent to the NHS clinic. On arrival the patient was shown a video for informed consent which detailed the possible problems that could be encoun­tered after surgery and which included a question and answer sequence. A ques­tionnaire was then provided to ensure that the patient had understood the video and this was then checked through by a nurse.

(2) Preoperative consultation included assessment of motivation, refraction, an explanation of likely acuity from 70% correction, assessment of ocular domi­nance, keratometry, applanation tonometry and optical pachymetry using the Mishima-Hedbys centering device.

(3) The surgical plan was based on the Thornton 2% formula I but grossly simplified. The basic myopia was calcu­lated by refraction then modified: For males -0.20 was added for every three years of age under 30 years and subs­tracted for every three years over 30 years. For females -0.20 was added for every three years under 34 years and sub­stracted for every three over 34 years up to the age of 42 years and thereafter a similar rule applied to that of men. -0.10 was added for every 2 mm Hg. of intraocular pressure below 13 mm Hg. and subtracted for every 2 mm Hg. above 17 mm Hg. was added for each keratometry dioptre below 43.00 and subtracted for each full dioptre above 44.20. -0.10 was also added for each 0.02 mm of central corneal thickness above 0.53 mm and subtracted for each 0.02 mm below 0.50 mm.

 In bilateral cases there was a general aim to undercorrect the non­ dominant eye by around 0.50 as the first pro­cedure and for a time interval of at least two months before surgery to the second eye. Only three optical zone (OZ) markers were used. The incision depth using an adjustable diamond blade on a micrometer screw, was aimed at over 95% corneal thickness by adding to the central pachymetry reading 0.07 mm for a 3.00 mm OZ, 0.08 for a 3.5 mm OZ and 0.09 mm for a 4.0 mm OZ. In 16 eyes an astigmatic element in excess of -1.50 necessitated additional relaxing incisions : these were placed perpendicular to the steepest meridian between but not touching the radial incisions and tangential to a 6.0 mm OZ.

(4) Stages of surgery. The fixation axis was marked as recommended in the PERK study, 2 cc lignocaine was injected retrobulbarly, the OZ was marked, 4 or 8 radial cuts were made from the OZ towards the periphery stopping short of the limbus. For cases requiring 're­ deepening', a second pass was made outside the 6.0 mm OZ together with central pricks at the OZ in order to maximise the depth of incision. Sixteen cuts were not practised as a primary pro­cedure. Six eyes required an additional 4 or 8 cuts to improve unaided vision, however the results from surgery in these eyes were recorded for purposes of the study before the additional cuts had been made.

(5) Postoperative care. The eye was padded for one night. The patient was discharged from hospital the following day using g.dexamethasone q.d.s. Patients returned for assessment of acuity, refrac­tion and postoperative symptoms at one month, six months, one year and two years. Assessment of symptoms, whether these were disabling and possible disap­ pointment with the result of surgery were made by an independent assessor using a written questionnaire.


Six months after surgery 86 eyes were seeing 6/12 unaided. 100% of eyes with preoperative myopia less than -3.5D and 94% of eyes with preoperative myopia less than -6.0D were seeing 6/12 unaided. The 14 eyes unable to see 6/12 unaided com­prised:-

(a) intended undercorrection of one eye in four presbyopes (6/18, 6/18, 6/36, 6/24)

(b) three eyes undercorrected where 4 addi­tional cuts later brought the acuity to 6/9 or better in two and to 6/18 in the third (6/18, 6/18, 6/36 respectively)

(c) one high myope who later required an additional 8 cuts to achieve 6/9 unaided (6/60) 

(d) two eyes that remained undercorrected despite in one an additional 8 cuts (6/60, 6/60) 

(e) four eyes with myopia in excess of -9.0D which were never expected to attain 6/12 (6/24, CF, 6/60, 6/60)

At twelve months 84% of the 60 eyes examined were seeing 6/12 or better.

Predictability was determined by analysing the number of results which fell within the range of planned treatment.  Refraction at six months showed that 16 eyes were outside the planned range : nine had been undercorrected (one by more than 1D.) and seven had been overcorrected (two by more than 1D.) There was a spread of unpre­dictability among all degrees of myopia. In bilateral cases where the first eye gave an unpredicted result the parameters were altered for the second eye as it was found that the behaviour of the second eye would pre­dictably follow that of the first eye.

Refraction remained relatively stable  and altered by less than 0.6D. Bet­ween the first and sixth postoperative month in 96 eyes (mean change -O.lD). In 4 eyes there was a decay in effect by more than -0.5D. Between the first and twelfth post­ operative months the refraction altered by less than 0.6D in 90% of the sixty eyes examined at one year and between the twelfth and twenty-fourth months in 100% of the 20 eyes examined at two years.

Symptoms of the 61 patients at six months are analysed in Table VI. In answer to the questionnaire, 39 patients admitted to symptoms although they were sometimes described as being present before surgery as well. Only two patients were disabled by symptoms in that the starburst effect pre­ vented them from driving at night. Six patients showed slight disappointment with the result of surgery and are examined in detail as they highlight well the problems that may confront the RK surgeon:

Case 1. Male aged 36: too long a transverse cut for astigmatism led to increased glare, loss of best cor­rected vision from 6/6 to 6/9 and still a need for dis­tance glasses. In addition he now has hypermet­ropic astigmatism so cannot read with this eye and feels his binocularity has suffered. The error was surgical.

Case 2. Female aged 24: -6. SD. was undercor­rected with 8 cuts and an additional 8 cuts increased her glare, prevented her from driving at night, reduced her best corrected vision from 6/5 to 6/6 and she still required glasses for clear dis­tance vision. Furthermore, the anisometropia resulted in headaches. She was left with the predi­cament of being unwilling to have surgery to the fellow eye which would relieve the headaches, because of possible increase in glare.

Case 3. Female aged 32: -S.SD. -S. 2SD. Bilateral undercorrection with further decay in effect over ensuing year. Left unaided vision improved from 6/36 to 6/18 with 4 additional cuts but associated with loss of best corrected vision from 6/5 to 6/6. Case 4. Female aged 38: -S. OD. Postoperative vis­ion 6/9 unaided but continued decay in effect by 2.SD. over a two year period, the cause appeared to be insufficient depth to the incisions. 

Case 5. Male aged 40: Obsessional neurosis undiagnosed before surgery, dissatisfied with an unaided vision of 6/6 right, 6/9 left.

Case 6. Female aged 2S: -4.SD. corrected to 6/9 unaided but complained of deterioration of vision towards the end of the day.

An increase in astigmatism is a known association of RK. The five cases suf­fering an increase of over l.OD. were highly myopic eyes requiring a 3.0 mm OZ. In no case did the induced astigmatism exceed 2.0D. and all patients were satisfied with the considerable improvement in unaided vision (6/12, 6/12, 6/18, 6/60, 6/12, respectively).

The highest myopes had the most to gain from surgery, but they also had the most to lose as is shown by Table VIII. Thirty three eyes required a 3.0 mm OZ and compared to those requiring 4. 0 or 3. 5 mm OZ, they showed a significantly higher predisposition to unstable refraction, increasing astig­matism, postoperative symptoms and disap­pointment with the result achieved. Although potentially time consuming, by using the methods described it was found that on average only 20 minutes of theatre time per week and three hours of outpatient time per month were required for the study of 100 operations over a two year period. The methods recommended are simple yet effec­tive and feasible in the NHS setting.

So where did it go wrong? The answer to this may be analysed subjectively from the written questionnaire concerning symptoms and disappointment and objectively from the incidence of complications. Fifty five of the 61 patients were pleased with the result of surgery despite the admission of symptoms in 60%. The symptoms were an anticipated part of the procedure, sometimes no worse than before surgery and provided informed consent is obtained should not be considered a reason for discarding the proce­dure. The aim to undercorrect the non­ dominant eye first tended to prevent disap­pointment and covered the eventuality of the unpredictable over-reaction to surgery in 7%  and left the options if under-reac­tion (9%) of either a further 4 or 8 cuts later or allowing residual myopia to offset future presbyopia. It also enabled an altering of parameters for greater accuracy in the second eye.

The six cases showing some disappoint­ment include all three cases of loss of best corrected vision, the only serious complica­tion encountered, and three of the four cases of decay in effect (the fourth was overcor­rected at one month and did not stabilise until three months). The salutary lessons to be learnt from these cases are the importance of surgical accuracy (Cases 1 and 4), the importance of patient education as to possi­ble undercorrection (Cases 2,3,4) and proba­ble postoperative symptoms (Cases 1,2,3,4,6), and the importance of correct assessment of motivation (Case 5). Case 2 also reflects the problems of induced anisometropia in a patient unwilIing to undergo surgery to the second eye.

Finally, it should be remembered that the higher myopes who have most to gain, also have the most to lose.

pasted-file-62 med
pasted-file-60 med
pasted-file-61 med
pasted-file-57 med
pasted-file-58 med
pasted-file-68 med
pasted-file-65 med
pasted-file-63 med

Sight-Threatening Complications After Radial Keratotomy

Before deciding to undergo an elective surgical procedure, patients should receive information regarding the possible benefits, risks, and alternatives to surgery. This is particularly true with keratorefractive surgery, in which an eye that has a visual acuity of 20/20 with spectacles or contact lenses is put under the knife (or laser, as the case may be). Individuals considering refractive surgery will often ask a question such as,"I know there are risks, but I can't go blind, can I?" The answer, of course, is, "Yes, there is a small risk of blindness."

It is clear that blinding complications may ensue after radial keratotomy (RK). One of these potentially blinding complications, globe rupture, is detailed by Vinger and associates in this ensue of the Archives. They reviewed the published literature and searched for additional cases of globe rupture. That incisions in the cornea or elsewhere in the eye predispose to globe rupture following blunt trauma has long been recognized, with the first report of rupture after RK published in 1987. Post surgical ruptures are not unique to incisional refractive surgery; dehiscence of penetrating keratoplasty wounds may result following trauma that occurs many years after the corneal transplant procedure. Similarly, ruptures at the limbus may occur in eyes that have had cataract extraction. Possibly, the frequency of this latter complication is reduced with small-incision surgery. That globe rupture may rarely occur after RK does not condemn this procedure, any more than the same risk condemns penetrating keratoplasty or cataract surgery.

There is no question that deep or full-thickness in cisions structurally compromise the eyewall, at least in the short term. How short is short term? An analysis of the 26 cases summarized by Vinger all for which the postoperative interval was known shows that half of all reported globe ruptures occur within 2 years of surgery, and about two thirds within 3 years of surgery. The median interval between surgery and globe rupture in this group of 26 eyes was 20 months. Visual inspection of this curve suggests that after 3 to 4 years, there does appear to be a decline in the risk for globe rupture. Unfortunately, in the absence of good data regarding the number of eyes that have had RK at different time points, we are not able to know the important denominators, such as the number of eyes at risk of globe rupture at each of these postoperative in tervals.

What about the predisposition to globe rupture from blunt trauma in the long term? Using biomechanical methods, Bryant  measured tensile strength of fully healed RK incisions in a human donor cornea obtained 8 years after surgery. There was no significant difference in tensile strength between the RK cornea strips and the control strips from unincised corneas. On the other hand, a persistent epithelial plug remained in at least one of the incisions, and such plugs appear to create stress concentrations at the incision sites, which may predispose the cornea to rupture. Thus, the increased susceptibility to rupture due to RK incisions may be hard to predict and may depend on factors such as the size of the epithelial plug and the strength of the wound collagen. In addition, incision length may be an important variable, with longer incisions offering less resistance to globe rupture. That these eyes appear to remain at risk for a very long period is supported by the asymptotic appearance of the curve with globe ruptures occurring as late as 6, 7, 8, and even 10 years after surgery. If one makes the assumptions that these globe ruptures originate within the corneal incisions (as stated by Vinger) and that the eyes would not otherwise have ruptured, it appears that eyes that have undergone RK remain at risk of globe rupture for as long as a decade.

At this point, there are no data available that allow us to state with confidence that there is a time beyond which an eye that has had incisional keratotomy is not at in creased risk of globe rupture compared with a normal eye.

As a practical matter, police departments, the military, and similar organizations often ask their physician advisors to determine when it is "safe" to hire a candidate who has had RK. Butler has proposed convales cent periods of 3 to 6 months following refractive surgery, including RK, for strenuous physical activities such as diving. One concern is legal liability, in the event that an employee is blinded as a result of trauma in the course of employment. The ophthalmologist is asked to determine when the risk for globe rupture is no longer in creased, relative to a normal, unoperated-on eye. I be lieve currently available data do not establish the postoperative interval required for return of rupture strength of the eye to preoperative levels. What is the likelihood that a patient will suffer the terrible complication of globe rupture after RK? Unfortunately, we do not have sufficient data to answer this question. The denominator, the number of eyes that have had RK, is unknown; frequently cited estimates range from 1 to 1.5 million individuals, or 2 to 3 million eyes. In addition, we do not know the numerator. There are at least 26 patients who have suffered this complication who now appear in the peer-reviewed literature; we have no way of knowing how many such ruptures have been observed but not reported. With a substantial number of cases of complications such as globe rupture already reported, there may be little incentive for journals to devote limited space to yet another report of such a complication.

There are other important unanswered questions regarding globe integrity after RK:

1. In how many eyes with globe rupture after RK was the trauma severe enough that it would have resulted in globe rupture even in the absence of the corneal incisions? Would the rupture of the RK incisions somehow prevent potentially more devastating complications of significant blunt trauma such as choroidal rupture? 

2. How many postrefractive keratotomy eyes have experienced substantial trauma (eg, sufficient to cause a hyphema) and not ruptured? In the PERK study, 5 in about 700 eyes followed for 10 years, no globe ruptures occurred among patients available for follow-up, although at least two eyes experienced trauma sufficient to cause hyphemas or vitreous hemorrhages.

3. How many eyes with refractive keratotomy have ruptured in a location other than the corneal excisions?

Vinger state that the potential RK patient should be informed of an increased risk of globe rupture following blunt trauma. Obviously, the duty of the ophthalmologist is to fully inform the patient regarding all significant risks, benefits, and alternatives to surgery. Vinger state that patients who have had refractive keratotomies should be advised to wear protective eye wear when engaged in ball sportsor other hazardous activities. The reality is that ophthalmologists already recommend the use of protective eyewear for activities such as squash and tennis, whether or not there has been previous ocular surgery. Unfortunately, one need only visit the local playground or health club to appreciate that these recommendations are routinely ignored.

Returning to the issue of informed consent, the level of risk that an individual will accept before undergoing an elective procedure is patient-specific. As part of the informed consent process, he or she should be told of the risks of undercorrection and overcorrection, glare, diurnal fluctuation, infection, progressive effect (hyper- opic shift), globe rupture, loss of best-corrected visual acuity, and other factors associated with the various surgical options. While we cannot quote exact numerical risks, we can tell our prospective patients that they may be at increased risk of globe rupture for at least a decade after surgery, and they should therefore take reasonable precautions.

PeterJ. McDonnell MD Los Angeles, Calif

Myopia and radial keratotomy: a survey among Norwegian ophthalmologists

1. Anna Midelfart M.D. Article first published online: 27 MAY 2009

Publication History

1. Issue published online: 27 MAY 2009

2. Article first published online: 27 MAY 2009

3. January 10th, 1990.


One hundred and eighty-nine of 200 ophthalmologists in Norway responded to a survey requesting them to report their age, sex, refractive state, use of corrective lenses, and if myopic, their view on radial keratotomy as a possible method to correct their own myopia. According to the answers, 32 (17%) females and 154 (82%) males, with mean age of 49 years, were registered. The reported refractive state was 26.5% emmetropy and 72.0% ametropy. The prevalence of myopia was 45%. The mean refractive status (equivalent sphere) in the right eye was -1.02 ± 2.28 D with a range from -8.5D to +7.25D (n = 184). Of the ametropes, 64.8% used spectacles, 15.3% used both spectacles and contact lenses, whilst 3.6% used only contact lenses for distance vision. With the exception of one, all myopes used corrective lenses. 

Only 2 myopic ophthalmologists responded positively to the question of whether they would consider having radial keratotomy to correct their own myopia. 


0001uS 2