LASIK flap thickness is trickier than you think
Regular ultrasound measurements may be deceiving you about the depth of your microkeratome cuts. Part 1 of a two-part article.
Like many surgeons, when I started performing LASIK in my own practice I derived my nomogram from the formula of Machat and Probst. Take the total corneal thickness, subtract 160 µm for the flap thickness, and subtract the predicted ablation depth. If the result is greater than 250 µm, proceed. If not, consider undercorrection, reduce the treatment zone, or advise the patient against LASIK.
At a conference, I heard several surgeons comment that they tried to make the initial flap thinner with the Automated Corneal Shaper (Bausch & Lomb) to allow for re-treatment if necessary. At the time the 60% rule was suggested as well, so that in a 500 µm cornea one would leave 300 µm.
As I faced scores of eyes in my practice that pushed these upper limits, I became increasingly uneasy that I was doing everything possible to meet these standards.
I was using the gas-turbine, manually turned Moria CB 130 because of my belief that superior flaps reduce microstriae. Using a manually turned microkeratome might be causing flap thickness variability. How thick was I really cutting?
According to Orbscan measurements, the postop thicknesses in many patients’ beds were much thinner than predicted. Although not supported by surface ultrasound, these measurements were disconcerting. (We subsequently learned that after LASIK, Orbscan is confused by the interface.)
Finding a pachymeter
I was beginning to lose sleep. I resolved to buy the best ultrasound pachymeter possible and to measure each patient intraoperatively to confirm what I was doing. Naively I pictured directly measuring the flap, and I began asking for a pachymeter that could measure down to 100 µm.
Everyone pointed me to DGH, which had developed a pachymeter for automated lamellar keratoplasty. This was an upgrade from normal devices, which are limited to measuring only down to the mid-200 microns. The standard method of subtracting the bed thickness seemed flawed to me, because of hydration of the bed from water pulled in with the microkeratome.
I purchased the DGH Pachette II and excitedly put it on top of my laser. It has never left that spot. In fact I have cancelled cases on 1 or 2 days when the probe was damaged and I could not use it.
The reason? Much to my dismay, the 160 µm flap I thought I was making with my slow, smooth pass with the 130 head was consistently thicker by as much as 50 µm. On enhancement, in the dry state, many of my beds were at the 250 µm limit, and I was forced to cancel the treatment. In the spring of 2002 I now have the experience of 11,000 cases, and I would like to summarize what we have learned and what we hope to know better.
In this article I will address the technique we developed and why we believe it may be the preferred method for studying flaps. In a second part of this article, I will address how this technique has informed us on our technique and has led us to advocate for thin flaps. We will also address the next generation of technology for studying flaps and their healing, the Artemis ultrahigh-frequency ultrasound.
The problem of measuring
First, let us discuss the problem of measuring cornea and stroma. The problem we all have when studying flaps is that we are trying to measure a moving target by averaging three measurements over what we are guessing to be the center, or the thinnest part of the bed. All microkeratomes have a tendency to break hemidesmosomal attachments and “slide” the epithelium. To avoid this, some lubrication is placed on the cornea just prior to the pass. This fluid is necessarily pulled into the interface.
Stromal tissue is “dry” and porous, teleologically designed for rapid transfer of nutrients and oxygen across its width. The next time you are doing an enhancement, place a drop of saline on the dry bed after ablation. You can see the area become raised as the water absorbs rapidly into the bed. Place a pachymeter on top of this drop and you will see changes of 15 µm to 40 µm in seconds.
All the data you have read about flap thickness is affected by this problem.
Because of the limits of the older pachymeters, the standard way of measuring of flap thickness has been subtraction. This, unfortunately, has confounded our data. If one uses this method intraoperatively, one will come to the conclusion that flaps are thinner than expected and that the bed after ablation is deeper than expected.
It is our contention that the stromal bed is the worst place to predict flap thickness immediately after the flap is made. One method to partially correct this problem is to the measure the surface, cut the flap, then mea-sure the surface again. The fluid absorbed by the cornea will partially reveal itself through the difference in measurements. You have cut through a sponge. The stromal bed is simply the thickest part of the sponge and therefore tends to lead to the greatest error.
On the other hand, the thinner stromal surface under the flap, though it will still lead to error regarding the exact thickness, tends to lead to a better prediction of the post-ablation depth of the dry bed. Remember, the porous stromal part of the flap does not begin until 50 µm to 60 µm depth.
We have found that the total error is somewhat proportional to the thickness of stroma on each of these surfaces, and the error is affected by the exposure time of fluid into the interface. Slower microkeratomes allow more time for absorbtion and increase the error. Perhaps other issues such as oscillation speed, pressure of the ring and pass speed also enter into the equation.
The error, however great, is always greater in the bed and less in the flap, even as the flap gets thicker. The flap error will overpredict rather than underpredict the post-ablation depth, as is the case by the subtraction method, which is preferable.
The errors generated by all of these methods beg us to be cautious when lasering below 300 µm on primary treatment. If one enhancement of a lifted flap one can measure an adequate bed, then proceeding beyond 300 µm is reasonable.
The use of a high-frequency pachymeter tip on the ablation area raises concerns about keratitis and about the ability of the tip itself to alter the hydration of the central ablation area. For several years, use of an alcohol sponge was our only defense against introducing pathogens into the central cornea. Alcohol is at best cleaning only; it would not protect against the spread of pathogens, especially a small-particle, slow virus. This led to our using a tonometer approach and the requirement that in our busy center we own two probes.
The probes are soaked 10 minutes in Opti-Cide and then rinsed with sterile water. This method reduces my concern about slow virus and larger pathogens being introduced by the probe. It also avoids the bed, and if there is a central effect from contact of the probe we have avoided another confounding point to ablation success. We are not sure that the tip drives fluid or squeezes fluid out of the stroma immediately after flap creation, but if we have an equal or perhaps superior method then we should avoid touching the central cornea.
In the second part of this article I will describe our technique (you can see a video on our Web site, www.glassesfree.com) and present our findings using this novel approach. This approach has led to us to view the dynamic environment of the cornea with a new respect.
The safety and science of our work will require an even better tool, such as the ultrahigh-frequency Artemis, which measures down to 1 µm in three dimensions in steady state, to solidify the findings we have reached using this crude approach. This device will provide some tantalizing data as we move into the era to come.
Note from the editors:
Look for part 2 of this article in an upcoming issue of Ocular Surgery News.For Your Information:
- Richard B. Foulkes, MD, is in private practice at 9730 S. Western Ave., Suite 225, Evergreen Park, IL 60805; (630) 920-5880; fax: (630) 920-8533; e-mail: email@example.com.