The combination of the VISX Star S4 Excimer Laser System and Wavescan Wavefront system (labeled as CustomVue Treatment) was recently approved for wavefront-guided LASIK to reduce myopia and myopic astigmatism. Specifically, the approval was for myopic astigmatism to –6.00D MRSE, with cylinder between 0.00D and –3.00D. Surgeons and staff will be asking questions about the approvals, particularly in relation to Alcon’s CustomCornea labeling.
Get your facts first, and then you can
distort them as much as you please.
Mark Twain (1835 - 1910)
Why the quote? Well, numbers can be misleading, and companies are always likely to report their results in the best light possible. It helps to be able to unravel the confusion by working your way back to the facts and evaluating them. That process provides answers that stand up to scrutiny and may invite questions as to the real meaning of the “reported figures”.
In that spirit, the following is an attempt to get back to the facts behind the VISX labeling. Some sample questions you are likely to hear, and some appropriate answers based on these facts, are provided below. In addition, there are several questions that you might want to ask the surgeon or staff.
All information presented here is drawn from the Professional Use Information manuals for the Alcon CustomCornea and VISX CustomVue LASIK procedures. Doesn’t VISX claim 98% of patients were 20/20 one year after their surgery in the clinical trial?
Doesn’t VISX claim 79% of patients had postoperative UCVA the same or better than their preoperative BCVA?
Doesn’t VISX claim improvements in postoperative symptoms?
Doesn’t VISX have a wider approval range?
Questions worth asking:
Why are wavefront aberrations never discussed in the VISX materials?
Why does the Wavescan only accept output from a 6.0 mm pupil?
Why does CustomVue not account for registration of the wavefront and/or patient rotation at the time of surgery?
Questions worth knowing the answer to:
1. Doesn’t VISX claim 98% of patients were 20/20 one year after their surgery in the clinical trial?
Wow! That’s impressive. Worth looking into… and worth questioning. It is important to note at the outset that no patients in the trial had a best-corrected VA worse than 20/20, or a PreVue (test lens) acuity of worse than 20/20+3. This is unlikely to be considered a ‘typical population’, so the percentage of 20/20 is likely to be higher than typical.
The number of patients evaluated in the time period in question (one year) was 86, just about 25% of the total number of patients enrolled in the clinical trial. A total of 248 eyes were not yet eligible for a one-year visit when the data were submitted, and 13 were discontinued due to retreatments over the course of the study. (Eliminating eyes that required retreatment is a pretty sure way to increase 20/20 percentages.)
The three month visit was the last with no discontinuations, and the percentage 20/20 was 88.4%. The six month visit was the last visit for which a reasonable percentage of patients had been followed (277/351, or about 80%). The percentage of 20/20 reported at that visit was 93.9%, though again 12 eyes were ‘discontinued’ due to retreatment. Presuming they were retreated because they weren’t 20/20, the percentage 20/20 at 6 months would drop to under 90% (consistent with the three month results).
The bottom line? Surgeons hoping for “percentage 20/20” in the high 90’s should operate on patients with excellent preoperative acuities, and consider leaving out of their calculations any patients that they have to later retreat. Neither of these restrictions seems to serve the interests of fair dealing.
In contrast to the misleading mathematics above, the Alcon CustomCornea efficacy data is straightforward and easy to interpret. It includes 139 spherical myopia patients with preoperative best-spectacle corrected VA of 20/25 or better. At 6 months all patients were accounted for, and 80% (111/139) enjoyed 20/20 or better UCVA, despite a known tendency in the FDA trial to undercorrect patients slightly. Recent commercial data suggest even better results. 2. Doesn’t VISX claim 79% of patients had postoperative UCVA the same or better than their preoperative BCVA?
This again is a statistic based on the elimination of retreated patients between the three and six month visits (coming from their 6-month data). Presuming that the 12 retreated patients did not fall into this category, the percentage would drop from 79% (218/277) to 75% (218/289).
This figure of 75% is respectable, but has been achieved by conventional LADARVision users for myopic astigmatism patients, with one surgeon reporting 83% of patients at 3 months with postoperative UCVA better than or equal to their preoperative BCVA.1
The CustomCornea data for myopic astigmatism patients is likely to be available for comparison near year-end.
A six-month evaluation of the clinical study participants showed that four times as many participants were very satisfied with their night vision after the VISX CustomVue procedure, compared to their night vision before with glasses or contacts.
The VISX data summarized for postoperative symptoms is interesting because:
a) 19 eyes did not fill out a survey, and
b) no attempt to measure changes in symptoms was made.
Since the elimination of 12 eyes because of retreatments has already been noted, and the number of patients seen at 6 and 12 months is well below the number originally enrolled, there is no way to make any real sense out of the information available. The people questioned in the preoperative survey were not the same people included in the postoperative survey.
The only true way to evaluate changes due to refractive surgery is to report the change in symptoms for each patient. If you want to generate statistics, it must be based on the statistics regarding the changes in each person, not the average responses from two different groups of patients. For instance, after Alcon’s CustomCornea treatment, 19% of patients reported at 6 months that night driving was better or significantly better than it had been preoperatively.
4. Doesn’t VISX have a wider approval range?
Questions worth asking:5. Why are wavefront aberrations never discussed in the VISX materials?
The CustomVue material indicates several times that the surgery is “wavefront-guided”, but no attempt is made to provide indications of how (or even whether) higher-order aberrations are affected by surgery. In contrast, the Alcon CustomCornea labeling includes a label that indicates “slightly improved optical quality”. In addition, 38% of CustomCornea patients had a reduction in higher-order aberrations pre- to postop at 6 months (vs only 14% of conventional patients), measured over a 6.5mm pupil.
Not only are similar data not reported for CustomVue, there are two relevant observations:
a) the WaveScan sensor measures the higher order aberrations only over the diameter of the patient’s pupil, to a maximum of 6.0mm (a minimum pupil size of 5.0mm is suggested)
b) And, from the labeling: “lthough the Wavescan… measures the refractive error and wavefront aberrations of the human eyes…., in the clinical study for this PMA,the average higher order aberration did not decrease after CustomVue treatment.” (bold italics added) The labeling related to VISX’s wavefront-guided ablation makes no mention of any change in optical quality or aberration reduction, either pre- to postoperative or relative to conventional surgery. 6. Why does the Wavescan only accept output from a 6.0 mm pupil?
The WaveScan device calculates the ablation profile on the basis of data obtained from a 6.0mm (maximum) pupil. Ablation patterns can be calculated for pupils as small as 5.0mm. Since it is well known that the Shack-Hartmann pattern is compromised at the very edge of the detection area, it is likely that useable data are available from a smaller diameter, perhaps 0.5mm smaller than the pupil size over which the wavefront is measured (5.5mm or 4.5mm respectively, in the cases above). Given the fact that the majority of important optical aberrations are in the periphery of the ablation zone, this is a serious limitation. This is even more critical when considering the potential for retreatments, though retreatments are not included in the labeling for either device at this point.
In contrast, the Alcon LADARWave calculates wavefront data from a dilated pupil, typically at least 7mm dilated. The wavefront is calculated over this interval, and the ablation pattern is generated over a 6.5mm zone, appropriately referenced to the center of the patient’s undilated pupil. This provides the best of both worlds: an appropriate reference point for the ablation pattern and an accurately calculated wavefront over a large pupil diameter.
does CustomVue not account for registration of the wavefront
and/or patient rotation at the time of surgery?