Secondary Logo

Journal Logo


Comparison of photorefractive keratectomy and laser in situ keratomileusis for the treatment of compound hyperopic astigmatism

El-Agha, Mohamed-Sameh H. MDa; Bowman, Wayne R. MDa; Cavanagh, Dwight MD, PhDa; McCulley, James P. MDa,*

Author Information
Journal of Cataract & Refractive Surgery: May 2003 - Volume 29 - Issue 5 - p 900-907
doi: 10.1016/S0886-3350(02)02039-4
  • Free


The safety and efficacy of excimer laser surgery for low to moderate hyperopia with astigmatism is well established. Both photorefractive keratectomy (PRK)1–3 and laser in situ keratomileusis (LASIK)4–9 have been successfully used to surgically correct compound hyperopic astigmatism. Photorefractive keratectomy for spherical hyperopia (H-PRK) involves removal of the corneal epithelium followed by removal of a negative meniscus of corneal stroma from the peripheral cornea to induce central corneal steepening.10 In PRK for compound hyperopic astigmatism (HA-PRK), the spherical component is addressed as described with additional removal of tissue in the flatter (more hyperopic) meridian to address the astigmatic component. In LASIK, whether for spherical hyperopia (H-LASIK) or compound hyperopic astigmatism (HA-LASIK), tissue removal is done in an identical manner but under a partial-thickness corneal flap.

Previous studies of HA-PRK and HA-LASIK have involved only 1 of the 2 procedures without direct comparison of the modalities in comparable groups. Such a direct comparison between PRK and LASIK has been reported for spherical myopia,11–13 myopic astigmatism,14 and spherical hyperopia.15 In this study, for the first time, we prospectively compared 2 groups of patients with comparable degrees of compound hyperopic astigmatism that had HA-PRK or HA-LASIK. All the surgery was performed in 1 center by the same 2 experienced keratorefractive surgeons using the same laser.

Patients and Methods

From September 1998 to April 1999, 41 eyes of 27 consecutive patients had HA-PRK as part of a prospective noncomparative trial for U.S. Food and Drug Administration (FDA) approval of the procedure. At this time, HA-LASIK was not available as an option to these patients. Two surgeons performed the HA-PRK procedure (R.W.B., 29 eyes; H.D.C., 12 eyes). From September 1999 to November 1999, 24 eyes of 15 consecutive patients had HA-LASIK, also as part of a prospective noncomparative FDA trial for approval of HA-LASIK. The same 2 surgeons performed the HA-LASIK procedure (R.W.B., 13 eyes; H.D.C., 11 eyes). Although patients were not randomized to either procedure, the eligibility criteria as well as the schedule of data collection were identical in the 2 groups.

Both prospective consecutive case series were performed with the approval of the University of Texas Southwestern Medical Center Institutional Review Board and in accordance with the Declaration of Helsinki guidelines for human research. All cases were performed at Zale Lipshy University Hospital Laser Center for Vision, University of Texas Southwestern Medical Center. Laser treatment was delivered in all cases using a Visx Star S2 excimer laser. In the HA-LASIK cases, the corneal flap was created using a Hansatome® microkeratome (Bausch & Lomb).

Preoperative Evaluation

Eligibilty criteria for HA-PRK and HA-LASIK were age greater than 21 years, stable refraction, normal ocular examination, spherical error ranging from +1.00 to +6.00 diopters (D), and refractive cylinder ranging from +0.50 to +4.00 D. No patient had a history of ocular surgery, corneal infection, systemic diseases, or medications that could affect corneal healing. Patients wearing contact lenses were required to remove soft contact lenses for 7 to 10 days and gas-permeable lenses for 2 to 3 weeks before the initial evaluation.

The preoperative examination included uncorrected visual acuity (UCVA), best spectacle-corrected visual acuity (BSCVA), manifest and cycloplegic refractions, corneal thickness by ultrasonic pachymetry, manual keratometry, and corneal topography by computerized videokeratography. When a significant difference was found between the manifest and cycloplegic refractions, postcycloplegic refraction was performed at a subsequent visit, aiming to “push” the manifest refraction closer to the cycloplegic refraction. Slitlamp examination, pneumotonometry, and dilated fundus examination were also performed.

HA-PRK: Surgical Technique and Postoperative Regimen

Topical anesthesia was achieved using 3 drops of proparacaine 1% several minutes apart. A Merocel® sponge (Medtronic Solan) soaked in proparacaine was applied directly to the corneal surface for 1 minute to loosen the corneal epithelium, after which the epithelium was removed using an Amoils brush. After epithelial removal (and during laser ablation), fluid visible on the corneal stroma was removed with a Merocel sponge. The laser was fired at a fluence of 160 mJ/cm2 and a repetition rate of 10 Hz, with an ablation diameter of 9.0 mm and an optical zone of 5.0 mm. Upon completion of laser ablation, ciprofloxacin 0.3% ophthalmic solution was instilled in the eye and a bandage contact lens was fitted to the cornea.

The patient was discharged on a postoperative regimen of ciprofloxacin 0.3% and rimexolone 1% ophthalmic solutions, both 4 times a day. Hydrocodone 5 mg/acetaminophen 500 mg tablets were prescribed, 1 tablet every 4 to 6 hours by mouth as necessary for pain.

Patients were seen on postoperative days 1 through 5 to ensure that epithelialization was proceeding normally. One week after surgery, ciprofloxacin was discontinued; rimexolone was gradually tapered and discontinued over the ensuing 3 months.

At 1 week and 1, 3, 6, and 9 months, UCVA, manifest refraction, BSCVA, keratometry readings (K-readings), and corneal status were determined. In the first week, patients were formally asked about the degree of pain they experienced. At each visit, they were asked to grade their pain as “no pain,” “mild,” “moderate,” or “severe.” The 4 grades were given an arbitrary numeric score of 0 to 3, respectively. The day(s) of maximum pain and the maximum pain score (ie, the highest pain score recorded at any visit) were recorded for each eye.

HA-LASIK: Surgical Technique and Postoperative Regimen

In all HA-LASIK patients, a nomogram adjustment was made to the desired refractive correction, which resulted in additional plus being entered into the laser. As in HA-PRK, topical anesthesia was administered and the cornea was centered and focused under the excimer laser. Flap alignment marks were placed on the cornea using a marker coated with methylene blue. A 9.5 mm Hansatome suction ring was used in all cases. The ring was positioned on the cornea with a slight superior decentration; suction was activated and intraocular pressure was verified to be at least 65 mm Hg with a Barraquer applanation tonometer or a pneumotonometer.

The microkeratome was used to create a 180 μm thick flap. The flap was reflected superiorly, exposing the stromal bed. Laser ablation was applied using the same parameters as in H-PRK. Care was taken to ensure that the stromal bed was dry before and during the ablation and that the undersurface of the flap was protected from the laser ablation. After laser ablation was completed, a drop of ciprofloxacin 0.3% was placed on the stromal bed and the corneal flap was replaced. The interface was irrigated with a balanced salt solution; the flap was painted back into position using a wet Merocel sponge and allowed to seat into place for 5 minutes. Adequate flap position was verified at the slitlamp immediately after surgery, and the patient was discharged with a clear plastic shield. Postoperative medication consisted of ciprofloxacin 0.3% and rimexolone 1% 4 times a day; both were discontinued at 1 week. Hydrocodone 5 mg/acetaminophen 500 mg tablets were prescribed in a dosage of 1 tablet every 4 to 6 hours by mouth as necessary for pain.

Patients were seen 1 day, 1 week, and 1, 3, 6, and 9 months after surgery at which times UCVA, manifest refraction, BSCVA, K-readings, and corneal flap status were determined. A numeric pain score (as described above) was obtained for each eye at the 1-day and 1-week visits.

In both groups, surgically induced astigmatism was calculated as described by Holladay and coauthors.16 Thus, for each patient, the astigmatism induced by the surgery was represented by a vector having a magnitude and angle. To assess deviation from the intended astigmatic correction, the magnitude error and angle of error were also calculated for each patient as described by Alpins.17 Using this method, positive values of the angle of error indicate a clockwise error, whereas negative values indicate a counterclockwise error. For the angle of error, both the absolute and the arithmetic means were calculated.

Statistical Analysis

The preoperative group means were compared using the Student t test. The postoperative means that were measured at only 1 time point were also compared using the Student t test. The repeated mean measurements of postoperative parameters (residual spherical equivalent [SE], mean average K-readings, residual astigmatism) were compared between groups and within each group using a 2-way repeated-measures analysis of variance (Bonferroni test). Proportions were compared using the Fisher exact test and the chi-square test. A P value less than 0.05 was considered statistically significant.


The demographic data and the mean preoperative refractive error in the 2 groups were comparable and are summarized in Table 1. The follow-up was 9 months in both groups with greater than 90% accountability at all time points (Table 2).

Table 1
Table 1:
Demographic data and preoperative refractive error
Table 2
Table 2:
Number of eyes (percentage of total) seen at each time point

The mean maximal pain score was 1.95 ± 1.19 in the HA-PRK patients and 0.84 ± 1.12 in the HA-LASIK patients; the difference was statistically significant (P=.0014). In HA-PRK patients, the pain usually began after a latent period of 1 to 2 days, peaked around 3 days (usually moderate to severe), and lasted for 5 to 7 days. In HA-LASIK patients, pain (if any) occurred a few hours after surgery, was usually mild, and lasted 6 to 8 hours. All HA-PRK patients required systemic analgesic agents for pain; HA-LASIK patients did not.

Figure 1 shows the frequency distribution of UCVA in the 2 groups at each postoperative evaluation and compares the proportion of eyes with a UCVA of 20/20 or better. The proportion of patients with this UCVA level was consistently smaller in the HA-PRK group. The difference between the 2 groups was greatest and most significant (P<.001) at 1 month. Over time, the difference diminished; by 9 months, the between-group difference was neither clinically nor statistically significant.

Figure 1.
Figure 1.:
(El-Agha) Uncorrected visual acuity over time.

Figure 2 shows the change in the mean residual SE over time. Overall, the differences between the 2 groups were statistically significant (P=.013). The P values of group comparisons at each time point are also shown. The HA-PRK procedure was associated with an initial myopic overshoot that peaked at 1 month (P<.001 at 1 month, P=.016 at 3 months). This was followed by regression of the surgical effect with a slow hyperopic drift. The HA-LASIK procedure had no myopic overshoot and stabilized by 1 month, with little change over the ensuing months.

Figure 2.
Figure 2.:
(El-Agha) Mean SE over time.

To assess stability, the P values were computed for the change over time within each group. With HA-PRK, there was a significant change in the mean SE from 1 to 3 months and from 3 to 6 months (P<.001 in both cases); from 6 to 9 months, the change was not statistically significant (P=.212). With HA-LASIK, there was no significant change during the study period (P=1.0 between all time points). From these data, it is evident that HA-LASIK stabilized within 1 month and HA-PRK was stable by 6 months.

Figure 3 shows the change in the mean K-readings in both groups over time. Overall, there was a statistically significant difference between the groups. The between-group difference was not statistically significant at any time point. There was an initial temporary steepening of the cornea with HA-PRK, which appeared to coincide with the initial myopic overshoot. However, over time this change was not statistically significant between time points (P=1.0). The mean K-readings after HA-LASIK were relatively stable.

Figure 3.
Figure 3.:
(El-Agha) Mean K-readings over time.

Figure 4 shows the achieved versus intended correction (in terms of SE) in both treatment groups at 9 months. Eighty percent of HA-PRK eyes and 86.4% of HA-LASIK eyes were within ±1.00 D of plano (intended) SE (P=1.00).

Figure 4.
Figure 4.:
(El-Agha) Achieved versus intended correction (SE) at 9 months.

Figure 5 shows the mean astigmatic error before and after surgery. There was a similar amount of residual astigmatism after both procedures (P=.562) with early stabilization. (There was no statistically significant change within each group over time [P=.611].) The results are summarized in Table 3.

Figure 5.
Figure 5.:
(El-Agha) Mean astigmatic error over time.
Table 3
Table 3:
Efficacy data for astigmatic component of treatment

Figure 6 shows the change in BSCVA at 9 months in both groups. A chi-square test was not feasible because of the distribution of the data.

Figure 6.
Figure 6.:
(El-Agha) Change in BSCVA at 9 months.

Peripheral haze developed in 8 HA-PRK eyes (19.5%) and no HA-LASIK eye. The haze was minimal to mild and was either transient or persistent. Table 4 shows the characteristics of eyes that developed haze and the final refractive outcome. No other complications or adverse events occurred in either treatment group.

Table 4
Table 4:
Characteristics of PRK patients who developed haze


Laser in situ keratomileusis has become the procedure of choice for patients and surgeons because of less pain and more rapid visual rehabilitation than with PRK.11–15 Nevertheless, many patients are not good candidates for LASIK because of anatomical considerations such as a small corneal diameter, an excessively flat or steep cornea, or epithelial basement membrane dystrophy. For these patients, PRK is the only keratorefractive surgical option. Therefore, PRK continues to have a role in keratorefractive surgery.

In our study, we were able to demonstrate that HA-PRK is associated with a higher degree and longer duration of pain, which incapacitated patients for several days. Pain after HA-LASIK, if any, was usually mild and disappeared within a few hours of surgery.

Recovery of visual acuity was more rapid after HA-LASIK than after HA-PRK, primarily because of the initial myopic overshoot that occurred after PRK. It is notable that at 9 months, however, the UCVA data in both groups were similar. Therefore, despite the initial lag of visual recovery after HA-PRK, there is a trend for the UCVA to improve over time.

We reported an initial myopic overshoot after H-PRK,15 and we have found similar changes after HA-PRK. This temporary myopic shift is probably the result of an aberrant epithelial healing response that results in increased central epithelial thickening or relative peripheral thinning, which reaches a steady state after a process of remodeling. Although this has not been documented in vivo, it would be possible to monitor corneal epithelial thickness after surgery using in vivo techniques such as tandem scanning confocal microscopy or high-frequency ultrasound. It is also notable that the mean K-readings were higher in the HA-PRK group at 1, 3, and 6 months but essentially the same as those in the HA-LASIK group at 9 months. This initial steepening of the K-readings thus coincided with the early myopic overshoot after HA-PRK.

Eyes having HA-PRK stabilized by 6 months, whereas those having HA-LASIK stabilized within the first month of surgery. It is striking that the initial myopic overshoot after HA-PRK involved only the spherical component, whereas the astigmatic component of the treatment appeared to stabilize early after both procedures.

The efficacy of astigmatic correction in both procedures was similar. Both were associated with some undercorrection of magnitude (mean ≈ 0.3 D). There was a slightly higher angle of error (absolute mean) associated with HA-LASIK, which may be a function of the LASIK flap. (Our most recent nomogram for LASIK takes into account the location of the hinge of the LASIK flap.) The arithmetic mean angle of error was less than 1 degree in both groups, indicating that there was no consistent error attributable to misalignment of the laser.

The BSCVA data in both groups are similar. However, slightly more HA-PRK eyes lost BSCVA, whereas slightly more HA-LASIK eyes gained BSCVA. The BSCVA data shown are for 9 months. The 5 eyes in the HA-PRK group that lost 2 lines of BSCVA were followed for 12 months. At that point, 4 of the eyes had gained 1 or 2 lines of BSCVA since the previous follow-up. This is consistent with the general trend for visual results to improve over time after PRK.

Mild peripheral haze, sparing the visual axis, occurred with HA-PRK but not with HA-LASIK. It tended to occur with higher attempted corrections and usually appeared at 3 to 6 months. It was either transient or persisted until the last follow-up. Persistent haze was not always associated with regression of the surgical effect.

Our study demonstrates that HA-LASIK is associated with less pain, faster visual recovery, and more rapid stabilization than HA-PRK. Some HA-PRK patients developed peripheral haze, which did not occur with HA-LASIK. The HA-LASIK technique also appears to be associated with less BSCVA loss, although we were unable to substantiate this statistically. A significant concern with both procedures, perhaps more with HA-PRK, is a continued hyperopic drift of the refractive SE refraction, which has also been shown to occur as part of the normal aging process. Based on our intermediate-term data, the current surgical procedure of choice for compound hyperopic astigmatism is HA-LASIK. Patients can be offered HA-PRK with the understanding that there will be more pain, delayed visual recovery, and a possibility of peripheral haze. The long-term stability of both procedures requires further study.


Patrick Ladage, PhD, assisted with the statistical analysis.


1. Dausch DGJ, Klein RJ, Schröder E, Niemczyk S. Photorefractive keratectomy for hyperopic and mixed astigmatism. J Refract Surg 1996; 12:684-692
2. Vinciguerra P, Epstein D, Radice P, Azzolini M. Long-term results of photorefractive keratectomy for hyperopia and hyperopic astigmatism. J Refract Surg 1998; 14:S183-S185
3. Vinciguerra P, Epstein D, Azzolini M, et al. Algorithm to correct hyperopic astigmatism with the Nidek EC-5000 excimer laser. J Refract Surg 1999; 15:S186-S187
4. Argento CJ, Cosentino MJ, Biondini A. Treatment of hyperopic astigmatism. J Cataract Refract Surg 1997; 23:1480-1490
5. Chayet AS, Magallanes R, Montes M, et al. Laser in situ keratomileusis for simple myopic, mixed, and simple hyperopic astigmatism. J Refract Surg 1998; 14:S175-S176
6. Knorz MC, Liermann A, Jendritza B, Hugger P. LASIK for hyperopia and hyperopic astigmatism–results of a pilot study. Semin Ophthalmol 1998; 13:83-87
7. Arbelaez MC, Knorz MC. Laser in situ keratomileusis for hyeropia and hyperopic astigmatism. J Refract Surg 1999; 15:406-414
8. Barraquer CC, Gutiérrez MAM. Results of laser in situ keratomileusis in hyperopic compound astigmatism. J Cataract Refract Surg 1999; 25:1198-1204
9. Lindstrom RL, Hardten DR, Houtman DM, et al. Six-month results of hyperopic and astigmatic LASIK in eyes with primary and secondary hyperopia. Trans Am Ophthalmol Soc 1999; 97:241-255; discussion, 255–260
10. Dausch D, Klein R, Schröder E. Excimer laser photorefractive keratectomy for hyperopia. Refract Corneal Surg 1993; 9:20-28
11. El-Maghraby A, Salah T, Waring GO III, et al. Randomized bilateral comparison of excimer laser in situ keratomileusis and photorefractive keratectomy for 2.50 to 8.00 diopters of myopia. Ophthalmology 1999; 106:447-457
12. El Danasoury MA, El Maghraby A, Klyce SD, Mehrez K. Comparison of photorefractive keratectomy with excimer laser in situ keratomileusis in correcting low myopia from −2.00 to −5.50 diopters; a randomized study. Ophthalmology 1999; 106:411-420; discussion by JH Talamo, 420–421
13. Steinert RF, Hersh PS. Spherical and aspherical photorefractive keratectomy and laser in-situ keratomileusis for moderate to high myopia: two prospective, randomized clinical trials; Summit Technology PRK-LASIK Study Group. Trans Am Ophthalmol Soc 1996; 96:197-221; discussion, 221–227
14. Fraunfelder FW, Wilson SE. Laser in situ keratomileusis versus photorefractive keratectomy in the correction of myopic astigmatism. Cornea 2001; 20:385-387
15. El-Agha MSH, Johnston EW, Bowman RW, et al. Excimer laser treatment of spherical hyperopia: PRK or LASIK? Trans Am Ophthalmol Soc 2000; 98:59-66; discussion, 66−69
16. Holladay JT, Cravy TV, Koch DD. Calculating the surgically induced refractive change following ocular surgery. J Cataract Refract Surg 1992; 18:429-443
17. Alpins NA. A new method of analyzing vectors for changes in astigmatism. J Cataract Refract Surg 1993; 19:524-533
© 2003 by Lippincott Williams & Wilkins, Inc.