comparison of central corneal thickness as measured by non-contact specular microscopy and...

Original Article

Comparison of central corneal thickness as measured bynon-contact specular microscopy and ultrasound pachymetrybefore and post LASIKMing-hui Zhao MD, Jun Zou MD, Wen-qing Wang MD and Jun Li MDDepartment of Ophthalmology, Shanghai Jiaotong University Affiliated Sixth People’s Hospital, Shanghai, China

ABSTRACT

Purpose: To compare central corneal thickness (CCT)measurements made by ultrasound pachymetry and non-contact specular microscopy in eyes before and post laser insitu keratomileusis (LASIK).

Methods: Forty eyes of 20 consecutive refractive surgerycandidates were included in this study. The CCT in botheyes was measured before and 3 months after LASIK witha non-contact specular microscope and a ultrasoundpachymeter.

Result: Both before and after LASIK, the difference of themean CCT between non-contact specular microscopy andultrasound pachymetry was statistically significant (P < 0.001before surgery, and P = 0.02 after surgery). The differencebetween them was -14.41 mm and -3.69 mm, respectively. Inboth situations, non-contact specular microscopy and ultra-sound pachymetry measurements were highly correlated (r =0.959, before surgery, and r = 0.979 after LASIK). Bland–Altman analysis showed that either before or post LASIK, the95% limits of agreement have a span of approximately 30 mm.Preoperatively, the mean standard deviation of repeatedmeasurements was 4.26 mm by ultrasound pachymetry, and4.02 mm by non-contact specular microscopy,whereas it was3.83 mm by ultrasound pachymetry, and 3.89 mm by non-contact specular microscopy, postoperatively. No statisticallysignificant differences were found between the meanstandard deviations by these two devices in both situations(P = 0.561 before surgery, and P = 0.849 after LASIK).

Conclusions: Both before and after LASIK, non-contactspecular microscopy is not in complete agreement with

ultrasound pachymetry in measuring CCT. Both devicesprovided comparable repeatability of measurements.

Key words: central corneal thickness, non-contact specularmicroscope, ultrasound pachymetry.

INTRODUCTION

Central corneal thickness (CCT) measurement has becomeincreasingly important in ophthalmic practice. In order toprevent keratoconus happening after the refractive surgery,corneal thickness measurement has become mandatory inclinical practice. Corneal thickness is also an indicator forthe corneal endothelial pump and barrier function.1 Withimproved diagnostic and therapeutic possibilities in glau-coma, the impact of corneal thickness variations on tonom-etry measurements has been emphasized. Patients with lowcorneal thickness may be underdiagnosed and undertreated,whereas the opposite exists in those with high cornealthickness.2

Currently, the most commonly used method for cornealthickness measurement is ultrasound pachymetry. It has beenused for over 30 years.3 Ultrasound pachymetry is an effi-cient and accurate way to measure corneal thickness, and isconsidered as the current gold standard. Its advantagesinclude easy to use, portability and low cost. However, it hasthe disadvantages of the probe directly contacting thecornea with need for topical anaesthesia. Moreover, its accu-racy is dependent upon the perpendicularity of the probeplaced on the cornea.

Non-contact specular microscopy is another commonlyused method for corneal thickness measurement. Ultrasoundpachymetry measurements depend on the reflection ofsound, whereas non-contact specular microscopy measure-ments depend on the reflection of light waves from the

� Correspondence: Professor Jun Zou, Department of Ophthalmology, Shanghai Jiaotong University Affiliated Sixth People’s Hospital, Shanghai, 200233, China.

E-mail: [email protected]

Received 11 June 2007; accepted 25 September 2007.

Clinical and Experimental Ophthalmology 2007; 35: 818–823doi: 10.1111/j.1442-9071.2007.01633.x

© 2007 The AuthorsJournal compilation © 2007 Royal Australian and New Zealand College of Ophthalmologists

mailto:[email protected]

anterior and the posterior corneal surfaces. Its advantagesinclude non-invasiveness, easy to operator, examiner inde-pendency and providing corneal endothelial cells’ imagessimultaneously.4,5

To avoid any kind of corneal epithelial damage or flapdisplacement, physicians usually do not perform ultrasoundpachymetry during the early postexcimer laser photoabla-tion stage. For its non-invasive advantage, non-contactspecular microscopy can be an ideal technique for studyingthe corneal thickness after photorefractive keractectomy. Inthis study, we compared CCT measurements before and postLASIK, using non-contact specular microscopy and ultra-sound pachymetry.

METHODS

Forty eyes of 20 consecutive refractive surgery candidateswith mean age of 26.56 � 5.22 years were included in thestudy. Ethical approval for the study was obtained from localmedical ethics committee. All subjects were volunteers andinformed consents were obtained.

The mean refractive error (spherical power + half cylinderpower) of all participants was -6.20 � 1.65 dioptres (D),ranges from -3.75 to -10.00 D. Exclusion criteria were asfollows: any abnormality in at least one eye (e.g. keratoconusor glaucoma); contact lens wear in the previous 2 weeks; andbefore ocular and/or corneal surgery.

LASIK had been performed by a single surgeon (JZ) at asingle centre using the Moria M2 microkeratome (MoriaCorp, Antony, France), and subsequent photoablation wasconducted with the Allegretto Wave Eye-Q excimer laser(Wavelight Laser Technologie AG, Erlangen, Germany).

The CCT in both eyes was measured before and 3 monthsafter LASIK with a non-contact specular microscope(Topcon SP-2000P, Topcon Ltd, Tokyo, Japan) and a ultra-sound pachymeter (Tomey SP-3000, Tomey Ltd, Nagoya,Japan). In the non-contact specular microscopy study, thesubject was positioned with his or her chin in a cup andforehead against a headband. CCT and endothelial celldensity were measured simultaneously. Only CCT readingswere used. Each eye was measured three times. The mean ofthree readings was obtained for further analysis. Immediatelyafterward, the cornea was anaesthetized with one drop oftopical proparacaine hydrochloride 0.5% and three consecu-tive CCT measurements were made by ultrasound pachym-etry, which was set at a frequency of 1641 m/s. The probewas aligned perpendicular to the centre of the cornea.The average of three readings was recorded as ultrasoundpachymetry result. All measurements were made by the sameophthalmologist.

To determine the reproducibility of measurements for thetwo machines, the standard deviation (SD) for the threerepeated measurements obtained from each eye was deter-mined, and then the mean of the SDs from all eyes for eachmachine was calculated. Comparison of the mean CCTvalues for the two instruments was conducted by t-test. Pear-son’s correlation analysis was used to investigate the corre-

lation between the two methods. A plot of differencesbetween the two techniques was carried out according to themethod described by Bland and Altman.6 Significance wasassumed at P < 0.05.

RESULTS

The average preoperative CCT was 543.04 � 24.71 mm forultrasound pachymetry, and 528.63 � 25.41 mm for non-contact specular microscopy. Postoperatively, the ultrasoundpachymetry average was 426.72 � 40.84 mm, whereas thenon-contact specular microscopy average was 423.03 � 39.11 mm. The tissue ablation amount measured by ultrasoundpachymetry was 116.32 � 35.07 mm, and 105.60 � 32.74 mm by non-contact specular microscopy. Good correla-tions were found between these ablation amount and theablation depth according to the manufacturer’s prediction(mean: 109.57 � 25.58 mm) (r = 0.900, P < 0.001 for ultra-sound pachymetry, and r = 0.859, P < 0.001 for non-contactspecular microscopy, Pearson’s correlation analysis).

In both pre- and post-LASIK measures, non-contactspecular microscopy results were less than the ultrasoundpachymetry, but excellent correlations were found betweenthem (r = 0.959, P < 0.001 before surgery, and r = 0.979,P < 0.001 post surgery, Figs 1,2). Preoperatively, the differ-ence between non-contact specular microscopy and ultra-sound pachymetry was -14.41 mm (95% limits of agreement,-28.44 to -0.38 mm. Postoperatively, the difference betweenthem was -3.69 mm (95% limits of agreement, -19.72to 12.34 mm. All differences were statistically significant(P < 0.001 before surgery, and P = 0.02 after LASIK). Fig-ures 3 and 4 were Bland–Altman plots of the differences inthe mean CCT values obtained by the two methods beforeand after LASIK, respectively. Both figures showed that there

Figure 1. A significant correlation was found between centralcorneal thickness measurements by non-contact specular micro-scopy and ultrasound pachymetry prior to surgery. The best fit lineis designated by the dotted line and the line of equivalence (y = x)by the solid line. LASIK, laser in situ keratomileusis.

Corneal pachymetry before and after LASIK 819


were significant differences between the two instruments,and no apparent relationships could be identified betweenthe differences and the mean CCT values. Preoperatively,Figure 3 clearly demonstrated a systematic bias towardlower CCT measurements using a non-contact specularmicroscope. Postoperatively, Figure 4 showed that the meandifference between the two devices was less than that inpreoperative data, The 95% limits of differences were

between -19.72 and 12.34 mm, which indicated that in somecases non-contact specular microscopy may give readingsless and in some greater than those obtained by ultrasoundpachymetry.

Preoperatively, the repeated measurements of the CCTresulted in a mean SD of 4.26 mm for ultrasound pachymetry,and 4.02 mm for non-contact specular microscopy. Postop-eratively, the repeated measurements of the CCT resulted ina mean SD of 3.83 mm for ultrasound pachymetry, and3.89 mm for non-contact specular microscopy. Both beforeand after LASIK, the difference between the ultrasoundpachymetry and non-contact specular microscopy wasnot statistically significant (P = 0.561 before surgery, andP = 0.849 after LASIK, a pairwise t-test).

DISCUSSION

In recent years, new instruments have been introduced forCCT measurements. They provide data in a convenient,non-contact way. Non-contact specular microscopy providesinformation of both central and peripheral cornea thick-ness, and allows for evaluation of endothelial cell densitysimultaneously. In this study, as a single investigator madeall the measurements, the variability of values in repeatedmeasurements of the same eye represented intra-observerreproducibility. Prior to surgery, reproducibility of bothinstruments was quite good. The non-contact specularmicroscope tended to show better repeatability than theultrasound pachymetry, although the difference was not sta-tistically significant (P = 0.561), which agrees with the resultof the Suzuki et al.’s study.7 Bovelle et al.4 reported that the

Figure 2. A significant correlation was found between centralcorneal thickness measurements by non-contact specular micro-scopy and ultrasound pachymetry post LASIK. The best fit line isdesignated by the dotted line and the line of equivalence (y = x) bythe solid line. LASIK, laser in situ keratomileusis.

Figure 3. Bland–Altman plots. The difference in central cornealthickness (CCT) measurements between non-contact specularmicroscopy and ultrasound pachymetry prior to LASIK plottedagainst the mean CCT. Mean CCT measurements = (mean CCTfrom ultrasound pachymetry + mean CCT from non-contact specu-lar microscopy)/2. The mean difference was -14.41 mm, with 95%limits of agreement from -28.44 to -0.38 mm. LASIK, laser in situkeratomileusis; SD, standard deviation.

Figure 4. Bland–Altman plots. The difference in central cornealthickness (CCT) measurements between non-contact specularmicroscopy and ultrasound pachymetry plotted against themean CCT post LASIK. Mean CCT = (mean CCT from ultra-sound pachymetry + mean CCT from non-contact specularmicroscopy)/2. The mean difference was -3.69 mm, with 95% limitsof agreement from -19.72–12.34 mm. LASIK, laser in situ keratom-ileusis; SD, standard deviation.

820 Zhao et al.


non-contact specular microscope provided more consistentmeasurements than the ultrasound pachymetry in normalhuman eyes, possibly as a result of the elimination ofobserver bias induced by probe placement with the ultra-sound pachymetry, whereas Tam and Rootman8 reported thatthe CCT measurements by specular microscopy were signifi-cantly less reproducible than those by ultrasound pachym-etry, but the error levels were both clinically acceptable. Inpost-LASIK eyes, we are not aware of any report of repro-ducibility of CCT measurements using non-contact specularmicroscopy and ultrasound pachymetry. In current study,both ultrasound pachymetry and non-contact specularmicroscopy displayed fairly good reproducibility. Therewas no statistically significant difference between them(P = 0.849).

Former studies have evaluated and compared the CCTvalues using different optical pachymeters and ultrasoundpachymetry, whereas the results were variable.5,7,9–11 Ingeneral, optical readings tended to be slightly thinner thanultrasound pachymetry readings. Bechmann et al.9 found thatthe mean CCT readings with standard retinal optical coher-ence tomography was 51 mm lower than that with ultrasoundpachymetry in normal eyes. Lackner et al.10 compared theCCT readings using Pentacam and ultrasound pachymetry,and found that the Pentacam measurements were comparableto those of ultrasound pachymetry in normal eyes, with aslight underestimation (9.8 mm). Whereas Chaidaroon11

found that the optical measurement of CCT was, on average,27 mm greater than those of ultrasound pachymetry. Modiset al.5 evaluated scanning-slit topography and ultrasoundpachymetry, and found that the mean CCT measured byscanning-slit topography was significantly higher than thatof the ultrasound pachymetry. However, this was becausethey did not use the acoustic equivalent correction factor(0.92). The corneal thickness readings by ultrasoundpachymetry and scanning-slit topography corrected with theacoustic equivalent correction factor were found to be similarin normal eyes.7,12

As an optical pachymeter, non-contact specular micro-scope had CCT readings that were reported to be lower thanthose obtained by ultrasound pachymetry.4,5,7 Bovelle et al.4

reported that in normal eyes, the CCT measurements madeby non-contact specular microscopy were, on average,31.6 mm lower than those by ultrasound pachymetry. Modiset al.5 reported that the mean difference was -33 mm. Suzukiet al.7 compared the CCT values obtained by these twodevices, and found that on average, the result obtained withthe Topcon SP-2000P was 22.8 mm lower than that obtainedwith the ultrasound pachymetry, but a good linear correla-tion was found between them (Pearson’s correlation coeffi-cient, r = 0.897, P < 0.001). In our study, the mean CCTvalues measured by a non-contact specular microscope wassignificantly smaller than that obtained by ultrasoundpachymetry in pre-surgery subjects, and the mean differencewas -14.41 mm. Bland–Altman plots suggested that in 95%of cases, the differences in measurements between these twodevices range from -28.44 to -0.38 mm.

A definitive reason for the lower corneal thicknessreported by the non-contact specular microscope has notbeen found. It can be partially explained by the differentstrategies used by each machine to make corneal thicknessmeasurements. Non-contact specular microscopy measure-ments depend on the reflection of light, whereas ultrasoundpachymetry measurements are based on the reflection ofultrasound from the anterior and posterior corneal surfaces.In ultrasound pachymetry, the exact posterior point of reflec-tion is not known. It is supposed to be located betweenDescemet’s membrane and anterior chamber. It is also provedthat the ultrasonic probe can displace 7–40 mm thick tearfilm, and even thinner than the epithelium.13

In this study, our second goal was to evaluate and comparethe pachymetry readings obtained with non-contact specularmicroscopy and ultrasound pachymetry in post-LASIK eyes.Similar to pre-surgery findings, the mean CCT valuesobtained by non-contact specular microscopy was thinnerthan that obtained by ultrasound pachymetry. The meandifference between them was -3.69 mm, which was less thanthe difference in pre-surgery data. Bland–Altman plots sug-gested that the 95% limits of agreement range from -19.72to 12.34 mm.

Several possibilities exist for explanation. The specularmicroscope functions as an optical pachymeter. It requiresclear reflections of the epithelial and endothelial surfaces toobtain a reliable thickness measurement. When the cornealmedium is not clear or has optical interfaces, the pathway oflight rays may be interrupted.14 Factors which may influenceoptical pachymetry system such as specular microscopy afterLASIK include stromal haze, stromal interface, changes inrefractive index and postoperative modifications of theanterior corneal contour.15 Another consideration is thatultrasound pachymetry offers less accurate corneal thicknessmeasurements in post-LASIK eyes. The increase in the pro-teoglycans and hyaluronic acid during corneal repairmentafter excimer laser photoablation provokes accumulation ofwater in the stroma, hence altering stromal hydration andmaking the corneal stroma more flexible.16 These cornealchanges could affect contact pachymetry techniques, such asultrasound pachymetry. Furthermore, in some pathologicalconditions, such as corneal oedema or after LASIK surgery,the speed of sound as well as the point of reflection may bechanged, which in turn affect the accuracy of the measure-ments of ultrasound pachymetry.17,18

In the present study, the tissue ablation amount measuredby ultrasound pachymetry correlated a little better than thosemeasured by non-contact specular microscopy with the abla-tion depth estimated by the laser system. It seems that ultra-sound pachymetry is more reliable than non-contact specularmicroscopy. In general, ultrasound pachymetry is consideredas the gold standard of corneal pachymetry. Wang et al.19

reported that optical pachymetry might underestimatecorneal thickness during corneal swelling. Thomas et al.12

found that postoperatively, the timing may have an impact onthe CCT readings by optical pachymetry. However, in thisstudy, ultrasound pachymetry and non-contact specular



microscopy measurements were only taken at prior to and3 months after LASIK, further long-term postoperativefollow-up research is required to confirm it.

Pierro et al.20 correlated CCT measurements obtained byultrasound pachymetry and specular microscopy, and found apoor correlation between them (r = 0.540, P < 0.001).Whereas, in our research, an excellent correlation was foundbetween these two instruments in either case, which inagreement with the results of other studies.4,5,7,21 However,according to Bland and Altman,6 a correlation does not meanthat the two methods agree, because correlation coefficientmeasures the strength of a relation between two variables,not the agreement between them. So we performed furthercomparisons of the two instruments using the Bland–Altmananalysis. The results showed that either before or postLASIK, the measurements of the two devices were not com-plete agreement. The 95% limits of agreement in both situ-ations have a span of approximately 30 mm. The differencesbetween them were statistically significant, but it could beargued that they are not clinically significant, especially inpost-LASIK cases, in which the differences between the twoinstruments readings will be less than �20 mm. It seems thatpostoperatively, for some practical purposes, measurementswith these two devices can be used interchangeably. Con-sidering its convenient, non-contact way of measurement,non-contact specular microscopy could be a promisingmodality for follow up after refractive surgery or other thera-peutic interventions.

In treating patients with the refractive surgery, cornealthickness measurement has become particularly important inclinical practice. It is generally accepted that during LASIKsurgery, at least 250 mm residual stromal bed should be pre-served after flap creation and stromal ablation.22 As non-contact specular microscopy is not in complete agreementwith ultrasound pachymetry in measuring CCT both beforeand after LASIK, one should compare values generatedby these two machines cautiously, because CCT values cansignificantly affect the decision of whether to proceed tosurgery or any other further enhancement repairs. Underes-timation of corneal thickness may lead to exclusion of someproper patients, and to use an excessively conservative treat-ment plan. By contrast, overestimation may increase the riskof keratectasia.

ACKNOWLEDGEMENTS

The authors thank Ying Huang MD (Shanghai Jiaotong Uni-versity Affiliated Sixth People’s Hospital, Shanghai, China)for her valuable suggestions, and Liang Wang MD (AustinDiagnostic Clinic, Austin, Texas, USA) for modifying thismanuscript.

REFERENCES

1. Mishima S. Corneal thickness. Surv Ophthalmol 1968; 13: 57–96.2. Argus WA. Ocular hypertension and central corneal thickness.

Ophthalmology 1995; 102: 1810–12.

3. Murphy GE, Murphy CG. Comparison of efficacy of longest,average, and shortest axial length measurements with a solid-tip ultrasound probe in predicting intraocular lens power.J Cataract Refract Surg 1993; 19: 644–5.

4. Bovelle R, Kaufman SC, Thompson HW, Hamano H. Cornealthickness measurements with the Topcon SP-2000P specularmicroscope and an ultrasound pachymeter. Arch Ophthalmol1999; 117: 868–70.

5. Modis L, Langenbucher A, Seitz B. Corneal thickness measure-ments with contact and noncontact specular microscopic andultrasonic pachymetry. Am J Ophthalmol 2001; 132: 517–21.

6. Bland JM, Altman DG. Statistical methods for assessing agree-ment between two methods of clinical measurement. Lancet1986; 1: 307–10.

7. Suzuki S, Oshika T, Oki K et al. Corneal thickness measure-ments: scanning-slit topography and noncontact specularmicroscopy versus ultrasonic pachymetry. J Cataract Refract Surg2003; 29: 1313–18.

8. Tam ES, Rootman DS. Comparison of central corneal thicknessmeasurements by specular microscopy, ultrasound pachymetry,and ultrasound biomicroscopy. J Cataract Refract Surg 2003; 29:1179–84.

9. Bechmann M, Thiel MJ, Neubauer AS et al. Central cornealthickness measurement with a retinal optical coherence tomog-raphy device versus standard ultrasound pachymetry. Cornea2001; 20: 50–4.

10. Lackner B, Schmidinger G, Pieh S, Funovics MA, Skorpik C.Repeatability and reproducibility of central corneal thicknessmeasurement with Pentacam, Orbscan, and ultrasound. OptomVis Sci 2005; 82: 892–9.

11. Chaidaroon W. The comparison of corneal thickness measure-ment: ultrasound versus optical methods. J Med Assoc Thai 2003;86: 462–6.

12. Thomas J, Wang J, Rollins AM, Sturm J. Comparison of cornealthickness measured with optical coherence tomography, ultra-sonic pachymetry, and scanning slit method. J Refract Surg 2006;22: 671–8.

13. Nissen J, Hjortdal JO, Ehlers N, Frost-Larsen K, Sorensen T. Aclinical comparison of optical and ultrasonic pachometry. ActaOphthalmol (Copenh) 1991; 69: 659–63.

14. Iskander NG, Anderson Penno E, Peters NT, Gimbel HV,Ferensowicz M. Accuracy of Orbscan pachymetry measure-ments and DHG ultrasound pachymetry in primary laser in situkeratomileusis and LASIK enhancement procedures. J CataractRefract Surg 2001; 27: 681–5.

15. Prisant O, Calderon N, Chastang P, Gatinel D, Hoang-Xuan T.Reliability of pachymetric measurements using Orbscan afterexcimer refractive surgery. Ophthalmology 2003; 110: 511–15.

16. Patel S, Aslanides IM. Main causes of reduced intraocular pres-sure after excimer laser photorefractive keratectomy (letter).J Refract Surg 1996; 12: 673.

17. Lackner B, Pieh S, Schmidinger G et al. Glare and halo phe-nomena after laser in situ keratomileusis. J Cataract Refract Surg2003; 29: 444–50.

18. Salz JJ, Azen SP, Bernstein J, Caroline P, Villasenor RA,Schanzlin DJ. Evaluation and comparison of sources of vari-ability in the measurement of corneal thickness with ultrasonicand optical pachymeters. Ophthalmic Surg 1983; 14: 750–4.

19. Wang J, Fonn D, Simpson TL, Jones L. Relation betweenoptical coherence tomography and optical pachymetry

822 Zhao et al.


measurements of corneal swelling induced by hypoxia. Am JOphthalmol 2002; 134: 93–8.

20. Pierro L, Conforto E, Resti AG, Lattanzio R. High-frequencyultrasound biomicroscopy versus ultrasound optical pachym-etry for the measurement of corneal thickness. Ophthalmologica1998; 212 (Suppl. 1): 1–3.

21. Kawana K, Tokunaga T, Miyata K, Okamoto F, Kiuchi T,Oshika T. Comparison of corneal thickness measurements

using Orbscan II, non-contact specular microscopy, and ultra-sonic pachymetry in eyes after laser in situ keratomileusis. Br JOphthalmol 2004; 88: 466–8.

22. Durairaj VD, Balentine J, Kouyoumdjian G et al. The predict-ability of corneal flap thickness and tissue laser ablation in laserin situ keratomileusis. Ophthalmology 2000; 107: 2140–3.



comparison of central corneal thickness as measured by non-contact specular microscopy and...

Documents