Secondary Logo

Journal Logo

Article

Accuracy of user-adjusted axial length measurements with optical biometry in eyes having combined phacovitrectomy for macular-off rhegmatogenous retinal detachment

Rahman, Rubina FRCOphth, FRCS*; Kolb, Simon FRCOphth; Bong, Chun Xiao MBChB; Stephenson, John PhD

Author Information
Journal of Cataract & Refractive Surgery: July 2016 - Volume 42 - Issue 7 - p 1009-1014
doi: 10.1016/j.jcrs.2016.04.030
  • Free

Abstract

Combined phacoemulsification and pars plana vitrectomy (phacovitrectomy) has become a common procedure for many vitreoretinal diseases, including as primary repair surgery for rhegmatogenous retinal detachment (RRD).1–4 This is a result of the continual advances in cataract surgery and vitrectomy surgery and favorable patient outcomes. The combined procedure reduces costs and offers quicker visual rehabilitation by avoiding the need for additional surgery and allowing a single recovery period.5–7 The aphakic state during surgery also gives excellent visibility, ensuring an unimpeded view for the treatment of the peripheral pathology common in RRD.8

As the anatomic success of combined surgery has improved, greater attention has been directed toward reducing refractive error to maximize postoperative visual function. Our previous large case series study1 found the refractive results of phacovitrectomy for RRD to be comparable with cataract surgery outcomes. Overall, although optical biometry was more accurate than ultrasound (US) (P = .040), we found that significantly more US-measured axial lengths (ALs) were preferentially selected over optical biometry measurements in the macula-off group than in the macula-on group (P = .016). We concluded that the biometry used for intraocular lens (IOL) power selection must be checked by comparing it with that in the fellow eye and the known refraction, especially in macula-off RRD cases.

The accuracy of AL measurement is crucial for IOL power calculation, and the presence of a bullous detached macula is likely to make AL measurements more challenging. It was our clinical observation that when used for AL measurements in macula-off RRD cases, optical biometry tended to underestimate the true AL (confirmed by the anterior position of the signal peak, even in the presence of a good signal-to-noise ratio [SNR]) (Figure 1). Hence, we performed a prospective analysis of the preoperative AL and postoperative AL measurements of all macula-off RRD cases having phacovitrectomy as the primary repair to evaluate a new technique of optimizing the accuracy of AL measurement with optical biometry termed the user-adjusted optical biometry measurement.

Figure 1
Figure 1:
Composite optical biometry showing a posterior retinal peak (AL = axial length; SNR = signal-to-noise ratio).

Patients and methods

The study evaluated consecutive macula-off RRD patients who had combined phacovitrectomy surgery performed by the same surgeon (R.R.) from November 2012 to August 2014.

The ALs were measured using optical biometry with partial coherence interferometry (PCI) (IOLMaster, version 5.4, Carl Zeiss Meditec AG) and US A-scan (Echoscan US-1800, Nidek Co. Ltd.) before phacovitrectomy at presentation of macula-off RRD. Skilled operators performed 10 reliable readings using optical biometry and US. When a posterior peak was not automatically selected in the primary optical biometry, AL measurements were manually adjusted by the biometry operator shifting the signal peak selection from the default anterior peak to a more posterior peak with an SNR of 2 dB or more (Figure 1). In this study, this method is termed user-adjusted optical biometry, and it has been previously described by Steel.9 When the optical biometry produced a scan with multiple peaks and no defined single posterior peak, the posterior peak correlating to the AL was guided by the fellow eye’s AL or the ipsilateral US AL.

All study eyes had the AL remeasured using optical biometry after at least 8 weeks postoperatively. The user-adjusted optical biometry was then compared with the preoperative US and with the postoperative optical biometry measurements and analyzed for statistical difference using paired-samples t tests. Assumptions for parametric testing were tested before the tests were performed. Secondary within-group comparisons and between-group comparisons of subsets of the full sample were done using nonparametric tests (Wilcoxon signed-rank test and Mann-Whitney U test) because of very small sample sizes.

Results

This retrospective case series analysis comprised 22 eyes of 22 patients. The patients had a mean age of 62.6 years, and 17 (77.3%) of the 22 were men. The mean postoperative AL measurement used for biometry calculation in our study was 25.6 mm ± 2.23 (SD) (range 20.97 to 29.64 mm).

Eighteen (81.8%) of the 22 patients had good-quality, interpretable optical biometry, comprising 13 user-adjusted and 5 primary unadjusted optical biometry measurements. The 5 patients with well-defined single peaks on optical biometry did not require user adjustment.

Thus, 13 eyes (59%) with user-adjusted optical biometry were included in the comparison of user-adjusted optical biometry and postoperative optical biometry. Eight of the scans had a well-defined second posterior peak that did not require assessment of agreement with the fellow eye’s AL or ipsilateral US AL scans. Five scans required some consideration of the fellow eye or ipsilateral US AL to shift to a correlating posterior peak among other more posterior peaks.

Four patients (18.2%) had optical biometry that was unusable, requiring US AL to calculate the IOL power. Two of these patients had a poor SNR (<2 dB). Two patients who had an unidentifiable second peak among multiple peaks were excluded from the analysis.

Axial Length Comparisons

User-adjusted optical biometry AL measurements were compared with preoperative US AL measurements using a paired-samples t test performed in the 13 eyes for which both readings were available. The mean preoperative AL recorded by the user-adjusted optical biometry method was 25.473 ± 1.755 mm. The mean preoperative AL recorded by the US A-scan method in these 13 eyes was 25.469 ± 1.826 mm. The mean difference between the 2 methods was 0.004 mm (4.0 μm ± 0.304 mm). There was no statistically significant difference at a significance level of 0.05 between the 2 methods of AL measurement compared with the postoperative optical biometry AL (95% confidence interval [CI], −0.180 to 0.188; t12 = 0.046; P = .964). The inference would not be affected by a correction for multiple comparisons. All distributional assumptions for the parametric testing were met by the data.

A Bland-Altman plot (Figure 2) shows moderate consistency between the 2 sets of readings, with 1 datapoint outside the limits of agreement (LoA) and a further datapoint on the LoA. The optical biometry readings were higher than the corresponding A-scan readings in the majority of cases; however, the absolute values of the difference measurements (excluding the 2 outlying values) were low. No relationship between the levels of agreement and the magnitude of data values was apparent.

Figure 2
Figure 2:
Bland-Altman plot indicating differences between preoperative optical biometry and A-scan AL measurements as a function of averages (AL = axial length; OB = optical biometry).

The differences between the user-adjusted optical biometry AL measurements were also compared with the postoperative optical biometry AL measurements using a paired-samples t test performed in the 13 eyes for which both readings were available. The mean postoperative AL recorded by the optical biometry method for these 13 eyes was 25.424 ± 1.800 mm. The mean difference between the preoperative and postoperative values was 0.049 ± 0.144 mm. There was no statistically significant difference at a significance level of 0.05 between the preoperative and postoperative optical biometry AL measurements (95% CI, −0.038 to 0.136; t12 = 1.23; P = .242). The inference would not be affected by a correction for multiple comparisons.

A Bland-Altman plot (Figure 3) shows good consistency between the 2 sets of readings, with no datapoints outside the LoA. The postoperative values were longer than the corresponding preoperative values in approximately half of the measured cases. No relationship between the levels of agreement and the magnitude of data values was apparent.

Figure 3
Figure 3:
Bland-Altman plot indicating differences between preoperative optical biometry and postoperative optical biometry AL measurements as a function of averages (AL = axial length; OB = optical biometry).

A further comparison was performed for the 13 cases for which the adjusted optical biometry values were available plus 5 additional cases for which the adjusted optical biometry AL values were not available but well-defined single-peaked readings on the unadjusted (default) optical biometry AL (a combined group) were available. In these cases, the appropriate adjusted or unadjusted preoperative optical biometry AL measurements were compared with the corresponding postoperative optical biometry AL measurements, again using a paired-samples t test. The mean preoperative AL measurement in the combined group was 25.329 ± 2.081 mm. The corresponding mean postoperative AL measurement in the same group was 25.284 ± 2.141 mm. The mean difference was 0.044 ± 0.131 mm. There was no statistically significant difference at a significance level of 0.05 between the preoperative and postoperative optical biometry AL measurements (95% CI, −0.021 to 0.110; t17 = 1.44; P = 0.170). The inference would not be affected by a correction for multiple comparisons.

The median preoperative AL measurements and postoperative AL measurements of the 5 patients with well-defined single peaks were 24.73 mm and 24.70 mm, respectively. A Wilcoxon signed-rank test found no significant difference between preoperative AL measurements and postoperative AL measurements (Z = −0.948, P = .343).

Of the 13 patients for whom user-adjusted optical biometry values were available, the RRD morphology was documented as bullous (n = 5) or shallow (n = 8), indicating the relative amount of subretinal fluid at the macula. In this group, the median AL measurements were 26.340 mm and 25.625 mm, respectively. A Mann-Whitney U test found no significant difference between the AL measurements in the 2 groups and hence no evidence to suggest that the amount of subretinal fluid affected the AL measurements performed with optical biometry (Z = −0.732, P = .464).

Potential Refractive Outcomes Difference

The IOL power calculated with user-adjusted optical biometry was within ±0.5 diopter (D) in 12 (92%) of 13 eyes (95% CI, 77.8 to 100.0) of the theoretical IOL power calculated using postoperative optical biometry measurements aiming for emmetropia. However, with IOL powers calculated using the preoperative US measurements, 10 (77%) of 13 eyes (95% CI, 54.0 to 99.8) had IOL power within ±0.5 D of the postoperative optical biometry AL theoretical IOL power for emmetropia. All cases of single-peaked, primary optical biometry IOL choices were within ±0.5 D of the postoperative optical biometry IOL calculations.

If the default optical biometry AL measurements were used rather than the measurements of the 13 patients with user-adjusted optical biometry, the difference in IOL power would range from 18.0 to 37.5 D, with a mean difference to postoperative optical biometry IOL choice of 6.96 D and a sample SD of 4.78 D. Using the user-adjusted optical biometry measurements for IOL calculations, the IOL power ranged from 12.0 to 25.0 D, with a mean difference in postoperative optical biometry IOL choice of 0.19 ± 0.43 D. Such a large difference in power implies the potential for a substantive error in the refractive outcome if the more accurate user-adjusted AL had not been selected.

Discussion

To our knowledge, this is the first study proposing a method of optimizing the accuracy of optical biometry measurements using user-adjusted optical biometry to calculate the IOL in macula-off RRD patients having phacovitrectomy as primary repair surgery.

Accurate measurement of the AL is essential for preventing significant refractive errors postoperatively. This can often be difficult in eyes with RRDs. Previous studies have shown that optical biometry is more accurate than US in general,10–13 except in macula-off cases in which the accuracy of the primary optical biometry measurements has to be assessed with caution by careful comparison with the known refraction and AL in the fellow-eye.1 In the macula-off group in our previous study, significantly more US-measured ALs were preferentially selected for IOL power estimation rather than the primary optical biometry measurements. However, in the current study, the mean preoperative AL values derived from the user-adjusted optical biometry showed no statistically significant difference from AL values derived from the US method. The Bland-Altman plot indicates moderate agreement between the 2 methods, providing further evidence of the acceptability of the user-adjusted method as an alternative to US measurements. Also, in our cohort more patients had IOL calculations based on user-adjusted optical biometry within ±0.5 D of the postoperative optical biometry IOL calculations. This might be because of possible inaccuracies in the US measurements from US probe applanation error on the cornea.

Optical biometry by the IOLMaster is based on the principle of dual-beam PCI.14 Its high precision, resolution, accuracy, and reproducibility of AL measurements in normal eyes have been reported.15,16 However, optical biometry is not without limitations. It is not suitable in dense ocular media and nonoptimum fixation.17,18 Lege et al.18 also reported the disadvantage of using optical biometry in RRD cases. In clinical practice, we have noticed that optical biometry can underestimate the true AL measurements in macula-off detachments, especially in cases with bullous RRD despite a good SNR. This was shown in our previous study1 of the subgroup of macula-off eyes that had the AL measured with the IOLMaster and Echoscan US-1800 devices. The default AL measured with the IOLMaster device without user adjustment was 0.98 ± 1.55 mm shorter than the corresponding recordings made with the Echoscan US-1800 device.

Several confounding factors might limit the accuracy of AL measurement with optical biometry. The patients’ ability to fixate is essential for accurate optical biometry measurements because this method evaluates the AL along the visual axis, whereas US biometry measures along the optical axis. Patients with RRD, in particular those with macula-off RRD, might have nonoptimum fixation because of reduced vision or dense ocular media caused by vitreous hemorrhage. Usually, optical biometry measures from the front of the cornea to the retinal pigment epithelium; however, sometimes an even higher peak (apparently corresponding to the inner limiting membrane) is detected anteriorly. In eyes with macula-off RRD, a similar strong interference from interfaces in the detached retina will show a good SNR measurement, even though the result is incorrect.19 Light scattering of the incoming and outgoing rays can also cause inaccurate AL measurements. The advent of optical biometry has not rendered US biometry obsolete. In addition, US biometry measures from the cornea to the internal limiting membrane and can be inaccurate because of applanation errors. In the 5 cases of user-adjusted optical biometry, we referred to the US AL to identify an appropriate posterior peak as well as other clinical data, such as the refraction and fellow-eye AL. We considered the margin of error of US biometry when identifying the posterior peak. In no case were the peaks so close together that the potential error of US biometry would have caused the identification of an erroneous posterior peak. There are advantages and pitfalls with all devices; therefore, the surgeon must be vigilant when interpreting the readings and disregard erroneous measurements accordingly.

Intraoperative aberrometry is a new technique for accurate assessment of refractive status. It could be used in similar and future studies to assess the accuracy of biometry techniques.

Although our results show a higher proportion of user-adjusted optical biometry IOL calculations within ±0.5 D of the postoperative optical biometry IOL calculation, these results are based on a small cohort, which is likely to be underpowered to detect a statistically significant difference between the 2 groups. However, the higher number of accurate IOL calculations in the user-adjusted group provide strong justification for further research with a larger cohort. Real-world postoperative refractive outcomes in a larger patient cohort might also help with further analysis of this new method.

In conclusion, user-adjusted optical biometry could be used as an alternative method for the measurement of AL in macula-off RRD with combined phacovitrectomy as the primary repair. It can simplify the process of biometry because some cases might not need US AL measurement. However, it would require agreement with the US AL in cases in which a posterior peak is not easily identifiable. We have also shown that user-adjusted optical biometry might outperform US AL when calculating IOL power; however, a larger study would be needed to confirm this.

The AL measurements should be correlated with those in the fellow eye and the known refraction. When there is clinical doubt regarding the validity of the optical biometry AL measurement, the US AL measurement should be obtained for comparison or for use as an alternative.

We acknowledge that the choice of AL measurement in macula-off RRD remains challenging; there will be cases in which neither optical biometry nor US are deemed accurate and the surgeon might have to resort to using the fellow eye’s AL measurements or abandon the option of combined surgery. This study provides an additional method for optimizing the accuracy of AL measurement in such a group. By introducing this new method of interpreting optical biometry scans in eyes with macula-off RRD, we can further optimize the accuracy of biometry and expand the usability of optical biometry in this challenging group.

What Was Known

  • Optical biometry is an accurate method for measuring the AL for IOL calculations.
  • Inaccuracies in measuring the AL for a macula-off RRD might be caused by subretinal fluid elevating the fovea or poor fixation because of reduced acuity. This potential inaccuracy might contribute to the deferral of cataract extraction and IOL implantation.

What This Paper Adds

  • User adjustment of optical biometry in macula-off RRD was comparable with postoperative optical biometry measurements with an attached retina.
  • Combined phacovitrectomy for macula-off RRD was performed with an accurate adjustment of optical biometry. This could facilitate intraoperative conditions for detachment repair, improve the patient’s visual recovery without a second surgery for cataract, and maintain accurate refractive outcomes.

References

1. Rahman R, Bong CX, Stephenson J. Accuracy of intraocular lens power estimation in eyes having phacovitrectomy for rhegmatogenous retinal detachment. Retina. 2014;34:1415-1420.
2. Thompson JT, Glaser BM, Sjaarda RN, Murphy RP. Progression of nuclear sclerosis and long-term visual results of vitrectomy and transforming growth factor beta-2 for macular holes. Am J Ophthalmol. 1995;119:48-54.
3. Jeoung JW, Chung H, Yu HG. Factors influencing refractive outcomes after combined phacoemulsification and pars plana vitrectomy: results of a prospective study. J Cataract Refract Surg. 2007;33:108-114.
4. Iwase T, Sugiyama K. Investigation of the stability of one-piece acrylic intraocular lenses in cataract surgery and in combined vitrectomy surgery. Br J Ophthalmol. 90, 2006, p. 1519-1523, Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1857524/pdf/1519.pdf. Accessed May 8, 2016.
5. Demetriades A-M, Gottsch JD, Thomsen R, Azab A, Stark WJ, Campochiaro PA, de Juan E Jr, Haller JA. Combined phacoemulsification, intraocular lens implantation, and vitrectomy for eyes with coexisting cataract and vitreoretinal pathology. Am J Ophthalmol. 2003;135:291-296.
6. Smiddy WE, Mady M, Anagnoste S. Acrylic intraocular lens placement in conjunction with pars plana vitrectomy. Am J Ophthalmol. 2001;131:748-750.
7. Scharwey K, Pavlovic S, Jacobi KW. Combined clear corneal phacoemulsification, vitreoretinal surgery, and intraocular lens implantation. J Cataract Refract Surg. 1999;25:683-698.
8. Suzuki Y, Sakuraba T, Mizutani H, Matsuhashi H, Nakazawa M. Postoperative refractive error after simultaneous vitrectomy and cataract surgery. Ophthalmic Surg Lasers. 2000;31:271-275.
9. Steel D., 2013. Refractive outcome and possible errors following combined phaco-vitrectomy. In: Lois N, Wong D, editors., Complications of Vitreo-Retinal Surgery. Lippincott Williams & Wilkins, Philadelphia, PA, pp. 260-266.
10. Findl O, Drexler W, Menapace R, Heinzl H, Hitzenberger CK, Fercher AF. Improved prediction of intraocular lens power using partial coherence interferometry. J Cataract Refract Surg. 2001;27:861-867.
11. Olsen T. Improved accuracy of intraocular lens power calculation with the Zeiss IOLMaster. Acta Ophthalmol Scand. 85, 2007, p. 84-87, Available at: http://onlinelibrary.wiley.com/doi/10.1111/j.1600-0420.2006.00774.x/pdf. Accessed May 8, 2016.
12. Wang JK, Hu CY, Chang SW. Intraocular lens power calculation using the IOLMaster and various formulas in eyes with long axial length. J Cataract Refract Surg. 2008;34:262-267.
13. Eleftheriadis H. IOLMaster biometry: refractive results of 100 consecutive cases. Br J Ophthalmol. 87, 2003, p. 960-963, Available at: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC1771781/pdf/bjo08700960.pdf. Accessed May 8, 2016.
14. Lee AC, Qazi MA, Pepose JS. Biometry and intraocular lens power calculation. Curr Opin Ophthalmol. 2008;19:13-17.
15. Findl O, Kriechbaum K, Sacu S, Kiss B, Polak K, Nepp J, Schild G, Rainer G, Maca S, Petternel V, Lackner B, Drexler W. Influence of operator experience on the performance of ultrasound biometry compared to optical biometry before cataract surgery. J Cataract Refract Surg. 2003;29:1950-1955.
16. Rajan MS, Keilhorn I, Bell JA. Partial coherence laser interferometry vs conventional ultrasound biometry in intraocular lens power calculations. Eye. 16, 2002, p. 552-556, Available at: http://www.nature.com/eye/journal/v16/n5/pdf/6700157a.pdf. Accessed May 8, 2016.
17. Hill W, Angeles R, Otani T. Evaluation of a new IOLMaster algorithm to measure axial length. J Cataract Refract Surg. 2008;34:920-924.
18. Lege BAM, Haigis W. Laser interference biometry versus ultrasound biometry in certain clinical conditions. Graefes Arch Clin Exp Ophthalmol. 2004;242:8-12.
19. Falkner-Radler CI, Benesch T, Binder S. Accuracy of preoperative biometry in vitrectomy combined with cataract surgery for patients with epiretinal membranes and macular holes; results of a prospective controlled clinical trial. J Cataract Refract Surg. 2008;34:1754-1760.
© 2016 by Lippincott Williams & Wilkins, Inc.