Influence of combined phacovitrectomy without tamponade on intraocular lens displacement and postoperative refraction

Introduction

Intra- and postoperative complications, duration of surgery and postoperative recovery times have decreased significantly with the development of modern vitrectomy (Eckardt 2005). However, occurrence of cataract after pars plana vitrectomy is a well-known complication (Faulborn et al. 1978). The combination of phacoemulsification and vitrectomy has become a safe and frequently performed procedure, sparing the patient a second surgical intervention while reducing the burden on the healthcare system (Seider et al. 2014).

Intraocular lens (IOL) power calculation depends on total corneal power, axial eye length as well as the postoperative IOL position and is essential for the prediction of the postoperative refraction and therefore patient satisfaction (Norrby 2008; Norrby et al. 2017). Studies have investigated whether phacovitrectomy influences the IOL position and therefore the postoperative refractive outcome compared to phacoemulsification alone, reporting inconsistent findings (Falkner-Radler et al. 2008; Manvikar et al. 2009; Frings et al. 2015; Vander Mijnsbrugge et al. 2018; Sato et al. 2020; Shiraki et al. 2020; Tranos et al. 2020). These inconsistencies may be caused by multiple factors including differences in study design (retrospective versus prospective with/without a control group), inclusion criteria (vitreoretinal conditions requiring air/gas versus fluid tamponades), operation techniques (core versus complete vitrectomy, 20-gauge versus 23-gauge vitrectomy and smaller) and IOL type. Increased macular thickness in vitreoretinal disease could be a further source of error in biometry because axial length measurements may be influenced (Frings et al. 2015). Myopic shifts have repeatedly been reported after phacovitrectomy without air/gas tamponade, although there are also reports that found no significant postoperative refractive error (Falkner-Radler et al. 2008; Manvikar et al. 2009; Frings et al. 2015; Sato et al. 2020; Shiraki et al. 2020; Tranos et al. 2020).

The purpose of this study was to prospectively analyse differences in postoperative ACD, refraction and macular thickness between eyes requiring phacovitrectomy without air/gas tamponade and their contralateral partner eyes requiring phacoemulsification alone.

Materials and Methods

This prospective study was conducted between 2017 and 2019 on 40 eyes of 20 patients in need of phacovitrectomy for epiretinal membranes without gas tamponade in one eye and phacoemulsification alone in the other eye. Eyes requiring air/gas tamponade, posterior polar cataracts, pseudoexfoliation syndrome, previous trauma and uveitis were excluded from the study. Written informed consent explaining all risks and benefits of the surgery were obtained from all participants. The study adhered to the tenets of the Declaration of Helsinki and was approved by the Ethics committee of the Medical University Graz, Graz, Austria (IRB00002556).

Main outcome measures were postoperative ACD (defined as the distance between the anterior surface of the cornea and the lens or IOL), mean absolute postoperative refractive error and postoperative macular thickness change. Autorefraction (Nidek ACCUREF-K 9001 by Shin-Nippon), subjective refraction and best corrected visual acuity were evaluated preoperatively and 8 weeks postoperatively. ACD was measured immediately after surgery, on the first postoperative day and 8 weeks postoperatively using a swept-source OCT device (IOL Master 700; Carl Zeiss Meditec AG, Jena, Germany). Macular OCT scans were performed with the Spectralis OCT (Heidelberg Engineering, Heidelberg, Germany).

All 20 patients received the same stepped, non-angulated, open-loop, single-piece hydrophobic acrylic IOL (Tecnis ZCB00; Johnson & Johnson Vision, Jacksonville, FL, USA). Surgeries were performed by three experienced surgeons (AH, CMX and OF). Both eyes were operated by the same surgeon in each patient. Topical anaesthesia was administered preoperatively for cataract eyes, whereas a sub-tenon block or a retrobulbar block was used for combined phacovitrectomy.

Cataract surgeries were performed using Oertli's Faros surgical platform (Oertli Instrumente AG, Berneck, Switzerland) or Stellaris (Bausch & Lomb, Rochester, NY, USA) phacovitrectomy machines. A 2.5 mm self-sealing corneal incision, injection of an ophthalmic viscosurgical device (OVD), continuous capsulorhexis, phacoemulsification, irrigation/aspiration of cortical material and injection of OVD into the capsular bag were performed in a standardized fashion. The IOL was implanted via injector into the capsular bag. Following the implantation of the IOL, the OVD was aspirated thoroughly from the anterior chamber, as well as retrolentally, to assure complete removal of the OVD.

Vitrectomy was performed using DORC's EVA 27-gauge vitrectomy system (DORC, Zuidland, the Netherlands) or Stellaris 25-gauge vitrectomy system (Bausch & Lomb). Sclerotomies were located superonasally, superotemporally and inferotemporally with a distance to the limbus of 3.5 mm. The infusion line was placed inferotemporally. Posterior vitreous detachment was induced before core vitrectomy, if necessary. Epiretinal membranes were stained with a dye and then removed together with the internal limiting membrane (ILM) or alone using a 27-gauge microforceps. The trochars were removed and no sutures were used. Postoperative therapy consisted of topical antibiotics and steroids.

Results

In total, 40 eyes of 20 patients (10 women; 50%) were included with a mean age of 78.5 years (SD 5.52; range 69–90 years). Descriptive analysis of preoperative optical biometry and the used IOL power are summarized in Table 1. There were no significant differences between the groups.

Descriptive analysis of pre- and postoperative ACD is summarized in Tables 1 and 2 respectively. Six eyes of three patients could not be measured at the 1 hr follow-up, but all patients were measured successfully 1 day and 8 weeks postoperatively. Mean and median intergroup differences at the 1 hr follow-up were 0.04 mm (SD 0.55, range −1.72 to 0.86) and 0.08 respectively. Mean and median intergroup differences at the 1-day follow-up were 0.01 mm (SD 0.25, range −0.71 to 0.35) and 0.02 respectively. Mean and median intergroup differences at the 8 weeks follow-up were 0.02 mm (SD 0.22, range −0.36 to 0.65) and 0.05 respectively. No difference between both groups could be demonstrated at any time-point (1 hr: p = 0.74, 1 day: p = 0.85; 8 weeks: p = 0.40). A post hoc power analysis for the 8 week inter-group difference resulted in a needed sample size of 997 eyes per group to find the measured difference as significant (two-tailed Wilcoxon signed rank test for a power (1-beta) of 0.8 and a significance level of <0.05).

Anterior chamber depth shifts from day 1 postoperatively to 8 weeks postoperatively in the cataract and the vitrectomy groups were −0.10 mm (SD 0.21, median −0.10, range −0.70 to 0.26) and −0.12 (SD 0.10, median −0.14, range −0.39 to 0.04) respectively (p = 0.36).

For IOL power calculation, constants from the Ocusoft/ULIB platform were used. To allow a fair comparison, a retrospective optimization of the constants was performed separately for both groups. For none of the formulae a relevant or significant difference was found between the groups with these optimized constants. The optimized constants for the cataract group were sf = 2.565 for Holladay I, pACD = 6.365 for HofferQ, A = 120.1 for SRK/T and a0 = 2.235 (a1 = 0.4 and a2 = 0.1) for Haigis. The optimized constants for the vitrectomy group were sf = 2.540 for Holladay I, pACD = 6.350 for HofferQ, A = 120.1 for SRK/T and a0 = 2.215 (a1 = 0.4 and a2 = 0.1) for Haigis.

Mean absolute error, median absolute error and percentages within 0.5 D and 1.0 D range are summarized in Table 3. There were no significant differences between the cataract and the phacovitrectomy groups (Holladay I p = 0.45; HofferQ p = 0.48; SRK/T p = 0.50; Haigis p = 0.87; Barrett p = 0.35).

Macular thickness change was 11.1 μm (SD 73.8, median 10.0, range 227–187) and −4.9 μm (SD 139.2, median −5, range 258–282) in the cataract and phacovitrectomy groups respectively (Table 4). This difference was not statistically significant (p = 0.48). In both groups, there was no significant correlation of macular thickness change and arithmetic refractive error (r² = −0.13, p = 0.58 and r² = −0.10, p = 0.68, respectively). Six eyes out of four patients showed postoperative intraretinal fluid at 8 weeks, out of which four eyes had a phacovitrectomy.

Discussion

We prospectively compared the postoperative ACD and refractive error in 20 eyes requiring phacovitrectomy without air/gas tamponade to their contralateral partner eyes requiring phacoemulsification alone. Both variables showed no significant differences between groups and no effect of macular thickness change on postoperative refractive error. The ACD serves as a surrogate for the effective lens position (ELP), a parameter describing the postoperative IOL position as if the lens had no thickness (Hamoudi & La Cour 2013). In our investigation using SS-OCT, the ACD showed no significant differences pre- or postoperatively between phacovitrectomized eyes and their partner eyes operated for cataract alone. Our findings agree with two recent investigations using the CASIA2 to measure ACD (Sato et al. 2020; Shiraki et al. 2020). Another study using the IOLMaster 700, however, found an increase in ACD after phacovitrectomy compared to cataract surgery, even though the distance between the anterior surface of the cornea and the IOL had to be remeasured manually in some cases due to inaccurate readings (Vander Mijnsbrugge et al. 2018). Another factor influencing their findings could be the implantation of a soft hydrophilic four haptic IOL. In our study, as well in the study from Sato et al. (2020), hydrophobic acryli c-loop haptic IOLs were implanted.

In theory, vitrectomy could indeed alter the vitreal counterpressure to the lens–iris diaphragm, and therefore the ELP. However, our data indicate that the ACD is not influenced by vitrectomy. The optimized constants for all formulae were slightly higher compared to the ULIB values. Further adjustment of the mentioned constants is not recommended and was done for this study merely to allow better comparison between both groups.

Similarly, to ACD, we found no differences in the postoperative refractive error between phacovitrectomized eyes and their partner eyes operated for cataract alone. This finding agrees with three investigations, but inconsistent with three other previous studies that found a myopic shift after combined phacovitrectomy without air/gas tamponade (Falkner-Radler et al. 2008; Manvikar et al. 2009; Frings et al. 2015; Sato et al. 2020; Shiraki et al. 2020; Tranos et al. 2020). Tranos et al. recently compared combined to sequential (vitrectomy followed by cataract surgery) phacovitrectomy and reported a greater myopic shift in the combined group. However, patients receiving fluid (epiretinal membranes) and gas tamponades (idiopathic macular holes) were not analysed separately (Tranos et al. 2020). Indeed, Shiraki et al. (2020) recently reported a myopic shift only in patients receiving gas tamponades compared to vitrectomized patients without gas tamponade and cataract controls.

The correlation between macular thickness change and refractive error was low in our investigation. Increased macular thickness has been reported to be a potential source of error in biometry because axial length measurements may be influenced (Frings et al. 2015). Partial coherence interferometry (PCI) has traditionally been employed for optical biometry but is increasingly replaced by swept-source optical coherence tomography (SS-OCT). Both techniques measure the axial eye length as the distance between the anterior surface of the cornea and the retinal pigment epithelium (RPE) (Fercher et al. 1988; Hitzenberger et al. 2016). Even though SS-OCT-based biometry is more effective in obtaining biometric measurements in eyes with posterior subcapsular and dense nuclear cataracts, axial length measurements show remarkable agreement between both techniques (Akman et al. 2016). The RPE is located adjacent to the focal plane of the eye (i.e. the photoreceptors) and the distance between the photoreceptors and the RPE is considered negligible and stable (Hamoudi & La Cour 2013). However, Frings et al. (2015) report that macular conditions involving anteroposterior traction may displace the RPE, thus inducing false short axial length measurements. In theory, these false short measurements may induce a myopic shift because the focal plane of the eye should move to its normal position after successful surgery (with the IOL calculated for its preoperative location). However, more studies are required because it is unclear how many vitreoretinal patients show preoperative RPE displacement and if the focal plane of the retina relocates entirely in all patients after surgery.

In patients receiving a retrobulbar or sub-tenon block fixation at the 1 hr, ACD measurement may be a limiting factor. A longer follow-up period than 8 weeks could have been beneficial but was found to be sufficient in previous studies. Furthermore, we did not investigate IOL tilt and decentration. Strengths of this study are its prospective design with bilateral comparison and the comparison of optimization of IOL constants.

In conclusion, there were no differences in ACD and postoperative refractive outcome between phacovitrectomized eyes and their contralateral partner eyes operated for cataract alone. Surgical-induced macular thickness change did not influence refractive outcome. Our data indicate that individual optimization of the already centrally (Ocusoft/ULIB platform) optimized IOL constants for phacovitrectomy is not beneficial. All investigated IOL biometry formulae showed comparable postoperative results, independent if cataract surgery alone or phacovitrectomy was performed.

Author Contributions

The corresponding author attests that all listed authors meet authorship criteria and that no others meeting the criteria have been omitted. CM-X, MG and NH: research design, data acquisition and analysis, interpretation of data, drafting the manuscript and critical revision of the manuscript. MG and FW: data acquisition and analysis and critical revision of the manuscript. AH and OF: research design, interpretation of data and critical revision of the manuscript.

Ethical Approval

Approval by the Ethics Committee of the Medical University Graz was obtained (EC-Nr: 29-100 ex 16/17).