Short-term postoperative changes in the choroidal vascularity index in patients with a unilateral epiretinal membrane

The idiopathic epiretinal membrane (iERM) is a fibrocellular growth at the vitreoretinal interface that can constrict and alter the structure of the retina, and its incidence increases with age [1, 2]. These changes can lead to vision loss, metamorphopsia, and aniseikonia [3, 4]. When indicated, ophthalmologists surgically remove the membrane and relieve the traction on the retina [5].

Previous studies have demonstrated that the choroid and retina, are affected in iERM, and that choroidal thickness decreased 3 months after surgery [6]. More recently, the choroidal vascularity index (CVI) was considered a new and more reliable metric for characterizing choroidal vascularity, which was computed by dividing the luminal area (LA) by the total subfoveal circumscribed choroidal area (TCA) based on enhanced depth imaging (EDI) optical coherence tomography (OCT) [7, 8]. Compared to CT, which is associated with multiple patient factors including age, axial length, intraocular pressure, and systolic blood pressure, CVI is a more robust and resistant choroidal parameter [8].

Short-term complications following surgery have been widely reported in patients with ERM [9, 10]. Datlinger et al. reported postoperative movement of the fovea one day after successful surgery for iERM [11]. Watanabe et al. reported that metamorphopsia and retinal displacement decreased significantly 3 months after epiretinal membrane surgery [12]. In addition, early postoperative variations in intraocular pressure, hypotony, choroidal detachment, and corneal decompensation have been reported. However, knowledge about the short-term changes in CVI after surgery in patients with ERM is limited. Furthermore, few studies have investigated changes in CVI in different regions in patients with ERM.

This study aimed to evaluate short-term postoperative changes in CT and CVI in eyes with ERM using EDI-OCT. In addition, we compared the differences in choroidal parameters between the affected and fellow eyes.

This study included 130 eyes of 65 patients with iERM. The mean age was 61.42 ± 11.30 years; 38 (58.5%) patients were male and 27 (41.5%) were female. Among eyes with ERM, there were 9, 31, and 25 in stages 2, 3, and 4, respectively. Eyes with ERM had worse visual acuity than the fellow eyes (logMAR, 0.688 ± 0.678 vs 0.155 ± 0.357, p < 0.001). Compared with the preoperative values, there was no significant difference in visual acuity at 1 week (logMAR, 0.653 ± 0.573, p = 0.928) after surgery in the ERM group. Visual acuity improved significantly at 1 month (logMAR, 0.416 ± 0.432, p < 0.01) and 3 months after surgery (0.346 ± 0.448, p < 0.01) (Table 1).

Compared to the fellow eyes, CVI was higher in the total macular (61.70 ± 5.17% vs 59.99 ± 5.26%, p < 0.01) and central regions (62.11 ± 6.84% vs 60.26 ± 6.52%, p < 0.01) in eyes with ERM (Table 2). CVI comparison with the fellow eye reduce interpersonal CVI difference, but an intrapersonal CVI difference still exists.

A significant difference was found in the total, N1, and central CVI and CT at T2 in eyes with ERM after surgery (p < 0.05, Table 3). Compared to preoperative values, the CVI in the total macula increased at one week after surgery (62.14 ± 5.02%). Furthermore, the total CVI decreased at one month (60.76 ± 4.97%) and three months (60.4 ± 4.83%) postoperatively. The CVI in N1 decreased at 1 and 3 months (60.31 ± 6.15% and 59.46 ± 7.22%, respectively) postoperatively compared to that preoperatively (62.51 ± 5.93%) and at 1 week (62.82 ± 7.74%) postoperatively. In the central macula, CVI decreased at 1 month (61.18 ± 6.37%) and 3 months (60.45 ± 8.20%) after surgery compared to that at one week after surgery (63.19 ± 7.09%). CT in T2 increased at 1 week (239.65 ± 72.98um) after surgery compared with that preoperatively (218.74 ± 62.61 μm), and decreased at 1 month (222.15 ± 71.91 μm) and 3 months (222.33 ± 65.72 μm) postoperatively compared to 1 week postoperatively (Table 3) (Fig. 2).

In this study, we observed a higher CVI 1 week after surgery and a lower CVI at 1 month and 3 months postoperatively in eyes with ERM. The change was pronounced in the nasal and central regions. Further, we observed that CT increased after surgery and subsequently declined in eyes with ERM. Better visual acuity was observed postoperatively. Moreover, the preoperative CVI was higher in eyes with ERM than in fellow eyes.

With the development of the SD OCT technology, choroidal structures can be visualized in vivo [16]. EDI-OCT improves the visibility of the choroid–sclera interface and enables quantitative CT evaluation. More recently, CVI, a new quantitative imaging biomarker, has been shown to be a viable tool for establishing an early diagnosis, monitoring disease progression, and evaluating post-operative recovery [7]. CVI has been found to change in many ocular diseases [17‐20] and altered after ophthalmic surgeries such as phacoemulsification [21, 22] and scleral buckling [23, 24]. In the current study, compared to CT, CVI had a higher sensitivity after ERM surgery. In addition, we calculated it using a self-developed semiautomatic software program with high repeatability [25].

Previous studies have evaluated CVI changes in patients with ERM after surgery. Rizzo et al. found that the CVI of eyes with ERM decreased at 1 and 3 months after vitrectomy [26]. Chun et al. reported a long-term decrease in CVI in the PPV-alone group and a short-term increase in the PPV combined with cataract surgery group [27]. However, these studies analyzed CVI changes over a longer postoperative period. Short-term complications after vitrectomy, such as hypotony and intraoperative retinal tears, have been reported [9, 28]. Therefore, short-term postoperative changes are also worthy of attention. Moreover, we measured CVI and CT in subdivisions according to the Early Treatment Diabetic Retinopathy Study rings using a semi-automatic software, which can provide choroidal data in specific regions.

A significantly higher CVI 1 week after surgery and a lower CVI at 1 month and 3 months postoperatively were the main OCT findings in this study. We hypothesized that this may be due to several mechanisms. First, early studies have reported that phacoemulsification can cause choroidal thickening due to surgical trauma-induced inflammation [21, 22]. After vitrectomy combined with phacoemulsification, the choroidal vascular area grows. This is possibly due to a disturbance in the blood–aqueous barrier, which enables inflammatory mediators from the aqueous to cross the vitreous and enter the choroid, causing structural alterations [7]. This could explain the increased CVI and CT at the early stage in this study. Second, vitreomacular tension may stretch the RPE and stimulate vascular endothelial growth factor release, resulting in an increased choroidal vascularity [29]. Vitrectomy eliminates vitreomacular traction, causing vascular endothelial growth factor levels to decrease. This explains why eyes with ERM had a greater CVI preoperatively and a lower CVI one month and three months postoperatively in our study. Third, the choroidal structure may change due to alterations in the retina, with the central and nasal choroids displaying higher effects than other regions. Hibi et al. reported a significant reduction in inner and middle retinal layers after ERM surgery [30]. Further, Park et al. reported progressive thickening of the nasal inner nuclear layer after ERM surgery [31]. These retinal changes might explain the pronounced central and N1 changes in the postoperative CVI. In addition, prominent changes in the nasal region of the macula may be related to the special anatomy of the retina, such as the papillomacular bundle. Different alignments of the papillomacular bundle between the nasal and temporal sides may lead to a more sensitive CVI on the nasal side. Although we cannot confirm that changes in retinal structure lead to changes in CVI, our results nonetheless suggest that the postoperative CVI of eyes with ERM, especially CVI in the central and nasal regions, may be a useful and objective method of assessing changes in choroidal vascular structure during the postoperative period. A sensitive technique for evaluating the choroidal vasculature may help clinicians to improve monitoring during follow-up.

This study had several limitations. First, the patient cohort was not sufficiently large. Second, this was a retrospective study, which carries a risk of sample bias. Third, we did not use swept-source OCT to obtain B-scan images with higher penetration. However, Agrawal et al. demonstrated that CVI measurements obtained using SS-OCT and SD-OCT concur with each other [16]. Fourth, because the location of anatomic check points (vessels and fovea) may change after surgery, to scan the same location of choroidal area during follow-up is difficult. This could be one of the reasons for the conflicting results in this literature regarding macular surgery and CVI. However, we used the follow up mode in OCT scanning, which could minimize this scanning deviation (Fig. 3). Moreover, we excluded subjects with obvious change in choroidal area during follow-up. Fifth, in this study, only half of the ERM patients had the angiographic examination such as fluorescein angiography or indocyanine angiography to exclude abnormal retinal or choroidal circulation. And none was reported to have obvious vasculature disruption in retina and choroid. Sixth, we used paired t-tests to compare CVI and CT between eyes with ERM and fellow eyes. CVI comparison with the fellow eye reduce interpersonal CVI difference, but an intrapersonal CVI difference still exists.

In conclusion, CVI of eyes with ERM increased in the early postoperative period and then decreased, especially in the central and nasal regions. Further prospective studies with larger sample sizes are required to confirm the effect of vitrectomy combined with cataract surgery on choroidal structure.

In this study, we investigate short-term choroidal structural and vascular changes after epiretinal membrane (ERM) surgery using enhanced depth imaging optical coherence tomography. We found a significantly higher CVI 1 week after surgery and a lower CVI at 1 month and 3 months postoperatively. The change was pronounced in the nasal and central regions. This may be related to postoperative inflammation and the release of macular traction. Therefore, CVI in the central and nasal regions, may be a useful and objective method of assessing choroidal blood-flow changes during the postoperative period. A sensitive technique for evaluating the choroidal vasculature may help clinicians to assess choroidal structural changes accompanying ERM and postoperative period.