Incidence and mechanism of conjunctival chemosis during phacoemulsification

Abstract

Background

This study aims to determine the incidence of conjunctival chemosis during phacoemulsification and to explore the possible mechanism and treatments.

Methods

Patients with or without chemosis during phacoemulsification by the same operator within two years were included. The initial site of chemosis, the chemosis originated time, and the degree of chemosis at the end of the surgery were recorded. The differences in phacoemulsification energy/time and irrigation volume were compared between two groups. Meanwhile, the distance between the clear corneal incision (CCI) and the end of the Bowman’s layer, surgically induced astigmatism (SIA), intraocular pressure and central corneal thickness were also compared between the two groups.

Results

The incidence of intraoperative chemosis was 9.17% (66/720). The distance between CCI and the end of the Bowman’s layer in the chemosis group was significantly longer than that in the non-chemosis group postoperatively (P < 0.0001). The initial sites of chemosis were located on both sides of the external opening of the CCI. The average time from the phacoemulsification probe introducing into the anterior chamber to the originating of chemosis was 30.23 ± 49.23s.

Conclusion

Intraoperative chemosis is related to the distance from CCI to the end of Bowmen’s layer. The residual conjunctiva around the incision wraps the phacoemulsification probe sleeve to form a passage, the leakage fluid enters the subconjunctiva through this passage, forming chemosis. Conjunctival incision on both sides of CCI can effectively prevent the development of chemosis.

Introduction

Cataract surgery using a clear corneal incision (CCI) is the preferred technique for most ophthalmologists [1]. However, there is a risk of intraoperative conjunctival chemosis if the conjunctiva is inadvertently incised during the creation of a clear corneal incision [2]. Even skilled operators, despite apparently well-placed CCIs, may encounter an incidence of intraoperative conjunctival chemosis as high as 9.8% [3].

Conjunctival chemosis, whether complete or incomplete, can result in the accumulation of irrigating solution on the corneal surface, causing significant disruption to the surgical field. In cases of prominent chemosis, it may even impede the entry of surgical instruments into the eye [4]. These problems are especially prominent in deep-set eyes [5]. While experienced surgeons may navigate these challenges with ease, for those less experienced, these factors can significantly amplify the surgical complexity and elevate the risk of intraoperative complications [5–7]. Furthermore, there is a theoretical concern that intraoperative conjunctival chemosis could increase the risk of intraocular infection [5, 8, 9].

Intraoperative conjunctival chemosis is typically attributed to the leakage of fluid through a conjunctival disinsertion, leading to the subconjunctival accumulation of irrigating fluid [4, 6]. Although several mechanisms have been proposed to explain its occurrence, there remains some controversy regarding its precise origins. The improper orientation of a “bevel-up” keratome, for instance, may inadvertently extend both sides of a correctly placed CCI posteriorly, potentially involving the conjunctiva [10].

Additionally, frequent passage of phacoemulsification and irrigation/aspiration (I/A) probes through the incision [7], along with the trapping conjunctiva inside the incision during the introduction of these probes [6], have also been implicated in this phenomenon. It’s worth noting that once conjunctival chemosis has become well-established, reducing it can be challenging and time-consuming. Therefore, early identification and prompt intervention are crucial for managing this issue effectively.

The purpose of this study was to explore the underlying mechanism of intraoperative conjunctival chemosis by evaluating all cases occurring during surgeries performed by the same operator within two years, which were then compared with instances where chemosis did not occur during the same period. Herein, we sought to identify early recognition strategies and effective intervention measures for managing conjunctival chemosis.

Patients and methods

Patients

This retrospective study included patients from the Eye & Ear, Nose, and Throat (EENT) Hospital of Fudan University between January 2020 and December 2021. The Ethics Committee of the EENT Hospital approved this study, and it adhered to the Declaration of Helsinki. All patients were informed in detail about the study, and written informed consent for publication was obtained from the participants. All patients were diagnosed with age-related cataract. Patients with other ocular diseases, a history of ocular trauma or previous intraocular surgery, and systemic diseases were excluded.

Patients with different degrees of conjunctival chemosis during surgery were included in the chemosis group, while age- and sex- matched non-chemosis patients who underwent cataract surgery at the same period were randomly included in the N-chemosis group.

Surgical technique

All surgeries were performed by a single surgeon (ZNZ) using the CENTURION vision system (Alcon laboratories, Inc., Fort Worth, TX, USA). After administering topical anesthesia with Oxybuprocaine Hydrochloride Eye Drops (Santen Pharmaceutical Co., Ltd., Japan), a single-plane clear corneal incision was manually created with a 2.6-mm “Bi-bevel” diamond keratome at the 11 o’clock position for all patients. This was followed by a 1-mm paracentesis incision at the 3 o’clock position. Both the main CCI and the side-port incision were made at the anterior margin of the limbal conjunctival vascular arcades [11]. Continuous capsulorhexis, hydrodissection, phacoemulsification, and irrigation/aspiration were sequentially performed. A foldable intraocular lens (ZCB00 or ZA9003, Johnson & Johnson Surgical Vision, Inc., Santa Ana, CA, USA) was implanted into the capsular bag. The incisions were hydrated with balanced salt solution. Intraoperative phacoemulsification parameters remained the same in both groups, and the same technique was used in both groups during and after surgery. The effective phacoemulsification time, the phacoemulsification energy (%), the total surgery time and the total irrigation volume used in each patient were documented. In the chemosis group, the time from introduction the phacoemulsification probe into the anterior chamber to the onset of conjunctival chemosis, the initial position of chemosis and the extent of chemosis at the end of the surgery were also recorded. The chemosis range at the end of surgery was divided into quadrants: grade 1, ≤ 1 quadrant; grade 2, ≤ 2 quadrants, > 1 quadrant; grade 3, ≤ 3 quadrants, > 2 quadrants; grade 4, ≤ 4 quadrants, > 3 quadrants.

Post-operatively, 1% Prednisolone Acetate Ophthalmic Suspension (Allergen Pharmaceutical Ireland, Ireland) and 0.5% Levofloxacin (Santen Pharmaceutical Co., Ltd., Japan) eye drops were administered four times daily for 2 weeks and 1% Pranoprofen (Senju Pharmaceutical Co. Ltd., Japan) eye drops four times daily for 1 month.

Ophthalmologic examination

All patients underwent complete ophthalmologic examination that included best corrected visual acuity (BCVA) in logMAR, slit-lamp biomicroscopy examination, cataract grading lens opacities classification III (LOCS III), and dilated fundus examination. The IOP was measured using Goldmann applanation tonometry; the anterior chamber deep (ACD) and axial length (AL) measurement was performed using an IOLmaster 700 (Carl Ziess, Inc., Jena, Germany). The central corneal thickness (CCT) and corneal keratometry (K) value was obtained using the corneal topography examination (Pentacam HR; OCULUS Optikgerate GmbH, Wetzlar, Germany), and the surgically induced astigmatism (SIA) was assessed using a method established by Alpins [12, 13]. Measurements of the AL, ACD, IOP, CCT, K value, and BCVA were obtained before surgery. The IOP, CCT, K value, and BCVA measurements were also obtained 1 day, 1 week, 3 months after surgery. The anterior segment optical coherence tomography (AS-OCT, RTVue-100; Optovue, Inc., Fremont, CA) was performed 1 day after surgery to obtained the distance from the end of the Bowman’s layer to the external wound opening of CCI.

AS-OCT Data Acquisition and Processing

A Fourier-domain OCT system with a corneal anterior module long adaptor lens (6-mm scan width and1.96-mm scan depth) was applied in current study. The cross-line scan mode was used to obtain images of CCI. During image acquisition, the patients were asked to fix their gaze at a peripheral target which opposite to the main incision to maintain perpendicularity with the OCT beam at the surface of the targeted tissue; this action was essential for accurate measurement. At least three scans of each patient were performed.

The measurements were performed by one trained operator and were based on a manual method [14] using the software built in to the apparatus (Fig. 1). The anterior of the limbus which was defined as the end of Bowman’s layer in AS-OCT [14]. The distance between the line that crossed the end of Bowman’s layer to the external wound opening of CCI was measured. The value was determined as the average of three independent measurements.

Fig. 1.

Measurement based on a manual method in an OCT cross-line scan image. First, one line crossing the ending point of Bowman’s layer at the peripheral cornea (line a) is drawn, which is perpendicular to the tangent of the external surface of the cornea. Then, a line running parallel to line a is drawn, crossing the external wound opening of CCI (line b). Finally, the distance between line a and line b was measured. CCI, clear corneal incision

Statistical analysis

Statistical analysis was performed using SPSS software (SPSS for Mac, Version 24.0; SPSS, Inc., Chicago, IL, USA). Comparisons between groups, sex and eye were conducted with Student’s t-test (or Mann-Whitney U test where appropriate). Differences in gender and eye between groups were analyzed using the chi-square test. The preoperative and postoperative measurements were analyzed using repeated measures analysis of variance tests. Spearman correlation was used to identify the factors correlated with the degree of conjunctival chemosis at the end of surgery. P values < 0.05 were considered statistically significant.

Results

A total of 720 cataract surgeries were performed in two years, 66 cases of intraoperative conjunctival chemosis were recorded and enrolled in the analysis, with an incidence rate of 9.17%. Figure 2 are representative images of intraoperative conjunctival chemosis. Sixty patients in the group without intraoperative conjunctival chemosis were included in the final analysis. The baseline characteristics between the two groups are shown in Table 1, and there was no significant difference in each parameter between the two groups before surgery (all P > 0.05). There were no intraoperative and postoperative complications in both groups.

Fig. 2.

Representative images of the intraoperative conjunctival chemosis. A, B and C are the initial stage of chemosis. D, E, and F show the chemosis status of A, B, and C at the end of surgery, respectively. The black arrow refers to the different initial position of chemosis

Table 1.

Patient demographics and baseline characteristics

Parameters	Chemosis Group (n = 66)	N-Chemosis Group (n = 60)	P Value
Age (y)	68.93 ± 11.71	70.14 ± 9.82	0.5331
Gender (F/M)	37/29	36/24	0.7193
Laterality (R/L)	28/38	27/33	0.7709
BCVA (LogMAR)	0.57 ± 0.26	0.63 ± 0.32	0.2485
IOP (mmHg)	14.28 ± 3.16	13.90 ± 2.76	0.4726
AL (mm)	23.69 ± 2.83	24.28 ± 3.81	0.3229
ACD (mm)	2.71 ± 0.37	2.80 ± 0.47	0.2325
Cataract hardness (LOCS III)
Cortical density	2.95 ± 1.09	3.24 ± 1.44	0.2024
Nucleus density	3.18 ± 0.77	3.15 ± 0.64	0.8134
Posterior subcapsular opacity	1.62 ± 0.74	1.55 ± 0.61	0.5656
Keratometry (D)
K1	43.37 ± 1.24	43.57 ± 1.27	0.3731
K2	44.14 ± 1.18	44.35 ± 1.27	0.3379
CCT (µm)	532.10 ± 31.22	536.22 ± 36.75	0.4977

Data are presented as the mean ± SD

F female, M male, R right eye, L left eye, BCVA best corrected visual acuity, IOP intraocular pressure, AL axial length, ACD anterior chamber depth, LOCS III lens opacities classification III, K1 flat keratometry, K2 steep keratometry, CCT central corneal thickness

Intraoperative parameters between the two groups are shown in Table 2. There were no significant differences in effective phacoemulsification time, average phacoemulsification energy, total operative time, and total irrigation volume between the two groups. AS-OCT results on the first day post-surgery showed that the distance from the end of Bowman’s layer to the external wound opening of the CCI in the chemosis group (772.6 ± 143.2 μm) was significantly longer than in the non-chemosis group (258.4 ± 54.36 μm) (P < 0.0001). After surgery, there was a significant improvement in BCVA and a decrease in IOP observed in both groups (all P < 0.05) was significantly improved and the IOP decreased in the two groups (all P < 0.05). However, no statistically significant differences were found between the two groups regarding BCVA, SIA, IOP and CCT at various postoperative follow-up time points. (Table 2, all P > 0.05).

Table 2.

Comparison of intraoperative and postoperative outcomes between two groups

Parameters / time points	Chemosis Group (n = 66)	N-Chemosis Group (n = 60)	P Value
Phacoemulsification energy (%)	24.53 ± 16.89	26.88 ± 15.63	0.4206
Effective phacoemulsification time (s)	45.47 ± 19.51	43.99 ± 20.14	0.6761
Surgery time (min)	9.12 ± 2.02	8.48 ± 1.80	0.0638
Irrigation volume (mL)	55.50 ± 12.60	52.20 ± 11.94	0.1348
Distance between CCI and the ending of Bowman’s layer (µm)	772.60 ± 143.20	258.40 ± 54.36	< 0.0001
BCVA (LogMAR)
1 D	0.24 ± 0.45	0.21 + 0.47	0.7151
1 W	0.11 ± 0.13	0.10 ± 0.13	0.667
3 M	0.06 ± 0.08	0.05 ± 0.07	> 0.9999
SIA (D)
1 D	0.80 ± 0.52	0.92 ± 0.51	0.1941
1 W	0.69 ± 0.46	0.75 ± 0.42	0.4475
3 M	0.53 ± 0.37	0.56 ± 0.29	0.6158
IOP (mmHg)
1 D	12.47 ± 2.80	12.08 ± 2.68	0.427
1 W	12.01 ± 3.13	11.35 ± 3.10	0.2373
3 M	11.92 ± 3.06	11.43 ± 2.88	0.3577
CCT (µm)
1 D	606.48 ± 56.82	604.34 ± 52.44	0.827
1 W	551.08 ± 45.41	548.26 ± 47.46	0.7339
3 M	534.50 ± 28.45	536.14 ± 37.71	0.7822

Data are presented as the mean ± SD

1D 1 day, 1 W 1 week, 3 M 1 month postoperatively, CCI clear corneal incision, BCVA best corrected visual acuity, SIA surgically induced astigmatism, IOP intraocular pressure, CCT central corneal thickness

Analysis of the distribution of chemosis in different ranges at the end of cataract surgery showed that there were 8 cases of grade 1 (12.12%), 43 cases of grade 2 (65.15%), 5 cases of grade 3 (7.58%), and 10 cases of grade 4 (15.15%). Further analysis of the chemosis group showed that the initial positions of intraoperative conjunctival chemosis were all located on both sides of the external opening of the CCI, with 26 cases (39.39%) starting from the lower left, 31 cases (46.97%) starting from the upper right, and 9 cases (13.64%) starting from both sides. There was no correlation between the initial position of chemosis and the final range of chemosis (r = 0.14, P = 0.27). The average time from the introduction of the phacoemulsification probe into the anterior chamber to the onset of conjunctival chemosis was 30.23 ± 49.23s, all of which occurred during the phacoemulsification stage, and the shortest time was immediately after the phacoemulsification probe entered, and these patients accounted for 42.42% of the total; This onset time was not correlated with the final range of conjunctival chemosis (r = -0.06, P = 0.63). Spearman correlation analysis also showed that there was no correlation between the final chemosis range and the parameters before, during and after surgery (all P > 0.05).

Discussion

During cataract surgery, chemosis of the conjunctiva may be observed in certain cases, particularly when a CCI is applied. In this study, the incidence of intraoperative chemosis was 9.17%. This finding is consistent with a study by Sugai et al., which employed an anterior margin CCI at the limbal vascular arch and reported a similar chemosis incidence of 9.8% [3]. In another study conducted by Ziakas NG et al., the incidence of chemosis during phacoemulsification via a superior scleral tunnel incision was 7.5% [15]. However, in a report by Dada et al. [7], the incidence of chemosis during phacoemulsification was significantly lower at 0.34%. Notably, Dada et al. did not provide specific details regarding the location of the CCI or the incision size in their report. This study was conducted by a less-experienced surgeon, having performed fewer than 1000 surgical cases. In addition, variations in patient demographics, incision location, size, and keratotomy bevel may also contribute to the differences in the incidence of intraoperative chemosis between studies.

There are various speculations about the mechanism of intraoperative conjunctival chemosis [6, 7, 10]. It is generally believed that breaching the conjunctiva during the incision is a necessary condition for chemosis [4–6]. The disinsertion of the conjunctiva from the limbus, which is necessary for creating the scleral tunnel and may result in minor incisional leakage, can lead to this condition [15]. Consequently, there is an accumulation of irrigating solution beneath the conjunctiva followed by subsequent ballooning. However, there is no consensus on how a correctly placed CCI forms a conjunctival disinsertion, and how the leaking fluid through and accumulated under the subconjunctiva. Poole et al. found that chemosis occurred even in corrected placed CCI, and suggested that a “bevel-up” keratome would cause both ends of the incision to extend posteriorly to the conjunctiva [10]. Dada et al. speculated that the frequent passage of phacoemulsification probe and I/A probe through the incision may be related [7]. Villada et al. believe that the introduction of the phacoemulsification or I/A probe into the anterior chamber trapping the conjunctiva into the incision is directly related to chemosis [6].

In this study, the “Bi-bevel " keratome was adopted, which prevented the posterior extension of the incision edges into the conjunctiva [16]. In addition, the average time from the introduction of the phacoemulsification probe into the anterior chamber to the onset of chemosis was 30.23 ± 49.23s, with all occurrences happening during the phacoemulsification stage. The earliest onset of chemosis was observed immediately after the phacoemulsification probe entered the anterior chamber, occurring in 42.42% of patients. Notably, these patients did not experience intraoperative conjunctival chemosis as a result of frequent passage of the phacoemulsification probe through the incision. Moreover, chemosis was observed to originate from the sides of the incision rather than from the center. Postoperative AS-OCT results revealed that the distance between the end of the Bowman’s layer to the external wound opening of CCI in the chemosis group was 772.6 ± 143.2 μm, which was significantly greater than that in the N-chemosis group. Therefore, it is speculated that CCI at this location may cause conjunctival rupture. Additionally, the labial conjunctiva surrounding the external CCI opening could wrap the phacoemulsification probe sleeve, creating a sealed passage. This allows fluid leakage from the anterior chamber to accumulate in the subconjunctival space, ultimately leading to chemosis (Fig. 3).

Fig. 3.

Intraoperative conjunctival chemosis formation process. A, B and C show the progression from initiation to expansion of chemosis, respectively; D shows that when chemosis initially occurred, the conjunctiva was incised immediately to prevent further development of chemosis. The black solid line arrows show the water passage formed by the ruptured conjunctiva around one side of the CCI wrapping the phacoemulsification probe sleeve, and the blue arrows show the flow direction of the leakage fluid from the anterior chamber into the subconjunctiva. The black dotted arrow shows the “7”-shaped conjunctival incision

Proper placement of the CCI is emphasized as a crucial measure to mitigate the risk of chemosis. However, there is currently a lack of standardized consensus or recommendations regarding the specific locations of CCI [2, 17]. Reported locations for CCI in the literature range from 0.5 mm to 1.5 mm anterior to the corneal limbus [17–19]. Additionally, some researchers use the anterior edge of the limbal vascular arcade as a reference for determining CCI placement [11]. In our study, the CCI was positioned using the anterior edge of the limbal vascular arcade as the anatomical landmark. However, the AS-OCT results on the first postoperative day indicated that the CCI was located posterior to the end of the Bowman’s layer in both chemosis and N-chemosis groups. On AS-OCT, the limbus was precisely defined as the range from the end of the Bowman’s layer to the scleral spur. Notably, within the Han population, the superior limbus exhibits an average width of approximately 1.4 mm [14]. It is suggested that the CCI is actually located in the limbus area in all patients included in current study. During slit-lamp examination, it was observed that conjunctival vessels extended into the clear cornea, projecting approximately 0.5 mm beyond the limbal edge [20]. However, the location of the limbal vessel edge varies among patients [17], thus leading to differences in the CCI location. In addition, for eyes with senile rings, the exact location of the limbus cannot be accurately determined using a slit-lamp [21], which may contribute to deviations in the CCI placement. Our study precisely quantifies the relationship between the incision placement and the risk of chemosis, highlighting that if the main incision positioned 770 μm beyond the end of Bowman’s layer, careful attention should be given to potential chemosis. Given these findings, preoperative evaluation with AS-OCT is recommended to account for anatomical variations in conjunctival attachment at the limbal region. This approach can enhance the precision of incision placement, thereby minimizing the risk of chemosis.

The present study is the first to report the distribution of conjunctival chemosis degrees at the end of surgery. Notably, the majority of cases (65.15%) exhibited chemosis in the 1–2 quadrant range, while only a small percentage (15.15%) displayed chemosis across all four quadrants. However, this study did not identify factors that might influence the degree of chemosis at the end of surgery. Although the distance between CCI and the end of the Bowman’s layer in the chemosis group was significantly greater than that in the N-chemosis group, it was not correlated with the degree of chemosis. Saleh et al. reported a case of CCI with incomplete intraoperative conjunctival chemosis (180 degrees superior conjunctiva) and believed that the pinhole of Tenon anesthesia in this case limited the further development of chemosis [22]. However, it’s important to note that all patients in this study underwent surgery under topical anesthesia without the presence of an additional external drainage channel on the conjunctiva. Therefore, factors affecting the degree of intraoperative chemosis need to be further confirmed by follow-up studies.

There were no intraoperative and postoperative complications recorded in this study, which is consistent with previous reports [3, 5, 7]. Compared with the non-chemosis group, the chemosis group had a slightly longer surgery time, but the difference was not statistically significant. There was no significant difference in CCT, IOP and BCVA between the two groups postoperatively. Previous studies have shown that the size and location of the corneal incision may affect SIA [23–25]. In this study, despite an average difference of 500 μm in CCI position observed between the two groups, no statistically significant difference in SIA was found postoperatively. Consistent with our findings, Wang et al. reported that there was no significant difference in postoperative SIA for CCI positions ranging from 0.5 mm to 1.5 mm anterior to the limbus [17].

During cataract surgery, the pooling of irrigating fluid on the operative field caused by conjunctival chemosis, making visualization difficulty and increasing the risk of complications. Various techniques have been proposed to deal with intraoperative conjunctival chemosis. Tilt the head towards the temporal side and add a drainage wick [5], make a small incision or insert a 26G needle at the highest point of the conjunctival mounds [7], or 5 mm behind the temporal limbus [4, 5, 15, 22], followed by blunt instrument extrusion to drain the fluid, seems to be the solutions to severe chemosis. Villada et al. suggested that pulling back the conjunctiva at the incision can reduce the incidence of chemosis when the phacoemulsification probe is introduced into the anterior chamber [6]. Poole et al. proposed that a 3.0 mm mini-peritomy of the conjunctiva at the incision site could significantly alleviate this problem [10]. In the initial stage of chemosis, we agree with the approach proposed by Poole et al. However, Ziakas NG et al. have raised concerns about its effectiveness, particularly in cases of scleral tunnel phacoemulsification [15]. Based on our findings, we suggest that proper proximity of CCI to the clear cornea can reduce intraoperative chemosis without increasing SIA. If chemosis does occur, we advise pausing the procedure and creating a 3.0 mm “7”-shaped conjunctival incision at the CCI site to disrupt the fluid passage, which can effectively prevent the aggravation of chemosis (Fig. 3). This approach aligns with the principle reported by Sugai et al., who found that a transconjunctival sclerocorneal incision with two small conjunctival cuts at both ends significantly reduced the incidence of intraoperative conjunctival chemosis compared to CCI [3]. Nevertheless, the effectiveness of this technique requires further validation through long-term, large-scale studies.

This study has certain limitations, including the involvement of only an inexperienced surgeon, which may introduce bias. In addition, AS-OCT measurements were limited to the central region of CCI, potentially overlooking critical information from the sides of the incision. The focus on the Han population further limits the generalizability of the findings due to variations in limbal width across different populations [14]. While we recognize the potential lack of novelty in this manuscript, we suggest that our findings have valuable clinical applications, including strategies for optimizing incision placement to reduce the risk of chemosis.

Conclusions

This study systematically recorded and analyzed varying degrees of intraoperative conjunctival chemosis, revealing a significant correlation between the location of the CCI and the occurrence of chemosis. Chemosis consistently originated bilaterally from the disinserted conjunctiva wrapping around the phacoemulsification sleeves, forming a closed channel. This channel allowed fluid leaking from the anterior chamber to accumulate beneath the conjunctiva, resulting in chemosis. Additionally, our study provided precise quantification of the relationship between the incision placement and the risk of chemosis, emphasizing the importance of closely considering potential chemosis if it extends beyond 770 μm from Bowman’s layer. Strategic placement of the CCI and careful management of conjunctival incisions could potentially reduce the incidence of chemosis.

Abbreviations

CCI

Clear corneal incision

SIA

Surgically induced astigmatism

I/A

Irrigation/aspiration

BCVA

Best corrected visual acuity

LOCS III

Lens opacities classification III

ACD

Anterior chamber deep

AL

Axial length

CCT

Central corneal thickness

K

Keratometry

Authors’ contributions

Y. Sun and W. Yuan contributed equally to this work. Y. Jiang was responsible for conception and design. W. Yuan collected the clinical samples. Y. Sun and W. Yuan collected the literatures, extracted and analyzed the data. The first draft of the manuscript was written by Y. Sun and all authors commented on previous versions of the manuscript. Z. Zhao was involved in reviewing the manuscript.

Funding

The study was supported by the Natural Science Foundation of Shanghai (Grant 20ZR1410100).

Availability of data and materials

The datasets generated or analyzed during the current study are included in this article. Further any enquiries can be directed to the corresponding author (atzzn1984@163.com).

Declarations

Ethics approval and consent to participate

This research was approved by the Ethics Committee of Eye & ENT Hospital of Fudan University and followed the tenets of the Declaration of Helsinki. All patients were informed about the study in detail. Informed consent has been obtained from the participants in this study.

Consent for publication

Not Applicable.

Competing interests

The authors declare no competing interests.