Introduction

Surgically induced astigmatism (SIA) is a prevalent cause of vision impairment following glaucoma surgery1. Trabeculectomy (LEC) remains the standard surgical intervention for glaucoma when intraocular pressure (IOP) cannot be effectively managed with pharmacological therapy or laser treatment; however, it is associated with a significant degree of SIA and consequent astigmatic changes, which can lead to visual deterioration and patient discomfort1,2. Minimally invasive glaucoma surgery (MIGS) has emerged as the preferred surgical option for patients with mild-to-moderate glaucoma who require additional intervention. Due to the minimal or nonexistent need for conjunctival sutures, antimetabolite therapy, or small incisions, the extent of SIA following MIGS is anticipated to be minimal1,2. For patients with mild to moderate open-angle glaucoma who also present visually significant cataract, combined cataract surgery (phacoemulsification with intraocular lens implantation) with minimally invasive glaucoma surgery (MIGS) is increasingly performed. This combined surgical approach has the clinical benefit of reducing patient risk and burden by avoiding two separate surgeries. Additionally, previous studies have indicated that lens extraction may improve trabecular outflow by deepening the anterior chamber, potentially providing additional intraocular pressure reduction effects when combined with MIGS3,4,5. We previously reported that in glaucoma patients undergoing concurrent cataract surgery, the first-generation trabecular micro-bypass stent (iStent, IS) and the second-generation iStent inject W (IW) significantly reduced IOP over one year, decreased the need for anti-glaucoma medications, and demonstrated a favorable safety profile6. Nonetheless, few studies have investigated the impact of MIGS on SIA.

In this study, we aimed to compare the SIA associated with three different MIGS techniques: IS, IW, and microhook ab interno trabeculotomy (µLOT).

Material and methods

This retrospective cohort study included the eyes of patients who underwent IS (N = 36) and μLOT (N = 36) from January 2018 to December 2019, and IW (N = 39) from October 2020 to June 2021 at our hospital. Common criteria for the three surgery groups were simultaneous cataract surgery and availability of recorded data including keratometry records obtained preoperatively and from 4 to 13 weeks postoperatively. The study adhered to the tenets of the Declaration of Helsinki, and the inclusion and exclusion criteria for IS and IW were in accordance with the Japanese Ophthalmological Society’s "Criteria for the Requirements for the Use of Trabecular Micro-bypass Stents with Phacoemulsification and Aspiration.” According to the criteria, the inclusion criteria for IS or IW were (1) mild to moderate open-angle glaucoma with cataract (mean deviation of − 12 dB or less on Humphrey visual field test), (2) IOP less than 25 mmHg with antiglaucoma medication, and (3) indication for cataract surgery with additional IOP reduction. Both eyes were included in the analysis of patients who underwent surgery. The IS or IW exclusion criteria were (1) patients who underwent other surgeries simultaneously; (2) secondary glaucoma, such as angle closure or inflammatory or neovascular glaucoma; and (3) a history of glaucoma surgery, such as LEC, goniosynechialysis, argon laser trabeculoplasty, selective laser trabeculoplasty, or glaucoma drainage device implantation. In addition, μLOT excluded two patients who had undergone LEC by the time of the postoperative keratometry measurement. Corneal astigmatism was measured using a keratometer (NIDEK ARK530A, Aichi, Japan) with a central diameter of 3 mm before and at 4–13 weeks postoperatively, and the SIA was examined by vector analysis. Additionally, we evaluated preoperative and postoperative IOP, medication scores, and refractive outcomes, including spherical equivalent (SE), refractive prediction error (RPE), absolute prediction error. RPE was calculated as the difference between actual postoperative refraction and predicted refraction using the Barrett Universal II formula.

Ethics approval

This study was conducted as part of a research protocol entitled “Clinical Study on Induced Astigmatism and Postoperative Outcomes of Minimally Invasive Glaucoma Surgery,” approved by the Institutional Review Board of Saitama Red Cross Hospital and complied with the tenets of the Declaration of Helsinki. The Saitama Red Cross Hospital Ethics Committee waived the need for obtaining informed consent for the use of patients’ medical record data in accordance with the provisions of the “Guidelines for Epidemiological Research” issued by the Japanese government and instead posted the protocol on its website to inform study participants. Only anonymized data were used for statistical analysis. Neither the patients nor public were involved in the design, or conduct, or reporting, or dissemination plans for this research.

Surgical procedures

All surgeries were performed on eyes with mydriasis induced using topical tropicamide phenylephrine eye drops and standard anterior chamber anesthesia with 1% lidocaine. After the completion of surgery, 1.5% moxifloxacin eye drops were administered.

Postoperatively, 0.5% moxifloxacin (ROHTO NITTEN Co., Ltd, Nagoya, Japan) and 0.1% betamethasone eye drops were applied topically four times daily for 3–4 weeks after surgery in all patients. IS, IW, and μLOT were performed simultaneously with cataract surgery. A single hydrophobic acrylate intraocular lens was inserted in all groups.

First- and second-generation trabecular micro-bypass stent

IS and IW were performed as previously reported6,7. Briefly, two 1-mm paracentesis incisions and a 2.4-mm upper-temporal clear corneal incision were created, and PEA + IOL was performed. The anterior chamber was then filled with an ophthalmic viscoelastic substance (OVD) (Healon V, AMO, Santa Ana, CA, USA), and one (IS) or two (IW) trabecular microbypass stents were inserted ab interno into the Schlemm’s canal through the lower nasal trabecular meshwork using a Hill Gonio lens; the OVD was removed, and the eye was filled with an isotonic solution.

Microhook ab interno trabeculotomy

Two 1-mm paracentesis incisions and a 2.4-mm temporal clear corneal incision were created, and μLOT was subsequently performed through the two paracentesis incisions after PEA + IOL was performed as previously reported8,9,10,11. A spatula-shaped microhook (M-2215; Inami & Co., Ltd., Tokyo, Japan) was designed for use during the μLOT. A viscoelastic material (1% sodium hyaluronate; Opegan Hi; Santen Pharmaceuticals or Healon V) was injected into the anterior chamber through paracentesis incisions formed in the inferior and superior auricular positions. A microhook was inserted into the anterior chamber through the paracentesis incisions using a Swan–Jacob gonioprism lens (Ocular Instruments, Bellevue, WA, USA) or an eye prism (Glaukos Corp. Laguna Hills, CA) to observe the angle opposite to the paracentesis incision. The tip of the microhook was then inserted into the Schlemm’s canal and moved circumferentially to incise the inner wall of the canal and trabecular meshwork over 2 h. The same procedure was followed with a microhook inserted through another paracentesis incision to perform a μLOT at the opposite angle. The μLOT spanned slightly over one-quarter of the circumference. After aspiration of the OVD, paracentesis and clear corneal incisions were closed by hydrating the corneal stroma.

Calculation of astigmatism parameters

Preoperative and postoperative astigmatism and SIA were calculated based on the SIA Calculator Version 2.1 developed by Sawhney and Aggarwal (http://www.insighteyeclinic.in/SIA_calculator.php)12 based on a vector analysis algorithm from keratometry values obtained preoperatively and 4–13 weeks postoperatively. The magnitude of the vector of preoperative and postoperative astigmatism and SIA, the meridian of the vector, and the mean arithmetic magnitude were determined for each surgical group. The American Society of Cataract and Refractive Surgery’s Astigmatism Double Angle Plot Tool version 1.1.0 (https://ascrs.org/tools/astigmatism-double-angle-plot-tool)13 was used to visualize the distribution of pre- and postoperative corneal astigmatism and SIA.

Statistical analysis

Patient demographics were compared using a generalized linear mixed model, with patients treated as random effects. The degrees of pre- and postoperative corneal astigmatism and SIA were analyzed using a linear mixed model in which patients were considered random effects. Linear mixed models were used to adjust for the hierarchical structure of data modeling by grouping measurements in subjects to reduce possible bias from including data from both eyes in one patient14. We explicitly employed linear mixed-effects models that included subject ID as a random intercept. This methodology is widely recognized for effectively managing intra-subject correlation in bilateral eye data15,16. By incorporating a random intercept for each subject, our analysis accounts for potential intra-subject correlations, thus appropriately addressing the non-independence of data from bilateral eyes. If significant differences were found between the three groups, then post-hoc pairwise comparisons were performed using the Tukey method.

All statistical analyses were performed using EZR (Saitama Medical Center, Jichi Medical University, Saitama, Japan),17 a graphical user interface for R (R Foundation for Statistical Computing, Vienna, Austria).

Unless otherwise noted, results are expressed as mean ± standard deviation, and a value of P < 0.05 was considered statistically significant.

Results

The demographic data of the participants, including age and sex, are presented in Table 1. While patients in the IW group tended to be younger, the difference was not statistically significant. No significant differences were found in sex distribution or laterality (left/right eye) among the groups (Table 1).

Table 1 Demographic characteristics of the study population.
Full size table

The astigmatic parameters are summarized in Tables 2 and 3. The arithmetic mean magnitudes of preoperative astigmatism were 0.92 ± 0.67 D, 0.9 ± 0.66 D, and 0.88 ± 0.56 D for the IS, IW, and µLOT groups, respectively, compared to 1.07 ± 0.68 D, 0.98 ± 0.6 D, and 1.01 ± 0.67 D postoperatively. No statistically significant differences were noted among the IS, IW, and µLOT groups in either the preoperative (P = 0.849) or postoperative (P = 0.86) measurements, as determined by the linear mixed-effects model (Table 2).

Table 2 Arithmetic mean of surgically induced astigmatism (SIA).
Full size table
Table 3 Centroid of surgically induced astigmatism (SIA).
Full size table

Although there was a slight postoperative increase in the arithmetic mean corneal astigmatism, this trend was not statistically significant across the IS, IW, and µLOT groups (P = 0.128, P = 0.511, P = 0.229, respectively; linear mixed-effects model) (Table 2). The arithmetic mean of SIA was 0.77 ± 0.45 D, 0.79 ± 0.41 D, and 0.76 ± 0.48 D for IS, IW, and µLOT, respectively, with no significant differences among the groups (P = 0.754, linear mixed-effects model) (Table 2).

Vector analysis indicated that the preoperative, postoperative, and SIA centroid sizes and confidence ellipses were comparable across the groups (Table 3, Fig. 1). The mean SIA vector was similar among the IS, IW, and µLOT groups, with values of 0.46 D at 52°, 0.37 D at 21°, and 0.15 D at 96°, respectively (Table 3, Fig. 1).

Fig. 1
figure 1

Scatter plots representing preoperative corneal astigmatism, postoperative corneal astigmatism, and surgically induced astigmatism (SIA) for each surgical group. The black dots in each figure represent the mean vector of preoperative corneal astigmatism, postoperative corneal astigmatism, and SIA for the group, while the yellow dots represent the corresponding values for each individual eye. Red circles indicate the 95% confidence intervals for the mean vector, and blue circles represent the 95% confidence intervals for the data set. All astigmatism values are expressed in plus cylinder format. Each ring represents 1.0 diopter (D). SIA, surgically induced astigmatism; IS, iStent; IW, iStent inject W; μLOT, microhook ab interno trabeculotomy.

Full size image

Preoperative intraocular pressure (IOP) differed significantly among the three groups (P = 0.04), with post-hoc analysis revealing that preoperative IOP was significantly higher in the μLOT group compared to the IW group (P = 0.04, Table 4). Additionally, IOP change significantly differed among the groups (P = 0.006, Table 4), with post-hoc analysis showing that the IOP reduction was significantly greater in the μLOT group compared to the IS group (P = 0.03) and the IW group (P = 0.008). However, postoperative IOP values did not significantly differ among groups (P = 0.16). While the magnitude of IOP reduction varied, the difference was not statistically significant (P = 0.16). Preoperative glaucoma medication scores were similar across groups (P = 0.25), but postoperative medication scores differed significantly (P = 0.02), with the IW group requiring fewer medications than the IS group (P = 0.015) (Table 4).

Table 4 Intraocular pressure (IOP), medication scores, and refractive profiles including spherical equivalents (SE) and refractive prediction errors.
Full size table

Regarding refractive outcomes, preoperative spherical equivalent (SE) values showed no significant differences among groups (P = 0.14). However, postoperative refractive prediction errors differed significantly (P = 0.025). Post-hoc pairwise comparisons indicated a significant difference between the IW and μLOT groups (estimated difference: 0.50 ± 0.18 D; P = 0.02), with the μLOT group exhibiting a slight myopic shift in refractive error, though both remained mild. Absolute prediction error did not significantly differ across groups (P = 0.062). Analysis of prediction accuracy within ± 0.5D and ± 1.0D also showed no significant differences among groups (Table 4).

The subgroup analysis based on glaucoma type—primary open-angle glaucoma (POAG), pseudoexfoliative glaucoma (PEX), and chronic angle-closure glaucoma (CACG)—is summarized in Table 5. Within the IS group, preoperative and postoperative IOP values, medication scores, SE, prediction errors, absolute prediction errors, and astigmatism did not significantly differ among glaucoma subtypes. In the IW group, absolute prediction error was the only parameter that significantly differed among subtypes (P = 0.034), but post-hoc pairwise comparisons revealed no significant differences between subtypes (CACG vs PEX, P = 0.89; CACG vs POAG, P = 0.20; PEX vs POAG, P = 0.095). In the μLOT group, no significant differences were observed among glaucoma subtypes for any parameter (Table 5).

Table 5 Additional information on the glaucoma patients including IOP changes in refractive profiles stratified by glaucoma subtype.
Full size table

Discussion

Previous studies have indicated that various MIGS generally does not influence refraction18,19,20. Specifically, the iStent trabecular micro-bypass stents, when combined with cataract surgery, have been shown to be refraction-neutral in terms of SIA18,19. A recent Korean study also reported that second-generation trabecular stents, when combined with cataract surgery, had only a minimal effect on refractive outcomes21. Similarly, µLOT is comparable to small-incision cataract surgery20, inducing less SIA than traditional LEC2. A systematic review by Chan and Kong1 reached a similar conclusion, noting that minimally invasive glaucoma surgery appears to be astigmatically neutral. A similar trend has also been observed in newer ab interno trabeculotomy procedures, such as Kahook Dual Blade (KDB). In a comparative study between phacoemulsification with KDB and phacoemulsification alone, the mean surgically induced astigmatism (SIA) was nearly identical between the two groups. The authors concluded that the addition of KDB does not cause extra astigmatism, with negligible postoperative corneal astigmatic changes observed in both groups22. Conversely, trabeculectomy consistently induces temporary astigmatism of approximately 1.0 D in the early postoperative period23,24,25.

In this study, the arithmetic means of SIA for IS, IW, and µLOT were 0.77 ± 0.45 D, 0.79 ± 0.41 D, and 0.76 ± 0.48 D, respectively, demonstrating mild and comparable values across the groups. Tanito et al., previously reported an arithmetic mean SIA of 0.62 ± 0.42 D after µLOT without cataract surgery, which was not significantly different from the SIA observed after microincision cataract surgery2,26. The mean SIA for microincision cataract surgery was 0.42 D with a 1.8-mm incision and 0.5 D with a 1.7-mm incision during coaxial phacoemulsification 20. Given that Tanito et al.2 performed µLOT without concurrent cataract surgery, while our study involved simultaneous cataract surgery with µLOT, the SIA results from our µLOT group seem reasonable. These findings suggest that IS, IW, and µLOT can be equally recommended to glaucoma patients with coexisting cataracts from a refractive perspective, because they are associated with mild SIA and only minimal postoperative increases in corneal astigmatism, with no significant differences observed among the groups. This finding is also consistent with a recent comparative study by Onoe et al.27, which compared ab interno trabeculotomy with microhooks and iStent inject W (both combined with cataract surgery) in patients with primary open-angle glaucoma (POAG). The study reported comparable surgical outcomes between the two groups, demonstrating significant reductions in intraocular pressure and decreases in the number of glaucoma medications required. Although their study primarily focused on IOP control and safety, the results also suggest that neither type of MIGS significantly impacts corneal optics. Our direct comparative data further confirm that both incisional angle procedures (e.g., µLOT) and implant-based procedures (e.g., iStent) result in minimal changes in postoperative refractive and astigmatic profiles.

It is also instructive to compare traditional MIGS procedures with subconjunctival gel stents such as the XEN gel stent, which is sometimes classified as a MIGS procedure despite involving subconjunctival vesicle formation through an ab interno approach. Even so, the effect on astigmatism is considerably less than with trabeculectomy. A retrospective study comparing cataract surgery combined with XEN versus cataract surgery combined with trabeculectomy reported significantly less induced corneal astigmatism and faster visual recovery in the XEN group28,29. Specifically, eyes receiving XEN gel stents combined with cataract surgery demonstrated a markedly lower incidence of higher-order myopic astigmatism and more rapid visual recovery compared to trabeculectomy. These findings further reinforce the notion that less invasive glaucoma surgeries generally have a gentler effect on corneal shape. Although XEN was not included in our current study, it is reassuring that all three MIGS procedures evaluated here (µLOT, IS, and IW) similarly induced only minimal astigmatism on average, with negligible changes in vector analyses and no statistically significant differences among techniques. These outcomes align well with the comparative results discussed above.

Multiple studies focusing on cataract surgery combined with MIGS have confirmed that adding a MIGS does not significantly impact refractive outcomes. Ioannidis et al.19 evaluated eyes undergoing cataract surgery combined with two iStent inject implants and reported a mean absolute error from the target refraction of 0.36 ± 0.25 D. In their cohort, 73.9% of eyes were within ± 0.50 D and 98.9% within ± 1.00 D of the target refraction. Additionally, 73.8% had postoperative residual refractive astigmatism ≤ 0.5 D, leading the authors to conclude that iStent combined with cataract surgery does not negatively affect refractive outcomes19. Similarly, Scott et al.18 demonstrated refractive results comparable to cataract surgery alone, with 80% and 95% of eyes achieving refractive accuracy within ± 0.5 D and ± 1.0 D, respectively. Rho et al.21 reported similar findings, noting a mean absolute refractive error of 0.33 ± 0.26 D, and 83.3% of eyes achieving spherical equivalent differences within ± 0.5 D when combining cataract surgery with iStent injections. In our previous study, µLOT also demonstrated a slight hyperopic tendency; however, both the refractive prediction error (0.09 ± 0.62 D) and the absolute prediction error (0.46 ± 0.42 D) were minimal and 63% of eyes achieving refractive accuracy within ± 0.5 D11. Our current study showed consistent outcomes, with 61% (µLOT), 71% (IS), and 74% (IW) within ± 0.5 D, and 86% (µLOT), 89% (IS), and 95% (IS) within ± 1.0 D, further supporting the minimal impact of MIGS on refractive accuracy.

The refractive neutrality of MIGS is likely due to its smaller incision size, absence of additional incisions when performed alongside cataract surgery, and minimal intraoperative incisional wound stretching, which reduces the likelihood of increased SIA. MIGS utilizes an ab interno or microincisional approach that does not require extensive scleral resection or suturing, thereby preserving the corneal shape. In contrast, trabeculectomy involves creating a scleral flap and placing sutures, often resulting in postoperative scarring and fibrosis. These changes can cause a temporary with-the-rule astigmatic shift ranging from 0.5 to 1.0 D during the early postoperative period23,24,25. Chan et al.1 noted that MIGS devices such as the iStent and CyPass stents, inserted via small incisions without conjunctival sutures or antimetabolite use, cause negligible astigmatic or refractive changes beyond those expected from cataract surgery incisions alone. Our findings align with these observations, showing minimal surgically induced astigmatism and refractive error after MIGS procedures (µLOT, IS, or IW). Ioannidis et al.19 also emphasized that the iStent uses existing small corneal incisions with minimal wound stretching, reducing alterations in corneal biomechanics. Overall, the minimal manipulation of corneal tissues in MIGS procedures likely explains their limited impact on surgically induced astigmatism.

Another important biomechanical factor is intraocular pressure (IOP) fluctuation. Conventional glaucoma surgeries often result in large fluctuations or significantly lowered IOP in the early postoperative period, potentially causing transient alterations in corneal curvature. In contrast, MIGS procedures typically achieve a more physiologic and gentle reduction in IOP. Hammel et al.30 demonstrated that the initial postoperative increase in corneal astigmatism associated with decreased IOP following Ex-PRESS shunt surgery resolved within three months as IOP stabilized. Similarly, Bormann et al.29 noted that changes in astigmatism due to postoperative hypotony were more pronounced in trabeculectomy compared with the XEN gel stent, reflecting differences in biomechanical impact between these surgical methods. MIGS procedures rarely induce extreme hypotony, thereby minimizing significant fluctuations in corneal shape. This characteristic likely explains the postoperative refractive stability observed in MIGS-treated eyes, as minor initial corneal shape changes rapidly normalize.

Importantly, MIGS procedures also support ocular surface health, indirectly contributing to refractive stability. Reducing dependence on glaucoma medications preserves the tear film and corneal surface integrity. Lee Jones et al.31 reported a significant reduction in glaucoma medication usage—from an average of 1.8 drops to 1.1 drops daily—after MIGS, accompanied by improved tear break-up time and reduced corneal staining at four months postoperatively. A healthier ocular surface environment reduces dry eye symptoms and irregular astigmatism related to ocular surface disease. Unlike trabeculectomy, which can exacerbate ocular surface dryness and irregular astigmatism due to filtering blebs or chronic medication use, MIGS techniques such as conjunctiva-sparing LOT and iStent procedures minimize these adverse effects. Our findings support this notion, as conjunctiva-preserving MIGS techniques appeared to prevent bleb-related corneal distortion and facilitated reduction or cessation of topical glaucoma therapy in some patients, further stabilizing the refractive outcomes.

In this study, we analyzed surgically induced astigmatism (SIA) and refractive prediction errors following IS, IW, and μLOT procedures in patients with POAG, PEX, and CACG. No significant differences were observed among these glaucoma subtypes. Due to the relatively small sample size, particularly in the CACG group, these results should be interpreted with caution; however, our findings align with the limited available literature. Although most existing studies predominantly focus on POAG, previous reports have also shown comparable astigmatic outcomes even when patients with PEX or angle-closure glaucoma were included. For instance, in the aforementioned study of Kahook Dual Blade (KDB) trabeculotomy combined with cataract surgery22, approximately 21% of eyes had secondary glaucoma (including PEX, pigmentary glaucoma, or PACG), but astigmatic outcomes were similar to POAG, and no abnormal corneal changes attributed to trabecular meshwork manipulation were noted.

These observations suggest that eyes with PEX or previously narrow angles respond equally well to MIGS as do eyes with POAG. Our findings support this: eyes with PEX in our cohort showed no additional surgery-induced astigmatism compared to POAG eyes. However, PEX syndrome is known to be associated with zonular weakness32 and potential intraocular lens (IOL) tilt or decentration, factors that could theoretically induce astigmatism or higher-order aberrations. Ioannidis et al.19 highlighted that pseudophakic eyes with PEX may have an increased risk of postoperative IOL mispositioning, such as mild lens tilt or subluxation33,34. Nonetheless, neither their study nor others have reported significant refractive surprises following MIGS in PEX eyes. In our study, careful surgical technique—including secure IOL insertion within the capsular bag—likely mitigated such risks, as indicated by the stable postoperative astigmatic outcomes observed in our PEX subgroup.

Thus, our data support the notion that MIGS procedures induce minimal astigmatism irrespective of the underlying glaucoma subtype. The consistent astigmatic neutrality across glaucoma types observed in this study reinforces our conclusion that minimal surgically induced astigmatism is a fundamental advantage of MIGS.

Nevertheless, this study has several limitations. First, its retrospective design limits the control over certain variables. Additionally, SIA was assessed at a single postoperative time point, and the keratometer used in this study has higher variability in detecting low levels of astigmatism, which may have contributed to variability in the results. Other limitations include differences in the timing of postoperative keratometry measurements across the groups, the involvement of multiple surgeons, and variations in incision types. However, these factors are likely minor and consistent with previous studies.

In conclusion, MIGS combined with cataract surgery induces minimal SIA, with no significant differences observed among IS, IW, and μLOT groups. Postoperative corneal astigmatism remained stable, and refractive predictability was maintained across all techniques. While μLOT was associated with a slightly greater IOP reduction and a minor myopic shift, the absolute prediction error did not differ significantly between groups. Additionally, subgroup analysis confirmed the refractive neutrality of MIGS across different glaucoma subtypes, including POAG, PEX, and CACG. These findings reinforce the safety and predictability of MIGS in preserving corneal optics when combined with cataract surgery. From a refractive standpoint, IS, IW, and μLOT can be equally recommended. Future studies with larger sample sizes and longer follow-up periods are warranted to further substantiate these findings and optimize surgical strategies for glaucoma patients undergoing cataract surgery.