Evaluation of Macular Thickness in Glaucoma Patients Using Prostaglandin Analog Eye Drops Undergoing Trabeculectomy with Mitomycin C: Prospective, Comparative, Randomized, Masked Examiner Study

According to the World Sight Report 2019, an estimated 2.2 billion people live with visual impairment or lack of vision, with more than 1 billion cases treatable and preventable. Among the blind world population, glaucoma is the leading cause of irreversible nature [1].

Early diagnosis and treatment are essential to prevent glaucoma morbidity. Lowering intraocular pressure (IOP) is the only proven means of interrupting or delaying the progression of glaucoma and preserving quality of life [2]. First-line treatment usually involves a topical medication (eye drops) to reduce IOP or through selective laser trabeculoplasty [3].

The choice of the most appropriate medication for each glaucoma patient may be influenced by cost, adverse effects and/or dosage schedules. Prostaglandin analog (PA) eye drops are the main choice for initial medical therapy. These drugs act mainly by increasing the outflow of aqueous humor via the uveoscleral pathway, resulting in a substantial reduction in IOP [4]. They are used on a large scale due to their efficacy, simplified dosage, especially local side effects and few or no systemic adverse effects [5]. However, there are reports of a direct relationship between its use and the development of macular edema (ME), which could lead to reduced visual acuity and harm to the user [6].

Trabeculectomy with mitomycin C (TREC with MMC) consists of the creation of a protected scleral fistula to divert aqueous humor drainage to the subconjunctival and subtenon spaces [7]. This surgery is widely used for the control of glaucoma, with the ability to produce low IOPs, thus being used even in eyes with a more rigorous target IOP requirement [8]. Among the described possible complications of TREC is ME, which can cause visual loss and thus affect the visual recovery of the patient undergoing this surgery [9, 10].

There are limited reports of macular evaluation and the development of cystoid ME after incisional surgery for glaucoma [10, 11]. The peripheral iridectomy performed in the usual surgical technique of TREC could lead to an increase in prostaglandins and cytokines in the anterior chamber. These inflammatory mediators could even reach the posterior pole of the eye and increase the permeability of the blood-aqueous barrier, also generating vascular leakage, which can lead to retinal thickening [10].

Thus, it is possible to formulate the hypothesis that the combination of TREC surgery itself with the usual use of PAs in patients with glaucoma could lead to a change in macular thickness in the postoperative period. However, there is no record in the literature (research in the MEDLINE/PubMed database with the keywords macular thickness and trabeculectomy, prostaglandin, and macular edema after trabeculectomy) on the influence of PA eye drops on macular thickness after TREC with MMC. The present study was designed to fill this gap.

This was a randomized, prospective, comparative clinical trial whose experimental design followed the Consolidated Standards of Reporting Trials (CONSORT) protocol. The included patients were from the Glaucoma Outpatient Clinic of the Goiás Eye Bank Foundation or the Reference Center in Ophthalmology of the Federal University of Goiás in Goiânia, Goiás, Brazil, between 2020 and 2022. If both eyes were eligible, only the eye with the first surgery was included in the study. In fact, just 5 patients had both eyes initially qualified for the study. Both the surgeon (single) and the technician responsible for the exams were masked to the study groups. All patients with glaucoma with an indication for TREC with MMC alone and in the clinical use of any PA eye drops (latanoprost, travoprost, bimatoprost, or tafluprost) in the eye to be operated on for at least 6 months were considered for inclusion.

The study adhered to the principles of the Declaration of Helsinki and was authorized by the Research Ethics Committee of UFG under the Certificate of Presentation for Ethical Assessment number 40055620.8.1001.5083. It was registered at http://clinicaltrials.gov under the number NTC06000280. All participants agreed to participate voluntarily by signing the informed consent form.

The inclusion criteria of the study were as follows: patients at least 18 years of age with primary glaucoma (primary open-angle glaucoma [POAG], primary angle close glaucoma, normal pressure glaucoma, pigmentary glaucoma, or pseudoexfoliative glaucoma), use of any commercially available PAs eye drops (latanoprost, travoprost, bimatoprost, or tafluprost) for at least 6 months, visual field changes – typical defect compatible with glaucomatous lesion [12] and/or typical anatomical impairment of the optic disc or retinal nerve fiber layer (such as Hoyt’s sign, optic disc cup ratio greater than 0.7, localized defect in the neural rim, or cup asymmetry) [13], who had an indication for TREC with MMC at the medical discretion (target IOP not established with maximum tolerable clinical medication and/or impossibility of using medications due to allergies and/or financial conditions). To be included, the operated eyes required IOP reduction by at least 20% compared to the baseline IOP on the last postoperative day (PO 30) and the absence of use of any ocular hypotensive medication.

Exclusion criteria were any pathology that could interfere with the test results, such as cataract, corneal edema (such as leukoma, ulcers, keratopathies); low quality of the optical coherence tomography (OCT) exam (signal strength ≤5/10) [14]; any previous macular pathology (such as macular hole, diabetic maculopathy, age-related macular disease); use of oral acetazolamide; advanced glaucoma with threat of or impaired fixation (due to the risk of PTA suspension in these cases); any surgical complications including hypotonia (IOP <6 mm Hg) in any evaluation; combined cataract and glaucoma surgery; history of cataract surgery less than 6 months before inclusion; need for surgical reintervention during follow-up for any reason; and need for reintroduction of topical antiglaucomatous therapy during follow-up.

Eligible volunteers were previously randomized through the website www.randomization.com into 2 groups: a study group (SG) and a control group (CG). The initial objective was to include 40 eyes, 20 per group. The sampling power was calculated from the mean difference (Δ) and the mean standard deviation (σ) of the parameters analyzed preoperatively and postoperatively in the groups with and without prostaglandin. For this purpose, the software G. Power^® 3.1 was used. To obtain the sample effect size, the study sample size (n = 40), alpha value of 0.05, desired sampling power of 0.80, and confidence interval of 0.95 were provided a priori. Thus, there was a minimum estimate of 21 eyes in total (sum of both groups) (95% CI: 13–29) to reach 80% power.

In the SG, the patients were instructed to continue using the eye drops until the day of surgery. In the CG, patients were instructed to discontinue PA between 30 and 60 days of the preoperative exams, which were performed no more than 15 days before surgery for both groups, with a minimum of 45 days of washout before surgery. In the CG, the patients were instructed to use 1 drop of 5 mg/mL carmellose sodium eye drop at the same time that they used the PA before its suspension. Pseudophakic patients for more than 6 months could be included in each group in the prior order of randomization. In the CG, the patients were informed to keep taking the eye drops and to add the other classes of ocular hypotensive drugs if they were not using them (except miotics and oral acetazolamide), with the aim of at least remedying the loss of the ocular hypotensive effect of the PA.

The subjects were evaluated, including best-corrected visual acuity (BCVA) using the Snellen chart at 6 meters, slit lamp biomicroscopy (XCELL 255, Reichert Inc., Depew, NY, USA), IOP measured in a calibrated Goldmann tonometer (CT210, Reichert Inc., Depew, NY, USA), gonioscopy with 4-mirror Goldman lens (Volk Optical Inc., Mentor, OH, USA), fundoscopy for mydriasis with 78D lens (Volk Optical Inc., Mentor, OH, USA), ophthalmoscopy indirect binocular with 2.0D lens (Volk Optical Inc., Mentor, OH, USA), and OCT with Cirrus 4000 (Zeiss Inc.), macular protocol (macular thickness and ganglion cell analysis) in the preoperative period (maximum 15 days before surgery), and postoperatively on 3 occasions: 1–3 days (“PO1”), 6–9 days (“PO7”), and 27–30 days (“PO30”). All exams were performed by the same technician, who was trained and experienced to perform them. In addition, the patients underwent visual field examinations (Humphrey Field, model HFA II ÿ750, Carl Zeiss-Meditec, Dublin, CA, USA) prior to surgery with the Swedish Interactive Threshold Algorithm (SITA) Standard 24–2 strategy up to 2 months before the procedure.

Macular thickness was assessed by spectral domain OCT obtained using a Cirrus 4000 apparatus (Zeiss Inc.) and macular scanning protocol (macular thickness and ganglion cell analysis). The calculations were based on data from the analytic software Ganglionar and Macular Thickness: Macular Cube (6 × 6 mm 512 A-scans × 128 B-scans centered around the fovea). The macular thickness map using the 1 mm, 3 mm and 6 mm circles from the Early Treatment Diabetic Retinopathy Study (ETDRS) was used to assess the thickness of the 9 subfields. The regions were designated as superior, inferior, temporal, and nasal parafoveal (subfields of the 3 mm rings) and perifoveal (subfields of the 6 mm rings) in addition to the central subfield. OCT images were excluded if the signal strength was ≤5/10, and a new test was performed immediately.

TREC with MMC consisted of opening the conjunctival base of the fornix and applying 0.4 mg/mL MMC for 2 min, as previously described [15]. The postoperative regimen in both groups included the use of 1% prednisolone eye drops (Ster^® União Química Farmacêutica Nacional S/A., Pouso Alegre, MG, Brazil) starting every 2 h, with weekly reduction for 6 weeks (4/4 h, followed by 6/6, then every 8 h, then 12/12 h and once daily), and moxifloxacin eye drops (Vigamox^®, AlconLabs, Fort Worth, TX, USA) every 6 h for 10 days.

Statistical analysis was performed using the Statistical Package for Social Science software (IBM Corporation, Armonk, USA) version 26.0. The characterization of the sample was performed using absolute frequency, relative frequency, mean and standard deviation. Student’s t test and χ² test were applied to the distribution of the macular profile and parameters in the groups with and without prostaglandin. The normality of the data was assessed using the Shapiro‒Wilk test. The delta and the percentage change were calculated on PO1, PO7 and PO30 compared to the preoperative period between the groups with and without prostaglandin. The Spearman correlation was performed between the deltas of the thickness with the mean deviation (MD) of the visual field and IOP in the groups with and without prostaglandin, and then the effect of the presence or absence of the lens was weighted (partial correlation). The adopted significance level was 5% (p < 0.05).

Forty eyes were randomized in the study initially, of which 17 were excluded because they did not meet the eligibility criteria of the study in the postoperative period. Among these patients, 4 patients did not undergo OCT on PO1, 4 patients did not undergo OCT on PO7, 2 patients did not undergo OCT on PO30, and 1 patient had a signal ≤5/10 OCT at all visits, even with several repetitions of the tests. Additionally, 3 patients underwent surgical reintervention during the follow-up (suture dehiscence), 2 patients had hypotonia (1 in each group), and 1 patient did not have a decrease in IOP of at least 20% from baseline. Thus, as there was a significant reduction in the number of patients initially estimated for the sample, new randomization was performed, using the same process of initial randomization described above to achieve this objective. In the end, the evaluated sample consisted of 35 eyes of 35 patients, with 17 eyes of 17 patients in the CG and 18 eyes of 18 patients in the SG (Fig. 1).

A total of 20 (57.1%) men (8 in the SG and 12 in the CG) and 15 (42.9%) women (10 in the SG and 5 in the CG) were included, p = 0.1. Regarding laterality, there were 14 right eyes (40%) (6 in the SG and 8 in the CG) and 21 left eyes (60%) (12 in the SG, 9 in the CG), p = 0.4. Thirty-three eyes (94.3%) had open angles (17 in the SG: 15 POAG, 1 juvenile glaucoma, and 1 pigmentary glaucoma; and 16 in the CG: 14 POAG, 1 juvenile glaucoma, and 1 pigmentary glaucoma), and 2 (5.7%) had primary angle closure glaucoma (1 in each group), p = 0.9. Regarding lens status, 26 eyes were phakic, 16 in the SG (88.9%) and 10 in the CG (58.8%), and 9 were pseudophakic, 2 in the SG (11.1%) and 7 in the CG (41.2%), p = 0.1. All eye drops used during the study were any of the commercial presentations available in Brazil. No preservative-free eye drop was used in any group. Travoprost was used in 9 eyes in the SG (8 in the CG); latanoprost in 1 eye in the SG (5 in the CG); and bimatoprost was used in 8 eyes in the SG (4 in the CG). No eye in either group was using tafluprost. Additionally, 17 eyes in both groups were in use of timolol maleate, 16 eyes in the SG were in use of brimonidine tartrate (12 in the CG), and 12 eyes in the SG were in use of topical carbonic anhydrase inhibitors (14 in the CG). The patients from the SG were in use of PA for 27.82 ± 26.39 months (26.94 ± 32.48 months in the CG, p = 0.8) at the time of inclusion in the study. Considering the other ocular hypotensive eye drops, the patients from the SG were in use for 29.06 ± 17.18 months (31.83 ± 34.85 months in the CG, p = 0.6) also at inclusion.

The mean age of the participants was 54.89 ± 15.76 years in the SG and 61.47 ± 15.26 years in the CG (p = 0.2). The visual field MD preoperatively was −20.17 ± 10.79 dB in the SG and −11.82 ± 15.92 dB in the CG (p = 0.08). The preoperative macular parameters of OCT in each group are shown in Table 1. The BCVA (LogMar) in the SG ranged from 0.16 ± 0.18 preoperatively to 0.24 ± 0.23 at PO1 (p = 0.02), 0.20 ± 0.20 at PO7 (p = 0.2), and 0.23 ± 0.24 at PO30 (p = 0.4). In the CG, the BCVA (LogMar) ranged from 0.17 ± 0.17 preoperatively to 0.26 ± 0.20 at PO1 (p = 0.1), 0.23 ± 0.16 at PO7 (p = 1.0), and 0.16 ± 0.17 at PO30 (p = 1.0). Regarding the difference in BCVA, there was no statistically significant difference between the groups in the preoperative period with PO1 (p = 0.8), PO7 (p = 0.8), and PO30 (p = 0.3).

There was an IOP reduction from 24.83 ± 7.86 mm Hg preoperatively in the SG to 15.75 ± 5.70 mm Hg at PO1 (p = 0.002), 13.31 ± 5.28 mm Hg at PO7 (p < 0.001), and 12.06 ± 3.09 mm Hg at PO30 (p < 0.001). In the CG, IOP ranged from 22.88 ± 6.01 mm Hg in the preoperative period to 11.89 ± 4.32 mm Hg at PO1 (p < 0.001), 11.47 ± 3.41 mm Hg at PO7 (p < 0.001), and 12.37 ± 3.58 mm Hg at PO30 (p < 0.001). There was no statistically significant difference between groups regarding the reduction in IOP compared to preop in PO1 (p = 0.8), PO7 (p = 0.6), and PO30 (p = 0.5).

Tables 2-4 show the absolute variation of each macular OCT variable compared to preop and its comparison between groups at different time points (Δ). There was equivalence in all of the comparisons evaluated. Figure 2 illustrates the change in macular parameters in percentage compared to preop of the 9 macular subfields in both groups. Taking into account the difference in percentage analyzing preop versus PO1, PO7, and PO30 showed no statistically significant difference between groups (p > 0.05). Correlations were performed between the deltas of the thickness of the macular parameters between the preoperative period and the PO30 within the groups with the possible confounders (MD and IOP) and controlling for the effect of the lens (partial correlation). There was no significant correlation in the tests performed or in relation to the possible influence of the lens on each of them (p > 0.05 for all tests).

The influence of PA eye drops on the onset of ocular inflammation or even on changes in macular thickness has been previously reported in the literature [16]. However, although it is possible to suppose that this causal effect can be magnified by any eye surgery, even by the inflammation caused, inherent in each procedure, there are no reports of its influence on macular thickness after TREC with MMC, a widely used surgery worldwide to control IOP in glaucoma patients.

The PA eye drops could, in theory, influence the effectiveness of TREC by increasing the inflammatory process. However, this was not an objective of the study, and therefore, there was no evaluation in this regard. Because postoperative inflammation is naturally expected, especially in the first postoperative days, and PA eye drops tend to be completely eliminated from the eyes after discontinuation within 4 weeks [17], macular evaluation was chosen up to 30 days after surgery, including evaluations from the first day (PO1) to the seventh day (PO7). Thus, it is believed that the period most likely to have a possible effect of PAs on macular thickness after TREC has been evaluated.

Previous studies have addressed changes in macular thickness after TREC and possible mechanisms that could contribute to its thickening [18, 19]. The reduction in IOP could lead to direct changes in the internal force of the retina and an indirect effect via subscleral deformations transmitted to adjacent compliant intraocular tissues, such as the macula itself [20, 21]. In addition, changes in the relationship between capillary pressure and interstitial fluid pressure could occur after the change in IOP expected in TREC [11, 18, 19]. Although some researchers have reported a considerable increase in macular thickness after surgical treatment for IOP reduction [11, 18, 22, 23], there are reports of persistent reduction in macular thickness after fistulizing surgery [24]. The possible isolated influence of IOP reduction on macular thickness is balanced between the groups in the present study, as the IOP reduction was significant and equivalent between them throughout the postoperative follow-up. Thus, it is believed that this possible confounding factor regarding macular thickness has been mitigated.

However, it has been previously proposed that macular thickness may remain unchanged after surgical IOP reduction. In a previous prospective study involving 45 eyes [25], the authors found that there was no significant correlation between changes in retinal thickness and IOP reduction after TREC or changes in mean extrafoveal retinal thickness after surgery, with or without adjuvant agents (MMC or 5-FU) in the evaluated interval. The researchers also concluded that the slight retinal thickening on the 2 postoperative days evaluated was lower than the 200 µm increase reported in a previous study [19] in hypotonic eyes and could be attributed to a prolonged subclinical inflammatory response of the retina. Thus, hypotonia itself could be a triggering factor for the change in macular thickness and perhaps not the expected decrease in IOP after successful TREC, and for this reason, it was considered an exclusion criterion from the study.

Even with a statistical equivalence in the severity of glaucoma measured by the MD of the visual field in the preop: −20.17 ± 10.79 dB in the SG versus −11.82 ± 15.92 dB in the CG (p = 0.08), there is no mistaking that more advanced glaucoma in the SG could influence the detection of eventual macular changes after TREC as it is widely known that patients with more advanced glaucoma may present changes in the macular region due to the loss of nerve fibers triggered by the glaucoma itself [26]. However, the groups were also equivalent in the preoperative macular parameters (Table 1), even with values of each variable very close to each other, at least limiting (or even preventing) the possible influence of this bias on the final results.

In the comparison between groups, when considering the variation between the preoperative period and the last observation performed (PO30), no OCT parameter showed statistically significant variation between groups (Tables 2-4). In fact, minimal change in the OCT parameters was observed, basically within the normal variability of the device used for the macular parameters studied [27]. Therefore, these results can be considered within the natural and expected fluctuation for the OCT test and thus can be interpreted as practically equivalent numbers. Even with a large number of tests performed to verify any influence of PA on macular thickness after TREC, and consequently incur a possible increase in alpha error [28], there was no significant variation between the groups considering the macular OCT parameters.

In a prospective study that evaluated 34 eyes undergoing fistulizing surgery with and without MMC, there was no significant difference in macular thickness change in patients using PA compared to those using other topical medications [11]. However, in that study, 25 eyes were evaluated using PA, and only 9 were evaluated without eye drops of that class. In addition, not all patients in the study used adjuvant MMC intraoperatively (used in 18 eyes), creating another confounding factor for the analysis, in addition to the large difference in “n” in each group. Thus, it is not possible to infer an accurate conclusion regarding the role of PAs in the variation in macular thickness after TREC in the results presented by Sesar et al. [11].

The surgical procedure itself may be responsible for changes in macular thickness. Vessani et al. [29] evaluated 54 eyes of patients with glaucoma divided into surgical (29 eyes) and control (25 eyes, with stable glaucoma) groups to assess fundus parameters by OCT. An increase in mean macular thickness of 4 µm was demonstrated 1 month after surgery and was more significant after surgery (TREC – 28 eyes or drainage implant – 1 eye) and was maintained for at least 2 months. This change did not correlate with the severity of glaucoma, similar to the findings of the present study [29]. In the study by Kadziauskiene et al. [18] with 106 eyes of 100 patients, which analyzed the macular thickness after TREC with and without 5-FU, the authors reported statistically significant changes in macular parameters measured by OCT but also within the variability of the device used. However, the interference of PAs between the treated eyes was not analyzed [18].

Finally, the correlations between the variations in the thickness of the macular parameters between the preoperative period and the PO30 within the groups were used to test the influence of possible confounders (glaucoma severity, as measured by the MD and IOP) in the model. No significant correlation was observed (p > 0.05 for all tests). Another possible confounding factor regarding the evaluations performed is the influence of the lens (or rather, the pseudophakia) as a direct synergistic effect in the induction of ME. This could occur because pseudophakic patients, in addition to a previous history of surgery and consequent inflammation (even in limited and controlled amounts, such as in phacoemulsification), may also have a breakdown of the blood-aqueous barrier [30], with a possible greater propensity for postoperative macular changes. Thus, a partial correlation was performed with the confounders tested above, with the presence of the lens as a control. There was no significant result in either group, suggesting the possible absence or at least a minimal influence of the lens in this regard. In a previous systematic review to evaluate the effect of the use of PAs on the development of cystoid ME after cataract surgery, there was no evidence to support the discontinuation of the use of eye drops before or during phacoemulsification to reduce this possible macular change [31].

Despite the strengths, this randomized, comparative study has some limitations, such as the relatively small sample size or the short-term follow-up time, despite apparently being adequate to assess OCT macular changes after TREC. Furthermore, objective measurement of cells in the anterior chamber and bleb morphology were not assessed and compared between groups since PA use theoretically may influence and exaggerate inflammation, directly affecting these parameters.

There was no change in the macular parameters of OCT after TREC during the 30-day postoperative follow-up, suggesting that the PA does not influence macular thickness in glaucomatous patients undergoing TREC.

The study was authorized by the Research Ethics Committee of the Federal University of Goias, assessment number 40055620.8.1001.5083. Written informed consent was obtained from all participants prior to the study.

The authors have no conflicts of interest to declare.

There was no sponsor or funder participation in the design, execution and analysis of the study or manuscript conception. There was no funding received for the study, just the authors’ own work.

S.C.A.B.S.: acquisition, analysis, interpretation of data for the work; conception and design of the work; and final approval of the version to be published. L.M.: conception of the work, drafting the work and reviewing it critically for important intellectual content, and final approval of the version to be published. M.P.A.: agreement to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.

Meeting presentations: This paper was presented as a poster at the XX Brazilian International Glaucoma Symposium, 2023, Porto de Galinhas, PE, Brazil.Trial registration: ClinicalTrials.gov; identifier: NCT06000280; August 17, 2023; Evaluation of Retina in Patients with Glaucoma Using Topical Prostaglandins Undergoing Trabeculectomy Surgery (TRAB).

The data that support the findings of this study are openly available in Biblioteca Digital de Teses e Dissertações da UFG at https://repositorio.bc.ufg.br/tede/.