The Efficacy of Glucocorticoids in the Treatment of Nonarteritic Anterior Ischemic Optic Neuropathy: A Systematic Review and Meta-Analysis.

Background: To evaluate the clinical effects and safety of glucocorticoids for patients with nonarteritic anterior ischemic optic neuropathy (NAION).
Methods: The databases MEDLINE, Embase, PubMed, Cochrane Database, and Web of Science were used to search for the relevant studies, and full-text articles that reported on the evaluation of glucocorticoids vs. no-treatment or placebo for patients with NAION. Review Manager 5.4 was used to estimate the pooled effects of the results among selected studies. Forest plots, funnel plots, and Begg's rank correlation were also performed on the included articles.
Results: A total of 983 patients were contained in the 9 studies that satisfied the eligibility criteria. The meta-analysis showed that, compared with the control group, the glucocorticoid group had significantly improved the VA (MD: -0.25, 95% CI [-0.45, -0.05], P = 0.02), VF (MD: -0.50, 95% CI [-0.94, -0.07], P = 0.02), and RNFL (MD: -14.10, 95% CI [-26.41, -1.79], P = 0.02) in NAION patients and had a high improvement rate of VA (RR 1.31, 95% CI [1.12, 1.52], P = 0.0005). No significant publication bias was observed in our study. Discussion. Our research preliminarily confirmed the effectiveness of glucocorticoids for NAION treatment, but more high-quality RCTs focusing on the hormone adverse reactions should be performed to verify our conclusions.

1. Introduction

Ischemic optic neuropathy (ION) is a vascular disease [1]. It occurs when the nutrient blood vessels of the optic nerve have circulatory disturbances [2]. The disease primarily occurs in one eye, but it can also occur in both eyes simultaneously or successively [3]. The incidence rate in the population is 0.5 per 100,000, and the disease most often occurs in people over 50, whose incidence is up to 2.3 ~ 10.2 per 100,000 [4, 5]. Pathologically, ION is divided into arteritis and nonarteritic ION, and 95% of ION is nonarteritic anterior ischemic optic neuropathy (NAION) [6, 7].

The pathological mechanism of NAION is still unclear [8, 9]. Some studies believe that the posterior ciliary artery that supplies the anterior part of the optic disc (i.e., the anterior and lamina area) has stenosis, occlusion, or perfusion pressure decrease, resulting in local microcirculation disorders, and the decreased optic papillary perfusion pressure will lead to insufficient blood supply to the optic disc, resulting in optic nerve ischemic disease [10, 11]. Optic disc edema often occurs in the acute phase. Although the edema will eventually disappear without any treatment, clinical observations have found that it does not change rapidly within 2 weeks [12]. The ischemic location and the degree of edema are important reasons that affect the central vision of NAION patients [13].

Glucocorticoids are anti-inflammatory and can reduce edema [14]. They can reduce capillary permeability and inhibit free radical injury. They have been used in the treatment of diseases with angiogenic edema as the main pathogenesis, such as intracerebral hemorrhage, brain tumor, and craniocerebral trauma [15, 16]. The pathological mechanism of NAION is often considered to be hypoperfusion or small vessel embolism in pathological mechanism. The optic disc edema caused by NAION cannot be completely defined as angiogenic or cytotoxic edema. Systemic hypoperfusion, nocturnal hypotension, local autoregulation dysfunction, vasospasm, venous obstruction, and thrombosis been identified as the inducing factors of NAION, and the application of hormones often leads to the rise of blood pressure and blood glucose. Therefore, clinicians use hormones to quickly reduce optic disc edema and restore visual function. At the same time, they often worry about the impact of its adverse effects on NAION patients and the accuracy of curative effect [17].

Literature has shown that the application of glucocorticoids in the treatment of NAION can achieve better curative effects [14, 18]. We used the meta-analysis method to systematically evaluate the clinical efficacy of glucocorticoids in the treatment of NAION and objectively evaluate the effectiveness and safety of glucocorticoids, so as to provide references for its further clinical research and application.

2. Methods

2.1. Literature Search Strategy

We used comprehensive databases (MEDLINE, Embase, PubMed, Cochrane Database, and Web of Science) to search for previous studies that investigated the effects of glucocorticoids on the clinical outcomes in patients with ION. The literature search was performed from inception up to November 30th, 2021, using the following keywords and MeSH terms: (1) glucocorticoids; (2) ION; (3) corticosteroid; and (4) optic neuropathy. Numerous combinations of words and strings were applied with Boolean operators “AND” and “OR” to broaden the search. Our literature search was carried out without any consideration for publication status limitations or language restrictions. The reference lists of the retrieved studies and review articles were examined manually to identify further relevant studies not identified by the search strategy.

2.2. Study Selection

A study was included if it satisfied the following criteria:

Researches compared patients who received glucocorticoids and other therapy

The study contained patients with ION

The study contained indicators evaluating efficacy between glucocorticoids and other therapy

The study was available in full text

The exclusion criteria were as follows:

Researches did not meet the inclusion criteria

The outcomes of interest were not reported or were impossible to use

The document was a review, abstract, letter, or a duplicate publication

2.3. Data Extraction

Data extraction was conducted independently by two reviewers (Zhou and Zhang), and the disagreements were resolved in consultation with a third reviewer. The number of patients in each study group was recorded. We also extracted data for study location, study design, intervention mode, patients' characteristics (age and gender), year of outset, and time of follow-up.

2.4. Quality Assessment

To qualitatively evaluate the quality of the included studies, the Cochrane Collaboration's tool was used to assess the quality of randomized controlled trials (RCTs), and the Newcastle–Ottawa Scale (NOS) was used to assess the quality of nonrandomized trials.

2.5. Statistical Analysis

The review manager (Version 5.4, Cochrane Collaboration, 2020) was used to estimate the pooled results in the selected studies. Chi-square test and I2 statistics were used to test the heterogeneity. An I2 value of 0%-50% indicated that heterogeneity was not relevant/important; 30%-50% suggested moderate heterogeneity; 50%-90% represents substantial heterogeneity; and 75%-100% represented considerable heterogeneity. A fixed effects model was applied in the absence of heterogeneity, while a random effect model was used when heterogeneity was observed. If there were more than 5 studies included in the meta-analysis, the data would be evaluated for publication bias by viewing the symmetry of the funnel plot and using the Begg rank correlation method.

3. Results

3.1. Search Process

The search yielded 1263 titles. After removal of duplicates, 1027 articles were identified. After manually inspecting the titles and abstracts, 920 articles were excluded. From these, 98 articles were further excluded due to various reasons including that they had a different study design, insufficient data available, or because they were review articles. Ultimately, 9 studies included in our present meta-analysis strictly met our selection criteria [19-27]. Figure 1 illustrates the search process, with the associated inclusion and exclusion criteria.

Figure 1

PRISMA flow diagram of study selection process.

3.2. Characteristics of Included Studies

The detailed characteristics of these 9 eligible studies are summarized in Table 1. In total, 983 patients were included, of which 504 were in an intervention group and 479 were in the control group; the age of most patients was over 50s. All included studies were published from 2007 to 2021. The sample size ranged from 10 to 613. These studies contained 6 retrospective cohort studies, two RCTs and 1 prospective cohort study.

Table 1

Characteristics of the eligible studies.

Study	Studydesign	Country	Treatment		No. of patients		Gender (M/F)		Age		Follow-up	Duration
			Intervention	Control	Intervention	Control	Intervention	Control	Intervention	Control
Kaderli 2007 [21]	Retrospective cohort study	Turkey	Triamcinolone intravitreal injection of 4 mg/0.1 mL	No treatment	4	6	2/2	3/3	56-74	56-74	>9 months	June 2004 to January 2005
Hayreh 2008 [20]	Retrospective cohort study	US	Prednisone 80 mg daily for 2 weeks, and then tapered down to 70 mg for 5 days, 60 mg for 5 days, and then cutting down by 5 mg every 5 days	No treatment	312	301	188/124	175/126	59.2 ± 12.6	62.0 ± 12.2	3 months	1973 to 2000
Rebolleda 2012	Retrospective cohort study	Spain	Prednisone 80 mg daily for 2 weeks, and then tapered down to 70 mg for 5 days, 60 mg for 5 days, and then cutting down by 5 mg every 5 days	No treatment	10	27	7/3	14/13	73.5 ± 11.6	68.0 ± 10.4	6 months	September 2008 to September 2009
Kinori 2014 [22]	Retrospective cohort study	Israel	Methylprednisolone (1 g/day) for 3 days, followed by oral prednisone (1 mg/kg) for 11 days	No treatment	24	24	14/10	16/8	54.4 ± 12.3	55.4 ± 9.6	6 months	—
Radoi 2014 [25]	Retrospective cohort study	France	Triamcinolone intravitreal injection of 4 mg/0.1 mL	No treatment	21	15	NR	NR	59-91	59-91	6 months	February 2009 to November 2012
Pakravan 2016 [23]	Randomized controlled trial	Iran	Methylprednisolone 500 mg twice a day for 3 days, followed by 2 weeks of oral prednisolone 1 mg/kg/day	Placebo	30	30	20/7	21/9	63.3 ± 4.7	64.1 ± 7.3	6 months	December 2010 to November 2013
Pakravan 2017 [24]	Prospective cohort study	Iran	Methylprednisolone 500 mg twice a day for 3 days followed with oral prednisolone 1 mg/kg for 10 days	Suggested treatments	43	30	32/11	21/9	63.2 ± 4.9	64.1 ± 7.3	6 months	May 2012 to October 2015
Saxena 2018 [27]	Randomized controlled trial	India	Prednisone 80 mg daily for 2 weeks, and then tapered down to 70 mg for 5 days, 60 mg for 5 days, and then cutting down by 5 mg every 5 days	Placebo	19	19	13/6	11/8	55.6 ± 4.2	58.0 ± 7.1	6 months	March 2015 to August 2016
Durbant 2021 [19]	Retrospective cohort study	France	Triamcinolone intravitreal injection of 4 mg/0.1 mL	No treatment	41	27	—	—	66	68	6 months	2009 to 2017

3.3. Results of Quality Assessment

After identifying the included articles, the abstract and full text of each article were carefully read and the publication's quality was evaluated according to Cochrane Collaboration's tools and NOS (Tables 2 and 3). For the 2 RCTs, there was no risk of bias. The risk of bias of the non-RCTs showed that all studies were rated over 6, which indicated no significant risk of bias.

Table 2

Risk of bias of randomized controlled trial studies.

Study	Random allocation	Hidden distribution	Blind method	Incomplete outcome data	Selective reporting of results	Other bias	Quality level
Saxena 2018 [27]	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	High
Pakravan 2016 [23]	Low risk	Low risk	Low risk	Low risk	Low risk	Low risk	High

Table 3

Risk of bias of cohort studies.

Study	Selection				Comparability of cohorts	Outcomes			Score
	Representativeness of cohort	Selection of nonexposed cohort	Ascertainment of exposure	Outcome lacking at the beginning		Outcome assessment	Sufficient follow-up time	Follow-up adequacy
Kaderli 2007 [21]	★	★	★	★	★★	☆	★	★	8
Radoi 2014 [25]	★	★	★	★	★★	★	★	★	9
Hayreh 2008 [20]	★	★	★	★	★☆	☆	★	★	7
Pakravan 2017 [24]	★	★	★	☆	★★	★	★	★	8
Rebolleda 2012	★	★	★	★	★☆	★	★	★	8
Kinori 2014 [22]	★	★	★	★	★☆	★	★	★	8
Durbant 2021 [19]	★	★	★	★	★★	★	★	★	9

3.4. Results of the Meta-Analysis for Outcomes

3.4.1. Visual Acuity

The pooled analysis indicated that, compared with the control group, the glucocorticoid group resulted in a significant improvement in the visual acuity (VA, calculated by the logarithm of the minimum angle of resolution, logMAR) with a mean difference (MD) of -0.25 (95% CI [-0.45,-0.05], P = 0.02; Figure 2). However, significant heterogeneity among the studies was detected (I2 = 89%, P < 0.00001).

Figure 2

Forest plot evaluating the outcomes of VA. VA: visual acuity.

3.4.2. Visual Field

A fixed effects model was used to evaluate the heterogeneity of visual field (VF, calculated by the mean deviation), as insignificant heterogeneity was found among the included studies (P = 0.46, I2 = 0%). The results showed that the glucocorticoid group had a better improvement in evaluation of visual field than the control group (MD = -0.50 with 95% CI [-0.94, -0.07], P = 0.02) (Figure 3).

Figure 3

Forest plot evaluating the outcomes of VF. VF: visual field.

3.4.3. Retinal Nerve Fiber Layer (RNFL)

For retinal nerve fiber layer (RNFL), 6 studies involving 303 patients reported it. Meta-analysis showed that compared to the control group, the glucocorticoid group had a higher decrease of RNFL (MD: -14.10, 95% CI [-26.41, -1.79], P = 0.02, fixed effects model), without significant heterogeneity (I2 = 0%, P = 0.61) (Figure 4).

Figure 4

Forest plot evaluating the outcomes of RNFL. RNFL: retinal nerve fiber layer.

3.4.4. Improvement Rate of VA

Change ≥3 lines in the Snellen VA chart was considered to be a significant change, which corresponded to a change in logMAR of at least 0.30 [20]. Some articles analyzed the improvement rate of VA (change ≥3 lines), we performed a pooled analysis on these studies. Overall, the pooled estimate showed that compared to the control group, the glucocorticoid group had a significantly higher improvement rate of VF (RR 1.31, 95% CI [1.12, 1.52]; P = 0.0005, fixed effects model), with insignificant heterogeneity among the included studies (I2 = 15%, P = 0.32) (Figure 5).

Figure 5

Forest plot evaluating the outcomes of improvement rate of VA. VA: visual acuity.

3.4.5. Adverse Reactions

Since there were not enough literature reports on the comparison of adverse reactions between the two groups, we could not make a combined analysis. We made a descriptive analysis about adverse reactions, and the specific results were shown in Table 4. Adverse reactions included gastrointestinal reactions, headache, weight gain, and anxiety. From the current results, there was no significant difference in adverse reactions between two groups.

Table 4

Summary of adverse reactions in the included studies.

Complication	Rate (%)		RR	95% CI
	Intervention	Control
Gastrointestinal symptoms	52.63	31.58	2.41	(0.64, 9.03)
Headache	26.32	21.05	1.34	(0.30, 6.02)
Weight gain	21.05	15.79	1.42	(0.27, 7.44)
Anxiety	10.53	0	5.57	(0.25, 124.19)

RR: risk ratio.

3.5. Publication Bias

The funnel plots for all four outcomes are shown in Figure 6. The shape showed some evidence of asymmetry, but the P value of the Egger test was not statistically significant (VA, P = 0.293; VF, P = 0.123; RNFL, P = 0.727; improvement rate of VA, P = 0.324). This indicated that there was no significant publication bias in our meta-analysis.

Figure 6

Funnel plot showing the publication bias. (a) visual acuity; (b) visual field; (c) retinal nerve fiber layer; and (d) improvement rate of visual acuity.

4. Discussion

Optic nerve edema is an important factor in the decline of visual function, and the timely application of glucocorticoids in a timely manner can help to eliminate edema, reduce the crowded state of optic disc, improve the blocking state of axial plasma flow, reduce the crowded pressure of capillaries at the optic nipple, improve the blood flow state, and improve the degree of nerve ischemia, so as to reduce some damage to visual function in a certain sense [14, 28, 29]. At the same time, glucocorticoids can also inhibit the expression of endothelin 21 [30, 31].

Our study systematically evaluated the therapeutic effect of glucocorticoids on NAION patients with VA, VF, RNFL, and adverse reactions as outcome indicators. A total of 9 studies were included in this paper. The results of meta-analysis showed that the application of glucocorticoids could improve VA (MD: -0.25, 95% CI [-0.45, -0.05], P = 0.02), VF (MD: -0.50, 95% CI [-0.94, -0.07], P = 0.02), and RNFL (MD: -14.10, 95% CI [-26.41, -1.79], P = 0.02) in NAION patients.

Chen et al. conducted a meta-analysis of the literature on the treatment of NAION with corticosteroids, and they found that corticosteroids did not significantly improve the vision of NAION patients [32]. However, our study strived to objectively analyze the effectiveness and safety of hormone treatment of NAION; therefore, some studies on hormone combined with other treatment methods and the literature that failed to clearly define arteritic or nonarteritic ION have been excluded to avoid the possible potential bias risk [33-35].

The adverse reactions of glucocorticoid drugs used in the treatment of NAION deserve clinicians' attention, such as monitoring blood glucose changes, nausea, indigestion, headaches, and weight gain, Anxiety and depression are easily overlooked, and intraocular hypertension is the most common adverse reactions [36, 37]. By summarizing the adverse reactions of included literatures and comparing the differences with the control group in a descriptive way, we found that the differences between the two groups were not significant.

This study had some limitations. First, among the included studies, 7 were nonrandomized studies and only 2 were RCTs. As the “gold standard” for clinical efficacy evaluation, RCTs have more objective guiding significance, which may reduce the reliability of the results of this study. Second, there were few reports of adverse reactions in the original studies included in the meta-analysis, so it was impossible to accurately evaluate the safety of glucocorticoids in NAION treatment.

5. Conclusions

Our meta-analysis found that the application of glucocorticoids in the treatment of NAION was safe and effective and can effectively improve VA and VF of NAION patients. However, the effect of glucocorticoids on the improvement of vision in NAION patients should be verified by more large sample RCTs and should focus on its hormone adverse reactions.