Ocular adverse events from pharmacological treatment in patients with multiple sclerosis—A systematic review of the literature

Purpose The aim of this study was to review the scientific evidence and describe the ocular treatment-emergent adverse events (TEAEs) related to pharmacological treatment in patients with multiple sclerosis. Methods A systematic review of literature was conducted according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis guidelines in the MEDLINE, LILACS, EMBASE, and COCHRANE databases. Articles were filtered based on title and abstract considering the selection criteria and subsequently filtered by full-text reading. The resulting articles were evaluated using the Joanna Briggs Institute Quality Tools. Study characteristics and results were extracted and presented in structured tables to conduct a narrative synthesis. Results A total of 2852 published articles were extracted using our strategy. After removing duplicates, 2841 articles were screened based on title and abstract, 102 articles were evaluated using quality tools, and 69 articles were filtered by full-text reading. Through this search strategy, 60 articles met all the inclusion criteria and seven articles, through a search update conducted in the same manner, were included. This resulted in 67 articles meeting the inclusion criteria, of which 11 were experimental and 56 were observational. The therapies related to ocular TEAEs were alemtuzumab, amantadine, fingolimod, steroids, CTLA-4 Ig, estriol, interferon β, natalizumab, hyperbaric oxygen, rituximab, siponimod, teriflunomide, and tovaxin. Fingolimod and siponimod were commonly associated with macular edema, interferon β was associated with retinopathy, alemtuzumab was associated with thyroid eye disease, amantadine was associated with corneal edema, and steroids were associated with acute retinal necrosis. Opportunistic infections were also found, and there was one life-threatening case. Conclusions Our search revealed different methodological assessments of the topic. However, longitudinal studies regarding ocular TEAEs related to multiple sclerosis therapy are necessary to provide evidence-based recommendations, especially in understudied regions such as Latin America and Africa. Physicians should monitor ocular symptoms in patients being treated for multiple sclerosis and consider an interdisciplinary approach. Systematic review registration PROSPERO ID CRD42020106886 Supplementary Information The online version contains supplementary material available at 10.1186/s13643-021-01782-7.


Background
Multiple sclerosis (MS) is a chronic autoimmune demyelinating disease with variable clinical, radiological, and pathological characteristics. Since its first description by Charcot, there have been significant efforts to clarify its pathophysiology and progression and determine the best strategies for appropriate management [1]. Currently, the therapy for MS comprises corticosteroids, immunosuppressants, and immunomodulators such as interferon (INF) and monoclonal antibodies. Although several cases of patients with ocular TEAEs have been reported over the years, there is a lack of clear understanding of the TEAEs that are strictly related to the disease and linked only to therapy [2].
Ocular manifestations in patients with MS are not rare and have been described as a major aspect of the disease. Any structure related to visual pathways could be involved in MS, and up to 20% of patients would display optic neuritis as an initial clinical manifestation. Visual field and color vision defects, relative afferent pupillary defect, and other ophthalmological findings could also be present [3].
According to the World Health Organization (WHO), the definition of drug-related adverse events corresponds to an involuntary harmful response to a medication used in a normal dose for a prophylactic, diagnostic, or therapeutic purpose [4,5]. However, it is necessary to consider that this harmful response, irrespective of how minor it is, must be explained to the patient before initiating any medication to predict, prevent, or treat any possible future event and avoid poor adherence to treatment [5]. Some of the TEAEs that have been described as related to therapy in MS are cataract, glaucoma, chorioretinopathy, visual acuity reduction, macular edema, retinopathy, among others [2].
The aim of this study was to review the scientific evidence and describe the ocular TEAEs related to pharmacological treatment in patients with MS to provide a theoretical basis and quality literature for ophthalmology specialists, neurologists, and general practitioners, which could help them in making appropriate clinical decisions.
Methods This review was written according to the Preferred Reporting Items for Systematic Reviews and Meta-analysis (PRISMA) extension statement for reporting systematic reviews. The protocol registration can be found under the PROSPERO ID CRD42020106886.

Study design
A systematic literature review focusing on ocular TEAEs in MS therapy was conducted using the MEDLINE, LI-LACS, EMBASE, and COCHRANE databases for articles published till November 2018. The search was updated on May 4, 2020. MEDLINE and COCHRANE databases were searched using Medical Subject Headings terms, EMBASE was searched using emtree terms, and LILACS was searched using Descriptores en Ciencias de la Salud terms, using boolean operators, as evidenced in Annex 1. We limited the search only to human studies, although no limits regarding language and period of publication were used.

Study selection
An initial search was conducted by JMO, JRG, LMM, and JARH, based on which a review was prepared to eliminate duplicates. Eligible studies were selected by screening the title and abstract by at least two reviewers independently, and discrepancies were resolved by an expert evaluator (ADLT). A study was included if (a) the abstract was available, (b) it contained original data, (c) the diagnosis of MS was made by a specialist, (d) the individual was being treated for MS, and (e) the ocular manifestations appeared after the initiation of MS treatment. Experimental and observational studies were also included. Articles were excluded from the analysis if they did not provide information regarding ocular TEAEs in individuals undergoing MS therapy.
The full texts of the selected articles were retrieved and classified according to the type of article/study by at least two reviewers independently, and discrepancies were resolved by an expert evaluator (MGS). They were then evaluated using the Joanna Briggs Institute Quality Tools (JBIQT). If an article was not available, contact with the author was made. The cascade of article selection was managed using the PRISMA flowchart [6]. The following information was extracted from the selected articles: study period, location, number of patients, type of medication, and ocular TEAEs. A dataset was constructed using the information mentioned ahead.

Data synthesis and presentation of results
Microsoft Excel (Microsoft Corp., Redmond, WA, USA) was used as a synthesis tool to organize the results of the search strategy and the information of the articles. The first sheet contained information about the initial search results without duplicates. The next six sheets were asigned to three pairs of authors (LMM vs JARH, WRC vs JRG, and JMO vs LOTG) to filter titles and abstracts in a peered manner. A color code was used to select the articles as follows: green to include, yellow to revise, and red to exclude. The next three sheets were used to evaluate concordance for each pair of authors. The subsequent sheet was used by the expert reviewer (ADLT) to resolve discrepancies from the peered review. An additional sheet was used to classify the methodological design of the studies and record JBIQT grades and the reasons for exclusion. The next sheet was used to record articles included after full-text reading. The final sheet included information regarding ID number, authors, year of application of the study, year of publication, journal, location, title, aim, sample size, population (gender and group characteristics), methodological design, MS type, therapy, time of presentation of TEAEs, symptoms after therapy suspension, statistical analysis, important details from results, important details from discussion, and observations. The same synthesis tool was used for the search strategy update. The extracted data were synthetized in tables and figures to present them in an organized manner throughout the article.

Risk of bias
We retrieved and classified the articles based on study design to evaluate their methodological quality using their respective JBIQT [7]. Each quality tool has a set of questions that are evaluated as yes, no, unclear, or not applicable domains. An expert in epidemiology (MGS) established a minimum cut-off point as the number of items on the tool marked as "Yes," including the major and minor criteria. Case reports with 5/8 items on the checklist were included; for case series, the minimum cut-off point was 6/10 items; for cross-sectional studies, it was 6/8; for randomized controlled trials (RCTs), it was 10/13; and for nonrandomized experimental studies, it was 6/9 (see Annex 2 for supplementary information).

General description
We extracted 2852 published articles (1756 from MED-LINE, 1093 from EMBASE, three from COCHRANE, and none from LILACS). After removing duplicates, 2841 articles were screened based on title and abstract, and 102 articles were evaluated using JBIQT. After quality evaluation, 33 articles were excluded, and 69 full-text articles were assessed for eligibility. Sixty articles met all the inclusion criteria. Figure 1 shows further detailed information regarding the inclusion of articles.
The search was updated on May 04, 2020, extracting 202 published articles (117 from MEDLINE, 85 from EMBASE, none from COCHRANE, and none from LI-LACS). After removing duplicates, 196 articles were screened based on title and abstract. A total of 11 articles were evaluated using JBIQT. After quality evaluation, two articles were discarded, and nine full-text articles were assessed for eligibility. Finally, after discarding two articles, as no ocular TEAEs were reported, seven articles met all the inclusion criteria ( Fig. 1).

Characteristics
Our final inclusion strategy yielded 67 published articles, including 20 in the United States of America (USA), 24 in Europe (Germany, Spain, France, Italy, Wales, England, Switzerland, Croatia, and Sweden), three in Canada, three in Australia, nine in Asia (Japan, Israel, Iran, and Turkey), and eight were multicentric across continents.
All studies were published between 1987 and 2020. Most of the studies included only adults, but one retrospective cross-sectional study included adults and children [8]. Of the 67 studies, 38 were case reports, four were case series, 14 were cross-sectional studies, and 11 were experimental studies. Table 1 summarizes the characteristics of the studies included in the present systematic review.
After JBIQT qualification, the average number of items marked "Yes" across the studies was 11/13 (range 10-13) for the RCTs, and one study was a nonrandomized controlled trial with 7/9 of the items marked "Yes." Six studies had a sample size ≤400 patients.
Adverse events related to each therapy TEAEs related to alemtuzumab Five studies from Germany, USA, and the United Kingdom (UK) referred to TEAEs when using alemtuzumab,     [53,59,61]. One of the studies was a case series extracted from two clinical trials (CARE-MS1 and CARE-MS2) where five patients developed Graves' disease (GD), but only two patients had TED [59]. According to Trinh et al [53], there was a female patient who developed bilateral ocular hyperemia, photophobia, proptosis, and upper eyelid retraction, consistent with TED, after three years of treatment [53]. A similar study from the UK reported 10 patients of 162 with GD, but only six patients presented TED, with a period of 26 months, 20 months, four years, 10 months, four years, and 48 months after the last infusion of alemtuzumab, respectively [61].
In a cross-sectional study conducted by Tuohy et al, thyroid autoimmunity developed in 35/87 patients (40.2%) of whom 22/35 (63%) had hyperthyroidism. Although the study reported GD, it was not clear whether the patients had any ophthalmological manifestation. Two patients in that study developed Herpes zoster ophthalmicus [70]. In another study, the use of alemtuzumab was related to thyroid gland autoimmunity (35%), but no TED was reported, and 1% of the patients presented conjunctivitis; however, the study did not specify the etiology of conjunctival inflammation [72].

TEAEs related to amantadine
We found two case reports from the USA that described four cases of bilateral corneal edema as a TEAE related to amantadine used for managing tremor and fatigue in patients with MS [30,35]. The duration of amantadine treatment before the onset of symptoms in all cases ranged from two months to six years, and the range of corneal thickening was 677 μm to >1000 μm [30,35]. In two cases, the corneal edema resolved approximately one month after amantadine was suspended [30,35]; in one case, the corneal edema resolved two months after amantadine suspension [35]; and in one case, the corneal edema required management with bilateral penetrating keratoplasty [35].

TEAEs related to fingolimod
A total of 25 studies reported ocular TEAEs in patients treated with fingolimod. Of these studies, 18 were case reports, mostly from the USA, Australia, Spain, and Germany; two were experimental studies, both multicentric; and five were cross-sectional studies, one from Canada, one from Italy, one from the USA, and two were multicentric.

TEAEs related to steroids
We found two case reports that described acute retinal necrosis as a TEAE after high-dose steroid treatment for relapsing-remitting multiple sclerosis (RRMS), one from Turkey and the other from the USA [45,49]. In the first report, the patient received an IV steroid pulse of 1000 mg/day for three days and was continued with 10 mg of daily prednisolone. After two months, the visual symptoms started and the diagnosis of acute retinal necrosis caused by the Varicella-zoster virus was made [45]. In the other case report, the patient received three courses of high-dose methylprednisolone (1000 mg/day for three days) for recurrent relapses five months before the presentation of visual symptoms. The diagnosis of Herpes simplex virus type 2 infection was made [49].

TEAEs related to CTLA-4 Ig
This immunoglobulin was used in a Phase 1 clinical trial conducted on 20 patients with RRMS from the USA, which evaluated its safety and tolerability in different doses. This study reported, a visual field defect in one patient receiving a 2 mg/kg dose of the drug within the first 24 h of infusion. During the long-term follow-up, 10% of the enrolled patients had blurred vision [17].

TEAEs related to estriol
A Phase 2 clinical trial conducted on patients with RRMS from 16 academic neurology clinics in the USA, evaluating the effect of estriol treatment in the reduction of MS relapses in women, reported TEAEs within 24 months of treatment. There were six events of visual defects (blurred vision or diplopia) reported in four patients receiving estriol compared to seven of the same visual defects reported in seven placebo patients [16].
One study was a case series that reported three patients with neuroretinitis after INF β therapy, of whom two had been undergoing INF therapy for around a year before the symptoms appeared, the other patient presented with a macular star exudate on funduscopic examination concurrently with the initiation of treatment [60].
There were five cross-sectional studies (one each from Japan, the USA, Canada, Croatia, and Italy). Those with previous optic neuritis (n = 3 patients, 3 eyes) exhibited visual evoked potential impairment with a delay in P100 latency before treatment. Moreover, some patients (n = 2 patients, 3 eyes) without previous impairment exhibited increased P100 latency after INF therapy. It is not clear whether this impairment is due to INF therapy or MS itself [65]. A multicentric pharmacovigilance study conducted to analyze 10 drugs used for MS treatment over a period of 24 months recorded 411 adverse reactions. Approximately 42% of the TEAEs were due to INF, from which approximately 9% were unexpected, and 5.8% were severe events. Two patients presented ophthalmological manifestations. The first patient, associated with peginterferon-β-1a, with previous use of dimethyl fumarate, presented mild-moderate eyelid edema. The second patient, associated with glatiramer acetate, with previous use of INFβ-1a, presented moderate-severe visual field defect, which resolved within a day [74].
Finally, we included two experimental studies. The first one was a double-blind RCT [77].

TEAEs related to natalizumab
We found four articles regarding natalizumab [13,25,56,62], including two case reports (one from the USA and the other from Switzerland), one cross-sectional study (from Sweden), and one RCT (from Japan).
Of the two case reports, one [25] reported a case of progressive multifocal leukoencephalopathy (PML) associated with John Cunningham virus (JVC) that started with a subacute onset of bilateral blindness after the 44th dose of natalizumab. The second study reported the first published case of ocular toxoplasmosis reactivation during natalizumab treatment [56].
A web-based MS registry cross-sectional Swedish national postmarketing surveillance study reported 19 patients with TEAEs related to ocular inflammation, herpes simplex, herpes zoster, urinary tract infection, enterovirus meningitis, increased infection susceptibility, and hepatitis C infection. However, the study did not clarify how many of the 19 patients presented the eye compromise. Furthermore, the study reported three cases of PML that occurred during these trial periods, including two cases in the open-label extension phase of the SENTINEL trial and one case in a clinical trial for Crohn's disease [78,79]. There was no information regarding the visual compromise in the follow-up of patients who developed PML [62].
The RCT reported a patient with retinal detachment 75 d after the first dose of natalizumab [13].

TEAEs related to hyperbaric oxygen
Of the 62 included articles, only one case report published in 1986 described TEAE due to hyperbaric oxygen in a patient with MS who presented slow-recovering central bilateral scotoma [63].

TEAEs related to rituximab
One case of ocular inflammation as a TEAE was found after the initiation of treatment, which was reported at the German Registry of autoimmune diseases (GRAID) in a multicentric retrospective study. The time between treatment onset and the TEAE was not determined [68].

TEAEs related to siponimod
We found two studies regarding siponimod. One was a multicentric Phase 2 clinical trial conducted on patients with RRMS from 73 specialized MS centers in Canada, the USA, Russia, and nine European countries, wherein the patients were divided into cohort 1 (n = 188) and cohort 2 (n = 109). It reported that one patient with a history of uveitis presented macular edema with the highest dose of the drug (10 mg), and one patient presented optic neuritis with 0.5-mg dose of the drug [14]. The second study was a Phase 3 clinical trial conducted on patients with secondary progressive multiple sclerosis (SPMS) from 292 hospitals, clinics, and specialized MS centers in 31 countries. It reported that 2% (n = 18) of patients receiving siponimod presented macular edema compared to <1% (n = 1) of those receiving placebo [12].

TEAEs related to teriflunomide
In an experimental study conducted on patients with RRMS from different countries of America, Europe, and Africa, 324 patients were divided into the following three treatment groups: INFβ-1a (n = 104), teriflunomide 7 mg (n = 109), and teriflunomide 14 mg (n = 111). The second group was treated for 66.6 weeks, and the third group was treated for 64.2 weeks. The ocular TEAEs reported were optic neuritis, macular edema, and ischemic optic neuropathy in the second group of patients who received 7 mg of teriflunomide. No ocular TEAEs were observed in groups 1 and 3 [15].

TEAEs related to tovaxin
We found an experimental study that evaluated the safety of tovaxin for 24 weeks in patients with RRMS from 30 to 40 different sites in the USA. Neither deaths nor discontinuations due to safety reasons were reported in that study. Diplopia was the only reported ocular event, and it was considered as a major reaction because of the necessity of in-hospital attention of the patient, but not because of the event itself. The authors did not consider this reaction as secondary to the drug and evaluated tovaxin as a safe treatment for the course of 24 weeks [11].

Methodological assessment of TEAEs
In 2016, PRISMA published an article about harmreporting in systematic reviews, where terms such as adverse drug reaction, adverse effect, adverse event, complication, harm, side effect, and toxicity were discussed [80]. However, there exists a large heterogeneity when using these terms and reporting TEAEs. This represents a difficulty at the time of reading and interpreting articles about drug safety. Moreover, the diversity between the different analytic and data collection strategies, the different forms of presentation of TEAEs, and the variety among the study designs represented additional challenges in this systematic review.
Two available strategies should be used to record information about TEAEs, i.e., active monitoring, when an event is known or suspected to be associated with an intervention, and spontaneous monitoring for new or unexpected events [81]. Different methodological designs were included in this systematic review because active monitoring generally evaluates long-term TEAEs in RCTs and short-term TEAEs in large cohort studies or case-control studies [81], whereas spontaneous monitoring commonly describes TEAEs in case series and case reports. Each study included in this systematic review had to be carefully evaluated, because spontaneous reports may lead to underreporting and false positives, RCTs may not identify new events, observational studies may require a large sample size, and case reports may not be well documented and have a long-time elapse until publication [82].

Geographical distribution
Developed countries have a higher incidence of MS. As the populations in these countries have better access to healthcare and early treatment strategies, it is expected that most of the studies that evaluated adverse ophthalmological events -and adverse events in general -had been conducted in these countries. This is consistent with our results, which indicated that studies were largely conducted in Europe, North America, and Japan. A few case reports were described in Australia and Israel. Some multicentric studies included a couple of Asian countries apart from Japan as well. There were no reports of adverse events from other countries or regions, which may result in a poor understanding of the entire picture worldwide.

Pregnancy and hormones in MS
MS relapses are decreased during pregnancy (primarily during the third trimester) and influenced by sex hormones such as estriol [16]. One RCT demonstrated that estriol was related to visual defects (blurred or double vision); nonetheless, the same visual defects were reported in the placebo group. In clinically isolated syndromes, the incidence of some of the typical features such as visual acuity and color vision was reduced [83]; therefore, the findings might not be directly related to estriolassociated TEAEs. These manifestations have also been described in acute optic neuritis, in which patients described blurred vision [84]. Furthermore, optic neuritis was related to different MS treatments such as INFβ-1b, siponimod, and teriflunomide [14,15,85]. However, the resolution of TEAEs after medication withdrawal was not described.

Follow-up recommendations
Optic neuritis is one of the primary manifestations of MS; therefore, ophthalmological follow-up is mandatory in these patients. Complete ophthalmological examination with attention to the efferent visual system and ancillary studies such as VEP and optic nerve optical coherence tomography (OCT) should be available for these patients [86]. Based on the present study, fingolimod-related TEAEs were highly common. We suggest the following American Academy of Ophthalmology recommendations for patients with MS being treated with fingolimod published in October 2011: 1) A screening evaluation for pre-existing uveitis or macular or retinal vascular disease before starting, or within the first few weeks of starting fingolimod. 2) A single re-evaluation at 3-4 months of therapy. 3) Patients should be advised that the incidence of macular edema is low (~2/1000), but if there is a history of uveitis, the incidence may be as high as 20%. 4) Visual acuity check and a complete eye exam, including a dilated fundus exam, is a proper ophthalmic screening protocol. 5) Patients with abnormalities on the exam or unexplained decreased visual acuity might benefit from diagnostic imaging with macular OCT [87]. These recommendations are supported by the methodology used in the study conducted by Zarbin et al. in 2013, who reported differences in different times of evaluation of these patients [73].
Regarding the use of alemtuzumab in patients with MS and thyroid disease, it was reported that up to 22% of patients treated with alemtuzumab would develop GD [53]. In almost 90% of the cases, GD and Graves ophthalmopathy (GO) co-occur; nevertheless, GO can develop after up to 1 year from being diagnosed with GD [88]. Therefore, we suggest an annual check-up by ophthalmologists, neurologists, and endocrinologists. This follow-up is important to ensure that patients undergo timely ophthalmological treatment while simultaneously ensuring the continued success of alemtuzumab treatment in the management of MS [53,59,70].
In contrast, one of the studies on patients treated with alemtuzumab reported Herpes zoster ophthalmicus infection [70]. Previous trials have identified Herpes simplex virus and Varicella-zoster virus systemic opportunistic infections are among the most common infections in patients treated with alemtuzumab and recommended preventive measures, including antiherpetic prophylaxis [89].
Studies evaluating the presence of corneal edema secondary to the use of amantadine in patients with MS corresponded to case reports. Considering that corneal findings can occur from two months to six years after medication initiation, we recommend that the neurologist treating the patient be aware of the warning signs to immediately consult the ophthalmology specialist and prevent irreversible endothelial failure [30,35].
Another study evaluated the presence of dry eye after the use of anticholinergics, based on which we recommend that the neurologist treating patients with MS must be aware of this TEAE. Although the symptoms should not be the reason to stop an efficient treatment, it is recommended to conduct an ophthalmological evaluation to treat dry eye [71].
The study that evaluated CTLA-4 Ig treatment in patients with MS demonstrated blurred vision in 10% of the patients. However, the cause of blurred vision was not clear, and the spectrum of MS manifestations included blurred vision secondary to optic neuritis and uveitis. This is why further studies are required to evaluate the safety of this drug in patients with MS, assess whether the ocular compromise corresponds to a TEAE, and generate recommendations on this issue [17].
The clinical trial comparing estriol combined with glatiramer acetate and placebo for women with RRMS reported six blurry and double vision events after six months of follow-up. However, the study did not report the etiology of the visual problems, and hence it is not possible to know whether it corresponds to the natural course of the disease. Further studies are necessary to evaluate the safety of this medication [16].
Two case reports described acute retinal necrosis caused by HSV-2 and Varicella-zoster virus in patients undergoing high-dose steroid therapy [45,49]. Therefore, we recommend a retina specialist evaluation for patients showing ocular symptoms such as floaters, blurred vision, and sudden vision loss after treatment with steroid pulses.
Patients undergoing INFβ-1a therapy showed multiple ocular TEAEs in a duration range of three months to 10 years (Table 1). Given the wide timeframe for the establishment of TEAEs, and that some of the pathologies may correspond to manifestations or associations of MS, we consider that the most appropriate recommendation for these patients is the evaluation by an ophthalmologist before starting treatment and at least once a year after the start of treatment.
We consider that there is insufficient information to provide screening recommendations regarding natalizumab, hyperbaric oxygen, rituximab, siponimod, teriflunomide, and tovaxin. Nevertheless, in the presence of any ocular symptoms, we strongly recommend an immediate evaluation by an ophthalmology specialist. Table 2 shows the referral recommendations for each TEAE reported in this systematic review, based on our clinical experience. However, it is important to consider that each patient's case should be individualized, and disease guidelines should be followed.

Conclusions
Considering that ophthalmological TEAEs are generally undermined in clinical trials, different methodological designs were included in this systematic review. In light of this, it is necessary to conduct longitudinal studies to provide evidence-based recommendations.
Despite the diverse geographical distribution of the studies, there were scarce data from Latin America and African countries. Therefore, we strongly encourage healthcare providers to conduct pilot studies in these regions.
We suggest that physicians perform a comprehensive interrogatory and evaluation for ocular signs and symptoms during the follow-up of patients receiving treatment for MS and refer when necessary. An interdisciplinary approach might be considered to evaluate the patient's requirements.