Tacrolimus

Efficacy of medical treatments for vernal keratoconjunctivitis: A systematic review and meta-analysis

Ines Roumeau, MD, a Adrien Coutu, MD,a Valentin Navel, MD,a,b Bruno Pereira, PhD,c Julien S. Baker, PhD,d Fred eric Chiambaretta, MD, PhD, a,b Dominique Bremond-Gignac, MD, PhD,e,f and Fred eric Dutheil, MD, PhD g,h,i

Abstract

Background: Vernal keratoconjunctivitis (VKC) is a severe type of allergic conjunctivitis for which treatment strategies are still under debate.
Objectives: This study sought to conduct a systematic review and meta-analysis to evaluate the efficacy of medical treatments for VKC.
Methods: The PubMed, Cochrane Library, Embase, and
ScienceDirect databases were searched to assess the efficacy of treatments for VKC. Random-effect meta-analyses on changes in clinical scores of symptoms and signs between baseline and after treatment, stratified on treatment classes, were computed. Meta-regressions were searched for potential influencing parameters.
Results: Included were 45 studies (27 randomized controlled trials and 18 prospective cohort studies), 1749 patients (78% were men; mean age, 11.2 years), and 12 different treatment classes. Mast cell stabilizers (MCSs; usually considered as firstline therapy), cyclosporine, and tacrolimus were the most studied drugs (in three-quarters of studies). Overall, all clinical scores improved. Total symptom and sign score decreased for MCSs (effect size, 23.19; 95% CI, –4.26 to 22.13), cyclosporine (effect size, 22.06; 95% CI, 22.72 to 21.40), and tacrolimus (effect size, 22.39; 95% CI, 23.36 to 21.43). No significant differences were shown depending on treatment classes, From athe Ophthalmology Department, Centre Hospitalier Universitaire Clermont-Ferrand, hthe Preventive and Occupational Medicine Department, WittyFit, Centre Hospitalier Universitaire Clermont-Ferrand, University Hospital of Clermont-Ferrand, Witty Fit, cthe Biostatistics Unit, Clinical Research and Innovation Direction, University Hospital of Clermont-Ferrand, bthe Translational Approach to Epithelial Injury and Repair Department, Centre National de la Recherche Scientifique, Institut National de la Sante et de la Recherche Medicale, Genetique Reproduction et Developpement, and gthe Physiological and Psychosocial Stress Department, Laboratoire de Psychologie Sociale et Cognitive, Centre National de la Recherche Scientifique, Universite Clermont Auvergne, Clermont-Ferrand; dthe Hong Kong Baptist University, Centre for Health and Exercise Science Research, Physical Education and Health, Kowlon Tong; e the Ophthalmology Department, University Hospital Necker Enfants Malades, and fthe Sorbonne Paris Cite University, Institut National de la Sante et de la Recherche Medicale, Unite Mixte de Recherche S1138, Team 17, From Physiopathology of Ocular Diseases to Clinical Development, Cordeliers Search Centre, Paris; and ithe Faculty of Health, School of Exercise Science, Australian Catholic University, Melbourne. concentration, age, sex, baseline activity scores, and atopy.
Sensitivity analyses demonstrated similar results. Conclusions: This study confirms the efficacy of MCSs in the treatment of VKC. Efficacy of cyclosporine and tacrolimus did not differ, suggesting that tacrolimus is a good alternative to cyclosporine for severe cases of VKC. Further studies are needed to compare other drugs and their precise place in treatment strategy. (J Allergy Clin Immunol 2021;nnn:nnn-nnn.)

Key words: Chronic disease, immunomodulators, topical drugs, atopy, inflammation, pediatrics

Introduction

Vernal keratoconjunctivitis (VKC) is a severe type of allergic conjunctivitis affecting mostly male children, with atopic condition.1 The prevalence of VKC strongly depends on the geographical origin, varying from 3.2 per 10,000 inhabitants in Western Europe2 to 500 per 10,000 inhabitants in Central Africa.3 Three forms are described: limbal VKC, tarsal VKC, and mixed VKC.1 In most cases, this eye condition is characterized by perennial ocular symptoms and signs with seasonal exacerbations.4 The physiopathology of VKC involves both IgEmediated allergic factors as well as nonallergic ones.4 The treatment of this pathology is a critical issue due to its wellknown impact on quality of life as well as its sightthreatening complications (shield ulcers, keratoconus, limbal stem cell deficiency, irregular astigmatism).5 Despite some proposals,6,7 therapeutic strategies are under debate. Nondrug therapies are the basis of the treatment (eye washing, solar protection, eviction of allergens with proven sensitization). Medical treatment is often necessary with artificial tears, mast cell stabilizing eye drops, and/or antihistaminic drugs as first-line therapies. In case of crisis or noncontrolled VKC, corticosteroids eye drops are usually added. Their effect is powerful, but they have potential serious side effects (glaucoma, cataract), leading to the use of cyclosporine (a calcineurin inhibitor) in case of cortico-dependence or resistance.4 The place of other immunomodulator drugs, such as tacrolimus (another calcineurin inhibitor) remains to be clarified.8 Numerous clinical scores attesting the VKC severity are available but none is consensual.9 In daily clinical practice, scores assessing the most frequent symptoms (itching, photophobia, discharge, tearing, foreign body sensation) and signs (hyperemia, keratitis, tarsal papillae, limbal inflammation) are simple and easy to apply.
Therefore, the objectives of this study were to assess the efficacy of all available medical treatments on symptoms and signs of VKC and to compare conventional drugs with new therapies.

METHODS

Literature search

We reviewed all interventional prospective studies measuring the efficacy of any treatment for active VKC. The PubMed, Cochrane Library, ScienceDirect, and Embase databases were searched in November 2020 with the following key words: ‘‘vernal keratoconjunctivitis’’ OR VKC AND therapy OR treatment OR therapeutic* (see Appendix E1 in this article’s Online Repository at www.jacionline.org). The search was not limited in time. Articles needed to be written in English, French, Spanish, or Portuguese. The studies had to report our primary outcome, an activity score (symptoms and/or signs) of VKC before and after any medical treatment. Studies needed to be primary research. Studies assessing severe allergic conjunctivitis without specification orpoolingpatientswithVKCandthosewithatopickeratoconjunctivitis,retrospective studies, and animal studies were excluded. We also excluded articles with only patients in nonactive phase of VKC. We imposed no limitation on the sample size and the regional origin. In addition, reference lists of all publications meeting the inclusion criteria were manually searched to identify any further studiesthat were not found with the key wordssearch. Thesearch strategy is presented in Fig 1. One author conducted all literature searches and collated the abstracts. Two authors separately reviewed the abstracts and, based on the selection criteria, decided the suitability of the articles for inclusion. Another author was asked to review the articles where consensus on suitability was debated. All authors then reviewed the eligible articles. We followed the PRISMA (Preferred Reporting Items for Systematic Reviews and Meta-Analyses) guidelines (see Appendix E2 in this article’s Online Repository at www.jacionline.org).

Data collection

The data collected included first author’s name, publication year, country, study design, aims, outcomes of included articles, sample size, number of eyes, age,percentageofmales,drugs(type,galenic,concentration,posology),activity scores of VKC (symptoms, signs, and total), measurement time (baseline and at re-evaluation), presentation of VKC (onset, severity, refractory characteristics, forms— limbic, tarsal, mixed), previous treatment with or without washout period, and other putative explaining variables (atopy, ethnicity).

Quality of assessment

We usedthe SIGN (ScottishIntercollegiate GuidelinesNetwork)checklists to assess the methodological quality of randomized clinical trials (10 items) and cohort studies (14 items).10 Items assessed the main causes of bias. We calculated an overall quality score by quoting each item (yes 5 1; no, cannot say, or not applicable 5 0).

Statistical considerations

Statistical analysis was conducted using Stata software (version 15; StataCorp, College Station, Tex).11-13 Baseline characteristics were reported for each study sample as number (percent) and as mean 6 SD for categorical and continuous variables, respectively. We conducted meta-analyses on effects of treatment at the most frequent time of re-evaluation for each study compared to baseline measures—or the closest time to the most frequent time. We described our results by calculating the effect size (ES; standardized mean differences). A positive ES denoted improved efficacy. A scale for ES has been suggested with 0.8 reflecting a large effect, 0.5 a moderate effect, and 0.2 a small effect.14 We conducted random-effect meta-analyses for each symptom (itching, photophobia, discharge, tearing, and foreign body sensation) and for each sign of VKC (hyperemia, keratitis, tarsal papillae, and limbal inflammation). When not available, we proportionally calculated a total subjective symptom score (TSSS), a total ocular sign score (TOSS), and a total symptom and sign score combining both scores—by computing the average of all individual scores transformed on a 0 to 100 scale. We conducted sensitivity analyses considering time of assessment, as well as on studies reporting the absence of concomitant treatments. We also computed meta-analyses on the effect of treatment versus a placebo in studies reporting the use of a placebo in the control group at re-evaluation time. For all aforementioned meta-analyses, we computed stratification on drugs and type of drugs (treatment class). When possible, meta-regressions were proposed to compare the benefits between drugs and type of drugs and to assess the influence of putative clinically relevant parameters such as concentration for the mostfrequentdrugs, ageof patients,sex, baselineactivityscores, or atopy. Results were expressed as regression coefficients and 95% CIs.

RESULTS

An initial search produced 1131 possibly corresponding articles (Fig 1). Removal of duplicates and the use of the selection criteria reduced the number of articles reporting the evaluation of treatment of VKC to 45 articles from 31 different journals,15-59 published between 197224 and 2019.9 All articles were written in English except 1 in French23 and 1 in Portuguese.29

Study designs

Twenty-seven studies were randomized controlled studies (14 placebo-controlled15,19,20,24-27,31,33,35,41,48,50,51 and 13 with another active treatment as comparator17,28,36,37,39,40,43,44,46,49,55,56,59), 2 were controlled studies,29,52 and 16 were single-group prospective cohort studies.16,18,21-23,30,32,34,38,42,45,47,53,54,57,58 All controlled studies were double masked except 4 that were single masked,17,29,46,55 and 3 that were not masked.33,40,52 Forty-two studies were monocentric,15-26,28-40,42-55,57-59 and 3 were multicentric.27,41,56 The majority of studies was carried out in the Middle East (n 5 17)15-19,21,25,28,32-35,42,46,47,54,59 and Europe (n 5 16),20,22-24,30,38-40,45,48-50,52,53,55,56 followed by Asia (n 5 6),31,36,37,51,57,58 South America (n 5 3)29,43,44 and North America (n5 2).26,27 Onestudywasheldin11countriesthroughoutEurope, Asia, Middle East, and North America.41

Quality of articles

Results of SIGN cohort studies criteria ranged from 66.6%32,34,38 to 100%45 for yes responses, with a mean score of 82.5 6 10.5%. Results of SIGN randomized controlled trials criteria ranged from 33.3%40 to 100%25,37,44,48,51 for yes responses, with a mean score of 80.2 6 15.5 (Fig 2, and see Appendix E3 in this article’s Online Repository at www. jacionline.org).

Inclusion criteria

Inclusion criteria were VKC without precision,20-22,25,28,29,42,47,50,51 18,24,27,30,36,37,39,40,44,46,55,56 active VKC, active severe VKC or refractory to medical treatment15,16, 18,19,23,25,31-35,41,43,45,48,49,52-54,57-59 48 (especially first-line drugs, resistance or dependency to corticoids,15,19,23,34,43,58,59 and resistance to cyclosporine16,49,54,57). Only 4 studies had age criteria with age required over 3,45 4,41,55 or 6 years old59 and under 16,45 17,55 or 18 years old.41,59 One study required the association of VKC and asthma.38

Exclusion criteria

The main exclusion criteria were related to concomitant or previous specific ocular disease such as uveitis,16,19,36,37,45,58 glaucoma,19,36,37,45 local infection (acute conjunctivitis, history of herpes),16,19,25,30,36,37,40,41,43,44,46,58,59 corneal diseases,16,19, 22,30,31,36,37,45,49,58 45,46 16,46 blepharitis, conjunctival pathologies, history of dry eye syndrome,46,51 cataract,19,45 or nonspecific ocular disease.15,20,21,25,27-29,31,36,37,40,41,45,55,59 Some studies also excluded patients with previous ocular surgery,16,21,30,40,41,45,46,58 concomitant systemic disease,15,28,31,36,37,40,45,49,55,59 or those using systemic corticosteroids, nonsteroidal anti-inflammatory drugs, antihistamines, immunosuppressive therapy medications,15,19,20,27,35,38,41,44,49,58,59 or any systemic therapy.30,31,45,46,55 Some studies excluded patients using any ophthalmic therapy for other pathologies29,41,45,46,55 or previous ocular therapy for VKC.21 Other criteria were cited, such as contact lens use,16,20,25,27,29,30,45,46,55,56 pregnancy, possible risk of pregnancy, or breastfeeding.21,28,29,46,55,56,59 Fifteen studies had no exclusion criteria.17,18,23,24,26,32,33,39,42,47,48,50,52-54,57

Population

Sample size. A total of 1749 patients diagnosed with VKC representing3496eyeswereincluded:1440patientsreceivedactive treatmentinbotheyes,107weretreatedwithactivetreatmentinone eye and placebo treatment in the other, 202 were treated with only a placebo in both eyes. Population sizes ranged from 7 patients30 to 197 patients53 with a mean of 39 6 40 patients per study.
Sex. The mean proportion of men was 78% (95% CI, 76% to 80%), varying from 30%21 to 95.8%,37 one study did not specify the proportion of men.50
Age. The mean age of VKC patients was 11.2 years (95% CI, 10.2 to 12.1 years), ranging from 3 years25,45 to 38 years.15,34,51 The minimal mean age in the included studies was 7.1 6 1.6 years36 and the maximal mean age was 26.6 6 16 years.21 Three studies didnot specify the mean age.52,53,58 All patients were children in 19 studies,17,23,29,30,32,35-37,41,42,44,45,47-49,52,53,55,59 whereas 22 studies evaluated both adults and children.15,16, 18-22,24-28,31,33,34,43,46,50,51,54,56,58
Four studies did not report the age range of patients and therefore whether some adults were included or not.38-40,57
Clinical form. Twenty-one studies22,23,26,27,29,33,36,37,41-49, 51,53,54,58 specified the proportion of limbal, tarsal, and mixed form of VKC. In these studies, VKC’s tarsal form was the most common (56.1%), followed by limbal (24.3%) and mixed form (19.6%).
Duration of disease. Twenty-two studies mentioned the mean duration of disease,15-17,19,22,25,27,31,34-37,41,42,45,46,49,54,55,57-59 ranging from 1.1 years36,58 to 12.6 years.46 Atopy. There was an overall percentage of 53% of patients who were atopic (95% CI, 49% to 57%) within the 20 studies reporting atopy,19,21,23,25-27,32,36-38,41,45-49,52,53,55,57 20% of patients with asthma (95% CI, 16% to 23%) within 11 studies,23,26,27,32,36-38,41,47,48,57 9% of patients with atopic dermatitis (95% CI, 7% to 12%) within 10 studies,23,25-27,32,36,37,45-47 and 33% of patients with allergic rhinitis (95% CI, 28% to 37%) within 9 studies.23,26,27,32,36,37,45,48,57 Other atopic factors retrieved were multiple allergic diseases, positive prick test, biological stigmas such as elevated serum total and specific IgE, eosinophil count, and eosinophil cationic protein level.

Aims and outcomes of included studies

All included studies, except 1,45 shared a similar main objective to evaluate the efficacy and safety of treatments for VKC based on clinical outcomes.15-22,24-26,28-40,42-44,46-55,57-59 Seven studies had an associated main objective of exploring the biological effects of treatments with a focus on changes in conjunctival impression cytology (in particular conjunctival immune cells and goblet cells),17,18,32 modifications of biological markers in tears (eosinophil cationic protein, cytokine levels, secretory mucins),39,40,47 and impacts on in vivo confocal microscopy images of dendritic cells.58 In 1 study, the main objective was the investigation of changes in biological markers after cyclosporine treatment in VKC.45

Characteristics of treatments

Type. A total of 20 different drugs grouped in 12 treatment classes were studied. All studies used local treatments except 2 assessing systemic treatments.38,42 Cyclosporine was assessed in 16 studies,18,19,23,32,33,35-37,41,45,47-50,52,53 mast cell stabilizers (MCSs) class in 11 studies (cromolyn sodium,17,24-27,39,44,55 lodoxamide,17,28,39,56 N-acetyl aspartyl glutamic acid [NAAGA],28,40 and nedocromil sodium25,55), and tacrolimus in 10 studies.16,34,36,37,43,44,49,57-59 Other drugs were used in fewer than 3 studies each: antihistaminic drugs (levocabastine);40,56 dual agents, MCS1 antihistaminic agents (olopatadine21,29 and ketotifen29); nonsteroidal anti-inflammatory drugs (topical ketorolac,51 topical diclofenac,22 oral aspirin42); corticosteroids46 (prednisolone, fluorometholone, loteprednol); mitomycinC;15,31 IFNa-2b;54,59 mipragoside;20 probiotic eye drop;30 and oral montelukast.38 Among the 14 studies using a placebo, the placebo was the vehicle in 8 studies,19,20,24,26,27,35,41,48 artificial tears or normal saline in 4 studies,15,31,33,51 and was not detailed in 2 studies.25,50

Meta-analyses on treatment efficacy compared to baseline

Overall meta-analyses showed that total symptoms and signs score improved globally (ES, 22.48; 95% CI, 22.88 to 22.09). Considering type of treatment, total symptoms and signs score decreased for MCSs (ES, 23.19; 95% CI, 24.26 to 22.13), cyclosporine (ES, 22.06; 95% CI, 22.72 to 21.40), and tacrolimus (ES, -2.39; 95% CI, 23.36 to 21.43). Antihistamines, dual agents, non-steroid anti-inflammatory drugs, IFNa-2b, mitomycin C, montelukast, probiotic eye drops, and mipragoside were seldom assessed, but all seemed to demonstrate improvement of total symptoms and signs score (Fig 3, and see Appendix E4 in this article’s Online Repository at www.jacionline.org). Meta-analysis on each symptom and sign demonstrated similar improvement for each treatment class (Fig 4 and see Appendixes E5 and E6 in this article’s Online Repository at www.jacionline.org). All symptoms were individually improved, with an ES <23 for itching (ES, 23.90; 95% CI, 24.67 to –3.13), foreign body sensation (ES, 23.05; 95% CI, 24.05 to 22.06), and TSSS (ES, 23.01; 95% CI, 23.47 to 22.55); an ES between 23 and 21.5 for tearing (ES, 22.29; ; 95% CI, 22.89 to 21.69) and discharge (ES, 21.99; 95% CI, 22.61 to 21.38); and an ES >21.5 for photophobia (ES, 21.39; 95% CI, 21.77 to 21.01) (Appendix E5). All signs were individually improved, with an ES <22 for limbal inflammation (ES, 22.81; 95% CI, 23.63 to 21.99), hyperemia (ES, 22.77; 95% CI, 23.43 to 22.12), and TOSS (ES, 22.26; 95% CI, 22.64 to 21.88); and an ES between 22 and 21 for tarsal papillae (ES, 21.28; 95% CI, 21.72 to 20.84) and keratitis (ES, 21.21; 95% CI, 21.61 to 20.8) (Appendix E6).

Meta-regressions and sensitivity analyses

Considering improvement in total symptoms and signs score, meta-regressions comparing the efficacy of treatments did not show significant differences depending on age, sex, baseline activityscoresofVKC,atopy, typeoftreatment,and concentration fortacrolimusandcyclosporine(Fig5,andseeAppendixE7inthis article’s Online Repository at www.jacionline.org). Metaregressions comparing the efficacy of treatments depending on each symptom or sign could not be performed due to collinearity. For rigor, funnel plots of all aforementioned meta-analyses were used to search for potential publication bias (partly presented in Appendix E8 in this article’s Online Repository at www. jacionline.org). All aforementioned meta-analyses were reperformed after exclusion of studies not evenly distributed around themeta-funnels andgavesimilarresults(data notshown).Finally, we performed the following additional sensitivity analyses (only on total symptoms and signs score in order to keep a sufficient number of studies). Using different time of re-evaluation gave similar results. The meta-analysis comparing the treatment groups to the placebo groups also demonstrated similar findings (see Appendix E9 in this article’s Online Repository at www. jacionline.org). Also, we performed a meta-analysis of treatment efficacy in the subgroup of 23 studies specifying the absence of concomitant treatment,15,16,20-22,27,29-35,37,42-44,46-49,51,58 and results did not differ from the main analyses.

DISCUSSION

The main findings were that clinical symptoms and signs improved with all tested treatments. Although transposition of our results in clinical practice should be interpreted with caution because of heterogeneity of patients between studies, efficacy of drugs did not differ in our metaanalysis between usual first-line therapy (MCSs), and immunomodulators (cyclosporine and tacrolimus that could be a good alternative in severe VKC). Other drugs were seldom assessed precluding robust conclusions. Underlined re-evaluation times were those included in the main analyses.

Conventional drugs

We demonstrated that all clinical scores showed great improvement with all treatments. Although meta-regressions were not feasible, tarsal papillae seemed the most resistant sign to treatment, in accordance with literature.48,50,56,58 Similarly, photophobia seemed the most difficult symptom to treat although results in literature are more discordant.19,21,54 MCSs were the most studied conventional drugs, especially cromolyn sodium that was historically the first described MCS.60 Despite insufficient data that precluded further comparisons among the 4 main MCSs, some investigators found better efficacy of nedocromil sodium and lodoxamide compared to that of cromolyn sodium.4 In case of absence of efficacy, they also proposed to try other MCSs before switching for another class, because of the specificities of each drug.4 Although other drugs (antihistamines, dual agents, nonsteroidal anti-inflammatory drugs, corticosteroids) are conventional in allergic conjunctivitis, efficacy of these drugs cannot be fairly assessed in our meta-analysis due to insufficient data. Likewise, no study on ‘‘over-the-counter’’ vasoconstrictors and oral antihistamines were reported. However, these 2 treatments are rarely used in VKC due to tachyphylaxis and inducedconjunctivitis for the first and more delayed and less efficient action than with topical presentation for the second.7

Place of topical immunomodulators

Cyclosporine and tacrolimus were the most studied drugs. Both drugs are calcineurin inhibitors that inhibit T-lymphocyte activation and the release of inflammatory cytokines.61 Ophthalmic application in VKC has been studied since the 1990s for cyclosporine,50 whereas the interest in tacrolimus is more recent.57 In our meta-analysis, patients with VKC benefiting from cyclosporine and tacrolimus often had a severe or refractory disease. We demonstrated similar benefits for both drugs, confirming that tacrolimus is a good alternative to cyclosporine for severe corticosteroid-dependent VKC, in line with literature.36,37,49 Efficacy also did not differ depending on concentration of cyclosporine or tacrolimus, for a similar number of applications per day, suggesting that low dose could be sufficient. It is particularly interesting considering that local tolerance of these otherwise well-tolerated drugs seems to be concentration-dependent.62 Initially, cyclosporine was only available as hospital-compound preparations at varying concentrations63 and now has commercialized forms at low dose (0.05% with only a marketing authorization in dry eye disease, and 0.1% concentrations, which is the only commercialized form with a marketing authorization for severe VKC). Several studies were in favor of 0.1% concentration.41,64 Similarly, tacrolimus is still mostly hospital-delivered, but is starting to be marketed in some countries.62 Tacrolimus 0.1% suspension received marketing authorization since 2008, only in Japan, with the indication: ‘‘Vernal conjunctivitis (in patients with inadequate response to anti-allergic agents)” and with the recommendation to use in patients with severe giant papillae. No other studies for marketing authorization have been conducted. The patients who may benefit the most from tacrolimus should be further investigated.

Other therapeutic options

Alongside these well-known treatments, some drugs were only reported in 1 (montelukast, mipragoside, probiotic eye drop) or 2 studies (IFNa-2b, mitomycin C) per drug in our meta-analyses, precluding robust conclusions and warranting further studies. There are also new emerging strategies of treatment of VKC. Atopic condition is an important risk factor of VKC, though not all patients with VKC present this condition—53% in our meta- analysis, which is in linewithliterature.65,66 Atopy couldalso be a prognostic factor for the severity of VKC.23 For patients who are atopic, specific therapies can be used, such as omalizumab or allergen immunotherapy. Omalizumab is an IgE-blocking agent commercialized for severe allergic asthma that could be an encouraging lead in VKC.67-69 Allergen immunotherapy, in case of proven sensitization, may as well be very promising,70,71 with even better results than with first-line topical medications.72 Subgroups of atopic patients in VKC should be considered, as some studies did.27,38 In addition to the atopy challenge, there may be a renewed interest in corticosteroids in VKC, although they are usually used sparingly because of their long-term side effects. Indeed, low-penetrating low-dose corticosteroids such as hydrocortisone, clobetasone, loteprednol, or fluorometholone in eye drops may warrant further studies in the treatment of VKC.1,46.

Quality of evidence and potential biases

Our study has some limitations. Even if our population of 1749 patients may not seem important considering the frequency of VKC in some countries,3 it is a considerable sample compared to existing literature in which 50% of included studies have fewer than 50 patients. It is also a huge step forward compared to the previous meta-analysis on VKC treatment that included 245 patients from 10 studies of which the most recent was from 2002 and which included few drugs.73 We updated the knowledge with all the new treatments. However, our meta-analysis is mainly salient in the evaluation of cyclosporine, tacrolimus, and MCSs, because other drugs were seldomly reported. Though there were similarities between patients, inclusion criteria were not identical between studies. In particular, even if severity assessment of VKC is not consensual6 and was not reported in most studies, a major limitation is heterogeneity in VKC severity across studies. It may have altered the comparison between classical first- and second-line medications—the latter mostly used in severe or refractory VKC. Severity assessment was also based on a subjectiveevaluation,limitingthequality ofthereported data. Similarly, the differential diagnosis of atopic keratoconjunctivitis could be questionable in studies reporting a mean age older than 20 years old,18,31,43,53 inducing a selection bias as both diseases could have specific treatment needs.74,75 There is also heterogeneity between studies regarding concomitant treatments. However, this corresponds to real life, and sensitivity analysis revealed similar results. Seasonal influence was not controlled in most studies. Moreover, symptoms and signs varied between studies, also inducing heterogeneity in total scores (total symptoms and signs score, TSSS, and TOSS). Even if we took the re-evaluation time closest to 1 month for each study, it varied across studies, limiting comparability. Not all of the included studies were randomized controlled studies, which may have induced heterogeneity in our results; however, we followed a rigorous methodology and computed sensitivity analyses by only keeping studies with the best studies designs. There was only 1 study for each stratification, thus the weight of studies did not require further ponderings. Anotherlimitationofourmeta-analysis isthelack ofdataoncharacteristics of patients such as history of previous treatments or duration of VKC, precluding further analyses. Finally, studies were mostly monocentric although all continents and all ethnicities were represented, favoring generalizability of our results.

Conclusions

We showed an overall improvement in clinical symptoms and signs with all treatments. The most frequent drugs were MCSs (usually considered as first-line therapy), cyclosporine, and tacrolimus. Efficacy of drugs did not differ, indicating that tacrolimus is a good alternative to cyclosporine for cases of severe VKC and promoting the possibility of using smaller doses for a similar effect in order to enhance local tolerance. Other drugs were seldom assessed and their places in VKC treatment strategy remain unclear.

REFERENCES

1. Leonardi A. Management of vernal keratoconjunctivitis. Ophthalmol Ther 2013;2: 73-88.
2. Bremond-Gignac D, Donadieu J, Leonardi A, Pouliquen P, Doan S, Chiambarretta F, et al. Prevalence of vernal keratoconjunctivitis: a rare disease? Br J Ophthalmol 2008;92:1097-102.
3. De Smedt SK, Nkurikiye J, Fonteyne YS, Tuft SJ, Gilbert CE, Kestelyn P. Vernal keratoconjunctivitis in school children in Rwanda: clinical presentation, impact on school attendance, and access to medical care. Ophthalmology 2012;119:1766-72.
4. Esposito S, Fior G, Mori A, Osnaghi S, Ghiglioni D. An update on the therapeutic approach to vernal keratoconjunctivitis. Paediatr Drugs 2016;18:347-55.
5. Feizi S, Javadi MA, Alemzadeh-Ansari M, Arabi A, Shahraki T, Kheirkhah A. Management of corneal complications in vernal keratoconjunctivitis: a review. Ocul Surf 2021;19:282-9.
6. Gokhale NS. Systematic approach to managing vernal keratoconjunctivitis in clinical practice: severity grading system and a treatment algorithm. Indian J Ophthalmol 2016;64:145-8.
7. Bielory L. Ocular allergy guidelines: a practical treatment algorithm. Drugs 2002; 62:1611-34.
8. Kumar S. Vernal keratoconjunctivitis: a major review. Acta Ophthalmol (Copenh) 2009;87:133-47.
9. Zicari AM, Capata G, Nebbioso M, De Castro G, Midulla F, Leonardi L, et al. Vernal keratoconjunctivitis: an update focused on clinical grading system. Ital J Pediatr 2019;45:64.
10. Harbour R, Miller J. A new system for grading recommendations in evidence based guidelines. BMJ 2001;323:334-6.
11. Benichou T, Pereira B, Mermillod M, Tauveron I, Pfabigan D, Maqdasy S, et al. Heart rate variability in type 2 diabetes mellitus: a systematic review and metaanalysis. PLoS One 2018;13:e0195166.
12. Benoist d’Azy C, Pereira B, Chiambaretta F, Dutheil F. Efficacy of different procedures of intra-corneal ring segment implantation in keratoconus: a systematic review and meta-analysis. Transl Vis Sci Technol 2019;8:38.
13. Navel V, Mulliez A, Benoist d’Azy C, Baker JS, Malecaze J, Chiambaretta F, et al. Efficacy of treatments for Demodex blepharitis: a systematic review and meta-analysis. Ocul Surf 2019;17:655-69.
14. Citrome L, Magnusson K. Paging Dr Cohen, Paging Dr Cohen . . . an effect size interpretation is required STAT!: visualising effect size and an interview with Kristoffer Magnusson. Int J Clin Pract 2014;68:533-4.
15. Akpek EK, Hasiripi H, Christen WG, Kalayci D. A randomized trial of low-dose, topical mitomycin-C in the treatment of severe vernal keratoconjunctivitis. Ophthalmology 2000;107:263-9.
16. Al-Amri AM, Mirza AG, Al-Hakami AM. tacrolimus ointment for treatment of vernal keratoconjunctivitis. Middle East Afr J Ophthalmol 2016;23:135-8.
17. Avunduk AM, Avunduk MC, Kapicioglu Z, Akyol N, Tavli L. Mechanisms and comparison of anti-allergic efficacy of topical lodoxamide and cromolyn sodium treatment in vernal keratoconjunctivitis. Ophthalmology 2000;107:1333-7.
18. Avunduk AM, Avunduk MC, Erd€ol H, Kapicioglu Z, Akyol N. Cyclosporine effects on clinical findings and impression cytology specimens in severe vernal keratoconjunctivitis. Ophthalmologica 2001;215:290-3.
19. Bleik JH, Tabbara KF. Topical cyclosporine in vernal keratoconjunctivitis. Ophthalmology 1991;98:1679-84.
20. Centofanti M, Schiavone M, Lambiase A, Taffara M, Giuffrida S, Bonini S. Efficacy of mipragoside ophthalmic gel in vernal keratoconjunctivitis. Eye Lond Engl 1996;10:422-4.
21. Corum I, Yeniad B, Bilgin LK, Ilhan R. Efficiency of olopatadine hydrochloride 0.1% in the treatment of vernal keratoconjunctivitis and goblet cell density. J Ocul Pharmacol Ther 2005;21:400-5.
22. D’Angelo G, Lambiase A, Cortes M, Sgrulletta R, Pasqualetti R, Lamagna A, et al. Preservative-free diclofenac sodium 0.1% for vernal keratoconjunctivitis. Graefes Arch Clin Exp Ophthalmol 2003;241:192-5.
23. Doan S, Gabison E, Abitbol O, Gatinel D, Chast F, Hoang-Xuan T. Efficacite de la ciclosporine A en collyre a 2 % en tant qu’epargneur de cortico€ıdes dans le traitement de la keratoconjonctivite vernale corticodependante. J Fr Ophtalmol 2007;30: 697-701.
24. Easty DL, Rice NS, Jones BR. Clinical trial of topical disodium cromoglycate in vernal kerato-conjunctivitis. Clin Allergy 1972;2:99-107.
25. el Hennawi M. A double blind placebo controlled group comparative study of ophthalmic sodium cromoglycate and nedocromil sodium in the treatment of vernal keratoconjunctivitis. Br J Ophthalmol 1994;78:365-9.
26. Foster CS, Duncan J. Randomized clinical trial of topically administered cromolyn sodium for vernal keratoconjunctivitis. Am J Ophthalmol 1980;90:175-81.
27. Foster CS. Evaluation of topical cromolyn sodium in the treatment of vernal keratoconjunctivitis. Ophthalmology 1988;95:194-201.
28. G€und€uz K, Uc¸akhan O, Budak K, Eryilmaz T, Ozkan M. Efficacy of lodoxamide 0.1% versus N-acetyl aspartyl glutamic acid 6% ophthalmic solutions in patients with vernal keratoconjunctivitis. Ophthalmic Res 1996;28:80-7.
29. Hida WT, Nogueira DC, Schaefer A, Dantas PEC, Dantas MCN. [Comparative study between 0.025% ketotifen fumarate and 0.1% olopatadine hydrochloride in the treatment of vernal keratoconjunctivitis]. Arq Bras Oftalmol 2006;69:851-6.
30. Iovieno A, Lambiase A, Sacchetti M, Stampachiacchiere B, Micera A, Bonini S. Preliminary evidence of the efficacy of probiotic eye-drop treatment in patients with vernal keratoconjunctivitis. Graefes Arch Clin Exp Ophthalmol 2008;246: 435-41.
31. Jain AK, Sukhija J. Low dose mitomycin-C in severe vernal keratoconjunctivitis: a randomized prospective double blind study. Indian J Ophthalmol 2006;54:111-6.
32. Keklikci U, Soker SI, Sakalar YB, Unlu K, Ozekinci S, Tunik S. Efficacy of topical cyclosporin A 0.05% in conjunctival impression cytology specimens and clinical findings of severe vernal keratoconjunctivitis in children. Jpn J Ophthalmol 2008;52:357-62.
33. Keklikci U, Dursun B, Cingu AK. Topical cyclosporine a 0.05% eyedrops in the treatment of vernal keratoconjunctivitis: randomized placebo-controlled trial. Adv Clin Exp Med 2014;23:455-61.
34. Kheirkhah A, Zavareh MK, Farzbod F, Mahbod M, Behrouz MJ. Topical 0.005% tacrolimus eye drop for refractory vernal keratoconjunctivitis. Eye Lond Engl 2011;25:872-80.
35. Kilic¸ A, G€urler B. Topical 2% cyclosporine A in preservative-free artificial tears for the treatment of vernal keratoconjunctivitis. Can J Ophthalmol 2006;41:693-8.
36. Kumari R, Saha BC, Sinha BP, Mohan N. Tacrolimus versus cyclosporine- comparative evaluation as first line drug in vernal keratoconjuctivitis. Nepal J Ophthalmol 2017;9:128-35.
37. Labcharoenwongs P, Jirapongsananuruk O, Visitsunthorn N, Kosrirukvongs P, Saengin P, Vichyanond P. A double-masked comparison of 0.1% tacrolimus ointment and 2% cyclosporine eye drops in the treatment of vernal keratoconjunctivitis in children. Asian Pac J Allergy Immunol 2012;30:177-84.
38. Lambiase A, Bonini S, Rasi G, Coassin M, Bruscolini A, Bonini S. Montelukast, a leukotriene receptor antagonist, in vernal keratoconjunctivitis associated with asthma. Arch Ophthalmol 2003;121:615-20.
39. Leonardi A, Borghesan F, Avarello A, Plebani M, Secchi AG. Effect of lodoxamide and disodium cromoglycate on tear eosinophil cationic protein in vernal keratoconjunctivitis. Br J Ophthalmol 1997;81:23-6.
40. Leonardi A, Bremond-Gignac D, Bortolotti M, Violato D, Pouliquen P, Delval L, et al. Clinical and biological efficacy of preservative-free NAAGA eye-drops versus levocabastine eye-drops in vernal keratoconjunctivitis patients. Br J Ophthalmol 2007;91:1662-6.
41. Leonardi A, Doan S, Amrane M, Ismail D, Montero J, Nemeth J, et al. A randomized, controlled trial of cyclosporine A cationic emulsion in pediatric vernal keratoconjunctivitis: the VEKTIS study. Ophthalmology 2019;126:671-81.
42. Meyer E, Kraus E, Zonis S. Efficacy of antiprostaglandin therapy in vernal conjunctivitis. Br J Ophthalmol 1987;71:497-9.
43. M€uller GG, Jose NK, de Castro RS. Topical tacrolimus 0.03% as sole therapy in vernal keratoconjunctivitis: a randomized double-masked study. Eye Contact Lens 2014;40:79-83.
44. M€uller EG, Santos MSD, Freitas D, Gomes JAP, Belfort R. Tacrolimus eye drops as monotherapy for vernal keratoconjunctivitis: a randomized controlled trial. Arq Bras Oftalmol 2017;80:154-8.
45. Nebbioso M, Sacchetti M, Bianchi G, Zicari AM, Duse M, Del Regno P, et al. Tear ferning test and pathological effects on ocular surface before and after topical cyclosporine in vernal keratoconjunctivitis patients. J Ophthalmol 2018;2018: 1061276.
46. Oner V, T€urkc€u FM, Tas¸ M, Alakus¸ MF, Is¸can Y. Topical loteprednol etabonate 0.5 % for treatment of vernal keratoconjunctivitis: efficacy and safety. Jpn J Ophthalmol 2012;56:312-8.
47. Oray M, Toker E. Tear cytokine levels in vernal keratoconjunctivitis: the effect of topical 0.05% cyclosporine a therapy. Cornea 2013;32:1149-54.
48. Pucci N, Novembre E, Cianferoni A, Lombardi E, Bernardini R, Caputo R, et al. Efficacy and safety of cyclosporine eyedrops in vernal keratoconjunctivitis. Ann Allergy Asthma Immunol 2002;89:298-303.
49. Pucci N, Caputo R, di Grande L, de Libero C, Mori F, Barni S, et al. Tacrolimus vs. cyclosporine eyedrops in severe cyclosporine-resistant vernal keratoconjunctivitis: a randomized, comparative, double-blind, crossover study. Pediatr Allergy Immunol 2015;26:256-61.
50. Secchi AG, Tognon MS, Leonardi A. Topical use of cyclosporine in the treatment of vernal keratoconjunctivitis. Am J Ophthalmol 1990;110:641-5.
51. Sharma A, Gupta R, Ram J, Gupta A. Topical ketorolac 0.5% solution for the treatment of vernal keratoconjunctivitis. Indian J Ophthalmol 1997;45:177-80.
52. Spadavecchia L, Fanelli P, Tesse R, Brunetti L, Cardinale F, Bellizzi M, et al. Efficacy of 1.25% and 1% topical cyclosporine in the treatment of severe vernal keratoconjunctivitis in childhood. Pediatr Allergy Immunol 2006;17:527-32.
53. Tesse R, Spadavecchia L, Fanelli P, Rizzo G, Procoli U, Brunetti L, et al. Treatment of severe vernal keratoconjunctivitis with 1% topical cyclosporine in an Italian cohort of 197 children. Pediatr Allergy Immunol 2010;21:330-5.
54. Turan-VuralE, Acar BT,Acar S.The efficacyof topicalinterferonalpha2btreatment in refractory vernal keratoconjunctivitis. Ocul Immunol Inflamm 2012;20:125-9.
55. Verin PH, Dicker ID, Mortemousque B. Nedocromil sodium eye drops are more effective than sodium cromoglycate eye drops for the long-term management of vernal keratoconjunctivitis. Clin Exp Allergy 1999;29:529-36.
56. Verin P, Allewaert R, Joyaux JC, Piozzi E, Koliopoulos J, Bloch-Michel E, et al. Comparison of lodoxamide 0.1% ophthalmic solution and levocabastine 0.05% ophthalmic suspension in vernal keratoconjunctivitis. Eur J Ophthalmol 2001;11: 120-5.
57. Vichyanond P, Tantimongkolsuk C, Dumrongkigchaiporn P, Jirapongsananuruk O, Visitsunthorn N, Kosrirukvongs P. Vernal keratoconjunctivitis: result of a novel therapy with 0.1% topical ophthalmic FK-506 ointment. J Allergy Clin Immunol 2004;113:355-8.
58. Wan Q, Tang J, Han Y, Wang D, Ye H. Therapeutic effect of 0.1% tacrolimus eye drops in the tarsal form of vernal keratoconjunctivitis. Ophthalmic Res 2018;59: 126-34.
59. Zanjani H, Aminifard MN, Ghafourian A, Pourazizi M, Maleki A, Arish M, et al. Comparative evaluation of tacrolimus versus interferon alpha-2b eye drops in the treatment of vernal keratoconjunctivitis: a randomized, double-masked study. Cornea 2017;36:675-8.
60. Ben-Eli H, Solomon A. Topical antihistamines, mast cell stabilizers, and dualaction agents in ocular allergy: current trends. Curr Opin Allergy Clin Immunol 2018;18:411-6.
61. Addis H, Jeng BH. Vernal keratoconjunctivitis. Clin Ophthalmol 2018;12:119-23.
62. Ghasemi H, Djalilian A. Topical calcineurin inhibitors: expanding indications for corneal and ocular surface inflammation. J Ophthalmic Vis Res 2019;14:398-9.
63. Lallemand F, Schmitt M, Bourges J-L, Gurny R, Benita S, Garrigue J-S. Cyclosporine A delivery to the eye: a comprehensive review of academic and industrial efforts. Eur J Pharm Biopharm 2017;117:14-28.
64. Nebbioso M, Alisi L, Giovannetti F, Armentano M, Lambiase A. Eye drop emulsion containing 0.1% cyclosporin (1 mg/mL) for the treatment of severe vernal keratoconjunctivitis: an evidence-based review and place in therapy. Clin Ophthalmol 2019;13:1147-55.
65. Bonini S, Bonini S, Lambiase A, Marchi S, Pasqualetti P, Zuccaro O, et al. Vernal keratoconjunctivitis revisited: a case series of 195 patients with long-term followup. Ophthalmology 2000;107:1157-63.
66. Lambiase A, Minchiotti S, Leonardi A, Secchi AG, Rolando M, Calabria G, et al. Prospective, multicenter demographic and epidemiological study on vernal keratoconjunctivitis: a glimpse of ocular surface in Italian population. Ophthalmic Epidemiol 2009;16:38-41.
67. Occasi F, Duse M, Nebbioso M, De Castro G, Di Fraia M, Capata G, et al. Vernal keratoconjunctivitis treated with omalizumab: a case series. Pediatr Allergy Immunol 2017;28:503-5.
68. Doan S, Amat F, Gabison E, Saf S, Cochereau I, Just J. Omalizumab in severe refractory vernal keratoconjunctivitis in children: case series and review of the literature. Ophthalmol Ther 2017;6:195-206.
69. Heffler E, Picardi G, Liuzzo MT, Pistorio MP, Crimi N. Omalizumab treatment of vernal keratoconjunctivitis. JAMA Ophthalmol 2016;134:461-3.
70. Sayed KM, Kamel AG, Ali AH. One-year evaluation of clinical and immunological efficacy and safety of sublingual versus subcutaneous allergen immunotherapy in allergic conjunctivitis. Graefes Arch Clin Exp Ophthalmol 2019;257: 1989-96.
71. Norris MR, Khaimov M, Bielory L. Use of allergen immunotherapy for treatment of allergic conjunctivitis. Curr Opin Allergy Clin Immunol 2020;20:609-15.
72. Mahdy RAR, Nada WM, Marei AA. Subcutaneous allergen-specific immunotherapy versus topical treatment in vernal keratoconjunctivitis. Cornea 2012;31: 525-8.
73. Mantelli F, Santos MS, Petitti T, Sgrulletta R, Cortes M, Lambiase A, et al. Systematic review and meta-analysis of randomised clinical trials on topical treatments for vernal keratoconjunctivitis. Br J Ophthalmol 2007;91:1656-61.
74. Leonardi A, Lazzarini D, Motterle L, Bortolotti M, Deligianni V, Curnow SJ, et al. Vernal keratoconjunctivitis-like disease in adults. Am J Ophthalmol 2013;155: 796-803.
75. Bremond-Gignac D, Nischal KK, Mortemousque B, Gajdosova E, Granet DB, Chiambaretta F. Atopic keratoconjunctivitis in children: clinical features and diagnosis. Ophthalmology 2016;123:435-7.