Topical Steroids and Antibiotics for Adult Blepharokeratoconjunctivitis (BKC): A Meta-Analysis of Randomized Clinical Trials

Purpose A meta-analysis was conducted to evaluate the efficacy and safety of topical treatments (including steroids and antibiotics) for adults with blepharokeratoconjunctivitis (BKC). Methods The following databases were searched for relevant randomised controlled trials (RCTs): China National Knowledge Infrastructure (CNKI), Web of Science, MEDLINE, PubMed, Embase, and Cochrane Central Register of Controlled Trials database (CENTRAL). Two reviewers selected studies and analyzed the risk of bias independently. The treatments were loteprednol 0.5%/tobramycin 0.3% (LE/T) and dexamethasone 0.1%/tobramycin 0.3% (DM/T). The efficacy outcome measures were change from baseline (CFB) in composite scores of ocular symptoms and signs; the CFB in the signs composite scores for blepharitis, conjunctivitis, and keratitis at each visit; the total ocular adverse event incidence (AEs); and the incidence of intraocular pressure (IOP) increase after treatment. Prepost mean differences (MDs) were compared for continuous outcome variables, and incidences were analyzed for dichotomous data. The pooled effect sizes were analyzed using 95% confidence intervals (CIs) in a fixed-effect model. Heterogeneity was evaluated using the Q-test and I2 statistic. Results The CFB to final visit in ocular symptoms and signs of BKC was not statistically different between the two treatments (95% CI, −0.33 to 1.50; MD = 0.58; P=0.21). The CFB in signs composite scores for blepharitis (95% CI, −0.16 to 0.48; MD = 0.16; P=0.32), conjunctivitis (95% CI, −0.55 to 1.76; MD = 0.61; P=0.30), and keratitis (95% CI, 0.00–0.28; MD = 0.14; P=0.05) was also similar with the two treatments. LE/T was a safer intervention than DM/T, with fewer overall adverse events (95% CI, 0.34–0.80; RR = 0.52; P=0.003) and significantly less elevation of intraocular pressure (IOP) (95% CI, 0.32–0.70; RR = 0.47; P=0.0002). Conclusions DM/T and LE/T are both effective treatments for BKC, but LE/T may be a safer intervention.


Introduction
Blepharokeratoconjunctivitis (BKC) is an inflammatory disorder of the eyelid margin with secondary conjunctivitis and keratopathy [1]. e main clinical manifestations include an inflamed palpebral margin, conjunctival congestion, conjunctival follicles, punctate corneal epithelial erosion, corneal stromal infiltration, corneal ulcer, and eventually scarring with consequent loss of vision and possibly blindness [2][3][4]. e severity of BKC can be categorized as mild, moderate, and severe, according to the depth and extent of keratopathy [5]. Inappropriate therapy or delayed diagnosis may cause serious corneal disorders [4].
Conventional treatments for BKC include nonsteroidal anti-inflammatory drugs (NSAIDs), antibiotics, and corticosteroids and are often prescribed in combination with artificial eye drops and eyelid hygiene, massage, or light therapy [6,7]. Several clinical trials have been performed to evaluate the safety and efficacy of topical drugs in the treatment of BKC in adults. No consensus, however, has been reached. Pranoprofen can significantly relieve ocular surface inflammation and improve the cure rate of BKC [8].
1% fluorometholone eye drops combined with levofloxacin eye drops were also effective [9]. Topical antibiotics eradicate positive cultures of bacteria from the palpebral margins, and different types of antibiotics (such as ciprofloxacin or tobramycin) have similar impact on the number of bacteria.
Topical steroids have also proven useful in relieving ocular symptoms, but ineffective in eliminating harmful bacteria from eyelids [10]. Clinically, patients with mild BKC are generally treated with nonsteroidal anti-inflammatory drugs, which are safer than glucocorticoids [8]. However, for moderate to severe BKC, glucocorticoids are generally used for their effectiveness as anti-inflammatory agents [8]. e present systematic review and meta-analysis was undertaken to evaluate the efficacy and safety of topical treatments for BKC. e evidence reviewed was from published RCTs on the effectiveness of topical treatments for adult BKC.

Study Selection.
Two reviewers (LZ and YJS) selected studies independently using predetermined selection criteria (see Section 2.3). e titles and abstracts were screened to assess relevance and eligibility, and the full article was accessed if necessary to check eligibility. If the two reviewers did not agree, discussion with an additional reviewer (ZQP) allowed a final decision to be reached.

Inclusion and Exclusion
Criteria. Studies meeting the following criteria were included: (1) the study included patients who were at least 18 years old; (2) patients had a diagnosis of BKC prior to the treatments; (3) the presence of ocular signs was evidenced by slit-lamp examination; (4) patients had received topical ocular treatments; and (5) the study design was RCT.
Trials were excluded if: (1) e patients had previously or concurrently taken oral or topical treatments (2) e data could not be included in statistical analysis due to incomplete reporting (e.g., lack of standard deviation).

Data Extraction.
Two reviewers (LZ and YJS) selected the studies based on the criteria outlined in Section 2.3. Each reviewer then extracted variables independently using a previously designed data extraction tool. e extracted variables included sample size, locations of studies, patients' mean age and gender balance, types of intervention, drug dosage, duration of follow-up, primary and secondary outcomes, and assessment methods. If these variables were not reported, the reviewers contacted the trial authors via email to request the data.

Assessment of Risk.
Risk of bias was accessed by two reviewers (LZ and YJS) independently based on the Cochrane Handbook criteria focusing on seven aspects as follows: allocation concealment, blinding of outcome assessors, incomplete outcome data, random sequence generation, participant and personnel blinding, selective reporting, and other sources of bias. Each included trial was critically appraised by the reviewers considering each of the above areas. Any unresolved disagreement between the two reviewers was resolved following review by author ZQP.

Statistical Analyses.
e primary outcome measure was CFB of the ocular symptoms and signs composite score at the final follow-up visit (day 15 ± 1). e secondary outcome measures were the CFB in composite score of ocular signs of blepharitis, conjunctivitis, and keratitis at each follow-up. e safety outcome measure was overall incidence of ocular AEs (including increased IOP). e weighted mean difference (WMD) of continuous data with 95% CI was determined using the inverse variance method. For dichotomous data, the risk ratio (RR) was calculated with 95% CI and the weighted summary RR was determined using the Mantel-Haenszel method. e Q-test and I 2 statistic were used to assess heterogeneity, with values of 25%, 50%, and 75% indicating low, medium, and high levels, respectively. At I 2 level greater than 50%, meta-analysis was conducted using the fixed-effects model, and at lower levels, the random-effects model. Statistical analyses in this study were conducted using Review Manager Software (RevMan 5.3; Cochrane Collaboration, Oxford, UK), and P values lower than 0.05 were considered significant.

Search Results and Study Inclusion.
e data searches yielded 587 studies ( Figure 1) of which 236 duplicated citations were excluded. After the exclusion of irrelevant citations based on titles and abstracts, 43 articles remained. On full-text checking, 41 of these were excluded due to their type of intervention (23 articles), nonhuman studies (4 articles), lack of reported outcome (1 article), review (1 article), letter (1 article), abstract (4 articles), case report (6 articles), or unqualified outcomes (1 article). e remaining 2 studies were included in the meta-analysis ( Figure 1).

Study Characteristics.
Characteristics of the included studies are summarized in Table 1. Both were published in English [11,12], one was conducted in China and the other in the U.S.A. In total, the studies involved 627 patients with BKC. e two studies used the same criteria for diagnosis of BKC and compared LE/T and DM/T ophthalmic solutions applied 4 times daily.

Risk of Bias.
Overall, risk of bias assessment found that the included studies were at low risk, and therefore, high quality ( Figure 2), although outcome, assessments were not blinded (Figures 2 and 3).

Effect Outcomes.
e primary outcomes of the metaanalysis are presented in Figure 4. Statistical heterogeneity (I 2 ) was zero for most outcome measures, and a fixed-effects model was used for these. e RCTs tested effects of the two interventions by comparing the ocular signs and symptoms composite score between the start point and endpoints. e endpoints included the score at the second visit (day 3 ± 1), the third visit (day 7 ± 1), and the fourth visit (day 15 ± 1).
e CFB in ocular symptoms and signs composite scores was not significantly different between the LE/T and DM/T groups at the second (95% CI: −0.40-0.98, I 2 � 0%; MD � 0.29) or the fourth follow-up visits (95% CI: −0.33 to 1.50, I 2 � 0%; MD � 0.58). However, the LE/T group showed a higher improvement from baseline than the DM/Tgroup at the third visit (95% CI: 0.03-1.69, I 2 � 0%; MD � 0.86) ( Figure 4). Statistically significant heterogeneity was found in the conjunctivitis composite score at the final visit (day 15 ± 1), so a random-effect model was used (Figure 7).

Discussion
BKC is an ocular disease that presents challenges to ophthalmologists in clinical practice. As its pathogenesis is not completely understood, clinical guidelines have not yet been provided. e purpose of BKC treatment is to modify meibomian gland function and the bacterial flora of the eyelid and conjunctiva and to reduce ocular inflammation [13]. Fortunately, with extensive research in recent years, knowledge about the disease is increasing. Yactayo-Miranda    et al. [14] reported a higher rate of positive thioglycolate broth cultures in eyes with chronic blepharoconjunctivitis than in eyes without the disease. Treatment with 0.5% levofloxacin reduced the rate of positive thioglycolate broth cultures as well as the number of bacteria harbored on the conjunctiva. Azithromycin 1% ophthalmic solution has also been evaluated as a treatment for blepharitis or blepharoconjunctivitis and has been found to significantly decrease the number of coagulase-negative staphylococci and coryneform bacteria and improve the symptoms and signs of blepharitis [15][16][17]. Compared with this intervention, however, a combination of 0.3% tobramycin and 0.05% dexamethasone allows faster and more effective reduction of inflammation in moderate to severe blepharitis/blepharoconjunctivitis [17]. Another improvement on azithromycin 1.0% has been demonstrated by combining it with dexamethasone 0.1% resulting in more effective relief of clinical symptoms and eradication of bacteria [18]. In addition, an antibiotic-steroid combination of 0.1% dexamethasone plus 0.5% moxifloxacin is effective in reducing the inflammation associated with bacterial blepharitis [19]. us, a range of treatment methods exist, with different outcomes and levels of effectiveness in the treatment of BKS.
Existing systematic reviews on the safety and efficacy of topical and systematic treatments for BKC in children found no high-quality evidence [6,20]    Journal of Ophthalmology antibiotics, and steroids have been considered as possible treatments for children and are available as eye drops or ointments [6]. e present meta-analysis included two articles involving the same intervention (antibiotics plus steroids) as well as the same outcome measures [11,12]. A third study testing this intervention was not selected for inclusion in the meta-analysis, due to a lack of standard deviation data [21]. e present analysis reveals no significant differences between LE/T and DM/T as treatments of BKC in terms of    Journal of Ophthalmology ocular signs and symptoms. ese results are inconsistent with those reported by Rhee and Mah [21] in which scores related to signs of blepharitis, conjunctivitis, and ocular discharge were reduced to a greater extent following DM/T treatment than LE/T treatment [21]. No AEs were reported, and the mean IOP change from baseline was similar in the two treatment groups [21]. e trial was a single-center study which included only 40 subjects with moderate or severe BKC, and the mean follow-up period was 3.5 days. In addition, the treatments were administered twice daily as opposed to four times per day in the two included trials. ese differences in methodology may explain, at least in part, the differences between findings.
Incidences of nonocular treatment-emergent adverse events were comparable between treatments (8/315 in LE/T; 7/315 in DM/T) in the current meta-analysis. e nonocular side-effects were mostly mild to moderate. Only one hypertension subject in the LE/T group and one with a headache in the DM/T group were considered severe. Fewer ocular adverse events occurred in the LE/T group (6/315) than in the DM/T group (14/316). Most ocular AEs were mild to moderate, but IOP was increased, with significantly higher frequency in the DM/T than LE/T group. In addition, changes in IOP found among Chinese patients were higher than those among Americans, according to these two RCTs, in agreement with previous findings [22]. However, any possible association between race and elevation of steroidrelated IOP has not yet been reported [23], but risk factors related to glaucoma, such as higher myopia, higher IOP and larger cup to disc area ratio are reportedly more common among Asians [24].
Corticosteroids may lead to multiple physiological changes involving the deposition of extracellular matrix material, the production of cross-linked actin fibers, inhibition of cell phagocytosis, and possibly to increased resistance to aqueous outflow and an elevation of IOP [25,26]. Due to the different molecular structures of dexamethasone and loteprenol etabonate, the latter is associated with relatively low IOP elevation, making it a safer treatment for ocular inflammation, especially for the patients with a higher risk of IOP [12,27].
Some    Journal of Ophthalmology treatments including longer remission periods than topical corticosteroids [28,29], but this advantage needs to be supported by more studies with large samples.
Blepharitis is a common condition, and its treatment is an important clinical issue [2]. For Demodex-infested patients, eyelid scrubbing with a mild shampoo is not effective in reducing the number of mites, while many studies have shown that scrubbing the eyelid margin with tea tree oil can achieve this and improve ocular discomfort [30,31]. Intense pulse light (IPL) is another effective and safe physical treatment for moderate to severe BKC adult patients and may be more useful in lessening eyelid margin inflammation than Meibomian gland expression [7].
Finally, dry eye secondary to BKC should also be considered. Artificial eye drops are an important treatment option because of the evaporative drying that accompanies meibomian gland disfunction and changes of the ocular surface microenvironment [32].
is review has some limitations. First, the number of clinical trials included is small. Only a few studies have focused on topical treatment of adult BKC. Second, this meta-analysis lacks subgroup analysis on different severities of BKC. Further high-quality randomised controlled trials with large stratified samples including different severity levels are warranted to determine the efficacy and limitations of the antibiotic plus steroid and other treatments for these different clinical categories of BKC.

Conclusion
is meta-analysis evaluated the evidence on efficacy and safety of topical antibiotics and steroids for adults with BKC. e results showed that LE/T and DM/T are similarly effective in the treatment of BKC, and that LE/T is associated with better safety.

Conflicts of Interest
e authors declare that they have no conflicts of interest.