Aims To report the clinical characteristics and correlation between meibomian gland dysfunction (MGD) and keratitis in young patients with ocular demodicosis.
Methods Observational case series of 60 patients younger than 35 years with ocular demodicosis, of which the diagnosis was based on microscopic counting of Demodex folliculorum and D. brevis of epilated lashes. Severity of keratitis and MGD was graded by photography and meibography, respectively, in a masked fashion.
Results MGD was detected in 54/60 (90%) patients with the loss of meibomian gland in the upper lid more than the lower lid (p<0.001). Blepharoconjunctivitis and a variety of corneal pathologies were noted in 47/60 (78.3%) and 39/60 (65%) patients, respectively. For a total of 120 eyes, normal cornea was noted in 53 (44.2%) eyes, superficial punctate keratitis or limbitis was noted in 17 (14.2%), while corneal stromal infiltration was found in 50 (41.7%) eyes. Both univariate and multivariate analyses showed that the severity of meibomian gland loss was significantly correlated with higher D. brevis count and more severe keratitis (all p<0.05). Rapid resolution of keratitis and blepharoconjunctivitis was accompanied by significant reduction of the Demodex count in 48 patients receiving lid scrub directed to kill mites.
Conclusions There is a significant correlation between MGD and keratitis in young patients with ocular demodicosis especially inflicted by D. brevis.
- ocular surface
Statistics from Altmetric.com
Among different species of mites, Demodex folliculorum and D. brevis are the only two species affecting the human skin.1–3 D. folliculorum infests the lash follicle, while D. brevis infests the lash’s sebaceous gland and the meibomian gland.3 Such infestation termed ocular demodicosis has been implicated in a number of external eye diseases such as blepharitis, eyelash disorders, conjunctivitis and blepharoconjunctivitis.2–8
Demodicosis is highly age dependent and generally regarded as common in asymptomatic adults, especially in elder people, but rare in children.9–11 Meibomian gland dysfunction (MGD) is also more prevalent in elder people but rare in children.12 In paediatric patients, we have reported that ocular demodicosis can be associated with chronic blepharoconjunctivitis5 and chalazia especially D. brevis infestation, another disease affecting meibomian glands.11 We thus wonder if ocular demodicosis may also play a role in MGD. To mitigate the concern of a high prevalence of both MGD and demodicosis in the elder population, we decided to conduct a prospective observational study by focusing on young adults. As there is notable keratitis in this cohort resembling herpes keratitis, we also summarise the clinical features that help differentiate from herpes keratitis.
Patients and methods
Following the tenets of the Declaration of Helsinki, we enrolled 60 patients younger than 35 years with ocular demodicosis at the Zhongshan Ophthalmic Center. The diagnosis of ocular demodicosis was based on lash sampling and microscopic mite counting as reported in all patients.5 13 14 Because demodicosis is not as common in the paediatric and young population,9 four instead of two lashes per lid were removed under slit-lamp microscope in all patients including two children (age less than 14), whose lash sampling were performed under general anaesthesia due to their poor cooperation as reported.5 11
All patients with demodicosis were instructed to receive the mite-killing therapy. Because there is no lid hygiene treatment based on tea tree oil in China, only 48 out of 60 patients could purchase Cliradex from Bio-Tissue (Miami, Florida, USA) that contains the active ingredient reported to kill Demodex mites.14 15 These 48 patients received Cliradex lid scrub twice daily for 3 months. For those presenting with corneal inflammation and conjunctival injection, 0.02% fluorometholone (Santan, Japan) was used twice daily for 3–10 days. Artificial tears were used 4–6 times daily as needed for patients presented with dry eye symptoms.
Diagnosis and grading of MGD
The diagnosis of MGD was made by slit-lamp examination to detect ≥1 lid margin abnormalities (irregular lid margin, vascular engorgement, plugged meibomian gland orifices and anterior or posterior replacement of the mucocutaneous junction) and poor meibum expression under digital pressure.16 The severity of MGD was graded by meibography using Keratograph 5 M (Oculus, Wetzlar, Germany) as reported17 to generate the meiboscore by combining the score from both upper and lower eyelids (see online supplementary figure), that is, 0 for no meibomian gland loss, 1–2 for less than one-third total gland loss and considered as ‘mild’, 3–4 for one-third to two-thirds total gland loss and considered as ‘moderate’ and 5–6 for more than two-thirds total gland loss and considered as ‘severe’.
Grading of corneal changes
We also routinely graded corneal pathologies by slit-lamp photographs as ‘0’ for no abnormal finding, ‘1’ for superficial punctate keratopathy (SPK) only or limbitis only and ‘2’ for stromal involvement such as infiltration or ulceration (figure 1). For those presenting with corneal ulcers, standard corneal scraping and microbiological culturing were also performed.
Grading of MGD and corneal changes was performed by two investigators under a masked fashion. Any inconsistent grading between the two was arbitrated by a third investigator, who had no knowledge about the other clinical information at the time. Meiboscore of the worse eye was used for comparison among different patients. The Pearson’s Χ2 test and p for trend test were used to compare categorical and ranked variables, respectively. Univariate regression was performed to evaluate the risk factors of keratitis and MGD. Those factors identified as probably significant (p<0.1) by univariate regression were subjected to multivariate regression. All statistical analyses were performed using SPSS software V.16.0 and reported as two-tailed probabilities, with p<0.05 being considered significant.
The study group consisted of 60 patients (19.1±7.5 years, 22 males and 38 females) with ocular demodicosis. The history disclosed chalazia in 29 (48.3%) patients. The Demodex count was 5.6±3.5 including 2.5±2.4 for D. brevis and 3.1±1.9 for D. folliculorum. Further examination revealed MGD (n=54, 90%), lash disorders (n=51, 85%), blepharitis (n=47, 78.3%), keratitis (n=39, 65%), allergic conjunctivitis (n=7, 11.7%) and phlyctenular conjunctivitis (n=3, 5%). The keratitis of all 39 patients had not responded to topical antiviral agents, antibiotics and steroid for a period up to 20 years under the clinical impression of herpes simplex keratitis (HSK, n=24), bacterial keratitis (n=3), severe dry eye (n=8), vernal keratoconjunctivitis (n=2) and limbitis (n=2).
MGD is prevalent and worse in upper lids
MGD was common (90%) and remarkable in 35 patients (35/60, 58.4%) presenting more than one-third meibomian gland loss (table 1). Intriguingly, MGD was worse in the upper eyelid than the lower lid (p<0.001, figure 2, representative cases). Consequently, the tear breakup time was 2.5±1.3 s in these patients. Meiboscore was significantly correlated with D. brevis count and keratitis grading (both p<0.001) but not correlated with D. folliculorum count, age, gender, blepharitis and disease duration (all p>0.05, table 1).
Keratitis is prevalent and severe
We then looked deeper into keratitis and noted different forms in 67 eyes of 39 study patients, including grade 1, that is, SPK only (figure 1A) or limbitis (figure 1B) in 17 eyes and grade 2, that is, corneal stromal involvement in 50 eyes (figure 1C-H). The corneal stromal involvement was associated with (19 eyes, figure 1C–E) or without (29 eyes, figure 1F–G) an epithelial defect and could be presented with perforation (two eyes, figure 1H), corneal neovascularisation (36 eyes) or scar (17 eyes). Keratitis was bilateral (76.9%) more than unilateral (23.1%) and located at inferior (32 eyes), central (25 eyes), diffusive (seven eyes) and superior (three eyes) corneas. For the 50 eyes with grade 2 keratitis, the stromal lesion extended to the limbus in all except two eyes (figure 1E). The microbial results of 19 cases presenting with corneal ulcers were all negative for aerobic microbes. An overwhelming majority, that is, 57 of 67 eyes (85.1%) with keratitis also presented with blepharitis, which was more evident in the central upper lid in 51 of 57 (89.5%) patients (figure 1D, representative cases). The severity of keratitis was significantly correlated with D. brevis count and meiboscore (both p<0.001) but not with D. folliculorum count, age, gender, blepharitis and disease duration (all p>0.05, table 1).
Significant correlation between MGD and keratitis
When the worse eye of each patient were compared, we noted a significant correlation between meiboscore and keratitis scores in 60 study patients (Spearman’s rho=0.97, p<0.001) (Figure 2, representative case 1). Multivariate regression analysis of the three variables that were found to be significant in univariate regression, that is, D. brevis count, keratitis grading and meiboscore, disclosed that both D. brevis count (OR 1.56; 95% CI 1.09 to 2.23, p=0.01) and higher keratitis scores (OR 8.72; 95% CI 3.21 to 23.67, p<0.001) were the two predictors of meiboscore, but meiboscore was the predictor of keratitis score (see online supplementary table). These results strongly indicated that patient with a higher D. brevis count tended to have more severe MGD, which was significantly correlated with keratitis severity.
Improvement of keratitis after lid scrub with Cliradex
We then investigated the possible correlation between demodicosis and keratitis in 48 of 60 patients who underwent Demodex mite-killing therapy. Results showed that both keratitis and ocular surface inflammation rapidly resolved within 2–3 weeks (Figures 3,4), including the two eyes with corneal perforation. At the end of 3 months of treatment, the Demodex count was significantly reduced to 0.5±0.7 (range 0–3, p<0.001). However, corneal scar remained in 35 eyes. Forty-six patients were followed up for at least 12 months, during which time four patients developed recurrent keratitis at 6, 8, 10 and 13 months, respectively, after discontinuation of treatment. They were all successfully controlled by the second run of Cliradex lid scrub. Interestingly, recurrent keratitis was noted in the ‘quiet’ fellow eye of three patients (representative case 2, figure 4).
The remaining 12 patients received baby shampoo lid hygiene and topical steroid and lubricants if needed. After treatment, the lid margin became clean and ocular surface inflammation was controlled temporarily, while Demodex count remained almost the same.
Representative case 1: coexistence of MGD and keratitis in one eye
A 6-year-old girl complained of redness and irritation in both eyes for 1 year. She was treated as HSK but failed to respond to antiviral and antibiotic therapies. On examination, her visual acuity was 20/300 and 20/40, respectively. Lash sampling revealed two D. brevis. The right eye had more blepharitis in the upper lid that was in contact with the corneal epithelial defect associated with stromal infiltrate and neovascularisation (Figure 3A). The left eye had milder blepharitis also worse in the upper lid with meibomian gland orifice plugging and inferior corneal SPK (Figure 3B). Meiboscore was four in the right eye (Figure 3E) and three in the left eye (Figure 3F). Inflammation was subsided in both eyes 2 weeks after lid scrub with Cliradex (Figure 3C; Figure 3D) together with healing of the keratitis and regression of vessels in the right eye (Figure 3C) and resolution of SPK in the left eye (Figure 3D). Her vision in both eyes was improved to 20/40 and 20/30, respectively. After 3 months of treatment, the Demodex count was 0 while both eyes were quiet.
Representative case 2: recurrence in a different eye
A 19-year-old woman complained of redness and irritation in both eyes and decreased vision in the right eye for 5 years. She was treated as HSK but failed to respond to antiviral and antibiotic therapies. On examination, her vision was 20/200 and 20/20, respectively. Lash sampling detected six D. brevis and two D. folliculorum. The right eye had blepharitis located at the same location of the corneal epithelial defect with stromal infiltration (figure 4A). The left eye was unremarkable except lash malalignment and irregular lid margin (figure 4B). Meibography showed meibomian gland loss that was more evident in both upper lids (figure 4E, F). After 2 weeks of treatment, the right keratitis and blepharitis subsided with dramatic regression of corneal vessels (figure 4C). Both eyes remained non-inflamed for 13 months until 1 day when the right eye remained stable while the left cornea presented with inferior stromal infiltration (figure 4D). Lash sampling revealed three D. brevis. This recurrent keratitis was controlled after 1 week of lid scrub with Cliradex.
Because both demodicosis and MGD are common in the elderly population,3 4 13 18 the potential causal relationship between the two is hard to resolve. Therefore, we focused on a cohort of young patients where both demodicosis and MGD are relatively rare. Our observational case series study discloses a strong correlation between ocular demodicosis especially by D. brevis and MGD and keratitis in young patients.
Univariate and multivariate analyses show that the severity of MGD was significantly correlated with demodicosis by D. brevis but not D. folliculorum in these young patients. This finding is consistent with the notion that D. brevis resides in sebaceous and meibomian glands and has been found to be significantly correlated with chalazia.11 We were surprised to note that demodicosis was associated with severe MGD that presented with more than 1/3 meibomian gland loss, especially in the upper lid, in such a young population. To further understand the impact of demodicosis on meibomian gland, a carefully designed case–control study or cohort study and disease animal model are warranted.
In six cases with ocular demodicosis, Kheirkhah et al,6 for the first time, reported corneal changes including superficial corneal vascularisation, marginal corneal infiltration, superficial corneal opacity, nodular corneal scar and phlyctenule-like lesion, suggesting the potential corneal involvement. Surprisingly, D. brevis is detected in four out of these six cases.6 Herein, we noted a variety of corneal changes in 39 out of 60 young patients with ocular demodicosis. Besides the aforementioned pathologies, we also detected SPK, limbitis, central infiltration and ulceration and even perforation. Hence, we surmise that ocular demodicosis should be considered in young patients presenting with sight-threatening keratitis.
The severity of keratitis was significantly correlated with that of MGD in our patients (figure 2 and table 1). Presumably, because of its predominant asymmetrical presentation (eg, figures 2 and 4), such keratitis could prompt one to suspect HSK as revealed in their histories. Our study disclosed the following clinical pearls that may help differentiate keratitis caused by demodicosis and that by herpetic infection. First, the history would suggest a refractory nature to antiviral, antimicrobial and topical steroid therapies but a rapid response by mite-killing lid scrub. Second, keratitis associated with demodicosis was strongly associated with ipsilateral blepharitis, lash malalignment and MGD, of which the upper lid tended to be worse than the lower. Third, keratitis associated with demodicosis, if recurred, could be in the contralateral eye.
The keratitis rapidly resolved after lid scrub with Cliradex, which contains terpinen-4-ol identified as the most active ingredient to kill mites.15 Recent studies revealed a potential microbial role in demodicosis. First, the mites may work as a vector carrying bacteria such as Staphylococci and Streptococci.19Second, mites can harbour symbiotic microbes in its intestine such as Bacillus oleronius, which can trigger a host immune reaction by producing proinflammatory proteins that can stimulate proliferation of peripheral blood mononuclear cells in patients with rosacea.20 We have reported a strong correlation among positive serum immunoreactivity to the 83 kDa and 62 kDa bacillus proteins, ocular Demodex infestation, facial rosacea and blepharitis.8 21 The microbial role is disclosed in ‘meibomitis-related keratoconjunctivitis’ reported by Suzuki and Kinoshita22 23 that exhibits similar clinical features such as blepharitis, chalazia, subepithelial corneal infiltration, corneal neovascularisation and conjunctival hyperaemia. They detected Propionibacterium acnes which normally inhabits human sebaceous follicles, in 60% of patients but 23.5% age-matched normal control subjects. Although systemic antimicrobial therapy is effective for curing ocular inflammation in meibomitis-related keratoconjunctivitis,22 23 refractory cases have also been reported. Their study did not disclose whether culture of P. acnes became negative on resolution of inflammation and whether there was concomitant demodicosis. Although we did not subject specimens to anaerobic microbial cultures, a number of our patients had not responded to systemic and/or topical antibiotics. Further study on the impact of demodicosis on the ocular surface microbiome and the potential pathogenic role between mites and microbes is warranted.
In summary, this prospective study demonstrates a strong correlation between ocular demodicosis especially by D. brevis and MGD and keratitis in young patients. Furthermore, we also summarise that the unique characteristics of MGD and keratitis associated with demodicosis can be differentiated from HSK.
Contributors LL and SCGT designed the study and reviewed the manuscript. LL, YL and HK researched and collected the data. LL, YL and XD wrote and edited the manuscript. XD analysed the data. CC contributed to discussion and reviewed the manuscript.
Funding This study is supported in part by a grant (81770892) from the National Natural Science Foundation of China, a grant (2014B020226003) from the Technological Project Foundation of Guangdong Province, a grant (201510010219) from the Technological Project Foundation of Guangzhou and a grant (R44 EY019586) from the National Eye Institute, the National Institutes of Health (to SCGT).
Disclaimer The sponsors or funding organisations had no role in the design or conduct of this research; collection, management, analysis and interpretation of the data and preparation, review or approval of the manuscript.
Competing interests SCGT has filed two patents for the use of tea tree oil and its ingredients for treating demodicosis. Cliradex is formulated by inclusion of the active ingredient identified through the support of grant R43 EY019586 (NEI, NIH).
Ethics approval This study is approved by the Ethics Committee of the Zhongshan Ophthalmic Center (Guangzhou, China).
Provenance and peer review Not commissioned; externally peer reviewed.
If you wish to reuse any or all of this article please use the link below which will take you to the Copyright Clearance Center’s RightsLink service. You will be able to get a quick price and instant permission to reuse the content in many different ways.