Background/aims To assess the efficacy and safety of oral azithromycin compared with oral doxycycline in patients with meibomian gland dysfunction (MGD) who had failed to respond to prior conservative management.
Methods 110 patients (>12 years old) with MGD were randomly assigned to receive either oral 5-day azithromycin (500 mg on day 1 and then 250 mg/day) or 1-month doxycycline (200 mg/day). They also continued eyelid warming/cleaning and artificial tears. A score comprising five symptoms and seven signs (primary outcome) was recorded prior to treatment and at 1 week, and 1 and 2 months after treatment. Total score was the sum of both scores at each follow-up. Side effects were recorded and overall clinical improvement was categorised as excellent, good, fair or poor based on the percentage of change in the total score.
Results Symptoms and signs improved significantly in both groups (p=0.001). While improvement of symptoms was not different between the groups, bulbar conjunctival redness (p=0.004) and ocular surface staining (p=0.01) were significantly better in the azithromycin group. The azithromycin group showed a significantly better overall clinical response (p=0.01). Mild gastrointestinal side effects were not significantly different between the groups except for the second visit, when the doxycycline group had significantly more side effects (p=0.002).
Conclusions Although both oral azithromycin and doxycycline improved the symptoms of MGD, 5-day oral azithromycin is recommended for its better effect on improving the signs, better overall clinical response and shorter duration of treatment.
Trial registration number NCT01783860.
- Ocular Surface
- Treatment Medical
- Eye Lids
Statistics from Altmetric.com
Meibomian glands have an important role in the health of the ocular surface by secreting a lipid and protein mixture into the tear film; this prevents evaporation of the tear film,1 ,2 even though the role of protein is less clear. Meibomian gland dysfunction (MGD) compromises the tear film and meibum lipids which results in evaporative dry eye.3 ,4 MGD is an extremely common and important cause of posterior blepharitis, the incidence of which is probably underestimated.5 ,6
The leading cause of MGD is obstruction of MG secondary to hyperkeratinisation of the duct epithelium and accumulation of meibum, resulting in inflammation and possibly an increased bacterial colonisation of the lid margins.7 MGD may be asymptomatic, only detectable by gland expression, or, more often, present with dry eye symptoms.8 Common symptoms are foreign body and burning sensation, red eye, tearing and photophobia.6 Common signs are loss of clarity and thickening of expressed meibum, pouting or plugging of meibomian gland orifices, meibomian gland drop out detectable by meibography, increased eyelid margin thickness and vascularity, eyelash loss, trichiasis and vascular invasion.9
Most cases usually require conservative management including warm compresses to provide appropriate meibum secretion, mechanical eyelid massage and cleansing with shampoo and cotton buds to remove excess debris, and artificial tears to continuously lubricate the ocular surface.10 ,11 In severe and refractory cases, however, antibiotics (topical and systemic) with anti-inflammatory properties are proposed.12 ,13 Tetracycline and its derivatives are antimicrobials with the ability to decrease inflammation and inhibit matrix metalloproteinases.14 Doxycycline is a long acting analogue of tetracycline which has been used to treat MGD through its antimicrobial, anti-inflammatory and anti-metalloproteinase properties, with fewer side effects than tetracycline.6 ,15 A few studies have shown the efficacy of azithromycin in chronic inflammatory diseases such as MGD.16 ,17 Azithromycin inhibits pro-inflammatory cytokines and is potent against Gram-negative microorganisms.14 Topical and oral azithromycin have recently been reported to improve the sign and symptoms of MGD and posterior blepharitis.12 ,14
Although previous studies have shown the efficacy of both oral doxycycline and oral azithromycin in the treatment of MGD, to the best of our knowledge there has been no study comparing their effects and side effects. Therefore, this randomised double masked open label clinical trial was designed to evaluate the efficacy (symptom and sign scores) and safety (side effects) of oral azithromycin compared with oral doxycycline in patients with posterior blepharitis who were unresponsive to the conservative management (eyelid warming/massage/cleaning and artificial tears).
Materials and methods
Sample size was calculated to detect at least 1.8 differences12 ,14 between the severity scores of two groups across the time periods of the first and last follow-up visit, considering type one error of 0.05 and power of 80%. This was 43 patients in each group. In order to overcome the problem of loss to follow-up, the aim was to recruit 55 patients in each group (15% more than the calculated sample size).
Block randomisation (four subjects per block) was used to assign patients to the treatment groups. Simple randomisation was used to assign the eye to the patient in the treatment group by writing digit numbers (1 to 110) and right/left words on separate sealed papers which were then randomly chosen (figure 1). One of the masked observers (AJF) secured the sealed papers and another (VK) performed the examination and scoring in all visits (figure 1).
All participants signed an informed consent. The Tehran University of Medical Sciences ethics committee approved the study. The study was performed at the Rassoul Akram Hospital. It was registered and released with ClinicalTrials.gov (NCT01783860). This research adhered to the tenets of the Declaration of Helsinki.
Included were consecutive patients (over the age of 12 years) with posterior blepharitis who had not responded to conservative management: eyelid warming/massage/cleaning (4–5 min) twice a day and artificial tears (four times a day). Exclusion criteria were: taking systemic or topical antibiotic within 1 month prior to inclusion, history of liver disease, pregnancy and breast feeding, contact lens wearing, allergy to azithromycin or cyclines, vernal and atopic keratoconjunctivitis, ocular and ocular adnexal surgery, altered lid anatomy for any reason, and incomplete follow-up (missing any of three pre-scheduled visits).
Severity of five main symptoms was measured on a 4-point categorical scale (0–3) according to patients’ response to questions: itching, burning, foreign body sensation, dryness and eyelid swelling (table 1). Slit lamp examination was performed to assess and record the severity of seven signs on a 4-point categorical scale: meibomian gland secretion, number of plugged gland orifices, conjunctival injection, lid margin redness, lid margin debris, tear break up time (TBUT), and ocular surface staining with fluorescein (table 1).
Meibomian gland expression was performed (digital pressure) on the central third of the lower eyelid and graded on a 0–3 scale based on the worst type of secretion (clear, cloudy, turbid clumps, solid paste).
Meibomian gland plugging was graded as follows: 0, clear orifice of meibomian glands in the central part of the lower eyelid; 1, less than one-third of the orifices contained turbid or oily secretion; 2, between one-third and two-thirds of the orifices contained turbid or oily secretion; 3, more than two-thirds of the orifices contained turbid or oily secretion.
Bulbar conjunctival redness was graded as none, pink, light red, and bright red based on slit lamp examination.
Lid margin redness was also graded as none, pink, light red, and bright red based on the colour of the lower eyelid margin on slit lamp examination.
Lid margin debris was scored based on the number of crusts at the lower eyelid margin, included in this study even though it is mostly present in anterior blepharitis.
TBUT was performed before ocular surface staining. Fluorescein impregnated (single drop of saline) paper was placed in the lower fornix, and patient was asked to gently blink a few times and then stop blinking while being observed (blue light of slit lamp) for the first break on the corneal surface tear film. The time of the first split was recorded and graded on a 4-point categorical scale: grade 0 (>10 s), 1 (8–10s), 2 (5–7s), and 3 (<5 s).
Ocular surface staining was immediately performed after recording the TBUT using a 4-point categorical scale panel (figure 2), which was a modification of panels introduced in the Oxford grading scale18 in order to be compatible with grading of the other signs. The patient's upper eyelid was raised to observe the entire cornea, and the temporal and nasal interpalpebral conjunctiva were assessed by asking the patient to look nasally and temporally, respectively. The panel which best represented the number and pattern of dots on the cornea and conjunctiva was selected and corresponding grade was recorded.
MGD was diagnosed based on having at least two symptoms and two signs (one must be the presence of meibomian gland signs) with a minimum severity score of 2 for each. All pretreatment and post-treatment assessment and examinations were performed by one observer who was masked to the type of treatment. Patients were randomly (block randomisation of four in each block) assigned to either a 5-day oral azithromycin (Kimidarou Co., Tehran, Iran) course (500 mg on the first day and then 250 mg per day for the following four days) or a 1-month oral doxycycline (Irandarou Co., Tehran, Iran) course (100 mg twice a day). Doxycycline was instructed to be taken with a full glass of water while sitting for a few hours before going to bed and keeping a couple of hours between any supplement and doxycycline. Patients were also instructed to continue conservative management: eyelid warming/massage/cleaning twice a day and artificial tears four times a day throughout the study.
The symptom and sign scores were recorded prior to treatment and three times after treatment: first visit (day 12; 1 week after 5-day azithromycin), second visit (day 36–37; 1 month after 5-day azithromycin and 1 week after 1-month doxycycline), and third visit (day 60; 1 month after 1-month doxycycline and 7 weeks after 5-day azithromycin) (figure 1). The original plan was to have four post-treatment visits in which the second (1 month after 5-day azithromycin) and third (1 week after 1-month doxycycline) visits were a couple of days apart. Since they were close to each other, the second visit was considered to be 1 month after 5-day azithromycin and 1 week after 1-month doxycycline.
Each patient's symptoms or signs were given a score of 0 to 3. The symptom score of each subject was calculated by adding the score (0–3) of five symptoms which resulted in a range of 0–15. The sign score of each patient was also calculated by adding the score (0–3) of seven signs which resulted in a range of 0–21. The total score (0–36) of each patent was calculated and recorded by adding the scores of symptoms (0–15) and signs (0–21) at each visit. Overall clinical responses were categorised based on the percentage of reduction of total score into four groups: excellent (76–100%), good (51–75%), fair (26–50%), and poor (1–25%). Side effects of treatments were also recorded at each visit.
The χ2 and Student t-test were used to compare demographic characteristics across the two intervention groups. The Student t-test was used for comparing symptoms, signs and total mean scores. The χ2 test was used to analyse the overall clinical response and side effects. A p value of less than 0.05 was considered as statistically significant and 95% CI was used as a measure of precision. Analysis was performed by a masked observer (MN).
There were 110 patients, of whom 10 did not complete the study according to the protocol (figure 1). Therefore, included for analysis were 100 patients (100 eyes) who were studied between February 2013 and July 2013. The two groups were matched for demographics (table 2).
Symptoms significantly improved in both groups (p=0.001, 95% CI −2.2 to −0.7). However, there was no significant difference between the groups at all post-treatment visits (table 3).
Both groups showed a significant improvement of clinical signs (p=0.001, 95% CI −2.8 to −0.8). Comparison of the second and third follow-up for each group showed only a significant improvement of mean sign score in the azithromycin group (p=0.01) and no significant difference in the other scores. While mean sign scores were not significantly different at the pretreatment and first post-treatment visits, they were significantly less (better response) in the azithromycin group at the second and third visits and consequently the mean total score was better at the same times (table 3). Although the last follow-up mean score of all seven signs was less (better response) in the azithromycin than in the doxycycline group, the difference was statistically significant in only conjunctival redness and ocular surface staining (table 4).
While the azithromycin group showed excellent (6%) and good (46%) responses in 52% of patients, the doxycycline group had an excellent response in none and a good response in 32% of patients (p=0.01). On the other hand, while there were more fair responses in the azithromycin (38%) than in the doxycycline group (34%), there were three times more poor responses in the doxycycline (34%) than in the azithromycin (10%) group.
Both groups had a few gastrointestinal side effects (nausea, abdominal cramp, diarrhoea and decreased appetite), which were not significantly different at all follow-up visits except for the second visit (immediately after termination of the doxycycline course), when there were significantly more side effects in the doxycycline (13/50, 26%) than in the azithromycin (2/50, 4%) group (table 5). All were temporary and did not result in stopping the medications.
This is, to the best of our knowledge, the first randomised clinical trial comparing oral doxycycline and azithromycin in the treatment of patients with MGD, in which there was a significant improvement of symptoms and signs in both groups. There was, however, no significant difference in improvement of patients’ symptoms between the two groups at the last follow-up visit (table 3). Although the last follow-up mean score of all clinical signs was better in the azithromycin group, a statistically significant difference was only observed for conjunctival redness and ocular surface staining (table 4). In general, the percentage of clinical improvement was significantly better in the azithromycin group at the last follow-up.
While the pathogenesis of MGD is still unclear and complex, both inflammation and bacterial colonisation are, to a different extent, playing a role.14 ,19 Although conservative management may be effective in the management of MGD, oral antibiotic/anti-inflammatory treatment is suggested when the improvement is slow or incomplete.12 More than normal colonisation of different bacteria synthesise lipolytic enzymes which produce irritating fatty acids and cholesterol to create tear film instability and inflammation.12 It is believed that some antibiotics with anti-inflammatory effects may help to control the bacterial colonisation and also eyelid inflammation.7 ,12 ,14 ,16 ,20
Although the role of doxycycline in the treatment of MGD has been shown previously,21 its side effects and subsequently low compliance of the patients sometimes result in treatment abortion. Its side effects include dermatological effects (macula, papules), gastrointestinal effects (oesophagitis, dyspepsia, diarrhoea, vomiting), and hypersensitivity (allergy, urticaria).20 However, this study showed a few gastrointestinal side effects in the doxycycline group. They were reported by 15 patients (30%) in the first week after starting the doxycycline, decreased to 26% (13/50) at the end of the 1-month treatment course, and then to 14% (7/50) at the end of the study.
Azithromycin has also been recently introduced; used in either its topical14 ,16 or oral12 ,22 forms it has been found to be effective in the treatment of MGD. Azithromycin is a semisynthetic macrolide antibiotic with a good intracellular penetration and long half-life; it has an anti-inflammatory and immunomodulatory effect that could help reduce eyelid and ocular surface inflammation.12 The oral formulation of azithromycin can suppress the production of some pro-inflammatory mediators like cytokines (TNFα, IL-1β), chemokines and metalloproteinases (MMP-1, MMP-3 and MMP-9).22 Its antibacterial and anti-inflammatory effects, coupled with a long half-life, make oral azithromycin a good treatment option for MGD.12 Results of this randomised clinical trial support the results of a recent publication on the beneficial effects of oral azithromycin on symptoms and signs of patients with posterior blepharitis.12 Such an effect has also been reported using topical azithromycin.16 The better effect of azithromycin on the symptoms and signs (table 3) in this trial could be attributed to a better antibacterial effect, inhibition of bacterial lipases and degradation of meibomian gland lipid, and/or its anti-inflammatory effects. However, we could not account for the fact that a significantly better improvement was observed in conjunctival redness and ocular surface staining (table 4).
Bakar et al23 reported that the side effects of systemic azithromycin were minimal and well tolerated in most patients treated for populo-pustular rosacea. Likewise, this series showed mild and temporary side effects which did not require treatment to be discontinued. The most common side effect was decreased appetite, which has also been reported by Greene et al.22 We, however, could not account for why three of the azithromycin patients still had decreased appetite 7 weeks after stopping the medication (table 5).
The cost of 1-month treatment with doxycycline is almost 50% more than 5-day treatment with oral azithromycin in our country. Since MGD is a chronic disease, multiple 5-day pulse treatment with azithromycin would be cheaper than long term daily oral doxycycline.
Strengths of this study were its randomised masked trial design, using a grading system to quantify the qualitative measures, and maintaining the conservative management of the MGD in both groups. One limitation might be the absence of a control group without any systemic medication; there has been debate24 on whether clinical trials essentially require a placebo control instead of having an active treatment control. However, since enrolled patients had not responded to conservative management and had moderate to severe MGD (two signs and two symptoms of at least grade 2 severity, of which one had to be meibomian gland signs), systemic medication (doxycycline or azithromycin) was deemed to be required. Another limitation would be considering all signs (meibomian gland, eyelid and ocular surface signs) with the same value in calculation and analysis of the mean sign score. To overcome such a limitation, a separate analysis was also performed to evaluate the change in each sign (table 4) after different treatments. Using one dosing regimen of oral azithromycin and doxycycline and not considering topical azithromycin is another limitation, which future studies are recommended to address.
In conclusion, this randomised double masked open label clinical trial demonstrated beneficial effects for both oral doxycycline and azithromycin in patients with MGD. The azithromycin group, however, had a relatively better effect with regard to symptom (statistically insignificant) and sign (statistically significant) scores and fewer side effects. Therefore, a 5-day course of oral azithromycin is recommended based on better clinical improvement, shorter duration of treatment, lower cost and less side effects. Since MGD is a chronic disease, a longer term regime (repeating the 5-day treatment course) would be often required.
Future studies are recommended to assess different doses of oral azithromycin, topical versus oral azithromycin, and the number and time interval of oral azithromycin courses in patients with MGD.
Contributors Conception or design of the work (MBK), the acquisition, analysis or interpretation of data (AJF, VK, MN, LG), drafting the work or revising it critically for important intellectual content (MBK, LG), and final approval of the version published (MBK). All agreed to be accountable for all aspects of the work in ensuring that questions related to the accuracy or integrity of any part of the work are appropriately investigated and resolved.
Funding This study was financially supported by the Eye Research Center, Iran University of Medical Sciences.
Competing interests None.
Ethics approval Tehran University of Medical Sciences ethic committee (no. 90-01-124-13076).
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.