Efficacy and Safety evaluation of a single thermal pulsation system treatment (Lipiflow®) on meibomian gland dysfunction: a randomized controlled clinical trial

Evaluate the efficacy and safety of LipiFlow® thermal pulsation treatment system compared with lid massage combined warm compress in Chinese patients with meibomian gland dysfunction (MGD). Patients (n = 100 eyes, 50 subjects) diagnosed with MGD were recruited for this prospective, randomized, 3-month clinical trial. In Lipiflow group, patients (n = 50 eyes) received a single LipiFlow® thermal pulsation system treatment. In warm compress group, patients (n = 50 eyes) underwent warm compress daily for two weeks after an initial manual lid massage. Patients’ symptoms were evaluated using Standard Patient Evaluation for Eye Dryness (SPEED) questionnaire. Safety parameters included best-corrected visual acuity (BCVA), intraocular pressure (IOP) and objective parameters including meibomian glands yielding lipid secretion (MGYLS) number, meibomian glands secretion (MGS) score, lipid layer thickness (LLT), tear-film breakup time (TBUT), corneal fluorescein staining (CFS) were measured and presented from baseline and to 3 months post-treatment. Baseline parameters in both groups were comparable (p > 0.05). SPEED score and TBUT improved in two groups from baseline to 3 months. MGYLS number, MGS score, LLT improved in LipiFlow group and these improvements were maintained with no significant regression at 3 months. CFS showed significant improvement in warm compress group at 1 month compared with LipiFlow group. Moreover, the correlation analysis indicated LLT was positively correlated with TBUT, MGS score, and MGYLS number. A single 12-min LipiFlow treatment is an effective therapy for MGD patients and can achieve improvements in symptoms alleviation and meibomian gland lipid secretion function lasting for at least 3 months.

tear film instability, hyperosmolarity, ocular surface inflammation and damage, and neurosensory abnormalities [1]. As one of the most commonly encountered disorders in ophthalmology, DED causes great impact on patients' life quality and increases the social economic burden [2].
The classification of DED is generally classified into aqueous deficient dry eye and evaporative dry eye [3], and meibomian glands dysfunction (MGD) is a leading cause of evaporative dry eye. The incidence of MGD varies from 38 to 68% among the middleaged and elderly, and should not be underestimated in pediatric and young population due to the longterm visual display use [2,4]. Based on meibomian gland secretion (MGS), MGD is further subdivided into low-delivery (hyposecretory meibomian gland dysfunction (HMGD) and obstructive meibomian gland dysfunction (OMGD)) and high-delivery states [5]. OMGD, the common form of MGD, is characterized by terminal duct obstruction and qualitative or quantitative changes of secreted lipids [6], which contribute to tear film instability, tear excessive evaporation and the vicious cycle of DED [7][8][9].
The goal of MGD treatment is to relieve meibomian gland obstruction and restore tear film stability. Traditional treatments including warm compress, eyelids mechanical massage, artificial lubricants, topical pharmaceutical therapies (e.g., topical antibiotics or anti-inflammatory agents, Ciclosporin A (CsA), Diquafosol) have limitations due to the requirements of long-term treatment cycle and good patient compliance. In recent years, novel treatment devices including LipiFlow® system, Intense pulsed light (IPL), MiBoFlo Termoflo® provide more treatment options [10].
Some randomized controlled trials (RCTs) have demonstrated the effectiveness of LipiFlow for MGD in improving dry eye symptoms, tear-film breakup time (TBUT) and meibomian gland function [11][12][13][14]. Unfortunately, the application of LipiFlow treatment has been limited in most regions of China because of the high cost, and eyelid warm compress, artificial lubricants and topical pharmaceutical agents are still the most common treatments. Based on facts that the prevalence of MGD in Asians appears to be higher than white individuals [9,15], and a population-based survey showed a high rate of lid margin telangiectasia in Chinese (68.0% in asymptomatic DED; 69.3% in symptomatic DED) [16], a comprehensive effective evaluation of LipiFlow treatment in Chinese population is necessary, however, to our knowledge, reports on this topic are limited.
Therefore, the aim of our study is to investigate the efficacy and safety of LipiFlow ® thermal pulsation system treatment on both symptoms and objective ocular surface signs of patients with MGD in China.

Study design
This prospective, randomized, observer-masked clinical trial was approved by the institutional review board of the Tianjin Medical University Eye Hospital. This study was conducted in compliance with the tenets of the Helsinki Declaration. Informed consent to participate in this study was obtained from each eligible patient.

Study criteria
Participants were allowed to join the study if they met the following inclusion criteria: age ≥ 18 years, SPEED score ≥ 6 at baseline, MGS scores of either eye ≤ 12 by evaluating 15 meibomian glands of the lower eyelids [17], being willing to complete an initial evaluation and follow-up visits after treatment.
Participants were excluded if there was any evidence of: ocular inflammation, infection, trauma, surgery; abnormalities of ocular surface function or eyelid function; a history of systemic immune disease; used systemic or topical glucocorticoids, immunomodulators within 3 months; have plans to receive ocular surgeries (e.g., cataract, myopic refractive surgery) within 3 months.

Interventions
Fifty participants who met the criteria at the Tianjin Medical University Eye Hospital were recruited and randomized into single LipiFlow treatment group (n = 50 eyes) and warm compress treatment group (n = 50 eyes) using a randomized digital table by the randomization manager. In LipiFlow treatment group, participants underwent a single LipiFlow® thermal pulsation system (TearScience Inc., Morrisville, NC, USA) treatment on the first visit: after instillation of anesthetic eye drops (Alcaine, proparacaine hydrochloride 5.0 mg/mL) in both eyes, sterile eye cups were placed on to the conjunctival sac as instructed by the manufacturer, after 12 min of upper and lower palpebral conjunctival surfaces heat while simultaneously graded pulsatile pressure applying, eye cups were removed slightly. In the warm compress group, participants received an initial warm compress and manual lid massage on the first visit, then started 15 min warm compress using the EyeGiene® Insta-Warmth™ patch (Eyedetec Medical, Inc., Danville, CA) after lid hygiene with wet cotton swabs daily for two weeks. Participants in any groups with a history of ocular lubricants use were allowed to continue after inclusion into the trial.

Study parameters
Baseline characteristics including age, gender, bestcorrected visual acuity (BCVA) and noncontact intraocular pressure (IOP) were recorded. The values of BCVA and IOP were presented as safety parameters at the end of follow-up.
SPEED questionnaire was used to evaluate subjective symptoms [18]. The total scores calculated as the sum of the frequency and severity scores for dry eye symptoms range from 0 (no symptoms) to 28 (severe symptoms).
Lipid layer thickness (LLT) of tear film and partial blinking (PB) rate were measured at LipiView ® Ocular Surface Interferometer (TearScience Inc., Morrisville, USA). Both eyes of participants were video-recorded for 20 s to capture the interference pattern of the tear film, and then the values of interferometric color units (ICUs) analyzed from patterns convert to nanometers of LLT (1 ICU approximately reflects 1 nm of the LLT). Simultaneously, the numbers of PB and total eye blinkings were recorded during the examination.
The measurements of tear-film breakup time (TBUT) and corneal fluorescein staining (CFS) used strips of fluorescein sodium (Jingming New Technological Development Co., Ltd., Tianjin, China). The tip of a strip was applied in the inferior conjunctiva to distribute fluorescein over the whole cornea uniformly, observation and calculation were completed under slit-lamp with cobalt blue illumination. The average of three consecutive calculations of TBUT was adopted as the final value, and CFS was scored as follows: the staining numbers of corneal superficial defects in each quadrant (supertemporal, inferotemporal, supranasal, and inferonasal) were scored from 0 to 3 points, no staining corresponds to 0, < 5 defects correspond to 1 point, 5-10 defects correspond to 2 points, > 10 defects correspond to 3 points [19].
The function of meibomian glands was quantified at the Meibomian Gland Evaluator (TearScience Inc., Morrisville, USA) as described by Kord and Blackie [20]. In brief, the hand-held instrument with a defined pressure was applied to nasal, central, and temporal regions of the lower eyelid, each region contained five consecutive meibomian gland orifices, a total of 15 glands were evaluated along the lower eyelid margin. The number of meibomian glands yielding liquid secretion (MGYLS) was counted, and the secretion quality was scored 0 to 3: 0 indicates no secretion, 1 point indicates toothpaste-like secretion, 2 points indicates opaque secretion, 3 points indicates clear secretion [14]. The total meibomian gland secretion (MGS) score was calculated as the sum of the points of 15 glands (the maximum value = 45 points). MG atrophy was observed at Keratograph 5 M (Oculus, Wetzlar, Germany) by capturing high-contrast infrared meibography images, subsequently, the proportion of MG dropout was calculated by ImageJ software (National Institutes of Health, Bethesda, Maryland) for upper and lower eyelids. (Fig. 1).
The evaluation parameters before treatments were collected as the baseline, and follow-up examinations were arranged at 1 month, 2 months and 3 months post-enrollment. Researchers who were responsible for data analysis were blinded to participant distribution.

Statistical analysis
Data analysis was performed using SPSS software for Windows version 17.0 (SPSS Inc., Chicago, IL, USA) and GraphPad Prism (GraphPad Software, La Jolla, CA). Enumeration data was performed with the chi-square test. Paired samples t test and independent samples t test were performed to compare the parameters within groups and intergroup, respectively. Repeated measures analysis of variance (ANOVA) was used to evaluate the effects of the two therapies on parameters before and after treatments. The linear Pearson correlation coefficient was used for correlation analysis. All outcomes were analysed with 95% confidance intervals, p value of < 0.05 was considered as statistically significant.

Results
Participants in both groups completed treatments and monthly follow-ups. LipiFlow treatment group (50 eyes) included 12 male and 13 female with an average age of (57.8 ± 12.2) years, warm compress treatment group (50 eyes) included 14 male and 11 female with an average age of (58.4 ± 12.9) years (p gender = 0.571, p age = 0.960). Parameters of baseline evaluation were comparable between two groups. (Table 1).

Safety assessment
There was no cornea or ocular surface damage in participants after LipiFlow treatment. BCVA and IOP, the safety indicators, showed no statistically change in LipiFlow group before and 3 months after treatment. (Table 2). Fig. 1 The semi-automated ImageJ tool is used to draw the borders of the total MG area (upper) and MG dropout area (lower) manually BCVA best-corrected visual acuity, IOP noncontact intraocular pressure.

Correlation analysis
The correlation analysis between tear-film LLT and other parameters including TBUT, MGYLS number and MGS scores are summarized in Table 3. The results showed LLT was positively correlated with TBUT, MGYLS number, MGS scores at baseline and each visit after treatments in all participants.

Discussion
This clinical trial was designed to evaluate the efficacy and safety of LipiFlow ® thermal pulsation system in Chinese MGD patients, including comprehensive ocular surface parameters with an attempt to investigate which would be improved significantly. Considering the effectiveness of warm compress rely on patients' compliance and a long treatment Fig. 4 Differences between two groups were labeled above the lines, MGS (A) and MGYLS numbers (B) showed statistical differences at two-and three-month. (*p < 0.05, **p < 0.01, ***p < 0.001) cycle, which are both difficult to be guaranteed, the participants in our control group underwent warm compress therapy for only two weeks instead of the entire study duration to model a real-world treatment status.
In this study, we found a single LipiFlow treatment and lid massage plus warm compress therapy both contribute to the improvements on SPEED scores and TBUT, which are consistent with previous reports [11][12][13]. These changes demonstrated warm compress with appropriate temperature (over 40℃) and sufficient therapeutic time [21,22] is still an effective choice for MGD, but the improvements in LipiFlow group were statistically greater at 2 and 3 months after treatment, we speculate it may be due to the recovery of meibomian glands function and this indicates LipiFlow® thermal pulsation system has the potential to consolidate a longer therapeutic effect on dry eye symptoms and tear-film stability.
MGS score and MGYLS number are related to the quality and quantity of lipid secretions, so they are essential on the assessment of meibomian gland function. The statistical difference between two treatment groups may be attributed to the operating principle of the LipiFlow ® system device, which safely delivers a therapeutic level of heat (42.5℃) to the surfaces of upper and lower eyelids directly over the meibomian glands, whereas warm compress devices transfer heat through eyelids skin, resulting in a decrease of ability to melt meibum and expel secretions [23,24]. Moreover, LipiFlow grades pulsatile pressure simultaneously evacuates the glands during heating, realizing a fully discharge of abnormal stored secretions in lower meibomian glands, subsequently regenerating new ones. Thus, the previously blocked dysfunctional meibomian glands get a chance to function properly, which will contribute to the ocular surface homeostasis.
Although we detected the improvements in meibomian glands function, we were unable to match the finding of glands structure increased after LipiFlow treatment relative to untreated controls reported by Hura et al. [25]. Given the lack of methods to revitalize atrophic glands currently, we figure that meibography evaluation is necessary for guiding MGD therapies and screening condition changes, interventions like LipiFlow at early stage may prolong the maintenance of normal structure and function of meibomian glands, but other factors including symptoms severity, glands structure, patient compliance and financial capacity should all be taken into the final consideration due to the rather high expenses of LipiFlow treatment.
It is informative to mention that LLT had a statistical increase in LipiFlow treatment group at each follow-up visit, which was not in warm compress group. Previous studies showed that LLT is related to subjective symptoms (SPEED), expressible meibomian glands numbers and meibomian glands losses [26][27][28]. And Isreb et al. reported a firm positive correlation among LLT, fluorescein tear film break-up time and Schirmer's test with anesthesia [29]. Here we found that LLT was positively correlated with TBUT, MGYLS number and MGS score in the whole research group. Lipids from the meibomian glands play an important role in the maintenance of the tear film's ocular surface, and form the most superficial layer which protects against excessive dehydration. The correlation analysis results confirm the improved meibomian glands function can make for the decrease in aqueous evaporation and relief symptoms like eye dryness, pain or foreign body sensation, which is consistent with SPEED scores outcome.
CFS scores at 1 month post-treatment in warm compress group are statistical better than in LipiFlow group, we speculate that operations of LipiFlow might cause extra but slight damage to ocular surface, which suggests practitioners to operate gently as much as possible during the treatment. No participant in either group experienced any adverse events involving damages in visual acuity or IOP throughout the treatment.
The main advantage of our study is the accomplishment of all participants' monthly visits and overall ocular surface reassessments. Furthermore, we added analysis of the correlation between LLT and MGS score, which further demonstrated the quality of secretions is vital to MGD manifestation and prognosis.
Our study has some limitations. The small number of recruited participants and the short duration of follow-up might weaken the reliability of results, long-term studies are needed to prove the efficacy and applied value of LipiFlow in future. Additionally, truly double-blind trail was hard to forward as participants could not be masked to the type of treatment, which might affect symptom evaluation. And observers in this study were masked to minimize bias largely. Measurement bias of some parameters is inevitable such as PB rate, considering the possibility of voluntary eyelids movement, particularly when patients have been instructed how to make an efficacious blink, analysis of PB rate in a small sample size study cannot explain the effective of different treatments, though the rate and duration of blinks are related to DED [30,31].

Conclusion
A single 12-min LipiFlow treatment can ameliorate subjective symptoms of MGD and markedly improve meibomian glands secretion function lasting for at least 3 months. It is necessary to perform meibography examination to guide treatment decisions, patients with structurally abnormal eyelid margins and meibomian glands or those who respond poorly to lid massage plus warm compress therapy, LipiFlow is an efficient and safe treatment option to be recommended.
Author's contribution Zhu Meng and Xiaoran Chu contributed equally to this work and should be regarded as co-first authors. Zhu Meng: study design, data curation, original draft writing; Xiaoran Chu: manuscript revision & edit, data analysis; Chen Zhang, Hui Liu, Ruibo Yang, Yue Huang: data collection & interpretation; Shaozhen Zhao: project supervision, study design & conception, data curation, manuscript revision. All authors finally approved the manuscript and agreed to be accountable.
Funding This work was funded by Natural Science Foundation of China (No.81970769) and Tianjin Key Medical Discipline (Specialty) Construction Project (TJYXZDXK-037A).

Data availability
The datasets generated during and analyzed during the current study are not public but are available from the corresponding author on reasonable request.

Declarations
Conflict of interest Zhu Meng, Xiaoran Chu, Chen Zhang, Hui Liu, Ruibo Yang, Yue Huang, Shaozhen Zhao declare that they have no conflict of interest.
Ethical approval This study was reviewed by the institutional review board of the Tianjin Medical University Eye Hospital (IRB No. 2016KY(L)-22) and followed the tenets of the Declaration of Helsinki.
Informed consent Informed consent was obtained from all individual participants included in the study.