Participants were recruited between July 2017 and October 2017 at the Department of Ophthalmology, Eye and ENT Hospital of Fudan University (Shanghai, China). The study adhered to the tenets of the Declaration of Helsinki and was approved by the Ethical Committee of the Fudan University Eye and ENT Hospital Review Board. All participants were fully informed and gave written consent for publication of this information.
We enrolled 37 right eyes of 37 participants (age 29.5 ± 6.3 years, 14 male and 23 female) who had experienced dry eye for more than 2 years following SMILE or LASEK. Among them, 25 underwent SMILE, and the remaining 12 underwent LASEK. All patients were routinely screened preoperatively and met the criteria for SMILE or LASEK, and all surgeries went smoothly and were performed by the same experienced surgeon (XTZ). We excluded patients who experienced new post-surgical severe ocular or systemic diseases or other long-term complications except dry eye.
We used the VisuMax femtosecond laser system (Carl Zeiss Meditec, Jena, Germany) to perform SMILE and the Mel-80 excimer laser system (Carl Zeiss Meditec AG, Jena, Germany) to perform was used to perform LASEK. We previously describe these procedures in detail. 8-9 The femtosecond laser settings were as follows: 500 kHz repetition rate, 130 nJ pulse energy, 110 to 120 µm intended cap thickness, 6 to 6.5 mm optical zone, 7.3 to 7.5 mm cap diameter, and a 2 mm side cut at the 12 o’clock position.
Safety and effectiveness parameters were assessed sequentially before and at 5 min after WC. Safety parameters: corneal curvature (K1, K2, Km) and central corneal thickness (CCT) measured by Pentacam HR (Oculus GmbH, Wetzlar, Germany), uncorrected distance visual acuity (UDVA), corrected distance visual acuity (CDVA) and spherical equivalent (SE). Effectiveness parameters: tear film lipid layer thickness (TFLLT), blink frequency, Meibomian secretary function scores, and tear film break-up time (TBUT). All patients were instructed to stop using eye drops such as artificial tears or wearing contact lenses 24 hours before examinations, and to avoid skin care products or make-up around the eyes. All examinations were performed between 11:00 am and 17:00 pm in the same room by a single pre-trained and experienced technician. The indoor temperature was set to 22 - 25 °C.
TBUT was measured using fluorescein sodium solution and a digital clock. The averages of three repeated measurements were recorded as final values.
All patients were treated for 20 minutes with a warm compress using a spontaneous heating eye mask (Zhenshiming Pharmaceutical Co., Ltd, Fuzhou, Jiangxi, CHN) according to the manufacturer's instruction. This eye mask is a standardized, fragrance free, over-the-counter eyeshade, certified after examination and verification by Shanghai Institute of Quality Inspection and Technical Research. At an ambient temperature of 25 ± 2 °C, it takes an average of 3 min to warm up to 35 °C, and maintains temperature over 35 °C for an average of 30 min. The mean temperature was 40.7 °C.
An interferometer (LipiView, TearScience® Inc, Morrisville, NC) was used for measuring the quantitative TFLLT and blink frequency dynamically by analyzing more than one million tear film data points. TFLLT was calculated using interferometry color described as interferometric color units (ICUs), which is equivalent to 1 nanometer (nm). Patients were instructed to hold their head in a comfortable position and look directly into the camera without deviation of eye position, and to blink freely throughout imaging. Natural light from the source passed through the tear film and was reflected back to the camera. The measurement region was the lower third of the cornea, approximately 1 mm above the inferior tear meniscus. The minimum (Min-), maximum (Max-), average (Ave-) and standard deviation (-std) TFLLT were analyzed automatically within preset 19.1 s. Simultaneously, total (TBF) and partial blink frequency (PBF) were also recorded, and were used to calculate partial blink rate (PBR) and complete blink rate (CBR). The upper cut-off of LipiView is 100 ICU; values higher than this are recorded as 100+ ICU. A conformance factor for imaging ≥ 0.7 was considered acceptable, but to ensure quality of our data, only conformance factor ≥ 0.8 were included in the final analysis.
Meibomian gland evaluator
Secretory function of the Meibomian glands was measured by a handheld Meibomian Gland Evaluator (TearScience® Inc, Morrisville, NC) and recorded as Meibomian gland scores. Gentle and blink-stimulated pressure (0.8 g/mm2 - 1.2 g/mm2) was applied to the Meibomian gland openings along the lower eyelid margin. The width of the evaluator covers an average of 5 gland openings. After gentle clean of the lower eyelid using clean cotton swabs, the evaluator was applied 1-2 mm below and in parallel with the eyelid, near the root of the eyelash. The eyelid was gently turned inside out slightly just until the gland opening was clearly visible. Fifteen glands in 3 regions (temporal, central, and nasal) were evaluated, with an average pressing time of 10-15 seconds. We graded the number and secretion characteristics of glands: 3 points (clear liquid), 2 points (cloudy liquid), 1 point (toothpaste-like), and 0 (no secretion). The following metrics were calculated: the total Meibomian gland secretion score (TMGS) of all 15 glands, ranging from 0 to 45; the number of glands secreting any liquid (MGL, clear or cloudy liquid, grade 2 or 3); and the number of glands secreting clear liquid (MGC, clear liquid, grade 3).
Statistical analysis was performed using SPSS ver. 22.0 (SPSS Inc, Chicago, IL, USA). Variables were described as averages ± standard deviations. The one-sample Kolmogorov-Smirnov test was used to test for normality. Paired t-tests or Wilcoxon tests were then used to analyze the changes from baseline in normal or non-normal distribution parameters, respectively. Variables were compared between SMILE and LASEK using independent sample t-tests or Mann-Whitney U tests. Pearson or Spearman correlation analyses were used to evaluate the linear relationships between different variables. P < 0.05 was considered statistically significant.