A total of 191 patients’ right eyes were enrolled in the study, including 95 females (49.7%) and 96 males (50.3%), with a median age of 53 years (range: 28–85 years). We collected the meibum before the first treatment (T0) and the third treatment (T2) All enrolled patients completed the whole examination and treatment; there was no pain or any discomfort reported during the tests. Among the patients, we chose 26 patients randomly to do the lipid analysis using LC-MS/MS. And we also recruited 10 volunteers who had no ocular surface disease and other systemic disease (The work plan is illustrated as follows) (Table 1).
Tear film and ocular surface analysis
Tear meniscus height was used to evaluate the quantity of the tear film. Following treatment with IPL, the tear meniscus height was markedly increased from 0.1658±0.003853 to 0.1818±0.004657 (P<0.01, t=2.642, df=367.1). At the same time, the BUT and average tear BUT (aBUT) were used evaluated the quality of the tear film; both increased following treatment from 3.90±0.10 to 5.02±0.10 (P<0.0001, t=7.9, df=380) and from 4.93±0.12 to 5.36±0.10 (P<0.01, t=2.774, df=374.5), respectively. Intriguingly, the corneal fluorescein staining (FL) of corneal epithelium integration and the bulbar redness of conjunctival (CR) inflammation were markedly improved, The FL score decrease from 0.43 ± 0.05 to 0.20 ± 0.03 (P<0.0001, t=4.147, df=380) and the CR scorce reduce from 1.24±0.04 to 0.59±0.03 (P<0.0001, t=12.95, df=380), respectively (Table 2).
Eyelid margin abnormalities
As previously reported, IPL may improve the eyelid margin; our findings suggest that there was no statistically significant difference after the IPL treatment in terms of orifice abnormality and MG dropout (from 0.9686±0.04571 to 0.9319±0.04625, P=0.5733, t=0.5636, df=380 and from 1.314±0.05266 to 1.314±0.05266, P>0.9999, t=0, df=380, respectively) (Table 2).
MGE and meibum quality
The scores of MGE and meibum quality were used to evaluate the quality and expressibility of the meibum. Following IPL treatment, the meibum changes from solid state to liquid state and becomes clear and easy to eliminate, the MGE score and meibum quality score were significantly decreased (from 1.34±0.06 to 0.94±0.04, P<0.0001, t=5.798, df=380 and from 1.27±0.05 to 0.86±0.03, P<0.0001, t=6.733, df=380, respectively) (Table 2).
Lipid metabolism of MG
As previously reported, IPL shows great therapeutic potential for patients with MGD and ocular surface diseases. It is well established that MGs are rich in lipids. The meibum and lipids of the MG were collected to perform a lipid analysis using LC-MS/MS. A total of 323 lipid species were identified and most of them were rich in triglyceride (TG, n=127), ceramide (Cer, n=86), phosphatidylcholine (PC, n=29), sphingomyelin (SM, n=28), simple Glc series G1 (CerG1, n=23), sphingosine (SO, n=9), lysophosphatidylcholine (LPC, n=8), lysophosphatidylethanolamine (LPE, n=8), monogalactosyldiacylglycerol (MGDG, n=5), phosphatidylserine (PS, n=5), digalactosyldiacylglycerol (DGDG, n=4), (O-acyl)-1-hydroxy fatty acid (OAHFA, n=4), phosphatidylethanolamine (PE, n=4), digalactosylmonoacylglycerol (DGMG, n=3), wax ester (WE, n=3), simple Glc series G2 (CerG2, n=2), coenzyme (CO, n=2), lysophosphatidylglycerol (LPG, n=2), phosphatidylglycerol (PG, n=2), diglyceride (DG, n=1), lysophosphatidylinositol (LPI, n=1), lysophosphatidylserine (LPS, n=1), lysosphingomyelin (LSM, n=1), monogalactosylmonoacylglycerol (MGMG, n=1), phosphatidylinositol (PI, n=1), and stigmasterol ester (StE, n=1) (Fig. 1A). When we compared the T0 to control and T2 to T0, there were significant differences observed between the groups in lipid quality, as shown by the volcano plot (Figs. 1B and C). Moreover, when we compared the T2 to T0, most of the different lipids were decreased in the former; when we compared the T0 to control group, most of the different lipids were increased in T0.
At the individual lipid compound level, the increased lipids in T0 (i.e., CerG1, CerG2, Co, DG, LPC, LPE, LPG, LPI, LPS, OAHFA, StE, and TG) were decreased following IPL treatment (T2) (Figs. 2A1–A12). There were no significant differences in Cer, DGDG, LSM, MGDG, PG, PI, PS, SM, SO, and WE that increased in T0 and T2 (Figs. 2B1–B10). The decrease in DGMG and MGMG in T0 could be reversed by IPL treatment (T2) (Figs. 2C1 and C2). Furthermore, the levels of PC and PE in T2 were higher than those measured in T0 (Figs. 2D1 and D2).
A total of 41 lipid species were significantly different in patients with MGD (T0) versus healthy controls. The different lipids were rich in TG (6/41), SM (6/41), Cer (6/41), OAHFA (3/41), PC (2/41), DGMG (2/41), SO (3/41), PS (2/41), LPE (1/41), LSM (1/41), CerG1 (1/41), CO (1/41), MGMG (1/41), and WE (1/41) (Table 3). Following IPL treatment (T2), 24 lipid species were significantly different compared with T0: TG (10/24), LPC (6/24), OAHFA (4/24), Cer (2/24), SM (1/24), and PE (1/24) (Table 4). The change of the lipid metabolic pattern was highly related to the improvement of the clinical index (Fig. 3). The correlation was rich in BUT1, CR and meibum quality and some lipids were cross related with them. LPC (18:2)+HCOO, OAHFA(18:1/34:1)-H, OAHFA(18:1/32:2)-H, OAHFA(18:1/32:1)-H, TG(16:0/18:2/20:5)+H, LPC(18:2)+H, OAHFA(18:1/30:1)-H, LPC(18:1)+H, TG(25:0/16:1/18:1)+NH4, TG(16:1/18:1/18:2)+NH4, TG(26:1/18:1/18:1)+NH4 and TG(18:1/18:2/20:5)+H were highly corelated with the change in the meibum quality. Cer(d17:0/2:0)+H, OAHFA(18:1/34:1)-H, OAHFA(18:1/32:1)-H, OAHFA(18:1/32:2)-H, TG(16:0/18:2/20:5)+H and OAHFA(18:1/30:1)-H were related with the change in the CR. The change of BUT1 may be induced by OAHFA(18:1/32:2)-H, OAHFA(18:1/34:1)-H, OAHFA(18:1/30:1)-H and OAHFA(18:1/32:1)-H. Cer(d18:1/18:0)+HCOO and Cer(d17:0/2:0)+H were related with the aBUT. TH was highly related with Cer(d18:1/18:0)+HCOO, SM(d36:1)+H, LPC(16:0)+HCOO, LPC(18:0)+HCOO and TG(18:1/18:2/20:5)+H. TG(16:0/18:2/20:5)+H was highly related with the MGE. Notably, 13 lipids were significantly increased in T0, whereas they were decreased following IPL treatment. The different lipids were OAHFA(18:1/34:1)-H(C52H97O4), OAHFA(18:1/32:1)-H(C50H93O4), OAHFA(18:1/32:2)-H(C50H91O4), LPC(18:1)+H(C26H53O7N1P1), LPC(18:0)+HCOO(C27H55O9N1P1), LPC(18:1)+HCOO (C27H53O9N1P1), LPC(16:0)+HCOO(C25H51O9N1P1), LPC(18:2)+H(C26H51O7N1 P1), TG(26:0/16:0/18:1)+Na(C63H120O6Na1), TG(18:1/18:2/20:5)+H(C59H99O6), TG(16:0/18:2/20:5)+H(C57H97O6), TG(16:1/18:1/18:2)+NH4 (C55H102O6N1), and Cer(d17:0/2:0)+H(C19 H40 O3 N1) (Table 5). We found that Cer(d18:1/18:0)+HCOO (C37H72O5N1) was decreased in T0, whereas it was increased in T2 (Table 5) (Fig. 4).