Abstract
Purpose.:
To analyze in vivo the dynamic changes induced by different artificial tears (ATs) in the precorneal tear film (PCTF) and lower tear meniscus (LTM) by using spectral-domain (SD) anterior segment optical coherence tomography (OCT).
Methods.:
We prospectively examined 42 normal human eyes by using SD-OCT imaging. On the day before OCT imaging, all enrolled subjects were evaluated for abnormalities of ocular surface. All tear film images were obtained before and after instillation of three different types of ATs (mucomimetic, lipid-based, and saline) in five serial scans: immediately (within 30 seconds), at the first, fifth, 10th, and 20th minute. Subjects received a drop of 35 μL in one randomly selected eye. All examinations were conducted in the same conditions of temperature, brightness, humidity, and time of day.
Results.:
Changes in the morphological pattern of both LTM and PCTF were associated with the type of artificial tear instilled on the ocular surface. Similarly, the radius of curvature (rc ), the height (h), and the depth (d) showed dynamic variations depending on treatment. Although by the 20th minute, both h and d returned to baseline values in all groups, a significant difference in rc (compared with baseline) was detected for mucomimetic ATs (P = 0.04) and lipid ATs (P = 0.02).
Conclusions.:
Spectral-domain OCT imaging has preliminarily proved to be a noninvasive tool to evaluate, in real time, the different changes induced by ATs instillation. An important stride in understanding the clinical response to various tear substitutes can be achieved by this objective and quantitative approach.
Normal tear film is important for the maintenance of ocular surface physiology and ocular comfort. Accordingly, tear film volumes and distributions are maintained in a dynamic balance between tear secretion and tear loss, so that corneal and conjunctival epithelia are protected while the eye is open. Any alteration in tear system may cause dysfunction and compromise the tear film integrity, potentially causing ocular discomfort and disease.
1–3
The current model of the tear film consists of a thick aqueous-mucin layer covered by the tear film lipid layer. Artificial tears (ATs) are commonly used to compensate for tear film inadequacy, so that the ocular surface is preserved. In actuality, the term “artificial tears” is a misnomer for most products because they do not mimic the composition of human tears.
4 Although the majority of these products function as lubricants (e.g., saline), some recent formulations claim to be more similar to the composition of human tears, such as lipid-based ATs and mucomimetic ATs, which mimic mainly the lipid layer and the aqueous-mucin layer, respectively.
Since ATs are available over-the-counter (i.e., they are directly sold to a consumer without a prescription from a healthcare professional), many individuals use them to self-medicate (e.g., for temporary symptoms or signs of ocular surface irritation).
If we consider that it is possible for ATs to be approved even without knowing their clinical efficacy (e.g., in the United States they are approved based on the US Food and Drug Administration monograph on over-the-counter products, 21 CFR 349) and their monographs can be free of a detailed description concerning inactive additives (in some cases, they are not listed) and solution parameters, it may be difficult to predict and verify their clinical outcome.
On the other hand, in clinical practice, it is difficult to evaluate and understand, in real-time and directly (noninvasively), the dynamic distribution and the residence time of ATs on the ocular surface. Therefore, despite their widespread use, their influence on the integrity of the tear film, on the basis of the different types, remains unclear.
Recently, optical coherence tomography (OCT) has been used to measure some parameters of the lower tear meniscus (LTM) after AT instillation in a quick, reproducible, noncontact, and noninvasive manner.
5–8 In these works, the height and the depth of the LTM, which are reliable indicators of the overall tear volume,
9–12 have shown significant changes only for 5 minutes.
8 Nevertheless, in general, patients usually report an improvement of their symptoms for much longer.
In this study, we used real-time OCT to evaluate the dynamic distribution of three different types of ATs on the tear film, by assessing the following qualitative and quantitative variables of lacrimal system: the morphological pattern of the LTM and PCTF, as well as the radius of curvature (rc ), the height (h), and the depth (d) of the LTM.
The research was conducted in accordance with the tenets of the Declaration of Helsinki. Informed consent was obtained from each subject, after ethics approval obtained from the Office of Research Ethics, University of Cagliari. Forty-two healthy subjects, 80.6% female, aged from 40 to 55 years (47.09 ± 4.07 [mean ± SD]) were recruited from patients of the Eye Clinic (Department of Surgical Sciences, Cagliari, Italy) and staff members of the School of Ophthalmology (University of Cagliari, School of Ophthalmology). All examinations were conducted in the same conditions of temperature (within a range of 15–25°C), humidity (within a range of 30%–50%) and time of day (between 3 and 5 PM) in a dimly lit consulting room.
On the day before OCT imaging, all enrolled subjects were tested in the following sequence: ocular surface disease index (OSDI),
13 which is a symptom questionnaire consisting of a set of questions assessing the level of discomfort and the functional impact of their irritation symptoms; fluorescein tear break-up time (FTBUT); fluorescein staining of the cornea and conjunctiva graded according to the Oxford system
14 ; Schirmer I test; and a slit lamp examination of the lid margins and meibomian glands. Schirmer I tear test was performed after a 15-minutes rest from previous tear tests and, similarly to the FTBUT, was timed with a stopwatch.
Based on the results of these tests, the inclusion criteria for healthy eyes were: no significant symptoms of ocular irritation (OSDI <12), FTBUT > 10 seconds, Oxford scheme ≤ panel A, and Schirmer 1 test score of more than 10 mm in 5 minutes.
Subjects who had worn contact lenses and those with external ocular diseases in the previous 6 months, abnormalities of the lid margins or meibomian gland disease, lax conjunctival folds (which may prevent reliable measurements of the LTM)
11 or any evidence of abnormal blinking, were excluded from the study. None of the normal subjects had ever used tear supplements before or had history of eye surgery.
Based on the tear layer that they mimic, ATs were classified in three different types: mucomimetic (MAT), lipid-based (LAT), and saline artificial tears (SAT). Thus, we used a drop of MAT (35 μL, sodium carboxymethylcellulose 0.5%, Optive UD, preservative-free; Allergan, Inc., Irvine, CA, USA), which has shown good mucomimetic and mucoadhesive properties
15–19 ; a drop of LAT (35 μL, polar phospholipids and mineral oil, Systane Balance; Alcon Laboratories, Inc., Fort Worth, TX, USA), which contains dimyristoyl phosphosphatidylglycerol (a polar phospholipid) and mineral oil that mimic the lipid layer of the tears
20,21 ; a drop of SAT (35 μL, saline 0.9%, BLUsal A; SOOFT, Montegiorgio, Italy).
With this in mind, the subjects were divided into three groups and only one type of artificial tear was used in each group. These subject groups were labeled by drop type (i.e., group 1, MAT; group 2, LAT; and group 3, SAT).
The LTM and PCTF of one randomly chosen eye in each patient was imaged by vertical and horizontal scans, respectively, on the vertical and horizontal axis passing across the corneal apex. The OCT scans were performed by using OCT equipment (Cirrus HD-OCT 4000; Carl Zeiss Meditec, Inc., Dublin, CA, USA). This system is a Fourier-domain OCT platform that works at a wavelength of 840 nm, takes 27,000 axial scans per second and has a 5-μm axial resolution. The cross-sectional ocular surface images were acquired using the anterior segment 5 line raster scanning protocol. The mode acquires a set of five parallel lines of equal length at 3 mm. Each line is composed of 4096 A-scans and each 5-line raster scan was taken approximately 0.75 seconds. The lines are separated by 250 μm. After image capture, the individual line with greatest clarity of detail was selected for our analysis. The subject was asked to blink normally during the examination period. Before each scan, patients were instructed to stare straight ahead. The ocular surface was imaged 2 seconds after a blink. The axial distance of patients was adjusted so that the LTM or the PCTF has been within the middle third of the scan.
22
Optical coherence tomography scans were performed at baseline and after instillation of ATs in five serial scans: immediately (within 30 seconds), at the first, fifth, 10th, and 20th minute.
The appearance of the LTM was described according to the morphological pattern in three categories (
Fig. 1): a hypoechoic type (areas of low or no reflectivity in the whole LTM, except the free surface); a constellation type (multiple hyperreflective points inside the LTM); and an enhanced reflectivity type (areas of high/moderate reflectivity in the whole LTM). Similarly, the PCTF appearance was described according three types (
Fig. 2): a linear pattern (a continuous line of homogeneous reflectivity onto the epithelial surface of the cornea); a double-band pattern
16 (a tear film of significant thickness—i.e., a two-layered structure localized onto the epithelial surface of the cornea, consisting of an outer band of high reflectivity and an inner band of low reflectivity); and a granular pattern (an outer line of high/moderate reflectivity segmented in different points).
As previously reported,
5,6 custom software was used to process OCT images to yield all metric variables (
Fig. 3): radius of curvature (
rc ), height (
h), and depth (
d) of the LTM. Briefly, operator inputs were provided to identify three touch points used to fit a circle that yielded the tear meniscus curvature. These three points included points where the tear surface touches the cornea and the lids and the middle point of the tear meniscus outer edge. After that, the software processed the image and yielded the results. At the same time,
h was measured from the cornea-meniscus junction to the lower eyelid-meniscus junction. On the other hand,
d was measured from the midpoint of the air-meniscus interface to the cornea–lower conjunctiva junction. Similar to others, the
h and
d of the LTM are used in our work as reliable indicators of changes in the overall tear volume.
9–12
All of the information was recorded by the researchers so that subjects could not be identified, directly or through identifiers linked to the subjects.
Statistical analysis was performed using statistical software (SPSS version 21.0; SPSS, Inc., Chicago, IL, USA).
Numeric data were summarized as mean ± SD for parametric data, while nominal data were summarized as percentage. Data were analyzed by Shapiro-Wilk test and Lilliefors test for normality.
The differences in age among the three groups were studied by means of one-way analysis of variance, and Tukey's range test for the post hoc analysis. Comparison among the three groups for sex, morphological pattern, and metric parameters (rc , h, d), was studied by Kruskal-Wallis statistic.
The variations over time of morphological patterns (compared with baseline) were calculated by Fisher's exact test, while the differences of rc , h, d, (compared with baseline) were analyzed by t-paired test (for normal distribution) and Wilcoxon signed-rank test (for nonnormal distribution).
No significant differences in baseline values for metric parameters (rc , h, d) were found among groups (Kruskal-Wallis statistic = 0.466, P = 0.792; Kruskal-Wallis statistic = 1.702, P = 0.427; Kruskal-Wallis statistic = 0.045, P = 0.978, respectively). Although by the 20th minute, both h and d returned to baseline values in all groups, a significant difference in rc (compared with baseline) was detected for mucomimetic and lipid ATs.
-
Group 1 (MAT)—At baseline, the
rc was 646.90 ± 349.77 μm,
h was 367.30 ± 152.15 μm, and
d was 170.06 ± 83.39 μm. After instillation of mucomimetic ATs, the
rc has shown a significant increase (compared with baseline) in all serial scans (
Table 3). An increase of the
h and the
d, was observed until the 10th minute.
-
Group 2 (LAT)—At baseline, the
rc was 563.10 ± 222.19 μm,
h was 317.57 ± 124.25 μm, and
d was 169.97 ± 79.44 μm. Immediately after lipid-based ATs instillation, the
rc increased significantly (1666.04 ± 909.38 μm) and, at the following scans, its values remained higher than baseline until the fifth minute (
Table 3). Then, at the 10th minute, the
rc came back to baseline value. Finally, a significant reduction of the
rc (473.76 ± 160.00 μm) was detected at the 20th minute (
Table 3).
The
h increased significantly after lipid-based AT instillation (812.32 ± 284.37 μm) and its values remained higher than baseline until the fifth minute (
Table 3). Similarly,
d increased significantly after instillation (322.07 ± 79.44 μm), but its elevation remained significant only until the first minute (
Table 3).
-
Group 3 (SAT)—At baseline, the rc was 496.59 ± 24.28 μm, h was 271.69 ± 68.73 μm, and d was 161.26 ± 72.80 μm. After saline AT instillation, an increase of the rc , h, and d, was observed until the following scans: at the 10th, fifth minute, and immediately, respectively.
In this study, we revealed for the first time by OCT simultaneous changes over time in the morphological patterns of the LTM and PCTF as well as in both the indices of the LTM volume (h and d) and in the rc , after AT instillation. With our approach, we demonstrated in real-time that specific variations in qualitative and quantitative parameters of human tears were associated with the type of artificial tear instilled on the ocular surface.
Of great interest were the differences in the morphological patterns that occurred with different AT compositions. We speculate that these observations may teach us about the activity of the ATs in vivo. For instance, in the constellation pattern (
Fig. 1B), it was possible to observe several points of different reflectivity inside the tear meniscus after MAT instillation. Since like all water-soluble polymers the particles of carboxymethylcellulose have the tendency to agglomerate or lump when first wet with water,
23 the multiple hyperreflective points of the constellation pattern may represent the result of chemical interactions between MATs and humans tears. Similar to others,
24 we could highlight the increase in reflectivity resulting from the increase in lipid concentration after LAT instillation in case of the enhanced reflectivity pattern (
Fig. 1C). This phenomenon is particularly interesting because it suggests, at least in part, that the OCT signal of high reflectivity elicited by the PCTF at the interface between air and the outermost surface layer of the tear film (
Fig. 2), is associated with the presence of lipids. Of note, the analysis of two morphological patterns, constellation and enhanced reflectivity patterns, has allowed us to detect the long residence time and the prolonged action of MAT and LAT, despite the decay back to baseline values for the dimensional parameters (
h and
d).
With regard to the morphological changes of the PCTF, OCT imaging has revealed that even under conditions of increased PCTF thickness, as in the double-band pattern, the layer with higher reflectivity is the outer one (
Fig. 2B). Since the tear film lipid layer is the outermost of the PCTF, the high intensity of OCT signal at this level may arise by both its qualitative characteristics (lipids = high reflectivity)
24 and its regular/smooth structure (guaranteed by uniform distribution of lipids and aqueous-mucin layer). As a consequence, the linear pattern implies a regular structure of the PCTF, whereas the granular pattern suggests a nonhomogeneous distribution of tears on the corneal surface. However, not all subjects demonstrated a linear pattern after instillation of LAT. This fact suggests that the simple addition of lipids may not be sufficient to lead to a regular distribution of the PCTF, at least in a minority of patients, and that other factors probably play an important role (i.e., the substances capable of spreading lipids over the aqueous layer). Interestingly, after SAT instillation, a significant increase of the rate of granular pattern was observed until the 20th minute in all patients, probably due to a
dilution of substances involved in determining the regular distribution of the PCTF (e.g., lipids or mucins).
Unlike the morphological pattern and the
rc , the indices of the LTM volume (
h and
d) have shown the return to baseline values in all groups by the 20th minute. Therefore, our results support previous data indicating the discrepancy between the period of the relief of symptoms and the time in which the tear meniscus shows an increase in volume.
4,8,11 Consequently, we believe that a quantitative analysis of the LTM volume is not appropriate to understand the improvement in symptoms after AT instillation. In contrast, the morphological pattern and the
rc have shown long-lasting changes from baseline, potentially permitting a better understanding of time-length of patient comfort.
4
Interestingly, the rc has shown a characteristic behavior depending on the type of ATs instilled. In particular, in the case of the LAT, we have demonstrated a significant reduction of the rc , despite the decay back to the baseline values for h and d. Conversely, in the case of MAT, the rc has initially shown a dramatic increase followed by a gradual decay with values that remained significantly higher than baseline (as opposed to the h and d). On the other hand, by instilling the SAT, we have observed a gradual return to the initial values of rc by the 20th minute.
The
rc is clearly an important parameter of the tear film (
Fig. 4). In fact, in physiological conditions, tear molecules close to the ocular surface interact with the conjunctival and corneal epithelia, and the surface tension distorts the tear surface forming the tear meniscus (
Fig. 4B).
25 Considering the Young-Laplace equation,
26 which describes the relationship (a direct proportionality) between the surface tension (γ) and the radius of curvature (
rc ) of meniscus, the changes in the latter (
Drc ), under condition of equal tear volume,
27 clearly represent in vivo an index of the surface tension variations (
Dγ) of tears:
Therefore, in case of a significant decrease in the final r c (at the 20th minute after AT instillation) with respect to the baseline value, and without significant differences between final and baseline values for tear volume (i.e., h and d), we clearly find a reduction in the surface tension. The opposite occurs with an increase in the final rc (with equal h and d).
In general, a reduction in
γ means a better wettability of the human tears, with equal volume, on the ocular surface, which implies a better distribution of the tear film and therefore a better protection for the epithelia. Particularly, the results of our study demonstrate that LAT instillation did not lead to a reduction in
γ in all patients, as theoretically suggested.
28 In fact, OCT imaging has allowed us to detect in vivo different changes in
γ of human tears after LAT administration on the basis of different interactions between supplemental tears and individual tear systems.
This fact suggests the importance of objectively evaluate the patient's response to the ATs in order to better understand and demonstrate therapeutic effects from different types of treatment modalities.
On the other hand, instillation of MAT resulted in a dramatically increase in
γ. Perhaps, this phenomenon occurred in vivo due to the high forces of cohesion between the molecules of MAT and the aqueous layer (determining high γ), as well as for an increased excretion toward lid skin of natural tear lipids, resulting in their dilution as suggested in previous models.
29 With this in mind, the relief of symptoms after MAT instillation is unlikely attributable to variations in γ, rather than to the attractive forces between MAT and aqueous-mucin layer, as well as between MAT and epithelia of ocular surface, which determine a prolonged long residence time of mucoadhesive polymer on the ocular surface.
30 By these chemical interactions, MATs are able to retain water molecules and form a protective structure, detected by OCT as a double-band structure on the epithelium of the cornea.
Similarly, we have documented an increase in γ, but only initially, after SAT instillation. It is likely that the washout of some solutes (e.g., tear lipids) may play a role also in this case.
All these OCT findings emphasize the possibility to evaluate the tear substitute most suitable for the individual patient, in order to be chosen among the high number of products available on the market. For all these reasons, we strongly believe that the final effect of tear substitutes should be directly evaluated on the patient to understand the morphological and metric variations induced on the tear system. In this sense, OCT imaging could be a useful tool to assess in vivo the dynamic interaction between ATs and human tears, noninvasively.
Interestingly, during OCT examination, a wave motion of the free surface of the LTM (
Fig. 5A) was observed immediately after each blink (less than 2 seconds). These oscillations describe a transfer of energy (derived from blinking) from one point to another of the ocular surface with little or no associated mass transport. Consequently, under these conditions, it is possible to observe a variation of the radii of curvature (a disturbance) along the free surface of the LTM (
Fig. 5A). As far as we know, our study is the first to discuss the problem of misrepresentation of tear meniscus recorded by OCT. Thus, in order to avoid this type of bias, we have performed OCT scans after the wave motion of the LTM (2 seconds after a blink). It is likely that the previous authors did not notice this artifact because of the lower resolution provided by their OCT devices (∼10 μm).
5,6
There are potential limitations when interpreting the results of our study. First, the identification of the tear meniscus points for the determination of the metric parameters was based on operator judgment. Therefore, the development of automatic software would be desirable to avoid possible biases (operator-dependent errors). Second, since conjunctivochalasis is a relatively common finding in the older dry eye population, it clearly creates limitations to the analysis of metric parameters of the LTM by OCT (
Fig. 5B). For this reason, the presence of lax conjunctival folds may represent an exclusion criterion for several old dry eye patients.
Similarly to others,
7,31,32 future studies should evaluate the relationship between specific dry eye symptoms and the appearance of the tear film by means of OCT, in order to ameliorate our understanding of the physiopathology of the ocular surface, as well as to predict the subjective efficacy of proposed topical lubricants.
In summary, in the current study, we have evaluated the retention time and the different dynamic distributions of ATs on the basis of their type, as well as a variable number of their possible effects. Obviously, tear substitutes can lead to various functional modifications by simply interacting with chemical or physical properties of lacrimal system. In this sense, OCT imaging has proved to be a noninvasive tool to evaluate, in real time, the different changes induced by supplemental tears, potentially permitting the development of truly effective ATs to improve the tear system. Further studies may analyze the morphological patterns and metric parameters of the tear film to identify in vivo the most suitable lubricants for various alterations of tears and ocular surface, as well as to customize the treatment to the individual case. Consequently, our research supports the use of OCT for evaluating the real-time response of the tear film to tear substitutes and for objectively quantify the dynamic efficacy of various dry eye treatments.
Presented in part at the annual meetings of the Association for Research in Vision and Ophthalmology, Fort Lauderdale, Florida, United States, May 2012, and Seattle, Washington, United States, May 2013.
Disclosure: P.E. Napoli, None; G.M. Satta, None; F. Coronella, None; M. Fossarello, None