Ability of non-invasive tear break-up time to determine tear instability in contact lens wearers

To evaluate the performance of precorneal and prelens non-invasive tear break-up time (NIBUT) measurements to determine tear instability in contact lens (CL) wearers. In this study, 50 eyes of 25 CL wearers were evaluated. Precorneal first and average NIBUTs and fluorescein tear break-up time (FBUT) were measured before wearing CLs in the morning. Those with FBUT less than 10 s were considered to have tear instability. After wearing CL, first and average prelens NIBUTs were measured at the 30th minute. The receiver operating characteristic (ROC) curve was analyzed to evaluate the performance of NIBUT measurements in the diagnosis of tear instability. The FBUT value had a significant correlation with first and average precorneal NIBUT values (p < 0.001, r = 0.653 and p < 0.001, r = 0.628, respectively). The FBUT value had no correlation with the prelens first and average NIBUT values (p = 0.542 and p = 0.263, respectively). To understand the relationship between the precorneal and prelens NIBUT values measured by the automated method, their correlation was evaluated. There was no significant correlation between the precorneal and prelens NIBUTs (for all; p > 0.05). The area under curve (AUC) in ROC curve for the first and average precorneal NIBUTs were 0.842 (p < 0.001) and 0.810 (p < 0.001), respectively. The AUC values for the first and average prelens NIBUTs at the 30th minute of CL wear were 0.586 and 0.619, respectively (p = 0.317 and p = 0.166, respectively). Precorneal NIBUT measurements may be useful in diagnosing tear instability. Prelens NIBUT values are not yet capable of adequately defining the tear film dynamics in CL wearers.


Introduction
Since some changes occur in the ocular surface and tear functions due to the use of contact lens (CL), it is necessary to evaluate the ocular surface and tear stability to prevent CL-related complications [1,2]. CLs separate the tear film into two layers, therefore they are considered to increase tear evaporation and cause dry eye [3,4]. Although CLs are very effective in correcting vision, nearly half of those who use CLs have dry eye symptoms by the end of the day due to tear instability, and some have to give up wearing CLs for this reason [1,5,6]. Failure to distribute the tear film evenly over the CL surface reduces contrast sensitivity and visual quality [7,8]. The evaluation of the quality of the precorneal and prelens tear film is, therefore, important in the follow-up of CL users, as well as guiding in the selection of lens material and determination of the wearing protocol.
Fluorescein tear break-up time (FBUT), which is the most commonly used method in determining tear film instability and expresses the time interval between a complete blink and the first dry spot in the precorneal tear film, is a standard clinical test for the diagnosis of tear dysfunction and evaluation of the effectiveness of related treatments [9,10]. Being simple and requiring no expensive equipment, FBUT measurements are widely used in clinical practice; however, this method cannot be used in CL users while they are wearing CLs on their eyes since fluorescein needs to be applied. Another way to measure tear instability is non-invasive tear break-up time (NIBUT), which is an automated method based on the principle of evaluating irregularities in the tear by projecting Placido rings onto the precorneal tear film. The advantage of this method is that it can measure tear instability without the need for a foreign substance, such as fluorescein drops or strips; thus, it provides a practical solution for the evaluation of prelens tear instability in people wearing CLs [11,12]. Although the NIBUT method is used by eye care practitioners, there is not yet sufficient evidence of its ability to assess tear instability in CL wearers [11,13].
It is important to measure prelens NIBUT since CL insertion increases evaporation. CL wear may lead to dry eye symptoms even in individuals with stable precorneal tear film [6]. This study aimed to evaluate the usability of precorneal and prelens NIBUT measurements with the Sirius multifunctional corneal topography device (Sirius, CSO, Florence, Italy) in the prediction and identification of tear instability in CL wearers. Thus, the ability of NIBUT to demonstrate precorneal or prelens tear film quality and whether it can be routinely used in CL users was investigated.

Methods
In this prospective study, 50 eyes of 25 CL wearers who presented to Van Training and Research hospital for routine control were evaluated. The research was carried out in accordance with the principles of the Declaration of Helsinki after receiving informed consent of the patients. Approval was obtained from the Van Training and Research ethics committee to conduct the study. The study was conducted between May and August 2019 in ophthalmology department of Van Training and Research Hospital. People with pterygium, keratoconus, or corneal dystrophy, ocular allergy, history of eye surgery, refractive surgery, or trauma, those with more than 0.75 diopters of astigmatism, and those using systemic drugs that could affect ocular surface or tear functions were excluded from the study.

Study protocol
Firstly, refractive errors and keratometric values and visual acuity of the participants were determined. Then, detailed biomicroscopic anterior segment examination was performed. To ensure standardization in the study, a fresh CL with the same material and design was tested in all participants. The suitability of this newly tested daily disposable CL (senofilcon A [Acuvue Oasys with HydraLuxe, Johnson & Johnson Vision Care] with 38% water content and 8.5 mm base curve/14.3 mm diameter) was confirmed by the same doctor (EM) who evaluated the centralization and movement of the lens, and the visual acuity of the participant. Eligible participants were told not to use their own CLs for a week, and attend an examination at the hospital at 9 a.m. the following week. After oneweek standardization, precorneal NIBUT and FBUT values were measured before wearing CL in the morning. After wearing CLs and waiting for 30 min to let the tear settle, the first and average prelens NIBUT values were measured over the CLs. NIBUT measurements were made by the same technician, and the FBUT measurements were undertaken by the same doctor (MSD). According to the internationally accepted limits (Dry Eye Workshop 2007 and 2017), FBUT values under 10 s are considered as tear instability [11,14]. For this reason, a baseline FBUT value of 10 s and above was considered as normal and less than 10 s as an indication of tear instability. The ability of NIBUT values to identify those with tear instability with respect to the baseline FBUT value obtained in the morning was investigated.

Non-invasive tear break-up time
The NIBUT of the participants was measured in accordance with the manufacturer's instructions using Sirius multifunctional topography device. Just before starting the measurement, the image was fully centered and sharpened. The patient is instructed to blink twice to spread the tear across the ocular surface, then keep the eyes open as much as possible. The device automatically starts measuring after the patient has blinked twice. The device evaluates time-dependent changes in the integrity of the tear film spread over the anterior corneal surface. The test is performed through videokeratoscopy, the video recording of the Placido disk ring projection on the patient's precorneal or prelens tear film. When the tear film is distributed homogeneously on the ocular surface, it causes a regular appearance in the rings, and any deformation or disruption in the tear film results in deformation in one or more rings at the measurement time and place. The system processes the video in real time and without user intervention. The first break up of tear film causes distortion or interruption in the Placido disk pattern, which is recorded as first NIBUT by the device. In the device, in order to assist and complement the diagnosis, after blinking, the average duration of all break-up intervals occurring in the precorneal or prelens tear film is recorded as average NIBUT. As an example, Fig. 1 shows the NIBUT maps of one of the participants in our study. Fluorescein tear break-up time A disposable fluorescein strip was applied over the inferior tear meniscus of the participants to measure the FBUT value. The participants were asked to blink several times for the homogeneous distribution of fluorescein on the ocular surface, and then to keep their eyes open as much as possible. The time elapsed from the blinking of the participants to the first black spot on the cornea was measured as the FBUT value using a cobalt blue filter on a biomicroscope. Those with a baseline FBUT of less than 10 s were considered to have tear instability, while 10 s or greater FBUT were considered to indicate a normal tear film [10,11,14].

Statistical analysis
Data obtained in the study were statistically analyzed using SPSS24 software (Statistical Package for the Social Sciences, IBM). Descriptive statistics were used for the variables of age, gender, and tear parameters. The normality of data distribution was evaluated by the Kolmogorov-Smirnov test. The relationship between FBUT and NIBUT (precorneal and prelens) was evaluated by Spearman correlation analysis since the data did not show normal distribution. Those with tear instability were determined based on internationally accepted cut-off values of basal FBUT measurements as mentioned earlier. Then, the receiver operator characteristic (ROC) curve analysis was conducted to measure the ability of first and average NIBUTs to identify tear instability [15]. The area under the curve (AUC), considered as an index of discriminating performance, assessed this ability. For first and average NIBUTs, the sensitivity, specificity, and optimum cut-off value for predicting tear instability were determined using ROC curve analyses based on the highest possible Youden index [16,17]. Evaluations were made at the 95% confidence interval, and p values less than 0.05 were considered to indicate statistically significant differences.

Results
The average age of the participants was 22.16 ± 4.83 years (range 14-31 years). The mean precorneal first and average NIBUT values of individuals were 7.92 ± 4.53 s and 9.41 ± 3.67 s, respectively, prior to CL wear. The mean FBUT was 6.68 ± 4.66 s before CL wear. Tear instability was observed in 32 (64%) eyes based on FBUT value. At the 30th minute of CL wear, the prelens first and average NIBUT values were 4.77 ± 4.16 s and 7.60 ± 3.36 s, respectively (Table 1).
When the correlation between the FBUT value and the precorneal and prelens NIBUT values are evaluated, a significant correlation was detected between the FBUT value and the precorneal first and average NIBUT values (p \ 0.001, r = 0.653 and p \ 0.001, r = 0.628, respectively), but no significant correlation was found between the FBUT value and the prelens first and average NIBUT values (p = 0.542 and p = 0.263, respectively). To understand the relationship between the precorneal and prelens NIBUT values measured by this new objective and automated method, their correlation was also evaluated. No significant correlation was detected between the precorneal and prelens NIBUT values (for all; p [ 0.05).
A ROC curve was designed to determine the ability of the automated first and average NIBUT measurements to detect tear instability with respect to FBUT. The AUC values in the ROC curve were 0.842 (CI: 0.731-0.953) for the precorneal first NIBUT and 0.810 (CI: 0.684-0.935) for the precorneal average NIBUT prior to CL wear (p \ 0.001; for both). The precorneal first and average NIBUT values prior to CL wear had a good ability to determine tear instability ( Fig. 2A).  (Fig. 2B). The ROC analysis suggested an ideal cut-off value of 8 s for the precorneal first NIBUT (maximum Youden index) with 78.1% sensitivity and 94.6% specificity in predicting tear instability. The ROC analysis suggested an ideal cut-off value of 12.65 s for the precorneal average NIBUT with 81.3% sensitivity and 77.8% specificity in predicting tear instability ( Table 2). The ROC analysis did not recommend any reliable cut-off values for prelens NIBUT values.

Discussion
This study examined the ability of the objective and automated precorneal and prelens NIBUT measurements to determine tear film instability in individuals using CLs. In this study, the previously internationally accepted FBUT cut-off value (less than 10 s as tear instability) was taken into consideration in identifying individuals with tear film instability [11,14]. We determined that the precorneal first and average NIBUT values automatically and objectively measured by the Sirius multifunctional topography device were able to successfully detect tear instability. However, we found that the reliability and diagnostic ability of prelens first and average NIBUT measurements over CL was very low. As suggested in the literature, the reason for this may be that the device  cannot distinguish between the thin areas and break-up areas in the prelens tear film since the prelens tear layer becomes very thin after CL wear [18]. Minor distortions caused by CL may also have limited the device's ability to reliably identify prelens tear film instability. Meşen et al. found that there was no significant difference between the measured precorneal NIBUT and FBUT values in individuals with dry eye symptoms [19]. Bhandari et al. found a good correlation between FBUT and precorneal NIBUT in patients with dry eyes and healthy individuals. They determined the precorneal NIBUT cut-off value as 6.2 s with 86.1% sensitivity and 81.1% specificity in the diagnosis of dry eye [13]. Best et al. found a significant correlation between FBUT and precorneal NIBUT before CL wear. They also stated that the basal NIBUT value could be used to predict CL drop-out, and those with a NIBUT value below 10 s should be followed up carefully [20]. Similarly, we found a significant correlation between FBUT and precorneal NIBUT in our study. In addition, we determined the optimum cut-off values for the precorneal first and average NIBUT measurements as 8 s and 12.65 s, respectively, in determining tear instability with high specificity and sensitivity. We consider that studies with larger series are needed to provide more reliable cut-off values. We believe that future technological and software-related developments will reduce variations in the results of conducted studies.
To the of best our knowledge, this is the first study to investigate the ability of precorneal and prelens NIBUT to detect tear instability with respect to FBUT in CL wearers. Thus, it becomes possible to evaluate whether this automated and objective method can be substituted for the traditional subjective method, which is the most commonly used method in clinical practice. Since tear function tests assess different properties of tear film, our study is important in terms of investigating the relationship of two different methods based on tear break-up time in CL wearers. In this study, we found no significant correlation between the prelens NIBUT values and the FBUT or precorneal NIBUT values, and the ROC analysis did not reveal a sufficiently reliable cut-off point for the prelens NIBUT values to predict tear instability. Gullion et al. found that prelens NIBUT is shorter than precorneal NIBUT. In addition, they detected that surface dehydration speed and area were much higher for the prelens than for the precorneal tear film [21]. Similar to our study, Llorens-Quintana et al. reported that there was no significant correlation between precorneal NIBUT and prelens NIBUT [22]. Pult et al. and Chui et al. similarly noted that prelens NIBUT was not a reliable predictive method for identifying symptomatic and asymptomatic CL users [23,24]. Glasson et al. found that the precorneal NIBUT value measured before CL wear showed a significant difference between the patients with and without CL intolerance, while the prelens NIBUT measured six hours after CL did not significantly differ between the two groups [5]. In contrast to these studies, Gullion et al. reported that prelens NIBUT was shorter in symptomatic CL wearers than in asymptomatic wearers [18]. Due to these contradictory results, prelens NIBUT cannot be said to be successful in detecting CL intolerance. Itokawa et al. reported that there is a correlation between precorneal NIBUT values measured using two different devices, but there is no correlation between prelens NIBUT values measured by that device [25]. The relationship between CL insertion duration and prelens NIBUT has been examined in the literature and conflicting results have been reported [26,27]. We detected no correlation between precorneal and prelens NIBUT, which were both measured by the same automated method. Studies on NIBUT and CLs are summarized in Table 3. These studies support the results of our study in terms of demonstrating the low reliability of prelens NIBUT. We think that the identification of prelens tear kinetics by NIBUT is not as practical and easy as precorneal NIBUT due to the formation of different tear kinetics in the presence of lenses.
It has been reported in the literature that there is a relationship between the distribution and stability of tear film and visual acuity [28]. Since NIBUT determines tear instability by evaluating topographic regularity indices, it also measures the anatomical function of the tear, as well as its physiological function. Since irregularities in the mires cause a decrease in image quality in the macula, it directly affects visual function [1,29]. This feature can help simultaneously evaluate both tear instability and its effect on vision quality using NIBUT in CL wearers with complaints about vision [1,30]. In our study, we determined that precorneal NIBUT was able to reliably detect tear instability. For this reason, we believe that this technique is useful in evaluating CL  (Table 3). But this is the first study to evaluate prelens and precorneal NIBUT using the Sirius multifunctional topography device in CL wearers. However, the number of patients in our study was limited and they had a relatively wide range of ages. Therefore, our results need to be confirmed by further studies with a larger sample size and comparing different age groups with each other.

Conclusion
It was shown in our study that the precorneal NIBUT values measured automatically in the follow-up of CL users may be a fast, reliable, non-invasive, and patientfriendly test based on a simple but effective imaging method in determining tear instability. In addition, this objective method may provide more detailed information than conventional methods about the tear film. Funding The authors declared that this study has received no financial support.
Data availability All authors declared that all data and materials as well as software application or custom code support their published claims and comply with field standards.

Declarations
Conflict of interest The authors declare that they have no conflict of interest.
Ethical approval All procedures performed in the studies involving human participants were in accordance with the ethical standards of the institutional and/or national research committee and with the 1964 Declaration of Helsinki and its later amendments or comparable ethical standards.
Consent to participate (ethics) Informed consent was obtained from all individual participants included in the study.
Consent to publish (ethics) Not applicable.