To estimate the astigmatic power and axis of the tear film over the central optical zone of the cornea by vector analysis of topographic data, in ocular surface disease (OSD) and controls, during blink suppression.
Video-keratoscopic images were captured on opening the eyes after a single blink 5, 10 & 15 s later during blink suppression in OSD patients (mixed aetiology, group 1 age 20 50 years, n = 12, group 2 > 50 years, n = 38) and controls (group 3, n = 19). The SimK and axis values were used to calculate the astigmatism (power and axis) that formed in the precorneal tear film during each period. Data were aggregated into 3 periods; T0-T5 (between 0 & 5 s after the blink), T5-T10 (5 & 10 s later, T10-T15 (between 10 & 15 s later).
Mean (± SD, 95%CI) astigmatic power (DC) formed in the tear film over each period was respectively : Group 1, −0.81 DC (0.99, −1.44 to −0.17), −2.65 DC(1.36, −3.52 to −1.79), −1.37 DC (2.15, −2.73 to −0.01). Group 2, −0.33 DC (0.38, −0.45 to −0.20), −0.57 DC (0.97, −0.91 to −0.24) −0.96 DC (2.10, −1.68 to −0.24), Group 3, −0.57 DC (0.55 −0.76 to −0.38), −0.56 DC (0.57, −0.76 to -0.37), −0.31 DC (0.44, −0.46 to −0.16). Changes were significant in groups 1 (p = 0.013) & 3 (p = 0.033) but not in 2 (p = 0.078). Intergroup differences were significant at all periods (p < 0.05). Significant correlations were detected following vector analysis, e.g Group2 between the astigmatism formed during T5-T10 (y) and ocular surface astigmatism at 5 s was y = 0.281x – 0.834 (r = 0.328, n = 38, p < 0.05). In all three groups apparent changes in the axes of astigmatism were not significant (p > 0.05).
Changes in the precorneal tear film after blinking are predominately astigmatic indicating that changes in the central region of the tear film following the natural blink are quasi-orthogonal.
The precorneal tear film is the first optical surface encountered by light on entering the eye. Thus, any disturbance in the optical quality of the surface of the tear film is expected to impact on the overall optical performance of the eye and vision. Following the blink, the optical performance of the eye improves reaching an optimum 3−10 s later followed by a gradual decline [ ]. The combined effects of the change in thickness distribution and modulation of tear film refractive index; from the outermost lipid layer of the tear film towards the corneal surface; occurring after the blink has the potential to impact on the optical performance of the eye. Disturbances in the tear film could cause uncertainties to arise during subjective or objective refraction [ ]. The overall quality of vision in dry eyes is reduced [ ]and there is evidence suggesting this reduction is partly driven by the tear film [ ]. It is not unusual to encounter difficulties when refracting the patient with ocular surface disease. The challenge is further accentuated when the patient with an unstable tear film is astigmatic. Information regarding the overall optical performance of the eye is helpful, but it would be useful to gain a further understanding of how the tear film could influence astigmatism.
If the thinning of the precorneal tear film after the blink was uniform over the central optical zone of the cornea then any effect on astigmatism would be negligible. Alternatively, if the dynamics of tear film thickness distribution leads to a gradation of orthogonality then the tear film would be akin to an astigmatic lens impacting on the overall astigmatism of the eye. The information generated by traditional Placido-disc based video-keratoscopy implies the change in ocular surface power following a blink, over the central optical zone of the cornea, is less than ± 0.50 DS [ ]. However, Erdélyi et al. [ ] reported mean astigmatic changes of 0.28 DC thirty seconds after the blink. Zhu et al. [ ] reported significant changes of astigmatism can occur after the blink and this was associated with the tear film. These reports were based on data obtained from small numbers of relatively young healthy subjects. It would be useful to gain a better understanding of any alterations of astigmatism, within the central optical zone of the cornea, driven by changes in the tear film after the blink in patients with ocular surface disease.
The aim of this study was, to determine if topographic changes in the central region of the tear film during exposure following the natural blink were orthogonal. If so, could the changes be modelled to predict the likely response following the blink in cases of ocular surface disease and in a control group of healthy individuals free of ocular surface disease.
Methods and materials
This was a prospective, consecutive, randomized, masked, non-interventional observational study conducted at the Eye Hospital/Department of Ophthalmology, Medical University of Silesia, Katowice, Poland. The study was approved by the Ethics Committee of Medical University of Silesia, Katowice, Poland. The tenets of the Helsinki agreement were followed throughout. Informed consent was obtained from all patients after they were fully informed about the study and procedures. All patients gave permissions to use their anonymized data.
Patients were separated into one of two groups, younger (up to 50 years) and older (>50 years) groups. A demarcation of 50 years was chosen as this was close to the median age of the bulk of patients treated in the Department. Over the course of this investigation a total of 69 subjects were enrolled. The subjects consisted of 50 patients with confirmed ocular surface disease (OSD) and 19 control subjects with absolutely no signs of OSD or history of any ocular or systemic condition known to affect the tear film. None of the OSD patients had any history of ocular surgery or signs of abnormal corneal topography. Of the OSD patients, twelve were aged between 20 and 50 years and thirty-eight were over 50 years of age.
Patients with OSD were recruited from the hospital outpatient clinic over the course of the investigation. These patients were defined as dry according to the TFOS/DEWS guidelines [ ]. All had a history of reduced tear stability, low tear volume, incidence of ocular surface staining, bulbar redness, subjective dry eye symptoms and OSDI scores > 12. The volunteer patients were asked to refrain from using any treatment prescribed for their OSD condition for at least 6 h prior to corneal topography assessment. None had any history of ocular surgery, partial or complete punctal occlusion. The age and gender matched controls free of systemic conditions known to affect the tear film, without any objective or subjective symptoms or signs of OSD, and OSDI score ≤10. None of the control subjects > 50 years of age totally fulfilled these conditions over the duration of this study. Thus, it was decided the control group would consist of subjects between 20–50 years of age.
Ocular surface topography
In all cases the ocular surface of the right eye was assessed using a single Placido-disc based video-corneal topographer. After making the initial adjustments, focusing on the right eye and familiarizing the subject with the procedure, the subject was asked to blink and look straight ahead at the fixation target, while keep both eyes open all times, until instructed to rest. A stop-clock was switched on once the subject opened her/his eyes and corneal topography maps were captured on four occasions after opening the eyes: Immediately after opening the eyes (time zero, T0) thereafter five (T5), ten (T10) and fifteen (T15) seconds later. In all cases topography was measured using a recently serviced and calibrated single video-corneal topographer (Tomey Version 4.2C, Tomey GmbH, Nürnberg, Germany). The captured images were stored in the memory banks of the same single device. The computed SimK values, and corresponding axes, over the central 3 mm optical zone of the cornea were harvested and transferred on to a spreadsheet at the end of the period of data acquisition. These values are produced from the analysis of captured topographic images by the algorithm built into topographer. The SimK values just describe the averaged sphero-cylindrical surface powers and axes over the central optical zone of the ocular surface. All measurements were obtained in a draft free environment of relative humidity 50–60 % and ambient temperature 20 22°c.
Data and statistical analysis
All data were entered on spread sheet for statistical analysis and separated into three groups as follows:
Group 1, subjects with ocular surface disease aged between 20 and 50 years (n = 12)
Group 2, subjects with ocular surface disease aged over 50 years (n = 38)
Group3, controls devoid of any signs or symptoms of OSD (n = 19).
Treatment of SimK data
Suppose the ocular surface astigmatism according to the SimK data changed from −1.00 DCx165 to −1.50 DCx165 between times T t and T t+5 then, it follows, the change in the astigmatic component of the tear film was −0.50 DCx165. Thus, the astigmatic power of the tear film formed between times T t and T t+x was −0.50 DCx165. However, simple subtraction cannot be used when there is a change in the axis of astigmatism. Methods based on vector analysis must be employed for the numerical assessment when changes in both astigmatic power and axis occur. The SimK data were subjected to vector analysis using two methods. Firstly, a simple adjustment [ ] of the procedure advanced by Alpins [ ]. Secondly, the procedure proposed by Thibos et al. [ ]. The first technique renders astigmatic data in a polar format while the second expresses the same data in a Cartesian format.
The first technique calculates the exact power and axis of any astigmatic component of the tear film formed between two points in time (T t and T t+5 ). In addition, the difference (Δθ) between the angle of the astigmatism at the ocular surface at T t and the axis of any astigmatic component of the tear film formed over the period until T t+5 can also be determined. This would indicate if the direction, either clockwise or anticlockwise, of any change in the astigmatic component of the tear film was skewed relative to the axis of any astigmatism at the ocular surface at a T t following the blink. This would show if there was a systematic shift of the axis of the astigmatic component of the tear film between two set time points following the blink. The astigmatic power of the change in the tear film was calculated for each time interval after the blink (T0-T5, T0-T10, T0-T15, T5-T10, T5-T15 and T10-T15) in all subjects.
The second technique requires the calculation of J 0 and J 45 vectors as described elsewhere [ ]. This process brings together individual pairs of astigmatic power and axis data into single figures (J 0 and J 45 ). Thus, the astigmatic data is made amenable to relatively simple statistical scrutiny. It is possible to compare the astigmatism at the ocular surface at any moment (T t ) following the blink and the exact power and axis of the astigmatic component of the tear film that was formed over a period after the blink in a relatively simple manner. The correlation between J 0 at time T t and J 0 at time T t+5 would be one if there was no absolute change in astigmatism. Hence, it was decided to compare the change (Δ) in the J 0 and J 45 vectors over the various periods T t and T t+5 with the J 0 and J 45 vectors at T t . This would point out if the magnitude of any astigmatism of the tear film formed during any time interval was predictable. Two separate pre-existing algorithms were used to calculate the two sets of vectors even though it is technically possible to derive the results of the first procedure from the second.
After a preliminary exploration of the data, a decision was made to concentrate on the astigmatic components of the tear film over the following periods T0-T5, T5-T10 and T10-T15. Data were then analyzed to determine the significance of changes in the astigmatism of the tear film formed within each group and between groups. Appropriate non-parametric tests were applied when data were not normally distributed (Friedman test for repeat measures, Kruskal Wallis test). The results obtained from the vector analysis of the SimK data within each group were further analysed as follows:
i) To determine the correlation between the change (ΔJ) in the astigmatic vector (J 0 and J 45 ) over each period (T0-T5, T5-T10, T10-T15) and astigmatic vector value at the start of the period within each of the groups (Pearson correlation).
ii) To determine the significance any correlation between exact power of the astigmatic component of the tear film formed over the period (T0-T5, T5-T10, T10-T15) and astigmatic power at the ocular surface at the start of each period (Pearson correlation).
iii) To determine the significance any correlation between the axis of the astigmatic component of the tear film formed over each period (T0-T5, T5-T10, T10-T15) and axis of the astigmatic power at the ocular surface at the start of each period (Pearson correlation).
And finally, data were scrutinized to determine the significance any rotation of the tear cylinder axis formed over the period T5-T10 from its orientation at T5, and over the period T10-T15 from its orientation at T10.
Changes and differences were considered statistically significant when p < 0.05.
The mean (± SD, range) ages (years) of the subjects recruited over the course of the study were 39.5 (± 7.3, 24 − 48) in group 1, 65.8 (±7.9, 51 − 81) in group 2 and 36.0 (±7.1, 22 − 45) in group 3. The main results are shown in Tables 1–3 and Figs. 1–3 . Glancing over Figs. 1 and 2 , the astigmatic power of the tear film formed was up to −6.00 DC over a period of 5 s in some instances the astigmatism was most prominent over T5-T10. In group 1, the astigmatism was most prominent over T5-T10. In group 2, the astigmatism was less pronounced over T0-T5. Fig. 3 shows, in group 3 the astigmatic power was generally under −0.50 DC exceeding −2.00 DC in a few instances. In all three groups there was a centrifugal shift from the origin after the blink. Astigmatic power appears to increase with continuing exposure.
|Power (DC) |
Mean, ±SD, CI 95%
|Axis (°) |
Mean, ± SD, CI 95%
|T0-T5||−0.81, 0.99, −1.44 to −0.17||87, 60, 49 to 125|
|T5-T10||−2.65, 1.36, −3.52 to −1.79||72, 68, 29 to 115|
|T10-T15||−1.37, 2.15, −2.73 to −0.01||111, 53, 78 to 145|
|T0-T5||−0.33, 0.38, −0.45 to −0.20||94, 52, 76 to 112|
|T5-T10||−0.57, 0.97, −0.91 to −0.24||98, 50, 81 to 115|
|T10-T15||−0.96, 2.10, −1.68 to −0.24||93, 46, 78 to 109|
|T0-T5||−0.57, 0.55, −0.76 to −0.38||77, 53, 51 to 102|
|T5-T10||−0.56, 0.57, −0.76 to −0.37||81, 54, 56 to 107|
|T10-T15||−0.31, 0.44, −0.46 to −0.16||61, 58, 33 to 89|