BASIC INVESTIGATION
Objective Assessment of Squamous Metaplasia of Conjunctival Epithelial Cells as Associated With Soft Contact Lens Wear Versus Non–Lens Wearers Michael J. Doughty, PhD
Purpose: To objectively assess the morphology of superficial conjunctival surface cells showing squamous metaplasia as seen in contact lens wearers.
Methods: Impression cytology samples were taken from interpalpebral bulbar conjunctiva from 23 young adult white European women with an average of 6 years of successful daily soft contact lens wear and 23 non–contact-lens-wearing women of the same age. Cell images were graded, and cell and nuclear dimensions and areas were measured by planimetry. Results: The contact lens wearers had a mean squamous metaplasia grade (Nelson) of 2.74 versus 0.35 for the non–lens wearers, with the cell longest dimension (LONG) being 54.4 6 5.6 mm versus 22.4 6 2.9 mm, respectively (P , 0.001). Similar differences were found for the cell shortest dimension (SHORT) and cell area values (P , 0.001). The cell nuclei in the contact lens wearers, based on measures of cell nuclear length (NUCLONG), were 12.7 6 1.7 mm versus 10.0 6 1.1 mm (P , 0.001) with no evidence of pyknotic (shrunken) nuclei. The nucleo-cytoplasmic ratios had mean values of 3.603 versus 1.282 for the cytoplasm-to-nucleus length (CYT/NUC length) ratio, and 0.075 and 0.319 for the NU/CYT length ratio for contact lens wearers and non-wearers, respectively.
Conclusions: As squamous metaplasia develops, at least in longterm soft contact lens wearers, the conjunctival surface cells become considerably enlarged, but the nuclear size of these cells also increases, rather than showing signs of pyknosis. Key Words: impression cytology, bulbar conjunctiva, human, contact lens, squamous metaplasia, nucleus-to-cytoplasm ratio
sity, that is, the transformation of secretory epithelium to nonsecretory or “keratinized” epithelium. As a result, the conjunctival nongoblet cells appear to enlarge and, as reviewed elsewhere,2 numerous investigators have reported such a change in contact lens wearers. Such assessments, however, have been largely qualitative with investigators assigning a grade for severity of the squamous metaplasia to the conjunctival cells visible in the impression cytology samples. It has usually been implied that such nuclear changes included the development of pyknosis (ie, a shrunken nucleus), although such a conclusion has not been generally supported by objective measures. A pilot conjunctival impression cytology (CIC) study on male contact lens wearers, compared with non-wearers, did not indicate such a nuclear change had occurred.3 Further studies on a small group of female contact lens wearers also revealed enlarged conjunctival cells but not shrunken nuclei,4 a result which further indicated that the pyknosis term has been used by others in a relative sense in that the nuclei appeared smaller relative to the overall size of the cells. With either of these studies,3,4 and across most CIC studies, only limited information is actually available to allow adequate conclusions to be drawn as to whether the nuclei in cells showing squamous metaplasia are “normal” in size, shrunken, or enlarged. The main goal of this study was to further define conjunctival cell morphology in samples where squamous metaplasia was expected to be present (ie, in long-term contact lens wearers), and now compared in detail to agematched female non–contact-lens wearers.
(Cornea 2014;33:1095–1102)
SUBJECTS AND METHODS
M
ost of the healthy conjunctival surface is generally composed of nonsquamous epithelial cells interspersed with goblet cells,1 but a transformation to squamous metaplasia can develop. This also involves a change in goblet cell den-
Received for publication March 21, 2014; revision received June 2, 2014; accepted June 10, 2014. Published online ahead of print July 22, 2014. From the Department of Vision Sciences, Glasgow Caledonian University, Glasgow, Scotland. The author has no funding or conflicts of interest to disclose. Reprints: Michael J. Doughty, PhD, Department of Vision Sciences, Glasgow Caledonian University, Cowcaddens Rd, Glasgow G4 OBA, Scotland (e-mail: [email protected]). Copyright © 2014 by Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Subjects A total of 46 healthy white European female subjects were recruited from optometry students at Glasgow Caledonian University. The protocol was approved by the institutional ethics committee, and a consent form was signed. The exclusion criteria were any systemic health conditions, any previous surgical procedures (including refractive surgery), history of substantial ocular trauma, any substantial corneal or conjunctival disease (eg, dry eye or infections), and also heavy use of cosmetic. Subjects were recruited at their convenience between 10.00 and 16.30 hours, and they were first asked to complete the Glasgow Caledonian University ocular comfort questionnaire,5,6 to obtain details of age, the www.corneajrnl.com |
1095
Doughty
presence of any symptoms and their severity, the use of any medications, and to check for any allergies. The questionnaire also asked for details of smoking, alcohol use, and cosmetic use. The latter was done by asking about the frequency of use of cosmetics based on the number of days per week and the amount based on the number of different products being used. In addition, the subjects were asked to complete a visual analog scale (VAS) to rate their ocular comfort for each eye separately.6 This has a horizontal scale, 14 cm long, marked at one end with very uncomfortable at the left side and with very comfortable at the right side. The subjects were carefully instructed (under supervision) to draw a vertical line as to their comfort level for their left and right eyes. Twenty-three contact lens wearers, with an average age of 22 years (range, 18–28 years) were recruited on the basis of having 4 or more years of successful daily contact lens wear, with no significant history of past or present problems and were only mildly symptomatic, at worse. They reported a successful history of 4 to 10 years of daily regular replacement soft contact lens wear, an average of 5.9 6 1.6 years. The most common history was of daily wear on a planned replacement strategy, usually of 3 months. Two lens wearers reported occasional use of a daily wear disposable option for their lenses. All used multipurpose solutions for their daily lens care. Three of the lens wearers indicated occasional use of contact lens rewetting drops. As initially examined while wearing their contact lenses, most had only slight lens surface coating or deposits, although for 5 subjects, there were some slight traces of mascara on the contact lens surface. Twenty-three control subjects were aged between 19 and 25 years (average, 21 years) and were randomly selected from ongoing and concurrent CIC studies on healthy young adult females who were expected to be largely asymptomatic.
External Eye Assessment and CIC Subjects chose which eye they wished to be sampled. Contact lens wearers removed their lenses a few minutes before the CIC being undertaken, having been advised to
Cornea Volume 33, Number 10, October 2014
bring spectacles for use in the period of a few hours after the procedure. CIC was performed as previously described,3,4,7 on the nasal bulbar conjunctival surface after topical anesthesia with preservative-free oxybuprocaine 0.4% or proxymetacaine 0.5% (both as Minims, Chauvin Pharmaceuticals, Kingston upon Thames, United Kingdom). The cells were collected onto a 10-mm diameter Biopore filter as available in a Millicell-CM unit 0.4 PICM 012550 (Millipore, Co, Cork, Ireland). After removal of 3 small plastic pegs on the unit edge, one side of the filter unit was marked with a vertical line to facilitate placement and later serve as a reference for the limbus-facing side of the collected material. The location aimed for placement of the center of the filter was approximately 5 to 10 mm from the limbus, that is, trying to avoid the limbus as far as possible. After CIC, a confirmation of the location of the CIC sampling site was assessed with sodium fluorescein (Fluorets strips, prewetted with Minims Saline).7 In addition, the overall staining bulbar conjunctiva was assessed using CCLRU (Cornea and Contact Lens Research Unit) grading schemes.8 The subjects were offered preservative-free artificial tears (a single Minims Saline or Minims Artificial Tears containing hydroxyethylcellulose) to use as needed once the effects of the anesthetic had worn off. All subjects were also asked on the next day whether they experienced any adverse reactions or notable discomfort.
Staining, Light Microscopy, and Cell Image Analysis of CIC Samples
Following previous protocol,3,4,7 the filters were air dried for a few minutes, a single drop of 2% glutaraldehyde fixative (in 80 mM cacodylate buffer, pH 7.2–7.4, 320 mOsm/kg) was then applied to the filter surface and left for 15 minutes at room temperature. The filter units were subsequently immersed in 99% methanol for 2 minutes, washed with distilled water for 1 minute, and then stained with Giemsa stain (Sigma product G3032, Sigma, Kinston upon Thames, United Kingdom) for 2 minutes at room temperature. After washing with tap water, the filter surface was examined using an Olympus Vannox light microscope attached to a JVC TK-1280E video camera.
FIGURE 1. Representative example of a Millicell-CM–obtained impression cytology sample from the interpalpebral (exposed) bulbar conjunctiva of a female non– contact lens wearer (A) and the overlay generated from the image. The cell outlines are shown with their nuclei (as inner smaller circles) where clearly discernable or just the cell nuclei (as isolated small circles). The arrows indicate regions where cell overlap and multilayering prevent cells or nuclei from being outlined on the overlay. The scale bar of 100 mm is shown as the length of the horizontally orientated white box (A) and as a horizontal line (B).
1096
| www.corneajrnl.com
2014 Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Squamous Metaplasia in Contact Lens Wearers
FIGURE 2. Representative example of a Millicell-CM–obtained impression cytology sample from the interpalpebral (exposed) bulbar conjunctiva of a female contact lens wearer showing grade 3 squamous metaplasia (A) and the overlay generated from the image. Other details as in Figure 1. The scale bar of 100 mm is shown as the length of the vertically oriented white box (A) and as a horizontal line (B).
From each filter, a representative region where a monolayer of cells predominated was photographed at 200· magnification. Color prints were graded according to a 4-point scale,9,10 which predominantly considers the size and shape of the cells. The assigned grades were further checked and subcategorized according to the nuclear size with 0 indicating completely normal and 3 indicating severe squamous metaplasia.7 A scale bar of 100 mm (including the sample ID number) was affixed to the print; a JPEG image file was generated and projected at a final magnification of approximately 1000·. An overlay was generated with the positions of the cell and nuclear borders (Figs. 1, 2). For all cells, several measurements were taken with a digitizer pad, namely the longest dimension of the cell (LONG), the shortest dimension (SHORT) and the longest dimension of the cell nucleus (NUCLONG), the area of the cells (AREAC), and the area of the nucleus (AREANUC). These are the primary morphometric data.
Data Analysis and Statistics Data were typed into a spreadsheets in Systat v. 11 (Systat Inc, Evanston, IL), adjusted for the scale marker, and then a series of calculations were made to produce secondary morphometric data. These were L:S RATIO, the nucleus-to-cytoplasm length, and cytoplasm-to-nucleus length ratios, as well as the nucleus-to-cytoplasm area ratio.4,11 The calculations for these secondary morphometric data are as detailed in Table 1. Standard descriptive statistics were generated (eg, average values 6 SD), as well as graphical output using Systat. The normality of all data sets was checked using the default Shapiro–Wilk option in Systat, with P . 0.05 being considered normal. Comparisons between data sets were
made, as appropriate, using a 2-sample Student t test or a Friedman rank-order test, with the level of statistical significance being set at P , 0.05.
RESULTS Subject Characteristics and CIC Sampling For all subjects, the bulbar conjunctiva showed no more than a mild hyperemia (grade 0.5–1) Any fluorescein staining was of a trace nature (grade 0.5 or less). Although many subjects, especially the contact lens wearers, had slight signs of fluorescein-highlighted microfolds across the bulbar conjunctiva close to the eyelid margins, these are considered to be normal features.12 There were no obvious signs of Meibomian gland dysfunction. Most subjects reported mild but occasional nonspecific symptoms of dryness or discomfort. From the VAS, the overall ocular comfort in the eye assessed by CIC was good to very good at 88 6 8 (mean 6 SD; range, 70–99 on a scale of 100), with a slight difference in favor of the non–contact-lens wearers (average, 87 6 9 vs. 84 6 10) (P = not significant). The contact lens wearers were thus only mildly symptomatic, and the VAS scores were similar in both eyes. The response to cosmetic use questions indicated lightto-moderate use only. Other than as-needed occasional use of oral analgesics (eg, paracetamol, ibuprofen) for headaches, the only oral medications being currently taken were oral contraceptives, reported by 5 non-wearers and 9 contact lens wearers. Four subjects reported the occasional use of nonprescription oral antihistamines for seasonal allergies (but were not currently taking them), and 1 non–lens-wearing subject reported occasional use of an oral nutritional supplement containing flaxseed oils. None of the subjects reported being
TABLE 1. Secondary Morphometric Measures From Impression Cytology Morphometric Parameter Cell shape Nucleus-to-cell length ratio Cytoplasm-to-nucleus length ratio Nucleus-to-cytoplasm area ratio Cytoplasm-to-nucleus area ratio
2014 Lippincott Williams & Wilkins
Abbreviation
Detail of Calculation
L:S ratio LNLONG ratio LONGLN ratio NU/CYT AREA ratio CYT/NUCAREA ratio
LONG/SHORT. Relative units NUCLONG/LONG. Relative units (LONG-NUCLONG)/NUCLONG. Relative units AREANUC/(AREAC-AREANUC). Relative units (AREAC-AREANUC)/AREANUC. Relative units
www.corneajrnl.com |
1097
Cornea Volume 33, Number 10, October 2014
Doughty
TABLE 2. Primary Morphometric Measures Group No lens wear Contact lens wear
LONG
SHORT
NUCLONG
AREAC
AREANUC
22.4 6 2.9 54.4 6 5.6
17.7 6 2.8 42.5 6 4.8
10.0 6 1.1 12.7 6 1.7
322 6 90 1836 6 377
69 6 13 117 6 21
All data as mean 6 SD.
smokers, whereas 27 indicated light-to-moderate alcohol consumption, usually reported at less than10 units per week. The combined use of the topical anesthetic and the CIC procedure caused no notable side effects. Most of the subjects accepted the offer of being given preservative-free artificial tears after the procedure, and verbally indicated on the next day that the use of these eye drops did improve comfort.
CIC Cells Image and Its Suitability for Morphometric Analysis For morphometry to be undertaken with CIC, regions of filter surface need to be selected where there are conjunctival epithelial cells in a monolayer, or mostly so. Overall, using the Millicell unit, a large area of at least 10 mm2 was usually covered. The middle region of the filter surface often contained multilayers of cells (unsuitable for epithelial cell morphometry) with monolayered regions a little off-center. In Figures 1 and 2 are shown 2 representative and suitable examples of the cells for a non–lens wearer and a soft contact lens wearer, respectively. The scale (magnification) of both images is the same. Figure 1A was from a non–contact-lens wearer where there are slight changes from a perfect grade of 0. The medium power field (200·) shows an example of a highdensity yield of cells, predominantly in a monolayer. There is, however, a fairly large region to the upper left where individual cells cannot be discerned because of multilayering (see the ring of arrows in Fig. 1B) and a number of other much smaller regions where there are just a few cells lying over each other (see single arrows on Fig. 1B). Although the individual cells may not always be clearly visible, their nuclei usually are, and therefore, in Figure 1B, all the nuclei that can be clearly discerned are drawn on the overlay. These
nuclei, where outlined, are nonoverlapping and have remarkably similar sizes. The example in Figure 1 was typical of a non–contactlens wearer in whom 75 to 100 cells could be fairly easily identified. Morphometric analysis of the cells in this example indicated the average cell longest dimension (LONG) to be 21.7 mm, the shorter cell dimension (SHORT) to average 19.0 mm, and the nuclear length (NUCLONG) to average 8.3 mm. Figure 2 shows an example of parts of a CIC sample, from a soft contact lens wearer of 8 years, containing mainly a single layer of cells. Most of the individual cells and their nuclei can be clearly identified over most of the image (see the overlay made of these cells in Fig. 2B). There are however parts of the image (upper and left hand side) where the overall staining is slightly greater (indicating more than 1 layer of cells), and it is simply not possible to make out the individual cell domains (see arrows on the overlay). The scale bar of 100 mm serves to emphasize that the longest dimension was substantial for many of the cells. The longest (LONG) dimensions averaged 65.3 mm, and the shorter dimension average was 47.1 mm. The nuclear longest dimension (NUCLONG) had an average value of 14.4 mm. Visual comparisons (ie, Fig. 2B versus Fig. 1B) and the morphometry show these values to all be obviously larger than those seen in a typical non–lens wearer. Of special note are the nuclear sizes, which are larger than in the non–lens wearer and do not include examples of nuclei that are smaller than in the non–lens wearer. Overall, these characteristics allow classification of the cells in Figure 2 as showing grade 3 squamous metaplasia. Overall, the (squamous metaplasia) cell grades in the non–lens wearers ranged from 0 to 1 with an average of 0.35 6 0.35 (mean 6 SD; median, 0.5). The contact lens wearers all had grade 2 or grade 3 squamous metaplasia, with a mean grade of 2.74 6 0.44 (median, 3).
FIGURE 3. Box plots to show distribution of data for cell longest dimension (LONG) (A), cell shorter dimension SHORT (B), and the cell nuclear length (NUCLONG) (C) for non–contact-lens wearers (NCL) and soft contact lens wearers (SCL). All dimensions are in micrometers.
1098
| www.corneajrnl.com
2014 Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Squamous Metaplasia in Contact Lens Wearers
FIGURE 4. Box plots to show distribution of data for cell area AREAC (A) and nuclear area AREANUC (B) for non–contact-lens wearers (NCL) and soft contact lens wearers (SCL). All dimensions are in square micrometer.
Objective Assessments of Cell and Nuclear Size in Contact Lens Wearers Versus Non–Lens Wearers The average number of cells measured/200· microscope field was 81 for the non–lens wearers and 38 for the contact lens wearers. The primary morphometric data (mean 6 SD) are given in Table 2 and Figures 3 and 4. The mean cell LONG dimension in the contact lens wearers was over 2 times that of the non–lens wearers (54.5 vs. 22.4 mm group mean values) as was the cell SHORT values (42.5 vs. 17.7 mm), with the box plots (Fig. 3A and B, respectively) showing no overlap. The NUCLONG measures were almost one third greater in the contact lens wearers, with group mean values of 12.7 vs. 10.0 mm, respectively (P , 0.001). There was slight overlap in the NUCLONG values from the 2 groups (Fig. 3C). The differences between the 2 groups were even more apparent when area measures were used for the comparisons (Table 2 and Fig. 4). The group mean AREAC value for the contact lens wearers was nearly 6 times that of the non–lens wearers (1836 vs. 322 mm2, P , 0.001; Fig. 4A), whereas the nuclei (AREANUC) values were 70% larger in the contact lens wearers (117 vs. 69 mm2, P , 0.001; Fig. 4B). There was almost no overlap in the AREANUC values when comparing the 2 groups (Fig. 4B), that is, the boxes (625% interquartile range) were separated but the whiskers (61.5 SD from the mean) slightly overlap. There were no outliers in the sets of AREANUC values, and there was no evidence of pyknotic (shrunken) nuclei in any of the cells measured. The lowest AREANUC values were 85 mm2 in the contact lens wearers and 53 mm2 in the non–lens wearers.
Objective Assessments of Size of Cell Nuclei in Relation to Cell Size
The “N/C” (or “N:C”) ratio has been given as a feature of conjunctival cell squamous metaplasia, for example N:C of
.6 for Nelson grade 3 squamous metaplasia.9 There are few indications in the published literature as to how the expected values were derived. Table 3 gives details of the secondary morphometric measures taken to try to elucidate the N/C ratios. If the subjective judgment is of the nuclear length (NUCLONG) in relation to cell length (LONG), then the calculations of the nucleus-to-cell length ratio (LNLONG) indicate a mean value of 0.458 for non–lens wearers and 0.236 for the contact lens wearers (P , 0.001), that is, as a fraction of cell size, that for the nucleus is lower in the squamous metaplasia (Fig. 5A). If considered in a slightly different way as the cytoplasm-to-nucleus length ratio (LONGLN), then the respective mean values were 1.282 and 3.603, respectively (P , 0.001), that is, much larger in the contact lens wearers. As shown in Figure 5B, there was no overlap in the LONGLN values, but 2 of the non–lens wearers did have rather higher LONGLN values (see asterisks on Fig. 5B). Another approach to considering the N/C ratio is to use area measures. Therefore, for example, if this assessment is based on considering the area of the nucleus in relation to the area of the surrounding cytoplasm, then the NU/CYT AREA ratios had mean values of 0.319 in the non–lens wearers and 0.075 in the contact lens wearers (P , 0.001). If the calculations are made the other way round (ie, cytoplasmic area to nuclear area, CYT/NUC AREA ratio), then the respective mean values were 3.77 and 15.79, respectively (P , 0.001). Stated another way, the NU/CYT AREA ratio was close to 4.25 times greater in the non–lens wearers, whereas the CYT/NUC AREA ratio was 4.25 times greater in the contact lens wearers.
DISCUSSION CIC is a simple and clinically applicable technique for assessing conjunctival cells to allow for a minimally invasive
TABLE 3. Secondary Morphometric Measures Group No lens wear Contact lens wear
L:S Ratio
LONGLN Ratio
NU/CYT AREA Ratio
CYT/NUC AREA Ratio
1.290 6 0.112 1.304 6 0.068
1.282 6 0.378 3.603 6 0.781
0.319 6 0.089 0.075 6 0.019
3.77 6 1.57 15.79 6 4.84
All data as mean 6 SD.
2014 Lippincott Williams & Wilkins
www.corneajrnl.com |
1099
Doughty
Cornea Volume 33, Number 10, October 2014
FIGURE 5. Box plots to show distribution of data for nucleus-to-cytoplasm length ratio LNLONG (A) and cytoplasm-to-nucleus length ratio (LONGLN) (B) for non–contact-lens wearers (NCL) and soft contact lens wearers (SCL).
investigation of the health of the conjunctival surface.9,13–15 CIC allows for assessments of the more superficial cells of the conjunctiva. If the conjunctival cells appear to be in a monolayer, then these have to be the cells that were at the surface when the filter was applied.16 The present studies were designed to further define cell size and nucleus-to-cytoplasm ratio changes in squamous metaplasia. Grading schemes for squamous metaplasia state or imply that normal conjunctival cells are small and “round” and that when squamous metaplasia develops, the cells can become elongated (and become enlarged). The present studies confirm the cell enlargement, but that substantial changes in cell dimensions (eg, in cell LONG values) are not accompanied by detectable or predictable changes in the L:S ratio (Fig. 6A). The conjunctival surface cells, in squamous metaplasia, therefore, enlarge without change in shape. The cytoplasm of the grade 3 cells generally stains fainter than for grade 0 cells, suggesting that they are thin (and flattened). The squamous metaplasia of the cells across the exposed portion of the bulbar conjunctiva seems to be very strongly associated with an increase in the nuclear size as well. There was no indication of small (shrunken) nuclei as based on subjective assessment and the morphometry. Moderately enlarged flattened cells can be found across the oculo-mucocutaneous cell junction region (Marx line) of the marginal (eyelid) conjunctiva.17,18 Based on morphometry, it has been estimated that as many as 40% of the cells sampled from this location had pyknotic (shrunken) nuclei with areas
of 25 mm2 or less.17 It remains to be established whether or not such cells can be reliably demonstrated to be present across the exposed bulbar conjunctival surface. At this time, at the very least, caution should be exercised before simply reiterating the grading scheme-based descriptors indicating that pyknotic nuclei are a general feature of squamous metaplasia of the bulbar conjunctiva. In squamous metaplasia of exposed bulbar conjunctiva, the cells enlarge (as do the nuclei) with the net result that the so-called N/C (or N:C) ratio changes. This change seems to be predictable when considered in relation to the overall increase in cell size. This is shown in Figure 6B for the cytoplasm-to-nucleus length ratio in relation to the cell LONG dimension. This N/C metric is the one recommended.4 Measures of cell dimensions (cell LONG) and nuclear length (NUCLONG) can be also be made using free (public domain) software such as Image J, and have acceptable reliability compared with cell overlay methods.19 With the common terminology being nucleo-cytoplasmic (or N/C or N:C) ratio, it might be argued that a metric should be developed that assesses this. For completeness, and in line with earlier studies,4,10 this is included in the present analyses. The primary and secondary morphometric data from the present studies are presented in terms of the squamous metaplasia grades in Table 4. Although an N:C ratio has been considered to be .1:6 in squamous metaplasia,9 no specific guidelines seem to exist as to how it should be estimated or calculated. For obviously enlarged squamous cells, calculations of the
FIGURE 6. Scatter plots to illustrate relationships between cell longest dimension (LONG) and either the cell shape L:S ratio (A) or the cytoplasmto-nucleus length ratio (LONGLN) (B).
1100
| www.corneajrnl.com
2014 Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Squamous Metaplasia in Contact Lens Wearers
TABLE 4. Primary and Secondary Morphometric Measures in Relation to Squamous Metaplasia Grading Squamous Metaplasia Grade (Nelson) Cell Feature Cell LONG Cell SHORT Nuclear length (NUCLONG) Cell area (AREAC) Nuclear area (AREANUC) Cell shape (L:S ratio) Nucleus-to-cytoplasm length ratio (LNLONG ratio) Cytoplasm-to-nucleus length ratio (LONGLN ratio) Nucleus-to-cytoplasm area ratio (NU/CYT AREA ratio) Cytoplasm-to-nucleus area ration (CYT/NUC AREA ratio)
0 19.8 15.2 9.6 240 60 1.327 0.449 1.097 0.372 3.04
6 6 6 6 6 6 6 6 6 6
1 1.3 1.7 0.7 46 3 0.107 0.051 0.239 0.079 0.72
26.5 21.0 10.6 442 79 1.287 0.406 1.617 0.251 5.27
6 6 6 6 6 6 6 6 6 6
2 0.8 2.2 2.0 60 23 0.134 0.082 0.617 0.108 3.12
49.2 36.9 12.6 1242 112 1.361 0.262 3.170 0.089 12.64
6 6 6 6 6 6 6 6 6 6
3 1.3 1.8 1.6 89 21 0.088 0.033 0.566 0.018 3.04
56.4 44.4 12.7 1981 120 1.283 0.226 3.755 0.069 16.89
6 6 6 6 6 6 6 6 6 6
5.3 3.9 1.7 327 21 0.081 0.032 0.803 0.017 4.93
All data as mean 6 SD.
cytoplasm-to-nucleus length ratios only yield values of close to 4, whereas those for cytoplasm-to-nucleus area ratios yield values of close to 16. It would be useful to obtain these data for truly squamous cells (eg, for the corneal epithelium) to assess whether the cytoplasm-to-nucleus length ratio might be even larger. Similarly, further data would be useful to confirm previous ideas,3,4 that this ratio should be close to 1.0 for normal (grade 0) cells. Although the main focus of this study was to compare conjunctival cell morphology when squamous metaplasia was present with cell morphology where it was not expected to be present, the contact lens wearers were deliberately recruited to be those with a longer history of lens wear compared with previous studies.4 This was performed to also be able to address whether or not the condition continues to develop or decline with further lens wear. The present studies further confirm that wearing of soft contact lens (and/or the use of the contact lens care solutions) can be expected to result in a notable squamous metaplasia of the exposed bulbar conjunctiva. All of the contact lens wearers showed substantial cell changes compared with non–lens wearers, with the mean squamous metaplasia grade being 2.74. The outcome of this study (as compared with an analysis of 16 other published studies),2 indicates that squamous metaplasia can continue to develop or persist after an average of nearly 6 years of successful soft contact lens wear. Recent studies also confirm the presence of low-grade squamous metaplasia of the bulbar conjunctiva after 2 years of soft contact lens wear.20 However, as noted previously,2 studies using what seem to be alternate subjective grading schemes have reported only slight squamous metaplasia even after 5 to 7 years of soft contact lens wear,21 or slight-to-moderate squamous metaplasia after an average of nearly 8 years of daily wear of silicone hydrogel lenses.22 However, the present objective analyses of subjects with an average of 6-year soft lens wear does indicate that moderate to substantial squamous metaplasia of the exposed bulbar conjunctiva can be common with such periods of soft contact lens wear. Analyses of the present data set do indicate a weak association between years of contact lens wear and the size of the bulbar conjunctival cells based on 2014 Lippincott Williams & Wilkins
the LONG measures (P = 0.031, r = 0.466). It should be stressed, however, that such cellular changes do not need to be accompanied by an obvious and notable fluorescein staining of the exposed bulbar conjunctiva. Although some slight diffuse staining was routinely noted, especially near the lens edge, specific desiccation-related (3 and 9 o’clock position) staining was not seen in any of the contact lens wearers. The ocular comfort scores are also generally consistent with a general lack of ocular surface staining. In conclusion, the present studies provide a basis for the objective assessment of substantial squamous metaplasia of the superficial cells of the bulbar conjunctiva, as can develop in contact lens wearers. Such assessments need to be done on individuals with squamous metaplasia from other causes (eg, dry eye disease in Sjogren syndrome).
ACKNOWLEDGMENTS The assistance of final year optometry students Alexia Bibollet, Louise Gess, Paula McIntyre, and Eimear McDaid for collection of the impression cytology samples is gratefully acknowledged. REFERENCES 1. Doughty MJ. Comparative anatomy and physiology of the cornea and conjunctiva. In: Martin Herran R, Corrales RM, eds. Ocular Surface. Baton Rouge, LA: CRC Press; 2012:32–78. 2. Doughty MJ. Contact lens wear and the development of squamous metaplasia of the surface cells of the conjunctiva. Eye Contact Lens. 2011;37: 274–281. 3. Doughty MJ, Naase T. Nucleus and cell size changes in human bulbar conjunctival cells after soft contact lens wear, as assessed by impression cytology. Cont Lens Anterior Eye. 2008;31:131–140. 4. Doughty MJ. Objective assessment of contact lens wear-associated conjunctival squamous metaplasia by linear measures of cell size, shape and nucleus-to-cytoplasm ratios. Curr Eye Res. 2011;36:599–606. 5. Doughty MJ, Blades KA, Ibrahim N. Assessment of the number of eye symptoms and the impact of some confounding variables for office staff in non-air-conditioned buildings. Ophthalmic Physiol Opt. 2002;22:143–155. 6. Doughty MJ, Lee CA, Ritchie S, et al. An assessment of the discomfort associated with the use of rose bengal 1 % eyedrops on the human eye: a comparison with saline 0.9 % and a topical ocular anaesthetic. Ophthalmic Physiol Opt. 2007;27:159–167.
www.corneajrnl.com |
1101
Doughty
7. Doughty MJ. Objective assessment of conjunctival squamous metaplasia by measures of cell and nucleus dimensions. Diagn Cytopathol. 2011;39: 409–423. 8. Terry RL, Schnider CM, Holden BA, et al. CCLRU standards for success of daily and extended wear contact lenses. Optom Vis Sci. 1993;70:234–243. 9. Nelson JD. Impression cytology. Cornea. 1988;7:71–81. 10. Blades KJ, Doughty MJ. Comparison of grading schemes to quantitative assessments of nucleus-to-cytoplasmic ratios for human bulbar conjunctival cells collected by impression cytology. Curr Eye Res. 2000;20:335–340. 11. Doughty MJ. Reliability of nucleus-to-cell and nucleus-to-cytoplasm calculations for conjunctival impression cytology specimens. Curr Eye Res. 2012;37:583–591. 12. Doughty MJ, Bergmanson JP. New insights into the surface cells and glands of the conjunctiva and their relevance to the tear film. Optometry. 2003;74:485–500. 13. Tseng SC. Staging of conjunctival squamous metaplasia by impression cytology. Ophthalmology. 1985;92:728–733. 14. Calonge M, Diebold Y, Sáez V, et al. Impression cytology of the ocular surface: a review. Exp Eye Res. 2004;78:457–472. 15. Lopin E, Deveney T, Asbell PA. Impression cytology: recent advances and applications in dry eye disease. Ocul Surf. 2009;7:93–110.
1102
| www.corneajrnl.com
Cornea Volume 33, Number 10, October 2014
16. Doughty MJ. Goblet cells of the normal human bulbar conjunctiva and their assessment by impression cytology sampling. Ocul Surf. 2012;10: 149–169. 17. Doughty MJ. Morphological features of cells along Marx’s line of the marginal conjunctiva of the human eyelid. Clin Exp Optom. 2013;96: 76–84. 18. Jalbert I, Madigan MC, Shao M, et al. Assessing the human lid margin epithelium using impression cytology. Acta Ophthalmol. 2012;90: e547–e552. 19. Doughty MJ. On the use of NIH image J for objective assessment of conjunctival cell and nucleus dimensions of impression cytology samples. Eye Contact Lens. 2011;37:50–56. 20. Diao H, She Z, Cao D, et al. Comparison of tacrolimus, fluorometholone, and saline in mild-to-moderate contact lens-induced papillary conjunctivitis. Adv Ther. 2012;29:645–653. 21. Bhatia RP, Lipika R, Garbyal RS, et al. Corneal and conjunctival impression cytology in soft contact lens wearers. Ann Ophthalmol (Skokie). 2006;38:117–120. 22. Sengor T, Aydin Kurna S, Ozbay N, et al. Contact lens-related dry eye and ocular surface changes with mapping technique in long term silicone hydrogel contact lens wearers. Eur J Ophthalmol. 2012;22(suppl 7):S17–S23.
2014 Lippincott Williams & Wilkins
Objective Assessment of Squamous Metaplasia of Conjunctival Epithelial Cells as Associated With Soft Contact Lens Wear Versus Non–Lens Wearers Michael J. Doughty, PhD
Purpose: To objectively assess the morphology of superficial conjunctival surface cells showing squamous metaplasia as seen in contact lens wearers.
Methods: Impression cytology samples were taken from interpalpebral bulbar conjunctiva from 23 young adult white European women with an average of 6 years of successful daily soft contact lens wear and 23 non–contact-lens-wearing women of the same age. Cell images were graded, and cell and nuclear dimensions and areas were measured by planimetry. Results: The contact lens wearers had a mean squamous metaplasia grade (Nelson) of 2.74 versus 0.35 for the non–lens wearers, with the cell longest dimension (LONG) being 54.4 6 5.6 mm versus 22.4 6 2.9 mm, respectively (P , 0.001). Similar differences were found for the cell shortest dimension (SHORT) and cell area values (P , 0.001). The cell nuclei in the contact lens wearers, based on measures of cell nuclear length (NUCLONG), were 12.7 6 1.7 mm versus 10.0 6 1.1 mm (P , 0.001) with no evidence of pyknotic (shrunken) nuclei. The nucleo-cytoplasmic ratios had mean values of 3.603 versus 1.282 for the cytoplasm-to-nucleus length (CYT/NUC length) ratio, and 0.075 and 0.319 for the NU/CYT length ratio for contact lens wearers and non-wearers, respectively.
Conclusions: As squamous metaplasia develops, at least in longterm soft contact lens wearers, the conjunctival surface cells become considerably enlarged, but the nuclear size of these cells also increases, rather than showing signs of pyknosis. Key Words: impression cytology, bulbar conjunctiva, human, contact lens, squamous metaplasia, nucleus-to-cytoplasm ratio
sity, that is, the transformation of secretory epithelium to nonsecretory or “keratinized” epithelium. As a result, the conjunctival nongoblet cells appear to enlarge and, as reviewed elsewhere,2 numerous investigators have reported such a change in contact lens wearers. Such assessments, however, have been largely qualitative with investigators assigning a grade for severity of the squamous metaplasia to the conjunctival cells visible in the impression cytology samples. It has usually been implied that such nuclear changes included the development of pyknosis (ie, a shrunken nucleus), although such a conclusion has not been generally supported by objective measures. A pilot conjunctival impression cytology (CIC) study on male contact lens wearers, compared with non-wearers, did not indicate such a nuclear change had occurred.3 Further studies on a small group of female contact lens wearers also revealed enlarged conjunctival cells but not shrunken nuclei,4 a result which further indicated that the pyknosis term has been used by others in a relative sense in that the nuclei appeared smaller relative to the overall size of the cells. With either of these studies,3,4 and across most CIC studies, only limited information is actually available to allow adequate conclusions to be drawn as to whether the nuclei in cells showing squamous metaplasia are “normal” in size, shrunken, or enlarged. The main goal of this study was to further define conjunctival cell morphology in samples where squamous metaplasia was expected to be present (ie, in long-term contact lens wearers), and now compared in detail to agematched female non–contact-lens wearers.
(Cornea 2014;33:1095–1102)
SUBJECTS AND METHODS
M
ost of the healthy conjunctival surface is generally composed of nonsquamous epithelial cells interspersed with goblet cells,1 but a transformation to squamous metaplasia can develop. This also involves a change in goblet cell den-
Received for publication March 21, 2014; revision received June 2, 2014; accepted June 10, 2014. Published online ahead of print July 22, 2014. From the Department of Vision Sciences, Glasgow Caledonian University, Glasgow, Scotland. The author has no funding or conflicts of interest to disclose. Reprints: Michael J. Doughty, PhD, Department of Vision Sciences, Glasgow Caledonian University, Cowcaddens Rd, Glasgow G4 OBA, Scotland (e-mail: [email protected]). Copyright © 2014 by Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Subjects A total of 46 healthy white European female subjects were recruited from optometry students at Glasgow Caledonian University. The protocol was approved by the institutional ethics committee, and a consent form was signed. The exclusion criteria were any systemic health conditions, any previous surgical procedures (including refractive surgery), history of substantial ocular trauma, any substantial corneal or conjunctival disease (eg, dry eye or infections), and also heavy use of cosmetic. Subjects were recruited at their convenience between 10.00 and 16.30 hours, and they were first asked to complete the Glasgow Caledonian University ocular comfort questionnaire,5,6 to obtain details of age, the www.corneajrnl.com |
1095
Doughty
presence of any symptoms and their severity, the use of any medications, and to check for any allergies. The questionnaire also asked for details of smoking, alcohol use, and cosmetic use. The latter was done by asking about the frequency of use of cosmetics based on the number of days per week and the amount based on the number of different products being used. In addition, the subjects were asked to complete a visual analog scale (VAS) to rate their ocular comfort for each eye separately.6 This has a horizontal scale, 14 cm long, marked at one end with very uncomfortable at the left side and with very comfortable at the right side. The subjects were carefully instructed (under supervision) to draw a vertical line as to their comfort level for their left and right eyes. Twenty-three contact lens wearers, with an average age of 22 years (range, 18–28 years) were recruited on the basis of having 4 or more years of successful daily contact lens wear, with no significant history of past or present problems and were only mildly symptomatic, at worse. They reported a successful history of 4 to 10 years of daily regular replacement soft contact lens wear, an average of 5.9 6 1.6 years. The most common history was of daily wear on a planned replacement strategy, usually of 3 months. Two lens wearers reported occasional use of a daily wear disposable option for their lenses. All used multipurpose solutions for their daily lens care. Three of the lens wearers indicated occasional use of contact lens rewetting drops. As initially examined while wearing their contact lenses, most had only slight lens surface coating or deposits, although for 5 subjects, there were some slight traces of mascara on the contact lens surface. Twenty-three control subjects were aged between 19 and 25 years (average, 21 years) and were randomly selected from ongoing and concurrent CIC studies on healthy young adult females who were expected to be largely asymptomatic.
External Eye Assessment and CIC Subjects chose which eye they wished to be sampled. Contact lens wearers removed their lenses a few minutes before the CIC being undertaken, having been advised to
Cornea Volume 33, Number 10, October 2014
bring spectacles for use in the period of a few hours after the procedure. CIC was performed as previously described,3,4,7 on the nasal bulbar conjunctival surface after topical anesthesia with preservative-free oxybuprocaine 0.4% or proxymetacaine 0.5% (both as Minims, Chauvin Pharmaceuticals, Kingston upon Thames, United Kingdom). The cells were collected onto a 10-mm diameter Biopore filter as available in a Millicell-CM unit 0.4 PICM 012550 (Millipore, Co, Cork, Ireland). After removal of 3 small plastic pegs on the unit edge, one side of the filter unit was marked with a vertical line to facilitate placement and later serve as a reference for the limbus-facing side of the collected material. The location aimed for placement of the center of the filter was approximately 5 to 10 mm from the limbus, that is, trying to avoid the limbus as far as possible. After CIC, a confirmation of the location of the CIC sampling site was assessed with sodium fluorescein (Fluorets strips, prewetted with Minims Saline).7 In addition, the overall staining bulbar conjunctiva was assessed using CCLRU (Cornea and Contact Lens Research Unit) grading schemes.8 The subjects were offered preservative-free artificial tears (a single Minims Saline or Minims Artificial Tears containing hydroxyethylcellulose) to use as needed once the effects of the anesthetic had worn off. All subjects were also asked on the next day whether they experienced any adverse reactions or notable discomfort.
Staining, Light Microscopy, and Cell Image Analysis of CIC Samples
Following previous protocol,3,4,7 the filters were air dried for a few minutes, a single drop of 2% glutaraldehyde fixative (in 80 mM cacodylate buffer, pH 7.2–7.4, 320 mOsm/kg) was then applied to the filter surface and left for 15 minutes at room temperature. The filter units were subsequently immersed in 99% methanol for 2 minutes, washed with distilled water for 1 minute, and then stained with Giemsa stain (Sigma product G3032, Sigma, Kinston upon Thames, United Kingdom) for 2 minutes at room temperature. After washing with tap water, the filter surface was examined using an Olympus Vannox light microscope attached to a JVC TK-1280E video camera.
FIGURE 1. Representative example of a Millicell-CM–obtained impression cytology sample from the interpalpebral (exposed) bulbar conjunctiva of a female non– contact lens wearer (A) and the overlay generated from the image. The cell outlines are shown with their nuclei (as inner smaller circles) where clearly discernable or just the cell nuclei (as isolated small circles). The arrows indicate regions where cell overlap and multilayering prevent cells or nuclei from being outlined on the overlay. The scale bar of 100 mm is shown as the length of the horizontally orientated white box (A) and as a horizontal line (B).
1096
| www.corneajrnl.com
2014 Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Squamous Metaplasia in Contact Lens Wearers
FIGURE 2. Representative example of a Millicell-CM–obtained impression cytology sample from the interpalpebral (exposed) bulbar conjunctiva of a female contact lens wearer showing grade 3 squamous metaplasia (A) and the overlay generated from the image. Other details as in Figure 1. The scale bar of 100 mm is shown as the length of the vertically oriented white box (A) and as a horizontal line (B).
From each filter, a representative region where a monolayer of cells predominated was photographed at 200· magnification. Color prints were graded according to a 4-point scale,9,10 which predominantly considers the size and shape of the cells. The assigned grades were further checked and subcategorized according to the nuclear size with 0 indicating completely normal and 3 indicating severe squamous metaplasia.7 A scale bar of 100 mm (including the sample ID number) was affixed to the print; a JPEG image file was generated and projected at a final magnification of approximately 1000·. An overlay was generated with the positions of the cell and nuclear borders (Figs. 1, 2). For all cells, several measurements were taken with a digitizer pad, namely the longest dimension of the cell (LONG), the shortest dimension (SHORT) and the longest dimension of the cell nucleus (NUCLONG), the area of the cells (AREAC), and the area of the nucleus (AREANUC). These are the primary morphometric data.
Data Analysis and Statistics Data were typed into a spreadsheets in Systat v. 11 (Systat Inc, Evanston, IL), adjusted for the scale marker, and then a series of calculations were made to produce secondary morphometric data. These were L:S RATIO, the nucleus-to-cytoplasm length, and cytoplasm-to-nucleus length ratios, as well as the nucleus-to-cytoplasm area ratio.4,11 The calculations for these secondary morphometric data are as detailed in Table 1. Standard descriptive statistics were generated (eg, average values 6 SD), as well as graphical output using Systat. The normality of all data sets was checked using the default Shapiro–Wilk option in Systat, with P . 0.05 being considered normal. Comparisons between data sets were
made, as appropriate, using a 2-sample Student t test or a Friedman rank-order test, with the level of statistical significance being set at P , 0.05.
RESULTS Subject Characteristics and CIC Sampling For all subjects, the bulbar conjunctiva showed no more than a mild hyperemia (grade 0.5–1) Any fluorescein staining was of a trace nature (grade 0.5 or less). Although many subjects, especially the contact lens wearers, had slight signs of fluorescein-highlighted microfolds across the bulbar conjunctiva close to the eyelid margins, these are considered to be normal features.12 There were no obvious signs of Meibomian gland dysfunction. Most subjects reported mild but occasional nonspecific symptoms of dryness or discomfort. From the VAS, the overall ocular comfort in the eye assessed by CIC was good to very good at 88 6 8 (mean 6 SD; range, 70–99 on a scale of 100), with a slight difference in favor of the non–contact-lens wearers (average, 87 6 9 vs. 84 6 10) (P = not significant). The contact lens wearers were thus only mildly symptomatic, and the VAS scores were similar in both eyes. The response to cosmetic use questions indicated lightto-moderate use only. Other than as-needed occasional use of oral analgesics (eg, paracetamol, ibuprofen) for headaches, the only oral medications being currently taken were oral contraceptives, reported by 5 non-wearers and 9 contact lens wearers. Four subjects reported the occasional use of nonprescription oral antihistamines for seasonal allergies (but were not currently taking them), and 1 non–lens-wearing subject reported occasional use of an oral nutritional supplement containing flaxseed oils. None of the subjects reported being
TABLE 1. Secondary Morphometric Measures From Impression Cytology Morphometric Parameter Cell shape Nucleus-to-cell length ratio Cytoplasm-to-nucleus length ratio Nucleus-to-cytoplasm area ratio Cytoplasm-to-nucleus area ratio
2014 Lippincott Williams & Wilkins
Abbreviation
Detail of Calculation
L:S ratio LNLONG ratio LONGLN ratio NU/CYT AREA ratio CYT/NUCAREA ratio
LONG/SHORT. Relative units NUCLONG/LONG. Relative units (LONG-NUCLONG)/NUCLONG. Relative units AREANUC/(AREAC-AREANUC). Relative units (AREAC-AREANUC)/AREANUC. Relative units
www.corneajrnl.com |
1097
Cornea Volume 33, Number 10, October 2014
Doughty
TABLE 2. Primary Morphometric Measures Group No lens wear Contact lens wear
LONG
SHORT
NUCLONG
AREAC
AREANUC
22.4 6 2.9 54.4 6 5.6
17.7 6 2.8 42.5 6 4.8
10.0 6 1.1 12.7 6 1.7
322 6 90 1836 6 377
69 6 13 117 6 21
All data as mean 6 SD.
smokers, whereas 27 indicated light-to-moderate alcohol consumption, usually reported at less than10 units per week. The combined use of the topical anesthetic and the CIC procedure caused no notable side effects. Most of the subjects accepted the offer of being given preservative-free artificial tears after the procedure, and verbally indicated on the next day that the use of these eye drops did improve comfort.
CIC Cells Image and Its Suitability for Morphometric Analysis For morphometry to be undertaken with CIC, regions of filter surface need to be selected where there are conjunctival epithelial cells in a monolayer, or mostly so. Overall, using the Millicell unit, a large area of at least 10 mm2 was usually covered. The middle region of the filter surface often contained multilayers of cells (unsuitable for epithelial cell morphometry) with monolayered regions a little off-center. In Figures 1 and 2 are shown 2 representative and suitable examples of the cells for a non–lens wearer and a soft contact lens wearer, respectively. The scale (magnification) of both images is the same. Figure 1A was from a non–contact-lens wearer where there are slight changes from a perfect grade of 0. The medium power field (200·) shows an example of a highdensity yield of cells, predominantly in a monolayer. There is, however, a fairly large region to the upper left where individual cells cannot be discerned because of multilayering (see the ring of arrows in Fig. 1B) and a number of other much smaller regions where there are just a few cells lying over each other (see single arrows on Fig. 1B). Although the individual cells may not always be clearly visible, their nuclei usually are, and therefore, in Figure 1B, all the nuclei that can be clearly discerned are drawn on the overlay. These
nuclei, where outlined, are nonoverlapping and have remarkably similar sizes. The example in Figure 1 was typical of a non–contactlens wearer in whom 75 to 100 cells could be fairly easily identified. Morphometric analysis of the cells in this example indicated the average cell longest dimension (LONG) to be 21.7 mm, the shorter cell dimension (SHORT) to average 19.0 mm, and the nuclear length (NUCLONG) to average 8.3 mm. Figure 2 shows an example of parts of a CIC sample, from a soft contact lens wearer of 8 years, containing mainly a single layer of cells. Most of the individual cells and their nuclei can be clearly identified over most of the image (see the overlay made of these cells in Fig. 2B). There are however parts of the image (upper and left hand side) where the overall staining is slightly greater (indicating more than 1 layer of cells), and it is simply not possible to make out the individual cell domains (see arrows on the overlay). The scale bar of 100 mm serves to emphasize that the longest dimension was substantial for many of the cells. The longest (LONG) dimensions averaged 65.3 mm, and the shorter dimension average was 47.1 mm. The nuclear longest dimension (NUCLONG) had an average value of 14.4 mm. Visual comparisons (ie, Fig. 2B versus Fig. 1B) and the morphometry show these values to all be obviously larger than those seen in a typical non–lens wearer. Of special note are the nuclear sizes, which are larger than in the non–lens wearer and do not include examples of nuclei that are smaller than in the non–lens wearer. Overall, these characteristics allow classification of the cells in Figure 2 as showing grade 3 squamous metaplasia. Overall, the (squamous metaplasia) cell grades in the non–lens wearers ranged from 0 to 1 with an average of 0.35 6 0.35 (mean 6 SD; median, 0.5). The contact lens wearers all had grade 2 or grade 3 squamous metaplasia, with a mean grade of 2.74 6 0.44 (median, 3).
FIGURE 3. Box plots to show distribution of data for cell longest dimension (LONG) (A), cell shorter dimension SHORT (B), and the cell nuclear length (NUCLONG) (C) for non–contact-lens wearers (NCL) and soft contact lens wearers (SCL). All dimensions are in micrometers.
1098
| www.corneajrnl.com
2014 Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Squamous Metaplasia in Contact Lens Wearers
FIGURE 4. Box plots to show distribution of data for cell area AREAC (A) and nuclear area AREANUC (B) for non–contact-lens wearers (NCL) and soft contact lens wearers (SCL). All dimensions are in square micrometer.
Objective Assessments of Cell and Nuclear Size in Contact Lens Wearers Versus Non–Lens Wearers The average number of cells measured/200· microscope field was 81 for the non–lens wearers and 38 for the contact lens wearers. The primary morphometric data (mean 6 SD) are given in Table 2 and Figures 3 and 4. The mean cell LONG dimension in the contact lens wearers was over 2 times that of the non–lens wearers (54.5 vs. 22.4 mm group mean values) as was the cell SHORT values (42.5 vs. 17.7 mm), with the box plots (Fig. 3A and B, respectively) showing no overlap. The NUCLONG measures were almost one third greater in the contact lens wearers, with group mean values of 12.7 vs. 10.0 mm, respectively (P , 0.001). There was slight overlap in the NUCLONG values from the 2 groups (Fig. 3C). The differences between the 2 groups were even more apparent when area measures were used for the comparisons (Table 2 and Fig. 4). The group mean AREAC value for the contact lens wearers was nearly 6 times that of the non–lens wearers (1836 vs. 322 mm2, P , 0.001; Fig. 4A), whereas the nuclei (AREANUC) values were 70% larger in the contact lens wearers (117 vs. 69 mm2, P , 0.001; Fig. 4B). There was almost no overlap in the AREANUC values when comparing the 2 groups (Fig. 4B), that is, the boxes (625% interquartile range) were separated but the whiskers (61.5 SD from the mean) slightly overlap. There were no outliers in the sets of AREANUC values, and there was no evidence of pyknotic (shrunken) nuclei in any of the cells measured. The lowest AREANUC values were 85 mm2 in the contact lens wearers and 53 mm2 in the non–lens wearers.
Objective Assessments of Size of Cell Nuclei in Relation to Cell Size
The “N/C” (or “N:C”) ratio has been given as a feature of conjunctival cell squamous metaplasia, for example N:C of
.6 for Nelson grade 3 squamous metaplasia.9 There are few indications in the published literature as to how the expected values were derived. Table 3 gives details of the secondary morphometric measures taken to try to elucidate the N/C ratios. If the subjective judgment is of the nuclear length (NUCLONG) in relation to cell length (LONG), then the calculations of the nucleus-to-cell length ratio (LNLONG) indicate a mean value of 0.458 for non–lens wearers and 0.236 for the contact lens wearers (P , 0.001), that is, as a fraction of cell size, that for the nucleus is lower in the squamous metaplasia (Fig. 5A). If considered in a slightly different way as the cytoplasm-to-nucleus length ratio (LONGLN), then the respective mean values were 1.282 and 3.603, respectively (P , 0.001), that is, much larger in the contact lens wearers. As shown in Figure 5B, there was no overlap in the LONGLN values, but 2 of the non–lens wearers did have rather higher LONGLN values (see asterisks on Fig. 5B). Another approach to considering the N/C ratio is to use area measures. Therefore, for example, if this assessment is based on considering the area of the nucleus in relation to the area of the surrounding cytoplasm, then the NU/CYT AREA ratios had mean values of 0.319 in the non–lens wearers and 0.075 in the contact lens wearers (P , 0.001). If the calculations are made the other way round (ie, cytoplasmic area to nuclear area, CYT/NUC AREA ratio), then the respective mean values were 3.77 and 15.79, respectively (P , 0.001). Stated another way, the NU/CYT AREA ratio was close to 4.25 times greater in the non–lens wearers, whereas the CYT/NUC AREA ratio was 4.25 times greater in the contact lens wearers.
DISCUSSION CIC is a simple and clinically applicable technique for assessing conjunctival cells to allow for a minimally invasive
TABLE 3. Secondary Morphometric Measures Group No lens wear Contact lens wear
L:S Ratio
LONGLN Ratio
NU/CYT AREA Ratio
CYT/NUC AREA Ratio
1.290 6 0.112 1.304 6 0.068
1.282 6 0.378 3.603 6 0.781
0.319 6 0.089 0.075 6 0.019
3.77 6 1.57 15.79 6 4.84
All data as mean 6 SD.
2014 Lippincott Williams & Wilkins
www.corneajrnl.com |
1099
Doughty
Cornea Volume 33, Number 10, October 2014
FIGURE 5. Box plots to show distribution of data for nucleus-to-cytoplasm length ratio LNLONG (A) and cytoplasm-to-nucleus length ratio (LONGLN) (B) for non–contact-lens wearers (NCL) and soft contact lens wearers (SCL).
investigation of the health of the conjunctival surface.9,13–15 CIC allows for assessments of the more superficial cells of the conjunctiva. If the conjunctival cells appear to be in a monolayer, then these have to be the cells that were at the surface when the filter was applied.16 The present studies were designed to further define cell size and nucleus-to-cytoplasm ratio changes in squamous metaplasia. Grading schemes for squamous metaplasia state or imply that normal conjunctival cells are small and “round” and that when squamous metaplasia develops, the cells can become elongated (and become enlarged). The present studies confirm the cell enlargement, but that substantial changes in cell dimensions (eg, in cell LONG values) are not accompanied by detectable or predictable changes in the L:S ratio (Fig. 6A). The conjunctival surface cells, in squamous metaplasia, therefore, enlarge without change in shape. The cytoplasm of the grade 3 cells generally stains fainter than for grade 0 cells, suggesting that they are thin (and flattened). The squamous metaplasia of the cells across the exposed portion of the bulbar conjunctiva seems to be very strongly associated with an increase in the nuclear size as well. There was no indication of small (shrunken) nuclei as based on subjective assessment and the morphometry. Moderately enlarged flattened cells can be found across the oculo-mucocutaneous cell junction region (Marx line) of the marginal (eyelid) conjunctiva.17,18 Based on morphometry, it has been estimated that as many as 40% of the cells sampled from this location had pyknotic (shrunken) nuclei with areas
of 25 mm2 or less.17 It remains to be established whether or not such cells can be reliably demonstrated to be present across the exposed bulbar conjunctival surface. At this time, at the very least, caution should be exercised before simply reiterating the grading scheme-based descriptors indicating that pyknotic nuclei are a general feature of squamous metaplasia of the bulbar conjunctiva. In squamous metaplasia of exposed bulbar conjunctiva, the cells enlarge (as do the nuclei) with the net result that the so-called N/C (or N:C) ratio changes. This change seems to be predictable when considered in relation to the overall increase in cell size. This is shown in Figure 6B for the cytoplasm-to-nucleus length ratio in relation to the cell LONG dimension. This N/C metric is the one recommended.4 Measures of cell dimensions (cell LONG) and nuclear length (NUCLONG) can be also be made using free (public domain) software such as Image J, and have acceptable reliability compared with cell overlay methods.19 With the common terminology being nucleo-cytoplasmic (or N/C or N:C) ratio, it might be argued that a metric should be developed that assesses this. For completeness, and in line with earlier studies,4,10 this is included in the present analyses. The primary and secondary morphometric data from the present studies are presented in terms of the squamous metaplasia grades in Table 4. Although an N:C ratio has been considered to be .1:6 in squamous metaplasia,9 no specific guidelines seem to exist as to how it should be estimated or calculated. For obviously enlarged squamous cells, calculations of the
FIGURE 6. Scatter plots to illustrate relationships between cell longest dimension (LONG) and either the cell shape L:S ratio (A) or the cytoplasmto-nucleus length ratio (LONGLN) (B).
1100
| www.corneajrnl.com
2014 Lippincott Williams & Wilkins
Cornea Volume 33, Number 10, October 2014
Squamous Metaplasia in Contact Lens Wearers
TABLE 4. Primary and Secondary Morphometric Measures in Relation to Squamous Metaplasia Grading Squamous Metaplasia Grade (Nelson) Cell Feature Cell LONG Cell SHORT Nuclear length (NUCLONG) Cell area (AREAC) Nuclear area (AREANUC) Cell shape (L:S ratio) Nucleus-to-cytoplasm length ratio (LNLONG ratio) Cytoplasm-to-nucleus length ratio (LONGLN ratio) Nucleus-to-cytoplasm area ratio (NU/CYT AREA ratio) Cytoplasm-to-nucleus area ration (CYT/NUC AREA ratio)
0 19.8 15.2 9.6 240 60 1.327 0.449 1.097 0.372 3.04
6 6 6 6 6 6 6 6 6 6
1 1.3 1.7 0.7 46 3 0.107 0.051 0.239 0.079 0.72
26.5 21.0 10.6 442 79 1.287 0.406 1.617 0.251 5.27
6 6 6 6 6 6 6 6 6 6
2 0.8 2.2 2.0 60 23 0.134 0.082 0.617 0.108 3.12
49.2 36.9 12.6 1242 112 1.361 0.262 3.170 0.089 12.64
6 6 6 6 6 6 6 6 6 6
3 1.3 1.8 1.6 89 21 0.088 0.033 0.566 0.018 3.04
56.4 44.4 12.7 1981 120 1.283 0.226 3.755 0.069 16.89
6 6 6 6 6 6 6 6 6 6
5.3 3.9 1.7 327 21 0.081 0.032 0.803 0.017 4.93
All data as mean 6 SD.
cytoplasm-to-nucleus length ratios only yield values of close to 4, whereas those for cytoplasm-to-nucleus area ratios yield values of close to 16. It would be useful to obtain these data for truly squamous cells (eg, for the corneal epithelium) to assess whether the cytoplasm-to-nucleus length ratio might be even larger. Similarly, further data would be useful to confirm previous ideas,3,4 that this ratio should be close to 1.0 for normal (grade 0) cells. Although the main focus of this study was to compare conjunctival cell morphology when squamous metaplasia was present with cell morphology where it was not expected to be present, the contact lens wearers were deliberately recruited to be those with a longer history of lens wear compared with previous studies.4 This was performed to also be able to address whether or not the condition continues to develop or decline with further lens wear. The present studies further confirm that wearing of soft contact lens (and/or the use of the contact lens care solutions) can be expected to result in a notable squamous metaplasia of the exposed bulbar conjunctiva. All of the contact lens wearers showed substantial cell changes compared with non–lens wearers, with the mean squamous metaplasia grade being 2.74. The outcome of this study (as compared with an analysis of 16 other published studies),2 indicates that squamous metaplasia can continue to develop or persist after an average of nearly 6 years of successful soft contact lens wear. Recent studies also confirm the presence of low-grade squamous metaplasia of the bulbar conjunctiva after 2 years of soft contact lens wear.20 However, as noted previously,2 studies using what seem to be alternate subjective grading schemes have reported only slight squamous metaplasia even after 5 to 7 years of soft contact lens wear,21 or slight-to-moderate squamous metaplasia after an average of nearly 8 years of daily wear of silicone hydrogel lenses.22 However, the present objective analyses of subjects with an average of 6-year soft lens wear does indicate that moderate to substantial squamous metaplasia of the exposed bulbar conjunctiva can be common with such periods of soft contact lens wear. Analyses of the present data set do indicate a weak association between years of contact lens wear and the size of the bulbar conjunctival cells based on 2014 Lippincott Williams & Wilkins
the LONG measures (P = 0.031, r = 0.466). It should be stressed, however, that such cellular changes do not need to be accompanied by an obvious and notable fluorescein staining of the exposed bulbar conjunctiva. Although some slight diffuse staining was routinely noted, especially near the lens edge, specific desiccation-related (3 and 9 o’clock position) staining was not seen in any of the contact lens wearers. The ocular comfort scores are also generally consistent with a general lack of ocular surface staining. In conclusion, the present studies provide a basis for the objective assessment of substantial squamous metaplasia of the superficial cells of the bulbar conjunctiva, as can develop in contact lens wearers. Such assessments need to be done on individuals with squamous metaplasia from other causes (eg, dry eye disease in Sjogren syndrome).
ACKNOWLEDGMENTS The assistance of final year optometry students Alexia Bibollet, Louise Gess, Paula McIntyre, and Eimear McDaid for collection of the impression cytology samples is gratefully acknowledged. REFERENCES 1. Doughty MJ. Comparative anatomy and physiology of the cornea and conjunctiva. In: Martin Herran R, Corrales RM, eds. Ocular Surface. Baton Rouge, LA: CRC Press; 2012:32–78. 2. Doughty MJ. Contact lens wear and the development of squamous metaplasia of the surface cells of the conjunctiva. Eye Contact Lens. 2011;37: 274–281. 3. Doughty MJ, Naase T. Nucleus and cell size changes in human bulbar conjunctival cells after soft contact lens wear, as assessed by impression cytology. Cont Lens Anterior Eye. 2008;31:131–140. 4. Doughty MJ. Objective assessment of contact lens wear-associated conjunctival squamous metaplasia by linear measures of cell size, shape and nucleus-to-cytoplasm ratios. Curr Eye Res. 2011;36:599–606. 5. Doughty MJ, Blades KA, Ibrahim N. Assessment of the number of eye symptoms and the impact of some confounding variables for office staff in non-air-conditioned buildings. Ophthalmic Physiol Opt. 2002;22:143–155. 6. Doughty MJ, Lee CA, Ritchie S, et al. An assessment of the discomfort associated with the use of rose bengal 1 % eyedrops on the human eye: a comparison with saline 0.9 % and a topical ocular anaesthetic. Ophthalmic Physiol Opt. 2007;27:159–167.
www.corneajrnl.com |
1101
Doughty
7. Doughty MJ. Objective assessment of conjunctival squamous metaplasia by measures of cell and nucleus dimensions. Diagn Cytopathol. 2011;39: 409–423. 8. Terry RL, Schnider CM, Holden BA, et al. CCLRU standards for success of daily and extended wear contact lenses. Optom Vis Sci. 1993;70:234–243. 9. Nelson JD. Impression cytology. Cornea. 1988;7:71–81. 10. Blades KJ, Doughty MJ. Comparison of grading schemes to quantitative assessments of nucleus-to-cytoplasmic ratios for human bulbar conjunctival cells collected by impression cytology. Curr Eye Res. 2000;20:335–340. 11. Doughty MJ. Reliability of nucleus-to-cell and nucleus-to-cytoplasm calculations for conjunctival impression cytology specimens. Curr Eye Res. 2012;37:583–591. 12. Doughty MJ, Bergmanson JP. New insights into the surface cells and glands of the conjunctiva and their relevance to the tear film. Optometry. 2003;74:485–500. 13. Tseng SC. Staging of conjunctival squamous metaplasia by impression cytology. Ophthalmology. 1985;92:728–733. 14. Calonge M, Diebold Y, Sáez V, et al. Impression cytology of the ocular surface: a review. Exp Eye Res. 2004;78:457–472. 15. Lopin E, Deveney T, Asbell PA. Impression cytology: recent advances and applications in dry eye disease. Ocul Surf. 2009;7:93–110.
1102
| www.corneajrnl.com
Cornea Volume 33, Number 10, October 2014
16. Doughty MJ. Goblet cells of the normal human bulbar conjunctiva and their assessment by impression cytology sampling. Ocul Surf. 2012;10: 149–169. 17. Doughty MJ. Morphological features of cells along Marx’s line of the marginal conjunctiva of the human eyelid. Clin Exp Optom. 2013;96: 76–84. 18. Jalbert I, Madigan MC, Shao M, et al. Assessing the human lid margin epithelium using impression cytology. Acta Ophthalmol. 2012;90: e547–e552. 19. Doughty MJ. On the use of NIH image J for objective assessment of conjunctival cell and nucleus dimensions of impression cytology samples. Eye Contact Lens. 2011;37:50–56. 20. Diao H, She Z, Cao D, et al. Comparison of tacrolimus, fluorometholone, and saline in mild-to-moderate contact lens-induced papillary conjunctivitis. Adv Ther. 2012;29:645–653. 21. Bhatia RP, Lipika R, Garbyal RS, et al. Corneal and conjunctival impression cytology in soft contact lens wearers. Ann Ophthalmol (Skokie). 2006;38:117–120. 22. Sengor T, Aydin Kurna S, Ozbay N, et al. Contact lens-related dry eye and ocular surface changes with mapping technique in long term silicone hydrogel contact lens wearers. Eur J Ophthalmol. 2012;22(suppl 7):S17–S23.
2014 Lippincott Williams & Wilkins
