Abstract
purpose. To compare self-reported dry eye disease across contact lens wearers, spectacles wearers, and clinical emmetropes (i.e., those not requiring refractive correction).
methods. The survey included two symptom questions (dryness and light sensitivity) that inquired about frequency and intensity at three times of day (morning, afternoon, and evening) and a self-perception question (i.e., Do you think you have dry eyes?). Dryness and light sensitivity scales were then calculated, summed, and scored, providing a dry eye classification. Logistic regression (binary and multinomial) and analysis of covariance were used to examine the relation between mode of refractive correction and dry eye status, frequency of symptoms, and diurnal change in symptom intensity.
results. Overall, 893 surveys were completed, and the age-adjusted frequency of dry eye in the sample was 28.7%, with 3.5% of the sample reporting severe symptoms (at least grade 4 of a possible 5 for both symptoms). Contact lens wearers were most likely to report dry eye disease (52.3%), followed by spectacle wearers (23.9%) and clinical emmetropes (7.1%). Adjustment for age and gender showed that contact lens (adjusted odds ratio = 12.37, 95% confidence interval = 7.55–20.26) and spectacle wearers (adjusted odds ratio = 2.06, 95% confidence interval = 1.12–3.80) were more likely than emmetropes to report dry eye problems. After adjustment for age and gender, contact lens wearers were shown to be more likely to experience frequent symptoms and an increase in symptoms throughout the day (F = 51.4, P < 0.0001).
conclusions. The frequency of self-reported dry eye is high, especially in contact lens wearers.
Many recent large-scale, population-based studies have used either self-report or a symptom-based definition of dry eye to examine factors associated with the disease.
1 2 3 4 5 6 7 Using symptom-based definitions, reports have suggested that overall, dry eye disease is prevalent in 0.39% to 33.7% of the population.
2 3 5 6 7 8 9 10 This estimate depends on several factors including the “definition” of dry eye disease and the target sample (e.g., young versus elderly, men versus women).
Studies have generally indicated that contact lens wearers report experiencing dry eye symptoms more frequently than do non-contact lens wearers. For example, McMonnies and Ho
11 were among the first to report that contact lens wearers report significantly more dry eye symptoms than do non-contact lens wearers. Doughty et al.
3 also reported that the percentage of symptomatic contact lens wearers (50.1%) was significantly greater than that of symptomatic nonwearers (21.7%). Begley et al.
12 and Guillon et al.
13 also showed that dry eye disease was more frequent or intense in contact lens wearers than in nonwearers. However, these studies generally were not
specific regarding modes of refractive error correction, other than comparing lens wearers and nonwearers. Thus, the relation between specific modes of refractive correction (other than contact lens wearers) and self-reported dry eye remains unclear. The purpose of this report was to compare self-reported dry eye disease in contact lens wearers, spectacle wearers, and clinical emmetropes.
This research was approved by the Biomedical Institutional Review Board, in accordance with the tenets of the Declaration of Helsinki. The survey consisted of three questions associated with dry eye. The first two questions were symptom-based and pertained to dryness and light sensitivity.
14 15 For these questions, the patients were asked about the frequency of symptoms and to choose one of the following responses: (1) never, (2) infrequently, (3) occasionally, (4) frequently, and (5) constantly. Unless the patient responds “never,” he or she was then asked about the intensity of the symptom at three time points during the day: (1) within the first 2 hours of inserting the lens, (2) in the middle of the day, and (3) at the end of the day. The intensity questions are coded from 1 (not intense at all) to 5 (very intense). The third question was a self-perception question (i.e., “Do you think you have dry eyes?” for spectacle wearers and emmetropes or “Do you think you have dry eyes while wearing your contact lenses?” for contact lens wearers). The response options for this question were yes, no, and unsure.
When scoring the survey, we scored the responses to the two symptom questions by multiplying the frequency of the symptom by the average intensity and a constant derived from a previously reported scoring algorithm.
14 These two composite scores were summed to create the scoring index. If a patients responded “yes” to the self-perception question and scored >0.03, he or she was classified as having dry eye disease; if a patient responded “no” or “unsure” to the self-perception question and scored >1.29, he or she was classified with dry eye disease.
In addition to these questions, age, gender, and type of refractive correction were also recorded. In this regard, individuals wearing spectacles or contact lenses or those who were clinically emmetropic (not requiring refractive correction per patient report) were eligible to complete the survey. Patients who had undergone postoperative refractive surgery were not eligible. Patients were recruited from a large, university-based ophthalmic clinic and from the surrounding community, and the questionnaire was self-administered. Patients attending the clinic who completed the questionnaire were those presenting only for routine ophthalmic care. Patients who came for follow-up or problem-based care were not eligible to complete the survey.
All statistical analyses were performed on computer (SAS, ver. 9.1; SAS Institute, Cary, NC). For all hypothesis testing, P < 0.05 was considered significant. Overall age-adjusted frequencies for self-reported dry eye disease were calculated with the U.S. Census statistics reported in 2000, using the direct adjustment method. The χ2 test was used to determine the relation between gender and mode of correction, whereas analysis of covariance (ANCOVA) was used to determine the relation between age and mode of correction.
Logistic regression was used to examine the relation between mode of refractive correction and self-reported dry eye disease, using the cutoff scoring algorithm, controlling for age and gender.
14 Adjusted odds ratio (aOR), 95% confidence interval, and probability are reported. The Hosmer-Lemeshow goodness-of-fit test was used to examine the final calibration of the model.
16 When the χ
2 value for this test is small (high probability), the model is considered well calibrated (i.e., the model provides accurate estimates of the probability of predicting an individual who report dry eye and one who does not). The discriminative ability of the models was evaluated using the area under the receiver operating characteristic (ROC) curve (i.e., can the model accurately discriminate between those who report dry eye and those who do not?). Discrimination was assessed using the following guidelines for area under the ROC curves: 0.5 indicated no discrimination, between 0.7 and 0.8 indicated acceptable discrimination, between 0.8 and 0.9 indicated excellent discrimination, and greater than 0.9 indicated outstanding discrimination.
16
The frequency of each symptom was reported, stratified by mode of correction. A multinomial logistic regression was then used to model the relation between the frequency of each symptom (i.e., dryness and light sensitivity) and mode of refractive correction. Analysis of covariance (ANCOVA) was used to determine the relation between daily fluctuations in symptom intensity (i.e., dryness and light sensitivity) and mode of refractive correction. All patients who responded “never” experiencing the symptom on the preceding frequency question were excluded from this analysis. Change in symptom intensity was determined as the difference between morning and evening self-reported intensity of each symptom and ANCOVA was then used to determine the relation between mode of refractive correction and mean change in intensity, while controlling for age and gender. Post hoc pair-wise comparisons were made using the Tukey method of multiple comparison, as appropriate.
Eight hundred ninety-three patients were included in the study. Three-hundred ninety-three (44.1%) wore contact lenses, 213 (23.9%) wore spectacles, and 287 (32.1%) were emmetropes wearing no correction. The mean age of the sample was 31.15 ± 12.7 years (range, 15–89). As described in
Table 1 , spectacle wearers were significantly older than contact lens wearers and emmetropes (
P < 0.0001). As expected, the number of females (66.1%) who wore contact lenses was significantly greater than the number of males (33.9%), whereas the distribution of gender was similar across spectacle wearers and emmetropes.
Overall, the age-adjusted frequency of self-reported dry eye in the sample was 28.7%, which was similar to the unadjusted frequency estimate of 29.7% (265 individuals).
Table 2presents multivariate logistic regression analyses of factors associated with dry eye. There were 52.7% of contact lens wearers who reported dry eye, whereas 17.4% of spectacle wearers and 7.3% of emmetropes reported dry eye (χ
2 = 183.3;
P < 0.0001). In multivariate logistic regression modeling
(Table 2) , the mode of correction was strongly related to self-reported dry eye, after controlling for both age and gender. Compared to clinical emmetropes, the aOR for spectacle wearers was 2.06 (95% CI: 1.12–3.80), whereas for contact lens wearers it was 12.37 (95% CI: 7.55–20.26). The aOR comparing contact lens wearers to spectacle wearers was 5.02 (95% CI: 3.83–9.43). Even after controlling for age and mode of correction, females were still nearly twice as likely to report dry eye (aOR: 1.96; 95% CI: 1.39–2.78). The area under the ROC curve was 0.80 for this model (excellent discrimination), and the Hosmer-Lemeshow goodness-of-fit test indicated good calibration (χ
2 = 12.0,
P = 0.15).
Overall, 43.3% of the sample reported dryness at least occasionally. Similarly, 33.6% of the sample reported light sensitivity at least occasionally.
Table 3displays the self-reported frequency of the dryness and light sensitivity symptoms stratified by lens correction. As can be seen, contact lens wearers reported significantly more dryness (68.1%, at least occasionally) than did spectacle wearers (31.9%, at least occasionally) and emmetropes (17.6%, at least occasionally). Mode of correction was strongly related to the reported frequency of dryness (χ
2 = 268.96;
P < 0.0001) after controlling for both age (χ
2 = 7.47,
P = 0.11) and gender (χ
2 = 15.59,
P = 0.004) using multinomial logistic regression. However, spectacle wearers reported experiencing significantly more light sensitivity at least occasionally (45.0%) than clinical emmetropes (33.6%) and contact lens wearers (27.3%). Again, the mode of correction was strongly related to the reported frequency of light sensitivity (χ
2 = 37.23,
P < 0.0001), when controlling for both age (χ
2 = 10.22,
P = 0.04) and gender (χ
2 = 19.46,
P = 0.001) using multinomial logistic regression.
For intensity of dryness, 13.4% of the overall sample reported at least a grade 3 (of a possible 5) in the morning, whereas 26.4% and 48.5% of the overall sample reported the same level of intensity at midday and in the evening, respectively. The intensity of light sensitivity seemed to be more evenly distributed over the course of the day. In this regard, 22.5% of the overall sample reported at least a grade 3 in the morning, and 25.8% and 21.3% reported the same level of intensity at midday and in the evening, respectively. Overall, 3.5% of the sample reported at least a grade 4 severity for both symptoms at the end of the day.
The diurnal change in the intensity of each symptom was calculated as the difference between the morning-to-evening intensity scores. For the dryness symptom, the average change in intensity for contact lens wearers was 1.34 ± 1.32, for spectacle wearers it was 0.94 ± 1.38, and for clinical emmetropes it was 0.51 ± 1.25. ANCOVA, controlling for both age (F = 0.18; P = 0.67) and gender (F = 16.6; P < 0.0001) showed a strong relation between mode of refractive correction and change in the dryness symptom (F = 51.4; P < 0.0001). Post hoc testing showed significant differences between all three pair-wise comparisons (contact lens versus spectacle, P < 0.0001; contact lens versus clinical emmetropia, P < 0.0001; and spectacle versus clinical emmetropia, P = 0.01). Fifteen percent of contact lens wearers reported “very intense” dryness at the end of the day, whereas 7% of spectacle wearers and 1% of clinical emmetropes reported this same severity of dryness.
For light sensitivity, the average diurnal change for contact lens wearers was 0.21 ± 1.22, whereas for spectacle wearers it was 0.02 ± 1.24 and for clinical emmetropes it was −0.26 ± 1.15. ANCOVA, with adjustment for both age (F = 23.5; P < 0.0001) and gender (F = 0.01; P = 0.93), showed a relation between mode of refractive correction and change in light sensitivity (F = 6.27; P = 0.002). Post hoc testing showed significant differences between two pair-wise comparisons (contact lens versus spectacle, P < 0.02; contact lens versus clinical emmetropia, P < 0.005; and spectacle versus clinical emmetropia, P = 0.99). At the end of the day, 9.4% of spectacle wearers reported “very intense” light sensitivity, compared with 7.7% of clinical emmetropes and 6.6% of contact lens wearers.
Data from this study suggest that approximately 29% of the sample experiences reported dry eye disease, and women are more likely than men to report these problems. Of these patients, 3.5% reported severe dryness and light sensitivity at the end of the day. These results seem to be in relatively good agreement with other reports. For example, Lin et al.
10 reported that 33.7% of a sample of Taiwanese patients reported a dry eye symptom often or all the time, and again, women were more likely than men to be affected. Shimmura et al.
1 reported that 33% of a Japanese sample believed they had dry eye disease. Schein et al.
17 showed 14.6% of an elderly sample were symptomatic of dry eye, although there was no association with either age or gender. McCarty et al.
8 reported that 5.5% of a sample in Australia reported “severe” dry eye symptoms, and that women were significantly more likely than men to report problems with dry eyes. Moss et al.
7 showed the prevalence of dry eye by self-report to be 14.4% in the Beaver Dam Eye Study; women were again significantly more likely to report problems with dry eye than men (16.7% vs. 11.4%, respectively). Chia et al.
6 showed that 16.6% of the sample from the Blue Mountains Eye Study reported at least moderate dry eye symptoms. Finally, Schaumberg et al.
5 reported in a cohort from the Women’s Health Study that the age-adjusted prevalence of dry eye among women was 7.8%. Again, we found that approximately 33% of women in this sample reported dry eye, which is much higher than the estimate of Schaumberg et al., probably because of the significant number of contact-lens–wearing females in the present study.
Of interest in this study is that even after controlling for age and mode of refractive correction, women were approximately twice as likely as men to report dry eye than were men. Others have also found that women are between 1.5 and 2 times more likely to report dry eye disease than are men.
6 7 8 10 That being said, no clinical testing (i.e., Schirmer’s test, tear breakup time, ocular surface staining) for dry eye was concurrently performed, making it difficult to rule out a gender-related response bias, wherein previous studies have shown that women are more likely to report symptoms in general than men.
18 19 20 21 This tendency has been shown to be true after adjustment for disease status differences through objective measures of disease.
22 Thus, although this and other studies have shown that women are more affected by dry eye than men, the finding may be a somewhat biased estimate.
Very little detail has been published on comparisons of the frequency of dry eye disease across modes of refractive correction, or more specifically, when comparing emmetropes to those requiring refractive correction through spectacle and contact lens wear. Overall, these data show that approximately 53% of contact lens wearers, 17% of spectacle wearers, and 7% of emmetropes report dry eye disease. Several studies have indicated that approximately 50% of contact lens wearers report dry eye, and indeed, our findings agree with that estimate.
3 12 23 24 After we controlled for age and gender, we found that contact lens wearers were approximately 12 times more likely than emmetropes and about five times more likely than spectacle wearers to report dry eye. However, after we adjusted for age and gender, we found that spectacle lens wearers were still twice as likely as emmetropes to report dry eye disease. Although this finding is not necessarily novel for contact lens wearers, it is noteworthy for spectacle wearers and difficult to explain. However, there were several potential explanations that can be put forward. First, it is likely that many of the spectacle wearers had tried and failed to adjust to contact lenses because of the problems with dryness and discomfort described in this study. Second, it is possible that spectacle wearers had sought more frequent vision care for refractive error (unlike the emmetropes), and this may play a role in explaining the relation between spectacle wear and dry eye. Spectacle lens wearers may have more of an awareness of their ocular health and status than an individual not requiring refractive correction. In addition, those wearing spectacles may have had a previous diagnosis of dry eye disease from their eye care providers, which may have influenced their self-reported dry eye status. Finally, one cannot rule out the potential relation between refractive error and dry eye. Although this was a cross-sectional survey study, it would have been interesting to examine the actual relation between measured refractive error and self-reported dry eye. To our knowledge, this relation has not been addressed in the literature before, and indeed this is an idea that needs further study.
When the responses to the questions regarding the frequency and intensity of individual symptoms were compared, 68% of the contact lens wearers, 32% of spectacle wearers, and 18% of emmetropes reported experiencing dryness, at least occasionally. However, light sensitivity was experienced by approximately 45% of spectacle wearers at least occasionally and by approximately 33.6% and 27.3% of emmetropes and contact lens wearers, respectively. The high percentage of light-sensitive spectacle wearers may be explained by potential glare or reflections associated with spectacle wear, and this increased scatter of light may be perceived by the patient as light sensitivity, although the temporal relation between symptoms and choice of correction cannot truly be determined from this study. For both symptoms, contact lens wearers were much more likely to report an increase in the intensity of the symptom toward the end of the day, which has also been reported by Begley et al.
12
In summary, these data show that a significant number of individuals are affected by dry eye symptoms. Contact lens wearers report very frequent dry eye problems, as do spectacle wearers, which is a novel finding. Progress needs to be continued in developing an understanding of the factors that explain the etiology of this problem.