Prevalence of dry eye symptoms by age in the Beaver Dam Eye Study, 1993 to 1995. P values represent a test of trend.
Prevalence of dry eye symptoms by age and sex in the Beaver Dam Eye Study, 1993 to 1995. P values represent a test of trend.
Customize your JAMA Network experience by selecting one or more topics from the list below.
Moss SE, Klein R, Klein BEK. Prevalence of and Risk Factors for Dry Eye Syndrome. Arch Ophthalmol. 2000;118(9):1264–1268. doi:10.1001/archopht.118.9.1264
Copyright 2000 American Medical Association. All Rights Reserved. Applicable FARS/DFARS Restrictions Apply to Government Use.2000
To examine risk factors for the prevalence of dry eye syndrome in a population-based cohort.
The prevalence of dry eye was determined by history at the second examination (1993-1995) of the Beaver Dam Eye Study cohort (N = 3722).
The cohort was aged 48 to 91 years (mean ± SD, 65 ± 10 years) and 43% male. The overall prevalence of dry eye was 14.4%. Prevalence varied from 8.4% in subjects younger than 60 years to 19.0% in those older than 80 years (P<.001 for test of trend). Age-adjusted prevalence in men was 11.4% compared with 16.7% in women (P<.001). After controlling for age and sex, the following factors were independently and significantly associated with dry eye in a logistic model: history of arthritis (odds ratio [OR], 1.91; 95% confidence interval [CI], 1.56-2.33), smoking status (past, OR, 1.22; 95% CI, 0.97-1.52; current, OR, 1.82; 95% CI, 1.36-2.46), caffeine use (OR, 0.75; 95% CI, 0.61-0.91), history of thyroid disease (OR, 1.41; 95% CI, 1.09-1.84), history of gout (OR, 1.42; 95% CI, 1.02-1.96), total to high-density lipoprotein cholesterol ratio (OR, for 1 unit, 0.93; 95% CI, 0.88-0.99), diabetes (OR, 1.38; 95% CI, 1.03-1.86), and multivitamin use (past, OR, 1.35; 95% CI, 1.01-1.81; current, OR, 1.41; 95% CI, 1.09-1.82). Nonsignificant variables included body mass; blood pressure; white blood cell count; hematocrit; history of osteoporosis, stroke, or cardiovascular disease; history of allergies; use of antihistamines, parasympathetics, antidepressants, diuretics, antiemetics, or other drying drugs; alcohol consumption; time spent outdoors; maculopathy; central cataract; and lens surgery.
The results suggest several factors, such as smoking, caffeine use, and multivitamin use, could be studied for preventive or therapeutic efficacy.
KERATOCONJUNCTIVITIS sicca, or dry eye syndrome, is a common complaint among middle-aged and older adults, even in the absence of diagnosed Sjögren syndrome, rheumatoid arthritis, and other autoimmune diseases.1-3 It can be a cause of great discomfort and frustration, yet very little is known about the epidemiology of dry eye syndrome.1,2
Thus, the purpose of this article is to estimate the prevalence of dry eye in the population of the Beaver Dam Eye Study and to explore its relationship with various risk factors. These factors include cardiovascular disease, medications, and lifestyle and environment.
The methods used to identify the Beaver Dam Eye Study population, reasons for nonparticipation, and comparisons between participants and nonparticipants were published previously.4,5 Briefly, a private census of Beaver Dam, Wis, was conducted from September 15, 1987, to May 4, 1988. The census identified 5924 residents between the ages of 43 and 84 years. During a 30-month period beginning on March 1, 1988, 4926 (83.1%) of the eligible residents were examined.4 Beginning March 1, 1993, 5-year follow-up examinations began. Of the 4541 surviving participants, 3684 were examined in the same order as at baseline. In addition, 38 eligible residents who had not participated in the baseline examination were examined at follow-up. Thus, 3722 subjects participated in the 5-year follow-up examination from 1993 to 1995.
Both the baseline and follow-up examinations followed a similar protocol. Informed consent was obtained from each participant at each examination. The examination included a medical history questionnaire, measurement of height, weight, and blood pressure, determination of refractive error and visual acuity, dilation of the pupils, stereoscopic color fundus photographs for evaluation of age-related maculopathy, slitlamp and retroillumination photographs of the lenses for evaluation of cataract, and collection of urine and blood for a series of standard laboratory tests.
Systolic and diastolic blood pressures were the averages of 2 measurements. Hypertension was defined as a systolic blood pressure of 160 mm Hg or greater, a diastolic blood pressure of 95 mm Hg or greater, or a history of hypertension with use of antihypertension medications. Body mass was defined as weight in kilograms divided by the square of height in meters. A subject was considered to have diabetes if he or she gave a history of diabetes mellitus, was treated with insulin or oral hypoglycemic agents or a specialized diet, or was diagnosed during the study period. The criterion for diagnosis was a glycosylated hemoglobin value greater than 2 SDs above the mean for a given age-sex group and a random blood glucose level of higher than 11.1 mmol/L (200 mg/dL). Arthritis, fractures, osteoporosis, gout, thyroid disorder, and stroke were determined by history. A history of cardiovascular disease was defined as a history of angina, heart attack, or stroke. Aspirin consumption was evaluated in terms of both overall usage (taking or not taking aspirin) and daily dosage (not taking aspirin, taking <1 aspirin every 2 days, taking 1 aspirin every 2 days, taking 1 aspirin every day, and taking ≥2 aspirin every day). Heavy drinking was defined as current or past consumption of 4 or more servings of alcoholic beverages daily. The average weekly consumption of alcohol in grams was computed as the sum of alcohol from each 0.355-L (12-oz) serving of beer, 0.118-L (4-oz) serving of wine, and 0.044-L (1.5 oz) serving of liquor or distilled spirits. Each serving of beer, wine, and liquor was considered to contain 12.96 g, 11.48 g, and 14.00 g of alcohol, respectively. A current or ex-smoker was an individual who had smoked at least 100 cigarettes in his or her life. Pack-years smoked was computed as the number of packs (20 cigarettes) smoked each day times the number of years smoked. The average daily consumption of caffeine in milligrams was computed as the sum of caffeine milligrams from each 0.237-L (8-oz) serving of brewed coffee (103 mg), instant coffee (57 mg), hot or iced tea (36 mg), hot chocolate (6 mg), and caffeine-containing soda (46 mg). The heating season was defined as the months of October through March, when indoor heating systems are used. Age-related maculopathy was determined from the stereoscopic fundus photographs by the Wisconsin Age-related Maculopathy Grading System.6 The presence of cataracts was evaluated from the slitlamp and retroillumination photographs. Central cataract was defined as nuclear cataract of grade 4 or 5 or cortical or posterior subcapsular cataract covering at least 25% of the central lens.7 Lens surgery was defined as the absence of the lens from either eye. Glaucoma was defined as a history of glaucoma or use of eye drops for glaucoma. Visual impairment was defined as a visual acuity of 20/40 to 20/200 in the better eye. Blindness was defined as 20/200 or worse.
The presence of dry eye at the time of the 5-year follow-up examination was determined by subject self-reported history of dry eye. History of dry eye was not determined at baseline. Dry eye was defined as a positive response to the question, "For the past 3 months or longer, have you had dry eyes?" For subjects needing further prompting, this was described as a "foreign body sensation with itching and burning, sandy feeling, not related to allergy." Because history of dry eye was not obtained until the 5-year follow-up examination, all analyses were based on data from that examination. Thus, the results are cross-sectional. Age- and sex-adjusted prevalence of dry eye was computed by multiple linear regression with indicator variables for sex and age groups 48 to 59, 60 to 69, 70 to 79, and 80 to 91 years. The proportion of males of 0.434 and the proportions for the 4 age groups of 0.348, 0.294, 0.254, and 0.104, were used in the calculations. Mantel-Haenszel procedures, stratified by age and sex, were used to test for trends and general associations in age- and sex-adjusted prevalences.8 Logistic regression was used to examine the association of several variables with the prevalence of dry eye.
The population examined varied in age from 48 to 91 years. The mean age (±SD) was 65 years (±10). Men comprised 43% of the population, and 99% of subjects were white.
Of the 3722 participants in the 5-year follow-up examination, 19 were missing information on dry eye. Of the remaining 3703, dry eye symptoms were present in 534 (14.4%) (95% confidence interval [CI], 13.3%-15.6%). Dry eye increased with age but changed little after age 70 years (Figure 1). Prevalence of dry eye was higher in women (17.0%) compared with men (11.1%; P<.001). This difference persisted across all ages (Figure 2). Adjusted for age, the prevalence was 11.4% in men and 16.7% in women (P<.001). We found no evidence for an age-sex interaction (P = .26).
Table 1 presents age- and sex-adjusted prevalence of dry eye by subject characteristics that show a significant or nearly significant (P≤.10) association with prevalence of dry eye. Among cardiovascular disease risk factors, serum total to high-density lipoprotein (HDL) cholesterol ratio was inversely associated with dry eye, and diabetes was directly associated. There was also a suggestion of an inverse association of serum total cholesterol with dry eye. Other cardiovascular risk factors that were not significantly associated with dry eye (P>.10) included body mass index, systolic and diastolic blood pressure, hypertension, HDL cholesterol, white blood cell count, hematocrit, history of stroke, and history of cardiovascular disease (data not shown).
Other medical history items associated with dry eye included history of arthritis, fractures, osteoporosis, gout and thyroid disorder (Table 1). People with a history of allergies did not have a significantly higher age- and sex-adjusted prevalence of dry eye (15.6%) compared with people without a history of allergies (14.1%; P = .28). In women, menstrual status and a history of hysterectomy with oophorectomy were not related to dry eye (data not shown). Among medications and supplements, only antidepressants, aspirin, and multivitamins were significantly or nearly significantly associated with age- and sex-adjusted prevalence of dry eye (Table 1). Other medications that were not related to dry eye included angiotensin-converting enzyme inhibitors, α- or β-antiadrenergic agents, antihistamines, antianxiety agents, calcium channel blockers, diuretics, antiemetics, parasympathetic agents, methyldopa, reserpine, and hormone use in postmenopausal women (data not shown). The joint relationship of arthritis and aspirin use with dry eye was examined. After adjusting for age and sex, it was found that arthritis and aspirin dose were each independently associated with dry eye. The age- and sex-adjusted prevalence of dry eye was 10.2% and 12.2% in nonusers and users of aspirin, respectively, in people without arthritis and 19.0% and 20.7%, respectively, in people with arthritis. No interaction was apparent. Among subjects with a history of gout, those not being treated had a significantly higher age- and sex-adjusted prevalence of dry eye (20.8%), whereas those being treated had a similar prevalence (14.6%) to those without gout (14.0%).
Ocular factors that were associated with age- and sex-adjusted prevalence of dry eye included lens surgery, visual acuity, and the use of glasses or contact lenses (Table 1). Age-related maculopathy, central cataract, and a history of glaucoma were not related to prevalence of dry eye (data not shown).
Lifestyle and environmental factors that were examined included alcohol consumption, cigarette smoking, caffeine consumption, climatic season, and time spent outdoors. A history of heavy drinking in the past was associated with a higher prevalence of dry eye (Table 1). However, the current amount of alcohol consumed was not (data not shown). Current cigarette smoking, and possibly past smoking, was related to dry eye, as were pack-years smoked and current packs smoked daily (Table 1). For the latter 2 variables, the association was largely between any reported amount smoked and none. Caffeine use was associated with a lower age- and sex-adjusted prevalence of dry eye (Table 1). Again, the reported amount consumed was largely irrelevant. Subjects who were examined during the months when indoor heating systems were in use reported more dry eye than those examined during the warmer months (Table 1). Finally, there was no association between dry eye and time spent outdoors during the winter or the summer (data not shown).
Logistic regression analysis identified variables that were significantly and independently associated with the prevalence of dry eye symptoms. After age and sex were included in the model, other factors were selected in stepwise fashion. Three levels of categorical variables, such as smoking status and multivitamin use, were represented by a pair of indicator variables comparing past use and current use with nonuse. The results of the analysis are presented in Table 2. The odds for dry eye increased 35% for each additional 10 years of age and increased similarly for women. The odds nearly doubled for persons with a history of arthritis. Also, smokers, past or current users of multivitamins, persons with a history of gout or thyroid disorder, and persons with diabetes were more likely to have dry eye. Persons consuming caffeine and persons with a higher total to HDL cholesterol ratio were less likely to have dry eye. After controlling for these additional correlates, the use of antidepressants and aspirin were no longer significant.
Large population-based studies of dry eye are few.1,2 Clinic-based studies are unlikely to give a true representation of the prevalence of dry eye because cases of the condition would be expected to be overrepresented. The present study has the advantages of being population-based, large, having a broad age range, and having a diverse assortment of correlates to examine.
Nevertheless, there are limitations to the present study. First, we used self-reports of dry eye with no objective testing. The tests commonly performed (fluorescein or rose bengal staining, Schirmer test, or tear film breakup time) lack sensitivity compared with self-reports.2,9 In addition, we found many associations that have previously been reported such as with sex, arthritis, and diabetes, but we did not find an association of dry eye with a history of allergies, a factor that could lead to misclassification. Thus, we believe using subject-reported symptoms is a valid approach. We also were limited in that we were not able to distinguish between dry eye due to deficiencies of tear production and dry eye owing to evaporation. Risk factors may differ between these 2 broad subcategories. However, the principal result of this limitation would be a weakening of any associations found. Another limitation is the cross-sectional design of the study. This precludes us from knowing the antecedent-consequent relationship between risk factor and end point. For instance, the relationship between multivitamin use and dry eye may be the result of people with dry eye taking vitamins in an attempt to affect the condition.
We found 14.4% of the population to have symptoms of dry eye. This compares favorably with the Salisbury Eye Evaluation (SEE) study, which reported the prevalence of dry eye based on symptoms to be 15%.1,10 However, there are differences between the studies. First, the questionnaire of the SEE study asked about the frequencies of 6 symptoms. Dry eye was considered to be present if 1 or more symptoms was reported often or all of the time.1,10 Second, the SEE population was older, with a mean age of 73.5 years and varying from 65 to 84 years, and 15% of the population was black. These differences may affect the comparability of the 2 populations. The Melbourne study reported a lower prevalence of dry eye based on a set of 6 symptoms that were different from those of the SEE study.2,10 That study found 7.4% reporting 2 or more symptoms and 5.5% reporting any severe symptom not attributed to hay fever.2 The Melbourne study population was younger, with a mean age of 59 years. However, this cannot entirely account for the difference in prevalence from the current study.
Our finding of an association between older age and an increase in dry eye symptoms is consistent with the Melbourne study.2 This is likely a result of normal changes in tear production and characteristics associated with advancing age. Reductions of tear volume and flow and increases in evaporation have been noted in older people.11 One study suggests it is increased evaporation and subsequent increase in tear film osmolarity with age that is the more important determinant.12 This further suggests meibomian gland dysfunction as the underlying etiologic factor. The absence of an age association in the SEE study may be due to the more restricted age range of the population, which began at 65 years.1
We found that women report more dry eye symptoms than men, a result that again agrees with the Melbourne study2 and differs from the SEE study.1 We note a convergence of prevalence between older women and older men in our data. This effect might account for the lack of a sex difference in the SEE study. However, we could find no evidence of an age-sex interaction. The reason for the differences among the studies remains unexplained.
Dry eye has been noted to be associated with other conditions including arthritis, diabetes, and thyroid disease.2,13 These associations have also been found in the current study, as well as associations with gout and osteoporosis. The association of dry eye with arthritis is independent of aspirin treatment. However, in the case of gout, the association is restricted to subjects who are not being treated for gout. Thus, it is probable that these are undiagnosed cases of arthritis.
The apparent protective effect of the total to HDL cholesterol ratio has not been reported previously, although the presence of cholesterol in the lipid layer of the tears has been noted.14 The lipid layer has been reported to be critical in maintaining tear film.15 It is possible that there is a correlation between serum cholesterol and cholesterol found in the lipid layer of tears.
It has been shown that topically applied xanthines can stimulate tear production, thereby decreasing tear film osmolarity and relieving dry eye symptoms.16 Caffeine is a xanthine, which may explain its protective effect. However, whether ingested caffeine has a similar stimulatory effect is not known. The major source of caffeine reported in this population was coffee. Many people with dry eye may abstain from coffee consumption to avoid the diuretic and drying effects. This would result in an apparent protective effect where none exists.
Cigarette smoking has not been reported as a risk factor for dry eye. After controlling for other independent correlates, we found a nearly 2-fold increase in the odds for dry eye in current cigarette smokers. The most obvious basis for the association is that cigarette smoke acts as a direct irritant in the eyes.
Medications have been implicated as being disposing factors in dry eye.17,18 We looked at the effects of diuretics, antihistamines, antianxiety drugs, antidepressants, antiadrenergic agents, and others. After controlling for age and sex, only antidepressants and aspirin were significantly associated with dry eye. However, after controlling for additional correlates in a multivariable model, these were no longer significant. It is possible that the relationship between these medications and dry eye was not controlled by age and sex alone, but by confounding variables. While people who are more sensitive to certain medications may experience aggravation of dry eye symptoms, these people may represent a subset of users that is too small to exhibit an association in a population study.
The relationship between sex hormones and dry eye is complex. While topical estrogen may be beneficial in treating dry eye,19 androgens may be more important in regulating the production of both the aqueous and lipid components of tears.20-22 This is consistent with our finding of an absence of an association between estrogen replacement treatment and dry eye in postmenopausal women.
In conclusion, symptoms of dry eye are common in the older population. The use of a person's history in studies of dry eye epidemiology is validated by the finding of established relationships such as between dry eye and arthritis. Several new relationships were found including total to HDL cholesterol ratio, cigarette smoking, and multivitamin and caffeine use. These findings may warrant further examination in other population-based and longitudinal studies.
Accepted for publication February 12, 2000.
This research was supported by National Institutes of Health grants EY06594 (Drs R. Klein and B. E. K. Klein) and AG11099 (Karen J. Cruickshanks), and in part by the Research to Prevent Blindness Inc, NY (Dr R. Klein, Senior Scientific Investigator Award).
Corresponding author: Scot E. Moss, MA, Dept of Ophthalmology and Visual Sciences, University of Wisconsin–Madison, 610 N Walnut St, 460 WARF, Madison, WI 53705-2397 (e-mail: email@example.com).
Create a personal account or sign in to: