- Email: [email protected]
Asteroid hyalosis in a population: the Beaver Dam eye study
Asteroid Hyalosis in a Population: The Beaver Dam Eye Study SCOT E. MOSS, MA, RONALD KLEIN, MD, AND BARBARA E. K. KLEIN, MD
● PURPOSE:
To estimate the prevalence of asteroid hyalosis and to examine correlates of asteroid hyalosis in a population-based cohort. ● METHODS: The population of Beaver Dam, Wisconsin, that was 43 to 86 years of age was examined from 1988 to 1990 (n ⴝ 4926). The population is predominantly white (99.4%) northern European. Asteroid hyalosis was determined from stereoscopic fundus photographs of three standard fields. ● RESULTS: Fundus photographs were gradable in 4747 subjects. Asteroid hyalosis was present in 1.2% (95% confidence interval, 0.9,1.5%). In subjects in which it was present, asteroid hyalosis was bilateral in 9%. Prevalence increased significantly (P < .001) with age from 0.2% in subjects 43 to 54 years to 2.9% in subjects 75 to 86 years. After adjusting for age, men were more likely (1.8%) to have asteroid hyalosis than women (0.8%). After adjusting for age and sex, asteroid hyalosis was significantly more likely to be found in subjects with greater body mass (P ⴝ .02) and higher alcohol consumption (P ⴝ .03). There were nonsignificant trends with systolic blood pressure (P ⴝ .07), serum cholesterol (P ⴝ .09), and serum albumin (P ⴝ .09). It was not significantly associated (P > .05) with diastolic blood pressure, hypertension, diabetes, cardiovascular disease history, high-density–lipoprotein cholesterol, serum calcium, cigarette smoking, physical activity, intraocular pressure, or refractive error. These relationships were confirmed in a multivariable logistic model. ● CONCLUSIONS: The current study documents the infrequency of asteroid hyalosis in the population as graded from three photographic fields of the fundus. It does not support previous observations of an association of asteroid hyalosis with diabetes or refractive error. The relevance of the new associations reported (body mass, Accepted for publication Feb 12, 2001. From the Department of Ophthalmology and Visual Sciences, University of Wisconsin Medical School, Madison, Wisconsin. This research is supported by National Institutes of Health grant EY06594 (Drs R Klein and BEK Klein) and, in part, by Research to Prevent Blindness (Dr R Klein, Senior Scientific Investigator Award). Inquiries to Scot E. Moss, MA, Department of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, 610 North Walnut St, 460 WARF, Madison, WI 53705-2397; fax: (608) 263-0279.
70
©
2001 BY
alcohol) remains to be determined. (Am J Ophthalmol 2001;132:70 –75. © 2001 by Elsevier Science Inc. All rights reserved.)
A
STEROID HYALOSIS HAS BEEN RECOGNIZED FOR
more than a century.1 It appears as cream- white spherical bodies distributed throughout the vitreous either randomly or in chains or sheets. It rarely causes any visual symptoms.1 However, earlier reports have associated it with hyperopia,2,3 diabetes,2– 8 hypercholesterolemia,4,5 hypertension,2,3,7 gout,9 and increased serum calcium.10 Some authors have recommended testing for diabetes and high cholesterol when asteroid hyalosis is observed in a patient.3–5,8 Other studies have not duplicated these results.10 –13 Many of these studies have been case series, case-control studies, or performed in clinic populations. Thus, the associations found may not reflect those of the general population. Because of the recommendations made for diagnostic testing when asteroid hyalosis is observed, it is important to confirm the associations on which these are based. Therefore, our purposes were to estimate the prevalence of asteroid hyalosis and to examine these previously reported relationships in a population-based cohort. In addition, we sought other unreported associations in the cohort.
METHODS PREVIOUS REPORTS OF THE BEAVER DAM EYE STUDY HAVE
described the methods used to identify the population and the examination procedures.14,15 Briefly, a private census of Beaver Dam, Wisconsin, was conducted from September 15, 1987, to May 4, 1988. The census identified 5924 persons between the ages of 43 and 84 years. Beginning March 1, 1988, and continuing through 30 months, 4926 of these individuals (83.1%) were examined. Of those participating, 99.4% were white and were predominately of northern European ethnicity. Reasons for nonparticipation and comparisons between participants and nonparticipants appear elsewhere.14 Briefly, nonparticipants were older and more likely to be men than participants. Five years later,
ELSEVIER SCIENCE INC. ALL
RIGHTS RESERVED.
0002-9394/01/$20.00 PII S0002-9394(01)00936-9
survivors were invited to participate in a follow-up examination.15 The examinations followed standard protocols. These included obtaining informed consent, measuring pulse and blood pressure, determination of refractive error and visual acuity, performing an anterior segment examination, measuring intraocular pressure, dilating the pupils, measuring height and weight, obtaining a medical history, performing fundus and lens examinations, obtaining slit-lamp and retroillumination photographs, obtaining stereoscopic fundus photographs of three standard fields of the posterior fundus,16 and collection of urine and blood for a series of standard laboratory tests. There was no protocol for noting asteroid bodies seen during ophthalmoscopy. The fundus photographs were graded for age-related maculopathy using a standard protocol.16 During this procedure, two gradings were performed during which the graders noted other lesions, such as asteroid hyalosis. These were graded as absent, questionably present, or present. The two gradings were compared, and discrepancies were adjudicated by a senior grader. For purposes of analysis, questionable asteroid hyalosis was regarded as absent. The two gradings agreed on the presence of asteroid hyalosis in 57 eyes and disagreed in 16. The senior grader considered asteroid hyalosis to be present in seven of the 16 eyes. In addition, asteroid hyalosis was considered to be present in a person if it was present in either eye. Systolic and diastolic blood pressures were the average of two 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 (kg) divided by the square of height (m). A subject was considered to have diabetes if he or she gave a history of diabetes mellitus, was treated with insulin, oral hypoglycemic agents, or diet, or was newly diagnosed during participation in the study. The criterion for diagnosis was a glycosylated hemoglobin value greater than two standard deviations above the mean for a given age-sex group and a random blood sugar level of greater than 11.1 mmol/l. A history of cardiovascular disease was defined as a history of angina, heart attack, or stroke. The average weekly consumption of alcohol in grams was computed as the sum of alcohol from each 0.355-liter (12-ounce) serving of beer, 0.118-liter (4-ounce) serving of wine, and 0.044-liter (1.5-ounce) 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. For analysis, alcohol consumption was divided into four groups: nondrinkers (0 g per week) and persons who consumed 40 g or less of alcohol per week, more than 40 g to 105 g per week, and more than 105 g per week. Heavy drinking was defined as the consumption of four or more servings of alcohol per day. A current or exsmoker was an individual who had smoked at least 100 cigarettes in his life. Pack-years smoked was computed as the number of VOL. 132, NO. 1
FIGURE 1. Prevalence of asteroid hyalosis by age and sex.
packs (20 cigarettes) smoked each day times the number of years smoked. Myopia and hyperopia were defined as refractive errors less than ⫺1 diopter and greater than ⫹ 1 diopter, respectively. Eyes that were aphakic or had intraocular lens implants were excluded. Age- and sex-adjusted prevalence of asteroid hyalosis was computed by multiple linear regression with indicator variables for sex and the age groups 43 to 54, 55 to 64, 65 to 74, and 75 to 86 years. The proportion of males of 0.439 and of 0.309, 0.267, 0.260, and 0.164 for the four age groups, respectively, were used in the calculations. MantelHaenszel procedures, stratified by age and sex, were used to test for trends and general associations in age- and sexadjusted prevalences.17 The test of trend was applied to variables with a natural ordering, such as categories of blood pressure or cholesterol. The test of general association was applied to variables without an ordering, such as categories of hypertension or smoking history. Logistic regression was used to examine the association of several variables with the prevalence of asteroid hyalosis. Because of the exploratory nature of the analysis, P ⫽ .10 was used as the entry criterion. Also, because of associations among the independent variables, the inclusion of one or several variables may result in others being removed. This may result in more than one model that adequately explains the data. To present the fullness of relationships explored, we include more than one multivariable model.
RESULTS OF THE 4926 PARTICIPANTS, 4747 (96.4%) HAD FUNDUS PHO-
tographs gradable for asteroid hyalosis. Subjects without gradable photographs were older (mean age, 71 years) than subjects with gradable photographs (62 years). After adjusting for age, there was no difference between these two groups with respect to systolic and diastolic blood pressure, body mass, or sex. However, persons without gradable photographs were more likely to have a history of diabetes and cardiovascular disease.
ASTEROID HYALOSIS
71
TABLE 1. Age- and Sex-adjusted Prevalence of Asteroid Hyalosis by Characteristics of the Population Prevalence Characteristic
Body mass (kg/m2)
Systolic blood pressure (mm Hg)
Diastolic blood pressure (mm Hg)
Hypertension Hypertension
Diabetes Cardiovascular disease history Gout history Serum albumin (g/l)
Serum calcium (mmol/l)
Ionized calcium (mmol/l)
Total cholesterol (mmol/l)
High-density–lipoprotein cholesterol (mmol/l)
Total/high-density–lipoprotein cholesterol ratio
Serum uric acid (mol/l)
Value
n
%
P Value
15.0–25.0 25.1–28.1 28.2–31.5 31.6–68.4 71–117 118–129 130–144 145–248 42–69 70–77 78–84 85–127 N Y Normotensive Untreated Treated, blood pressure normal Treated, blood pressure high N Y N Y N Y 23–43 44–46 47–49 50–60 1.40–2.34 2.35–2.42 2.43–2.50 2.51–3.34 0.94–1.20 1.21–1.23 1.24–1.26 1.27–1.66 2.45–5.25 5.30–5.95 6.00–6.70 6.75–15.50 0.15–1.00 1.05–1.25 1.30–1.55 1.60–4.10 1.46–3.63 3.64–4.67 4.68–5.97 5.98–39.00 60–280 290–340 350–420 430–3030
1160 1203 1182 1173 1171 1175 1252 1147 1101 1292 1158 1194 2993 1747 2993 324 1123 296 4313 414 3799 877 4346 383 857 1396 1542 938 1197 1189 1076 1271 961 1510 1432 830 1165 1191 1212 1164 1196 1207 1062 1264 1178 1196 1168 1185 1216 1185 1209 1124
0.55 1.37 1.41 1.53 1.23 0.87 0.86 1.99 1.24 0.79 1.67 1.27 1.06 1.52 1.06 2.08 1.30 1.76 1.24 1.20 1.11 1.83 1.10 2.25 1.81 1.45 0.78 1.09 1.40 1.40 1.17 0.96 1.74 0.89 1.50 0.80 1.20 0.79 1.20 1.74 1.32 0.88 1.47 1.28 1.20 1.35 1.05 1.33 1.01 1.43 1.30 1.18
.02
.07
.37
.11 .24
.98 .18 .05 .09
.29
.25
.09
.80
.80
.60
Continued on next page.
72
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2001
TABLE 1. (Continued) Age- and Sex-adjusted Prevalence of Asteroid Hyalosis by Characteristics of the Population Prevalence Characteristic
White blood cells (⫻109)
Smoking history
Pack-years (years)
Alcohol consumption (g/week)
Heavy drinking Lifestyle Use of steroid drops Refractive error, RE
Refractive error, LE
Intraocular pressure, RE (mm Hg)
Intraocular pressure, LE (mm Hg)
Visual acuity, RE
Visual acuity, LE
Value
n
%
P Value
2.2–5.9 6.0–7.1 7.2–8.4 8.5–43.7 Nonsmoker Exsmoker Current 0 ⬍15 ⬍38 38⫹ 0 ⬍40 ⬍105 105⫹ Never Ever Active Sedentary Never Ever Myope Emmetrope Hyperope Myope Emmetrope Hyperope 1–13 14–15 16–17 18⫹ 1–13 14–15 16–17 18⫹ ⱕ20/40 20/20–20/32 20/16–20/10 ⱕ20/40 20/20–20/32 20/16–20/10
1174 1194 1164 1201 2107 1690 947 2120 786 962 851 2379 813 769 766 3920 817 1122 3625 4404 175 868 1846 1455 840 1883 1494 1281 1218 1098 1169 1355 1251 1057 1110 380 2225 2156 334 2075 2362
0.91 1.21 1.76 1.05 0.90 1.43 1.61 0.92 1.24 1.24 1.92 0.90 1.33 1.22 1.96 1.12 1.76 1.26 1.22 1.24 0.74 0.95 0.73 0.71 0.64 0.57 0.94 0.75 0.36 0.37 0.99 0.71 0.97 0.41 0.66 0.02 0.71 0.65 1.07 0.78 0.60
.47
Asteroid hyalosis was present in 57 participants for an overall prevalence of 1.2% (95% confidence interval, 0.9,1.5%). Among the participants in which it was present, asteroid hyalosis occurred only in the right eye in 42%, only in the left eye in 49%, and bilaterally in 9%. It increased significantly (P ⬍ .001) with age from 0.2% in the 43-year to 54-year age group to 2.9% in the 75-year to 86-year age group (Figure 1). Also, it was more common in men at all ages (Figure 1). The age-adjusted prevalence of asteroid hyalosis in men was 1.8% compared with 0.8% in women (P ⬍ .001). VOL. 132, NO. 1
.21
.10
.03
.13 .96 .60 .53
.32
.49
.57
.46
.55
Five years later, 3684 of the cohort participated in the follow-up examination. Of these, 3458 had photographs gradable for asteroid hyalosis at both examinations and were free of the condition at baseline. Nine persons (0.3%) had developed asteroid hyalosis 5 years later. The 5-year incidence was 0.1%, 0.5%, 0.3%, and 0% in persons aged 43 to 54, 55 to 64, 65 to 74, and 75 to 86 years, respectively, at baseline. The incidence was 0.4% (six of 1511) in men and 0.2% (three of 1947) in women (P ⫽ .19). Because of the small number of incident cases, no further analysis of incidence was pursued.
ASTEROID HYALOSIS
73
Table 1 presents age- and sex-adjusted prevalence of asteroid hyalosis by subject characteristics. Persons with greater body mass, a history of gout, and higher alcohol consumption were more likely to have asteroid hyalosis. The trends were statistically significant (P ⬍ .05) in all cases. In addition, there was some evidence of trends (.05 ⬍ P ⬍ .10) with respect to systolic blood pressure, serum albumin, total cholesterol, and smoking. Further division of the group with diabetes by glycosylated hemoglobin level or duration of diabetes did not reveal any trends with asteroid hyalosis (data not shown). Logistic regression was used to evaluate the independent association of several variables with asteroid hyalosis. Age and sex were included in every model, whereas additional variables from Table 1 were selected in stepwise fashion. Table 2 presents the results of two models that are similarly adequate in explaining the data. In one model, asteroid hyalosis is more likely with older age, in men, with greater body mass, and with higher alcohol consumption. In an alternate model, systolic blood pressure replaces body mass. Initially, alcohol consumption and systolic blood pressure were studied as four groups. However, the analysis indicated that the two middle groups of alcohol consumption and the first three groups of systolic blood pressure did not differ. Therefore, these categories were combined to produce more parsimonious models. Serum albumin, total cholesterol, gout history, and pack years smoked did not reach statistical significance in any model (smallest P ⫽ .18).
serum cholesterol,4,5 hypertension,2,3,7 high serum calcium,10 hyperopia,2,3 and gout.9 We found no evidence to suggest a relationship between asteroid hyalosis and diabetes. The cohorts included in the earlier studies were largely from eye clinic populations where people with diabetes may be over-represented. Similarly, persons with hyperopia may also be over-represented in some clinic populations and produce biased results. Because asteroid hyalosis is composed of complexes of calcium phospholipids, its association with serum cholesterol and calcium is credible.18 –20 However, we found a negative trend with total and ionized serum calcium and only a weak positive association with total serum cholesterol. In addition, there was no association with serum phosphorus. Because we have few cases of asteroid hyalosis, we cannot rule out a modest relationship with cholesterol or gout, but it is unlikely to be related to serum calcium or phosphorus. We also found an association of asteroid hyalosis with systolic blood pressure but not with hypertension. When subjects with hypertension were divided into those who were untreated, treated but still had high blood pressure, or treated and had normal blood pressure (Table 1), it is suggested that people with successful treatment of hypertension have asteroid hyalosis at a frequency more like people without hypertension. The association we found between body mass and asteroid hyalosis is unexplained. To our knowledge, no other research has examined this relationship. We have also found no other reports showing alcohol consumption to be a risk factor for asteroid hyalosis. Furthermore, we are lacking a biologic basis for the connection. Studies have shown the alcohol concentration in the vitreous to be highly correlated with that of blood.21,22 Also, an in vitro study has documented the formation of calcium phosphate particles induced by ethanol.23 However, the ethanol concentrations used in that study were far greater than any physiologic concentrations. Also, alcohol consumption may be a surrogate marker for other behaviors, such as diet, that were not measured. In the Beaver Dam population, asteroid hyalosis was bilateral 9% of the time. Other estimates of the prevalence of bilateral asteroid hyalosis vary from 7% to 22%.2–13 The risk factors we found are all systemic in nature, whereas the unilateral nature of asteroid hyalosis argues for an eyespecific precipitating factor. None of the eye-specific factors we examined, refractive error, ocular trauma, lens surgery, or intraocular pressure, were related to asteroid hyalosis. The basis behind its unilateral nature remains unexplained. Despite the advantages of this population-based study, it does have some limitations. First, the presence of asteroid hyalosis was determined from stereoscopic fundus photographs of a limited number of fields. Because asteroid bodies may be present only in peripheral areas or because the photographs were focused on the retina rather than the vitreous, they might not be detected. Had there been a
DISCUSSION PREVIOUS STUDIES OF ASTEROID HYALOSIS WERE IN CLINIC
populations,2,3,9,11 were case series,4,5,7 or were small casecontrol studies.6,8,10,12,13 Case series or case-control studies cannot provide estimates of prevalence. Prevalence estimates from clinic populations may be biased, because such groups do not represent the general populace. Thus, population-based studies are the best source for prevalence estimates in a general population. To our knowledge, this is the first large population-based study reporting on the prevalence of asteroid hyalosis. The prevalence of asteroid hyalosis in Beaver Dam, 1.2%, is higher than the frequency reported previously in other studies that have varied from 0.24% to 0.90%.2,3,9,11 Other studies have commented on asteroid hyalosis being found primarily in the elderly but without giving age-specific prevalence. One study that examined agespecific frequencies found it to increase with age, a result that we confirm.9 We also found a higher prevalence in men. Several studies duplicate this result,3– 6,10,12 but others have noted no difference between the sexes.7,9,11,13 The pathogenesis of asteroid hyalosis is poorly understood, and few factors have consistently been reported to be associated with this condition. Factors suggested to be associated with asteroid hyalosis include diabetes,2– 8 high 74
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2001
TABLE 2. Odds Ratio and 95% Confidence Interval for Prevalence of Asteroid Hyalosis for Specified Values of Subject Characteristics in Multivariate Logistic Regression
Characteristic
Model 1 Age Sex Body mass group
Alcohol consumption Model 2 Age Sex Systolic blood
Value
Odds Ratio
95% Confidence Interval
10 years Male 25.1–28.1 kg/m2* 28.2–31.5 kg/m2* 31.6–68.4 kg/m2* ⱕ105 g/week† ⬎105 g/week†
2.20 1.68 2.68 3.22 3.57 1.54 2.65
1.66, 2.93 0.92, 3.05 0.96, 7.46 1.17, 8.85 1.30, 9.79 0.81, 2.93 1.24, 5.67
10 y Male ⱖ145 mm Hg‡
2.04 2.05 2.10
1.53, 2.71 1.14, 3.70 1.21, 3.64
1.44 2.28
0.76, 2.72 1.07, 4.87
pressure group Alcohol consumption
†
ⱕ105 g/week ⬎105 g/week†
*Compared with 15.0 –25.0 kg/m2. † Compared with nondrinkers. ‡ Compared with ⱕ144 mm Hg.
protocol for noting asteroid hyalosis seen during the ophthalmoscopic examination, such cases may have been included. Thus, we are probably underestimating the prevalence of asteroid hyalosis. However, this would bias the risk factor relationships only if the ability to detect the asteroid bodies was also related to the risk factors. Otherwise, the only effect would be to weaken any associations. Another limitation is the cross-sectional design of the study. This prevents us from knowing the antecedentconsequent relationship between the risk factors and the end point. Thus, we do not know if the risk factor preceded, occurred simultaneously with, or followed the appearance of asteroid bodies. Finally, the small number of cases with asteroid hyalosis makes it difficult to detect associations with other factors, such as cholesterol and hypertension. The current study found a prevalence of asteroid hyalosis of 1.2% in a middle-aged to elderly population. Prevalence increased with age and was higher in men, results confirming earlier research. However, we could not confirm previous observations of an association of asteroid hyalosis with diabetes. Thus, based on these data, further testing for diabetes in persons in whom asteroid bodies are observed should not be considered. Further research is needed to corroborate the new associations reported with body mass and alcohol consumption. VOL. 132, NO. 1
REFERENCES 1. Duke-Elder S. Diseases of the lens and vitreous. Glaucoma and hypotony. In: Duke-Elder S, editor. System of ophthalmology. Volume 11. London: Kimpton, 1969:324 –326. 2. Potter JW, Newcomb R. Prevalence of asteroid bodies in a VA optometry clinic. J Am Optom Assoc 1980;51:19 –25. 3. Bergren RL, Brown GC, Duker JS. Prevalence and association of asteroid hyalosis with systemic diseases. Am J Ophthalmol 1991;111:289 –293. 4. Smith JL. Asteroid hyalitis. Incidence of diabetes mellitus and hypercholesterolemia. JAMA 1958;168:891– 893. 5. Bard LA. Asteroid hyalitis. Relationship to diabetes and hypercholesterolemia. Am J Ophthalmol 1964;58:239 –242. 6. Wasano T, Hirokawa H, Tagawa H, Trempe CL, Buzney SM. Asteroid hyalosis: posterior vitreous detachment and diabetic retinopathy. Ann Ophthalmol 1987;19:255–258. 7. Lamba PA, Shukla KN. Vascular factors in asteroid hyalopathy. Clinical and biochemical observations. Canad J Ophthalmol 1971;6:46 –51. 8. Cockburn DM. Are vitreous asteroid bodies associated with diabetes mellitus? Am J Optom Physiol Opt 1985;62:40 – 44. 9. Safir A, Dunn SN, Martin RG, Tate GW, Mincey GJ. Is asteroid hyalosis ocular gout? Ann Ophthalmol 1990;22:70 – 77. 10. Jervey ED, Anderson WB Jr. Asteroid hyalitis: a study of serum calcium levels in affected patients. Southern Med J 1965;58:191–194. 11. Hatfield RE, Gastineau CF, Rucker CW. Asteroid bodies in the vitreous: relationship to diabetes and hypercholesterolemia. Proc Mayo Clin 1962;37:513–514. 12. Luxenberg M, Sime D. Relationship of asteroid hyalosis to diabetes mellitus and plasma lipid levels. Am J Ophthalmol 1969;67:406 – 413. 13. Stanley JA, Martin DG. Asteroid hyalosis and the blood glucose. NC Med J 1970;31:183–185. 14. Klein R, Klein BEK, Linton KLP, DeMets DL. The Beaver Dam Eye Study: visual acuity. Ophthalmology 1991;98: 1310 –1315. 15. Klein R, Klein BEK, Lee KE. The changes in visual acuity in a population. The Beaver Dam Eye Study. Ophthalmology 1996;103:1169 –1178. 16. Klein R, Davis MD, Magli YL, et al. The Wisconsin age-related maculopathy grading system. Ophthalmology 1991;98:1128 –1134. 17. Mantel N. Chi-square tests with one degree of freedom: extensions of the Mantel-Haenszel procedure. JASA 1963; 58:690 –700. 18. Topilow HW, Kenyon KR, Takahashi M, Freeman HM, Tolentino FI, Hanninen LA. Asteroid hyalosis. Biomicroscopy, ultrastructure, and composition. Arch Ophthalmol 1982;100:964 –968. 19. Streeten BW. Vitreous asteroid bodies. Ultrastructural characteristics and composition. Arch Ophthalmol 1982;100: 969 –975. 20. Miller H, Miller B, Rabinowitz H, Zonis S, Nir I. Asteroid bodies-an ultrastructural study. Invest Ophthalmol Vis Sci 1983;24:133–136. 21. Winek CL, Esposito FM. Comparative study of ethanol levels in blood versus bone marrow, vitreous humor, bile and urine. Forensic Sci Int 1981;17:27–36. 22. Stone BE, Rooney PA. A study using body fluids to determine blood alcohol. J Anal Tox 1984;8:95–96. 23. Larsen MJ, Thorsen A, Jensen SJ. Ethanol-induced formation of solid calcium phosphates. Calcif Tissue Int 1985;37: 189 –193.
ASTEROID HYALOSIS
75
● PURPOSE:
To estimate the prevalence of asteroid hyalosis and to examine correlates of asteroid hyalosis in a population-based cohort. ● METHODS: The population of Beaver Dam, Wisconsin, that was 43 to 86 years of age was examined from 1988 to 1990 (n ⴝ 4926). The population is predominantly white (99.4%) northern European. Asteroid hyalosis was determined from stereoscopic fundus photographs of three standard fields. ● RESULTS: Fundus photographs were gradable in 4747 subjects. Asteroid hyalosis was present in 1.2% (95% confidence interval, 0.9,1.5%). In subjects in which it was present, asteroid hyalosis was bilateral in 9%. Prevalence increased significantly (P < .001) with age from 0.2% in subjects 43 to 54 years to 2.9% in subjects 75 to 86 years. After adjusting for age, men were more likely (1.8%) to have asteroid hyalosis than women (0.8%). After adjusting for age and sex, asteroid hyalosis was significantly more likely to be found in subjects with greater body mass (P ⴝ .02) and higher alcohol consumption (P ⴝ .03). There were nonsignificant trends with systolic blood pressure (P ⴝ .07), serum cholesterol (P ⴝ .09), and serum albumin (P ⴝ .09). It was not significantly associated (P > .05) with diastolic blood pressure, hypertension, diabetes, cardiovascular disease history, high-density–lipoprotein cholesterol, serum calcium, cigarette smoking, physical activity, intraocular pressure, or refractive error. These relationships were confirmed in a multivariable logistic model. ● CONCLUSIONS: The current study documents the infrequency of asteroid hyalosis in the population as graded from three photographic fields of the fundus. It does not support previous observations of an association of asteroid hyalosis with diabetes or refractive error. The relevance of the new associations reported (body mass, Accepted for publication Feb 12, 2001. From the Department of Ophthalmology and Visual Sciences, University of Wisconsin Medical School, Madison, Wisconsin. This research is supported by National Institutes of Health grant EY06594 (Drs R Klein and BEK Klein) and, in part, by Research to Prevent Blindness (Dr R Klein, Senior Scientific Investigator Award). Inquiries to Scot E. Moss, MA, Department of Ophthalmology and Visual Sciences, University of Wisconsin-Madison, 610 North Walnut St, 460 WARF, Madison, WI 53705-2397; fax: (608) 263-0279.
70
©
2001 BY
alcohol) remains to be determined. (Am J Ophthalmol 2001;132:70 –75. © 2001 by Elsevier Science Inc. All rights reserved.)
A
STEROID HYALOSIS HAS BEEN RECOGNIZED FOR
more than a century.1 It appears as cream- white spherical bodies distributed throughout the vitreous either randomly or in chains or sheets. It rarely causes any visual symptoms.1 However, earlier reports have associated it with hyperopia,2,3 diabetes,2– 8 hypercholesterolemia,4,5 hypertension,2,3,7 gout,9 and increased serum calcium.10 Some authors have recommended testing for diabetes and high cholesterol when asteroid hyalosis is observed in a patient.3–5,8 Other studies have not duplicated these results.10 –13 Many of these studies have been case series, case-control studies, or performed in clinic populations. Thus, the associations found may not reflect those of the general population. Because of the recommendations made for diagnostic testing when asteroid hyalosis is observed, it is important to confirm the associations on which these are based. Therefore, our purposes were to estimate the prevalence of asteroid hyalosis and to examine these previously reported relationships in a population-based cohort. In addition, we sought other unreported associations in the cohort.
METHODS PREVIOUS REPORTS OF THE BEAVER DAM EYE STUDY HAVE
described the methods used to identify the population and the examination procedures.14,15 Briefly, a private census of Beaver Dam, Wisconsin, was conducted from September 15, 1987, to May 4, 1988. The census identified 5924 persons between the ages of 43 and 84 years. Beginning March 1, 1988, and continuing through 30 months, 4926 of these individuals (83.1%) were examined. Of those participating, 99.4% were white and were predominately of northern European ethnicity. Reasons for nonparticipation and comparisons between participants and nonparticipants appear elsewhere.14 Briefly, nonparticipants were older and more likely to be men than participants. Five years later,
ELSEVIER SCIENCE INC. ALL
RIGHTS RESERVED.
0002-9394/01/$20.00 PII S0002-9394(01)00936-9
survivors were invited to participate in a follow-up examination.15 The examinations followed standard protocols. These included obtaining informed consent, measuring pulse and blood pressure, determination of refractive error and visual acuity, performing an anterior segment examination, measuring intraocular pressure, dilating the pupils, measuring height and weight, obtaining a medical history, performing fundus and lens examinations, obtaining slit-lamp and retroillumination photographs, obtaining stereoscopic fundus photographs of three standard fields of the posterior fundus,16 and collection of urine and blood for a series of standard laboratory tests. There was no protocol for noting asteroid bodies seen during ophthalmoscopy. The fundus photographs were graded for age-related maculopathy using a standard protocol.16 During this procedure, two gradings were performed during which the graders noted other lesions, such as asteroid hyalosis. These were graded as absent, questionably present, or present. The two gradings were compared, and discrepancies were adjudicated by a senior grader. For purposes of analysis, questionable asteroid hyalosis was regarded as absent. The two gradings agreed on the presence of asteroid hyalosis in 57 eyes and disagreed in 16. The senior grader considered asteroid hyalosis to be present in seven of the 16 eyes. In addition, asteroid hyalosis was considered to be present in a person if it was present in either eye. Systolic and diastolic blood pressures were the average of two 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 (kg) divided by the square of height (m). A subject was considered to have diabetes if he or she gave a history of diabetes mellitus, was treated with insulin, oral hypoglycemic agents, or diet, or was newly diagnosed during participation in the study. The criterion for diagnosis was a glycosylated hemoglobin value greater than two standard deviations above the mean for a given age-sex group and a random blood sugar level of greater than 11.1 mmol/l. A history of cardiovascular disease was defined as a history of angina, heart attack, or stroke. The average weekly consumption of alcohol in grams was computed as the sum of alcohol from each 0.355-liter (12-ounce) serving of beer, 0.118-liter (4-ounce) serving of wine, and 0.044-liter (1.5-ounce) 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. For analysis, alcohol consumption was divided into four groups: nondrinkers (0 g per week) and persons who consumed 40 g or less of alcohol per week, more than 40 g to 105 g per week, and more than 105 g per week. Heavy drinking was defined as the consumption of four or more servings of alcohol per day. A current or exsmoker was an individual who had smoked at least 100 cigarettes in his life. Pack-years smoked was computed as the number of VOL. 132, NO. 1
FIGURE 1. Prevalence of asteroid hyalosis by age and sex.
packs (20 cigarettes) smoked each day times the number of years smoked. Myopia and hyperopia were defined as refractive errors less than ⫺1 diopter and greater than ⫹ 1 diopter, respectively. Eyes that were aphakic or had intraocular lens implants were excluded. Age- and sex-adjusted prevalence of asteroid hyalosis was computed by multiple linear regression with indicator variables for sex and the age groups 43 to 54, 55 to 64, 65 to 74, and 75 to 86 years. The proportion of males of 0.439 and of 0.309, 0.267, 0.260, and 0.164 for the four age groups, respectively, were used in the calculations. MantelHaenszel procedures, stratified by age and sex, were used to test for trends and general associations in age- and sexadjusted prevalences.17 The test of trend was applied to variables with a natural ordering, such as categories of blood pressure or cholesterol. The test of general association was applied to variables without an ordering, such as categories of hypertension or smoking history. Logistic regression was used to examine the association of several variables with the prevalence of asteroid hyalosis. Because of the exploratory nature of the analysis, P ⫽ .10 was used as the entry criterion. Also, because of associations among the independent variables, the inclusion of one or several variables may result in others being removed. This may result in more than one model that adequately explains the data. To present the fullness of relationships explored, we include more than one multivariable model.
RESULTS OF THE 4926 PARTICIPANTS, 4747 (96.4%) HAD FUNDUS PHO-
tographs gradable for asteroid hyalosis. Subjects without gradable photographs were older (mean age, 71 years) than subjects with gradable photographs (62 years). After adjusting for age, there was no difference between these two groups with respect to systolic and diastolic blood pressure, body mass, or sex. However, persons without gradable photographs were more likely to have a history of diabetes and cardiovascular disease.
ASTEROID HYALOSIS
71
TABLE 1. Age- and Sex-adjusted Prevalence of Asteroid Hyalosis by Characteristics of the Population Prevalence Characteristic
Body mass (kg/m2)
Systolic blood pressure (mm Hg)
Diastolic blood pressure (mm Hg)
Hypertension Hypertension
Diabetes Cardiovascular disease history Gout history Serum albumin (g/l)
Serum calcium (mmol/l)
Ionized calcium (mmol/l)
Total cholesterol (mmol/l)
High-density–lipoprotein cholesterol (mmol/l)
Total/high-density–lipoprotein cholesterol ratio
Serum uric acid (mol/l)
Value
n
%
P Value
15.0–25.0 25.1–28.1 28.2–31.5 31.6–68.4 71–117 118–129 130–144 145–248 42–69 70–77 78–84 85–127 N Y Normotensive Untreated Treated, blood pressure normal Treated, blood pressure high N Y N Y N Y 23–43 44–46 47–49 50–60 1.40–2.34 2.35–2.42 2.43–2.50 2.51–3.34 0.94–1.20 1.21–1.23 1.24–1.26 1.27–1.66 2.45–5.25 5.30–5.95 6.00–6.70 6.75–15.50 0.15–1.00 1.05–1.25 1.30–1.55 1.60–4.10 1.46–3.63 3.64–4.67 4.68–5.97 5.98–39.00 60–280 290–340 350–420 430–3030
1160 1203 1182 1173 1171 1175 1252 1147 1101 1292 1158 1194 2993 1747 2993 324 1123 296 4313 414 3799 877 4346 383 857 1396 1542 938 1197 1189 1076 1271 961 1510 1432 830 1165 1191 1212 1164 1196 1207 1062 1264 1178 1196 1168 1185 1216 1185 1209 1124
0.55 1.37 1.41 1.53 1.23 0.87 0.86 1.99 1.24 0.79 1.67 1.27 1.06 1.52 1.06 2.08 1.30 1.76 1.24 1.20 1.11 1.83 1.10 2.25 1.81 1.45 0.78 1.09 1.40 1.40 1.17 0.96 1.74 0.89 1.50 0.80 1.20 0.79 1.20 1.74 1.32 0.88 1.47 1.28 1.20 1.35 1.05 1.33 1.01 1.43 1.30 1.18
.02
.07
.37
.11 .24
.98 .18 .05 .09
.29
.25
.09
.80
.80
.60
Continued on next page.
72
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2001
TABLE 1. (Continued) Age- and Sex-adjusted Prevalence of Asteroid Hyalosis by Characteristics of the Population Prevalence Characteristic
White blood cells (⫻109)
Smoking history
Pack-years (years)
Alcohol consumption (g/week)
Heavy drinking Lifestyle Use of steroid drops Refractive error, RE
Refractive error, LE
Intraocular pressure, RE (mm Hg)
Intraocular pressure, LE (mm Hg)
Visual acuity, RE
Visual acuity, LE
Value
n
%
P Value
2.2–5.9 6.0–7.1 7.2–8.4 8.5–43.7 Nonsmoker Exsmoker Current 0 ⬍15 ⬍38 38⫹ 0 ⬍40 ⬍105 105⫹ Never Ever Active Sedentary Never Ever Myope Emmetrope Hyperope Myope Emmetrope Hyperope 1–13 14–15 16–17 18⫹ 1–13 14–15 16–17 18⫹ ⱕ20/40 20/20–20/32 20/16–20/10 ⱕ20/40 20/20–20/32 20/16–20/10
1174 1194 1164 1201 2107 1690 947 2120 786 962 851 2379 813 769 766 3920 817 1122 3625 4404 175 868 1846 1455 840 1883 1494 1281 1218 1098 1169 1355 1251 1057 1110 380 2225 2156 334 2075 2362
0.91 1.21 1.76 1.05 0.90 1.43 1.61 0.92 1.24 1.24 1.92 0.90 1.33 1.22 1.96 1.12 1.76 1.26 1.22 1.24 0.74 0.95 0.73 0.71 0.64 0.57 0.94 0.75 0.36 0.37 0.99 0.71 0.97 0.41 0.66 0.02 0.71 0.65 1.07 0.78 0.60
.47
Asteroid hyalosis was present in 57 participants for an overall prevalence of 1.2% (95% confidence interval, 0.9,1.5%). Among the participants in which it was present, asteroid hyalosis occurred only in the right eye in 42%, only in the left eye in 49%, and bilaterally in 9%. It increased significantly (P ⬍ .001) with age from 0.2% in the 43-year to 54-year age group to 2.9% in the 75-year to 86-year age group (Figure 1). Also, it was more common in men at all ages (Figure 1). The age-adjusted prevalence of asteroid hyalosis in men was 1.8% compared with 0.8% in women (P ⬍ .001). VOL. 132, NO. 1
.21
.10
.03
.13 .96 .60 .53
.32
.49
.57
.46
.55
Five years later, 3684 of the cohort participated in the follow-up examination. Of these, 3458 had photographs gradable for asteroid hyalosis at both examinations and were free of the condition at baseline. Nine persons (0.3%) had developed asteroid hyalosis 5 years later. The 5-year incidence was 0.1%, 0.5%, 0.3%, and 0% in persons aged 43 to 54, 55 to 64, 65 to 74, and 75 to 86 years, respectively, at baseline. The incidence was 0.4% (six of 1511) in men and 0.2% (three of 1947) in women (P ⫽ .19). Because of the small number of incident cases, no further analysis of incidence was pursued.
ASTEROID HYALOSIS
73
Table 1 presents age- and sex-adjusted prevalence of asteroid hyalosis by subject characteristics. Persons with greater body mass, a history of gout, and higher alcohol consumption were more likely to have asteroid hyalosis. The trends were statistically significant (P ⬍ .05) in all cases. In addition, there was some evidence of trends (.05 ⬍ P ⬍ .10) with respect to systolic blood pressure, serum albumin, total cholesterol, and smoking. Further division of the group with diabetes by glycosylated hemoglobin level or duration of diabetes did not reveal any trends with asteroid hyalosis (data not shown). Logistic regression was used to evaluate the independent association of several variables with asteroid hyalosis. Age and sex were included in every model, whereas additional variables from Table 1 were selected in stepwise fashion. Table 2 presents the results of two models that are similarly adequate in explaining the data. In one model, asteroid hyalosis is more likely with older age, in men, with greater body mass, and with higher alcohol consumption. In an alternate model, systolic blood pressure replaces body mass. Initially, alcohol consumption and systolic blood pressure were studied as four groups. However, the analysis indicated that the two middle groups of alcohol consumption and the first three groups of systolic blood pressure did not differ. Therefore, these categories were combined to produce more parsimonious models. Serum albumin, total cholesterol, gout history, and pack years smoked did not reach statistical significance in any model (smallest P ⫽ .18).
serum cholesterol,4,5 hypertension,2,3,7 high serum calcium,10 hyperopia,2,3 and gout.9 We found no evidence to suggest a relationship between asteroid hyalosis and diabetes. The cohorts included in the earlier studies were largely from eye clinic populations where people with diabetes may be over-represented. Similarly, persons with hyperopia may also be over-represented in some clinic populations and produce biased results. Because asteroid hyalosis is composed of complexes of calcium phospholipids, its association with serum cholesterol and calcium is credible.18 –20 However, we found a negative trend with total and ionized serum calcium and only a weak positive association with total serum cholesterol. In addition, there was no association with serum phosphorus. Because we have few cases of asteroid hyalosis, we cannot rule out a modest relationship with cholesterol or gout, but it is unlikely to be related to serum calcium or phosphorus. We also found an association of asteroid hyalosis with systolic blood pressure but not with hypertension. When subjects with hypertension were divided into those who were untreated, treated but still had high blood pressure, or treated and had normal blood pressure (Table 1), it is suggested that people with successful treatment of hypertension have asteroid hyalosis at a frequency more like people without hypertension. The association we found between body mass and asteroid hyalosis is unexplained. To our knowledge, no other research has examined this relationship. We have also found no other reports showing alcohol consumption to be a risk factor for asteroid hyalosis. Furthermore, we are lacking a biologic basis for the connection. Studies have shown the alcohol concentration in the vitreous to be highly correlated with that of blood.21,22 Also, an in vitro study has documented the formation of calcium phosphate particles induced by ethanol.23 However, the ethanol concentrations used in that study were far greater than any physiologic concentrations. Also, alcohol consumption may be a surrogate marker for other behaviors, such as diet, that were not measured. In the Beaver Dam population, asteroid hyalosis was bilateral 9% of the time. Other estimates of the prevalence of bilateral asteroid hyalosis vary from 7% to 22%.2–13 The risk factors we found are all systemic in nature, whereas the unilateral nature of asteroid hyalosis argues for an eyespecific precipitating factor. None of the eye-specific factors we examined, refractive error, ocular trauma, lens surgery, or intraocular pressure, were related to asteroid hyalosis. The basis behind its unilateral nature remains unexplained. Despite the advantages of this population-based study, it does have some limitations. First, the presence of asteroid hyalosis was determined from stereoscopic fundus photographs of a limited number of fields. Because asteroid bodies may be present only in peripheral areas or because the photographs were focused on the retina rather than the vitreous, they might not be detected. Had there been a
DISCUSSION PREVIOUS STUDIES OF ASTEROID HYALOSIS WERE IN CLINIC
populations,2,3,9,11 were case series,4,5,7 or were small casecontrol studies.6,8,10,12,13 Case series or case-control studies cannot provide estimates of prevalence. Prevalence estimates from clinic populations may be biased, because such groups do not represent the general populace. Thus, population-based studies are the best source for prevalence estimates in a general population. To our knowledge, this is the first large population-based study reporting on the prevalence of asteroid hyalosis. The prevalence of asteroid hyalosis in Beaver Dam, 1.2%, is higher than the frequency reported previously in other studies that have varied from 0.24% to 0.90%.2,3,9,11 Other studies have commented on asteroid hyalosis being found primarily in the elderly but without giving age-specific prevalence. One study that examined agespecific frequencies found it to increase with age, a result that we confirm.9 We also found a higher prevalence in men. Several studies duplicate this result,3– 6,10,12 but others have noted no difference between the sexes.7,9,11,13 The pathogenesis of asteroid hyalosis is poorly understood, and few factors have consistently been reported to be associated with this condition. Factors suggested to be associated with asteroid hyalosis include diabetes,2– 8 high 74
AMERICAN JOURNAL
OF
OPHTHALMOLOGY
JULY 2001
TABLE 2. Odds Ratio and 95% Confidence Interval for Prevalence of Asteroid Hyalosis for Specified Values of Subject Characteristics in Multivariate Logistic Regression
Characteristic
Model 1 Age Sex Body mass group
Alcohol consumption Model 2 Age Sex Systolic blood
Value
Odds Ratio
95% Confidence Interval
10 years Male 25.1–28.1 kg/m2* 28.2–31.5 kg/m2* 31.6–68.4 kg/m2* ⱕ105 g/week† ⬎105 g/week†
2.20 1.68 2.68 3.22 3.57 1.54 2.65
1.66, 2.93 0.92, 3.05 0.96, 7.46 1.17, 8.85 1.30, 9.79 0.81, 2.93 1.24, 5.67
10 y Male ⱖ145 mm Hg‡
2.04 2.05 2.10
1.53, 2.71 1.14, 3.70 1.21, 3.64
1.44 2.28
0.76, 2.72 1.07, 4.87
pressure group Alcohol consumption
†
ⱕ105 g/week ⬎105 g/week†
*Compared with 15.0 –25.0 kg/m2. † Compared with nondrinkers. ‡ Compared with ⱕ144 mm Hg.
protocol for noting asteroid hyalosis seen during the ophthalmoscopic examination, such cases may have been included. Thus, we are probably underestimating the prevalence of asteroid hyalosis. However, this would bias the risk factor relationships only if the ability to detect the asteroid bodies was also related to the risk factors. Otherwise, the only effect would be to weaken any associations. Another limitation is the cross-sectional design of the study. This prevents us from knowing the antecedentconsequent relationship between the risk factors and the end point. Thus, we do not know if the risk factor preceded, occurred simultaneously with, or followed the appearance of asteroid bodies. Finally, the small number of cases with asteroid hyalosis makes it difficult to detect associations with other factors, such as cholesterol and hypertension. The current study found a prevalence of asteroid hyalosis of 1.2% in a middle-aged to elderly population. Prevalence increased with age and was higher in men, results confirming earlier research. However, we could not confirm previous observations of an association of asteroid hyalosis with diabetes. Thus, based on these data, further testing for diabetes in persons in whom asteroid bodies are observed should not be considered. Further research is needed to corroborate the new associations reported with body mass and alcohol consumption. VOL. 132, NO. 1
REFERENCES 1. Duke-Elder S. Diseases of the lens and vitreous. Glaucoma and hypotony. In: Duke-Elder S, editor. System of ophthalmology. Volume 11. London: Kimpton, 1969:324 –326. 2. Potter JW, Newcomb R. Prevalence of asteroid bodies in a VA optometry clinic. J Am Optom Assoc 1980;51:19 –25. 3. Bergren RL, Brown GC, Duker JS. Prevalence and association of asteroid hyalosis with systemic diseases. Am J Ophthalmol 1991;111:289 –293. 4. Smith JL. Asteroid hyalitis. Incidence of diabetes mellitus and hypercholesterolemia. JAMA 1958;168:891– 893. 5. Bard LA. Asteroid hyalitis. Relationship to diabetes and hypercholesterolemia. Am J Ophthalmol 1964;58:239 –242. 6. Wasano T, Hirokawa H, Tagawa H, Trempe CL, Buzney SM. Asteroid hyalosis: posterior vitreous detachment and diabetic retinopathy. Ann Ophthalmol 1987;19:255–258. 7. Lamba PA, Shukla KN. Vascular factors in asteroid hyalopathy. Clinical and biochemical observations. Canad J Ophthalmol 1971;6:46 –51. 8. Cockburn DM. Are vitreous asteroid bodies associated with diabetes mellitus? Am J Optom Physiol Opt 1985;62:40 – 44. 9. Safir A, Dunn SN, Martin RG, Tate GW, Mincey GJ. Is asteroid hyalosis ocular gout? Ann Ophthalmol 1990;22:70 – 77. 10. Jervey ED, Anderson WB Jr. Asteroid hyalitis: a study of serum calcium levels in affected patients. Southern Med J 1965;58:191–194. 11. Hatfield RE, Gastineau CF, Rucker CW. Asteroid bodies in the vitreous: relationship to diabetes and hypercholesterolemia. Proc Mayo Clin 1962;37:513–514. 12. Luxenberg M, Sime D. Relationship of asteroid hyalosis to diabetes mellitus and plasma lipid levels. Am J Ophthalmol 1969;67:406 – 413. 13. Stanley JA, Martin DG. Asteroid hyalosis and the blood glucose. NC Med J 1970;31:183–185. 14. Klein R, Klein BEK, Linton KLP, DeMets DL. The Beaver Dam Eye Study: visual acuity. Ophthalmology 1991;98: 1310 –1315. 15. Klein R, Klein BEK, Lee KE. The changes in visual acuity in a population. The Beaver Dam Eye Study. Ophthalmology 1996;103:1169 –1178. 16. Klein R, Davis MD, Magli YL, et al. The Wisconsin age-related maculopathy grading system. Ophthalmology 1991;98:1128 –1134. 17. Mantel N. Chi-square tests with one degree of freedom: extensions of the Mantel-Haenszel procedure. JASA 1963; 58:690 –700. 18. Topilow HW, Kenyon KR, Takahashi M, Freeman HM, Tolentino FI, Hanninen LA. Asteroid hyalosis. Biomicroscopy, ultrastructure, and composition. Arch Ophthalmol 1982;100:964 –968. 19. Streeten BW. Vitreous asteroid bodies. Ultrastructural characteristics and composition. Arch Ophthalmol 1982;100: 969 –975. 20. Miller H, Miller B, Rabinowitz H, Zonis S, Nir I. Asteroid bodies-an ultrastructural study. Invest Ophthalmol Vis Sci 1983;24:133–136. 21. Winek CL, Esposito FM. Comparative study of ethanol levels in blood versus bone marrow, vitreous humor, bile and urine. Forensic Sci Int 1981;17:27–36. 22. Stone BE, Rooney PA. A study using body fluids to determine blood alcohol. J Anal Tox 1984;8:95–96. 23. Larsen MJ, Thorsen A, Jensen SJ. Ethanol-induced formation of solid calcium phosphates. Calcif Tissue Int 1985;37: 189 –193.
ASTEROID HYALOSIS
75