Prevalence of asteroid hyalosis and systemic risk factors in United States adults

Abstract

Background/Objective

Asteroid hyalosis (AH) is an uncommon clinical entity of unknown aetiology that is associated with older age. Previous epidemiologic studies have reported various systemic and demographic risk factors for AH but remain limited due to regional constraints of their study populations. Additionally, Hispanic and Non-Hispanic black populations remain under sampled. The aim of this study is to examine the prevalence of asteroid hyalosis in the United States and identify associated factors at a national level.

Subjects/Methods

This is a population-based, cross-sectional study of 5578 subjects aged 40 and older from the 2005 to 2008 National Health and Nutrition Examination Survey (NHANES). The primary outcome measured was asteroid hyalosis on retinal imaging in any eye. Evaluated risk factors included patient demographics, medical history, body measures, serum markers, and fundus photography findings.

Results

Prevalence of asteroid hyalosis was 0.86% overall, 0.86% in Caucasians, 0.79% in African-Americans, and 0.88% in Hispanics. Asteroid hyalosis was associated with older age (p < 0.0001, 95% confidence interval [CI], 0.06–0.12; odds ratio [OR], 1.09) but not ethnicity or sex. After adjusting for age, greater bodyweight (p = 0.049; 95% CI, 0.001–0.04; OR, 1.02), and history of myocardial infarction (p = 0.022; 95% CI, 0.07–1.55; OR, 2.36) were also found to be significant risk factors.

Conclusion

Asteroid hyalosis is a rare entity in the US associated with older age, greater body weight, and prior history of MI. A potential relationship between AH and cardiovascular disease remains plausible.

Introduction

Asteroid hyalosis (AH) is an idiopathic clinical entity associated with multiple yellow-white calcium and lipid-containing spherical particles, or asteroid bodies, within the vitreous [1]. AH is often unilateral and may limit the view of the fundus. However, even in severe fundus-obscuring cases, vision is rarely affected. Thus, AH is generally considered benign and not requiring of intervention, though treatment with vitrectomy is rarely performed if complications develop [1, 2].

Previous epidemiologic studies have reported prevalence of AH ranging from 0.3 to 2% of the population [1]. While the pathophysiology of AH remains largely unknown, the relationship between AH and older age is well described in the literature [3–6], and recently it has been suggested that prevalence of AH may be expected to increase with the aging population [7]. Given that there are currently over 46 million United States citizens over the age of 65 and that number is expected to nearly double by 2060 [8], a greater understanding of AH pathology and the disease process is necessary to adapt with this changing population.

While previous epidemiologic analyses have attempted to study the pathophysiology of AH by identifying associated risk factors for the development of AH, little has been done to describe this clinical entity at a national level in the United States (US). Besides old age, other previously suggested associations include male sex [4–6], hypertension [3, 9], stroke [3], serum triglycerides [3], serum low-density lipoprotein (LDL) [3], hypercalcemia [10], atherosclerotic vessel disease [9], diabetes [9], high body mass [5], and alcohol consumption [5]. Collectively, these may imply some association between AH and metabolic or cardiovascular disease. However, results from these previous cross-sectional studies are variable due to differing sample sizes and specific study populations, and the majority of these studies evaluated patients within only a specific geographic region in the US or internationally. Furthermore, prior case-control and population studies have primarily observed Asian and White male populations while non-Hispanic black and Hispanic populations remain under sampled with respect to the United States population.

The purpose of this study is to describe the prevalence and associated risk factors for AH in United States adults using the National Health and Nutrition Examination Survey (NHANES) [11]. Here we add to the clinical literature of case-control and population-based studies of AH by describing the prevalence of this condition at a national level. We describe risk factors and associations of AH in a more diverse and demographically-representative United States (USA) population.

Methods

The present population-based study is a cross-sectional study of 5578 adults aged 40 and older from NHANES data collected from the years 2005 to 2008. This research adhered to the tenets of the Declaration of Helsinki and was conducted in accordance with regulations set forth by the Health Insurance Portability and Accountability Act. Institutional review board approval was not pursued as this study utilized publicly available data.

NHANES is a program of the National Center for Health Statistics (NCHS), and its purpose is to assess the health and nutritional status of adults and children in the United States through interviews and physical examinations [11]. Health interviews are conducted in respondents’ homes, and examinations are performed by teams of physicians, medical and heath technicians, and dietary experts in specially designed and equipped mobile centres. The sample for the survey aims to represent the U.S. population of all ages, and data is collected on demographic, socioeconomic, dietary, and health-related information. Our data comes from the Retinal Imaging subsection of the Ophthalmology Component, which assessed for the presence of multiple ocular pathologies using two forty-five degree non-mydriatic digital retinal images of each eye using an ophthalmic digital imaging system [11]. Inclusion criteria included age equal to or over 40 as fundus imaging was only available in this subset. Exclusion criteria included lack of available fundus imaging.

Asteroid hyalosis was assessed and diagnosed via fundus images for each eye by observing for the presence of soap-like crystals floating in the vitreous cavity in front of the retina, which can obstruct the view of retinal lesions. The degree of asteroid hyalosis presence was classified as 0 = No, 1 = Questionable, 2 = Yes, 3 = Yes in the centre, or 4 = no data. The responses of 0 and 1 were then combined to define the absence of asteroid hyalosis and responses of 2 and 3 were combined to define the presence of it according to the NHANES suggested groupings [11]. Observed information included patient demographics, body measures, other fundus photography findings, serum markers, and past medical history data collected via questionnaire.

Using this data, statistical analysis was performed with R (Version 3.5.1). Serial univariate and multivariate logistic regression was carried out to identify associations between specific factors and the presence of asteroid hyalosis. Odds ratios were calculated for each observed variable with respect to those with fundus imaging consistent with and without asteroid hyalosis and adjusted for age. A p-value of  ≤ 0.05 was considered statistically significant.

Results

The study population included 5578 adults (49.9% males and 50.1% females) aged 40 and older. The mean age of all subjects was 59.53 ± 12.45 years. Of all subjects, 22.39% were Hispanic, 20.33% were Non-Hispanic Black, 54.03% were Non-Hispanic White, and 3.24% were listed as Other (Table 1).

Table 1.

Characterization of study population.

		Without Asteroid Hyalosis	With Asteroid Hyalosis	Total
Subjects*		5530 (99.1)	48 (0.860)	5578 (100)
Mean Age (Years)		59.43 ± 12.42	71.13 ± 10.65	59.53 ± 12.45
Sex**	Men	2759 (99)	27 (0.97)	2786 (50)
	Women	2771 (99)	21 (0.75)	2792 (50)
Ethnicity**	Hispanic	1238 (99)	11 (0.88)	1249 (22)
	Non-Hispanic White	2988 (99)	26 (0.86)	3014 (54)
	Non-Hispanic Black	1125 (99)	9 (0.79)	1134 (20)
	Other	179 (99)	2 (1.10)	181 (3.26)

*n (%) of total population, **n (%) of subgroup.

The overall prevalence of asteroid hyalosis in our study population was 0.86% (48 subjects). A total of 56 eyes had asteroid hyalosis with 8 eyes in 4 patients representing bilateral asteroid hyalosis (14.3%). Of the patients with unilateral presentation, 28 subjects were affected in the right eye and 20 were affected in the left eye. The average age of subjects with asteroid hyalosis was 71.13 ± 10.65years compared to 59.43 ± 12.42 years in subjects without AH. Gender specific prevalence was 0.97% in males and 0.75% in females. Ethnicity specific prevalence was 0.88% in Hispanics, 0.79% in African-Americans, and 0.86% in Caucasians (Table 1).

A multivariate logistic regression was performed on several risk factors to determine statistically significant risk factors. Asteroid hyalosis was associated with older age (p < 0.0001, 95% confidence interval [CI], 1.06–1.12; odds ratio [OR], 1.09). There were no significant associations with ethnicity or sex.

After adjusting for age, asteroid hyalosis was associated with history of myocardial infarction (MI) (p = 0.02; OR, 2.36) and bodyweight in kilograms (p < 0.05; OR, 1.02) (Table 2). Waist circumference (p = 0.07; OR, 1.02), serum uric acid in mg/dL (p = 0.07; OR, 1.19), personal history of coronary artery disease (p = 0.08; OR, 2.04), and personal history of high cholesterol (p = 0.09; OR, 1.79) approached but did not reach statistical significance (Tables 2,  3). There was no statistically significant association between asteroid hyalosis and diabetes (p = 0.43), hypertension defined as > 140/90 (p = 0.83), BMI (p = 0.15), calcium levels (p = 0.18), or prior stroke (p = 0.20) (Tables 2, 3). Additionally, other fundus findings identified in our image analysis such as diabetic retinopathy, central diabetic macular oedema, arteriovenous nicking, or vertical cup to disc ratio were not associated with AH (p > 0.005) (Table 3).

Table 2.

Age-adjusted demographic and examination Associations with Asteroid Hyalosis.

		Odds Ratio	95% Confidence Interval	P-value
Demographics	Age	1.09	1.06–1.12	<0.0001
	Female Gender	0.79	0.45–1.41	0.43
Ethnicity	Hispanic	0.33	0.05–2.56	0.23
	Non-Hispanic White	1.39	0.47–5.90	0.6
	Non-Hispanic Black	1.99	0.59–9.04	0.3
	Other	3	0.39–18.41	0.23
Body Measures	Systolic blood pressure	0.99	0.97–1.00	0.18
	Hypertension (Blood pressure > 140/90 mmHg)	0.9	0.36–2.73	0.83
	Weight (kg)	1.02	1.00–1.04	<0.05
	BMI (kg/m2)	1.05	0.98–1.11	0.15
	Obesity (BMI > 30 kg/m2)	1.33	0.62–2.67	0.43
	Waist Circumference (cm)	1.02	1.00–1.04	0.07
Fundus Photography	Mild Nonproliferative Diabetic Retinopathy (NPDR)	1.18	0.35–2.98	0.76
	Moderate/Severe NPDR	3.13	0.50–10.71	0.12
	Central Diabetic Macular Oedema	3.62	0.20–17.82	0.21
	Early Age Related Macular Degeneration	1.62	0.53–4.03	0.34
	Arteriovenous Nicking	0.89	0.26–2.26	0.83
	Vertical Cup to Disc Ratio	4.84	0.45–48.01	0.18

Statistically significant p-values are in bold.

Table 3.

Age-adjusted Serum Markers and Questionnaire Associations with Asteroid Hyalosis.

		Odds Ratio	Confidence Interval	P-value
Serum Markers	High Density Lipoprotein (HDL) (mg/dL)	0.98	0.96–1.00	0.15
	Triglycerides (mg/dL)	1	0.99–1.00	0.93
	Low Density Lipoprotein (LDL) mg/dL	1.01	1.00–1.02	0.1
	Total Cholesterol mg/dL	1	1.00–1.01	0.53
	C-Reactive Protein (CRP) mg/dL	1.08	0.79–1.27	0.51
	Haemoglobin A1c	1.11	0.83–1.40	0.41
	Calcium (mg/dL)	0.59	0.27–1.27	0.18
	Phosphorus mg/dL	1.27	0.73–2.13	0.38
	Uric Acid (mg/dL)	1.19	0.98–1.44	0.07
Questionnaire	Ever had 5 or more drinks every day	0.79	0.36–1.91	0.57
	Smoked at least 100 cigarettes over entire life	0.85	0.46–1.56	0.61
	High blood pressure	0.91	0.50–1.65	0.76
	High Cholesterol	1.79	0.94–3.61	0.09
	Diabetes	1.32	0.63–2.57	0.43
	Arthritis	1.14	0.63–2.11	0.67
	Congestive Heart Failure	0.67	0.28–1.96	0.4
	Coronary Artery Disease	2.04	0.86–4.33	0.08
	Heart Attack	2.36	1.07–4.74	0.02
	Stroke	0.56	0.25–1.51	0.2
	Thyroid Problem	1.05	0.48–2.65	0.91
	Cancer	1.25	0.60–2.95	0.58
	Osteoporosis	1.12	0.46–3.35	0.83

Statistically significant p-values are in bold.

Discussion

AH is an idiopathic condition readily apparent on clinical exams as multiple spherical white bodies within the vitreous and associated with aging [1]. While prior epidemiologic studies have suggested a handful of associations between AH and systemic conditions, there is limited generalizability with respect to the entire US population as previous study populations have consisted primarily of non-Hispanic white or Asian populations within specific geographic regions [1, 3–7, 9].

This study is unique in that it is the first epidemiological study of asteroid hyalosis at a national level in a more ethnically diverse study population with greater Hispanic (22.5%) and Non-Hispanic Black (20.4%) representation (Table 1). We report an overall national prevalence of AH of 0.86% as well as ethnicity-specific prevalence rates for Hispanics and Non-Hispanic Blacks of 0.88% and 0.79% respectively. To our knowledge, this study represents the most highly powered analysis of AH and Hispanic and non-Hispanic black ethnicity to date and finds no association of these ethnicities with AH development in the US.

The existing epidemiologic literature evaluating medical risk factors for AH has offered conflicting evidence. Bergren et al. first described an increased prevalence of systemic hypertension, atherosclerosis, and hyperopia in 101 eyes with asteroid hyalosis [9]. Conversely, a study of AH in an autopsy population found only a male gender and posterior vitreous detachment to be associated with AH when adjusted for age [6]. The Blue Mountain Eye Study in Australia then confirmed the association between AH and male gender but found no association with hypertension, heart disease or diabetes [4]. Meanwhile, a study from Seoul, Korea conducted 5 years after the Blue Mountain Study reported a significant association with hypertension and prior stroke [3]. While collectively the existing evidence appears to suggest some relationship between AH and cardiovascular or metabolic disease, prior epidemiologic studies of AH have been variable in design and study population.

The association between AH and greater body weight identified in this study is in line with the findings of one of the largest epidemiologic studies evaluating AH in the United States to date, the Beaver Dam eye study, which examined a population of 4926 patients from Beaver Dam, Wisconsin [5]. As participants in that study were primarily white Northern Europeans (99.4%), our similar finding in a more diverse population adds validity and generalizability to the association of greater body weight with AH. On the other hand, our results indicating a relationship between AH and a prior history of MI is less supported by the literature and has never been directly evaluated.

Collectively, our findings remain consistent with a potential relationship between AH and cardiovascular disease, as both greater bodyweight and history of MI are known to be independently associated with heart disease in the US [12, 13]. This is further supported by a trend towards higher prevalence of coronary artery disease and dyslipidaemia in patients with AH. Admittedly, however, these variables were not found to be statistically significant, nor were other known risk factors for cardiovascular disease such as hypertension. One potential explanation for this discrepancy is the wide clinical spectrum encompassing cardiovascular disease. Patients with prior MI history may represent those with more advanced cardiovascular disease whereas a diagnosis of coronary artery disease could encompass patients with earlier and less contributory disease (Table 3). At the very least, our analysis of AH risk factors using national data cannot exclude the possibility of a relationship between asteroid hyalosis and cardiovascular disease.

Mechanistically, dystrophic calcification may represent a shared pathophysiologic mechanism between AH and cardiovascular disease. It has been suggested that the development of AH involves changes in calcium and lipid levels within the eye and consequent deposition within the vitreous [2, 9]. This is supported by histopathologic analyses of AH vitreous bodies confirming the presence of primarily calcium-associated phospholipids [14]. In comparison, deposition of calcium and lipids within the walls of blood vessels is a hallmark of atherosclerosis, and the relationships between calcification with age and cardiovascular disease are also well documented [15, 16]. Another proposed mechanism of AH relates to increased permeability of basal membranes in diabetic eyes leading to calcium and phospholipid extravasation and formation of asteroid bodies within the vitreous [1, 2, 17]. However, this is not supported in our cross-sectional study of the US population as neither a reported history of diabetes nor evidence of diabetic retinopathy on fundus imaging were associated with AH. As previous studies reporting an association between diabetes and AH have evaluated only patients presenting for ophthalmic evaluation, we hypothesize there may be a selection bias for diseases more prevalent among patients presenting for ophthalmic evaluation [1]. NHANES’ method of selecting a randomized, diverse, and nationally representative study population may ameliorate this type of referral bias.

Further research evaluating the relationship between AH and cardiovascular disease is warranted. Additionally, further research may evaluate differences in patients with bilateral versus unilateral AH given the suggested systemic associations with cardiovascular disease. The major limitations of this study are its cross-sectional design and the potential for recall bias on answers to questionnaire-based data. Our study population is also limited as only subjects aged 40 and older underwent retinal imaging per the NHANES protocol. In addition, the ability to fully evaluate the fundus and peripheral vitreous is limited with imaging alone, potentially leading to an underestimation of AH prevalence in this study. No imaging data is provided if a sufficiently high-quality image is unable to be obtained for a given patient, and our evaluation of the association between retinal disease and AH is also hindered when view of the fundus is obscured in severe central AH. Additionally, milder or peripheral cases of AH may be relatively underestimated as we considered all patients with Fundus photography grade 1 (questionable findings) as negative cases in this study and photography is unable to capture the most peripheral vitreous.

Conclusion

In conclusion, asteroid hyalosis is an idiopathic benign clinical entity associated with older age. This study supports a new association between AH and history of MI in a large US population and validates a previously reported association between AH and greater body weight [5]. Future investigations may be indicated to explore possible shared pathophysiologic mechanisms between the development of AH and cardiovascular disease.

Summary

What was known before

• Asteroid hyalosis is a rare clinical entity of unknown aetiology with an estimated prevalence 0.3–2% according to cross-sectional studies surveying mostly Caucasian and Asian study participants within specific regions of the United States (US).

• Various associations with systemic conditions have been suggested in the literature including hypertension, stroke, hypertriglyceridemia, hypercalcemia, diabetes, and coronary artery disease.

• While associations between AH and various cardiovascular risk factors are frequently mentioned, the existing evidence has been variable and inconsistent among studies.

What this study adds

• Prevalence of AH in the US 0.86% when evaluating a more diverse and representative study population of the US.

• Specific Prevalence for Hispanics and Non-Hispanic Black patients were 0.88% and 0.79%, respectively. No significant association exists between AH and ethnicity or gender.

• Greater body weight and prior history of MI are associated with AH development after adjusting for age at the national level. A potential relationship between AH and cardiovascular disease remains plausible.

Author contributions

Contributors RD devised the study, collected data, contributed to the analysis, discussion, main interpretation of results and writing of the manuscript. XC collected data, contributed to the analysis, discussion, main interpretation of results and writing of the manuscript. NA collected data, contributed to the analysis, discussion, main interpretation of results, and writing of the manuscript. ES devised the study, collected data, contributed to the analysis, discussion, main interpretation of results, and writing of the manuscript. RS devised the study, collected data, contributed to the analysis, discussion, main interpretation of results, and writing of the manuscript.

Data availability

The dataset analysed during the current study are available in the NHANES repository [https://www.cdc.gov/nchs/nhanes/index.htm].

Competing interests

The authors declare no competing interests.