a b s t r a c t

Background: Ultrasonography is often used in the evaluation of patients with ocular concerns; however, several pathologic conditions and even some age-related changes can have similar sonographic appearances. One approach that clinicians use is to assume that unilateral findings visible at normal gain are acute, whereas bilateral findings requiring high gain are chronic, especially in the elderly population. To date, no studies have systematically evaluated this assumption.

Objectives: The objectives are to determine the prevalence of monocular and binocular mobile vitreous opacities (MVOs) in the vitreous chamber in an asymptomatic population at normal and high gain levels and to determine its prevalence with higher age stratifications.

Methods: We conducted a cross-sectional survey using 2-dimensional ultrasonography with a high-frequency

transducer of 105 asymptomatic subjects aged 20-89 years and evaluated each subject’s eyes for the presence of MVOs at both normal and high gain levels in progressive age stratifications.

Results: Ultrasonographic scans were obtained on 105 subjects. At normal gain levels, MVO was present in only 1

subject (0.95%; 95% confidence interval, 0.0%-5.0%). At high gain levels, MVO was present in 28.6% (30/105) of subjects. Of the subjects with MVO at high gain, 60% (18/30) had unilateral MVO. Mobile vitreous opacity was found more frequently with advancing age, being present in 23 subjects older than 59 years, compared with 7 subjects 59 years and younger (51.1% vs 11.7%, P b .001).

Conclusions: Mobile vitreous opacity in the vitreous chamber visualized at high gain levels is relatively common and may not be pathologic, even if unilateral and occurring at a relatively young age.

(C) 2015

Introduction

Ocular trauma and vision changes are regular causes of emergency department (ED) visits, accounting for 2% of ED visits [1]. eye injuries account for 3.5% of all occupational injuries in the United States, and about 2000 US workers injure their eyes each day [1]. In many acute ocular conditions, however, the physical examination can be difficult–especially because specialized equipment and ophthalmologic expertise are often unavailable in the ED. Eye injuries also pose risk to the provider because misdiagnosis of ocular pathology can lead to loss of vision, significant morbidity, and significant legal risk [2].

In this setting, bedside ultrasonography has quickly become an ad-

junct to standard examination techniques, allowing clinicians the ability to evaluate for retinal detachment (RD), posterior vitreous detachment (PVD), vitreous hemorrhages (VHs), lens dislocations, and other eye

? Prior presentations: none.

?? Funding sources/disclosures: none.

? Acknowledgments: none.

* Corresponding author. Tel.: +1 413 794 2424; fax: +1 413 794 8070.

E-mail address: [email protected] (G. Budhram).

problems that have been traditionally difficult to diagnose using stan- dard ED methods and tools [3]. The majority of ED ocular ultrasono- graphic studies have so far have focused on RD, so test characteristics for PVD, VH, and other entities are unfortunately not well described. Re- garding RD, initial investigations by Blaivas et al [3] in 2002 were prom- ising, showing a 100% sensitivity and 100% specificity in 61 patients with ocular trauma or other visual concerns. In subsequent investiga- tions, however, test characteristics have been less robust. In 2011, Shi- nar et al [4] demonstrated a 92% specificity, and in 2010, Yoonessi et al [5] found only an 83% specificity of ultrasonography for RD. It is likely that these lower specificities are at least partly due to sonographic similarities between RD and other disease entities including PVD and VH. Certain types of VH, namely, acute mild VH, can also appear similar to chronic conditions including asteroid hyalosis. Although differentiat- ing ultrasound characteristics have been described in the literature for more than 20 years [6], novice sonographers or clinicians with an understanding of only the basic ultrasonographic applications might easily confuse one clinical entity with another. In some instances, misdiagnosis could significantly impact treatment strategies and patient outcomes.

http://dx.doi.org/10.1016/j.ajem.2015.09.017

0735-6757/(C) 2015

Fig. 1. Membranous MVO. A, Retinal detachment in an ED patient after a skiing injury at normal gain. B, posterior vitreous detachment at high gain in an ED patient with acute, atraumatic Vision loss. C, Subacute, organized VH at high gain in an ED patient after a fall 4 days prior and partial monocular Visual loss. D, Membranous MVO from an asymptomatic study subject at high gain (arrows indicate membranes).

Retinal detachment is typically identified on ultrasonography as a unilateral, brightly echogenic line that “floats” in the vitreous, moves with eye movement, and is visible at normal gain levels (Fig. 1A). It should also be tethered anteriorly to the ora serrata and posteriorly at the optic disk, creating a V or funnel shape. Sonographers typically use this fact to differentiate it from a PVD, which is not tethered posteriorly. Posterior vitreous detachment occurs when the gel-like vitreous substance contracts and pulls its posterior edge away from the retina and optic disk. Posterior vitreous detachment alone can be asymptom- atic and does not necessitate emergent ophthalmologic referral and treatment, but as the vitreous retracts from the retina, it may cause tears, hemorrhage, and detachment of the retina–a condition which does require ophthalmologic referral. Sonographically, both chronic and acutely symptomatic vitreous detachments appear as a thin, mobile, C-shaped concave upward membrane that is seen at high gain

levels (Fig. 1B). Unlike RD, it is not attached at the optic disk.

The sonographic appearance of VH depends on its age and severity. In fresh mild hemorrhages, small areas of particulate MVO are seen in the posterior chamber (Fig. 2A). As the hemorrhage ages, particularly with severe hemorrhages, the blood first organizes and then forms mobile membranes that are visible in the posterior chamber (Fig. 1C).

Asteroid hyalosis is typically an asymptomatic phenomenon, occur- ring in patients older than 60 years and requiring no specific treatment. The asteroid bodies represent calcium-laden lipids suspended within the vitreous. These may be progressive but never lead to severe vision loss [7]. Asteroid hyalosis is of concern to emergency physicians because it has a similar appearance on ultrasonography to acute mild VH: that of pinpoint “stars in the night sky” (Fig. 2B).

Although the sonographic descriptions above seem discrete, in prac- tice, there can be some ambiguity in the ultrasound findings leading to confusion for emergency physicians evaluating patients with acute inju- ry. Retinal detachment, subacute VH, acute PVD, and chronic PVD can all appear as mobile vitreous opacities (MVOs) that are membranous in na- ture and swirl with eye movement (Fig. 1). Acute mild VH maybe be confused with asteroid hyalosis because both appear as particulate MVOs that swirl with eye movement [8] (Fig. 2). Indeed, MVO in the

vitreous chamber is now thought to be a relatively common and some- times asymptomatic phenomenon, especially in the aging eye [9]. This fact has been recently highlighted by Schott et al [10] in a case report demonstrating PVD being confused for RD on ultrasonography in the ED.

To decide whether ultraSonographic findings are acute or chronic, clinical sonographers often use gain and lateralization: findings that are unilateral and visible at normal gains are considered acute, whereas findings that are bilateral and require high gain are more likely to be chronic, especially in the elderly population [9]. Although this approach seems intuitive, to date, no research has evaluated the accuracy of this assumption. The hypothesis of this study is that the sonographic finding of MVO in the vitreous chamber is a binocular phenomenon that is observed primarily at high gain levels and increases in prevalence with increasing age. We hope that results of this study will help clinicians differentiate acute, vision-threatening pathology from normal age-related changes.

Methods

Study design

This was a cross-sectional observational study that enrolled a conve- nience sample of patients aged 20-89 years presenting to an urban aca- demic ED between March and June 2013. Patients were stratified into decades of age with 15 patients included in each stratum. Patients were eligible to participate in the study if they presented to the ED with a nonocular concern, were between the ages of 20 and 89 years, were able to provide informed consent, and were able to speak and/or read English. They were excluded from participation if they were cur- rently experiencing ocular symptoms or if they had any condition which would make ultrasonography uncomfortable for them, a history of allergic reaction to the coupling gel or adhesive, any history of eye surgery, “flashers” or “floaters” in any visual fields, artificial eye, or his- tory of significant eye pathology or trauma, excluding myopia (near- sighted) or hyperopia (farsighted) that is correctable to normal visual

Fig. 2. Particulate MVO. A, Acute, mild VH at high gain in an ED patient with visual changes after motor vehicle collision. B, An ED patient with known asteroid hyalosis at high gain. C, Particulate MVO from an asymptomatic study subject at high gain.

acuity. If their visit to the ED involved a condition requiring immediate attention, they were also excluded. Electronic medical record review was also used to screen for known ocular diagnoses as well as to document chronic medical conditions. Written informed consent was obtained from each participant. The local Institutional Review Board approved this study.

Study setting and population

This study was performed at a large academic ED with approximately 130,000 annual patient visits.

Study protocol

The primary sonographer was an emergency ultrasound fellow who had previously completed 3 years of emergency medicine residency with general ultrasonography training and was in the process of ad- vanced training in ultrasonography. Her training in ocular ultrasonogra- phy included a 1-hour didactic session followed by 1 hour of hands-on training. This physician surveyed the electronic ED Tracking system for potential study participants. Patients aged 20-89 years presenting with- out an eye-related concern were approached and screened for eligibility in the study. Demographic information (age, sex, ethnicity) was obtain- ed from the medical record. Patient history and the medical record were also used to document any chronic medical conditions including diabe- tes, hypertension, hyperlipidemia, end-stage renal disease , cir- rhosis, congestive heart failure (CHF), Atrial fibrillation , peripheral vascular disease, and Deep vein thrombosis /pulmo- nary embolism (PE).

For the purposes of the study, mobile vitreous opacities were defined as bright, hyperechoic material visible only in the vitreous chamber of the eye (and not extending beyond its borders) and that swirled when the subject looked left and right. This material could be particulate or membranous in appearance.

The ocular ultrasound examination was performed by the emergen- cy ultrasound fellow according to standard, defined ocular Imaging techniques [9]. First, a thin, transparent film dressing (“Tegaderm”; 3M, St Paul, MN) was placed over the closed eyelid for protection. Ultra- sound gel was deposited on the film dressing, and a high-frequency (5- 10 MHz) linear transducer was lightly placed on the eye, with care to minimize pressure on the eye. A Sonosite M-Turbo ultrasound machine (FUJIFILM Sonosite, Bothell, WA) and L25x 13-6-Hz linear transducer was used to obtain all images. The machine was preset to the factory “ophthalmic” examination settings, and a standard imaging depth of

3.5 cm was used. Once a clear image was obtained using Sonosite’s pro- prietary “autogain” setting, the patients were asked to move their eyes left and right continuously as a 30-second video clip was recorded. The eye was imaged at an “autogain” setting for the first 5 seconds of the clip. During the next 25 seconds, the gain was slowly increased until MVO became visible or until the maximum gain setting was reached. This level was defined as high gain. After the examination, the transparent film and gel were removed from the first eye, and the procedure was then repeated on the other eye. Each patient had 2 videos randomly labeled with 2 distinct identification numbers corre- sponding to the right and left eyes, but no identifying patient informa- tion was recorded with the videos. All of the video clips were uploaded and stored on an off-site video archive system (“Ultralinq”; Ultralinq Healthcare Solutions, New York, NY) for later review. Supple- mental Video is an example of a 30-second video clip from a study sub- ject where MVO can be seen as the gain is gradually increased.

Two Attending emergency physicians, each with residency training in emergency medicine, Fellowship training in emergency ultrasonogra- phy, and greater than 5 years in clinical practice, later reviewed all video clips. The reviewers were blinded to age and were unaware which eye clips were paired in a single subject. Each reviewer independently re- ported whether or not MVO was visible on the initial autogain setting

and/or on a high gain setting. Any disagreements between the re- viewers were resolved by consensus of an adjudication committee com- posed of the 2 original reviewers and a third emergency physician with the same credentials.

Key outcome measures

The key outcome measure was the presence or absence of MVO at both autogain and high gain settings for each eye.

Statistical analysis

Descriptive statistics were used to describe the prevalence of MVO by age strata and gain levels. We calculated proportions and 95% confi- dence interval (CI) regarding the prevalence of MVO. Linear trends were examined using methods described by Vittinghoff et al [11]. We tested the hypothesis that MVO increased approximately linearly with age using the Cochran-Armitage test of trend [12]. Departures from linearity were examined using visual plots. Logistic regression analysis then was used to calculate the relative odds of MVO in older age stratum vs youngest and to derive marginal estimates. All tests of hypothesis were 2-sided, with a critical level of <= 5%.

We calculated ? for agreement between the original sonographer obtaining the images and the reviewing sonographers.

Statistical calculations were done in Stata/MP 13.1 for Windows (StataCorp, Union Station, TX).

Sample size

No previous estimates exist regarding the proportion of eyes with a finding of MVO. As we expected that the proportion of patients exhibiting MVO would increase with age, we estimated that the 7 age strata (20s to 80s) would have proportions of patients with MVO of ap- proximately 0, 0, 0.2, 0.4, 0.6, 0.6, and 0.6, respectively. A priori, we de- termined that a total sample of 105 would provide 81% power to detect a linear trend of similar magnitude using a 2-sided Z test with continuity correction and a significance level of .05.

Results

One hundred and five patients were initially enrolled in the study, but 3 were excluded after medical record review revealed a history of macular degeneration (2) or diabetic retinopathy (1). Three additional subjects were enrolled, for a total of 105 subjects consisting of 47 men and 58 women, with 15 subjects in each age stratum. Baseline charac- teristics of our study population are described in Table 1.

A total of 210 eyes were scanned. A summary of the eyes and sub- jects with MVO is presented in Table 2 and Fig. 3. Mobile vitreous opac- ity was present in 30 subjects (28.6%) at high gain but in only 1 subject at both the autogain and high gain level, a male aged 77 years (0.95%; 95% CI, 0%-5%), and was unilateral in that subject. This supports the hy- pothesis that MVO is preferentially visible at high gain level.

Thirty patients demonstrated MVO at high gain (28.6%; 95% CI, 20.2%-38.2%), with 18 of 105 patients (17.1%; 95% CI, 10.5%-25.7%) hav-

ing unilateral MVO and 12 (11.4%; 95% CI, 6%-19.1%) having bilateral MVO, refuting the hypothesis that MVO is usually bilateral.

The average age of all patients with MVO (either unilateral or bilat- eral) was 69.5 years, whereas the average age of patients without MVO was 49.2 years. Regarding age distribution, MVO was found more fre- quently with advancing years (Cochran-Armitage test of trend is signif- icant ?²[1df] = 15.0, P b .001). A nonlinear, threshold effect is also apparent, as MVO’s prevalence increases precipitously among patients older than 60 years. It was present in 23 of 45 of subjects older than 59 years compared with only 7 subjects 59 years and younger (51.1% vs 11.7%, P b .001). Compared with patients in the youngest age strata, patients 60 years and older are 9.4 times as likely to have MVO, and

Table 1

Baseline characteristics

Characteristic n (%)

Age (categorical), y

20-29	15 (14.3%)
30-39	15 (14.3%)
40-49	15 (14.3%)
50-59	15 (14.3%)
60-69	15 (14.3%)
70-79	15 (14.3%)
80-89	15 (14.3%)
Sex
Male	44 (41.9%)
Female	61 (58.1%)
Race
White	77 (73.3%)
African-American	13 (12.4%)
Hispanic, white	13 (12.4%)
Hispanic, black	1 (1%)
Other	1 (1%)
Medical history Diabetes	21 (20.0%)
Hyperlipidemia	39 (37.1%)
Hypertension	49 (46.7%)
ESRD	2 (1.9%)
Cirrhosis	3 (2.9%)
CHF	12 (11.4%)
AF	13 (12.4%)
Peripheral Vascular Disease	4 (3.8%)
DVT/PE	5 (4.8%)

they are 3.8 times as likely as those aged 40 to b 60 years to have MVO. This supports the hypothesis that MVO is increasingly prevalent with advancing age. Notably, however, MVO was also found in 20.0% (3/15) of subjects in their 30s and was present in each age stratum thereafter. Table 3 demonstrates the incidence of chronic medical conditions per age strata and divided by presence/absence of MVO. Mobile vitreous opacity was not preferentially identified in patients with diabetes, hyperlipidemia, hypertension, ESRD, cirrhosis, CHF, peripheral vascular

disease, or DVT/PE.

Regarding interrater reliability, the sonographer and the overreading physicians agreed on all assessments of MVO except 1, the case with MVO at normal gain: in this case, 1 physician disagreed. Per protocol, 2 physicians overread the sonographer’s images, and in the case that they disagreed, a third overreading physician settled the dispute, in this case agreeing that MVO was present. ? was therefore 1.0.

Discussion

Although there is an abundance of case reports describing ocular ul- trasound in patients with vision loss or ocular trauma, surprisingly little primary literature exists in either emergency medicine or ophthalmolo- gy that systematically evaluates Ocular ultrasonography. Several studies describe sensitivities and specificities for RD and PVD, but ultrasono- graphic findings in the normal eye are not well described. For clinicians evaluating patients with traumatic or atraumatic vision loss at the bed- side, this can present some difficulty if the sonographer is not keenly aware of the defining pathologic characteristics of RD, PVD, and VH.

This study demonstrates a higher than expected prevalence of MVOs

in the vitreous chamber of asymptomatic patients, occurring in 28.6% of

our study population when viewed at high gain. Despite the assumption that MVO would be rare in younger patients, MVO was a relatively com- mon finding starting in the fourth decade. Regarding laterality, MVO was found to be unilateral as often as it was found to be bilateral, refut- ing our hypothesis that MVO is a binocular phenomenon and suggesting that laterality may not be useful in differentiating an acute from a chron- ic finding.

Reassuringly, MVO was absent in 99.0% of patients at normal gain

levels. This is helpful when visualizing RD, which should be present at normal gain levels. It is less helpful, however, when attempting to iden- tify other acute pathology that normally requires high gain to be seen. Acute PVD and acute VH both cause vision loss and usually require high gain, but chronic PVD and asteroid hyalosis are asymptomatic and are also only seen at high gain.

Unfortunately, MVO was not preferentially found in patients with di- abetes, hyperlipidemia, hypertension, ESRD, cirrhosis, CHF, atrial fibril- lation, peripheral vascular disease, or DVT/PE.

Regarding morphology, the authors observed that MVO could be characterized in the study population as either membranous or particu- late. Although an evaluation of these distinctions in our population was not part of this study, it is worthwhile to note that the morphology of both types of MVO could mimic actual ocular pathology. Particulate MVO (Fig. 2C) could represent asteroid hyalosis in asymptomatic pa- tients or acute minor VH in patients with vision loss or eye trauma. Sim- ilarly, the membranous type of MVO (Fig. 1D) may be confused in the acute setting with RD acute PVD, or subacute organized VH. Clinicians should be aware of the nuances used to differentiate these entities if using bedside ultrasound in patients with acute vision loss or ocular trauma.

It is also worth noting that MVO in this study does not represent an imaging artifact. Imaging artifacts are the appearance of structures which are not actually present and are created by the sound beam re- flection, refraction, enhancement, or other effects. Artifacts, however, typically extend outside the organ of interest or disappear when the transducer beam angle is changed. To exclude imaging artifact, a defin- ing characteristic of the MVO in this study was movement and swirling with eye movement and location not extending outside the globe.

Limitations

Although all study participants denied a history of ocular injury, eye problems, or vision changes, it is possible that they may have had undi- agnosed ocular sequelae of diabetes, coagulopathy, or other chronic ill- ness that were identified as MVO. The particulate type of MVO could represent undiagnosed asteroid hyalosis, whereas the membranous MVO could have been chronic asymptomatic PVD or fibrinous vitreous bands which occur with age. The authors do not think that this adverse- ly affected the study because a clinician would have to differentiate these findings from acute injury in any previously asymptomatic patient presenting with trauma or vision loss. Reassuringly, also, MVO was not preferentially found in patients with diabetes or any of the other screened medical conditions.

Our study attempted to exclude patients with ocular pathology based on history, medical record review, and presenting concern. A comprehensive eye examination was not performed because we felt that many occult ocular pathologies could not be accurately diagnosed by tools available in the ED.

Table 2

Subjects with MVO at high gain by age strata

	20-29 n = 15	30-39 (n = 15)	40-49 (n = 15)	50-59 (n = 15)	60-69 (n = 15)	70-79 (n = 15)	80-89 (n = 15)	All subjects (N = 105)
Eyes with MVO	0(0%)	4(13.3%)	4(13.3%)	3(10%)	9(30%)	13(43.3%)	9(30%)	42(20%)
Subjects with MVO	0(0%)	3(20%)	2(13.3%)	2(13.3%)	8(53.3%)	9(60%)	6(40%)	30(28.6%)
Subjects with unilateral MVO	0(0%)	2(13.3%)	0(0%)	1(6.7%)	7(46.7%)	5(33.3%)	3(20%)	18(17.1%)

Fig. 3. Subjects with MVO at high gain by age strata.

Our study used the “autogain” function available on the Sonosite M-turbo ultrasonography machine. Although this function is similar to the “auto-optimize” on other machines, the settings will not be standardized across platforms. Decreasing the overall gain on any ultra- sonography machine until the vitreous substance in the posterior chamber appears anechoic will usually approximate the “autogain” or “auto-optimize” function.

A limiting factor to the external generalizability of our study is that a single sonographer performed all the ultrasonographic exam- inations, introducing the possibility that operator skill played a role in the detection of MVO. The recorded video clips, however, were overread by staff physicians with near-perfect agreement for the presence of MVO.

To do a proper ocular ultrasonographic examination and detect MVO, subjects were asked to continuously direct their gaze left and right while imaging. Some subjects redirected their gaze only minimal- ly, possibly limiting our ability to detect MVO.

Our study evaluated only patients without acute eye concerns; we did not evaluate the test characteristics of ultrasonography for diagnosing ocular pathology, which has been studied before [3-5,13,14].

Conclusions

The results of our study indicate that MVO in the vitreous chamber is uncommon at normal gain levels but relatively common at high gain settings, and found both unilaterally and bilaterally. Although it does occur more frequently with advancing age, it was present in our popu- lation in all age groups older than 30 years.

Supplementary data to this article can be found online at http://dx. doi.org/10.1016/j.ajem.2015.09.017.

References

1. Xiang H, Stallones L, Chen G, Smith GA. Work-related eye injuries treated in hospital emergency departments in the US. Am J Ind Med 2005;48(1):57-62.
2. Tintinalli JE, Stapczynski J, Ma OJ, Cline DRC, Meckler G, editors. Tintinalli’s emergen- cy medicine: a comprehensive study guide. New York: McGraw-Hill; 2011.
3. Blaivas M, Theodoro D, Sierzenski PR. A study of bedside ocular ultrasonography in the emergency department. Acad Emerg Med 2002;9(8):791-9.
4. Shinar Z, Chan L, Orlinsky M. Use of ocular ultrasound for the evaluation of retinal detachment. J Emerg Med 2011;40(1):53-7.
5. Yoonessi R, Hussain A, Jang TB. bedside ocular ultrasound for the detection of retinal detachment in the emergency department. Acad Emerg Med 2010;17(9): 913-7.

   Table 3

   Medical history in patients with and without MVO

   	20-29 (n = 15)	30-39 (n = 15)	40-49 (n = 15)	50-59 (n = 15)	60-69 (n = 15)	70-79 (n = 15)	80-89 (n = 15)	All subjects (N = 105)
   Diabetes
   MVO	0	1	0	0	3	3	2	9
   No MVO	0	0	3	3	2	3	1	12
   Hyperlipidemia MVO	0	0	2	1	6	7	3	19
   No MVO	0	1	2	4	5	3	5	20
   Hypertension
   MVO	0	0	1	2	7	7	4	21
   No MVO	1	1	5	6	4	4	7	28
   ESRD MVO	0	0	1	0	0	0	0	1
   No MVO	0	0	0	0	1	0	0	1
   Cirrhosis
   MVO	0	0	1	1	0	0	0	2
   No MVO	0	0	0	1	0	0	0	1
   CHF MVO	0	0	0	0	4	2	1	7
   No MVO	0	0	0	0	2	2	1	5
   MVO	0	0	0	1	3	1	1	6
   No MVO	0	0	0	1	2	1	3	7
   PVD MVO	0	0	0	0	1	0	1	2
   No MVO	0	0	0	0	0	0	2	2
   DVT/PE
   MVO	0	0	0	0	0	0	1	1
   No MVO	0	0	2	2	0	0	0	4

   AF

   McNicholas MM, Brophy DP, Power WJ, Griffin JF. Ocular sonography. AM J Roentgenol

   1994;163(4):921-6.

   Moss SE, Klein R, Klein BE. Asteroid hyalosis in a population: the Beaver Dam eye study. Am J Ophthalmol 2001;132(1):70-5.

6. Allan PLP, Baxter Grant M, Weston Michael, editors. Clinical ultrasound. 3rd ed. Edinburgh: Churchill Livingstone; 2011.
7. Lyon M, Jehle D. Ocular ultrasound. In: Ma OJ, Mateer J, Reardon R, Joing S, editors. Emergency ultrasound. 3rd ed. New York: McGraw Hill; 2014.
8. Schott ML, Pierog JE, Williams SR. Pitfalls in the use of ocular ultrasound for evaluation of acute vision loss. J Emerg Med 2013;44(6):1136-9.
9. Vittinghoff E, Glidden DV, Shiboski SC, McCulloch CE. Regression methods in biostatistics linear, logistic, survival, and repeated measures models. New York: Springer Science & Business Media Inc.; 2005.
10. Gart JJ, Tarone RE. The relation between score tests and approximate UMPU tests in exponential models common in biometry. Biometrics 1983;39(3):781-6.
11. Blaivas M. Bedside emergency department ultrasonography in the evaluation of ocular pathology. Acad Emerg Med 2000;7(8):947-50.
12. Vrablik ME, Snead GR, Minnigan HJ, Kirschner JM, Emmett TW, Seupaul RA. The diagnostic accuracy of bedside ocular ultrasonography for the diagnosis of retinal detach- ment: a systematic review and meta-analysis. Ann Emerg Med 2014;65(2):199-203.