The Journal of Emergency Medicine, Vol. 40, No. 1, pp.

53–57, 2011
Copyright © 2011 Elsevier Inc.
Printed in the USA. All rights reserved
0736-4679/$–see front matter

doi:10.1016/j.jemermed.2009.06.001

Ultrasound in
Emergency Medicine

USE OF OCULAR ULTRASOUND FOR THE EVALUATION

OF RETINAL DETACHMENT
Zachary Shinar, MD,* Linda Chan, PHD,† and Michael Orlinsky, MD*

*Department of Emergency Medicine, LAC⫹USC Medical Center, University of Southern California Keck School of Medicine and
†Office of Biostatistics and Outcome Assessment, University of Southern California Keck School of Medicine, Los Angeles, CA
Reprint Address: Michael Orlinsky, MD, Department of Emergency Medicine, LAC⫹USC Medical Center, 1200 North State Street,
Los Angeles, CA 90033

e Abstract—Background: Retinal detachment is an ocu- e Keywords— ocular; ultrasound; evaluation; retinal;

lar emergency posing diagnostic difficulty for the emer- detachment
gency practitioner. Direct fundoscopy and visual field
testing are difficult to perform and do not completely
rule out retinal detachment. Ophthalmologists use ocular INTRODUCTION
ultrasound to enhance their clinical acumen in detecting
retinal detachments (RD), and bedside ultrasound capa- The use of ultrasonography by emergency physicians has
bility is readily available to many emergency practitio- greatly expanded over the last 10 years, and one of its
ners (EP). Study Objective: Our study sought to assess potential applications in the emergency department (ED)
whether ocular ultrasound would be a helpful adjunct is in the diagnosis of intraocular disease. Evaluation for
for the diagnosis of RD for the practicing EP. Methods: retinal detachment, vitreous hemorrhage, and vitreous
This was a prospective observational study with a con- detachment are all possible using a standard 10-MHz
venience sample of patients. As part of a general course linear probe. Other less common diseases also seen with
on emergency ultrasonography, practitioners received a ocular ultrasound include ocular tumors, intraocular for-
30-min training session on ocular ultrasound before be- eign bodies, globe rupture, and retrobulbar hemorrhage.
ginning the study. Trained practitioners submitted ultra- Recently, there has been some literature to suggest that
sound scans with interpretation on patients with signs intracranial pressure can also be evaluated by measuring
and symptoms consistent with retinal detachment. Re-
the diameter of the optic nerve sheath (1).
sults: Thirty-one of the 72 practitioners trained submit-
Although many of these intraocular diseases are emer-
ted ocular ultrasound reports on patients presenting to
gent conditions, most of them have significant physical
the Emergency Department with concerns for retinal
detachments. EPs achieved a 97% sensitivity (95% con-
findings that can be diagnostic without the need for
fidence interval [CI] 82–100%) and 92% specificity (95% ocular ultrasound. Retinal detachment (RD), on the other
CI 82–97%) on 92 examinations (29 retinal detach- hand, has less obvious physical findings, yet if missed,
ments). Disc edema and vitreous hemorrhage accounted often results in devastating sequelae. The most emergent
for false positives, and a subacute retinal detachment RD is when the retina is detached, but the macula is not.
accounted for the only false negative. Conclusion: These These “mac-on” RDs have more subtle physical findings
data show that trained emergency practitioners can use than “mac-off” RDs, because the visual center is still
ocular ultrasound as an adjunct to their clinical assess- functional. Both the risk of a missed diagnosis and the
ment for retinal detachment. © 2011 Elsevier Inc. benefit of emergent intervention are significant.

RECEIVED: 7 October 2008; FINAL SUBMISSION RECEIVED: 9 May 2009;

ACCEPTED: 1 June 2009
53
54 Z. Shinar et al.

Current practice using direct fundoscopic examination

and visual field testing, if used solely, allows for only a
limited view of the retina and would miss a large number
of retinal detachments (2). A method to more accurately
evaluate for retinal detachment would significantly im-
pact emergency practice by making an earlier diagnosis,
leading to a more rapid Ophthalmology consult, and
definitive care for the patient.
Because ophthalmologists have been using ultrasound
for over four decades to assess a number of different
disease processes including retinal detachment, it is rea-
sonable that emergency physicians could be taught to do
this in the ED (3). Ophthalmologic literature documents
that retinal detachments can be accurately characterized
using ultrasonography, and one study found that ultra-
sound correctly identified the size of the detachment
within a small sector of the eye (3 clock hours) in 33 of
35 (94.2%) patients (4,5).
Blaivas et al. published a study assessing the accu-
racy of ocular ultrasonography in the ED. This study
showed that 60 of 61 intraocular diseases were accu-
rately detected by emergency sonographers trained in
ocular ultrasound. Nine retinal detachments were di-
agnosed in that study (6).
Our study sought to evaluate whether emergency
practitioners minimally trained in the use of intraocular
ultrasound could accurately evaluate retinal detachments
during their normal daily practice.

MATERIALS AND METHODS

This was a prospective observational study performed on

a convenience sample of patients in a large urban teach-
ing hospital ED with over 136,000 ED visits per year.
This study was approved by the institutional review
board at our hospital.
Seventy-two practitioners (8 attending physicians, 54
resident physicians, 10 physician assistants) from the ED
received a 30-min lecture on ocular ultrasound. All study
practitioners had begun their ultrasound education and
practice within the last year and were required to attend
the ocular lecture in order to participate. The ocular
portion included the method of examination, contraindi- Figure 1. (A) Ocular ultrasound of a retinal detachment. (B)
cations to examination, and potential diseases encoun- Ocular ultrasound of a large vitreous hemorrhage with no
tered during assessment. The lecture emphasized the retinal detachment. (C) Ocular ultrasound of a vitreous de-
tachment with no retinal detachment.
different findings of the three intraocular diseases: retinal
detachment (RD), vitreous hemorrhage (VH), and vitre-
ous detachment. A retinal detachment appears as a taut rhage results in complete opacification of the vitreous
linear opacity within the vitreous chamber that moves in chamber. A vitreous detachment occurs when the vitre-
conjunction with eye movements (Figure 1A). A vitreous ous humor detaches from the posterior retina (Figure
hemorrhage consists of wavy linear or curved strands 1C). This results in a mobile “swaying seaweed” appear-
connecting with the retina that sway as the eye moves ance on ultrasound, where the vitreous appears separated
from side to side (Figure 1B). Severe vitreous hemor- from the retina. During the lecture, the practitioners were
Ocular Ultrasound 55

Table 1. Demographic and Clinical Characteristics of the RESULTS

Study Population

Characteristic Finding Ocular ultrasounds were obtained on 101 patients over 1

year. Of these patients, follow-up records were attained
Age
n 90
for 90 patients. The patients were 81% Hispanic, 71%
Mean ⫾ SD 45.5 ⫾ 26.1 male, 39% diabetic, and 20% hypertensive. Their aver-
Minimum 18 age age was 45.5 years (range 18 – 80 years, SD 26.1)
Maximum 80
Gender
(Table 1). Thirty-one practitioners contributed cases
n 90 (n ⫽ 101) to the study, of which 36 (36%) were done by
% (#) Male 71% (64) first-year Emergency Medicine (EM) residents, 55 (54%)
% (#) Female 29% (26)
Race
by second-year EM residents, and 6 (6%) by third-year
n with data 89 EM residents. One attending and three physician assis-
% (#) Hispanic 81% (72) tants also submitted one case each.
% (#) White 10% (9)
% (#) African-American 4% (4)
Two patients had examinations of both eyes second-
% (#) Asian 4% (4) ary to bilateral vision loss. Both patients had correct
Comorbidity interpretations with ocular ultrasound. Results were re-
n with data 85
% (#) no comorbidity 52% (44)
ported per eye; demographic data were reported per
% (#) diabetes mellitus (DM) only 28% (24) patient. Of the 92 eye examinations, 29 (32%) were
% (#) hypertension (HTN) only 8% (7) diagnosed with retinal detachments by Ophthalmology.
% (#) high cholesterol (CHOL) only 0% (0)
% (#) HTN ⫹ DM 8% (7)
Of these, we correctly identified 28 by ocular ultrasound
% (#) HTN ⫹ CHOL 1% (1) (sensitivity 97%; 95% CI 82–100%). Of the 63 exami-
% (#) HTN ⫹ DM ⫹ CHOL 2% (2) nations Ophthalmology stated did not have retinal de-
tachment, we correctly identified 58 examinations (spec-
ificity 92%; 95% CI 82–97%) (Table 2). The positive
predictive value was 85% (95% CI 68 –95%) and the
also instructed that the signs/symptoms of retinal detach-
negative predictive value was 98% (95% CI 91–100%).
ment included sudden loss of vision, visual field defects,
Of the six incorrect diagnoses, two were scanned by
or flashing/flickering lights in vision.
first-year EM residents, three by second-year EM resi-
Entry criteria included patients with complaints con-
dents, and one by an attending.
cerning to the practitioner for retinal detachment. Each
practitioner was instructed to submit a data collection
sheet that included their printed ultrasound scans
along with their ultrasonographic assessment for reti- DISCUSSION
nal detachment. Only patients who received ophthal-
mology follow-up were included in the study. Our results suggest that emergency practitioners with
Ocular examinations were performed using an Aloka minimal ultrasound training can use ocular ultrasound to
SSD-1400 (Aloka America, Wallingford, CT) 10-MHz detect retinal detachments. Given the high sensitivity
linear probe and a closed eyelid technique. Assessment (97%) and negative predictive value (98%) found in this
of retinal detachment included a three-step approach.
First, the emergency physician (EP) placed the linear
10-MHz probe on the closed eyelid of the eye of concern. Table 2. Comparison between ED Ultrasound and
Ophthalmology Findings
Second, the EP adjusted the contrast and brightness such
that small irregularities within the vitreous were visible, Ophthalmology Findings
creating an optimal image for assessment. Third, the EP RD⫹ RD⫺ Total
asked the patient to look left and right while the physi-
cian watched the ultrasound screen for movement pat- ED ultrasound
RD⫹ 28 5 33
terns consistent with different intraocular diseases. RD⫺ 1 58 59
We used the ophthalmologists’ evaluations as the gold Total 29 63 92
standard and compared our assessment with theirs. Their Sensitivity 97% (28/29); 95% CI 82%, 100%
Specificity 92% (58/63); 95% CI 82%, 97%
evaluation was not blinded to our interpretations. Their Positive predictive value 85% (28/33); 95% CI 68%, 95%
evaluation included indirect fundoscopic examination Negative predictive value 98% (58/59); 95% CI 91%, 100%
with dilatated pupils and intermittent use of ultrasound as Accuracy 93% (86/92); 95% CI 86%, 98%
an adjunct. We estimated the sensitivity and specificity ED ⫽ emergency department; RD ⫽ retinal detachment; CI ⫽
using 95% exact confidence intervals (CI). confidence interval.
56 Z. Shinar et al.

study, ocular ultrasound can potentially be used as an large busy academic setting. At our institution, a bed-
adjunct in the clinical assessment of retinal detachment. side ultrasound is readily available, and scans gener-
Comparing the results of our study to the current ED ally take ⬍ 2 min to perform. Under these conditions,
practice of direct fundoscopy alone, there is an advantage the results of our study show that ocular ultrasound is
to using ocular ultrasound. Siegel et al. found a 38% miss a useful skill for emergency practitioners to learn.
rate of retinal diseases that required intervention when Because the practitioners in this study were mostly
non-dilatated direct fundoscopy was used alone (2). junior residents with minimal ultrasound experience, it is
Given our results, ocular ultrasound could potentially conceivable that our data would be even better with more
improve this number significantly. experienced operators.
Analysis of the false positives and false negatives of
our study has lead to some interesting conclusions. The
one patient where an emergency practitioner misidenti- Limitations
fied a true retinal detachment (false negative) involved a
patient with diabetes mellitus. This patient presented to Our study was an observational study with several lim-
the ED with decreased vision over 2 months in her left itations. Our confidence intervals were wide and, despite
eye, which was previously treated seven times with laser having a large number of retinal detachments, a larger
for VH and RD. The EP diagnosed no pathology and sent study would be necessary to decrease these intervals and
the patient to the Ophthalmology clinic urgently. The strengthen our results. Another limitation was patient
ophthalmologist diagnosed a pre-retinal hemorrhage follow-up. Eleven (11%) patients either did not return for
with a small mac-on retinal detachment. The patient, follow-up or their charts were not available for review.
given her extensive previous treatments and subacute Additionally, our study suffered from potential bias in
course, was not given laser treatments but was referred to that practitioners who felt more comfortable with ultra-
retina clinic for follow-up the next month. sound may have been more likely to enroll patients.
Concerning the false positives, two involved subhya- However, a large number of the trained practitioners
loid heme. Subhyaloid heme is hyperechoic, occurs be- (n ⫽ 31, 43%) did contribute to the study. A final
tween the retina and the vitreous, and can mimic retinal concern was lack of blinding to the clinical presenta-
detachment because the border between the hemorrhage tion of the patient for both the emergency practitioner
and the vitreous humor may be mistaken for the retina. and ophthalmologist.
Usually, the dense hyperechoic retina with hypoechoic fluid
between the retina and choroid is distinctive enough to
differentiate retinal detachment from subhyaloid heme. CONCLUSION
Two other false positives had disc edema. Although the
images submitted do not show evidence of retinal Our study shows that ocular ultrasonography by emer-
detachment, one can postulate that significant disc gency practitioners is a potential adjunct in the evalua-
tion of retinal detachment.
edema could mimic a retinal detachment by increasing
the apparent retinal thickness.
An additional point of emphasis is that retinal tears,
REFERENCES
which are difficult to diagnose with ocular ultrasound,
can develop retinal detachment in a delayed fashion. 1. Tayal VS, Neulander M, Norton HJ, Foster T, Saunders T, Blaivas
There was one case in our study in which a patient was M. Emergency department sonographic measurement of optic nerve
seen emergently by an EP and an ophthalmologist for a sheath diameter to detect findings of increased intracranial pressure
in adult head injury patients. Ann Emerg Med 2007;49:508 –14.
traumatic eye injury. Both the EP’s ultrasound and initial 2. Siegel BS, Thompson AK, Yolton DP, Reinke AR, Yolton RL. A
ophthalmologist’s assessment showed no RD. In outpa- comparison of diagnostic outcomes with and without pupillary
tient ophthalmology follow-up 2 days later, the patient dilatation. J Am Optom Assoc 1990;61:25–34.
3. Coleman DJ, Silverman RH, Lizzi FL, Rondeau MJ. Ultrasonogra-
had developed an RD as fluid had accumulated behind phy of the eye and orbit, 2nd edn. Philadelphia: Lippincott Williams
the retinal tear. Consequently, patients with symptoms & Wilkins; 2006.
such as flashes and floaters who do not have an RD on 4. Byrne SF, Green RL. Ultrasound of the eye and orbit. St. Louis, Mo:
Mosby Year Book; 1992.
initial ED examination, do require timely outpatient oph- 5. Blumenkranz MS, Byrne SF. Standardized echography (ultrasonog-
thalmology follow-up for further evaluation. raphy) for the detection and characterization of retinal detachment.
Our practitioners included emergency medicine resi- Ophthalmology 1982;89:821–31.
6. Blaivas M, Theodoro D, Sierzenski P. A study of bedside ocular
dents, physician assistants, and an attending physician ultrasonography in the emergency department. Acad Emerg Med
who were working under normal shift conditions at a 2002;9:791–9.
Ocular Ultrasound 57

ARTICLE SUMMARY
1. Why is this topic important?
This topic is important because retinal detachment is
an emergency that is often difficult to diagnose in the
Emergency Department.
2. What does this study attempt to show?
This study attempts to show if emergency practitioners
can be trained to use ocular ultrasound to detect retinal
detachment.
3. What are the key findings?
The study shows that emergency practitioners with
minimal training can detect retinal detachment with a
high degree of sensitivity and specificity.
4. How is patient care impacted?
Patient care would be impacted by allowing for proper
disposition of ocular conditions, specifically separating
emergent from non-emergent referrals to Ophthalmology.
More patients with retinal detachment would receive
timely treatment to save their vision.