Proposal for Study Assessing Spontaneous Retinal Vein Pulsation
a multinational e-mail/web-based study
Copyright ©1999-2002, CARE-SRVP
* Completed June 1998 *

Proposed Primary Clinical Question

In patients with clinically suspected high intracranial pressure (high-ICP) and intact crania, does the presence of spontaneous retinal vein pulsation (SRVP) rule-out clinically important high-ICP?
(That is, can you SnNout high-ICP if you see SRVP?)

Proposed Secondary Clinical Questions

• If you see SRVP, can a lumbar puncture (LP) be done without causing brain herniation through the foramen magnum?
• Can you safely omit CT headscans for ruling-out high-ICP prior to lumbar puncture if you see SRVP?

Proposed "Biological" Question

In any patient undergoing CTHS, does SRVP rule-out high-ICP on that scan?

Clinical Relevance

1. A&E/ Emergency Room clinicians, Infectious Disease clinicians, General Internists, Neurologists, Neurosurgeons, Oncologists, and Paediatricians frequently see patients:
   • in whom it's reasonable to suspect high ICP as an explanation for their clinical state (headache, confusion, coma)
   • in whom a lumbar puncture is indicated (to detect CNS infection, bleed, or tumour)
2. Despite the paucity of evidence on its necessity/appropriateness, it has become customary in most places (and compulsory in some) to obtain CT head scans to rule-out clinically important high-ICP before performing lumbar puncture. The opportunity costs from compulsory head CT are big, both in money and in the time we spend hanging around waiting for it to be performed and interpreted.

Justification for why SRVP might be a useful indicator of ICP

(Based on CATs from my service plus a MEDLINE search from 1965-present on the MeSH terms: "RETINAL-VEIN"/ all subheadings AND "INTRACRANIAL-PRESSURE"/ all subheadings. The search and yield [11 citations] is appended to this draft as "medline.doc")

Justification from studies of normal individuals

First, Walsh^* et al confirmed SRVP in a convenience sample of 9 Neurology patients and then watched their SRVP as they underwent lumbar puncture and the Queckenstedt manoeuvre (in which compressing the jugular veins raises ICP and, if all is normal, "communicates" this rise to the spinal fluid and raises the CSF level in a manometer attached to the LP needle). He documented that:
• SRVP disappeared when spinal fluid pressure rose past 204 (+/- 17) mm H2O
• SRVP reappeared when spinal fluid pressure fell past 201 (+/- 17) mm H2O

^* [Walsh TJ, Garden JW, Gallagher B: Obliteration of retinal venous pulsations during elevation of cerebrospinal-fluid pressure. Amer J Ophthalmology 1969;67:954-6.]

Second, Lorentzen^† documented that SRVP was present in one or both eyes in 91% of females and 88% of males ages 10-79 (no decline with advancing age!). SRVP was slightly more likely to be present in the right (87%) than left (75%) eye.

^† [Lorentzen: Acta Ophthalmologica 1970;48:765-770]

Justification from studies of patients with clinically-suspected increased ICP

Levin^* examined SRVP in 43 patients he was convinced really had high ICP (by lumbar puncture > 190 mm H20 [oops!], findings at surgery, or convincing clinical evidence of brain herniation) and in 189 patients he was convinced really had normal ICP (by the absence of any signs, symptoms, or suspicion that they might have high ICP). The results are in Table 1. No patient with high ICP had SRVP. That is, if we think of the loss of SRVP as the abnormal sign, it had a sensitivity of 100% for high ICP in this small study, and the presence of SRVP "SnNouts" clinically important high-ICP. 88% of the 189 "normal" patients had SRVP (specificity = 88%), so the absence of SRVP is only suggestive, but not diagnostic, of high-ICP (it doesn't "SpPin" this diagnosis).

^* [Levin BE: The clinical significance of spontaneous pulsations of the retinal vein. Arch Neurol 1978;35:37-40]

Table 1: SRVP and ICP in the Levin Study

Sensitivity = a/(a+c) = 100% (93% to 100%)
Specificity = d/(b+d) = 88% (83% to 92%)
LR+ = sens/(1-spec) = 8
LR- = (1-sens)/spec = 0
Prevalence = (a+c)/(a+b+c+d) = 23%
Positive Predictive Value = a/(a+b) = 70%
Negative Predictive Value = d/(c+d) = 100%

I have asked every neurologist/neurosurgeon I know about this sign. All but one knew about it, most used it, and only one reported ever seeing SRVP in a patient with high-ICP, but he was a dean (I mean the doctor, not the patient!).

The current crude data suggest that the sign has the properties of a clinically "safe" finding

The anecdotal experience (that needs refutation or validation!) on our clinical service is that uncooperative patients and non-expert examiners result in under-reading rather than over-reading SRVP (that is, they report SRVP to be absent when, in fact, it is present). Since patients would bypass CT scanning only when SRVP was present, the errors fall on the side of clinical safety (albeit inefficiency).

** Queries for consideration and feedback marked with asterisks **

Proposed Primary Clinical Question

In patients with clinically suspected high intracranial pressure (high-ICP) and intact crania, does the presence of spontaneous retinal vein pulsation (SRVP) rule-out clinically important high-ICP?
(That is, can you SnNout high-ICP if you see SRVP?)

Proposed Secondary Clinical Questions

• If you see SRVP, can a lumbar puncture (LP) be done without causing brain herniation through the foramen magnum?
• Can you safely omit CT headscans for ruling-out high-ICP prior to lumbar puncture if you see SRVP?

Proposed "Biological" Question

In any patient undergoing CTHS, does SRVP rule-out high-ICP on that scan?

** Are these sensible questions? **
** other questions you'd like to include in this (or a subsequent) study? **

At any stage of training and experience.
** Should they view Sackett's video on SRVP before they begin? Or should this be part of a before-after study of the effects of the video? **

Any consenting patients who are about to undergo CT head scans (CTHS) (lumbar puncture with manometry (LP+M) is optional, but all have to undergo CTHS)
** Okay to reject patients who undergo LP+M but NO CTHS? Of concern is that we might miss some patients who undergo emergency LP (admittedly, mostly without manometry) for meningitis, subarachnoid haemorrhage, etc **
** should we exclude patients whose eyes "wander" so that we can't get a good look at their retinal veins (but keep a log of them and their CT results)? **
** What should we do about patients with cataracts (but keep a log of them and their CT results)? **

Patients will be of two sorts:
1. consecutive patients with clinically-suspected high ICP from any cause (as long as their fontanels are closed and their crania are intact); cause would be specified. This sample will answer the primary clinical question and infer answers to the secondary clinical questions.
2. consecutive patients undergoing CTHS for any reason. This sample will answer the biological question and build a large data set almost totally comprised (I think!) of patients with normal ICP.

** other sorts of patients we should include? **

Fundoscopy

1. By whatever method that is in routine use at the study site.
2. Performed "blind" to any CT or LP result or to any prior fundoscopy
3. Recommend (but don't insist on) starting with the right eye and proceeding to the left if SRVP NOT found in the right eye.
4. Use of mydriatics is optional, but should be recorded.
5. Examination by a second clinician ("blind" to the findings of the first) is recommended but optional.
6. Examination by an ophthalmologist/optometrist ("blind" to other fundoscopy) is ideal but not essential.
7. Reported as:
   • SRVP present
   • SRVP absent

** Should we also have a category for "SRVP indeterminate" when examiner really can't tell? This will provide a refuge, but also will give an indication of confidence and competency in finding SRVP**

CT Head Scan

1. Performed within 4 hours of fundoscopy.
2. No ICP-lowering Rx between the exam and the CTHS.
3. Performed "blind" to fundoscopy results.
4. Reported as:
   • high-ICP present
   • high-ICP absent

** what time gap between fundoscopy and CTHS should we permit? **
** should reporting categories be: high-ICP vs. "safe to LP"? **
** should we offer CT readers an "indeterminate" category **
** should we specify CT criteria for high-ICP and provide a tick-off form? **
** I will seek local input from Radiology on this, and you are welcome to ask your Radiologists as well. **

As with the COAD studies, via Internet to our Toronto Website with instant editing and entry into the data set.

(dummy tables)

Description of Study Clinicians

	Level of Training and Experience
	Student	PGY1	PGY 2-4	PGY 5-10	Practitioner
Country and Centre

 

Analysis for Trends in Skill with Training and Experience

(This will be a descriptive table and we'll cut the data any way that appears most informative)

	Level of Training and Experience
	Student	PGY1	PGY 2-4	PGY 5-10	Practitioner
SRVP Absent
Can't tell
SRVP Present

(This will tell us whether skill [measured by % of patients judged to have SRVP present] rises [and falls!] with training and experience; will stratify by whether they're familiar with SRVP before joining the study, and by whether they've looked at the training video])

 

Analysis for Agreement

		Second Examiner
		SRVP Absent	Can't tell re: SRVP	SRVP Present	Totals
First Examiner	SRVP Absent
	Can't tell re: SRVP
	SRVP Present
	Totals

(This series of tables will generate kappas for agreement above and beyond chance)

 

Analysis for Primary and Secondary Questions

		CTHS Result
		High-ICP	Can't tell	Normal ICP	Totals
SRVP	Absent
	Can't tell
	Present
	Totals

(This series of tables will examine the overall results and the results for the 2 patient subgroups [clinically-suspected high-ICP and consecutive CTHSs])

Please send any comments or suggestions to dave sackett at: sackett@bmts.com.

CARE Home	CARE Members	Status of Studies