The College of Optometrists

how accurate wud u say the van herricks technique is?

also from past questions there was 1 that said would u refer from van herricks alone?

surely you would have to chek pressures and fields?

can any1 explain where such a question is coming from?

you could refer on Van Herricks alone if angles were narrow or near closed for possible prophylactic iridotomy. or maybe if there has been an acute angle closure attack in the other eye previously.

van herrick on own - not too bad - combine with nazal shaddow test - result better!

yes - we refer for narrow angles - even if pressures normal and feilds okay. You can't predict when it will go. And angle goes narrow over the years.

Can only assess (and urgency) after pressures are known.

Look at retinal nerve fiber layer as well for damage and also lens for glaucomaflecken. The iris would be atrophied in chronic sub acute.

I would say the accuracy of VH depends on what your hoping to use it for.

All it really tells you is that the patients angles "might" be narrow.

From my own experience:

At VHG4 the angle is highly likely to be open/wide open(shaffer III and IV) when viewed with gonioscopy.

At VHG3 the angle is still probably open but may occasionally be narrow.

At VHG2/VHG1 the angle could be very narrow, narrow, moderately narrow or open. You just can't tell without gonioscopy.

You should refer a patient if you think they have very narrow angles but if you use van herricks for this be prepared for a significant number of false positive referrals.(you'll also probably miss some patients with very narrow angles as well)

I find VH's a useful technique to identify which patients I'm going to perform gonioscopy on.

Interestingly a patient I induced acute glaucoma on actually had angles I graded at VHG3. I didn't have a goniolens at that time but wasn't concerned about angle closure before I dilated him.

Due to the inaccuracy is there any point in even grading van herricks. surely it would be enough to check if they weren't closed and recorded it as such. I say this as an opthalmolgist at a LOC conference in sunderland said that as far as he was concerned it was better that we induced closed angle glaucoma in a clinical setting where something could be done about it, instead of when the Px was at home or somewhere where help wasn't readily available.

Do you err on the side of caution and not dilate when a Px has a VHG1 or do you say it is more beneficial to dilate as long as you tell the Px the risk of induced CAG and to seek medical attention immediately if acute painful red eye.

For that matter if a Px had closed angles he/she would know about it wouldn't they. So should I even worry about grading the angle at all with this inaccurate technique or should I dilate no matter what and worry about the consequences after. After all if i do induce an attack and the Px follows my instructions they will present at the hospital and have an iridotomy so reducing the risk of future attacks.

What's the worst that could happen?

Appendectomy?

Sorry to disagree. The problem is not rise in IOP shortly after that is the concern but the mid dilated pupil block a few hours after that leads to a closed attack. People suffering a closed attack are not always in acute pain. Lucidity of thought in an eighty year is not always what it should (?) be.

Homer,

If you're not prepared to learn/use gonioscopy then using van herricks/shaddow test is the best option you've got left.

Dilating without doing either and not worring about the consequences is an option. However, if you do this and don't warn your patient of the symptoms of angle closure I don't see how you can be considered to have "done them a favour". The patient may sit at home and think to themselves "my vision's blurred and my eye hurts but my optician said my eyes were o.k this morning so it will probably be better by the morning"

If their pre-warned (and so likely to seek help) then perhaps you have "done them a favour". Although try telling the patient that!!! Also despite prompt treatment some damage may occur or the patient may be left with chronic glaucoma.

If you go down the route of using van herricks then if you think a patient has very narrow angles (VHG1)and is at risk of angle closure following dilation then you have two options:

1. Refer them for gonioscopy.

2. Dilate them.

If you choose to dilate it should be done under controlled clinical conditions i.e.

1. You should be able to moniotor the IOPs post dilation if required.

2. Leaflet given to patient telling them to return or seek immediate medical attention if they experience problems when they get home etc etc.

Say you dilate them and all is o.k. They still may have very narrow angles and need prophlactic iridotomies.(the patient I induced acute glaucoma on had been previously dilated by me without incident).

Without using gonioscopy it is not easy to know what to do. However, the chances of it happening to you are small but I don't think they can be ignored. Having a patient sat infront of you with an IOP of 78mmHg that had been 20mmHg four hours ago is a very sobering experience I can tell you!

Always remember seeing a patient in my pre-reg year who had had a briefly painful eye in the night, which had then felt OK but her vision was a bit blurred. Saw her optometrist the next day and was referred in urgently, with a fixed dilated pupil, slightly clouded cornea and IOPs of 70 mmHg. The patient sat quite happily in the consultant's room, no pain at all, he reckoned the acute rise in pressure had probably destroyed all the pain receptors. VA was LP.

1) Accuracy wise,

VH is adequate for differentiating narrow from wide open angles, but occassionaly (in fact very rarely) caughtout by plateu iris.

Poor at differentiating closed from narrow but open.

2) Referral based solely on VH,

No definitive answer. VH is poor with narrow angles, most narrow angles do not go on to develop ACG, and thus referral will have a low specificity.

Assuming gonioscopy is not available, and even if it is - there is no agreed criteria of when to create a prophylactic PI. My descision would consider other risk factors, particularly symptoms (but often no symptoms if angle closure chronic) or previous ACG in other eye; but also central anterior chamber depth (Smith technique) and the contour of the iris (prominant midperipheral bowing more at risk that flatter); and top lesser extant, asian race (or eskimos but these are rare where I work), age, female gender, and cataract.

At the risk of being accused of pedantry,

"VH is poor with narrow angles, most narrow angles do not go on to develop ACG, amd thus referral will have a low..."...sensitivity.

Specificity will be quite high because almost all closed angles are narrow by VH, but not all angles narrow by VH are closed.

John, I value your input highly, and have thus questioned (and confused) myself, but still believe I was correct in the wording.

Sensitivity refers to the proportion of people with disease who have a positive test result.

The sensitivity of VH referral is high, as it correctly identifies almost everyone with an occludable angle, i.e. few false -ves.

Specificity refers to the proportion of people without disease who have a negative test result.

The specificity of VH referal is comparitively low, as many of the people referred with the technique will not have an occludable angle, i.e. many false +ves.

I think...

We may be trying to say the same thing Michael, so I will try not to make any assumptions in clarifying the wording.

You are correct in your statement that sensitivity is the proportion of people with a disease that test positive.

Van Herrick’s test does not adequately differentiate between patients who have a narrow but open angle and those that are at risk of angle closure. Many patients with angles narrow by VH do not have occludable angles. If we take 100 people with narrow angles by VH and only 30 of them really have an occludable angle, the sensitivity of the test is 30%.

The test is not accurate in identifying those with disease and will result in many false positives (70%) because it has low sensitivity.

Specificity is the proportion of people without disease who have a negative test result.

An angle that is wide open by VH is highly unlikely to be occludable. If we take 100 people with a wide angle by VH and 1 of them has an occludable angle the specificity of the test is 99%.

The test is very accurate at identifying those without disease and will not give very many false negatives (1%) because it has a high specificity.

It is the sensitivity of a test that dictates the false positive rate, and the specificity that dictates the false negative rate.

on much lighter note - a dodgy angle changes slowly and can be detected by measuring it every year and jumping in when you notice it narrowing. A VHerrick is okay for this.

John, at least we agree in our opinion regarding the utility of VH.

The crux of why we quibble on wording is based on what we consider the +ve test and the disease.

Me:

Narrow angle with VH = +ve test

Occludable angle = disease

You:

Narrow angle with VH = disease

Occludable angle = +ve test

For example,

You stated that you agree, sensitivity refers to the proportion of people with disease who test positive. Using your scenario,

"If we take 100 people with narrow angles by VH and only 30 of them really have an occludable angle"

You: 100 people with disease, 30 people with +ve test = 30% sensitivity

Me: 100 people with +ve test, 30 people with disease = 30% chance of +ve VH test result correctly identifying an occludable angle, i.e. VH test has 30% specificity

To avoid differences in terminology distracting from my message, I will reword my sentiments thus,

If an angle is occludable, VH is very good in detecting it, but is does not do this accurately/efficiently, because it often incorrectly flags angles that are unlikely to close.

Sorry, Michael there is still some confusion.

I am not saying 100 people with disease, 30 test positive. I am saying exactly the same as you:

i.e. 30 people with disease out of 100 who test positive. This gives 30% sensitivity not specificity.

not quite sure what your getting at here palfi?

The 'test' is an indicator of the 'disease'