Intermittent Exotropia of the Divergence Excess Type - A View Point

Jeffrey Cooper, M.S., O.D., FAAO

Associate Clinical Professor, SUNY

Introduction

Classical teaching suggests that intermittent exotropia of the divergence excess type (DE) has the following characteristics: intermittency; normal stereopsis; rare amblyopia; good convergence amplitudes; and deep suppression during deviation. These patients align their eyes whenever stereoscopic information is placed in front of them. Some clinicians believe that the deviation seems to first appear at or around age six and that this deviation is a result of "socially compulsive near tasks" where there is an unchecked drive to release accommodation and thus convergence at distance. The result is divergent strabismus at distance. The etiology of DE, according to some clinicians, is an accommodative anomaly. However, this conclusion is untenable when one reviews the salient clinical findings. This is important since treatment should be consistent with both clinical and experimental findings. The following is a summary of a more comprehensive review on the subject which appeared in the Summer edition of Binocular Vision and Eye Muscle Quarterly (1993). This paper has 218 references encompassing most of the literature on DE including a statistical analysis of the various treatment options.

Sensorimotor Findings

Duane originally identified two distinct types of (intermittent) exotropes - convergence insufficiency and divergence excess (DE). Each type of exotrope has distinct sensory-motor findings and responds differently to treatment. Burian described a third type of (intermittent) exotropia in which he believed that a divergence excess acquired a convergence insufficiency. He labeled the third variety as basic exotropia. Both basic and divergence excess types of intermittent exotropia have similar the sensory motor characteristics, therefore, they may be thought as variants of the same anomaly. Costenbader provided an early, clinically accurate description of DE: exophoria or exotropia at distance, normal near point of convergence, adequate convergence amplitude, intermittency, equal vision, good stereopsis, and anomalous correspondence (ARC) when deviated. Costenbader reported that the deviation was first noted in the majority of DE before 18 mos. Progression often occurred until about age 6, at which time the strabismus became more noticeable. Only 6% of DEs are first observed after 5 years of age. DE is found more commonly in women and blacks. In addition this strabismus has a strong hereditary predisposition. The refractive distribution is similar to the normal population, not skewed towards myopia as originally suspected.

It is often reported that the ACA ratio is high in DE. This is a logical conclusion when one calculates the ACA using distance and near measurements. The average deviation is approximately 30 prism diopters at distance and 5 at near. Using these measurements one may calculate the average ACA using the formula: (distance deviation - near deviation + pd in cm)/near accommodative demand [(30-5+15)/2.5] or 16/1. This by definition results in a high ACA. However, there are too many clinical findings which mitigate against a high ACA ratio. For example, if DE patients really have a high ACA ratio they should complain about blur whenever they move their eyes in from the deviated position to alignment via the CAC cross link. Also, occlusion for 45 minutes in over 60% of the DE patients results in a substantial increase in the near deviation so that the angular measurement approximates the distant deviation. This would cause a significant reduction in the ACA ratio. Since occlusion does not effect accommodation it should not effect the ACA ratio. Studies by Ogle and Dyer who used gradient fixation disparity methods to measure the ACA in DE found them to be about 3.5/1, a relatively low ACA ratio. Von Noorden found normal ACAs using traditional gradient measurements. Lastly, if ACA ratios were really high one would expect a consecutive esotropia to occur at near after surgical intervention if orthophoria was created at distance. Happily for the surgeon this usually does not occur.

The distance-near ACA ratio measurements are tainted by proximal convergence, depth of field errors, blur interpretation, vergence aftereffects, and pupillary changes due to the near triad. To measure objective ACAs; Cooper, Ciuffreda, and Kruger used an infrared optometer to accurately and instantaneously measure accommodation, and an infrared eye movement system to measure eye position. They found that in both true and simulated DE the ACAs were normal and equal, i.e., 5.9/1. Kushner, on the basis of these findings measured gradient ACA in 83 DE patients. He reported that over 90% of the DE had normal ACAs and the small percentage that did have high ACAs resulted in surgical over-correction at near.

As previously mentioned, over 60% of DE patients who are occluded for over 45 minutes show a dramatic increase in their near deviation and a small increase in the distance deviation. This is termed simulated DE, while those who do not show this change in deviation with occlusion are termed true DE. The increase in the deviation after sustained occlusion demonstrates that occlusion does not immediately result in a loss of all vergence related responses. This phenomena is not unique to the intermittent exotrope. It is also observed in normal individuals. For example, if a normal individual wears a 25D base-out prism for a 45 minutes, and then is occluded the occluded eye will become esotropic with a slow regress towards orthophoria. If the occluder and prism are removed before orthophoria is regained, the subject will often report diplopia and an esotropia which decreases slowly over time. However, if fusion occurs immediately after the removal of the prism and occluder, there is an immediate esophoria which rapidly decreases. This demonstrates the strong effect that fusion has on the apparent position of the eyes. This effect has been called vergence aftereffects, prism adaptation, or slow vergence. If the subject goes to sleep with the prisms in place, and they are removed during sleep the subject will wake up with an esotropia with a slow recovery of normal binocular vision. This vergence mechanism is responsible for orthophorization in normal individuals. In addition, it is probably responsible for the smaller exo deviation at near, which is held in check by strong stereo cues initiating binocular alignment in DEs. Thus the slow vergence mechanism helps remove the load of sustained vergence (fast vergence) created by the large exophoria/tropia. Thus, simulated DE have strong vergence aftereffects while the true DE have strong proximal convergence. Strong proximal convergence accounts for the discrepancy between distance near ACA and the gradient ACA in true DE.

It is also noteworthy that most DE and basic exotropes are intermittent, even though, they both tend to deviate more upon distance fixation. The time of intermittency is not dependent on the angle of deviation. Whereas 20 ft is assumed to be infinity in most ophthalmic examinations this is not the case with DE. The deviation increases significantly in both amount and duration when fixation changes from 20 ft to 200 ft. When in bright sunlight most intermittent exotropes usually close one eye or deviate. Some clinicians believe that the closing of one eye is pathognomonic of a DE. When DE have amblyopia it is usually mild and related to an anisometropic error. Fusional amplitudes may be measured with a prism bar. A prism bar measurement allows the clinician to view the patient’s eyes and check for suppression or anomalous responses. Surprisingly, DEs have normal convergence fusional amplitudes, on the other hand, their divergence amplitude is usually less than the phoria measurement. This mitigates against the common perception that DE is the result of an inability to sustain fusional convergence or a result of a latent deviation becoming manifest. Fusion can be initiated by the presentation of stereoscopic detail, while deviation can be initiated by the presentation of simultaneous perception targets. Thus, the position of their eyes at any specific time is stimulus dependent. Said another way, the DE become binocular when there is an advantage, i.e., in the presence of stereoscopic information, and they deviate when there is no advantage - in the absence of stereoscopic information. Somehow DEs know apriori that a stimulus contains retinal disparity information.

When DE patients deviate a variety of things can happen: they can suppress, they can see double with normal correspondence (NRC), or they can change their retinal correspondence to one of anomalous correspondence (ARC). Cooper and Feldman investigated how the DE "sees" during deviation. They used a translucent hemisphere to present peripheral and foveal visual stimuli and monitored eye position with an EOG during both alignment and deviation. They found that during deviation the DE had an extension of the binocular field known as panoramic viewing. There was also a shift in retinal projection which matched the spatial world; harmonious anomalous retinal correspondence. In addition, they found that neither fovea suppressed during deviation. The foveas had different retinal motor values during deviation, but had the same retino-motor values during alignment. Thus, patients with DE have dual retinal correspondence. If the stimulus conditions are changed ARC may be eliminated to produce NRC with or without suppression. Therefore, retinal correspondence is not wired in but variable.

This phenomenon can not be explained by Burian's classical adaptation theory. Burian believed that ARC was a learned phenomena related to the age when the strabismus was first noted. Morgan explained the phenomenon with changes in egocentric localization of afterimages with changes in oculomotor position. For example, Urist reported that if one puts an after-image on one eye and if that person makes a version movement the image moves with the eye resulting in a change in egocentric localization of the afterimage, i.e., the afterimage moves to the side with eye movement (see fig 1). On the other hand, if one makes a vergence movement there is no change in egocentric localization, i.e., the afterimage appears straight ahead but smaller in size. In other words if the deviation is initiated by a version like defect there is a change in egocentric localization which would match the objective angle resulting in harmonious retinal correspondence (HARC). Re-alignment, as predicted by Morgan, results in normal retinal correspondence (NRC). If we use the same reasoning with retinal correspondence, there would be a change in egocentric localization with version induced movements. Thus, during alignment there would be NRC, while during alignment projection would change to correspond with the motoric movement or HARC.

INSERT FIG. 1 ABOUT HERE

What causes the DE?

Bielchowsky thought that the DE was a result of anatomical divergence of the orbits. Others have maintained that the DE is a result of abnormal insertions of the extra ocular muscles. Surgical exploration has not provided physiological support for mechanical theories. In addition, exotropia can not be surgically induced in monkeys even if the medial rectus is disinserted. After disinsertion the medial rectus often spontaneously reinserts with the development of normal concomitant movements. Duane thought that the deviation in the DE was a result of active divergence. There is strong evidence supporting Duane's hypothesis. Blodi and Van Allen have shown that during deviation the lateral rectus increases its rate of firing with a concomitant decrease in the firing rate of the medial rectus. Also, the deviation is greater than the maximum fusible divergence amplitude (phoria is larger than the divergence amplitude). This strongly suggests that the position of the eyes during the tropia phase is not due to a relaxation of convergence or an inability to converge rather to active divergence.

Posner suggested that the DE was a result of phylogenic development. The exotropia occurs with decerebralization of the cortex and a return to a lower level of binocularity (cerebellar function). Contrary to Posner, ocular placement of the eyes, straight ahead or laterally, is not related to phylogeny but to function. Animals which hunt have frontal position of their eyes for accurate depth location of their prey; while non-carnivorous animals, the hunted, have lateral position of the their eyes to increase the peripheral field to detect danger. The early onset and strong genetic predisposition of intermittent exotropia supports the theory that DE is neurologically programmed and not a learned phenomena.

I believe that the current information suggests that DE is a genetic anomaly that actually creates a purposeful deviation of the eyes. Chavase noted that during the course of evolution from vertebrates to primates the eyes began to move frontally with the final placement being related either to aggression or protection. Protective mechanisms were dependent on the largest possible field of vision while predation was based upon exact localization of the prey. Hunting animals such as man, monkey, wolves, tigers, cats, lions etc. have frontal position of the eyes while hunted animals such as rabbits, cows, horses, etc. have lateral placement of their eyes.

The DE seems to be a functional compromise of these two visual systems. These individuals have both stereoscopic vision when there is an advantage, and panoramic viewing when there are few stereoscopic cues. Panoramic viewing increases the teleological sensing system to a full 300 degrees and serves to expand the motion detection system of the eyes. During near viewing when stereoscopic cues are plentiful the eyes are usually aligned. An experienced clinician will attest to the fact that intermittent exotropes rarely deviate during stereoscopic testing. They demonstrate normal stereopsis on both contour targets such as the Titmus stereo fly test and on random dot stereograms. Thus, DEs have normal binocular disparity detectors in the cortex. The deviation is reduced by a feedback loop from the fast disparity vergence system to the slow vergence system. Sustained utilization of slow vergence reduces the load on fast system and decreases the apparent phoria. At distance, where stereo cues are less abundant, deviation occurs which decreases the output to the slow vergence system resulting in an apparent increase in the angle of deviation. When deviation occurs there is a shift in retino-motor values as predicted by Morgan's motor theory and ARC ensues. I believe that this chameleon like theory of being binocular with resultant stereopsis when advantageous, and deviating at distance viewing or when stereoscopic information is lacking is the true cause of DE.

Some clinicians have advocated that the basic etiology of DE is accommodative in origin. However, as previously noted the DEs have a normal ACA, relatively normal accommodative dynamics, and are first noted way before any socially compulsive near tasks occur. All of the abnormal findings reported are vergence related. Thus, it is difficult for me to accept an accommodative etiology.

Treatment

The cooperative patient

I have found that treatment which is not consistent with the physiological findings has not been very successful. For example, treatment to build fusional amplitudes in the DE who has normal fusional amplitudes does not result in a lasting cure. Treatment to eliminate the deviation by surgery does not result in lasting cure. Surgical treatment is the most successful when treatment is directed at creating a significant overcorrection so that the drive to re-develop original divergence excess deviation becomes mechanically impossible. Unfortunately the surgeon can not create a predictable over correction.

Treatment consistent with my theory employs behavior modification whereby the visual system is initially stimulated with targets which elicit binocular alignment (stereoscopic targets). Figure 2 depicts in flow chart form both diagnosis and treatment of the DE. After therapy using stereoscopic targets, the cues which stimulate binocular alignment (disparity and similarity) are sequentially faded out while providing reinforcement. The end result is alignment in the absence of any fusion stimulus. Brock was the first to advocate the use of large, peripheral stereoscopic stimuli to initiate alignment. He suggested that therapy begin at near viewing and then slowly move to distance where the stimuli become smaller, the retinal disparity induced by the stimuli decreases, and the patient is more likely to be exotropic. Before one moves from distance to near, the targets should be made progressively smaller. Cooper et al has experimentally shown, that as the stimulus becomes smaller, fusion amplitudes and the ability to maintain binocularity decreases. When the patient demonstrates a consistent binocular alignment response, the disparity cues are decreased, i.e. flat fusion targets are used. The last step of treatment uses simultaneous perception targets which are devoid of vergence cues. They are initially presented in free space at 40 cm and then moved to distance viewing. There is always reinforcement for alignment. Some of the techniques I employ use a vertical prism for dissociation to obtain ocular alignment. Some other examples of simultaneous perception targets include: Cheiroscopic and Amblyopia series with the Liquid Crystal Version of the Computer OrthopterÔ ; mirror transfer techniques; and right view of the Clown VectogramÔ with the left view of the SpirangleÔ . The last phase of treatment reinforces alignment in instrument in the absence of any cues for fusion. For some unknown reason, binocular alignment without suppression is more difficult when targets are presented in a stereoscope.

INSERT FIG. 2 ABOUT HERE

The hardest task for the typical DE is cheiroscopic tracing. For some reason it is easier for the DE to initially eliminate suppression while using a mirror cheiroscope vs. a Brewster stereoscope. Initially, cheiroscopic tracings are performed in the exo position the goal being no suppression and with normal projection. The ultimate goal is to maintain an ortho or aligned position, while performing a cheiroscopic drawing. One of the most effective techniques utilizes the liquid crystal version of the computer orthopter. The butterfly and the box are used in Auto Vergence, random pattern, of the Computer Orthopter. This target like the cheiroscope uses first degree targets to disrupt suppression mechanisms. The patient's goal is to match the stimulus position with their eyes. i.e., using voluntary vergence to keep the butterfly in the box as the vergence demand changes. The final goal of treatment is the maintenance of binocular alignment in the absence of any visual cues, i.e. a blank visual field.

Interestedly, the general tenets of this treatment plan were taught to me by some of those who believe that DE is a result of an accommodative problem. However, I never understood why the primary emphasis of the treatment program, advocated by those who believe that the etiology of the DE was the accommodative system, was the vergence system if the DE was a result of an accommodative anomaly. It should be noted that this therapeutic regimen employs minimal convergence amplitude training.

I personally believe that if one can obtain divergence fusional amplitudes greater than the measured objective angle, that the deviation will probably disappear. This is because you will have eliminated all ARC and suppression responses. In any case, if convergence therapy is employed it must be balanced by divergence therapy. Personally, I do not advocate the use of plus lenses at near for the typical DE. I have not seen a difference in the long term success between those patients prescribed glasses and those not.

One of the problems that I encountered in my early career was that after treating patients with the previously described regimen, it was not unusual to have infrequent reports of the exotropia occurring during periods of fatigue, illness, or prolonged non-visual attention. The most vivid one I recall was a patient with a DE on whom I had just completed a vision therapy re-evaluation. The parent was very pleased, because the child’s eyes no longer deviated. Cover testing revealed orthophoria at distance and near. Prolonged cover testing did not elicit a deviation. There was no suppression on cheiroscope etc. While I was patting myself on the back and completing the record, I noticed the patient from the corner of my eye. He was deviating without knowing it. Then I tried to teach pathological diplopia so to provide him/her with an additional cue that he/she was exotropic. This failed because I could not elicit a tropia during therapy.

Subsequently, I have changed my treatment regimen. I now employ pathological diplopia awareness at the onset of therapy. The technique involves darkening a room, putting a red lens in front of the fixating eye, and then occluding until deviation occurs. I then interpose a muscle light or pen light to obtain diplopia. If this doesn't work, I oscillate the pen light or use vertical prism to elicit diplopia. If I can't get diplopia at this point I probably never will, and the prognosis for a cure decreases somewhat. A cure occurs when the examiner can't elicit a deviation with occlusion and the parent/patient state that the deviation rarely or never occurs. If I can get diplopia then I proceed to do the following in order: decrease the oscillation, increase the room illumination, increase the transmission of the filter, move from near to far, and lastly dim the brightness of the transilluminator. When you finish this procedure most of these patients will have pathological diplopia or diplopia upon deviation. Your patient may complain of double vision, that's good. I always warn both the patient and the parent that during this phase of training the patient may complain of diplopia. We will eliminate it later on. Also, during this phase of training no fusional training should be given, since we want deviation and diplopia. Once this is accomplished I then begin therapy as previously described stressing postural alignment. Although, we do improve fusional amplitudes, voluntary vergence and cross-linking of both ACA and CAC with binocular accommodative rock training and accommodative rock with variable prism demands; the thrust of training is postural alignment in the absence of cues.

Treatment initially results in the reduction in the amount of time of deviation. Secondarily, this presumably strengthens the slow vergence system to hold the eyes in alignment through a feedback loop. Demer has recently discovered that within Tenon's capsule there is a fine muscular circular pulley system in which the extraocular muscles pass through. This pulley system is controlled by numerous smooth muscles. This may provide the mechanical mechanism for vergence adaptation. Over time there are probably changes in the length of the extraocular muscles. The lateral rectus lengthens and the medial rectus muscle shortens. There is strong histological evidence for this. Goldspink and his co-workers have shown that striated mammalian muscles in both adults and children can adapt by changing to a new length. This occurs by the addition or subtraction of sarcomeres at the tendon-myofibril junction. Thus, continuous vergence adaptation or slow vergence may cause a progressive change in the number of sarcomeres with a change in muscle length. The final result is permanent orthophorization. Lack of treatment should result in either no change in the deviation over time or a decreased ability to maintain sensory fusion over time. If the deviation occurs more frequently current theories would predict a decrease in vergence adaptation with secondary muscular changes due to elimination or addition of sarcomeres. Sacromere changes can occur within 7-15 days. The result would be in an increase in both the size and frequency of the deviation. Thus, sensory fusion changes ultimately affect both neural and muscular function.

Non-cooperative or unsuccessful patient

Though, I believe the regimen described has the greatest chance of success in treating the DE, I will employ any other technique which I think that might help. That includes patching, minus lenses, prismatic correction, and/or surgery. Young children, non-communicative patients, non-responsive patients, and patients who refuse vision training are provided passive vision therapy. Usually I begin by alternately patching each eye for 4-5 hours per day for 3 mos. I warn the patient that the deviation might increase with this procedure, but not to worry. If patching decreases either the apparent angle or the amount of time of deviation, I continue patching for another 3 mos. Surprisingly, this has a positive effect in over 60% of the cases. I may, also, arbitrarily prescribe -2.00 OU over their distance prescription. When over minusing works, suprisingly it is independent of the ACA ratio. Rutstein and London have shown that over minusing probably does not result in an increase in myopia. If there is a residual defect left I might add prism to eliminate the deviation with the goal of slowly tapering the prism over time. Lastly, if both passive and active vision therapy are unsuccessful I believe surgery should be performed to eliminate the motor deviation. Optimally, the surgeon strives for a 10 diopter over correction as long as the gradient ACA is within normal limits. Both you and the surgeon must be aware that if the gradient ACA is high over correction techniques are contra-indicated. Over correction techniques are also contra-indicated in adults.

Success Rates

There is controversy over which is the best method to cure the intermittent exotrope. This probably means that there are many ways to correct the DE, all of which are effective, or no one single treatment is always effective. Review of the literature suggests that orthoptics/vision therapy has the best success rate of between 60-90% while surgery only has a 45-80% success. It should be remembered that neither surgery nor vision therapy have ever undergone a rigorous double-blind, clinical trial to determine efficacy of treatment. Also, it is possible that the vision therapy and surgical population differ in magnitude of deviation and intermittency. I personally believe the best way to treat the DE is to use all your tools to increase the chances of success. I often will begin with patching, minus lenses and/or prism therapy before initiating active vision therapy. I then begin intense red lens therapy, followed by the therapeutic regimen that I have previously described. If we do not meet the patient's expectations, I will employ the skills of a well trained strabismus surgeon. Immediately after surgery, we will resume a short term vision therapy program. I believe this provides the best care for the patient.

Remember, no one gets 100% success of anything. Also, remember the DE is relatively asymptomatic other than cosmesis. His/her binocular system may be functionally better than the non DE individual. Treatment eliminates certain binocular advantages, specifically an increased field of view during deviation. Treatment should be undertaken only when cosmesis or asthenopia are of concern to the patient or parent. We advise that treatment takes approximately nine months of once a week therapy with home therapy as a supplement. In my experience home based therapy for any exo deviation including convergence insufficiency rarely works. One needs to monitor, reinforce, and alter therapy on the basis of performance. There is no control with home based therapy programs.