The Incidence and Waveform Characteristics of Periodic Alternating Nystagmus in Congenital Nystagmus

Abstract

purpose. To investigate the incidence and waveform characteristics of periodic alternating nystagmus (PAN) in congenital nystagmus (CN).

methods. In a prospective study, 18 patients with CN without associated sensory defects agreed to undergo eye movement documentation using binocular infrared oculography. Two of the 18 had a diagnosis of suspected PAN before entering the study. The patients sat in a dimly lit room and viewed an LED (4 mm in diameter) located in the primary position, at a distance of 100 cm. During an 8-minute recording, patients were read a story of neutral interest to hold attention at a constant level. PAN was defined as a left-beating nystagmus, a transition phase, a right-beating nystagmus, and a final transition phase; the sequence was then repeated.

results. Seven of the 18 patients had PAN (median cycle: 223 seconds, range 180–307 seconds). The periodicity of the cycles for each adult patient was regular, although the phases within a cycle were often asymmetric. Six of the seven patients had an anomalous head posture (AHP), and in five the AHP was in only one direction. Except for one patient, the PAN waveforms had an increasing slow-phase velocity in at least one phase of the cycle; in the other phase they were linear.

conclusions. The occurrence of PAN in CN is not as rare as previously thought and can be missed because of the long cycles and the use of only one AHP. The AHP was dependent on, and could be predicted from, the waveforms containing the longest foveation times. Although the waveforms and foveation times may differ among the phases of the PAN cycle, the periodicity of the cycle was usually regular and therefore predictable. Identification of PAN is essential in cases in which surgical treatment is considered for correction of AHPs.

Congenital periodic alternating nystagmus (CN + PAN) is a specific form of infantile nystagmus that is involuntary and conjugate and involves regular epochs of “active” and “quiet” phases. ¹ A cycle of PAN includes a left-beating nystagmus, a quiet or transitional phase, a right-beating nystagmus, and a second transitional phase. Patients with congenital PAN, as with CN, rarely report oscillopsia. ^{2 3 4}  

Congenital PAN had been thought to be a rare occurrence. ^{5 6 7} However, two recent reports, using eye movement recordings as the basis for diagnosis, have shown that PAN is prevalent in CN associated with albinism ¹ and in CN patients with and without sensory defects. ⁸ Abadi and Pascal ¹ reviewed the eye movements of 32 CN patients with oculocutaneous or ocular albinism and found that 12 patients (37.5%) had PAN. In general, the nystagmus intensity varied over the PAN cycle. None of the 12 patients had an anomalous head posture (AHP) or were able to null the nystagmus with convergence. Gradstein et al. ⁸ performed a retrospective analysis of clinical findings from approximately 200 CN patients with and without sensory defects and found 18 patients (9%) with a diagnosis ofPAN. Five of these 18 had albinism; all but two used AHPs. Both studies concluded that congenital PAN is underdiagnosed. 

AHPs are used by patients with CN to maximize visual acuity by placing the eyes in the orbital location where nystagmus is least. When an AHP causes physical or social discomfort, surgery should be considered to move the null point to the center of gaze. In the case of PAN, the AHP is more difficult to detect, because the null point changes orbital location and could result in intermittent or more than one AHP. 

The focus of this study was to investigate the incidence of PAN, including a description of cycle time and waveform characteristics in patients with PAN who do not have sensory defects. It is this group of patients in which misidentification of the null point may lead to inappropriate surgery. 

Methods

Subjects

Eighteen CN patients who did not report oscillopsia, ranging in age from 11 to 57 years (13 males and 5 females), participated in this study (Table 1 . As far as possible, an effort was made during patient recruitment to select only CN patients without sensory defects referred by colleagues from surrounding medical centers. 

All participants underwent ophthalmologic examination before testing. Patients with visual acuity less than 0.5 and no obvious sensory deficit underwent further electrophysiological testing. Visual evoked potential examinations were performed to rule out albinism and optic nerve lesions, and flash electroretinograms (ERGs) or pattern ERGs were performed to test for the presence of possible inherited retinal degeneration or dystrophies. Neuroimaging (computed tomography or magnetic resonance imaging) of the posterior fossa was performed on the children in the study to rule out optic nerve compression. 

Investigations were performed according to the guidelines of the Declaration of Helsinki. Subjects were fully informed about the nature of the procedures and gave their written consent before the beginning of the testing. The parents of the two children in the study signed the consent form. 

Eye Movement Recordings

The patients sat in a dimly lit room and were instructed to fixate a stationary LED (4 mm in diameter) located in the primary position, at a distance of 100 cm. Head movements were minimized using chin and head rests. Each subject’s eye movements were recorded for an average duration of 8 minutes, except in the case of the two children, who underwent eye movement recordings for 4 to 5 minutes. During the session, patients were read stories of neutral interest to hold attention at a constant level. Binocular, horizontal eye movements were recorded using an infrared limbal reflection technique (resolution, 0.1°; range, ± 20°). Eye movements were calibrated monocularly and binocularly before and after the 8-minute recording session. The eye position signals were digitized at 250 Hz and stored on computer for off-line analysis. 

Data Analysis

The following features of the nystagmus were examined: amplitude (peak-to-peak slow-phase displacement), frequency (number of oscillations per second), expanded foveation time (the period when the slow-phase velocity was between 0° and 10°/sec), beat direction of the fast phase, and, when found, cycles of PAN. 

Cycles of PAN

Fast-phase components of the PAN recordings were detected automatically using velocity and acceleration criteria. The remaining slow-phase regions were checked manually, to eliminate artifacts and undetected fast phases; then, the mean slow-phase velocity over each second of the eye movement recording was found. Plots of these mean slow-phase velocities versus time provided a concise illustration of the PAN cycles. 

For foveation time to have a meaningful relationship to acuity, the target image must lie within the foveal area, and retinal slip velocity must be low. When we measured the amount of time (beat-to-beat throughout the PAN cycle) that velocity was within a ±4°/sec window, few waveforms had periods of 60 msec or longer. Thus the usefulness of the method was less obvious. For clarity, plots were based on a velocity criterion of ±10°/sec to assist identification of best foveation times. This method describes the portion of the cycle that afforded best visual acuity and indicates a cause for a preferred AHP. In two cases, (patients 15 and 17), in which linear or decreasing slow-phase velocities occurred, the analysis was referred to as“ low-velocity” time. 

Results

Seven of the 18 CN patients (39%) had PAN. The PAN cycle duration ranged from 180 seconds to 307 seconds, with a median cycle duration of 223 seconds. PAN and CN waveform characteristics are summarized in Tables 2 and 3 , respectively. With the exception of one child (patient 17) and one adult (patient 4) who were referred with a clinical diagnosis of suspected congenital PAN, the patients entered this investigation with a diagnosis of CN only. 

Patient 1

A 30-year-old man with infantile esotropia underwent three surgeries (at ages 2, 10, and 11 years) to correct strabismus before entering this study. Funduscopy and visual evoked potential testing for chiasmal misrouting were normal. At 25 years of age, he was unsuccessfully treated with botulinum toxin to correct residual right-alternating esotropia (near 30 D and far distance 20 D). 

Eye movement recordings showed PAN, with one complete cycle lasting 185 seconds (Fig. 1) . At the time of the eye movement recordings, the patient had a constant 10° left AHP. This placed the eyes in a right orbital position, which was optimal for nulling the left-beating nystagmus, the phase of the PAN that had the shortest foveation times. 

Patient 3

This 40-year-old journalist had a normal fundoscopy and visual evoked potential. Eye movement recordings showed a PAN cycle of approximately 180 seconds (Fig. 2) . The constant left AHP, which placed his eyes in a right orbital position, was optimal for nulling the left-beating nystagmus. 

Patient 4

A 27-year-old man with normal visual acuity underwent two Kestenbaum ⁹ –Anderson ¹⁰ procedures at 12 and 14 years of age to correct an AHP. He continued to have a slight to moderate right AHP. He never used a left AHP and reported double vision in lateral gaze that was caused by a postsurgical limitation in abduction of the right eye. Eye movement recordings showed PAN with a mean cycle time of 236 seconds (Fig. 3) . 

Patient 10

A 38-year-old pharmacist with inherited CN and exophoria used an intermittent right AHP but never used a left AHP. Eye movement recordings showed PAN with a mean cycle of 293 seconds. The AHP placed the eyes either in the primary position or in left orbital gaze, which was optimal for nulling the right-beating nystagmus (Fig. 4) . 

Patient 15

A 12-year-old boy with suspected X-chromosomal recessively inherited CN and exophoria had PAN with a mean cycle duration of 221 seconds. The patient had an intermittent left AHP (0–10°) and never used a right AHP. Results of neurologic and ophthalmologic examinations, ERG, and magnetic resonance imaging were normal (Fig. 5) . 

Patient 17

An 11-year-old girl with infantile strabismus and CN, which was observed in the first few weeks of life, showed PAN with a cycle duration of 307 seconds (Fig. 6) . She used an alternating, intermittent AHP, left when fixating right and right when fixating left. Extensive neurologic and ophthalmologic examinations, including electroencephalogram and a computed tomographic scan, did not detect any abnormalities. 

Patient 18

A 29-year-old man who was an audiovisual engineer had good visual acuity and stereopsis and stopped using an alternating AHP as a teenager. Eye movement recordings showed a PAN with a mean cycle of 223 seconds (Fig. 7) . 

Discussion

Thirty-nine percent of the sample had PAN, which is consistent with the percentage reported in CN patients with albinism. ¹ One important clinical difference between the two sample groups is that none of the CN patients with albinism had an AHP, which would exclude them as candidates for surgical correction. It is not surprising that patients with sensory deficits (in this case, albinism) may not use AHPs, because increasing foveation time by way of a head turn would not markedly improve visual acuity. 

Gradstein et al. ⁸ and earlier investigators ^{5 6 7 11 12} reported a lower percentage of PAN in CN than either this study or the study by Abadi and Pascal. ¹ Some factors that may have influenced this difference beyond the inherent limitations of a retrospective analysis, are as follows.

1.  
   PAN may have been missed because of the long duration of a cycle. Gradstein et al. ⁸ used 2 to 3 minutes of continuous straight-ahead recording time. Our study showed that median PAN cycles were approximately 4 minutes and could be as long as 5 minutes.
2.  
   The absence of control for attention during the recording session. This factor has been pointed out by Gradstein et al.8p927 It can affect arousal and thereby modify nystagmus intensity. Although recording sessions were long in our study, it was not difficult to maintain the attention of the patients with the aid of storytelling. All participants, including the two children, could easily comply with the task demand and listen to the stories.Further, the recordings from the adult patients provided evidence that PAN found in CN can have regular, predictable cycles, which is critical when surgery is indicated.
3.  
   PAN may have been missed in the past, because patients may have used only one head turn. Patients with CN and PAN deploy strategies to use the waveform that affords the longest foveation times and thereby optimizes visual acuity. For example, as in patient 1 of our study, an AHP to the left was optimal for left-beating nystagmus. This is because a left AHP puts the eyes in right gaze position where left-beating nystagmus is minimal. The left-beating nystagmus could not be nulled when the eyes wereplaced in left orbital gaze. Therefore, a patient with PAN may use only one head turn, as was observed in five of the seven patients in this study, and this may be an important reason why the disorder had been missed in the past.

Findings from our study not only underscore the importance of diagnosing PAN but also show the advantage of describing the waveform characteristics that may be useful in determining the optimal surgical procedure. Review of the mean and peak slow-phase velocities along with the foveation time of each slow phase during the recording indicates which part of the cycle yields the best visual acuity and predicts the AHP that would most benefit the patient. 

Although studies with larger sample sizes are needed to provide the evidence necessary to suggest how long a patient with suspected PAN should be observed to determine whether PAN is present, our study provides two guidelines:

1.  
   If the predominant waveform is bidirectional (for example, pseudopendular), unidirectional with an alternating direction, or a mixture of bidirectional waveforms, the chance that PAN is present is negligible.
2.  
   In the case of a predominant jerk waveform, the methods presented in this study should be adequate to detect most PAN cycles.

The role of eye muscle surgery in patients with congenital PAN is still uncertain. Gradstein et al. ⁸ recently compared the outcome of the Kestenbaum ⁹ –Anderson ¹⁰ procedures with large recessions of all four horizontal muscles ¹³ and found that the latter procedure yielded satisfactory results for correction of AHPs. However, this may not always be the operation of choice for PAN, as suggested by the authors. A patient may have one consistent AHP as well as foveation characteristics that suggest a procedure that attempts to move the null zone from eccentric to primary gaze, as in patient 3 of our study who had a 20° to 25° left AHP and showed PAN that almost completely cycled within a left-beating nystagmus (Fig. 2) . Using an animal model, Dell’Osso et al. ¹⁴ suggested that an extraocular muscle tenotomy may be sufficient to damp nystagmus and improve visual acuity when AHPs are inconsistent and convergence does not diminish nystagmus intensity. The advantage of this procedure may be the elimination of double vision in extreme lateral gaze, which sometimes occurs after large rectus muscle recession procedures. This technique has yet to be used in patients in a controlled trial. 

In summary, the occurrence of PAN in CN is not as rare as previously suggested and can be missed because of long cycles and the patient’s preference for only one AHP. Although phases within a PAN cycle can be asymmetric (greater duration of beating in one direction over the other), evidence indicates that the cycles of PAN in CN are usually regular and that the motion of the null is not susceptible to modification. This is in contrast to the waveform and foveation characteristics of the nystagmus. Identification of PAN and its waveform characteristics are essential in cases in which surgical treatment is considered for correction of AHPs. 

 Submitted for publication December 21, 1998; revised May 27, 1999; accepted June 15, 1999.

 Commercial relationships policy: N.

 

Table 1.

 

Clinical Status

Patient	Age	Sex	Visual Acuity (Snellen)	Stereopsis*	Diagnosis
1	30	M	0.3; 0.3	Nil	CN, PAN+ infantile esotropia
2	28	M	0.5; 0.5	50	CN
3	40	M	0.67; 0.67	140	CN, PAN
4	27	M	1.2; 1.0	40	CN, PAN
5	33	M	0.5; 0.5	TNO: 480	CN
6	19	M	0.5; 0.33	800	CN
7	57	F	0.67; 0.67	TNO: 240	Autosomal dominantly inherited CN
8	26	F	0.5; 0.25	TNO: 800	Autosomal dominantly inherited CN
9	47	M	0.33; 0.33	80	CN
10	38	M	0.25; 0.25	200	Inherited CN, PAN
11	38	F	0.5; 0.5	50	Inherited CN
12	26	F	0.67; 0.67	40	CN
13	20	M	0.67; 0.67	200–100	CN
14	55	M	0.67; 0.5	100	CN
15	12	M	0.25; 0.33	Nil	CN, PAN+ exophoria
16	25	M	0.33; 0.5	60	CN
17	11	F	0.33; 0.33	Nil	CN, PAN+ infantile strabismus
18	29	M	0.5; 0.5	100	CN, PAN

*  Titmus test, unless otherwise indicated (disparity range, 3000–40 seconds of arc); TNO test (disparity range, 480–15 seconds of arc). Data are in seconds of arc.

Table 2.

 

PAN Waveform Characteristics

Patient	Anomalous Head Posture	Mean PAN Cycle (sec)	Predominant Waveform	Mean Phase Duration (sec)	Mean Amplitude (deg)	Mean Frequency (Hz)	Peak Slow-Phase Velocity (°/sec)
1	10° Left	185 sec	Left jerk	70	8.8	3.0	60
			Transition (PP)	18	6.9	4.0	75 (spv left) 37 (spv right)
			Right jerk	75	3.0	4.0	25
3	20–25° Left	180 sec	Left jerk	84	4.7	4.0	57
			Transition (nulled)	6	0	0	0
			Right jerk	10	2.5	2.0	8
4	10° Right	236 sec	Left jerk	157	6.0	3.0	25
			Transition (nulled)	9	0	0	0
			Right jerk	96	5.1	5.0	55
10	0–10° Right intermittent	293 sec	Left jerk	188	9.5	4.0	44
			Transition (pendular/triangular)	5	4.8	2.0	5
			Right jerk	178	8.5	4.0	39
15	0–10° Left intermittent	221 sec	Left jerk	140 (lin spv)	6.5	3.0	83
			Transition (nulled)	0	0	0	0
			Right jerk	178 (dec spv)	5.9	4.0	61
17	0–10° Left alternating	307 sec	Left jerk	192 (lin spv)	5–10	5.0	107
			Transition (pendular)	96	0.65	8.0	49
			Right jerk	340	3.8	4.0	55
18	Nil	223 sec	Left jerk	90	3.0	3.0	38
			Transition (pendular)	13	3.6	3.0	44
			Right jerk	93	4.0	5.0	55

 PP, pseudopendular; lin spv, linear slow-phase velocity; dec spv, decreasing slow-phase velocity.

Table 3.

 

CN Waveform Characteristics

Patient	Anomalous Head Posture	Predominant Waveform	Mean Amplitude (deg)	Mean Frequency (Hz)	Peak Slow-Phase Velocity (°/sec)
2	Nil	Pseudopendular right and left	4.2	3.5	47 (spv left)
					50 (spv right)
5	10–15° Right	Pseudo-pendular left	4.38	3.0	32 (spv right)
6	10° Right tilt	Pseudo-pendular right and left	2.0	6.0	21 (spv left)
					35 (spv right)
7	Nil	Pseudo-pendular right and left	5.0	6.0	46 (spv left)
					26 (spv right)
8	Nil	Pseudo-pendular right and left	1.5	6.0	34 (spv left)
					14 (spv right)
9	0–10° Left	Right jerk; elongated foveation	5.7	4.0	57 (spv left)
11	Nil	Bidirectional jerk	5.0	4.0	80 (spv left)
					33 (spv right)
12	Nil	Alternating direction right	0–2	0–4.0	4 (spv left)
		Alternating direction left	1.3–4	0–4.0	20 (spv right)
13	10–15° Left	Alternating direction right	5.0	4.0	50 (spv left)
		Alternating direction left	4.0	4.0	69 (spv right)
14	Nil	Right jerk; elongated foveation	5.0	3.0	68 (spv left)
16	Nil	Alternating direction right	1.2	3.0	28 (spv left)
		Alternating direction left	1.1	3.0	11 (spv right)

Figure 1.

 

(Top) The mean slow-phase velocity (spv) of PAN in patient 1 as a function of time. Positive spv indicates movement to the subject’s right. Each data point represents the mean spv over a period of 1 second. The PAN cycle had a regular periodicity, with greater slow-phase velocities occurring during phases of left-beating nystagmus (up to 60°/sec) than right-beating nystagmus (up to 20°/sec). (Middle) The expanded foveation time of each nystagmus beat during the recording. The expanded foveation time was defined as the duration for which the spv of each beat was less than ± 10°/sec. For this subject, the shortest foveation times (<60 msec) occurred during periods of left-beating nystagmus, which could be the basis of the patient’s poor visual acuity. The longest foveation times occurred during the transition phases (indicated by the vertical bars between the top and middle plots) and shortest during left-beating nystagmus. (Bottom) Representative infrared right eye movement recordings for each phase of the PAN cycle (right-beating nystagmus phase, pseudopendular, bidirectional jerk transition phase, left-beating nystagmus phase) had increasing slow-phase velocities. The upper trace shows eye position (R, right; L, left); the lower trace shows the eye velocity.

Figure 2.

 

Notation as in Figure 1 . (Top) The PAN cycle of patient 3 has a left-beating bias with a minimal right-beating nystagmus. (Middle) During periods of left-beating nystagmus, occasional, long-duration slow phases produced intermittent long foveation times. For this subject, consistently long foveation times occurred during the transition and right-beating nystagmus (e.g., at approximately 200 seconds into the recording) and could explain good visual acuity. Foveation times of 600 msec and more are represented on the y-axis at the 600-msec mark. (Bottom) Representative right eye movement recordings of a left-beating nystagmus phase, a transition phase with one of the intermittent long-duration slow-phase velocity waveforms, and a right-beating phase.

Figure 3.

 

Notation as in Figure 1 . (Top) The peak spv of the right-beating nystagmus reached 55°/sec, whereas peak spv of the left-beating nystagmus remained below 25°/sec. (Middle) Longest foveation times occurred during transition and left-beating nystagmus phases. (Bottom) Representative right eye movement recordings for each phase of the PAN cycle, with increasing spv waveforms in both left- and right-beating phases and a nulled transition phase.

Figure 4.

 

Notation as in Figure 1 . (Top) PAN of patient 10 with a regular 5-minute cycle. (Middle) Foveation times were relatively short for both right- and left-beating nystagmus phases and could account for this patient’s poor visual acuity. (Bottom) Representative right eye movement recordings for each phase of the PAN cycle, with increasing spv waveforms in both left- and right-beating phases and a damped transition phase.

Figure 5.

 

Notation as in Figure 1 . (Top) The plot shows part of a relatively long PAN cycle from patient 15. (Middle) The low slow-phase velocity times were short, except during the transition phase when this patient had better visual acuity. (Bottom) Representative right eye movement recordings show that the left-beating nystagmus had mostly linear waveforms with a decreasing spv component that did not return to zero velocity before a saccade was executed. The right-beating nystagmus consisted of a decreasing spv waveform. Both waveforms had no acceptable foveation times. The longest low-velocity times occurred during and just before the transition phases. These times were often less than 60 msec, during the periods of both right- and left-beating nystagmus and could account for the patient’s poor visual acuity.

Figure 6.

 

Notation as in Figure 1 . (Top) The plot shows a section of a relatively long PAN cycle in patient 17. (Middle) Foveation times were generally below 60 msec during right-beating nystagmus and were negligible during the main phase of left-beating nystagmus. (Bottom) Representative right eye movement recordings of the left-beating phase, the transition phase with pendular nystagmus, and the right-beating nystagmus phase. The right-beating nystagmus, with slow phases of increasing velocity, had short foveation times, and the linear slow phase left-beating nystagmus and pendular transition phases had no foveation time. These waveforms yielded foveation times that would not support good visual acuity.

Figure 7.

 

Notation as in Figure 1 . (Top) Two complete cycles of PAN (patient 18). The cycles were regular, approximately 4 minutes long. (Middle) This patient had relatively long foveation times during all phases of PAN, which can account for the patient’s good visual acuity. (Bottom) Representative right eye movement recordings show that both right- and left-beating nystagmus had increasing spv waveforms that yielded relatively long foveation times. During transitions, the waveforms damped increasing foveation times.