The differential diagnosis of unilateral and bilateral superior oblique palsy (SOP) is sometimes difficult, and paralysis of an unoperated eye may become apparent after surgery on the more affected eye.1–3 The most definite diagnostic points of bilateral SOP are an alternating hypertropia (a right hypertropia in left gaze and a left hypertropia in right gaze) and a positive Bielschowsky test with head tilted to either side.4 The other signs of bilateral SOP include large excyclotropia in excess of 10° to 15°,5, 6 although this has been disputed,7 bilateral objective excyclotorsion of the globes on fundus examination,3 a significant increase of excyclotropia in downgaze,4 a traumatic origin with symptomatic cyclotropia,4 a significant V pattern of 15 prism diopters (PD) or more, chin depression,7 and a smaller amount of vertical deviation in the primary position.7, 8 However, bilaterality may not always be diagnosable using these criteria alone, because several patients with bilateral involvement did not show these findings.1–3 Kushner advocated that the assessment of oblique fields of gaze is essential for diagnosing bilateral SOP and that a qualitative cover test should be performed in these fields to determine the presence of hypertropia reversal.3 However, accurate prism and cover testing in the four oblique fields can be technically demanding. Thus, we considered the usefulness of the Lancaster red–green test (LRGT) (Luneau, France) to record the deviation, including torsion in such fields, and its ability to show excyclotorsion, V pattern, and the other criteria. Because the usefulness of the LRGT in the diagnosis of SOP has not been reported in the literature, we undertook this study to determine its usefulness both in the diagnosis of SOP and in the differentiation of unilateral versus bilateral SOP.
The tenets of the Declaration of Helsinki were followed throughout in this research. A necessary approval was secured from the ethics committee for human studies of Seoul National University Hospital Clinical Research Institute. Informed consent was obtained from all subjects after the study had been explained.
The medical records of 266 patients 15 years of age or older presenting with diplopia, who had undergone LRGT as described previously9 between January 2001 and December 2002, were retrospectively reviewed.
The LRGT was performed in a dark, quiet room. The patients were seated 1 m away in front of a rectilinear grid of black dots on the wall. The patients wore red–green goggles with the red glass over the right eye and the green glass over the left eye and held a flashlight projector with a green or red filter. The examiner held a flashlight of red filter and projected the streak light vertically on the central dot. Then the patient projected the green streak light to align the red streak on the wall. In this way, ocular deviations in the nine diagnostic positions of gaze with each eye fixing were obtained.
Sixty-three patients who showed cyclotropia by LRGT were enrolled. Forty-two patients were diagnosed as having SOP after a thorough history-taking, a review of previous photographs, and a full ophthalmologic examination, which included the alternate prism and cover test, an evaluation of ductions and versions, and the Bielschowsky head tilt test. In most cases, the Maddox double rod test at 50 cm and fundus photography had also been performed, and a few selected cases had undergone magnetic resonance imaging or computed tomography of the brain.
A diagnosis of bilateral SOP was made using one of the following criteria: 1) bilateral weakness of superior oblique muscles on duction/version testing, 2) a bilaterally positive head tilt test, 3) alternating hypertropia, 4) reversal of hypertropia in the oblique field of gaze, or 5) bilateral excyclotorsion on fundus photographs. With ancillary criteria of a significant V pattern (>15 PD) or a history of trauma, the possibility of bilateral SOP was also carefully evaluated.
LRGT results were analyzed using a computer (Fig. 1). After scanning the Lancaster charts at the same resolution and amplification ratio, the coordinates of red and green bars were obtained using the Photoshop program version 6.0 (Adobe, San Jose, California, USA).
Parameters, including degree of cyclotropia in primary position and downgaze, the presence of alternating hypertropia, horizontal and vertical deviations in the primary position, and the amount of V pattern, were evaluated in the unilateral SOP, bilateral SOP, and non-SOP groups. The absolute degree of cyclotropia was measured in the primary position (P) and in downgaze (D). For calculation, we indicated the excyclotropia as a positive value and the incyclotropia as a negative value. The difference of the degree of cyclotropia in the primary position and in downgaze (D-P) was calculated from the value of P and D. The alternating hypertropia was evaluated to be present when the right hypertropia in the left gaze and left hypertropia in the right gaze were shown in the Lancaster chart. The reversal of hypertropia in the oblique field of gaze was evaluated to be present when the hypertropia in the primary position was changed into hypotropia in any of the oblique fields of gaze was shown in the Lancaster chart. The amount of horizontal and vertical deviation was also calculated from the distance of the central points of the red and green bars. Because the horizontal deviation had directional component of exotropia and esotropia, the amount of horizontal deviation was indicated as a positive value for exotropia and a negative value for esotropia. In the asymmetric chart showing secondary deviation of paralytic strabismus, the chart of the secondary deviation was adopted. The degree of V pattern was defined as the subtracted value of horizontal deviations between downgaze and upgaze: H1 – H2 in Figure 1.
To check the reliability of LRGT, the degree of cyclotropia in the primary position (P) measured with LRGT was compared with the results measured with Maddox double rod test. The statistical analyses were performed using Excel 2002 program and SPSS 11.0 for Windows program (SPSS Inc., Chicago, Illinois, USA).
Among the 63 patients who showed cyclotropia on the LRGT, 33 patients were diagnosed as having unilateral SOP and nine patients as having bilateral SOP. The remaining 21 patients had skew deviation (six patients), thyroid orbitopathy (three), sixth nerve palsy (two), third nerve palsy (one), blowout fracture (one), inferior oblique palsy (one), inferior rectus palsy (one), previous strabismus surgery (one), exotropia with hypertropia (one), vertical gaze palsy after neurosurgery (one), internuclear ophthalmoplegia (one), and a cyclotropia of indefinite origin (two). The typical features of LRGT results in bilateral SOP, unilateral SOP, and non-SOP cyclotropia are presented in Figure 2.
Torsional diplopia was reported by eight of the 63 patients (13%): four of nine bilateral SOP (44%), four of 33 unilateral SOP (12%), and one of 21 non-SOP group (5%). The bilateral SOP group more frequently reported torsional diplopia than the other groups (p = 0.049 for unilateral SOP and 0.019 for non-SOP by Fisher exact test).
Degree of Cyclotropia
All the patients with SOP showed excyclotropia. Seventeen non-SOP patients showed excyclotropia and four non-SOP patients showed incyclotropia. The degree of cyclotropia for each of the three parameters (P, D, and D-P) was calculated for each subject group (Fig. 3). The degree of P was larger in bilateral SOP (16.3 ± 6.7° of excyclotropia) than in unilateral SOP (7.4 ± 4.7° of excyclotropia) (p = 0.003, Mann-Whitney test). The degree of D was also larger in bilateral SOP (20.3 ± 9.0° of excyclotropia) than in unilateral SOP (10.7 ± 3.7° of excyclotropia) (p = 0.013, Mann-Whitney test). However, the value of D-P was not different for the unilateral (3.4 ± 5.2°) and bilateral (4.0 ± 4.4°) SOP groups (p = 0.739, Mann-Whitney test).
Comparison between the SOP and non-SOP groups showed that degrees of P and D were not significantly different. The SOP group showed 9.3 ± 6.3° of P and 12.8 ± 6.5° of D. The non-SOP group showed 10.5 ± 6.1° of P and 10.4 ± 4.5° of D (p = 0.474 in P and 0.093 in D, t test). However, D-P was significantly larger in SOP (3.5 ± 5.0°) than in non-SOP (-0.01 ± 4.7°) (p = 0.008, t test).
The mean degree of cyclotropia (P) was 10.3 ± 7.2° measured by the LRGT and 7.2 ± 8.3° measured by the Maddox double rod test. The mean absolute difference of each degree of cyclotropia between LRGT and Maddox double rod test was 6.3 ± 4.3°. However, the difference between the two tests results was not statistically significant (p = 0.052 by paired t test). The degree of cyclotropia (P) measured with LRGT showed statistically significant correlation (p = 0.003, correlation coefficient of 0.589 by correlation analysis) with the degree of cyclotropia measured with Maddox double rod test (Fig. 4A). To assess agreement between the two methods, the statistical methods proposed by Bland and Altman10 were applied (Fig. 4B). The mean difference of cyclotropia between LRGT and Maddox double rod test was 3.0 ± 7.1°. The limit of agreement was −11.2° to 17.2°. The 95% confidence interval for the difference was −0.3° to 6.1°. The 95% confidence interval for the lower limit of agreement was −16.4° to −5.9° and for the upper limit of agreement 12.0° to 22.5°.
Alternating Hypertropia and Reversal of Hypertropia in the Oblique Field of Gaze
Four (44%) of nine patients with bilateral SOP and none of 33 patients with unilateral SOP showed an alternating hypertropia and reversal of hypertropia in the oblique field. The bilateral SOP group showed significantly larger frequency of alternating hypertropia or reversal of hypertropia in the oblique field of gaze (p = 0.001, Fisher exact test). None of the patients without alternating hypertropia showed reversal of hypertropia in the oblique field.
Horizontal and Vertical Deviation in the Primary Position
The absolute amount of horizontal deviation was 5.8 ± 7.1 PD in unilateral SOP and 5.6 ± 5.3 PD in bilateral SOP (p = 0.469, Mann-Whitney test) (Fig. 5). After adjustment by directional component of horizontal deviation of exotropia and esotropia, the value was -2.1 ± 8.9 PD in unilateral SOP and −1.6 ± 7.7 PD in bilateral SOP (p = 0.833, Mann-Whitney test). There was no significant difference in the horizontal deviation in the primary position between unilateral and bilateral SOP. The amount of vertical deviation was 8.4 ± 5.0 PD in unilateral SOP and 3.9 ± 2.6 PD in bilateral SOP. The bilateral SOP group showed a significantly smaller amount of vertical deviation than the unilateral SOP group (p = 0.007, Mann-Whitney test).
The Amount of V Pattern
The amount of V pattern was calculated by subtracting the horizontal deviation in downgaze from that in upgaze. The mean amount of V pattern showed a tendency of larger amount of V in bilateral SOP (mean, 11.3 ± 9.1 PD) than in unilateral SOP (mean, 6.5 ± 6.2 PD), but it was not statistically significant (p = 0.128, Mann-Whitney test). In the non-SOP group, four patients showed an A pattern and 17 patients showed a V pattern with various features.
LRGT comparisons of the SOP and non-SOP groups showed that the degree of cyclotropia in the primary position and downgaze was not different. The degree of cyclotropia in downgaze was larger than in the primary position in the SOP group, whereas in the non-SOP group, they were not significantly different.
LRGT comparisons of the unilateral and bilateral SOP group showed that the degree of cyclotropia in the primary position and in downgaze was larger in the bilateral SOP group. The two groups were similar in terms of the degree of cyclotropia difference in the primary position and in downgaze. Alternating hypertropia or reversal of hypertropia in the oblique field was more frequent in the bilateral SOP group. The vertical deviation in the primary position was larger in the unilateral SOP group, whereas the horizontal deviation was similar in the two groups. Finally, V pattern values were similar in the two groups.
Graphic analyses on LRGT findings have been previously performed. Zee et al.11 analyzed the Lancaster charts of paralytic strabismus by drawing lines on Lancaster charts. Gonzalez et al.12 introduced the computer-assisted Lancaster test and demonstrated its accuracy, reproducibility, and quantitative ability. In the present study, we performed computer graphic analysis of LRGT using a similar method to that by Gonzalez et al.12 In addition, we used several LRGT graphic analysis parameters (P, D, P-D, V, horizontal and vertical deviation) to differentiate unilateral and bilateral SOP. Using this graphic technique, we found that LRGT was both convenient and useful for analyzing cyclotropia and for measuring ocular deviation in nine different directions of gaze in SOP. We found that the evaluation of the degree of cyclotorsion was useful in the differential diagnosis of cyclotropia. Degrees of cyclotorsion in the primary position and in downgaze were found to be good differential diagnostic points for unilateral, bilateral SOP, and non-SOP. Although the Maddox double rod test or the Bagolini striated glasses test can also measure the degree of cyclotorsion in downgaze,5 the LRGT can yield more reliable results because the position of downgaze is constant and it can be extended by the examiner.
The comparison between LRGT and Maddox double rod test showed a significant correlation in the measurement of the degree of cyclodeviation. Our previous study13 regarding the evaluation of cyclotropia with LRGT also showed a good correlation between LRGT and fundus photography. Compared with fundus photography, the limitation of the LRGT in measuring cyclodeviation is that only the amount of cyclotropia and not the absolute amount of cyclodeviation in each eye can be measured.
With regard to comparisons between the SOP and non-SOP groups, the D-P value was found to be a good differential point. The LRGT D-P value may be taken as a marker of the functional integrity of the superior oblique muscle. In terms of unilateral and bilateral SOP, according to Kushner’s study,3 alternating hypertropia or reversal of hypertropia in the oblique field was the most differentiating feature. If alternating hypertropia or reversal of hypertropia in the oblique field was present in a patient with SOP, then the patient probably had bilateral SOP. Interestingly, in the present study, none of the patients without alternating hypertropia showed reversal of hypertropia in the oblique fields of gaze. If this is true in a large number of the patients with bilateral SOP, we may not need to perform a qualitative cover test in the oblique fields.
Torsional diplopia was significantly more frequent in bilateral SOP than in unilateral SOP. This finding may have been caused by a larger degree of cyclotropia in the bilateral SOP group. However, our previous research on cyclotropia showed that degrees of cyclotropia (P, D, and D-P) are not significantly related to the presence of torsional diplopia.13 Past studies have reported that all patients with bilateral SOP showed torsional diplopia that increased in downgaze.6, 7 Our results confirmed that the frequent presentation of torsional diplopia may be a diagnostic feature of bilateral SOP.
Compared with the number of patients with unilateral SOP (33), the number of patients with bilateral SOP (9) was relatively small in the present study. This small number of patients with bilateral SOP might have minimized the impact of some of the comparisons in our study and future research, including a higher number of patients with bilateral SOP, is necessary to confirm the results of the present study.
In conclusion, we found that the LRGT was useful for the diagnosis of SOP and for the differentiation of unilateral SOP and bilateral SOP.
This work was supported by grant no. R01-2005-000-10,875-0 from the Basic Research Program of the Korea Science & Engineering Foundation.
Department of Ophthalmology
Seoul National University Bundang Hospital
300, Gumi-dong, Bundang-gu, Seongnam
Gyeonggi-do 463-707, Korea