Fixation disparity (FD) is a small misalignment, often <10 min of arc (′) of the visual axes, that occurs while an individual maintains single binocular vision (within Panum's fusional area).1 Measuring a patient's FD through a series of test prisms, to generate a forced-vergence FD curve (FDC), can assist with diagnosis of a binocular vision disorder. Certain FDC variables are associated with asthenopic symptoms1,2 as well as with anomalies of binocular vision.1 Specifically, FDC criteria associated with symptoms are a non-type I curve,2,3 a slope in the central region of the curve of greater (in absolute value) than −1 min of arc per prism diopter (′/Δ),3–5 and a y intercept >4′ eso or >6′ exo.3
The Sheedy Disparometer (Fig. 1) has been used to measure FD for more than two decades, but it is no longer commercially available. The Saladin Near Point Balance Card (Saladin card, Bernell Corp., Mishawaka, IN) is a newer instrument for measurement of FD (Fig. 2). A previous comparison of these two devices6 showed that the Saladin card generally yields FD measurements that are more repeatable and of smaller magnitude compared with those obtained by the Disparometer. Measurements from both instruments tended to be in the same direction (eso or exo). In contrast, Ngan et al.7 found that the Saladin card and the Wesson FD Card did not produce sufficiently similar data to warrant use of the same norms. Corbett and Maples8 demonstrated that the Saladin card has good test-retest reliability. The purpose of this study was to further compare FD measurements obtained by the Saladin card and the Disparometer, including measurements through larger prism values than those used in the previous study of these two instruments,6 and to determine if the same clinical norms developed for the Disparometer can now be applied to the newer Saladin card.
Participants were 44 individuals aged 23 to 34 years, who were students at the Illinois College of Optometry (ICO). Inclusion criteria were best-corrected near visual acuity of 20/20 or better in each eye, no history of strabismus or amblyopia, no diplopia or eyestrain when performing near tasks for at least 1 h, and no use of medications that could affect accommodation. The study followed the tenets of the Declaration of Helsinki and was approved by ICO's Institutional Review Board. Informed consent was obtained from all subjects.
Three identical testing stations were set up in ICO's preclinical laboratory, with a different examiner at each to perform one test: dissociated phoria with Modified Thorington test (Star Ophthalmic Instruments, Willowbrook, IL), FD with Saladin card, or FD with Disparometer. The three examiners were all third-year optometry students at the time of the study and were specifically trained in the testing procedures by the first author. Subjects were randomized to begin at one of the three examiners' stations and then they proceeded to each in turn. At each station, subjects wore their habitual near correction (spectacles or contact lenses). They were seated comfortably in the examination chair. Consistent target lighting was provided by full room illumination and a 50 W stand lamp placed directly above the subject's forehead and directed toward the testing device. Standardized instructions were given to the subjects at each station. Examiners were masked to results from the other stations.
Modified Thorington Phoria
At one station, EM measured subjects' dissociated nearpoint horizontal phoria using the Modified Thorington card and a Maddox rod before the OD. The Modified Thorington card, containing an oblique line of numbers and letters with a small central hole through which a transilluminator light shone, was held at a 40 cm working distance. Subjects were instructed to fixate the white light, while keeping the letters and numbers on the card clear. Subjects were then instructed to report the letter or number that the vertical red line passed through. If the red line was reported as moving, subjects were asked to close their eyes and report the location of the red line when they first opened their eyes. The dissociated phoria was recorded as ortho or in prism diopters eso or exo.
At a second station, CDT measured horizontal FD through a series of prismatic demands using the Saladin card at 40 cm. Subjects wore polarized glasses and held the Saladin card at eye level. They were instructed to keep the words surrounding the polarized lines clear. A penlight held 2.5 cm behind the card illuminated each circle in turn as subjects reported which circle contained the set of vertical lines that appeared perfectly aligned. The endpoint was bracketed, and the corresponding FD value was recorded from the card. FD measurements were made in this manner through a sequence of no prism, 3Δ base-in (BI), 3Δ base-out (BO), 6Δ BI, 6Δ BO, 9Δ BI, 9Δ BO, 16Δ BO, and 20Δ BO. To help prevent prism adaptation, loose prisms were held by the examiner before each subject's OD for no longer than 15 s. Subjects were instructed to close their eyes for 15 s between each measurement. Subjects were to report if either polarized line disappeared (indicating suppression). If suppression occurred, the penlight was flickered off and on in an attempt to break suppression. If subjects failed to obtain fusion through a prism within 5 s, testing was stopped for that prism direction and no FD value was recorded for that test prism or for any higher prism value of that base direction.
At a third station, PE measured horizontal FD using the Disparometer without prism, followed by the same series of loose prisms in the order described above. Subjects again wore polarized glasses and held the Disparometer at 40 cm at eye level. They were instructed to keep the words surrounding the polarized lines clear, and to report if either polarized line disappeared (indicating suppression). The examiner held loose prisms before each subject's OD for no longer than 15 s, and subjects closed their eyes for 15 s between measurements. The instrument's dial was rotated until subjects reported that the vertical lines appeared perfectly aligned, and the corresponding FD value was recorded. Endpoints were bracketed. If subjects reported alignment at more than one line position, the center value of the range of alignment was recorded. Testing was completed for a given prism direction if any suppression or diplopia could not be resolved within 5 s. As with the Saladin card, no FD value was recorded for any test prism that could not be fused.
Because none of the FD data were normally distributed, non-parametric tests were used (Sigmaplot version 11, Systat 2008, Chicago, IL). One outlier was identified, who exhibited an FD value of 15′ exo through no prism, >2 standard deviations from the mean value. This subject's data were excluded from further statistical analysis. Because the maximum FD value measureable with the Disparometer is 25′, if a subject's FD was recorded as >25′, a value of 30 was substituted in the statistical analysis. Similarly, because the maximum FD value measureable with the Saladin card is 18′, if a subject's FD was recorded as >18′, a value of 20′ was substituted in the analysis. We did not believe it would be prudent to re-analyze our data without those subjects who had any measurements of >18′ with the Saladin card or >25′ with the Disparometer, because we would have had to eliminate 23 subjects.
FDCs were plotted for each subject to determine their type. The distribution of curve types was evaluated using the Mann-Whitney rank sum test. FDC slopes were obtained by linear regression using the FD values measured through 3Δ BI, 0Δ, and 3Δ BO with the Disparometer and the Saladin card.2,4 Medians and minimum/maximum values were calculated for the slopes as well as for the FD values measured through each vergence demand with each instrument. Medians were used rather than means to minimize the ceiling effect that could result from the maximum measureable FD being reached in a number of cases. In addition, medians and minimum/maximum values were used because the data were not normally distributed. The Wilcoxon signed rank test was used to compare the FDC slopes and FD values at each vergence demand with each instrument. In addition, a Bland-Altman analysis9 was used to evaluate the level of agreement between the Disparometer and the Saladin card, both for FDC slopes and FD values obtained at each vergence demand.
Based on review of FDCs plotted for each subject, the number of each Ogle curve type found for each instrument is shown in Table 1. There was no statistically significant difference between numbers of curve types found with the two testing methods (p = 1.0). Non-type I curves were found for six of 43 subjects with the Saladin card and for five subjects with the Disparometer. Only three subjects showed different curve types with the two instruments. In nearly every case, a subject's curves from the two instruments were very similar in shape. However, measurements with the Disparometer tended to be more eso (curves displaced upward) than those with the Saladin card, particularly measurements through BI prism.
The median slope (minimum and maximum values) in the central region of the FDCs (between 3Δ BI and 3Δ BO) was −1.1 (−5.67, 0.89) for the Disparometer, whereas the median slope for the Saladin card was −0.7 (−3.42, −0.17). A slope of greater absolute value than 1′/Δ was found for 21 of 43 subjects with the Disparometer and for 14 subjects with the Saladin card. The Wilcoxon signed rank test found a statistically significant difference (p = 0.048) between the slopes obtained with each testing method. Fig. 3 is a Bland-Altman difference vs. mean plot for the slope, demonstrating limited agreement between the two methods. The mean (95% confidence interval) of the difference (Disparometer − Saladin card) in FDC slope was −0.4 (−0.7, −0.06). The mean difference was significantly different from 0 (one-sample t-test, p = 0.022).
The Disparometer found the median FD without prism (the y intercept of the FDC) to be more eso/less exo than did the Saladin card (Table 2). A Bland-Altman difference vs. mean analysis (Table 3 and Fig. 4) demonstrated poor agreement between the FD measures with the difference between the means being significantly different from 0. Without prism, 23 of our subjects showed exo FD with the Saladin card, whereas only three subjects showed exo FD with the Disparometer. In contrast, without prism 38 subjects showed eso FD with the Disparometer vs. 10 with the Saladin card. (The remaining subjects showed a 0 FD.) All but one subject showed more eso or less exo FD without prism when measured with the Disparometer when compared with the Saladin card.
The median FD was more eso/less exo for the Disparometer than for the Saladin card for every prism demand except 6Δ BO and 9Δ BO (Table 2), the medians for which were very similar with the two instruments. The minimum/maximum values showed great variability in the measurements, particularly for those through BO prism. The Wilcoxon signed rank test found a statistically significant difference (p < 0.0001) between the FD values obtained by the two instruments through zero prism and all the BI values. No statistically significant differences were found between the two instruments for the measurements through BO prisms. Table 3 and Figs. 5 to 8 show the Bland-Altman difference vs. mean analysis for FD through the vergence demands. The mean difference was significantly different from 0 for all the BI prisms (demonstrating poor agreement between the two methods) but not for the BO prisms. The great variability in the difference between results obtained, as seen in the plots, points to difficulties in comparing these methods using BO prisms. Note that the Bland-Altman limits of agreement increase abruptly from ∼±4′ for the zero demand and BI prisms to ∼±12′ for BO prism demands.
The mean near horizontal phoria value measured by Modified Thorington was 2.66Δ exo, with a standard deviation of 3.10Δ, ranging from 12Δ exo to 1.5Δ eso. The direction of the Saladin card measurement without prism was in the same direction as the Modified Thorington near phoria in 26 of the 43 subjects analyzed (often exo). In contrast, the Disparometer frequently found eso FD without prism. The phoria was in the same direction as the FD in 20 subjects with Saladin card only, 6 subjects with both Saladin card and Disparometer, 5 subjects with Disparometer only, and 12 subjects with neither.
The proportion of Ogle FDC types found with each testing method was similar, but both instruments found higher numbers of type I curves and fewer type II curves among our subjects than Sheedy and Saladin previously found with the Disparometer1,10 (refer to Table 1). Although the numbers of FDC types obtained with the Disparometer and the Saladin card correlate with each other, the slopes found with the Disparometer tended toward larger absolute values than those found with the Saladin card, and the Bland-Altman analysis demonstrated limited agreement between the two testing methods. These differences were often clinically significant. Interestingly, a majority of subjects (who were originally selected because they were asymptomatic) did show slopes of <1′/Δ3–5 with the Saladin card, which would be considered normal. Fewer showed normal slopes with the Disparometer. In addition, despite the finding that the y intercept of the FDC with the Disparometer was significantly more eso than with the Saladin card, all but seven of our subjects' y intercept values measured with the Disparometer fell within the expected 4′ eso to 6′ exo range (these seven being >4′ eso). These same seven subjects had Saladin card measurements that were less eso/more exo and did fall within the 4′ eso to 6′ exo range, with the mean difference between their y intercept values from the two instruments being 5.1′ (range, 3 to 8′). Additionally, all but two subjects' y intercept values measured with the Saladin card fell within the 4′ eso to 6′ exo range (the two that did not were >6′ exo).
Our finding that the Disparometer tended to yield FD measurements that were more eso/less exo may be due to differences in the instruments' designs. Goss and Patel11 have shown that the Disparometer tends to yield more eso FD measurements because of its nonius lines being positioned behind the rest of the target. If the patient's lines of sight intersect in the plane of the letters adjacent to the polarized lines, the patient may display more eso, or less exo, FD than truly exists. We measured the nonius lines on our Disparometer to be 4.8 mm behind the plane of the letters. In contrast, the nonius lines on our Saladin card are 0.7 mm behind the plane of the letters surrounding the target window. Using the published formula to calculate the errors induced by these target offsets,12 we found a difference between instruments of 5.4′ (Disparometer measuring more eso than Saladin card). This error helps to account for the tendency of FD measurements with the Disparometer to be more eso for most vergence demands. This design difference may help to explain why Zurakowski et al.6 similarly found smaller FD measurements with the Saladin card than with the Disparometer (explaining at least those through BI prism, which generally produces eso FD). Clearly, another explanation is needed for the finding of Zurakowski et al.6 of smaller (less exo) FD values through BO prism with the Saladin card, and our finding of no statistically significant differences in FD through BO prism values.
An additional physical difference between the Saladin card and the Disparometer FD targets is the fusion lock. The Saladin card has a thin horizontal line visible by both eyes across the center of the target window, between the two polarized vertical lines that constitute the FD target. The Disparometer has no central fusion lock. Ukwade13 showed that FD measurements are more precise using both a central and peripheral fusion lock (as with the Saladin card) than with a peripheral fusion lock alone (as with the Disparometer). Given the subjects' expected lag of accommodation,14 accommodation may be postured even further behind the target without the fine central fusion lock in the Disparometer. This condition would decrease input from accommodative convergence and consequently increase the demand on fusional convergence when fusing the target through BO prism. Greater BO demand tends to result in greater exo FD. Thus, FD measurements through BO prism might be expected to be more exo with the Disparometer than with the Saladin card, counteracting the eso-producing effect of the Disparometer's nonius lines being recessed. The presence of a central fusion lock may also help to explain the decreased variability (smaller range) of Saladin card measurements through BO prism compared with measurements with the Disparometer.
Additionally, the lack of significant differences between the two instruments for FD measurements through BO prisms may be due to the greater variability of measurements through BO prism vs. BI prism with both instruments. Our finding of larger variability for BO measurements than for BI is similar to that of Corbett and Maples.8 It should also be noted that the higher the prism demand used, the more subjects showed suppression and thus there were less data to analyze. This situation was particularly evident for the 16Δ and 20Δ BO vergence conditions using the Saladin card (Table 2).
The mean and standard deviation of our subjects' near horizontal phoria with the modified Thorington card were similar to published normative values of 3Δ exo ±3.14 It is not surprising that a number of subjects showed a dissociated phoria in the opposite direction of their habitual FD, as Saladin and Sheedy10 found 38% of their subjects had a nearpoint phoria and FD in opposite directions using the Disparometer. Interestingly, in this study, the Saladin card found these measurements to be in opposite directions for 39% of subjects, but the Disparometer found them opposite for 74% of subjects.
We chose not to present an analysis of associated phoria data (x intercept of the FDC) because some subjects' curves never crossed the x axis, and we would have had to interpolate from the data collected to find the x intercept for many other subjects. Associated phoria commonly is determined by direct prism neutralization of FD using a device such as the Mallett unit, which we did not attempt to do in this study.
We conclude that in nearly all cases, the Disparometer and the Saladin Near Point Balance Card both generate the same FDC type for a given subject. However, FDC slopes differ in many cases, tending toward smaller absolute values with the Saladin card. In addition, measurements with the Disparometer tend to be more eso/less exo than those with the Saladin card, although measurements through BO prism are variable with both instruments. The Saladin card often produces FDC slopes and y intercepts within the normal range (as published for the Disparometer) for asymptomatic subjects. However, slopes and y intercepts obtained by the Saladin card are not sufficiently similar to those obtained by the Disparometer to warrant use of the same norms. Thus, we do not believe that norms developed for the Disparometer can be applied directly to the Saladin card. We do believe that the Saladin card is a useful clinical instrument, but further study is needed to establish appropriate norms for interpretation of its findings.
We thank Drs. Yi Pang and James Saladin for their helpful comments on this project.
Kelly A. Frantz
Illinois College of Optometry
3241 S. Michigan Avenue
Chicago, Illinois 60616
1. Sheedy JE. Actual measurement of fixation disparity and its use in diagnosis and treatment. J Am Optom Assoc 1980;51:1079–84.
2. Sheedy JE, Saladin JJ. Association of symptoms with measures of oculomotor deficiencies. Am J Optom Physiol Opt 1978;55:670–6.
3. Saladin JJ. Phorometry and stereopsis. In: Benjamin WJ, Borish IM, eds. Borish's Clinical Refraction, 2nd ed. St. Louis, MO: Butterworth Heinemann, Elsevier; 2006:899–960.
4. Sheedy JE, Saladin JJ. Phoria, vergence, and fixation disparity in oculomotor problems. Am J Optom Physiol Opt 1977;54:474–8.
5. Sheedy JE. Fixation disparity analysis of oculomotor imbalance. Am J Optom Physiol Opt 1980;57:632–9.
6. Zurakowski T, Keszo N, Saladin JJ. Repeatability of fixation disparity measurements: a comparison of the Saladin Near Point Card and the Sheedy Disparometer. Optom Vis Sci 2003;80(Suppl.):37.
7. Ngan J, Goss DA, Despirito J. Comparison of fixation disparity curve parameters obtained with the Wesson and Saladin fixation disparity cards. Optom Vis Sci 2005;82:69–74.
8. Corbett A, Maples WC. Test-retest reliability of the Saladin card. Optometry 2004;75:629–39.
9. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;1:307–10.
10. Saladin JJ, Sheedy JE. Population study of fixation disparity, heterophoria, and vergence. Am J Optom Physiol Opt 1978;55:744–50.
11. Goss DA, Patel J. Comparison of fixation disparity curve variables measured with the Sheedy Disparometer and the Wesson Fixation Disparity Card. Optom Vis Sci 1995;72:580–8.
12. Jaschinski-Kruza W, Schubert-Alshuth E. Variability of fixation disparity and accommodation when viewing a CRT visual display unit. Ophthalmic Physiol Opt 1992;12:411–9.
13. Ukwade MT. Effects of nonius line and fusion lock parameters on fixation disparity. Optom Vis Sci 2000;77:309–20.
14. Scheiman M, Wick B. Clinical Management of Binocular Vision: Heterophoric, Accommodative, and Eye Movement Disorders, 3rd ed. Philadelphia, PA: Wolters Kluwer Health/Lippincott Williams & Wilkins; 2008.