Purpose
To develop and validate an improved measure of control in intermittent exotropia (XT).
Design
Prospective, noninterventional case series.
Methods
Twelve children with intermittent XT were evaluated during 4 sessions (2 hours apart) over a day, on 2 separate days (8 sessions per child). Control was standardized using a scoring system and quantified 3 times during each examination. Overall control for a day was calculated as the mean of all 12 measures. Single measures of control and the mean of 2 (double) and 3 (triple) measures over the examination were compared with the respective day mean, and first-day measures were compared to the second-day mean.
Results
At distance, 17% (49/287, 95% confidence interval [CI] 13% to 22%) of single measures differed from the day mean, whereas only 8% (16/191, 95 CI 5% to 13%) of double measures and 5% (5/95, 95% CI 2% to 12%) of triple measures differed. Comparing day 1 measures to overall mean for day 2, 17% (24/143, 95% CI 11% to 24%) of single measures and 17% (22/130, 95% CI 11% to 24%) of double measures differed by more than 1 level, whereas 11% (5/47, 95% CI 4% to 23%) of triple measures differed.
Conclusions
The mean of 3 assessments of control during a clinic examination better represents overall control than a single measure.
In intermittent exotropia (XT), severity is often judged by assessing a patient’s ability to control the underlying exodeviation. Assessment of control is most often based on observations of spontaneous motor fusion, or lack of spontaneous motor fusion, either by the clinician or by the parent. Some authors have stated that “poor control” should be an indication for surgery, but “poor control” was not well defined. Recently, attempts have been made to quantify control using control scales. Nevertheless, these control scores rely on a single assessment during the clinical examination and in a previous study we found considerable variability when comparing isolated measures of control, including change from spontaneous tropia to phoria (and vice versa). In our previous study, changes in control occurred both over the course of 1 day (approximately 2 hours between assessments) and from one minute to the next when no intervening testing or dissociation had occurred. Such variability makes it difficult to reliably distinguish between a patient with consistently poor control and a patient whose control assessment happens to coincide with a rare episode of poor control. In a previous report, we postulated that multiple measures of control might better represent overall control for an individual patient. In the present study, we compared 1, 2, and 3 assessments of control during a single office examination session to an overall assessment of control during the day.
Methods
Patients
Parents/legal guardians of eligible children were contacted about the study and 12 children (median age 7, range 6 to 13 years) were prospectively enrolled. Children were considered eligible if they had basic, pseudo, or true types of intermittent XT; convergence insufficiency type intermittent exotropia (near angle more than 10 prism diopters [pd] greater than distance), sensory exotropia, paralytic exotropia, or coexisting developmental delay were excluded. Due to the testing burden, the minimum age for inclusion was set at 5 years old. The median visual acuity was 20/20, range 20/15 to 20/30 in the right eye and median 20/22, range 20/20 to 20/30 in the left eye. No patient had any interocular visual acuity difference of more than 2 logMAR lines. Nine (75%) were female, consistent with previously reported female preponderance in intermittent XT, and race was reported as “white” for 10 (83%). The median angle of deviation by prism and alternate cover test (PACT) was 20 (range 14 to 30) pd at distance (3 meters) and 12 (range 6 to 25) pd at near.
Control Assessment
Assessment of control was standardized using a previously described 0- to 5-point control scale scored separately for distance fixation (3 meters) and then for near fixation (1/3 meter). Using this control scale, any spontaneous exotropia was graded during a 30-second period of observation (constant exotropia = score 5, exotropia more than 50% of the time = score 4, exotropia less than 50% of the time = score 3). In the absence of any spontaneous exotropia during the 30-second observation period, exodeviation recovery was rated as the worst of 3 10-second dissociations (more than 5 seconds to recovery = score 2, 1–5 seconds to recovery = score 1, less than 1 second to recovery = score 0).
Examinations
Children were evaluated at 4 separate examinations over the course of a day, with at least 2 hours between each examination. Assessments were repeated in the same manner on a second day (median 21 days, range 7 to 41 days later), yielding a total of 8 examinations per child. It was not possible to standardize the amount of time between first and second days because families often had many other conflicting commitments. All examination procedures were performed by a certified orthoptist. At each examination, testing was performed according to a predetermined testing protocol. Control was assessed 3 times during each examination—at the start of the examination, after the PACT, and at the end of the examination—yielding a total of 12 control assessments over the day. The median duration of each examination was 21 minutes (range, 10 to 55 minutes).
Analysis
For the purposes of this study, an overall control score for the day was calculated for each patient as the mean of all 12 control scores (day mean). Therefore, over the 2 days there were a total of 24 control scores per child and 2 day mean scores per child. One child was unable to complete 1 assessment of control at 1 examination because of waning cooperation and therefore the total number of control assessments over the 2 days was 287 out of a potential 288. Single control scores (n = 287) were compared to the respective day mean, calculating proportions differing by more than 1 level. Double control scores were calculated as: 1) the mean of first and second control scores during a single examination session and 2) the mean of first and third control scores during a single examination session. These combinations for double scores were used in order to reflect clinical scenarios where control might be measured at the beginning and end of the examination or at the beginning and middle of the examination. Triple control scores were calculated as the mean of the 3 control assessments across a single examination session. Double control scores (2 per examination; n = 191) and triple control scores (1 per examination; n = 95) were then compared to the respective day mean, calculating proportions differing by more than 1 level. Distance and near control were analyzed separately.
In order to evaluate whether single, double, or triple control scores at 1 examination session represent an individual patient, we compared single (n = 143), double (n = 130), and triple (n = 47) scores on day 1 to the overall mean (of all 12 control scores) for day 2, calculating proportions differing by more than 1 level. Day 1 and day 2 mean control scores (mean of 12 measures across the day) were compared using intraclass correlation coefficients (ICC).
Results
Variability Over the Day and Over the Examination
Consistent with our previous study of variability of single control scores, 11 of 12 patients (92%) showed a change (>1 level) over the day at either distance or near on either day 1 or day 2. Seven of 12 patients (58%) changed at distance or near on both day 1 and day 2. Consistent with our previous report of variability over an examination, 8 of 12 patients (67%) showed a change in single measures of control on at least 1 examination at distance and 6 of 12 (50%) at near.
Single, Double, and Triple Measures of Control at Distance
For distance, 17% (49 of 287, 95% confidence interval [CI] 13% to 22%) of single measures and 8% (16 of 191, 95 CI 5% to 13%) of double measures differed by more than 1 level from the respective day mean ( Table 1 ), whereas only 5% (5 of 95, 95% CI 2% to 12%) of triple measures differed ( Table 1 ).
| Single Control Scores | Mean of 2 (Double) Control Scores | Mean of 3 (Triple) Control Scores | |
|---|---|---|---|
| Distance | 17% (49/287, 95% CI 13% to 22%) | 8% (16/191, 95 CI 5% to 13%) | 5% (5/95, 95% CI 2% to 12%) |
| Near | 12% (34/287, 95% CI 8% to 16%) | 4% (8/191, 95% CI 2% to 8%) | 1% (1/95, 95% CI 0% to 6%) |
a Table shows the proportion of single, double, and triple scores differing more than 1 level from the respective day score.
Single, Double, and Triple Measures of Control at Near
For near, 12% (34 of 287, 95% CI 8% to 16%) of single measures and 4% (8 of 191, 95% CI 2% to 8%) of double measures differed by more than 1 level from the respective day mean ( Table 1 ), whereas only 1% (1 of 95, 95% CI 0% to 6%) of triple measures differed ( Table 1 ).
Correlation of Mean Control Over Day 1 and Day 2
At distance, day 1 and day 2 means were highly correlated (ICC = 0.88, 95% CI 0.64 to 0.96, Figure , Top). At near, day 1 and day 2 means were moderately correlated (ICC = 0.58, 95% CI 0.06 to 0.85, Figure , Bottom).
