Differentiating Acute and Subacute Vertical Strabismus Using Different Head Positions During the Upright-Supine Test

Key Points

Question  Is acute or subacute vertical strabismus influenced by head position during the upright-supine test?

Findings  In this study of 19 patients with skew deviation and 18 patients with fourth nerve palsy, only 1 patient with skew deviation (5%) and no patients with fourth nerve palsy showed more than a 50% decrease in vertical misalignment in the supine position relative to the upright position with the head centered. On head tilt toward the hypotropic eye, 10 patients with fourth nerve palsy (56%) and no patients with skew deviation showed more than a 50% increase in vertical misalignment while supine.

Meaning  The upright-supine test is not a sensitive method to separate acute or subacute skew deviation from fourth nerve palsy.

Importance  Accurate clinical differentiation between skew deviation and fourth nerve palsy (4NP) is critical in the acute and subacute settings.

Objective  To determine the sensitivity and specificity of the upright-supine test to detect vertical misalignment changes using different head positions for the diagnosis of acute or subacute skew deviation vs 4NP.

Design, Setting, and Participants  This multicenter study enrolled consecutive patients from Coimbra University Hospital Centre, Coimbra, Portugal, and Michigan State University, Lansing, within 2 months of presenting with vertical diplopia and diagnosed as having skew deviation or acquired unilateral 4NP. The study used nonmasked screening and diagnostic test results from June 1, 2013, to December 31, 2016. Data were analyzed from January 1, 2017, to June 30, 2017.

Main Outcomes and Measures  A 50% or greater change in vertical misalignment between the upright and supine positions, with the head centered and tilted to either side. Measurements included the alternate prism and cover (APC) test, the double Maddox rod test, the APC test change index ([measurement upright − measurement supine] / [measurement upright + measurement supine]), and the APC test sensitivity and specificity.

Results  Of the 37 included patients, the mean (SD) age was 58 (14) years, and 26 (70%) were male. We enrolled 19 patients (51%) with skew deviation and 18 (49%) with 4NP. Eighteen patients with skew deviation (95%) showed additional ocular motor and/or neurological signs. When moving to the supine position, only 1 patient with skew deviation (5%) showed more than a 50% decrease of hypertropia with the head centered (APC test: sensitivity, 5%; specificity, 100%). Three patients with 4NP (17%) showed more than a 50% decrease of hypertropia with the head tilted toward the hypertropic eye, and 10 patients with 4NP (56%) showed more than a 50% increase of hypertropia with the head tilted toward the hypotropic eye. Change indexes were different between the skew deviation and 4NP groups for head tilt to the hypotropic eye (difference, −0.33 prism diopters; 95% CI, −0.43 to −0.20; P < .001). Cyclotorsion worsened in the supine position only in patients with skew deviation (hypertropic eye: difference, −7.6 prism diopters; 95% CI, −13.00 to −0.75; P = .01; hypotropic eye: difference, 8.2 prism diopters; 95% CI, 0 to 15.75; P = .03).

Conclusions and Relevance  The upright-supine test with the head centered is not a sensitive method to separate acute or subacute skew deviation from 4NP. Conversion of an incomitant vertical deviation in the upright position to a comitant vertical strabismus in the supine position in all head positions, as well as the absence of additional ocular motor and/or neurologic signs, may constitute a more useful clue.

Vertical strabismus is a common symptom in the neuro-ophthalmology practice and may result from several conditions, including skew deviation and acquired fourth nerve palsy (4NP).¹ Skew deviation reflects damage to supranuclear vestibular pathways and may constitute the initial manifestation of a life-threatening disorder, such as brainstem stroke.² On the other hand, acquired isolated 4NP is often caused by presumed microvascular ischemia and trauma and rarely represents a neurological emergency.³ While skew deviation is commonly accompanied by additional ocular motor (eg, nystagmus) and neurological (eg, ataxia) signs that aid in diagnosis, it may also occur in isolation.^4-6 Therefore, accurate clinical differentiation between the strabismic pattern of 4NP and skew deviation becomes critical, particularly in the acute setting.

Several tests have been devised for this purpose. Evaluation of strabismus comitancy is useful.^7,8 Generally, isolated unilateral 4NP produces an incomitant hypertropia that worsens in contralateral and downward gaze, while skew deviation is classically comitant; however, exceptions often occur.^8,9 A positive Parks-Bielschowsky 3-step test with increased hypertropia in contralateral gaze and ipsilateral head tilt (relative to the hypertropic eye) is the standard for diagnosing 4NP but shows limited sensitivity (70%) and may have only about 50% specificity.^9,10 Thus, occasionally, patients with skew deviation also show a positive Parks-Bielschowsky test.^7,8 Objective cyclotorsion measurement of the hypertropic eye may further help differentiate between 4NP (excyclotorted) and skew deviation (incyclotorted).^7,11,12

In 2011, the upright-supine test was proposed by Wong et al¹³ as a fourth step to differentiate skew deviation from 4NP based on the degree of hypertropia in the upright and supine positions with the head centered. Hypertropia improved at least 50% in the supine position only in patients with skew deviation (sensitivity, approximately 80%; specificity, 100%).^6,13 According to the authors, a reduction in the overall activity of the utriculo-ocular vestibular pathways while supine may have minimized the imbalance caused by the asymmetric damage of these pathways in patients with skew deviation, which in turn led to a reduction of vertical strabismus.¹³ Importantly, all patients with skew deviation had chronic strabismus in the preliminary study⁶ (n = 10; mean [minimum] duration of symptoms, 48 [15] months) and in a subsequent extension of this study¹³ (n = 25). Therefore, to our knowledge, data on the upright-supine test during acute or subacute presentation of vertical strabismus are not available, a setting where the information derived from this test would be most useful.

In this study, we aimed to evaluate the sensitivity and specificity of the upright-supine test in diagnosing skew deviation vs 4NP in the acute and subacute settings and to determine whether adding head tilt positions in the supine position improves diagnostic accuracy. Because neck movements seem to influence the vestibular tone and may indeed substitute for loss of vestibular information after vestibular disease, we hypothesized that in patients with skew deviation (but not 4NP), an enhanced cervico-ocular reflex could influence the magnitude of vertical misalignment during head tilts in the supine position.^14,15

This prospective, observational study was conducted at Michigan State University, Lansing, and Coimbra University Hospital Centre, Coimbra, Portugal, between June 1, 2013, and December 31, 2016. Written informed consent was obtained from all participants, according to a protocol conforming to the tenets of the Declaration of Helsinki and approved by the Michigan State University and Coimbra University Hospital Centre institutional review boards.

We consecutively examined 19 patients with skew deviation (12 [63%] male) and 18 patients with unilateral 4NP (14 [78%] male) (χ²₁ = 0.95; P = .48) less than 2 months from diplopia onset (Table). Only 3 patients (8%) had diplopia for more than 1 month. The mean (SD; range) age of patients with skew deviation was 57 (17; 23-85) years and was 58 (12; 38-85) years for patients with 4NP (Mann-Whitney U = 167.0; P = .91). The mean (SD; range) interval from diplopia onset to testing was 9 (9; 1-30) days for patients with skew deviation and was 18 (18; 1-60) days for patients with 4NP (Mann-Whitney U = 134.0; P = .26). All patients underwent axial and sagittal T1-weighted and T2-weighted, fluid-attenuated inversion recovery, proton density, and diffusion-weighted brain magnetic resonance imaging (MRI) with gadolinium enhancement (slice thickness, 5 mm). The median (interquartile range) time from symptom onset to MRI was 14 (29) days and 30 (147) days for patients with skew deviation and 4NP, respectively (Mann-Whitney U = 68.0; P = .40).

Diagnosis of skew deviation was based on the following criteria: (1) comitant or incomitant hypertropia, with or without associated head tilt or subjective ocular torsion deviation; (2) presence of a causative lesion on MRI; and (3) absence of depression limitation in adduction. The presence of additional neurologic and/or ocular motor signs, although constituting supportive criteria, was not mandatory for skew deviation diagnosis because skew deviation may occur in isolation.^4-6 The cause of skew deviation was ischemic brainstem stroke in 13 patients (68%), multiple sclerosis in 5 patients (26%), and non-Hodgkin lymphoma in 1 patient (5%). Diagnosis of unilateral 4NP was based on the following criteria: (1) incomitant hypertropia that increased with adduction of the hypertropic eye and/or with head tilt toward the hypertropic eye, with or without associated head tilt or subjective ocular torsion deviation; (2) absence of a causative lesion on MRI; (3) limited depression in adduction; and (4) absence of any other neurologic and/or ocular motor signs. The cause of 4NP was presumed microvascular ischemia in 15 patients (83%) and head trauma in 3 patients (17%). A positive Parks-Bielchowsky test result fulfilling the 3-step criteria was not mandatory for diagnosing 4NP because in about 30% of cases, only 2 of 3 steps are positive.⁹ Exclusion criteria included (1) history of diplopia and/or strabismus since childhood, unclear time of diplopia onset, photographic evidence of longstanding head tilt, and/or increased vertical fusion amplitudes; (2) previous surgery for strabismus; (3) reversal of hypertropia in head tilt; (4) other causes of vertical strabismus; and (5) visual acuity worse than 20/50 OU.

All patients underwent a complete neuro-ophthalmic and neurologic examination performed by 1 of the 3 coauthors (J.L., A.S., or E.E.) before MRI scans. Vertical misalignment was measured with the alternate prism and cover (APC) test, and subjective ocular torsion was measured using the double Maddox rod test. Clinical and demographic characteristics of patients are summarized in the Table.

The APC test used an occluder placed alternately in front of each eye while the patient maintained fixation on an accommodative target (ie, optotype letter) located 0.3 m away. The magnitude of vertical strabismus was estimated by placing a prism of increasing strength in front of one eye.¹⁶ This was performed with all 3 head positions—head centered, tilted 30° to the right, and 30° to the left—both in the upright and supine positions.

The double Maddox rod test used trial frames in which red and white Maddox rods were placed over the right and left eyes, respectively. A transilluminator located 1 m from the patient was used to create a red and white horizontal line in front of each eye, respectively. If any tilting of the line(s) from the horizontal was noted, patients were asked to rotate the rod(s) until the line(s) became horizontal.¹⁶ Measurements of subjective cyclotorsion were repeated 2 times, and the results per participant were averaged. This test was performed both in the upright and supine positions.

Categorical data are presented as frequencies (percentages) and were compared using the χ² test. Ordinal or discrete variables are represented using mean values and compared using Mann-Whitney U test and Wilcoxon test. Changes in vertical strabismus between the upright and supine position were analyzed using a change index ([measurement upright − measurement supine] / [measurement upright + measurement supine]), as used in previous literature.⁶ Positive and negative change index values reflected improvement and worsening, respectively, of misalignment from the upright to the supine position (1 if complete; 0 if no change). Additionally, the sensitivities and specificities of the APC and Parks-Bielchowsky tests were calculated. A clinically relevant APC test was considered when there was a 50% or greater decrease or increase in the vertical misalignment from the upright to the supine position. Calculations were performed separately for the 3 head positions (head centered, tilted 30° to the right, and 30° to the left). All statistical analyses were performed using SPSS version 20 (IBM Corporation). All P values were 2-tailed, and significance was set at P < .05.

There was right hypertropia in 12 patients with skew deviation (63%) and 8 patients with 4NP (44%) (difference, 19%; 95% CI, 13-50; χ²₁ = 1.30; P = .33). Six patients with skew deviation (32%) showed an increase of vertical misalignment either with adduction of the hypertropic eye or with head tilt toward the hypertropic eye, and only 1 patient with skew deviation (5%) fulfilled the 3-step criteria for a positive Parks-Bielchowsky test result. In contrast, 16 patients with 4NP (89%) had a positive Parks-Bielchowsky test result (sensitivity, 88%; specificity, 95%). Eighteen patients with skew deviation (95%) showed additional ocular motor signs (eg, spontaneous and gaze-evoked nystagmus, internuclear ophthalmoplegia, and slow saccades) and/or neurological signs (eg, ataxia, hemiparesis, and facial palsy) (Table).

In the skew deviation group, the mean (SD) vertical misalignment in the upright and supine positions was 8.0 (6.3) prism diopters (PD) and 7.6 (6.5) PD, respectively (difference, 0.4; 95% CI, −0.50 to 1.00; z = −0.37; P = .71; change index, 0.02; SD, 0.27). When moving to the supine position, only 1 patient with skew deviation (5%) showed a decrease of 50% or greater in hypertropia, and 2 patients showed an increase of 50% or greater. In the 4NP group, the mean (SD) vertical misalignment in the upright and supine positions was 5.3 (2.9) PD and 5.1 (2.0) PD, respectively (difference, 0.2; 95% CI, −0.50 to 1.00; z = −0.64; P = .52; change index, 0.01; SD, 0.12). When moving to the supine position, no patient with 4NP showed a decrease or increase of 50% or greater in hypertropia (APC test for skew deviation diagnosis with head centered: sensitivity, 5%; specificity, 100%) (Figure 1). The change indexes were not different between the skew deviation and 4NP groups (difference, 0.01; 95% CI, −0.05 to 0.06%; Mann-Whitney U = 165.5; P = .87).

In the skew deviation group, the mean (SD) vertical misalignment in the upright and supine positions was 8.3 (6.1) PD and 7.3 (6.7) PD, respectively (difference, 0.6; 95% CI, 0 to 2.00; z = −1.55; P = .12; change index, 0.09; SD, 0.28). When moving to the supine position, only 2 patients with skew deviation demonstrated a decrease of 50% or more in hypertropia, and no patient with skew deviation showed an increase of 50% or more. In the 4NP group, the mean (SD) vertical misalignment in the upright and supine positions was 7.8 (5.3) PD and 5.5 (2.2) PD, respectively (difference, 2.3; 95% CI, 0.50 to 3.50; z = −2.26; P = .02; change index, 0.12; SD, 0.19). When moving to the supine position, 3 patients with 4NP demonstrated a decrease of 50% or more in hypertropia (APC test for skew deviation diagnosis with head tilted toward the hypertropic eye: sensitivity, 10%; specificity, 82%), and no patient with 4NP showed an increase of 50% or greater (Figure 2). The change indexes did not differ between the skew deviation and 4NP groups (difference, 0; 95% CI, −0.05 to 0.06; Mann-Whitney U = 126.5; P = .17).

In the skew deviation group, the mean (SD) vertical misalignment in the upright and supine positions was 7.0 (6.4) PD and 7.2 (6.7) PD, respectively (difference, −0.2; 95% CI, −0.50 to 0; z = −0.37; P = .55; change index, −0.01; SD, 0.09). When moving to the supine position, no patient with skew deviation demonstrated a decrease or increase of 50% or more in hypertropia. In the 4NP group, the mean (SD) vertical misalignment in the upright and supine positions was 2.8 (1.6) PD and 5.2 (2.2) PD, respectively (difference, −2.4; 95% CI, −3.50 to −1.00; z = −3.35; P < .001; change index, −0.32; SD, 0.26). When moving to the supine position, a substantial number of patients with 4NP (n = 10 [55%]) demonstrated an increase of 50% or more in hypertropia (APC test for 4NP diagnosis with head tilted toward the hypotropic eye: sensitivity, 55%; specificity, 100%), and no patient with 4NP showed a decrease of 50% or more (Figure 3). The change index was higher in the 4NP group (difference, −0.33; 95% CI, −0.43 to −0.20; Mann-Whitney U = 45.5; P < .001).

From the upright to supine position, incyclotorsion in the hypertropic eye and excyclotorsion in the hypotropic eye increased in patients with skew deviation (hypertropic eye: mean [SD] deviation: upright, −7.4 [7.8] PD; supine, −15.0 [13.9] PD; difference, −7.6; 95% CI, −13.00 to −0.75; z = −2.37; P = .01; hypotropic eye: mean [SD] deviation: upright, 8.0 [10.2] PD; supine, 16.2 [13.9] PD; difference, 8.2; 95% CI, 0 to 15.75; z = −2.07; P = .03). In patients with 4NP, excyclotorsion in the hypertropic and hypotropic eye showed no changes (hypertropic eye: mean [SD] deviation: upright, 3.3 [3.5] PD; supine, 1.8 [7.1] PD; difference, 1.5; 95% CI, −3.50 to 2.50; z = −0.36; P = .72; hypotropic eye: mean [SD] deviation: upright, 1.5 [3.3] PD; supine, 1.6 [7.8] PD; difference, 0.1; 95% CI, −2.50 to 1.00; z = −0.95; P = .34) (Figure 4).

In this study, we investigated the clinical utility of using different head positions to differentiate 2 common causes of vertical strabismus, ie, skew deviation and 4NP, in the acute and subacute settings. We found that vertical misalignment did not change between the upright and supine positions in either the skew deviation or 4NP groups when the head was centered. Only in patients with 4NP did we document a change in hypertropia from the upright to supine position when the head was tilted to either side (more so with tilt to the hypotropic side). Subjective cyclotorsion worsened in the supine position only in the skew deviation group, although cyclotorsion changes were highly variable in both groups.

Our data are in contrast to previous work showing improvement of vertical strabismus in the supine position in patients with skew deviation, an effect postulated to result from position-induced reduction in the activity of the utriculo-ocular vestibular pathways.^6,13 One plausible explanation for this discrepancy is the symptom duration in our study. Most (91%) of these patients were studied during the acute period (ie, less than 1 month), while in the series by Parulekar et al,⁶ the median duration of symptoms at the time of the study was 4 years. It is possible that several adaptive mechanisms take place over time in patients with asymmetric otovestibular disease, such as skew deviation. As an example, acute asymmetric vestibular disease presenting with spontaneous nystagmus may demonstrate a change in nystagmus pattern and direction over time, suggesting a shift in the vestibular imbalances from initial disruption into late disinhibition¹⁷; similarly, central vestibular dysfunction may add features over time, like in progressive ataxia and oculopalatal tremor.¹⁸ Thus, we hypothesize that following the acute phase of skew deviation, an initially large vestibular imbalance gradually settles down into a more predictable moderate-sized imbalance in the chronic phase that allows for adaptive mechanisms to take place, ultimately contributing to positional modulation of hypertropia.^6,13

In patients with 4NP in the upright position, both the increase and decrease of hypertropia precipitated by ipsilateral and contralateral head tilt, respectively, were partially mitigated when performing head tilt in the supine position. Sydnor et al¹⁹ found equivalent results. In patients with 4NP, changes in hypertropia magnitude precipitated by head tilt, classically measured in the upright position, have been attributed to the effect of gravity’s pull on utriculo-ocular pathways. Specifically, in a normal individual, head tilt increases the output of the ipsilateral utricle to the ipsilateral superior rectus and inferior oblique muscles as well as to the contralateral superior oblique and inferior rectus muscles, promoting a small counterroll (torsional) movement of the eyes but almost no vertical eye movement (ocular counterrolling reflex).⁵ However, in patients with 4NP, head tilt to the paretic side stimulates counterroll but is forced to rely on the intact superior rectus, increasing the preexisting hypertropia, a component of the Parks-Bielschowsky test.¹⁶ We share the opinion of Sydnor et al¹⁹ that in the supine position, gravity’s influence on head tilt is probably mitigated, and thus the effect of head tilt on hypertropia magnitude seen in the upright position is cancelled in the supine position.¹⁹ Of note, although upright vs supine changes in hypertropia were noted in patients with 4NP during head tilt positions, the actual average magnitude of the change was minimal (approximately 2 PD) and may be difficult to detect in the average patient.

Subjective cyclotorsion worsened in patients with acute and subacute skew deviation in the supine position. Interestingly, in patients with chronic skew deviation, cyclotorsion improved in the supine position, which again raises the possibility of central adaptive mechanisms taking place over time.^6,13 One additional nonmutually exclusive reason for cyclotorsion worsening in the supine position regards previous data showing that patients with acute (but not chronic) skew deviation show an unstable torsional eye angle and position, which in a normal condition is directly derived from horizontal and vertical eye positions (Listing law).²⁰ Of note, changes in subjective cyclotorsion were highly variable in both groups, and thus caution must be taken when interpreting these results. The double Maddox rod test is a subjective test, and patients with an acute neurologic insult may find it difficult to perform.

Our work has limitations, including the limited number of participants, the lack of a healthy control group, and data derived from 2 geographically distant clinical sites. These may limit the generalization of our results to a wider-ranging vertical strabismus group.

In our acute and subacute cohorts, vertical misalignment and ocular torsion in the supine position with the head centered (upright-supine test) was not useful for distinguishing skew deviation from 4NP. Conversion of an incomitant vertical deviation in the upright position to a comitant vertical strabismus in the supine position in all head positions, as well as the absence of neurologic signs, may constitute a more useful clue to differentiate 4NP from acute and subacute skew deviation.

Accepted for Publication: December 21, 2017.

Corresponding Author: João Lemos, MD, PhD, Neurology Department, Coimbra University Hospital Centre, Praceta Mota Pinto, Coimbra 3000-075, Portugal (merrin72@hotmail.com).

Published Online: February 15, 2018. doi:10.1001/jamaophthalmol.2017.6796

Author Contributions: Drs Lemos and Eggenberger had full access to all of the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

Study concept and design: Lemos, Subei, Cunha, Glisson, Eggenberger.

Acquisition, analysis, or interpretation of data: Lemos, Subei, Sousa, Nunes, Eggenberger.

Drafting of the manuscript: Lemos.

Critical revision of the manuscript for important intellectual content: All authors.

Statistical analysis: Lemos, Sousa.

Administrative, technical, or material support: Lemos, Sousa.

Study supervision: Lemos, Eggenberger.

Conflict of Interest Disclosures: All authors have completed and submitted the ICMJE Form for Disclosure of Potential Conflicts of Interest and none were reported.