Post‐concussion clinical findings of oculomotor function in paediatric patients with persisting symptoms compared to healthy controls

Abstract

Objective

Oculomotor deficits in vergence and accommodation can arise in paediatric patients with persistent concussion symptoms, although the profile is not well established. This study aimed to describe the frequency of these deficits in persistently symptomatic concussed paediatric patients and identify effective screening tools.

Methods

This was a prospective cohort study conducted at three clinical sites across the United States. Participants aged 8–18 years with diagnosed concussion were recruited within 9 months of injury through concussion clinics or referral to a vision provider. Participants without concussion were recruited through the local community and eye clinics. Clinical measures of ocular alignment, vergence and accommodation were collected. Group comparisons were assessed using Welch's t‐test, Mann–Whitney U test and Fisher's exact test with Bonferroni correction. The diagnostic value of near point of convergence (NPC) and accommodative amplitude (AA) for identifying persistently symptomatic concussed participants was evaluated using logistic regression and receiver operating characteristic curve analysis.

Results

Seventy‐one participants were recruited, including 34 concussed participants (mean age 14.3 [SD 2.4] years; 74% female, 26% male; median time since concussion 107 [IQR 80–118] days) and 32 controls (mean age, 12.7 [SD 2.1] years; 56% female, 44% male). Concussed participants scored significantly worse or had higher failure rates than controls on all vergence and accommodative tests (p < 0.05) except ocular alignment and monocular accommodative facility. Concussed participants had a higher frequency of diagnoses (vergence: 62% vs. 3%; accommodation: 76% vs. 3%; p < 0.001). NPC and AA were significant predictors for concussion in individual models (NPC: OR = 2.16 [95% CI: 1.52–3.61], p < 0.001, mean AUC [SD] = 0.88 [0.13]; AA: OR = 0.46 [95% CI: 0.29–0.64], p < 0.001, mean AUC [SD] = 0.88 [0.15]).

Conclusion

The oculomotor profile of persistently symptomatic concussed paediatric participants shows a high frequency of vergence and accommodative deficits, for which NPC and AA are effective screening tools. Further investigation should examine oculomotor deficits in acutely concussed paediatric patients.

Key points.

• In this prospective study, a significant presence of vergence and accommodative deficits was found in concussed paediatric patients, and effective post‐concussion vision screening tools were identified.

• Vergence and accommodative deficits occurred in over 60% of concussed paediatric patients compared to non‐concussed controls. The near point of convergence and push‐up amplitude of accommodation tests were found to be effective predictors for identifying concussed patients with persistent symptoms.

• Near point of convergence and accommodative amplitude may be effective screening tools for assessing oculomotor function post‐concussion.

INTRODUCTION

Concussions are prevalent in the paediatric population, with lifetime estimates ranging from 6.5% to 18.3% in those 13–17 years of age in the United States. ¹ Concussion, or mild traumatic brain injury, results from mechanical force transmitted to the brain, causing functional but non‐structural neurological disturbance. ² , ³ Oculomotor dysfunction is one of the five identified concussion subtypes, ⁴ , ⁵ with patients reporting symptoms such as light sensitivity and double or blurry vision ⁶ , ⁷ , ⁸ , ⁹ , ¹⁰ , ¹¹ , ¹² , ¹³ , ¹⁴ , ¹⁵ which are often associated with deficits in vergence, accommodation and saccades. ⁹ , ¹⁰ , ¹¹ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ Receded near point of convergence (NPC), abnormal accommodative amplitude (AA) and oculomotor dysfunction correlate significantly with higher symptom reporting, prolonged recovery and persistent concussive symptoms. ⁶ , ⁹ , ²² , ²³ , ²⁴ Although the exact aetiology remains unclear, these deficits likely stem from disruptions in afferent, efferent and central visual pathways. ⁸ , ²⁵ , ²⁶ , ²⁷ Given the high level of cognitive demands in school and daily life for the adolescent population, it is imperative to thoroughly characterise these oculomotor deficits in concussed paediatric patients to inform treatment plans and facilitate recovery. ⁸ , ²⁸ , ²⁹

Previous studies have reported a high frequency of vergence and accommodative deficits in concussed adolescents. However, these studies have been either retrospective, thereby limiting comparability due to methodological differences, or prospective but lacking a control group. ⁹ , ¹⁰ , ¹¹ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ³⁰ Here, a prospective study was conducted, comparing symptomatic concussed patients with age‐matched controls to characterise oculomotor deficits and assess the effectiveness of NPC and AA, two of the most common optometric tests used as screening tools by non‐eye care providers. We hypothesised that these two groups would have distinct oculomotor profiles.

METHODS

Institutional review board approval

This study was approved by Boston Children's Hospital, Stanford University and Salus University Institutional Review Boards. Written parental permission was obtained from the participants' parent or guardian, and written assent was obtained from all participants prior to the start of the study.

Patient recruitment and eligibility

Participants were recruited from Boston Children's Hospital, Stanford University and Salus University's Pennsylvania College of Optometry. Control participants were recruited from all three sites through word of mouth, children of department staff, local community and patients who came to the eye care clinics for routine eye examinations. Concussed participants were recruited from concussion clinics at Boston Children's Hospital and Stanford University, or referred to our two senior authors' (TLR and AR) vision care practices for post‐concussion evaluation. Those recruited from vision care practices were referred due to having overall persistent concussion symptoms, with the purpose of ruling out vision involvement, not necessarily because they were exhibiting specific vision symptoms.

Concussed participants were eligible if they were diagnosed with a concussion by their treating physician in accordance with the Berlin Consensus Statement on Concussion in Sport, ² and if the duration from concussion to the time of the study visit was between 4 weeks and 9 months. ³¹ Control participants were eligible if they had no prior history of concussion or known history of vergence or accommodation deficits. Participants ranged in age from 8 to 18 years, had a best‐corrected distance visual acuity of 0.10 logMAR or better in each eye, and wore appropriate refractive correction as follows:

Participants were required to wear correction if their cycloplegic refraction revealed ≥2.00 dioptres (D) of hyperopia, ≥1.00 D of myopia, ≥1.00 D of anisometropia or ≥1.25 D of astigmatism. If required to wear a refractive correction, this had to be as follows: hyperopic sphere power symmetrically reduced by no more than 1.50 D spherical equivalent (SE), myopia and SE anisometropia within 1.00 D of full correction, astigmatism within 1.00 D of full correction and axis within 10 degrees for magnitudes of ≥1.00 D.

Participants were excluded for any history of amblyopia, strabismus, reports of constant diplopia, in‐office vision therapy or ocular injury that could affect vision or the oculomotor system, abnormality of the cornea, lens or central retina, constant or intermittent esotropia at distance or near, constant or intermittent exotropia at distance, constant exotropia at near, vertical heterophoria ≥2 prism dioptres (Δ) at distance or near, manifest or latent nystagmus, or any disease known to affect vergence, accommodation or ocular motility. Participants were excluded if they were unable to perform study‐related clinical tests.

Testing procedures

Participants underwent a visual function examination by a paediatric optometrist, which included measures of visual acuity at distance and near, stereopsis, ocular motility, ocular alignment, vergence and accommodation. Ocular alignment was measured using the cover and uncover test, and magnitudes were assessed using the prism and alternate cover test. Vergence tests included NPC, near vergence facility and near convergence and divergence fusional amplitudes. NPC was measured using a Near Point Rule (Gulden Ophthalmics, guldenophthalmics.com) from the forehead using a 0.18 logMAR vertical column of letters. The letters were moved toward the participant until they reported diplopia (break value), then slowly moved away from the participant until they reported single vision (recovery value). Near vergence facility was measured at 40 cm using a hand‐held prism (Vergence Facility Stick Prism, Gulden Ophthalmics, guldenophthalmics.com) with 3 Δ base‐in or 12 Δ base‐out prisms placed in front of the right eye while the participant viewed a 0.18 logMAR vertical column of letters positioned at the midline. Participants were asked to report when the letters appeared clear and single for each prism value. The number of cycles (1 cycle represents clearing both base‐in and base‐out prisms) in 1 min was recorded. Convergence and divergence fusional amplitudes were measured using a horizontal prism bar with the following prism increments: 1 Δ, 2 Δ to 20 Δ in 2 Δ steps and 20 Δ to 45 Δ in 5 Δ steps (EZ View Horizontal Prism Bar, Gulden Ophthalmics, guldenophthalmics.com) at near (40 cm) when viewing a 0.18 logMAR vertical column of letters. The prism was introduced one step at a time, and participants were asked to report when the target appeared blurry (blur value) or double (break value). The prism then was reduced one step at a time until the participant reported single vision (recovery value). Measurements were recorded as the blur, break and recovery values.

Accommodation tests included monocular AA and monocular accommodative facility. AA was measured with the Near Point Rule using the push‐up method with a 0.18 logMAR vertical column of letters. The letters were moved toward the participant until the participant noted sustained blur. Monocular accommodative facility was measured using a ±2.00 D flipper lens with a 0.18 logMAR vertical column of letters (±2 D Flipper, Gulden Ophthalmics, guldenophthalmics.com). Participants viewed the letters alternately through the plus and then the minus lenses. Participants were asked to report when the letters appeared clear for each lens. The number of cycles in 1 min was recorded.

Greater difficulty for either base‐out (convergence demand) or base‐in prisms (divergence demand) for vergence facility and minus (increased accommodative demand) or plus lens (decreased accommodative demand) for accommodative facility was assessed by the tester and validated by the participant when necessary.

Clinical cutoffs and diagnostic criteria

Clinical diagnoses and deficits were determined using standard diagnosis criteria cutoffs (Tables 1 and 2). ¹⁰ , ¹⁸ , ³² , ³³ , ³⁴ , ³⁵ , ³⁶ For monocular tests such as AA and monocular accommodative facility, values from the worse eye were used. For fusional ranges, the blur value was used. If blur was not reported by the participant, the break value was used.

TABLE 1.

Test failure criteria for clinical tests at near.

Clinical tests at near	Failure criteria
Esophoria	≥3 Δ of esophoria at near
Exophoria	Near exophoria ≥4 Δ than distance
Near point of convergence	>6 cm
Divergence fusional amplitudes	Blur/break a <8 Δ or if Sheard's criterion is not met
Sheard's criterion	Blur/break a <2 × near esophoria
Convergence fusional amplitudes	Blur/break a ≤15 Δ or if Sheard's criterion is not met
Sheard's criterion	Blur/break a <2 × near exophoria
Vergence facility	≤9 cycles per minute
Amplitude of accommodation	<11 D
Monocular accommodative facility	≤6 cycles per minute

^a Blur/break refers to the point at which the participant reports that the target is blurry, or if there is no blur, when it splits into two. Δ, prism dioptre.

TABLE 2.

Vergence and accommodation diagnosis criteria.

Clinical diagnosis and findings	Criteria
Vergence diagnoses
Convergence insufficiency	First criteria and one other must be met
Near deviation	≥4 Δ near exophoria more than distance
Near point of convergence	>6 cm
Convergence fusional amplitudes at near	≤15 Δ blur/break a or failing Sheard's criterion b
Vergence facility (3ΔBI/12ΔBO)	≤9 cpm, BO prism harder
Convergence deficit	First criteria and one other must be met
Near point of convergence	>6 cm
Convergence fusional amplitudes at near	≤15Δ blur/break a or failing Sheard's criterion b
Vergence facility (3ΔBI/12ΔBO)	≤9 cpm, BO prism harder
Convergence excess	First criteria and one other must be met
Near Deviation	≥3 Δ esophoria at near
Divergence fusional amplitudes at near	<8 Δ blur/break a or failing Sheard's criterion b
Vergence facility (3ΔBI/12ΔBO)	≤9 cpm, BI prism harder
Divergence deficit	Must meet both criteria
Divergence fusional amplitudes at near	<8 Δ blur/break a or failing Sheard's criterion b
Vergence facility (3ΔBI/12ΔBO)	≤9 cpm, BI prism harder
Fusional vergence dysfunction	Must meet first two criteria OR fail third
Convergence fusional amplitudes at near	≤15 Δ blur/break a or failing Sheard's criterion b
Divergence fusional amplitudes at near	<8 Δ blur/break a or failing Sheard's criterion b
Vergence facility (3ΔBI/12ΔBO)	≤9 cpm, BI and BO prisms both difficult
Accommodative diagnoses
Accommodative insufficiency	Must meet one criteria
Accommodative amplitude	<11 D
Accommodative facility (±2D)	≤6 cpm, (−) lens difficult
Accommodative excess	Must meet both criteria
Accommodative Amplitude	≥11 D (normal)
Accommodative Facility (±2D)	≤6 cpm, (+) lens difficult
Accommodative infacility	Must meet criteria
Accommodative facility (±2D)	≤6 cpm, (+) and (−) lens equally difficult
Accommodative dysfunction	Most meet both criteria
Accommodative Amplitude	<11 D
Accommodative Facility (±2D)	≤6 cpm, (+) lens difficult
Accommodative insufficiency and accommodative infacility	Must meet both criteria
Accommodative amplitude	<11 D
Accommodative facility (±2D)	≤6 cpm, (+) and (−) lens equally difficult

Abbreviations: BI, base‐in prism; BO, base‐out prism; cpm, cycles per minute; D, dioptre; Δ, prism dioptre.

^a Blur/break refers to the point at which the participant reports that the target is blurry, or if there is no blur, when it splits into two.

^b Sheard's Criterion: Compensating vergence range (positive or negative fusional vergence) of at least two times near heterophoria.

Statistical analysis

Descriptive statistics for deficits and diagnoses were reported as mean and 95% confidence intervals for continuous variables and as frequency for categorical variables. Normality for continuous variables was determined using the Shapiro–Wilk test with an alpha value of 0.05, and group differences were assessed using Welch's t‐test or Mann–Whitney U test as appropriate (Table S1). Fisher's Exact Test was used to assess group differences in categorical variables. Bonferroni adjustment was applied to account for multiple comparisons for clinical test measures, frequency of test failure and frequency of deficits separately. ³⁷

Logistic regression models were used to classify individuals with and without concussion using either NPC or AA as predictors. The combination of NPC and AA as predictors was also used in a multivariable logistic regression model since NPC and AA are common screening tests often utilised by non‐eye care providers. Receiver operating characteristic (ROC) curves were constructed for each model, and the area under the curve (AUC) was calculated to evaluate model performance and assess screening utility. ³⁸ , ³⁹ Ten‐fold cross‐validation, a resampling technique, was performed to assess the robustness of the AUC values and to reduce concerns of model overfitting. ⁴⁰ Optimal NPC and AA threshold values (i.e., cutoff points) were determined to classify individuals with and without concussion from the apparent ROC curves using Youden Index to maximise sensitivity and specificity. ⁴¹

All analyses were conducted using R (version 4.3.2, R Core Team 2023, r‐project.org) with an alpha value of 0.05.

RESULTS

In total, 37 concussed and 34 control participants were recruited, of which 34 concussed (25 [74%] female, 9 [26%] male; mean [SD] age 14.2 [2.4] years) and 32 controls (18 [56%] female, 14 [44%] male; mean [SD] age 12.7 [2.1] years) met the inclusion criteria. Participants were excluded due to reports of constant diplopia on the visual function examination (n = 1), ineligible distance visual acuity (n = 3) and presence of esotropia (n = 1). Within the concussed group, the median [IQR] time since concussion was 107 [80, 118] days. Twenty‐four (71%) concussed participants had sustained concussions through sports, one (3%) through a motor vehicle accident and nine (26%) through other mechanisms, such as falls.

Clinical tests of vergence and accommodation

As shown in Table 3, concussed participants had significantly worse mean measures than the controls on vergence tests, including NPC (9.3 vs. 3.9 cm; p < 0.001), convergence fusional amplitudes (14.9 vs. 24.1 Δ, p < 0.001), divergence fusional amplitudes (10.1 vs. 14.4 Δ, p < 0.001) and vergence facility (9.9 vs. 13.8 cycles/min, p = 0.004). The mean magnitude of exodeviation was not significantly different between the concussed and control participants at distance (0.6 vs. 0.4 Δ; p > 0.99) or near (2.2 vs. 2.1 Δ, p > 0.99).

TABLE 3.

Means, 95% confidence intervals and statistical values for tests of ocular alignment, accommodation and vergence.

Test	Concussed		Control		p‐value
	Mean	95% CI	Mean	95% CI
Distance exodeviation (Δ)	0.6	(0.2–1.0)	0.4	(0.1–0.7)	>0.99
Near exodeviation (Δ)	2.2	(0.8–3.5)	2.1	(1.1–3.1)	>0.99
Amplitude of accommodation (D)	9.2	(8.2–10.2)	12.9	(12.5–13.4)	<0.001
Monocular accommodative facility (cycles/min)	5.5	(3.7–7.3)	8.7	(6.6–10.7)	0.13
Near point of convergence (cm)	9.3	(7.7–10.8)	3.9	(3.4–4.5)	<0.001
Convergence fusional amplitudes (Δ)	14.9	(12.4–17.3)	24.1	(21.7–26.5)	<0.001
Divergence fusional amplitudes (Δ)	10.1	(9.0–11.1)	14.4	(13.4–15.4)	<0.001
Vergence facility (cycles/min)	9.9	(8.1–11.6)	13.8	(12.5–15.2)	0.004

Abbreviations: CI, confidence interval; D, dioptre; min, minute; Δ, prism dioptre.

For accommodation tests, concussed participants had reduced mean performance on AA (9.2 vs. 12.9 D, p < 0.001). There was no significant difference between the groups for monocular accommodative facility (5.5 vs. 8.7 cycles/min, p = 0.13).

Frequency of test failure based on normative values

Concussed participants failed vergence tests at significantly higher rates than control participants (Table 4) using the standard failure criteria cutoffs (Table 1). NPC showed the largest group difference in failure rates, with 22 (65%) concussed participants and just one (3%) control failing the test (p < 0.001). Similarly, 20 (59%) concussed and zero (0%) control participants failed convergence fusional amplitudes (p < 0.001), while 15 (44%) concussed and three (9%) control participants failed vergence facility (p = 0.03). However, group differences were not significant for those failing, with difficulty in convergence demand (12 [35%] concussed, two [6%] controls; p = 0.08), difficulty in divergence demand (two [6%] concussed, zero [0%] controls; p > 0.99), nor for equal difficulty with both demands (one [3%] concussed, one [3%] control; p > 0.99). Five (15%) concussed and zero (0%) controls failed divergence fusional amplitudes, but this difference between groups was not significant (p = 0.75; Table 4).

TABLE 4.

Participant counts for test failure in the concussed and control groups.

	No. (%)		p‐value
	Concussed (n = 34)	Control (n = 32)
Exophoria	8 (24)	9 (28)	>0.99
Esophoria	3 (9)	0 (0)	>0.99
Near point of convergence	22 (65)	1 (3)	<0.001
Divergence fusional amplitudes	5 (15)	0 (0)	0.75
Convergence fusional amplitudes	20 (59)	0 (0)	<0.001
Vergence facility	15 (44)	3 (9)	0.03
Difficulty with converging	12 (35)	2 (6)	0.08
Difficulty with diverging	2 (6)	0 (0)	>0.99
Equal difficulty	1 (3)	1 (3)	>0.99
Amplitude of accommodation	26 (76)	1 (3)	<0.001
Monocular accommodative facility	23 (68)	11 (34)	0.18
Difficulty with plus lens	20 (59)	11 (34)	0.75
Difficulty with minus lens	1 (3)	0 (0)	>0.99
Equal difficulty with both lenses	2 (6)	0 (0)	>0.99

Accommodative amplitude yielded significantly higher failure rates (Tables 1 and 4) in the concussed compared to the control group, with 26 (76%) concussed participants and one (3%) control failing AA (p < 0.001). However, monocular accommodative facility had similar failure rates between groups, with 23 (68%) concussed and 11 (34%) control participants failing monocular accommodative facility (p = 0.18). Both groups had a high failure rate with difficulty clearing the plus lens (relaxing accommodation), as 20 (59%) concussed and 11 (34%) control participants failed, exhibiting difficulty with the plus lens. Comparatively fewer participants across groups had greater difficulty clearing the minus lens (stimulating accommodation), or had equal difficulty with both lenses, with one (3%) concussed and zero (0%) control participants failing with minus difficulty and two (6%) concussed and zero (0%) control participants failing with equal difficulty. There were no significant differences between groups for plus (p = 0.75), minus (p > 0.99), or both lenses equally difficult (p > 0.99; Table 4).

Frequency of deficits

The concussed group had more vergence or accommodation diagnoses than the controls, using standard diagnosis cutoff criteria (Table 2). Thirty‐two (94%) concussed participants had at least one diagnosis, with 11 (32%) being diagnosed with only accommodative deficits and the remaining 21 (62%) diagnosed with both vergence and accommodative deficits. On the other hand, only 12 (37%) control participants had a diagnosis, with one (3%) having a vergence deficit and 11 (34%) having an accommodative deficit (Table 5, Figure 1a,b). Frequency of diagnosis was significantly higher in concussed participants compared to the controls for convergence deficit, accommodative insufficiency and accommodative dysfunction. Furthermore, significantly more concussed than control participants had a vergence (p < 0.001) or accommodative (p < 0.001) diagnosis (Table 5). Because of an unexpected high failure rate in controls, the proportion of accommodative deficits was investigated without using monocular accommodative facility as a diagnostic criterion. The number of those with accommodative deficits fell from 32 (94%) to 26 (76%) for concussed participants and from 11 (34%) to one (3%) for controls; all under the accommodative insufficiency subtype (Figure 1c,d).

TABLE 5.

Count and frequency of accommodation and vergence diagnoses in control and concussed cohorts.

Diagnosis	No. (%)		p‐value
	Concussed (n = 34)	Control (n = 32)
Vergence diagnosis	21 (62)	1 (3)	<0.001
Convergence insufficiency	7 (21)	0 (0)	0.14
Convergence deficit	12 (35)	0 (0)	0.002
Convergence excess	1 (3)	0 (0)	>0.99
Divergence deficit	0 (0)	0 (0)	>0.99
Fusional vergence dysfunction	1 (3)	1 (3)	>0.99
Accommodation diagnosis	32 (94)	11 (34)	<0.001
Accommodative insufficiency	10 (29)	0 (0)	0.01
Accommodative excess	6 (18)	10 (31)	>0.99
Accommodative dysfunction	14 (41)	1 (3)	0.003
Accommodative infacility	0 (0)	0 (0)	>0.99
Accommodative insufficiency and accommodative infacility	2 (6)	0 (0)	>0.99

FIGURE 1.

Proportional Venn Diagram of participants diagnosed with deficits in vergence (blue), accommodation (pink) or none (grey), using amplitude of accommodation (AA) and monocular accommodative facility as diagnostic criteria for (a) concussed and (b) control participants. Figures (c) and (d) show the deficits for concussed and control participants, respectively, when monocular accommodative facility was excluded from the diagnosis.

Clinical utility of NPC and AA as predictor variables of concussion

Continuous values for NPC and AA were used as individual and combined predictor variables to classify concussed and control participants. As individual models, both NPC (p < 0.001) and AA (p < 0.001) were significant predictors. In the combined model, NPC (p = 0.02) but not AA (p = 0.06) was a significant predictor (Table 6). Further, AUC measurements showed that NPC and AA, both individually and combined, had high and comparable diagnostic value in differentiating between concussed and control participants (Figure 2). Apparent and cross‐validated AUC values were 0.88 (both) for the NPC model, 0.89 and 0.88 for the AA model and 0.91 and 0.90 for the NPC + AA combined model. Optimal cutoff points using the Youden Index were 6.25 cm and 10.88 D for NPC and AA, respectively (Figure 2a). These cutoff values closely correspond to the diagnostic criteria of the present study (Table 1).

TABLE 6.

Logistic regression models with accommodative amplitude (AA) and near point of convergence (NPC) as predictor variables (shown in italics) individually and combined in a multivariable model, with the outcome variable being concussed versus control.

Variable	β	Standard error	Odds ratio	95% CI	p‐value
AA	−0.78	0.20	0.46	(0.29–0.64)	<0.001
NPC	0.77	0.22	2.16	(1.52–3.61)	<0.001
NPC + AA
AA	−0.39	0.20	0.68	(0.42–0.98)	0.06
NPC	0.58	0.25	1.79	(1.17–3.12)	0.02

Abbreviation: CI, confidence interval.

FIGURE 2.

ROC curves for logistic regression models, including (a) AA (pink), NPC (blue) and AA and NPC combined (green) models. Apparent area under the curve (AUC) with 95% CI, cutoff based on Youden's index and cross‐validated mean AUC and SD are reported. (b) Confusion matrices (AA, NPC, combined) show the generated predictions for each model. AA, amplitude of accommodation; CI, confidence intervals; NPC, near point of convergence; ROC, receiver operating characteristics; SD, standard deviation.

DISCUSSION

This study examined the vergence and accommodative profile of concussed paediatric participants. It is the first study to our knowledge that incorporates a control group within a prospective design, allowing comparison of a comprehensive profile of visual function performance between persistently symptomatic concussed and healthy paediatric patients. The diagnostic utility of common clinical tests of vergence and accommodation in identifying concussed patients who may benefit from a detailed visual function assessment was also examined.

Vergence

This study showed a significant presence of vergence deficits in concussed paediatric subjects, with 21 (62%) symptomatic concussed participants possessing a vergence diagnosis compared to only one (3%) control. Symptomatic concussed participants had a higher failure rate than controls with vergence tests including NPC, convergence fusional amplitudes and vergence facility. Notably, although 65% of symptomatic concussed participants failed NPC, only 56% had a convergence‐related diagnosis. These results align with our previous retrospective report that NPC failure is not entirely indicative of a vergence deficit and must be used in conjunction with other tests to arrive at a definitive diagnosis. ¹⁸

Notably, some symptomatic concussed participants had difficulty with divergence. Fifteen percent of the concussed participants failed divergence fusional amplitudes and had a significant reduction in average values compared to control participants, as well as a significant reduction in average values for divergence fusional amplitudes in concussed compared to control participants. These results suggest a heterogeneous visual profile in the concussed population, and point to a generalised vergence dysfunction in both convergence and divergence.

The predominant vergence deficit subtype in symptomatic concussed participants was convergence deficit (35%), followed by convergence insufficiency (21%). Previously, Master et al. ¹⁹ reported a much higher frequency of convergence insufficiency, with 49% of concussed paediatric patients having the diagnosis. However, since their convergence insufficiency diagnosis required a receded NPC, but not necessarily greater exophoria at near than distance as in the current study, their diagnostic category of convergence insufficiency likely included participants with a convergence deficit diagnosis. We previously coined the diagnosis of convergence deficit as a response to a subgroup of concussed patients that lacked greater exophoria at near than distance, but who would otherwise meet the criteria for the classic two‐sign and three‐sign convergence insufficiency. ¹⁸ Given that one would not expect a concussion to cause a change in ocular alignment (heterophoria), it is reasonable to conclude that convergence deficit represents a sizable subtype in this symptomatic concussed group. Other studies that have adopted these diagnostic criteria separating convergence deficit from convergence insufficiency have reported that 18% to 35% of symptomatic concussed participants had convergence insufficiency and 10% to 14% had convergence deficit. ¹⁰ , ¹¹ Recent studies have also used receded NPC, along with either failing vergence facility or poor positive fusional vergence ranges to diagnose convergence insufficiency following concussion. However, these studies did not provide data on near heterophoria, making it unclear whether their findings truly align with the convergence deficit diagnosis referenced here. ⁴² , ⁴³ The convergence deficit subtype indicates a concussed patient profile and aetiology that are distinctly different from naturally occurring convergence insufficiency, possibly arising from disturbances to central neural visual pathways following concussion.

Accommodation

Symptomatic concussed participants failed AA at much higher rates compared to controls (26 [76%] concussed, 1 [3%] control), which suggests that AA could be an effective clinical test in identifying accommodation deficits in concussed patients, with accommodative insufficiency as the predominant subtype. A surprising result in this prospective study concerns the high failure rates with the plus lens (difficulty relaxing accommodation) in controls (11 [31%] participants failed) on the monocular accommodative facility test. The high failure rate and variability could also possibly be due to the high test–retest variability of monocular accommodative facility, which may necessitate a habituation period. ⁴⁴ As it was applied in this study, monocular accommodative facility has limited clinical utility in identifying accommodative deficits in a concussed population. Given the high failure rate in an otherwise visually healthy control group, future studies should revisit the established normative values in a large control population.

Without monocular accommodative facility, only a single clinical test (push‐up AA) was used to diagnose accommodative deficits, whereas at least two or three clinical findings were needed to diagnose vergence deficits. Therefore, studies utilising objective measures of accommodation are needed to confirm and detail the profile of accommodative deficits in the concussed population.

NPC and AA as diagnostic tools for concussion screening

NPC and AA are two tests that have gained popularity with paediatricians, neurologists and sports medicine providers. ⁸ This study found the tests to be predictive in classifying persistently symptomatic concussed participants (Table 6). With high AUC values for the NPC, AA and combined models, both tests could be effective tools in identifying symptomatic concussed as well as non‐concussed control patients when used individually, with a small improvement in performance when combined. Furthermore, clinical cutoff scores determined using the Youden Index from the apparent ROC curves of the individual NPC and AA models were 6.25 cm and 10.88 D, respectively. These scores correspond closely to the respective diagnostic criteria of 6.0 cm and 11 D (Figure 2a, Table 1) and are similar to the cutoff scores used by eye care providers and in randomised clinical trials. ³² , ³³ , ³⁴

Clinical implications

In this study, significantly higher rates of both vergence and accommodative deficits were found in the concussed group, which validated the efficacy of NPC and AA as screening tools to be used by non‐eye care providers.

Accordingly, the cutoff scores of 6 cm for NPC and 11 D (9 cm) for monocular push‐up AA, when measured from the forehead, can be used by clinicians to identify concussed paediatric and adolescent patients with visual function deficits who could benefit from referral to a vision specialist for further comprehensive binocular vision and accommodation evaluation.

Strengths and limitations

The major strength of this study is the inclusion of an age‐matched control group, which serves as a baseline to quantify the degree and frequency of visual function deficits in concussed patients, and allows for diagnostic analysis of screening tools.

There are several limitations to address in this study. Firstly, the concussed cohort was mainly referred to vision providers for persistent concussion symptoms, although not specifically for vision symptoms. As such, this cohort did not represent the general concussed population. Future studies should investigate concussed individuals, regardless of the presence of persistent symptoms, along with an age‐matched control group, to determine the frequency of oculomotor deficits and confirm the cutoff test values.

The regression and ROC analyses performed here lacked external validation due to a small sample size, which presents a concern for overfitting. Although cross‐validation allows assessment of the stability of the model, an external validation group is necessary for a robust evaluation of the diagnostic utility of NPC and AA as screening tools for oculomotor deficits in persistently symptomatic concussed participants. Nonetheless, these analyses resulted in findings similar to clinical cutoff values currently used by eye care providers and in a previous randomised clinical trial. ⁴⁵

CONCLUSIONS

This prospective study found a significantly higher frequency of vergence and accommodative deficits in persistently symptomatic concussed paediatric patients, compared with an age‐matched control group. The oculomotor profile of the symptomatic concussed cohort was distinct from an age‐matched cohort without concussion and was characterised by vergence and accommodative deficits, which deviated from the classic convergence insufficiency profile. The monocular accommodative facility test had low clinical discriminatory utility for accommodation deficits. NPC and monocular AA were found to be effective diagnostic tools which could be used to screen concussion patients who would benefit from a detailed visual function examination by an eye care provider.

AUTHOR CONTRIBUTIONS

Carissa H. Wu coordinated and collected data, carried out all analyses, supervised data collection across sites, drafted the manuscript and revised the manuscript. Sophia Marusic coordinated and collected data, aided data visualisation, reviewed analyses, critically reviewed and revised the manuscript and supervised the study the first 1.5 years across sites. Dr. Jennifer X. Haensel designed the study, collected data, aided data visualisation, reviewed analyses and critically reviewed and revised the manuscript. Dr. Aparna Raghuram and Dr. Tawna L. Roberts conceptualised and designed the study, collected data, reviewed analyses, reviewed and revised the manuscript and supervised the study. Dr. Mitchell Scheiman designed the study and critically reviewed and revised the manuscript. Dr. Isdin Oke aided data visualisation, reviewed analyses and critically reviewed and revised the manuscript. Joellen Leonen collected data, supervised data collection and critically reviewed and revised the manuscript. Dr. Amir Norouzpour critically reviewed and revised the manuscript. Kristin E. Slinger, Neerali Vyas, Christabel A. Ameyaw Baah, Amber Hu and Drs Caitlyn Y. Lew, Gayathri Srinivasan, Erin Jenewein and Siva Meiyeppen collected data and critically reviewed and revised the manuscript. All authors approved the final manuscript as submitted and agree to be accountable for all aspects of the work.

FUNDING INFORMATION

This study was funded by the following: AAO Career Development Award (AR); Boston Children's Hospital Ophthalmology Foundation Discovery Award (AR); Stanford Maternal & Child Health Research Institute (JXH); NEI P30‐EY026877 (Byers Eye Institute at Stanford University); Research to Prevent Blindness (Byers Eye Institute at Stanford University). The sponsors were not involved in the study design, collection, analysis, interpretation, writing or decision to publish this paper.

CONFLICT OF INTEREST STATEMENT

The authors have no conflicts of interest to disclose.

Supporting information

Table S1:

Wu CH, Marusic S, Haensel JX, Oke I, Slinger KE, Vyas N, et al. Post‐concussion clinical findings of oculomotor function in paediatric patients with persisting symptoms compared to healthy controls. Ophthalmic Physiol Opt. 2025;45:1597–1607. 10.1111/opo.70010

DATA AVAILABILITY STATEMENT

The data presented can be shared upon reasonable request to the corresponding author.