Pre-operative vestibular pattern and balance compensation after vestibular schwannoma surgery

Neuroscience 172 (2011) 285–292

PRE-OPERATIVE VESTIBULAR PATTERN AND BALANCE COMPENSATION AFTER VESTIBULAR SCHWANNOMA SURGERY C. PARIETTI-WINKLER,a,b,c G. C. GAUCHARD,a,b C. SIMONc AND P. P. PERRINa,b,c*

eral vestibular deafferentation effects. © 2011 IBRO. Published by Elsevier Ltd. All rights reserved.

a National Institute for Health and Medical Research (Inserm), U 954, Thematic group “Neurodegenerative Diseases, Neuroplasticity, Cognition”, Faculty of Medicine, Vandoeuvre-lès-Nancy, France

Key words: vestibular schwannoma, balance control, vestibular reflectivity, vestibular compensation, neuroplasticity.

b Nancy-Université, Balance Control & Motor Performance, UFR STAPS, Villers-lès-Nancy, France

Postural control in humans is a complex sensorimotor function requiring the central processing of information from the visual, somatokinesthetic and vestibular system, leading to a context-specific motor response (Nashner, 1976; Keshner et al., 1987). This response results in a stabilization of anti-gravity activity and gaze and allows the adjustments of static and dynamic postures (Massion and Woollacott, 1996). Sensory information is complementary and partially redundant allowing a fine-tuned postural control during sensory conflict situations (Curthoys and Halmagyi, 1995). Damage to any of these balance regulation inputs influences the output of the postural system, resulting in an increased risk of postural instability and fall. Vestibular schwannoma (VS) is a benign tumour from Schwann cells surrounding the vestibular nerve, which slowly grows within the internal auditory canal and then into the cerebellopontine angle (Wiegand et al., 1996; Matthies and Samii, 1997), leading to a gradual vestibular dysfunction. The slowly progressive alteration of vestibular function allows the gradual implementation of central adaptive mechanisms called vestibular compensation, which minimizes VS-related symptoms, such as a perceptual syndrome with vertigo or dizziness and clinical signs like body and limb deviation or nystagmus seen in acute vestibular lesions (Curthoys and Halmagyi, 1999; Curthoys, 2000). The total unilateral vestibular deafferentation induced by the surgical tumour removal suddenly leads to a decompensation of this previously compensated situation, which explains why most patients report severe vertigo immediately after surgery and which is responsible for perturbations of the postural control. After surgery, a progressive implementation of central adaptive mechanisms leads to a restoration and even improvement of vestibular and balance performances, associated with a lower number of falls, development of more appropriate sensorimotor and/or behavioural strategies and better resolution of sensory conflicts (Jenkins, 1985; Parietti-Winkler et al., 2006). However, the impact of tumoral growth on vestibular function seems to vary from one subject to another, with regard to the increasing pressure, which the slowly growing tumour exerts on the VIII cranial nerve and its vessels. The lesion of the vestibular end organ and nerve is usually visible in a reduced caloric response ipsilateral to the tumour. A vestibular paresis before surgery at the side of

c

University Hospital of Nancy, Department of Oto-Rhino-Laryngology, Head and Neck Surgery, Nancy, France

Abstract—This longitudinal study aimed to assess the sensorimotor balance strategies before and after vestibular schwannoma (VS) surgery according to the degree of preoperative vestibular lesion. Thirty-eight VS patients were split in three groups according to caloric vestibular test results before surgery; nine had a symmetrical vestibular response (vestibular normoreflexy), 19 with a decreased response of more than 20% of the affected side (vestibular hyporeflexy) and 10 with an absent caloric response on the side of the affected labyrinth (vestibular areflexy). They underwent pendular rotary vestibular testing (RVT), allowing to evaluate gain and directional preponderance of the vestibulo-ocular reflex, and a sensory organisation test (SOT), evaluating balance control in six conditions (C1 to C6). These tests were performed shortly before, and 8 and 90 days after surgery. Directional preponderance performances of patients with vestibular normoreflexy or hyporeflexy followed a classical time-course with a huge asymmetry just after surgery and a recovery to pre-operative performances at 90 days; patients with vestibular areflexy were relatively stable in time. Variation in SOT performances of patients with vestibular normoreflexy, especially in the more complex C4 to C6, followed a classical time-course with an important postural degradation just after surgery and a recovery to pre-operative performances at 90 days. Patients with vestibular areflexy showed no balance degradation just after surgery and a marked increase in performances at 90 days after surgery, especially in C5 and C6. Performances of patients with vestibular hyporeflexy were intermediate, close to performances of patients with vestibular normoreflexy before surgery and close to performances of patients with vestibular areflexy at 8 and 90 days after surgery. Pre-operative vestibular function alteration triggers an adaptive process, characterized by a restoration of the symmetry of the vestibular nuclei activity and by sensory substitution and new behavioural strategies, allowing the anticipation of unilat*Correspondence to: P. P. Perrin, Equilibration et Performance Motrice, Nancy-Université, UFR STAPS, 30, rue du Jardin Botanique, 54 600 Villers-lès-Nancy, France. Tel: ⫹33-383-682-929; fax: ⫹33-383-154-647. E-mail address: [email protected] (P. P. Perrin). Abbreviations: AS8, eight days after surgery; AS90, ninety days after surgery; BS, before surgery; CoG, centre of gravity; EP, endogenous pacemakers; ES, equilibrium score; IQR, interquartile range; NP, nonpacemaker; NS, non significant; RVEST, vestibular ratio; RVIS, visual ratio; RVT, rotary vestibular test; SOT, sensory organization test; VS, vestibular schwannoma.

0306-4522/11 $ - see front matter © 2011 IBRO. Published by Elsevier Ltd. All rights reserved. doi:10.1016/j.neuroscience.2010.10.059

285

286

C. Parietti-Winkler et al. / Neuroscience 172 (2011) 285–292

the tumour in caloric testing is only found in 61 to 87% of patients and complete in 30 to 60% of cases (Bergenius and Magnusson, 1988; Pfaltz et al., 1991; El-Kashlan et al., 1998; Darrouzet et al., 2004). Thus, the presurgical degree of vestibular compensation, which sets in together with the progressive loss of vestibular function due to tumoral growth, also varies. Therefore, the consequences of the unilateral vestibular deafferentation on postural control after surgery are expected to be more important in patients without vestibular paresis before surgery. This hypothesis could explain results, in which patients who had reduced pre-operative caloric response appear to have less post-operative dizziness compared with patients with normal pre-operative vestibular testing (El-Kashlan et al., 1998; Humphriss et al., 2003). To our knowledge, most of the studies on VS investigating the post-operative relationships between balance control, posturography and vestibular compensation were not interested to the magnitude of pre-operative vestibular paresis. The studies, which have taken into account the pre-operative vestibular reflectivity, were interested to a subjective evaluation of balance by questionnaire. So far no study was interested in the pre-operative magnitude of vestibular paresis and its effect on the post-operative control evaluated measured by posturography. This longitudinal study aimed to assess the evolution of pre- to postoperative postural strategies in vestibular schwannoma surgery according to pre-operative vestibular pattern.

EXPERIMENTAL PROCEDURES Patients Thirty-eight patients (median (M) age⫽55.0 years, interquartile range (IQR)⫽16.0 years, 24 women) with unilateral VS, who were scheduled for surgical ablation, volunteered to participate to this study. Each participant had clinical otoneurological examination, instrumental audiological and vestibular tests and magnetic resonance imaging (MRI) with gadolinium contrast. Tumours were measured from the pre-operative MRI realized between 1 and 3 months before surgery and dispatched according to the Koos classification in four stages (Koos et al., 1976); the stage of each tumour was confirmed by the per-operative findings. Stage I corresponds to a tumour with a size less than 10 mm confined to the internal auditory meatus; stage II, to a tumour with a size less than 20 mm penetrating the cerebello-pontine angle without contact with the brainstem; stage III, to a tumour with a size greater than 20 mm, filling in the cerebello-pontine angle but not compressing the brainstem; stage IV, to any compressing tumour whatever its size (Darrouzet et al., 2004). Stage IV tumours have been excluded from this study because of potential brain stem participation to the postural dysfunction. All VS ablations were performed by the same surgeon using the translabyrinthine approach. Both inferior and superior vestibular nerves were resected in all subjects. Except for transient post-operative facial palsy, all patients presenting post-operative complications, such as meningitis or cerebrospinal fluid leak, were excluded from the study. After surgery, the patients were mobilized as early as possible and all underwent an active post-operative postural training program. The rehabilitation program was composed of different exercises designed to improve balance, mobility and coordination, beginning one day after surgery and ending just before being discharged from the hospital centre, that is for 8 to 10 consecutive days. The two first days consisted in the recovery of verticality, initially sitting

posture and then standing posture, with and without ocular fixation. From the third day, balance and gait exercises were introduced, combining ocular fixation or head movements, forward and backward, with direction and speed changes. Besides, eye-head fixation and coordination exercises were introduced in sitting position. From the fifth day, balance and gait exercises were performed in eyes closed condition too, and the eye-head coordination exercises were introduced during gait with random orders. No specific rehabilitation program has been given to the patients after discharge from the hospital.

Vestibular evaluation and patient distribution Two tests were performed with videonystagmography (Synapsys, France). (i) The bithermal caloric vestibular testing allowed to evaluate the degree of asymmetry according to the Jongkees formula (Jongkees and Philipzoon, 1964; Gauchard et al., 2003) and to split participants in three groups, those with absent caloric response on the side of the affected labyrinth (vestibular areflexy), those with a decreased response, on the side of the affected labyrinth, of more than 20% (Bergenius and Magnusson, 1988) (vestibular hyporeflexy) and those with a symmetrical response (vestibular normoreflexy) (Table 1). (ii) The pendular rotary vestibular test (RVT) allowed to calculate both the gain of the vestibulo-ocular reflex, which was determined by the ratio of the amplitudes of the eye velocity to those of the chair velocity, and the directional preponderance, which was determined by the mean slow phase eye velocity over the duration of the stimulus (Parietti-Winkler et al., 2008).

Posturography The sensory organization test (SOT, EquiTest®, Neurocom®, OR, USA) evaluates the patient’s ability to make effective use of visual, vestibular and somatosensory inputs separately and to suppress sensory information that is inappropriate. To give inadequate information, somatosensory and visual cues are disrupted by using a technique commonly referred to as sway-referenced, which involves tilting the support surface and/or the visual surround to directly follow the anterior–posterior sways of the subject’s centre of gravity (CoG). The patient’s task is to maintain an upright stance, as stable as possible, during the three 20 s trials in six conditions that combine three visual conditions with two platform conditions (Table 2). An equilibrium score (ES) was calculated by comparing the subject’s anterior–posterior sway during each 20 s SOT trial to the maximal theoretical sway limits of stability, which is based on the individual’s height and size of the base of support. It represents an angle (8.5° anteriorly and 4.0° posteriorly) at which the person can lean in any direction before the centre of gravity would move beyond a point that allows him/her to remain upright (i.e., point of falling). The following formula was used to calculate the ES: [12.5°–((␪max–␪min)/12.5°)]⫻100, where ␪max indicates the greatest AP CoG sway angle displayed by the subject while ␪min indicates the lowest AP CoG sway angle. Lower sways lead to a higher ES, indicating a better balance control performance (a score of 100 represents no sway, while 0 indicates sway that exceeds the limit of stability, resulting in a fall). ES were calculated for every condition: the ES were C1ES in condition 1, C2ES in condition 2, C3ES in condition 3, C4ES in condition 4, C5ES in condition 5, and C6ES in condition 6. Composite equilibrium score (CES) was calculated to evaluate global balance performances. It was calculated by adding every ES from conditions 1 and 2 and three times every ES from sensory conditions 3, 4, 5 and 6, and finally dividing that sum by the total number of trials, that would smooth the data across all conditions. Besides, each ES was adjusted to C1ES to identify the significance of each sensory system influencing postural control (Table 2) (Nashner and Peters, 1990; Black et al., 1995; Herdman et al., 1995).

C. Parietti-Winkler et al. / Neuroscience 172 (2011) 285–292

287

Table 1. Characteristics of the groups constituted by patients with vestibular normoreflexy (group NR), hyporeflexy (group HR) or areflexy (group AR). Intergroup comparisons were performed with the Kruskall-Wallis H-test for the quantitative parameters (expressed in median associated with the interquartile range (IQR)), such as age, height, weight and body mass index (BMI), and with the ␹2-test for the qualitative parameters (expressed in number of subject (n=) presenting the character), such as gender, tumour stage or transient post-operative facial palsy

Age (y) Gender Women Men Height (m) Weight (kg) Body mass index (kg/m2) Tumour stage Stage 1 Stage 2 Stage 3 Transient post-operative facial palsy

Group NR, n⫽9 median (IQR)/n=

Group HR, n⫽19 median (IQR)/n=

Group AR, n⫽10 median (IQR)/n=

Kruskall-Wallis or ␹2 tests

Intergroup comparisons NR vs. HR

NR vs. AR

HR vs. AR

51.0 (15.0)

55.0 (16.5)

55.5 (7.0)

NS

—

—

—

5 4 1.70 (0.15) 78.0 (18.3) 25.2 (3.7)

11 8 1.66 (0.15) 78.0 (19.5) 27.8 (4.7)

8 2 1.64 (0.11) 72.5 (17.0) 24.0 (6.5)

NS

—

—

—

NS NS NS

— — —

— — —

— — —

NS

—

—

—

NS

—

—

—

3 2 4 3

4 7 8 5

2 3 5 1

NS, non significant.

Procedure Tests were carried out in the Laboratory for the Analysis of Posture, Equilibrium and Movement (LAPEM) of the University Hospital of Nancy (Ministry of Health agreement for research). All patients gave informed consent prior to the study and were then submitted to RVT and SOT 3 days before surgery (BS) and two times after surgery, at short term (8 days, AS8) and at middle term (90 days, AS90).

Statistical analysis The statistics were produced using non-parametric tests considering the relatively small sample size in the three groups. For the

intra-group comparisons between the three stages (BS, AS8 and AS90), the Friedman (␹2, overall heterogeneity) and Wilcoxon (z, pairwise comparisons) tests were used for all RVT and SOT parameters. For the inter-group comparisons, Kruskall-Wallis (overall heterogeneity) and Mann and Whitney test (pairwise comparison) were used for age, anthropometric parameters and RVT and SOT parameters. The Yates corrected ␹2-test (distribution comparison) was used for gender, tumour stage and facial palsy parameters. A probability level of Pⱕ0.05 was used as an indicator of statistically significant results for overall heterogeneity and Bonferroni procedure was applied to pairwise comparisons by adjusting “familywise alpha” to a significance level of Pⱕ0.05/

Table 2. Sensory organization test (EquiTest®, Neurocom®, Clackamas, OR, USA): determination of the six conditions and significance of sensory ratios (according to Nashner and Peters, 1990; Black et al., 1995; Herdman et al., 1995) Sensory organisation test Conditions Name Condition 1 (C1) Condition 2 (C2) Condition 3 (C3) Condition 4 (C4) Condition 5 (C5) Condition 6 (C6)

Situation Eyes open, fixed support Eyes closed, fixed support SR surround, fixed support Eyes open, SR support Eyes closed, SR support SR surround, SR support

Available cues Vision, vestibular, somatosensory Vestibular, somatosensory Vestibular, somatosensory Vision, vestibular Vestibular Vestibular

Ratios Name Somatosensory (RSOM)

Pair C2/C1

Significance Question: Does sway increase when visual cues are removed? Low scores: Poor use of somatosensory references. Question: Does sway increase when somatosensory cues are removed? Low scores: Poor use of visual references. Question: Does sway increase when visual cues are removed and somatosensory cues are inaccurate? Low scores: Poor use of vestibular cues or vestibular cues unavailable. Question: Do inaccurate visual cues result in increased sway compared to no visual cues? Low scores: Reliance on visual cues even inaccurate. Question: Do inaccurate somatosensory cues result in increased sway compared to accurate somatosensory cues? Low scores: Poor compensation for disruptions in selected sensory inputs.

Visual (RVIS)

C4/C1

Vestibular (RVEST)

C5/C1

Visual preference (RPREF)

(C3⫹C6)/(C2⫹C5)

Altered proprioceptive information management (RPMAN)

(C4⫹C5⫹C6)/(C1⫹C2⫹C3)

SR, sway-referenced.

Unavailable or altered cues — No vision Vision altered Somatosensory altered No vision, somatosensory altered Vision altered, somatosensory altered

288

C. Parietti-Winkler et al. / Neuroscience 172 (2011) 285–292

Table 3. Rotary vestibular test (RVT): Median results, associated with interquartile range (IQR), of the gain and directional preponderance (DP) for the patients with vestibular normoreflexy (group NR), hyporeflexy (group HR) or areflexy (group AR) just before (BS), and eight (AS8) and ninety (AS90) days after vestibular schwannoma surgery RVT

Group NR Gain DP (°/s) Group HR Gain DP (°/s) Group AR Gain DP (°/s)

BS median (IQR)

AS8 median (IQR)

AS90 median (IQR)

Friedman test P-value

0.50 (0.25) 0.70 (1.73)

0.43 (0.18) 7.60 (4.98)

0.37 (0.23) 1.50 (3.00)

0.53 (0.40) 0.60 (1.18)

0.38 (0.26) 3.70 (4.08)

0.35 (0.28) 0.80 (1.30)

0.30 (0.30) 2.15 (2.40)

Wilcoxon z-test/post-hoc Bonferroni correction BS vs. AS8 P-value

BS vs. AS90 P-value

AS8 vs. AS90 P-value

NS P⫽0.013

— P⫽0.015

— NS

— P⫽0.013

0.56 (0.46) 2.20 (2.73)

P⫽0.065 P⫽0.001

— P⬍0.001

— P⫽0.023

— P⫽0.005

0.31 (0.28) 0.95 (1.30)

NS NS

— —

— —

— —

P-values in italics⫽borderline significance; P-values in normal typography⫽statistical significance; NS⫽non significant; (—) non calculated.

3⫽0.017 in accordance with the three possible intra-group or inter-group comparisons; in the same way, borderline significance was defined at a level of Pⱕ0.10/3⫽0.033.

RESULTS Data on age, gender, height, weight, body mass index, tumour stage or transient post-operative facial palsy for the three groups are presented in Table 1. No significant difference was observed for all these parameters between the three groups. For the RVT analysis (Table 3), intra-group comparisons showed that gain values were similar in the three groups, no statistical heterogeneity being observed between the three stages. On the other hand, while the values of directional preponderance were stable in patients with vestibular areflexy, they showed more variation in patients with vestibular normoreflexy and hyporeflexy, statistical significance being observed between AS8 and both BS and AS90 in these two groups. No statistical significance was observed between BS and AS90. Inter-group comparisons showed no heterogeneity between the three groups for gain values for the three stages and for values of directional preponderance in BS and AS90. On the other hand, an overall heterogeneity was observed at AS8 (P⫽0.035), borderline significance being observed between patients with vestibular normoreflexy and areflexy (P⫽0.027); on the other hand, no statistical significance was observed between patients with vestibular hyporeflexy and both patients with vestibular normoreflexy and areflexy. For the SOT analysis and ratios (Table 4A, B, Fig. 1), intra-group comparisons showed a high overall heterogeneity between the three stages for C4ES, C5ES, C6ES, CES, RVEST and RPMAN in the three groups and additionally, in patients with vestibular normoreflexy, for C1ES, C2ES and RVIS and, in patients with vestibular hyporeflexy, for RVIS and RPREF. Concerning patients with vestibular normoreflexy, performances were mainly characterized by a high decrease in values at AS8 and by a recovery of BS values at AS90, statistically significant differences being observed between AS8 and both BS and AS90 for C4ES, C5ES, C6ES,

CES, RVEST and RPMAN. No difference was observed between BS and AS90. Concerning patients with vestibular hyporeflexy, performances were mainly characterized by decrease in values at AS8 and a marked high increase in values at AS90, which were higher than those observed BS; statistically significant differences were observed between BS and AS8 for C5ES, CES, RVEST and RPREF and between AS90 and both BS and AS8 for C4ES, C5ES, C6ES, CES, RVIS, RVEST and RPMAN. Concerning patients with vestibular areflexy, performances were stable at AS8 and increased at AS90 compared to BS, statistically significant differences being observed between AS90 and both BS and AS8 for the C4ES, C5ES, C6ES, CES, RVEST and RPMAN. Inter-group comparisons in BS showed a high heterogeneity for the main ES (C5ES: P⫽0.004; C6ES: P⫽0.021; CES: P⫽0.014) and for ratios (RVIS: P⫽0.058; RVEST: P⫽0.005; RPMAN: P⫽0.002), patients with vestibular areflexy displaying lower performances than patients with vestibular normoreflexy (C5ES: P⫽0.001; C6ES: P⫽0.004; CES: P⫽0.006; RVEST: P⫽0.002; RPMAN: P⬍0.001) and hyporeflexy (C5ES: P⫽0.020; RVEST: P⫽0.026; RPMAN: P⫽0.031); no statistically significant differences were observed between patients with vestibular normoreflexy and hyporeflexy. In AS8, a high heterogeneity was found for ES (C1ES: P⫽0.095; C2ES: P⫽0.022; C5ES: P⫽0.043; C6ES: P⫽0.027; CES: P⫽0.013) and ratios (RSOM: P⫽0.032; RVEST: P⫽0.041; RPMAN: P⫽0.020), patients with vestibular normoreflexy displaying lower performances than patients with vestibular hyporeflexy (C5ES: P⫽0.028; C6ES: P⫽0.009; CES: P⫽0.005; RVEST: P⫽0.028; RPMAN: P⫽0.007) and areflexy (C2ES: P⫽0.006; C5ES: P⫽0.018; CES: P⫽0.018; RSOM: z⫽⫺2.8, P⫽0.005; RVEST: P⫽0.018; RPMAN: P⫽0.027); no statistically significant differences were observed between patients with vestibular normoreflexy and hyporeflexy. At AS90, a high heterogeneity was found for ES (C2ES: P⫽0.067; C3ES: P⫽0.014) and RSOM (P⫽0.016), patients with vestibular normoreflexy displaying lower performances than patients with vestibular hyporeflexy (C3ES: P⫽0.005; RSOM: P⫽0.018) and areflexy (RSOM: P⫽0.029).

C. Parietti-Winkler et al. / Neuroscience 172 (2011) 285–292

289

Table 4. Sensory organization test (SOT): Median results, associated with interquartile range (IQR), of (Part A) the equilibrium scores (ES) for the six conditions (C1, C2, C3, C4, C5 and C6) and for the composite equilibrium (CES) and strategy (CSS) scores and of (Part B) the ratios of the somesthetic (RSOM), visual (RVIS), vestibular (RVEST) functions, of visual preference (RPREF) and of altered proprioceptive inputs (RPMAN) for the patients with vestibular normoreflexy (group NR), hyporeflexy (group HR) or areflexy (group AR) just before (BS), and eight (AS8) and ninety (AS90) days after vestibular schwannoma surgery

Part A SOT ES Group NR C1ES C2ES C3ES C4ES C5ES C6ES CES Group HR C1ES C2ES C3ES C4ES C5ES C6ES CES Group AR C1ES C2ES C3ES C4ES C5ES C6ES CES Part B SOT ratios Group NR RSOM RVIS RVEST RPREF RPMAN Group HR RSOM RVIS RVEST RPREF RPMAN Group AR RSOM RVIS RVEST RPREF RPMAN

BS median (IQR)

AS8 median (IQR)

AS90 median (IQR)

Friedman test P-value

94.7 (1.5) 92.7 (2.7) 90.0 (4.6) 87.0 (5.0) 63.7 (13.4) 64.7 (11.5) 78.9 (5.7)

93.0 (2.5) 87.3 (5.2) 90.0 (7.3) 80.7 (9.0) 0.0 (0.0) 0.0 (9.1) 50.0 (6.1)

94.0 (1.7) 89.7 (5.9) 89.3 (3.4) 87.0 (3.3) 59.0 (6.0) 70.0 (14.3) 78.9 (6.3)

95.3 (2.0) 92.0 (2.8) 91.3 (2.9) 84.3 (5.3) 56.3 (27.1) 59.3 (15.0) 75.8 (7.0)

94.3 (2.3) 91.7 (4.8) 89.7 (5.4) 83.7 (8.2) 5.0 (48.8) 56.7 (56.0) 64.7 (20.1)

93.7 (2.7) 91.0 (5.0) 88.7 (9.7) 83.2 (5.7) 34.7 (12.3) 47.3 (21.0) 67.2 (6.2)

Wilcoxon z-test/post-hoc Bonferroni correction BS vs. AS8 P-value

BS vs. AS90 P-value

AS8 vs. AS90 P-value

P⫽0.038 P⫽0.015 NS P⫽0.007 P⫽0.002 P⬍0.001 P⫽0.001

NS NS — P⫽0.011 P⫽0.008 P⫽0.008 P⫽0.008

NS NS — NS NS NS NS

NS NS — P⫽0.012 P⫽0.012 P⫽0.008 P⫽0.008

94.7 (2.3) 92.3 (3.2) 91.7 (1.7) 88.3 (3.5) 65.3 (8.8) 71.7 (14.8) 81.8 (6.0)

NS NS NS P⫽0.001 P⬍0.001 P⬍0.001 P⬍0.001

— — — NS P⫽0.002 NS P⫽0.016

— — — P⫽0.004 P⫽0.002 P⫽0.001 P⬍0.001

— — — P⬍0.001 P⬍0.001 P⬍0.001 P⬍0.001

94.2 (2.3) 92.0 (1.3) 90.7 (6.3) 85.0 (5.0) 37.0 (60.0) 44.0 (58.7) 64.1 (26.5)

94.7 (3.7) 91.8 (2.7) 90.5 (3.7) 87.0 (6.3) 62.2 (10.7) 72.2 (16.3) 80.5 (5.3)

NS NS NS P⫽0.042 P⫽0.002 P⫽0.002 P⬍0.001

— — — NS NS NS NS

— — — P⫽0.005 P⫽0.005 P⫽0.007 P⫽0.005

— — — NS P⫽0.007 P⫽0.005 P⫽0.005

0.97 (0.04) 0.92 (0.06) 0.67 (0.15) 1.02 (0.07) 0.76 (0.06)

0.94 (0.04) 0.86 (0.09) 0.00 (0.00) 1.04 (0.14) 0.30 (0.10)

0.94 (0.04) 0.91 (0.04) 0.63 (0.06) 1.04 (0.09) 0.78 (0.07)

NS P⫽0.025 P⫽0.002 NS P⫽0.001

— P⫽0.033 P⫽0.008 — P⫽0.008

— NS NS — NS

— P⫽0.021 P⫽0.011 — P⫽0.008

0.97 (0.02) 0.90 (0.05) 0.61 (0.27) 1.02 (0.11) 0.73 (0.15)

0.97 (0.04) 0.89 (0.07) 0.05 (0.52) 1.13 (0.37) 0.54 (0.34)

0.97 (0.02) 0.94 (0.02) 0.69 (0.10) 1.04 (0.05) 0.80 (0.07)

NS P⫽0.010 P⬍0.001 P⫽0.013 P⬍0.001

— NS P⫽0.002 P⫽0.002 P⫽0.019

— P⫽0.007 P⫽0.001 NS P⬍0.001

— P⫽0.005 P⬍0.001 P⫽0.017 P⬍0.001

0.98 (0.04) 0.89 (0.04) 0.37 (0.12) 1.08 (0.14) 0.62 (0.09)

0.98 (0.01) 0.90 (0.05) 0.39 (0.63) 0.98 (0.10) 0.55 (0.41)

0.98 (0.02) 0.91 (0.05) 0.67 (0.14) 1.05 (0.10) 0.79 (0.14)

NS P⫽0.097 P⫽0.002 NS P⬍0.001

— — NS — NS

— — P⫽0.005 — P⫽0.005

— — P⫽0.007 — P⫽0.005

P-values in italics⫽borderline significance; P-values in normal typography⫽statistical significance; NS⫽non significant; (—) non calculated.

DISCUSSION This longitudinal study has shown that postural performances evolved in different ways according to pre-operative vestibular reflectivity. The variation in performances of patients with vestibular normoreflexy followed a classical time-course with a huge postural degradation just after surgery and a recovery to pre-operative performances 3 months after surgery; in patients with vestibular areflexy,

there was no balance degradation just after surgery and a big increase in performances 3 months after surgery. Performances of patients with vestibular hyporeflexy were intermediate, close to those of patients with vestibular normoreflexy before surgery and close to those of patients with vestibular areflexy just after surgery and 3 months after. Thus, whereas postural BS performances were lower in patients with vestibular areflexy than in the two other

290

C. Parietti-Winkler et al. / Neuroscience 172 (2011) 285–292

Fig. 1. Representation of the changes over time for the C5ES (black symbol, dotted line) and CES (white symbol, full line) parameters for the patients with vestibular normoreflexy (group NR, lozenge), hyporeflexy (group HR, square) or areflexy (group AR, triangle) just before (BS), and eight (AS8) and ninety (AS90) days after vestibular schwannoma surgery.

groups, they became similar for patients with vestibular areflexy and hyporeflexy and lower in patients with vestibular normoreflexy at AS8. At AS90, although postural performances improved for all patients, some values of patients with vestibular normoreflexy were lower than those of the two other groups. Concerning directional preponderance, whereas performances of patients with vestibular normoreflexy and hyporeflexy followed a classical timecourse with a huge asymmetry just after surgery and a recovery of pre-operative performances at 3 months, patients with vestibular areflexy showed stable performances between BS and AS90. Some features of our study should be considered, especially regarding the surgical approach, localization of VS on the vestibular nerve branches or the determination of groups according vestibular testings. Concerning surgery, all surgeries have been performed by a single surgeon. Moreover, only patients operated with translabyrinthine approach were included. Patients who underwent other surgical techniques such as retrosigmoid or middle fossa approaches, which could affect balance performances in a different way than the translabyrinthine approach, were not included in the study. Therefore, these results cannot be generalized to all surgical techniques. Moreover, although it has been shown a relationship between post-operative posturographic results and the superior or inferior origin of the tumour on the vestibular nerve (Gouveris et al., 2006), this study has not split the data according to the nerve of origin of the tumour. It could constitute a new field of investigation in future studies. Thus, the determination of vestibular pattern only based on through bithermal caloric testing is maybe an incomplete approach. However, the addition of rotary testings for this determination would be optimal, but would lead to a too

important increase in participant number regarding the number of potential groups (six instead of three), that is not compatible with our sample size. In studies based on the measures of post-operative symptoms and on questionnaire evaluating the consequences of dizziness on daily life (El-Kashlan et al., 1998; Humphriss et al., 2003), patients who had a reduced caloric response before surgery appeared to present less post-operative dizziness compared with patients who had normal pre-operative vestibular testing. These results seem to be in agreement with our posturographic study results, vestibular function pattern appearing predictive for the immediate post-operative symptoms and postural performances. Moreover, the fact that the degree of tumour related vestibular dysfunction is predictive for the postoperative postural performances could be explained by the Ris hypothesis (Ris et al., 1997), who proposed a subdivision of the intact vestibular nucleus neurons into two categories: firstly, endogenous pacemakers (EP) neurons, whose resting activity is relatively little susceptible to unilateral vestibular deafferentation, and secondly, non-pacemaker (NP) neurons which become silent after unilateral vestibular deafferentation. Indeed, according to this hypothesis, in patients with vestibular normoreflexy, the activity of the vestibular nuclei ipsilateral to the VS before surgery should be based on both functional EP and NP neurons. After the surgical tumour removal, the deprived NP neurons become silent as opposed to the EP neurons. The dramatic degradation of postural performances observed immediately after surgery could be related to the sudden unilateral suppression of vestibular information originating from NP neurons. Spontaneous restoration of resting activity in the deprived NP vestibular neurons could then occur by intrinsic mech-

C. Parietti-Winkler et al. / Neuroscience 172 (2011) 285–292

anisms, such as increased excitability or sensitivity to neuromediators (Darlington and Smith, 1996; Guilding and Dutia, 2005), or extrinsic mechanisms, such as modifications in synaptic connections (Li et al., 2002). Those intrinsic or extrinsic phenomenons, triggered by unilateral vestibular deafferentation, lead to a recovery of the neural activity balance between the two vestibular nuclei (Curthoys and Halmagyi, 1995; Curthoys, 2000) and to the improvement of the postural control performances observed in our study. However, Ris showed a weak correlation between the resolution of postural symptoms and return of balance in the medial vestibular nuclei, suggesting that, at the same time of this activity restoration in the NP vestibular neurons deprived of their inputs, other central adaptive mechanisms are triggered by surgery, such as substitution by extra-vestibular sensory afferences, efference copy/re-afference copy reconciliation and new behavioural strategies (Ris et al., 1997; Curthoys and Halmagyi, 1999; Curthoys, 2000), and play also a role in the improvement of postural performances. Indeed, before surgery, patients with vestibular normoreflexy did not have to implement central adaptive mechanisms because of a vestibular function which remained symmetric. All these phenomena (restoration of activity in the NP deprived vestibular neurons and sensorial substitution, efference copy/ re-afference copy reconciliation and new behavioural strategies altogether) lead to a reduction in postural tone asymmetry with the recovery of the descending vestibulo-spinal activity and allow behavioural vestibular compensation. For patients with vestibular areflexy, the activity of the vestibular nuclei ipsilateral to the VS before surgery is based on EP neurons and on NP neurons which were deafferented because of the tumour and the resting activity of which had been restored even before unilateral vestibular deafferentation. The surgical removal of the tumour thus does not seem to modify the activity of the vestibular nuclei. The tumour related deafferentation could trigger the recovery of the resting activity of the NP neurons and then anticipate the effects of unilateral vestibular deafferentation on NP neurons. Concurrently to the mechanisms leading to the restoration of the activity of deafferented NP neurons (Darlington and Smith, 1996; Li et al., 2002; Guilding and Dutia, 2005), sensory substitution and new behavioural strategies could also be implemented before surgery and so constitute an experience of vestibular compensation for patients, who would not have to implement the adaptive process de novo after surgery. This kind of preoperative anticipation of the effects of the surgical unilateral vestibular deafferentation related to the tumoral deafferentation could explain why patients with vestibular areflexy present a globally efficient balance control before surgery, although their postural performances were significantly worse than those of patients with vestibular normoreflexy in test conditions using preferentially vestibular information. This could also explain the absence of a dramatic degradation of postural performances immediately after surgery. The patients with vestibular hyporeflexy could represent an intermediate situation. Indeed, we can hypothesize

291

that the deafferentation of NP neurons is less pronounced for patients with vestibular hyporeflexy than for those with vestibular areflexy. The surgical removal of the tumour completes the vestibular deafferentation started by the tumoral growth. Indeed, surgery leads to a deafferentation of the remaining functional NP neurons which become silent, but does not induce any modifications for those of NP neurons which were already deafferented by the tumour and the resting activity restoration of which had occurred before surgery. In these patients, an asymmetry between the activities of the two vestibular nuclei is induced by the surgery but to a lesser extend than for patients with vestibular normoreflexy. This could explain that the postural performances degradations observed immediately after surgery in patients with vestibular hyporeflexy were less important than those of patients with vestibular normoreflexy and more important than those of patients with vestibular areflexy. At 3 months after surgery, whatever the pre-operative pattern of vestibular function, we observed an improvement of postural performances compared to immediately after surgery. This improvement is more pronounced for patients with vestibular areflexy and hyporeflexy, with significant improvement compared to their performances before surgery, as opposed to patients with vestibular normoreflexy who returned to their pre-operative level. We could expect that this difference fades away with time and that the postural performances of patients with vestibular normoreflexy reach the level of patients with vestibular hyporeflexy and areflexy at longer delay than 3 months (Parietti-Winkler et al., 2010). Indeed, as seen above, patients with vestibular normoreflexy did not experience vestibular compensation before surgery and thus had to implement the adaptive process de novo after surgery. We can suppose that the implementation of the vestibular compensation process is longer for these patients than for patients with vestibular areflexy, and, to a lesser degree, patients with vestibular hyporeflexy, in whom the tumoral growth anticipated partially the effect of the surgical unilateral vestibular deafferentation on postural control regulation. It would be interesting in the future to compare the postural performances according to the vestibular function pattern longer after surgery, especially after one year (Parietti-Winkler et al., 2010).

CONCLUSION In conclusion, the pre-operative vestibular pattern seems to play an important role in the balance control compensation after vestibular schwannoma surgery. An alteration of the vestibular function related to the tumoral growth triggers the implementation of adaptive processes, characterized by a restoration of the symmetry of activity of both vestibular nuclei but also a sensory substitution and new sensorimotor strategies, which allow to anticipate partially the effects of complete surgical unilateral vestibular deafferentation on balance regulation.

292

C. Parietti-Winkler et al. / Neuroscience 172 (2011) 285–292

REFERENCES Bergenius J, Magnusson M (1988) The relationship between caloric response, oculomotor dysfunction and size of cerebello-pontine angle tumours. Acta Otolaryngol (Stockh) 106:361–367. Black FO, Paloski WH, Doxey-Gasway DD, Reschke MF (1995) Vestibular plasticity following orbital spaceflight: recovery from posflight postural instability. Acta Otolaryngol Suppl 520:450 – 454. Curthoys IS (2000) Vestibular compensation and substitution. Curr Opin Neurol 13:27–30. Curthoys IS, Halmagyi GM (1995) Vestibular compensation: a review of the oculomotor, neural, and clinical consequences of unilateral vestibular loss. J Vestib Res 5:67–107. Curthoys IS, Halmagyi GM (1999) Vestibular compensation. Adv Otorhinolaryngol 55:82–110. Darlington CL, Smith PF (1996) The recovery of static vestibular function following peripheral vestibular lesions in mammals: the intrinsic mechanism hypothesis. J Vestib Res 6:185–201. Darrouzet V, Martel J, Enee V, Bébéar JP, Guérin J (2004) Vestibular schwannoma surgery outcomes: our multidisciplinary experience in 400 cases over 17 years. Laryngoscope 114:681– 688. El-Kashlan HK, Shepard NT, Arts HA, Telian SA (1998) Disability from vestibular symptoms after acoustic neuroma resection. Am J Otol 19:104 –111. Gauchard GC, Gangloff P, Jeandel C, Perrin PP (2003) Physical activity improves gaze and posture control in the elderly. Neurosci Res 45:409 – 417. Gouveris H, Akkafa S, Lippold R, Mann W (2006) Influence of nerve of origin and tumor size of vestibular schwannoma on dynamic posturography findings. Acta Otolaryngol 126:1281–1285. Guilding C, Dutia MB (2005) Early and late changes in vestibular neuronal exitability after deafferentation. Neuroreport 16:1415– 1418. Herdman SJ, Clendaniel RA, Mattox DE, Holliday MJ, Niparko JK (1995) Vestibular adaptation exercises and recovery: acute stage after acoustic neuroma resection. Otolaryngol Head Neck Surg 113:77– 87. Humphriss RL, Baguley DM, Moffat DA (2003) Change in dizziness handicap after vestibular schwannoma excision. Otol Neurotol 24:661– 665. Jenkins HA (1985) Long-term adaptive changes of the vestibulo-ocular reflex in patients following acoustic neuroma surgery. Laryngoscope 95:1224 –1234. Jongkees B, Philipzoon AJ (1964) Electronystagmography. Acta Otolaryngol 198:55.

Keshner EA, Allum JHJ, Pfaltz CR (1987) Postural coactivation and adaptation in the sway stabilizing responses of normals and patients with bilateral vestibular deficit. Exp Brain Res 69:77–92. Koos WT, Spetzler RF, Böck FW (1976) Microsurgery of cerebellopontine angle tumors. In: Clinical microneurosurgery (Koos WT, Böck FW, Spetzler RF, Ammerman B, eds), pp 91–112. Stuttgart and Acton, Mass: Georg Thieme. Li H, Dokas LA, Godfrey DA, Rubin AM (2002) Remodeling of synaptic connections in the deafferented vestibular nuclear complex. J Vestib Res 12:167–183. Massion J, Woollacott MH (1996) Posture and equilibrium. In: Balance, posture and gait (Bronstein AM, Brandt T, Woollacott M, eds), pp 1–18. London: Arnold. Matthies C, Samii M (1997) Management of 1000 vestibular schwannomas: clinical presentation. Neurosurgery 40:1–10. Nashner LM (1976) Adaptating reflexes controlling the human posture. Exp Brain Res 26:59 –72. Nashner LM, Peters JF (1990) Dynamic posturography in the diagnosis and management of dizziness and balance disorders. Neurol Clin 8:331–349. Parietti-Winkler C, Gauchard GC, Simon C, Perrin PP (2006) Sensorimotor postural rearrangement after unilateral vestibular deafferentation in patients with acoustic neuroma. Neurosci Res 55:171– 181. Parietti-Winkler C, Gauchard GC, Simon C, Perrin PP (2008) Visual sensorial preference delays balance control compensation after vestibular schwannoma surgery. J Neurol Neurosurg Psychiatry 79:1287–1294. Parietti-Winkler C, Gauchard GC, Simon C, Perrin PP (2010) Long term effects of vestibular compensation on balance control and sensory organization after unilateral deafferentation due to vestibular schwannoma surgery. J Neurol Neurosurg Psychiatry 81:934 – 936. Pfaltz CR, Ura M, Allum JH, Gratzl O (1991) Diagnosis and surgery of cerebellopontine-angle tumors. ORL J Otorhinolaryngol Relat Spec 53:121–125. Ris L, Capron B, de Waele C, Vidal PP, Godaux E (1997) Dissociations between behavioural recovery and restoration of vestibular activity in the unilabyrinthectomized guinea-pig. J Physiol 500: 509 –522. Wiegand DA, Ojemann RG, Fickel V (1996) Surgical treatment of acoustic neuroma (vestibular schwannoma) in the United States: report from the acoustic neuroma registry. Laryngoscope 106: 58 – 66.

(Accepted 21 October 2010) (Available online 28 October 2010)