Dizziness and balance outcomes after two different postoperative rehabilitation approaches following neck surgery: analyses of a multicenter randomized controlled trial

ABSTRACT Background Dizziness and balance problems are common symptoms in patients with cervical radiculopathy. Objective To evaluate the effect of neck surgery postoperatively combined with either structured rehabilitation or standard approach in patients with cervical radiculopathy and dizziness and/or balance problems, and investigate factors influencing dizziness and balance at 6-month follow-up. Methods Individuals (n = 149) with cervical radiculopathy and dizziness and/or balance problems were randomized preoperatively to structured postoperative rehabilitation or standard postoperative approach. Outcomes were intensity of dizziness and subjective balance, and clinical measures of balance. Results Self-reported measures improved at three months (p ˂ 0.001 to p = .007) and the standing balance at six months (p = .008). No between-group differences. Baseline values, neck pain, and physical activity level explained 23–39% of the variance in 6-month outcomes for self-reported measures. Baseline values and physical activity level explained 71% of the variance in walking balance, and lower baseline scores were significantly associated with standing balance impairments (OR 0.876). Conclusion Patients improved significantly in dizziness and subjective balance intensity shortly after surgery, and in standing balance at 6 months, independent of postoperative rehabilitation. Neck pain, physical activity, and neck muscle function influenced dizziness and balance, although preoperative values and neck pain were of most importance for 6-month outcomes.


Introduction
Patients with cervical radiculopathy (CR) due to degenerative disc disease most often present with arm and neck pain, sensory and motor impairments, self-rated disability, and reduced overall health (Daffner et al., 2003;Henderson, Hennessy, Shuey, and Shackelford, 1983). Surgery via an anterior or posterior approach is a common intervention with the aim to reduce neurological symptoms and pain by decompressing the nerve root. Dizziness, as well as joint position sense and/or balance impairments are frequently found in patients with CR Karlberg, Persson, and Magnusson, 1995;Peng et al., 2018;Wibault et al., 2013), and in other neck disorders (Ruhe, Fejer, and Walker, 2011;Stanton, Leake, Chalmers, and Moseley, 2016;Treleaven, Jull, and Sterling, 2003). Patients with dizziness due to a variety of causes show reduced quality of life (Neuhauser et al., 2008). Individuals with neck pain combined with dizziness report lower quality of life than individuals with only neck pain do (Ferrer-Pena et al., 2019).
The main reason for dizziness due to cervical disorders is thought to be altered afferent signals from the proprioceptors of the neck. Proprioceptors are present in joints, ligaments, and deep cervical muscles (Boyd-Clark, Briggs, and Galea, 2002). Altered proprioception in patients with neck disorders may be caused by inflammation, as well as ischemia, or swelling to cervical structures (Thunberg et al., 2001). Cervical interverte-bral discs are supplied with different types of mechanoreceptors and a higher number of Ruffini corpuscles ingrown into inner annulus fibrosus has been reported (Yang et al., 2017). Increased number of Ruffini corpuscles was found to correlate with the presence of dizziness in patients with cervical degenerative disc disease (Yang et al., 2017). Both inflammation due to disc degeneration, and the increased number of Ruffini corpuscles can lead to erroneous proprioceptive signals and result in dizziness sensations. It can also lead to increased muscle tension and consequently to erroneous proprioceptive signals from muscle spindles (Liu, Liu, and Peng, 2021). Pain from cervical structures may also alter proprioceptor sensitivity thus modulate cervical afferent input (Johansson et al., 2003;Thunberg et al., 2001), and influence body perception at the central level (Haggard, Iannetti, and Longo, 2013). Altered neck muscle function such as muscle fatigue and structural changes can influence proprioception (Röijezon, Clark, and Treleaven, 2015) and postural control (Mitsutake et al., 2016;Stapley, Beretta, Toffola, and Schieppati, 2006) in individuals with neck pain and CR. Neck pain and altered neck muscle function are present in patients with CR (Halvorsen, Abbott, Peolsson, and Dedering, 2014;Henderson, Hennessy, Shuey, and Shackelford, 1983), as well in patients after anterior cervical decompression and fusion (ACDF) surgery (Andresen et al., 2018;Peolsson, Vavruch, and Öberg, 2002;Wibault et al., 2018).
To reduce dizziness and improve balance in patients with neck-disorders caused by possible cervical somatosensory dysfunction, local interventions to reduce neck pain, and improve neck muscle functioning, combined with rehabilitation programs for sensorimotor and postural control has been recommended (Treleaven, 2008). Neckspecific exercises targeting the deep cervical muscles together with a behavioral approach has shown to be effective for reducing dizziness in individuals with whiplash-associated disorders (WAD) (Treleaven et al., 2016).
Improvements of surgical interventions on postural performance in patients with cervico-brachial pain (Persson, Karlberg, and Magnusson, 1996) and dizziness in patients with cervical radiculopathy and/or myelopathy (Peng et al., 2018) were found. In both of these studies, surgery was superior to unspecified conservative treatment. However, evaluations after ACDF show remaining dizziness and impaired balance both short-term (Peolsson, Kammerlind, and Ledin, 2004) and long-term after surgery (Hermansen, Cleland, Kammerlind, and Peolsson, 2014;Hermansen, Peolsson, Hedlund, and Kammerlind, 2020). None of these previous studies evaluated the added benefit of a structured postoperative rehabilitation program, including specific neck muscle exercises and pain management, on dizziness and balance problems. We hypothesize that structured postoperative rehabilitation may affect dizziness and balance by further reducing pain, neck muscle impairment, and may improve the somatosensory component of the postural control system with subsequent improvements in dizziness and balance problems and performance. The aim of the study was to investigate the effect of neck surgery with two different postoperative rehabilitation approaches in patients with cervical radiculopathy focusing on selfreported dizziness and balance problems and clinical measures of balance. A second aim was to investigate factors influencing 6-month outcomes in dizziness and balance.

Methods
This study consists of additional analyses of a multicenter randomized controlled trial evaluating two different postoperative rehabilitation programs for patients with CR. Primary outcomes of pain and neck-related disability have been previously published (Wibault et al., 2018).

Participants
Two-hundred and two patients with CR due to cervical disc disease were randomized at inclusion (preoperatively) to either a structured postoperative rehabilitation program or standard postoperative approach (Figure 1). Patients were recruited from four surgical clinics in the southern parts of Sweden and underwent surgery between February 2009 and November 2012 (Wibault et al., 2018). Randomization was performed by the project leader using a computer-generated list developed by a statistician with an allocation ratio of 1:1. The project leader did not participate in any of the treatments or test-situations. One patient was excluded due to canceled surgery (Wibault et al., 2018).
Inclusion criteria for the main study were individuals between 18 and 70 years of age, with symptoms and clinical findings of cervical radiculopathy, assessed by a neurosurgeon/neuro-orthopedic surgeon, with MRI verified nerve root compression, and scheduled for ACDF or posterior cervical foraminotomy (PCF). Exclusion criteria for the main study were myelopathy, systemic disease or trauma contraindicated to perform the treatment program or clinical tests, known drug use/ abuse, and patients who were not fluent in the Swedish language (Wibault et al., 2018).
To assess the effects of structured rehabilitation on dizziness and balance problems, a sub-group of 149 (74% of the original sample) patients who preoperatively (at baseline) reported dizziness (at rest or movement) and/or self-reported balance problems defined as more than zero on the 100 mm visual analog scale (VAS) were included ( Figure 1). Table 1 presents patient characteristics and background data. This study was approved by the Regional Ethics Review Board in Linköping, Sweden (Dnr.M126-08), and registered at ClinicalTrials.gov (identifier: NCT01547611). Patients provided informed consent prior to inclusion (Wibault et al., 2018). The tests and the rehabilitation programs were expected to do no harm. Slight muscle ache could be experienced as a result of exercises and/or test procedures. When working with balance rehabilitation exercises, a certain amount of dizziness or imbalance is expected as part of the rehabilitation program. Patients not randomized to take part of the neck-specific exercise group received customary postoperative care and rehabilitation.

Surgery and postoperative care
In ACDF, the disc and osteophytes were removed and the segments were fused using the standard cage at each spinal center (Wibault et al., 2018). PCF was performed without fusion and with or without laminectomy. Postoperative care during the first 6 weeks after surgery included advice regarding good posture and ergonomics, activities and movements to avoid, and shoulder mobility exercises. This treatment did not differ between groups. At 6 weeks postoperatively, patients returned for routine visits to the surgeon and a physical therapist. The physical therapist introduced active cervical mobility exercises. No further follow-ups at the spinal clinics were planned after that, and patients were referred to their assigned study treatment (Wibault et al., 2018).

Standard approach
Patients assigned to standard approach did not receive any further scheduled follow-ups or structured rehabilitation within the study. At the time of discharge from the spine center, patients were recommended to contact a primary health-care provider if needed. Thus, these individuals may have received physical therapy treatment but not designed for this specific study (Wibault et al., 2018).

Structured rehabilitation
Patients in the structured postoperative rehabilitation group received physiotherapy in outpatient physiotherapy clinics including exercise therapy for the neck, trunk, and scapula combined with a behavioral approach to pain and stress management. Exercise therapy for the neck aimed to target the deep cervical muscles to improve neuromuscular control and endurance of the neck muscles. The exercise program with dosage and progression is described elsewhere (Wibault et al., 2018). Patients in the structured rehabilitation group who expressed having dizziness or balance problems were instructed in vestibular exercises in addition to the exercise therapy for the neck. Exercises from a standardized vestibular rehabilitation program included dizziness provoking eye and head movements, balance exercises in standing on different surfaces with ACDF = anterior cervical decompression and fusion, CI = confidence interval, NDI = neck disability index, NME = neck muscle endurance, PCF = posterior cervical foraminotomy, Rehab = structured postoperative rehabilitation program, Standard = standard postoperative approach, VAS = visual analogue scale eyes open and closed, and balance exercises in walking, for example, walking on a line with simultaneous head movements.

Data collection
Participants completed ratings and measures of dizziness and balance at baseline, three months and six months after surgery. The researcher collecting the questionnaires and the physiotherapists conducting the clinical tests were blinded for randomization.

Self-reported measures of dizziness and balance problems
Dizziness intensity at rest and during movement, and the intensity of balance problems were rated on a 100mm VAS (0 = no symptoms, 100 = worst imaginable symptoms) (Kammerlind, Bergquist Larsson, Ledin, and Skargren, 2005). Self-reported intensity of dizziness and balance problems using the 100-mm VAS has previously been used in patients with: vestibular dysfunction (Kammerlind, Bergquist Larsson, Ledin, and Skargren, 2005); dizziness in patients with WAD (Treleaven et al., 2016); and cervicogenic dizziness (Reid et al., 2015).

Clinical measures of balance
Standing balance performance was measured using a tandem stance test with eyes closed (Kammerlind, Bergquist Larsson, Ledin, and Skargren, 2005). The test was measured in seconds, and was terminated if participants opened their eyes, reached for additional support, changed foot position, or completed the maximum value of 30 seconds (Kammerlind, Bergquist Larsson, Ledin, and Skargren, 2005). Test scores were dichotomized as normal or impaired according to age-related cutoff values for standing balance (Table 2). Walking balance performance was tested with the "walking in a figure-of-eight" test, with the speed controlled by a metronome (Johansson and Jarnlo, 1991). The number of incorrect steps (i.e. walking on, or outside the drawn lines) was recorded. Three trials of the balance tests were performed, and the mean of all three trials was used. If the participant reached the maximum value of 30 sec or no incorrect steps respectively at the first or second trial, no further trials were performed and the maximum value was assigned to the remaining trials (Kammerlind, Bergquist Larsson, Ledin, and Skargren, 2005).

Complementary outcome measures
Age, gender, levels of surgery, frequency of dizziness, neck pain, and neck-related disability were collected as background data at baseline. Neck-related disability was measured using the neck disability index (NDI) score value of 0-100 (0 = no disability, and 100 = complete disability) (Vernon and Mior, 1991). Neck pain intensity was measured using a 100 mm VAS (0 = no pain, 100 = worst imaginable pain) (Carlsson, 1983).
Physical activity level was rated at baseline and at 6-month follow-up. A 4-point score of activity level (1 = inactivity, 2 = low, 3 = moderate, 4 = high) was calculated using two questions; time spent doing everyday physical activity and time performing exercise and sports (Peolsson et al., 2007). Scores were dichotomized into either low (1-2) or moderate/high levels (3-4) of physical activity.
Clinical measures of neck muscle endurance (NME) for ventral and dorsal neck muscles were assessed in supine and prone, respectively, and measured in seconds. A 2 kg weight for women and a 4 kg weight for men were used when assessing dorsal NME (Peolsson et al., 2007).

Data analyses
An NDI score difference of 10 units was used in sample size estimations in the original RCT, NDI being the primary outcome measure, and 202 patients were recruited (Wibault et al., 2018). The sample size of the present study was determined by the individuals out of these 202 who reported any dizziness at baseline.
Demographics and background variables were calculated using parametric or non-parametric statistics depending on the measures and normality of data. A 3 × 2 mixed design analysis of variance (ANOVA) was conducted to assess the impact of two different interventions on balance and dizziness across three time periods. Greenhouse-Geisser correction was used when the sphericity assumption was not met. Additional repeated contrast analyses were performed if main effects were significant. One hundred and thirteen participants completed self-reported outcome measures and 112-116 participants completed balance tests at all three time-points and were included in the ANOVA. A p-value of ˂ 0.05 was considered statistically significant. Effect sizes were considered small if Cohen's d was >0.2, intermediate if >0.5, and large if >0.8 (Cohen, 1988).
Simple linear regression analyses were performed to evaluate the impact of each independent variable (i.e. intervention group, neck pain, NME ventral and dorsal, level of physical activity, gender, and age) on each dependent variables (i.e. dizziness at rest and during movement, subjective balance problems, and walking in a figure-ofeight) at six-month follow-up. Multiple regression analyses were performed to evaluate the relative influence on outcomes. Results of the multiple regressions were expressed as standardized beta coefficients that exhibited the importance of each independent variable on the 6-month outcomes, and adjusted R 2 to describe how much of the variability in 6-month outcomes are accounted for by the independent variables. Regression analyses were performed by complete case analysis (i.e. all analyses were performed on those individuals who responded to all variables included in the full model (n = 105-108). Multiple regression analyses were adjusted for preoperative values.
Dichotomized variables of the tandem stance test were used in simple and multiple logistic regression analyses to assess the influence of independent variables on standing balance. Dichotomizing was due to the built-in ceiling effect of 30 seconds as maximum value in this test. Results are expressed as Odds ratios (OR) with 95% confidence intervals (CI) and Nagelkerke pseudo R 2 . The beta coefficients and ORs were considered statistically significant if their associated p-values were < 0.05. IBM SPSS Statistics version 26 was used for all analyses.

Results
Descriptive values of the included participants are presented in Table 3. There were no between-group differences at baseline in any outcome measure (p = .117-0.831). Individuals excluded did not differ from the included group regarding age, sex, levels of fusion, randomization, or neck pain. Excluded individuals, without dizziness, had significantly lower baseline neck disability scores (p ˂ 0.001), and included a higher proportion of patients with PCF (p = .041) than did the included individuals with dizziness.

Exercise compliance
Compliance with the rehabilitation program (i.e. participation in at least 50% of the rehabilitation sessions) was 68% with no report of exercise compliance in 5 patients.

Results of self-reported measures over time and between intervention groups
There was no significant interaction effect between group and time for dizziness at rest. There was no significant difference in main effect between the two intervention groups (p = .622). There was a total main effect over time F(1.6, 176.5) = 5.81, p = .007, Cohen´s d = 0.46. Contrast analysis revealed significant improvements from baseline to 3-month follow-up (p = .021), but not between 3 and 6 months (p = .530).
There was no time by group interaction effect for dizziness during movement. No between-group differences were found between the two intervention groups (p = .767). There was a total main effect over time, F(1.8, 200.1) = 19.478, p ˂ 0.001, Cohen's d = 0,84. Contrast analyses showed significant improvements from baseline to 3-month follow-up (p ˂ 0.001), but not between 3 and 6 months (p = .55).
There was no time by group interaction effect for selfreported balance problems. No between-group differences between the two intervention groups were found (p = .937). There was a total main effect over time, F(1.6, 182.4) = 15.17, p ˂ 0.001, Cohen's d = 0.74. Contrast analyses showed significant improvements from baseline to 3-month follow-up (p ˂ 0.001), but not between 3 and 6 months (p = .72).

Results of clinical measures of balance, over time and between intervention groups
No between-group differences were found in the clinical balance measures (p = .611, standing and p = .753, walking). There were no time by group interaction effects in either clinical measure. There was a total main effect over time in standing balance, F(1.8, 202.6) = 5.2, p = .008, Cohen's d = 0.43. Contrast analyses showed that improvements were not significant at 3 months (p = .057) or between 3 and 6 months (p = .141). There was no significant main effect over time in walking balance (p = .27).

Variables influencing 6-month outcomes in dizziness and balance
Simple regression analyses showed that neck pain influenced both dizziness at rest and during movement, and balance problems. In addition, lower NME scores of dorsal and ventral neck muscles increased dizziness during movement (Table 4). Adjusted multiple-regression analyses showed that neck pain, lower physical activity levels, and preoperative values of dizziness at rest explained 23% (Adj. R 2 0.23, p ˂ 0.001) of the variance in dizziness at rest at 6 month follow-up. Neck pain in combination with preoperative values of each respective dependent variable explained 29% (Adj. R 2 0.29, p ˂ 0.001) and 39% (Adj. R 2 0.39, p ˂ 0.001) of the variance of dizziness during movement and balance problems at 6 months (Table 4).
An increase in neck pain resulted in higher odds of being impaired in standing balance (p = .047) in the simple logistic regression analyses. Preoperative standing balance score was the only variable that influenced the outcome of the multiple logistic regression. Increased scores resulted in lower odds at being impaired in standing balance at 6-month follow-up (OR 0.876, p ˂0.001) ( Table 5). Simple regression analyses showed that being a man, increasing age, and having lower dorsal NME were related to worse walking balance. In the multiple regression analysis, level of physical activity and preoperative balance scores explained 71% of the variance in walking balance at 6 months (Table 5).

DISCUSSION
This study is unique in evaluating the added benefit of a structured postoperative rehabilitation program on dizziness and balance problems in patients after cervical decompression surgery. The main results of this study were that subjective ratings of dizziness and balance problems along with standing balance significantly improved over time with moderate to large effect sizes, but without differences between postoperative rehabilitation groups over time. Patients improved significantly between baseline and 3-month follow-up in selfreported outcomes. Standing balance showed significant improvements at 6-month follow-up. Neck-pain and preoperative values significantly influenced all three selfreported dizziness and balance outcomes at 6 months. Preoperative values also significantly influenced outcome of both standing and walking balance. Group allocation did not influence any of the models.
The results on self-reported measures with significant improvements at three months, without further improvements after that and without between-group differences, are similar to previously presented results on pain, disability, and clinical measures of neck-related function (Wibault et al., 2018(Wibault et al., , 2017. These results indicate that the early effects on subjective dizziness and balance might be due to surgery alone. However, the lack of differences in dizziness and balance between groups with different postoperative rehabilitation approaches may be more complex. In the present study, patients of both groups received surgery, as well as shoulder mobility exercises before discharge and neck Table 3. Descriptive values of self-reported and clinical measures of dizziness and balance at 3 time-points and for the 2 intervention groups. 8.0 7.7 8.5 (95% CI) (6.7-10.5) (6.3-12.5) (5.5-10.1) (6.7-10.6) (6.4-12.1) (5.3-10.8) (6.0-10.0) (4.9-10.4) (5.4-11.5) CI = confidence interval, Rehab = structured postoperative rehabilitation program, Standard = standard postoperative approach, VAS = visual analogue scale, § Standing balances as measured with tandem stance test (eyes closed), # Walking balance measures with walking in a figure-of-eight mobility exercises at six weeks as part of routine postoperative care. As similar outcomes were seen in both groups, the results indicate that reduced pain and improved functioning following surgery and an early exercise regimen (Wibault et al., 2018(Wibault et al., , 2017 might be enough to improve the subjective ratings of dizziness and balance. Other studies on patients with CR showed positive results of surgery, without specific postoperative rehabilitation, on postural performance (Persson, Karlberg, and Magnusson, 1996), and dizziness intensity and frequency (Peng et al., 2018). Both studies evaluated the effects of surgery compared to conservative treatment, that is, physiotherapy and/or cervical collar, with superior effect of surgery on postural control evaluated with posturography (Persson, Karlberg, and Magnusson, 1996) and self-reported dizziness outcomes (Peng et al., 2018). Significant improvements were seen after 3 months (Persson, Karlberg, and Magnusson, 1996) and immediately at 1-month follow-up (Peng et al., 2018) strengthening the implication that early improvements in the present study were due mainly to the surgical intervention. Some improvements from the neck-specific exercise program could have been expected based on previous studies Treleaven et al., 2016). A similar neck muscle exercise program to the one in the present study showed a few reductions in dizziness intensity and dizziness-related disability in patients with WAD, although this was combined with a behavioral approach (Treleaven et al., 2016). One other study on patients with neck pain showed improvements in postural performance, evaluated with posturography, from physiotherapy treatment of the neck problems without balance or vestibular exercises  although in the study by Karlberg et al. (1996) patients had a higher dizziness intensity rating than in the present study and patient selection was based on patient's cervicogenic dizziness classification.
However, at three-month follow-up, the ratings of dizziness, particularly at rest but also during movement and of balance problems, were largely reduced and therefore low. On a scale, such as the VAS, low scores provide less opportunity to improve further. In addition, approximately two-thirds of patients in the structured rehabilitation group completed 50% or more of the intervention sessions, which may have reduced the impact of rehabilitation on outcomes. Balance performance differed from the self-reported outcomes showing significant changes in standing balance at 6-month follow-up and no main effect over time for walking balance. A possible explanation might be that when the stress on spinal joints and pain from the neck reduced or diminished after surgery, normal signals were transported from the somatosensory receptors, reducing dizziness and the subjective feeling of being off balance. Balance impairments, however, might need some training to challenge the balance system before measurable improvements can occur.
Neck pain influenced all self-reported outcome measures at six months in both univariate analyses and multiple-regression analyses. These findings are concordant with the theory implicating pain as one factor in cervicogenic dizziness (Johansson et al., 2003;Thunberg et al., 2001). Previous studies on CR patients after surgery have shown correlations between neck pain and dizziness at rest and during movement (Hermansen, Peolsson, Hedlund, and Kammerlind, 2020), and some associations between neck pain and postural performance measured with posturography (Peolsson, Kammerlind, and Ledin, 2004;Persson, Karlberg, and Magnusson, 1996). In addition, the higher neck disability scores at baseline in the included individuals with dizziness compared to the excluded Adj. R 2 = adjusted R-squared, β = standardized beta coefficients, NME = neck muscle endurance, § Multiple regression models are adjusted for variable-specific preoperative values, * p-value ˂.05 is considered significant individuals without dizziness strengthen the presumed relationship between neck problems and dizziness in this population.
In the simple regression analyses, dorsal and ventral NME had a significant negative association with dizziness during movement, and NME dorsal to walking balance, but in the multiple models, these values were not significant. In daily activities, there is high demands on neck stabilizing muscles possibly causing fatigue in these muscles, which could explain the findings of dizziness during movement. When walking in a "figure-ofeight," the participant must look down on the floor, which may be the reason dorsal NME influence that test.
Age and gender affected the walking balance measure in the simple regression analysis. Higher age is correlated with impaired balance (Osoba, Rao, Agrawal, and Lalwani, 2019), and men were expected to have more incorrect steps because they generally have bigger feet and therefore it is harder to step within the lines in the test. However, physical activity and baseline balance scores were the only variables significantly affecting walking balance in the multiple model.
Multiple regression analyses showed that the included explanatory variables explained a rather small amount of the variance in self-reported outcomes (23-39%) at 6-months, and the strongest, or only, variables to influence 6-month scores were their baseline values. This proposes that factors other than those included in the study could be of greater importance for these outcomes. Such factors could be cervical range of motion (cROM), or psychosocial variables. Studies of patients with cervicogenic dizziness have shown that they have less cROM and worse psychosocial scores than healthy individuals (Micarelli et al., 2021). In addition, cROM correlated with scores on the dizziness handicap inventory (DHI), when DHI scores were classified as low, and psychosocial ratings were correlated with DHI scores when these were high (Cuenca-Martinez et al., 2018). Dizziness and balance ratings correlated with reduced health-related quality of life in patients long term after ACDF surgery (Hermansen, Peolsson, Hedlund, and Kammerlind, 2020).
Patients in the present study were included due to their arm and neck pain and neurological symptoms of CR (Wibault et al., 2018), and not their dizziness; therefore, one limitation of the present study is that we do not know if causes other than from the neck might be responsible for their dizziness or balance symptoms. Patients with cervical myelopathy were, however, not included in the original study sample which eliminates the impact of, for example, reduced somatosensory reception from the feet on dizziness and balance problems. Eye movements in all movement planes were performed as part of the initial exercises to activate the deep cervical muscles (Wibault et al., 2018); however, no further specialized sensorimotor exercises were included in the rehabilitation program. Complementary vestibular rehabilitation exercises were recommended for individuals with dizziness or balance problems at baseline. A limitation of this study was the lack of registration of participants in the intervention group who received these complementary exercises. Sensorimotor exercises, together with local treatments aimed to reduce neck pain and restore function, and vestibular rehabilitation, have been recommended to treat cervicogenic dizziness (Treleaven, 2008). In future studies, it would be interesting to evaluate the effect of a rehabilitation program specifically combining these three exercise types on patients who presented with remaining dizziness and balance problems after neck surgery due to cervical radiculopathy. Adj. R 2 = adjusted R-squared, β = standardized beta coefficients, NME = neck muscle endurance, OR (95% CI) = Odds ratio (95% confidence interval), § Multiple regression models are adjusted for variable-specific preoperative values, # Standing balance test results dichotomized as impaired or not impaired based on age-related reference values, age is therefore excluded as an explanatory variable in the logistic regression analyses, & Nagelkerke pseudo R-squared, * p-value ˂.05 is considered significant

Clinical implications
Balance and dizziness are important factors to focus on during rehabilitation after neck surgery. Improvements seem to be slower in balance outcomes compared to selfreported outcomes demanding longer follow-ups in clinical practice. Neck pain intensity and physical activity levels are related to dizziness and balance outcomes after neck surgery.

Conclusion
Patients with cervical radiculopathy and dizziness improved significantly in dizziness intensity and subjective balance intensity shortly after surgery independent of participating in neck-specific rehabilitation or receiving standard postoperative rehabilitation regimen. Standing balance significantly improved over time in both rehabilitation groups, but not until 6 months after surgery. Neck pain, physical activity, and function of neck muscles were factors that influenced dizziness and balance outcomes, although preoperative values were the factors of most importance for 6-month outcomes.

Disclosure statement
No potential conflict of interest was reported by the author(s).