Incidence of C5 palsy in anterior cervical decompression & fusion, posterior cervical decompression & fusion and laminoplasty for degenerative cervical myelopathy: systematic review and meta-analysis of 21,231 cases

Literature search

A comprehensive literature search was conducted on October 15, 2024, across five databases (Google Scholar, Embase, PubMed, Web of Science, and Cochrane Library). Studies published through 2024 that reported on the incidence of C5P in patients undergoing cervical decompression were reviewed. The search utilized keywords combined with Boolean operators: (((ACDF) OR (cervical fusion) OR (laminoplasty) OR (circumferential fusion) OR (posterior decompression)) AND (C5 palsy)).

Inclusion and exclusion criteria

Studies reporting on the occurrence, management, or recovery of C5P following decompressive surgery for degenerative cervical myelopathy were considered. Narrative or systematic reviews, opinions, letters to the editor, and non-English publications were excluded. The article selection criteria are summarized in Table 1.

Table 1.

Inclusion and exclusion criteria of selection of articles to be included in the review

Manuscript selection and data extraction

The search results were downloaded, imported into EndNote (Clarivate, Philadelphia, PA, USA), and deduplicated. Two authors independently screened titles, followed by a separate screening of manuscripts based on the aforementioned criteria. Full-text manuscripts were then reviewed in duplicate, and final article selection was completed. For studies with overlapping patient cohorts from the same center, only the most relevant sample was included. Discrepancies among the authors were resolved through discussion with the senior author.

Quality assessment

The methodological quality of the included studies was assessed using the Newcastle Ottawa Scale for non-randomized studies (Table 2). Only studies with adequate methodological quality were included in the analysis.

Table 2.

Quality assessment of the included studies

Statistical analysis

The meta-analysis was conducted using Stata ver. 16.0 software (Stata Corp., College Station, TX, USA). The reported incidence of C5P was pooled, and the mean incidence was calculated with a 95% confidence interval (CI). A random-effects model was used for data synthesis when heterogeneity was high (I²>50% and p<0.10). Otherwise, a fixed-effects model was applied. Sensitivity and subgroup analyses were performed when heterogeneity was detected. Publication bias was assessed using Egger’s regression test. A p-value less than 0.05 was considered indicative of statistical significance for all statistical analyses.

Etiopathogenesis of C5P

Wang et al. [17,21] described C5P as a disorder or insult to the nerve root or segmental spinal cord. Five major etiopathogenic mechanisms were suggested, namely inadvertent nerve root injury, nerve root traction secondary to cord shift, segmental cord disorder, spinal cord ischemia, and reperfusion cord injury [12,17,21-23]. The main purported hypotheses underlying the higher prevalence of C5P in posterior decompressive surgeries include the “tethering phenomenon” (or posterior shift of spinal cord leading to C5 nerve root traction) and spinal cord expansion secondary to acute canal decompression [17,24].

The C5 nerve roots are particularly vulnerable due to their short length, relatively less laxity, location at the apex of cervical lordosis (which exposes them to the highest strain during spinal cord shift), proximity to the C5 superior articular process (which is more anteriorly protruded), and horizontal orientation [4,25]. These factors increase their risk of injury from traction or stretch compared to the other subaxial roots [10,13,26- 29]. Bydon et al. [10] found a 3-fold higher posterior cord shift at C4–5 level in patients with C5P, limiting the nerve root’s accommodation to the shift. Pennington et al. [13] demonstrated an axonotmesis-like nerve injury in these palsies, which explains the mean 6- to 12-month recovery period for weakness.

Incidence of C5P

The pooled incidence of C5P from 97 studies (21,231 patients) was 7% (95% CI, 4%–10%) (Fig. 2), analyzed using a random-effects model due to heterogeneity across studies.

Fig. 2.

Forest plot showing the pooled incidence of C5 palsy following surgery for cervical myelopathy. ACDF, anterior cervical decompression and fusion; PCDF, posterior cervical decompression and fusion; LP, laminoplasty; CI, confidence interval; REML, restricted maximum likelihood.

Subgroup analysis

A subgroup analysis was performed based on the surgical procedure, including ACDF, posterior cervical decompression and fusion (PCDF), LP, and combined anterior and posterior fusion (circumferential fusion). Upon subgroup analysis, the incidence rates were 4% (95% CI, 3%–5%) for ACDF (Fig. 3), 6% (95% CI, 5%–7%) with LP (Fig. 4), 10% (95% CI, 8%–12%) for PCDF (or laminectomy and fusion [LF]) (Fig. 5), and 16% (95% CI, 8%–24%) for circumferential fusion (Fig. 6).

Fig. 3.

Forest plot showing the incidence of C5 palsy following anterior cervical decompression and fusion surgery for cervical myelopathy. ES, effect size; CI, confidence interval.

Fig. 4.

Forest plot showing the incidence of C5 palsy following laminoplasty surgery for cervical myelopathy. ES, effect size; CI, confidence interval.

Fig. 5.

Forest plot showing the incidence of C5 palsy following posterior decompression and fusion surgery for cervical myelopathy. ES, effect size; CI, confidence interval.

Fig. 6.

Forest plot showing the incidence of C5 palsy following global decompression and fusion surgery for cervical myelopathy. ES, effect size; CI, confidence interval.

Risk factors for C5P

Various studies have examined the risk factors for C5P in patients undergoing cervical spinal surgeries [16,30,31]. Pennington et al. [13] found that mean foraminal diameter (p<0.001) and gait disturbance at the initial consultation (p=0.008) were significant predictors of postoperative C5P. A preoperative foraminal diameter <2 mm was associated with a twofold higher incidence of C5P. In addition, preoperative symptoms such as tingling or numbness (p=0.08) and inadvertent dropping of objects (p=0.09) tended toward significance, but were not statistically significant in multivariate analysis. In the study by Wang et al. [17,21], the C5P rate was substantially higher in males (5.9% vs. 4.1%). They also found unilateral involvement to be substantially more common.

Several radiological parameters have been purported to predict C5P with a high specificity and sensitivity, including reduced anteroposterior canal dimension, decreased intervertebral foraminal size, and high cervical lordosis angle (open angle of lamina) [32-34]. Lubelski et al. [35] studied 98 patients undergoing anterior or posterior decompression at the C4–5 level and found a significant impact of the severity of preoperative C4–5 foraminal stenosis on the occurrence of postsurgical C5P. Notably, each 1 mm increase in foraminal size reduced the risk of developing palsy by 50-fold.

Other potential risk factors include inadequate restoration of cervical lordosis or alignment, severity of myelopathy (measured by Japanese Orthopaedic Association scores) and high-intensity spinal cord signal on T2-weighted image. However, the impact of these factors remains debated [36-39]. Egger’s regression test showed no significant publication bias among the included studies (p=0.521).

Definitions of C5P

There is a significant variation in the definitions for C5P across studies, which can potentially impact the reported incidence rates. Imagama et al. [4] defined C5P as a postoperative deltoid muscle strength of 0–2 on the MMT, with or without weakness of biceps brachii, while other muscles remain unaffected. Nakashima et al. [43] defined it as a ≥1 grade deterioration in deltoid muscle on MMT, with or without biceps brachii involvement. Nassr et al. [14] described C5P as loss of motor power in the deltoid and/or biceps, sensory impairment in the C5 dermatome, and increased pain in the C5 distribution. Nakamae et al. [44] defined C5P as a palsy in the deltoid and biceps musculature with a ≥1 grade decrease in MMT without sensory impairment.

Incidence of C5P

In the study by Shou et al. [45], the mean incidence of C5P was 5.3% (95% CI, 4.6%–6.0%), while the incidence varied between 0 and 30% in the study by Sakaura et al. [12] In the meta-analysis by Wang et al. [17], the pooled mean incidence of C5P was 6.3% (95% CI, 5.7%–7.9%), with the mean rate ranging between 1% and 29% across the reviewed studies. The meta-analysis by Wang et al. [17,21] found C5P incidence rates of 6.2% for posterior cervical approaches and 5% for anterior cervical spinal approaches. Most studies suggest a relatively higher incidence of C5P with posterior cervical procedures. In the increasing order of frequency, the reported incidence of C5P in patients undergoing LP, ACDF, anterior cervical corpectomy with discectomy and fusion (ACCDF)–hybrid procedure, anterior cervical corpectomy and fusion (ACCF), and LF were 4.4%, 5.5%, 6%, 7.5% and 13%, respectively. In our meta-analysis, the overall pooled mean incidence of C5P was 7% (95% CI, 4%–10%).

Underlying pathological diagnosis and C5F

Wang et al. [17,21] compared the incidence of C5P across diverse cervical degenerative pathologies. They found a significantly higher incidence of C5P in patients with ossified posterior longitudinal ligament (OPLL) at 8.1%, compared to those with cervical spondylomyelopathy (CSM) at 4.8%. In their study, patients with OPLL had higher C5P rates than those with CSM when undergoing ACDF (5.5% vs. 4.7%) and LP (8.1% vs. 3.1%), but similar rates when undergoing LF (13% vs. 13.1%) [17,25]. The higher incidence of C5P in OPLL patients has been attributed to the greater spinal cord shifting, resulting in a stronger tethering effect on the nerve roots [16]. A meta-analysis by Shou et al. [45] involving 13,621 patients from 79 studies, identified male sex and LF as key risk factors for C5P. Another meta-analysis identified male sex, preoperative intervertebral foraminal stenosis, underlying OPLL, laminectomy, and excessive spinal cord shift as significant risk factors for C5P in patients undergoing posterior cervical procedures [25].

Surgical approaches and C5P

The multicenter, prospective, randomized phase-3 CSM-protect trial found that the posterior approach was independently associated with at least 4-fold greater risk of developing postoperative C5P compared to the anterior approach [24]. In our study, subgroup analysis showed ACDF had the lowest incidence of C5P (4%; 95% CI, 3%–5%), while circumferential fusion (combined anterior-posterior) carried the highest risk for C5P (16%; 95% CI, 8%–24%]). In a meta-analysis by Takase et al. [8,11] involving patients who underwent 3 to 6 level anterior cervical decompression, the incidence of late C5P was 4.2%. The incidence rates following ACDF, ACCDF-hybrid, and ACCF procedures were 2.3%, 3.9%, and 6.4%, respectively. Based on network meta-analysis, they concluded that ACCF was associated with a substantially higher risk of C5P compared to ACDF, recommending ACDF as the preferred option for long-segment cervical fusions when feasible. Broadly, our study also demonstrated relatively higher rates of C5P following posterior decompressive surgeries, compared to anterior approaches.

The meta-analysis by Wang et al. [17,21] found that LF was associated with a significantly higher incidence of C5P (p=0.004) and higher Neck Disability Index (p<0.001) compared to LP. Based on a single-surgeon experience, Kim et al. [3] concluded that LP was associated with a significantly lower incidence of C5P compared to LF (p=0.01) [7]. However, analysis of data from the National Registry showed no significant difference between LP and LF in this respect, suggesting that surgeon experience and training may impact complication rates [46]. Our study also found relatively higher rates of C5P following LF or PDCF, in comparison with LP (6% [95% CI, 5%–7%] for LP vs. 10% [95% CI, 8%–12%] for PDCF or LF).

The types of LP described include cervical expansile open-door LP (EOLP), Hirabayashi, Z-type, double-door, and spinous-process splitting-type LP. A large-scale study by Levy et al. [47] on EOLP found C5P to be the most common complication, occurring in 2.6% of patients. Notably, C5P in LP was more common on the open-door side of the lamina compared to the hinge side, likely due to the potentially greater tethering effect on the open side. They recommended creating the hinge on the patient’s dominant arm side to minimize the risk of loss of dominant arm function. A systematic review by Gu et al. [25] found that open-door LP had a higher incidence of C5P, compared to double-door LP, likely due to cord rotation from asymmetric spinal decompression, resulting in greater tethering on the open side [4,5,25,48]. On the contrary, a large-scale, multicenter study by Imagama et al. [4] found no significant difference in C5P rates between open-door and double-door LP.

C4–5 foraminotomy and its role in C5P

Sasai et al. [49] identified the presence of preoperative C5 radiculopathy (which may correlate with the degree of C4–5 foraminal stenosis) as a crucial predictor of palsy. In this context, the impact of C4–5 foraminotomy on postoperative C5P has been widely debated. Studies by Katsumi et al. [50], Komagata et al. [51], and Sasai et al. [49] supported the role of prophylactic C4–5 foraminotomy in reducing the risk of C5P, attributing the benefit to the reduction in the nerve root anchorage at the site of intervertebral foramen. However, other studies have demonstrated significantly increased C5P rates after intraoperative manipulation of the nerve root and C4–5 foraminotomy. Pennington et al. [13] suggested that the C5 nerve root’s susceptibility to injury during decompression may be due to the small size of the C4–5 foramen and the horizontal orientation of the C5 nerve root, which can put it in line with the long axis of the Kerrison’s rongeur during decompression, increasing the risk of impingement by the instrument [52].

Predicting the outcome following C5P

In the study by Lubelski et al. [53], deltoid strength improvement was found to predict recovery from C5P. They observed that patients who experienced complete (overall 60%) or partial recovery (29% of cases) demonstrated improvement in motor power by at least one Medical Research Council (MRC) grade at approximately 6 weeks following the deficit. They underlined the significance of examination at 6 weeks’ time point to predict any meaningful recovery, since the motor power of grade 4/5 or greater at 6 weeks was predictive of complete recovery. On the other hand, patients who gained strength of 3/5 or below in the antigravity muscles alone exhibited only partial recovery. In short, the cohort with little or no recovery within 6 weeks of the onset of C5P was unlikely to experience good long-term outcomes. In addition, electrophysiological testing, such as electromyography, can help identify patients with postoperative C5P who are unlikely to experience meaningful recovery.

A study by Saadeh et al. [54] found that among patients with severe C5P (defined by antigravity strength of MRC grade ≤2) at 3 months postoperatively, 50% regained useful strength by 12 months. However, for those with persistently severe C5P at 6 months, only 25% recovered sufficient strength by 12 months. Patients with MRC grades 0 or 1 at 6 months did not regain useful strength by 1 year. In their series, female sex was associated with good recovery, while those with diabetes mellitus had significantly poorer outcomes.

Lim et al. [2] investigated 36 patients who developed C5P following cervical decompression surgery. They found that 50% of patients (among whom, 91.7% recovered between 6 months and 2 years, while 8.3% did not recover until 2 years) required longer than 6 months to experience useful neurological recovery. The factors associated with longer recovery (>6 months) included motor grade ≤2 (p<0.001), multi-segment paresis involving segments apart from C5 root (p=0.002), extent of posterior spinal cord shift (p=0.04), and the absence of somatic sensation with pain (p=0.008).

Pennington et al. [13,55] found that patients who underwent C4–5 foraminotomy were more likely to develop permanent C5 deficit (p=0.004). Other radiological parameters, such as mean cord-lamina angle and length of laminectomy, showed trends toward significance (p=0.06 and p=0.08, respectively) but did not reach statistical significance.

Factors underlying delayed or compromised neurological recovery

Hashimoto et al. [56] suggested that pre-existing, asymptomatic damage to the anterior horn cells in the spinal cord gray matter may contribute to the development of severe postoperative C5P. Severe C5 palsies are known to have slower and poorer neurological recovery. Additionally, multilevel paresis has been associated with poor outcomes, potentially due to focal reperfusion injury to the spinal cord following spinal decompression [38]. Similarly, significant sensory involvement (>50%) and intractable pain may indicate severe spinal cord injury or ischemia/reperfusion injury, which are associated with poor prognosis in patients with C5P.

Limitations

Our study has limitations, including the exclusion of non-English publications, which may have led to missed relevant articles. Additionally, the predominantly retrospective nature of the included studies reduces the level of evidence. However, given the focus on complications from surgical interventions, retrospective case-control studies are often the primary source of data, limiting the availability of higher-quality evidence from randomized controlled trials. The study’s results may be influenced by the overlapping nature of some of the reviewed parameters, such as circumferential fusion being usually performed on patients with more severe pathologies. Additionally, certain factors like aggressive shoulder retraction or shoulder taping during positioning, which could impact outcomes, were beyond the purview of this review.