C5 nerve palsy is a common complication after cervical decompression with an incidence of 0%-30%^[1]. The symptoms of C5 nerve palsy, including deltoid and biceps paralysis, sensory loss, and shoulder weak intractable pain, are usually unilateral, and bilateral symptoms are rare ^{[2, 3]}.

The etiology of C5 nerve palsy remains unclear, but iatrogenic injuries of the spinal cord during cervical surgery (such as spinal cord displacement and ischemiareperfusion injury) are often thought to contribute to its occurrence. C5 nerve palsy can occur in different manners during decompression surgeries^[4-6], but so far no studies have been reported to compare its incidence after multi-segmental cervical decompression through the anterior, posterior, and combined approaches ^[7].

The purpose of this study was to elucidate whether the incidence of postoperative cervical nerve root palsy differed after 4 different surgical approaches for multi-segmental cervical decompression, namely (1) anterior corpectomy and fusion fixation, (2) anterior corpectomy and fusion fixation + posterior laminectomy, fusion and internal fixation, (3) posterior laminectomy, fusion and internal fixation, and (4) posterior laminoplasty fusion and internal fixation.

We retrospectively analyzed the clinical data of 375 patients who underwent cervical decompression spine surgery between January, 2006 and December, 2012. These patients received multilevel anterior cervical corpectomy and autograft fibular strut fusion, anterior corpectomy and fusion followed by posterior fusion, posterior laminectomy and fusion, or posterior laminoplasty procedures for treatment of cervicai spinal stennosis. All the patients had at least two cervical segments decompressed. The patients were excluded from the study who had no adequate follow-up data, received a previous spinal cord injury, were unable to exercise preoperative or postoperative motor function testing, or received a decompression not involving the C5 level.

The surgical approaches for decompression (anterior, posterior, or combined) were determined by experienced spinal surgeons. Four approaches for multi-segmental cervical decompression were used fortreatment of cervical stenosis, including (1) anterior corpectomy and fusion fixation, (2) anterior corpectomy and fusion fixation + posterior laminectomy, fusion, and internal fixation, and (3) posterior laminectomy, fusion, and internal fixation, and (4) posterior laminoplasty fusion fixation. In this study, C5 nerve palsy is defined as deltoid and/or biceps muscle weakness, C5 dermatome area hypoesthesia, or increased pain in the C5 distribution as compared with the preoperative status. The residual symptoms are defined as persistent symptoms (failure of full motor strength recovery in either the deltoid or biceps brachii, sensory deficits in the C5-innervated area, or increased pain in the C5-innervated area as compared with the preoperative pain intensity) after the end of the follow-up period. The patients' data including age, gender, preoperative physical condition, diabetes, smoking, and number of decompressed levels were collected and statistically analyzed to identify the risk factors contributing to C5 nerve palsy following the decompression surgeries.

The data are presented as Mean±SD and analyzed using Pearson χ² test, Fisher's exact test, and two-sample t test as appropriate. A P value less than 0.05 were considered to denote a statistically significant difference. All the statistical analyses were performed using SPSS 13.0 statistical software.

Of the 375 patients, 60 patients were excluded due to failure of regular follow-up (34 cases), a reduced scope of the C5 level (15 cases), or injury prior to the surgery or carotid C5 nerve palsy associated with spinal cord injury or shoulder discomfort (11 cases). In the remaining 315 patients, 128 patients were treated with anterior corpectomy, 77 had combined anterior treatment, 58 underwent posterior laminectomy decompression, and 52 received posterior lamina-forming treatment.

These 315 patients, including 146 women and 169 men with a mean age of 57.65±15.74 years (range 39-72 years), were followed up for a mean of 10 months. The detailed data of each group are shown in Tab. 1.

表 1(Table 1)

Table 1 Summary of demographic data and surgical outcomes of the patients available for analysis Outcome Corpectomy Corpetomy with posterior Laminoplasty Laminectomy and fusion Overall Patients 128 77 52 58 315 Male/Female 66/62 34/43 35/17 34/24 169/146 Age (year) 54.72±11.83 57.91±10.34 60.03±9.12 63.32±10.45 57.65±15.74 Incidence of C5 palsy 6 (4.69%) 6 (7.79%) 2 (3.85%) 5 (8.62%) 19 (6.03%) Symptom onset time (week) 2.51±0.49 4.23±1.77 3.54±2.33 2.12±0.33 2.59±2.74 Residual deficit 1(16.67%) 1(16.67%) 0(0%) 1(20%) 3(15.79%) Time to maximal improvement (week) 18.11±2.13 28.51±3.12 21.82±1.23 18.31±2.31 20.86±2.52

Table 1 Summary of demographic data and surgical outcomes of the patients available for analysis

Tab. 1 shows no significant difference among the 4 groups in the incidence of C5 nerve root palsy (P=0.28). To detect the difference in the incidences ranging from 3.85% to 8.62% with an 80% power Pearson χ² test, at least 256 patients shall be needed in each group, but the power of our test was 18%.

Five (3.42%) of the 146 women and 14 (8.28%) of the 169 men developed C5 nerve palsy (P < 0.05). Two of these 19 patients (10.52%) had bilateral symptoms, while 1 (5.26%) had paralysis involving more than one segment. The time of initial onset of the C5 nerve root palsy symptoms ranged from immediately following the surgery to 8 weeks after the surgery, and the symptoms completely resolved in the majority of patients within 24 weeks. But 15.79% of the patients still had residual pain after the follow-up period ended. The laminoplasty group had the lowest rate (0) of residual symptoms, while the posterior laminectomy and fusion fixation group had a rate up to 20%, but no significant difference was found in the rates between the latter two groups. Cervical nerve root palsy symptoms disappeared as soon as 1 week postoperatively or until as long as 2 years after the onset, with a mean of 21 weeks; the symptoms disappeared within 6 months in 71.4% of the patients. The mean age of the patients with postoperative cervical nerve root palsy was 57.6 years, similar to that of patients without C5 nerve palsy (58.6 years, P=0.558).

Tab. 2 shows the possible risk factors of C5 nerve palsy associated with the 4 surgical procedures. In the anterior and posterior fusion group and the laminectomy and fusion group, none of the factors tested appeared to contribute to C5 nerve palsy. In the anterior corpectomy group, the number of levels decompressed was significantly greater in the 6 patients with C5 nerve palsy (P=0.002), among whom 5 received 3-level cervical corpectomy. Age was also a risk factor in this subgroup, and the symptomatic patients had a significantly greater age than those without symptoms (62.43±10.31 vs 54.32±13.38 years, P=0.01). The other variables listed were not statistically significant in the corpectomy group. In the posterior laminoplasty group, age was also a significant factor related to the development of C5 nerve palsy, and the mean age of patients developing C5 nerve palsy was 47 years compared to 60 years in the asymptomatic patients (P=0.04).

表 2(Table 2)

Table 2 Summary of subgroup statistical analysis by potential risk factor based on surgery type (Mean±SD or number of cases) Risk Factor Palsy Corpectomy Anterior/Posterior Fusion Laminoplasty Laminectomy and fusion Age (year) Yes 62.43±10.31 58.47±10.13 47.13±8.21 55.56±14.75 No 54.32±13.38 55.42±11.34 60.34±10.33 56.75±15.47 P 0.01* 0.44 0.04* 0.14 Gender ratio (M/F) Yes 3/3 3/4 1/1 3/2 No 63/59 31/39 34/16 31/22 P 0.37 0.26 0.15 0.19 Smoking Yes 3 4 1 4 No 61 46 25 42 P 0.32 0.24 0.35 0.57 Previous cervical surgery Yes 0 1 0 0 No 2 3 1 1 P 0.40 0.31 1.00 0.26 Previous upper extremity surgery Yes 1 1 0 1 No 7 5 3 2 P 0.50 0.11 0.38 1.00 Diabetes Yes 3 2 1 2 No 22 10 15 13 P 0.35 0.39 0.57 0.49 Preoperative weakness Yes 1 0 0 1 No 6 2 2 4 P 1.00 0.50 0.35 0.40 Number of levels decompressed Yes 3.01±1.11 1.52±0.51 2.13±1.04 2.52±1.49 No 1.23±0.49 1.49±0.62 2.21±1.13 2.63±1.43 P 0.002** 0.27 0.41 0.22 The data were compared using Pearson χ2 test or two-sample t test in each group.

Table 2 Summary of subgroup statistical analysis by potential risk factor based on surgery type (Mean±SD or number of cases)

In this study, we found that the overall incidence of C5 nerve palsy following cervical decompression of multiple adjacent segments was 6.03%, close to previously reported results^[8]. No significant difference was found in the incidences of C5 nerve palsy among patients receiving different surgical procedures. According to our data, there was no significant difference in the risk of developing postoperative C5 nerve palsy based on the type of surgery.

C5 nerve palsy occurs commonly within the first 2 weeks after the surgery, and the symptoms disappear mostly within 24 weeks. This study revealed that this complication could occur up to 2 months postoperatively; 15.79% (3/19) of the patients with C5 nerve palsy still had residual symptoms at the end of the follow-up period and were considered to have residual deficits. These symptoms improved within 21 weeks in most of the patients (84.21%).

The risk factors for postoperative C5 nerve palsy included the male gender and the number of corpectomy levels. Men were more likely to develop C5 nerve palsy than women (8.28% vs 3.42%), and an increasing number of corpectomy levels was associated with a significant increase in C5 nerve palsy rates (P < 0.05). In the subgroup analysis, age was identified as a significant risk factor in the corpectomy and laminoplasty groups. In the corpectomy group, an older age was associated with a higher incidence of C5 nerve palsy, but in the laminoplasty group, the patients with C5 nerve palsy had a younger age than those free of symptoms. This discrepancy was most probably incidental due to the relatively small number of patients in each subgroup. In the overall patients, no association was found between the patients' age and the development of C5 nerve palsy.

Researchers previously thought that intraoperative electromyographic (EMG) monitoring was useful for detecting C5 nerve paralysis during surgery ^{[9, 10]}. Intraoperative EMG is sensitive in detecting motorevoked potentials (MEP), but not all C5 nerve palsy can be found by intraoperative EMG monitoring ^[11-13], although it does help to understand why some of the patients' symptoms disappear soon after surgery^[14]. A recent study found that somatosensory-evoked potentials (SSEP) monitoring did not alert the surgeon of any impending C5 nerve palsy during the surgical procedures. MEP or EMG monitoring was thus not routinely used in our patients ^[15-17].

Another method of avoiding C5 nerve palsy is to perform prophylactic foraminotomies during either laminopasty or laminectomy, but this remains controversial since its effect has not yet been confirmed ^[18-21]. The principle is that posterior cervical spinal decompression can cause posterior shifting of the spinal cord due to C5 nerve stretching^{[22, 23]}. In this study, posterior decompression did not include preventive foramen decompression.

In preoperative imaging studies, the presence of carotid stenosis of the C3-4 level and C5 neural foramen stenosis increased the risk of C5 nerve palsy in posterior cervical spinal surgery; if C5 nerve stretch was accompanied by C5 neural foramen stenosis, the symptoms could be worse ^[24]. Imagama et al reported that a narrow C5 nerve hole formation significantly increased the incidence of postoperative C5 nerve palsy ^[25].

The limitations of this study include its retrospective design and its lack of detailed imaging data and documentation of clinical signs. Future studies are warranted to determine the association of the severity of preoperative stenosis with imaging findings and preoperative clinical findings. In addition, the lack of preoperative and postoperative radiographic contrast prevents imaging-based prediction of the occurrence of C5 nerve palsy ^[26]