Complications of anterior surgery in cervical spine trauma: An overview

2.1. Patient population 

Patients with unstable lower cervical spine injuries who underwent anterior surgery over a 15-year period (1990–2005) were identified for the study. Indication for surgery was set by the instability criteria of White and Panjabi [7]. White and Panjabi have developed a checklist to quantitatively assess cervical stability, which is based on radiological and clinical findings. Depending on the degree of injury of its supporting structures, they have assigned a value of either 1 or 2 if present, or 0 if absent. Patients having 5 or more points were considered to have an unstable injury.

Radiographic measurements were done in a standardized manner that is the X-ray tube to film distance was 1.8m, to avoid magnification errors. Kyphotic angulation was measured by drawing lines tangent to the posterior vertebral body margins on each vertebra, then the relative rotation angle created by adjacent tangents was recorded. Similarly, horizontal translation of one vertebra on the other was measured as the perpendicular distance between their two tangents.

2.2. Surgical technique 

A Southwick and Robinson left-sided anterolateral approach between the thyroid gland and the neurovascular bundle using a transverse incision was performed in all our patients. Sixty-five patients had anterior decompression and iliac bone grafting fashioned in a typical Smith–Robinson way. Nine patients underwent corpectomy and cervical spine reconstruction with titanium mesh cage; in 6 patients the Harms cage (DePuy AcroMed, Raynham, MA, USA) was used, and in the remaining 3 the Pyramesh cage (Medtronic Sofamor Danek, Memphis, TN, USA). The cage was then filled with morselized autograft from the corpectomy site. All patients were instrumented using an anterior cervical plate. Three patients had the initial Orozco plate requiring bicortical purchase, 5 had the recent Zephir™ (Medtronic Sofamor-Danek, Memphis, TN, USA) plate with unicortical titanium screws, and the remaining 66 patients were plated using an H-shaped AO/ASIF plate. In 12 out of the 66 patients, we used an AO/ASIF plate and special 3.5mm cortical screws with titanium plasma covering not needing penetration of the posterior wall, while in the remaining 54 patients a Cervical Spine Locking Plate-CSLP™ (Synthes Spine, Paoli, PA, USA) with self-locking screws was applied.

2.3. Data collection and assessment 

Demographic information, initial neurological examination, surgical reports, imaging findings and follow-up records were evaluated. Imaging studies included flexion–extension views, along with clinical reexamination of the patients. Preoperative MR imaging was available in 32 patients. Fusion was assessed radiographically at regular intervals. Mean follow-up time was 6.4 years and at least 1-year follow-up was necessary for all patients.

Data acquisition was done prospectively for 48 patients, whereas for the remaining patients it was initially based on a retrospective review of their charts and their latest imaging examinations. Additional reexamination took place for these patients and they were requested to come for clinical and radiological follow-up.

The American Spinal Cord Injury Association (ASIA) scale was utilized to evaluate the neurological status of the patients.

2.4. Postoperative care 

Postoperatively, the patients were immobilized in a hard collar for 6–12 weeks except for one patient with multilevel injury, where a halo vest was placed for additional stability.

3.1. Patient data 

Seventy-four patients had adequate follow-up to be included in the study. There were 49 male and 25 female patients. The average age of the patients was 36 years (range 17–82 years).

3.2. Type of injuries and number of operated levels 

There were 28 (38%) patients with predominantly vertebral body fractures and 46 (62%) patients with lesions involving mainly the posterior elements (Table 1, Table 2).

Table 1. Classification of injuries

	Vertebral body fractures	Posterior elements lesions
	7 compression wedge (9.5%)	32 unilateral facet fracture-dislocations (43.2%)
	13 burst (17.6%)	6 unilateral facet dislocations (8.1%)
	8 teardrop (10.8%)	8 bilateral facet dislocations with or without fracture (10.8%)

Table 2. Affected levels of injury

	Fractured vertebra	Number
	C3	1
	C4	3
	C5	11
	C6	4
	C7	9

	Affected level in dislocations	Number
	C3–C4	6
	C4–C5	9
	C5–C6	25
	C6–C7	11
	C7–T1	1

Forty patients underwent 1-level fusion, 31 had a 2-level fusion and in 2 patients 3-levels were operated on. One of the patients with a non-contiguous injury involving 4-levels had lost his diaphragmatic respiration and permanent mechanical ventilation was required; he was operated on using a long CSLP plate and underwent C4-T1 fusion.

3.3. Classification of complications 

The observed complications are summarized in Table 3.

Table 3. Complications

	Intraoperative complications (technique-related)
	Failure to provide lordosis (neutral or <100 kyphosis)	4
	Inadequate reduction of the dislocation	2
	Misplacement of the graft (malrotated position)	2
	Obliquity of the plate	3
	Minor screw intrusion in the disc	4
	No purchase of the screws (i.e., completely in the disc)	1 (revision)
	Postoperative complications
	Angulation deformity (kyphosis >100)	2
	Screw breakage (early AO plate)	1 (required removal)
	Backout of the screws (early AO plates)	4 (1 removal due to dysphagia)
	Postoperative hematoma	1
	Adjacent level ossification	8
	Temporary neural palsies	3

Modified with permission from Lambiris et al. [6].

Failure to provide lordosis, i.e. neutral alignment or immediately postoperative kyphosis of <100 was observed in 4 (5.4%) neurologically intact patients. The patients reported no symptoms at their latest follow-up. Similarly, inadequate reduction of the dislocation in 2 (2.7%) neurologically intact patients had no clinical consequence. Misplacement of the graft was noticed in 2 patients (2.7%), where the graft was placed in a malrotated position at the site of intended fusion, i.e. without providing enough cortex support; both these 2 patients achieved a solid fusion. Seven patients of the study (9.5%) were observed to have less-than-ideal positioning of the plate and/or the screws. In 4 out of them, there were screws with minor penetration of the adjacent disc, but this did not prevent healing of the fusion. The remaining 3 patients had their plates positioned obliquely, which had no clinical implications. One patient, whose cranial screws had no purchase of the vertebral body being completely in the intervertebral space, underwent early revision of the osteosynthesis.

Late angulation deformity (kyphosis >100) was noticed in 2 patients (2.7%) at their follow-up, which was not present immediately postoperatively. We had one screw breakage in a neurologically intact patient, noticed at her 6-month follow-up. The patient was asymptomatic and in the dynamic views there was no evidence of motion. The patient was informed about the broken screw and following her desire, she was operated on for implant removal.

In four patients (5.4%) there was a backout of the screws when an unrestricted backout plate was used. One of them, plated with an early AO plate, required removal of the materials due to dysphagia, when sound fusion was present. The remaining three patients had a backout of two screw threads and were advised keeping the hard collar for a longer period; two of them eventually went on to fusion without further complications, but they were reluctant to undergo a new surgery for implant removal. The remaining one was a quadriplegic polytrauma patient who remained in the ICU for 4 months postoperatively and it was considered adequate to closely follow his course radiographically, than performing a revision osteosynthesis. At the 1-year follow-up, the patient had no evidence of further screw migration and a solid fusion.

Postoperative drains were used in all patients. A postoperative hematoma developed in a patient where the drain was accidentally removed, which however subsided uneventfully. Neural complications consisted of one superior laryngeal nerve injury, one recurrent laryngeal nerve injury and one Horner syndrome, all of which recovered completely. Adjacent level ossification was observed radiographically in 8 patients (10.8%), yet no one of them reported any symptoms.

3.4. Radiological and neurological results 

Radiographic evidence of bridging bone across the entire length beneath the plate was observed in 67 patients (90.5%).

Twenty-three patients were neurologically intact (ASIA E). Neurological improvement was one ASIA scale on average for incomplete injuries; no complete injury did improve.

3.5. Morbidity and mortality 

We had no wound infections or complications from the donor site. Three patients with quadriplegia died due to pneumonia in the postoperative period while in the ICU. Massive atelectasia occurred in another two ICU patients, which required aggressive treatment with physiotherapy and bronchodilatation.

There were also two patients with persistent cardiovascular instability. One of them was a polytrauma patient who was intubated promptly and transferred to the ICU. Before intubation the patient was reported to have a complete lesion below C6 level associated with priapism. The patient was operated on the 9th post-injury day. He had persistent bradyarrythmias and hypotension, which were not controlled with medication and eventually died on the 8th postoperative day.

4.1. Complications related to the surgical approach 

The anterior approach to the cervical spine has the risk of significant complications, such as injury to the thyroid gland, the neurovascular structures of the neck, injury or rupture of the esophagus, the trachea or the thoracic duct. Pneumothorax has also been reported in cases of lesions near the cervicothoracic junction [8]. Esophageal and tracheal ruptures are serious and life-threatening incidents, which however are rather uncommon (0.25–1.49%) [1], [2], [9], [10]. Intraoperative injury to the neurovascular structures of the neck ranges from neural palsies and medium vessel injuries, to vocal cord paralysis and major vessel injury. Injury of a major vessel such as the carotid artery, the vertebral artery or the jugular vein can be a real problem, yet it is infrequent [11]. A recent study using magnetic resonance angiography (MRA) reported a 17% incidence of traumatically induced vertebral artery occlusion, though it rarely produces any symptoms due to the supply from the contralateral vertebral artery and the circle of Willis [12]. This finding deserves serious consideration however, since potential intraoperative injury of the contralateral vertebral artery may significantly compromise the injured patient, especially the elderly. We had no such complication in our series.

Voice changes and difficulty swallowing are well known issues after an anterior approach to the cervical spine [13], [14] and are in most instances temporary, although they may persist for more than 6 months in up to 12% of patients [2], [15].

Recurrent laryngeal nerve injury is the most common nerve injury after an anterior cervical approach [13]. The reported complication rate varies from 1.8 to 3.3% [16], [17], [18], [19], [20], [21], but rates as high as 10% have been reported, particularly in revision cases [22]. We had 1 patient (1.3%) with transient vocal cord paresis, who had a full recovery before discharge from hospital.

Fewer complications have been associated with the left sided approach to the lower cervical spine; the left recurrent laryngeal nerve has a longer and less oblique course and lies deep in the tracheoesophageal groove [23], thus being less susceptible to injury from overstreching during retraction [24]. This has led to strong surgeon preference regarding the laterality of approach [22].

Since all our patients underwent a left-sided approach, we cannot comment on this issue.

Hematoma, as well as injury to the pharyngeal plexus or to the hypoglosseal nerve [25], [26] can also cause dysphagia. If corpectomy is performed, postoperative drainage is particularly obligatory, since the hematoma can develop slowly and cause symptoms of a space-occupying lesion even 1 week postoperatively, as it happened in one of our patients.

It should be also noted that the first report of the CSRS in 1989, concluded that although the majority of surgical procedures were anterior ones (64%), there was a noticeably lower incidence of neurologic complications with anterior than with posterior approaches (0.64% vs. 2.18%), the reverse of what the literature experience would indicate [4].

4.2. Complications related to the technique 

Complications of unistrumented anterior cervical fusion are reported to reach up to 50% in circumferential instability after trauma [27], [28], [29], [30], [31]. The introduction of anterior cervical spine instrumentation by Boehler in 1967 was followed by the development of various anterior plates and screws, as a supplement to anterior cervical fusion techniques. In fact, there has been a continuous improvement in the implants and the instrumentation used for anterior stabilization over the last 2 decades, and several clinical studies have underlined the in vivo adequacy of anterior plate constructs, even in the presence of associated posterior lesions [16], [18], [19], [20], [32], [33], [34], [35], [36], [37]. This is further confirmed by very recent studies [6], [38], [39], [40], [41].

Less-than-ideal positioning of the implants may occur in the beginning of the learning curve and complication rates as high as 13% have been reported [19]. Reoperation rates for screw backout subsequent to disc intrusion vary from 1.1 to 3.7%, and generally concern unrestricted backout systems such as the early AO/Orozco plates or the Caspar system [18], [19]. Minor disc penetration however has no serious sequelae [32], unless there is completely no bony purchase of the screws, as occurred in one patient of our study who finally underwent revision of the osteosynthesis (Fig. 1, Case 1).

View Large Image Download to PowerPoint View Large Image Download to PowerPoint

Fig. 1. (Case 1) Multilevel injury—adjacent level ossification. A 21-year old male with fracture-dislocation of C6–C7, teardrop fracture of C6 and lateral vertebral body fracture of C7, along with undisplaced fracture of the C4 left facet (A). Initial C5–C7 stabilization was inadequate since the cranial screws were inserted into the C5–C6 intervertebral space, without having any bony purchase (B). Revision surgery followed two days later and a new C5–C7 stabilization was performed. Had the fractures of C4 and of C7 been properly evaluated, stabilization would have included the C4–C5 and C7–T1 levels to avoid instability. However, this did not occur until postoperatively and the patient wore a halo vest for 2 months, followed by a hard collar for another month. Even with these problems, postoperative radiograph at 4-years follow-up (C) shows fusion, which was however accompanied by anterior ossification at the C4–C5 level. Revision surgery, in addition to instability, is likely the cause for adjacent-level ossification at the cranial level in this patient.

Screw loosening and backout is a well known complication of earlier anterior cervical plates at a rate of 3.7–5% [16], [18], [19]. In the present study, screw backout occurred in four patients (5.4%) plated with the early AO plates and required closer follow-up; in one patient reintervention due to dysphagia was necessary to remove the materials. Initial screw anchoring by one of the caudal screws was inadequate, but it should be also noted that this patient underwent a C3–C4 fusion which might further explain the backout. One could thus assume that more than usual neck extension during the operation may be a contributing factor, because of the technical difficulties in operating at this high cervical level.

Modern implants however are restricted backout systems [42], which permit locking of the screws to the plate and they are shown to have a reduced backout rate along with improved rigidity, thus obviating the need for bicortical purchase [43], [44]. Albeit the use of newer and improved implants which allow for adequate bone anchoring, it should be also noted that our patients were mainly young individuals with good bone quality. Therefore, no concern was raised about the possibility of underlying osteoporosis and poor screw purchase.

Using the reduction technique for facet dislocations suggested by Cloward and others [32], [33], [45], [46], [47], we failed to completely reduce the dislocation in 2 (2.7%) patients. This can happen in the case of a concomitant fracture of the pedicle or the facet joint as in our two patients, or because of a long standing dislocation before the attempted reduction [48]. Given that there was partial reduction and both two patients remained neurologically intact, we did not consider that this could greatly influence their functional outcome and suggested no further intervention. The patients have solid fusion at 5 and 9 years of follow-up without any neurological symptoms (Fig. 2, Case 2).

View Large Image Download to PowerPoint

Fig. 2. (Case 2) Failure to completely reduce a dislocation. A 20-year old female with a unilateral facet fracture-dislocation (A) who was neurologically intact before surgery and her status remained unchanged postoperatively. The patient underwent anterior open reduction of the dislocation which was partially successful and a CSLP was placed. Asymptomatic solid fusion at 5 years of follow-up (B), without canal encroachment as revealed by the MRI (C).

There was one other patient not included in the study, who had an irreducible locked bilateral facet dislocation and eventually underwent posterior reduction. In these rare cases, posterior reduction and stabilization after anterior discectomy should be the treatment of choice.

Moreover, redislocation has been reported in several studies [16], [19], [32], [48], and a recent study reported a rate of 19% in high stage distractive-flexion injuries [17]. We had no complications of that kind.

On the other hand, it is well known that when narrowing of the neuroforamen occurs, nerve roots become compressed; therefore the goal is to restore the intervertebral space in lordosis. Failure to provide lordosis occurred in 4 patients (5.4%) of our study, mainly at the beginning of the learning curve or due to poor grafting technique. None of these patients however had a kyphosis over 10°, therefore we considered it as clinically insignificant.

Kyphosis exceeding 10° is considered significant by most investigators [49], but is seldom encountered in instrumented anterior fusion. This could reflect either delayed union, or residual instability especially in multilevel lesions (3 or more levels). The reported rate using older stabilization systems ranges from 1.4 to 4.7% and screw loosening is often the cause [18], [36], [50], but one should also be careful in order not to miss any instability, which would eventually lead to kyphosis if left untreated [16]. We had 2 patients (2.7%) with postoperative kyphosis, who sustained a multilevel injury and the reconstruction did not include a titanium mesh cage. In such cases, more secure postoperative bracing than a hard collar, i.e. a halo vest, may be necessary. Therefore, particularly for multilevel injuries, the posterior approach seems more appropriate [51], [52]. Alternatively, a titanium mesh cage could be used as an adjunct to the anterior stabilization [53].

Nonunion is difficult to detect in spine surgery and further imaging studies such as CT may be necessary [54]. Nonunion is usually revealed by its complications, late angulation deformity or implant failure because of material fatigue [44]. It is reported that if the 6-month limit is taken into account, there is a 5–6% rate in single-level procedures, rising to 15% in two-level fusions [5], [55]. Its clinical significance is however questioned, because a patient may be asymptomatic despite radiologic nonunion and may not require intervention [56]. We did not have any evidence of motion on dynamic views, or any associated pain. Fusion was demonstrated radiographically in 90.5% of our patients, which is consistent with the reported rate in the literature (Table 4).

Table 4. Overview of complications reported in the literature concerning anterior instrumented fusion for cervical spine injuries

	Study	Number of patients	System	Posterior lesions (%)	Neurological status preop/postop	Mean f.up	Levels	Shape of graft	Fusion	Postoperative immobilization	Complications—reoperation	Remarks
	Böhler et al. [33]	26	AO/Orozco	N/A	5 intact, no deterioration	2 y	N/A	Iliac crest, wedged	100%	N/A	None reported	Anterior open reduction for dislocations
	de Oliveira et al. [34]	40	Special plate	72	13 intact, no deterioration	3 y	31 (1-), 9 (2-)	Iliac crest, disc-shaped	100%	Hard collar for 8–10 w	Infected superficial hematoma (1), shoulder pain (1), reoperation (1): root irritation from a screw	Special plate, 4.5mm AO cortical screws, disc-shaped graft, anterior open reduction for dislocations
	Tippets et al. [20]	19	Caspar	N/A	4 intact, no deterioration	4.9 mo	8 (1-), 10 (2-), 1 (3-)	Iliac crest, rectangular	N/A	Hard collar, 3 patients halo	Vocal paresis (1), deep infection (1), reoperation (1): deep infection→incision and drainage, antibiotics	Caspar technique, rectangular graft, methyl-methacrylate use advocated in case of poor bony purchase of screws, anterior open reduction for dislocations
	Caspar et al. [16]	60	Caspar	50	13 intact, no deterioration	1 y	27 (1-), 33 (2-)	Iliac crest, rectangular	100%	Soft collar for 4–8 w	Missed instability (2), recurrent instability (1), screw loosening (3), dysphagia (1), vocal paresis (1), deep infection (1), reoperation (4): missed instability (2), recurrent instability (1), inadvertent level plating (1)	Caspar technique, rectangular graft
	Goffin et al. [48]	41	Caspar	50	15 intact, no deterioration	1 y	25 (1-), 15 (2-), 1 (3-)	Iliac crest, wedged	100%	Soft collar for 3 mo	Incomplete reduction (3), redislocation (2), reoperation (2): redislocation	Long-term follow-up of 25 patients revealed 15/25 asymptomatic ALOD [57]
	Ripa et al. [19]	92	AO/Orozco	66	16 intact, no deterioration	19.3 mo	31 (1-), 60 (2-), 1 (3-)	Iliac crest, custom- not wedged	98.9%	Most patients SOMI for 3 mo	Malposition of hardware (12), translation (1), screw loosening (4), recurrent nerve palsies (2), reoperation (1): dyphagia due to screw backout	Right-sided incisions (1 left-sided), methyl-methacrylate use advocated in case of poor bony purchase of screws
	Aebi et al. [32]	86	AO/Orozco	74	43 intact, no deterioration	40 mo	57 (1-), 28 (2-), 1 (3-)	Iliac crest, wedged	100%	Collar (hard and soft) for 6 w	Screw loosening (1), screw breakage (1), redislocation (1), dysphagia (1), Horner (1), reoperation (1): redislocation, due to small graft	Right-sided incisions, special 3.5mm cortical screws with titanium plasma covering, anterior open reduction for dislocations
	Randle et al. [18]	54	Caspar	100	8 intact, no deterioration	1 y	18 (1-), 36 (2-)	Iliac crest, rectangular	100%	Hard collar or Yale brace for 12 w	Screw loosening (2), vocal paresis (1), kyphosis (1), deep wound infection (1), donor site infection (1), CSF leak (1), reoperation (2): infection and implant removal, screw loosening	Caspar technique, rectangular graft
	Garvey et al. [35]	14	Caspar	100	3 intact, no deterioration	30 mo	6 (1-), 7 (2-), 1 (3-)	Iliac crest, wedged	100%	11/14 hard collar, 3 halo for 3 mo	Dysphagia (2), screw breakage (1), screw backout (2)	5/14 asymptomatic ALOD
	Kostuik et al. [36]	14	Morscher	N/A	N/A, no deterioration	18 mo	2 levels on average	Iliac crest, N/A	100%	Soft collar, Philadelphia, four-poster orthosis	Screw loosening (3), screw breakage (3), disc intrusion (3), plate migration (2), reoperation (1): kyphosis due to plate migration	The series included 42 patients treated with anterior plate fixation for various causes, out of whom 14 were cervical spine injuries
	Blauth et al. [61]	102	Morscher	72	20 intact, no deterioration	For 87/102 patients, N/A	N/A	Iliac crest, wedged	100%	N/A	Screw loosening (4), kyphosis (4), reoperation (3): screw loosening	Series comprises also another group of 89 patients treated with earlier systems, who had more complications. 45/87 asymptomatic ALOD
	Razack et al. [40]	22	CSLP	100	3 intact, no deterioration	32 mo	22 (1-)	Strut allograft (12), iliac crest, autograft (10)	100%	Miami J for	Plate breakage (1)	Hard collar and external bone graft stimulator were used for the broken plate and a solid fusion was reported
	Lifeso et al. [38]	18	CSLP?	100	7 intact, no deterioration	2 y	18 (1-)	Iliac crest, keystone	100%	Soft collar	None reported	Specific group of injuries (included only compression–extension stage 1 injuries)
	Brodke et al. [62]	20	CSLP	55	0 intact, no deterioration	6 mo	N/A	Iliac crest, wedged	90%	Minerva or Miami J for 8–10 wks	Instrumentation failure (1), pneumonia (1), ARDS (1), neck pain (7)	Prospective study, specific group of injuries (excluded: patients with radiculopathies or neurologically intact), no attempt for anterior open reduction
	Henriques et al. [17]	36	CSLP	100	15 intact, no deterioration	15 mo	32 (1-), 4 (2-)	Iliac crest, N/A	75%	Hard collar for 6 w	Redislocation (7), dysphagia (1), recurrent laryng. palsy (1), lateral cutaneous nerve palsy (1), reoperation (5): due to symptomatic nonunion	No attempt to reduce preoperatively, anterior open reduction for dislocations
	Daentzer et al. [60]	95	Codman	79	45 intact, no deterioration	14.9 mo	58 (1-), 36 (2-), 1 (3-)	Iliac crest, N/A	100% in 86/95	Philadelphia for 6 w	Screw loosening (1), recurrent laryngeal nerve paresis (4), pull-out (1), instability (3), reoperation (19): mainly due to instability. 10 out of 19 patients underwent additional posterior stabilization	Different design plate (semi-constrained)
	Reindl et al. [41]	41	AO (13), CSLP (28)	100	22 intact, 1 deterioration recovered at 12 mo	14.6 mo	36 (1-), 4 (2-), 1 (3-)	Iliac crest, N/A	100%	Soft collar for 6 w	Dysphagia (8), wound infection (1), donor site infection (2), missed injury (1), reoperation (1): missed injury	Two patients underwent posterior reduction and fusion due to irreducible dislocations; one of them required supplemental anterior fusion
	Kwon et al. [63]	20	CSLP or peak plate	100	14 intact, no deterioration	1 y	20 (1-)	Iliac crest, wedged	100% in 18/20	Guilford brace	Donor site infection (1)	Prospective, single-level injuries, patients with SCI were excluded
	Lambiris et al. [6]	74	AO (3), CSLP (66), Zephir (5)	73	23 intact, no deterioration	6.4 y	40 (1-), 31 (2-), 2 (3-), 1 (4-)	Iliac crest, wedged (65), cage (9)	90.5%	Hard collar for 6–12 w, 1 patient halo	Disc intrusion with no purchase of the screws (1), screw backout (4), kyphosis (2), screw breakage (1), hematoma (1), reoperation (3): screw backout and dyphagia, no purchase of the screws, screw breakage	Anterior open reduction for dislocations, 7/74 asymptomatic ALOD

ALOD: Adjacent-level ossification development.

To date, very few studies have reported on the rate of adjacent-level changes in cervical spine injuries. Garvey et al. [35] had 5 (35.7%) out of 14 patients who underwent anterior cervical plating for trauma, whereas in the study of Goffin et al. [57] with long-term follow-up of 25 patients, adjacent-level changes occurred in 15 (60%) of the patients. The reported rate of 10.8% (8/74 patients) herein is lower than the abovementioned ones and there is an ongoing study in our institution aiming at elucidating the possible reasons for the development of these changes. Consequently, whether adjacent-level ossification develops as a result of instability [35], because of small disc-to-plate distance [58], or due to the anterior surgical procedure with excessive periosteal stripping [59] is subject to further investigation and cannot be answered by the present report, but in at least one of our patients adjacent level ossification could be attributed to missed instability.

In the end, while one can note “complications” on the radiographs, they usually have no or minor clinical consequences for the patient. On the other hand, we identified 9 patients with clinically significant complications as follows: no purchase of the screws, late angulation deformity, screw breakage, backout of the screws and postoperative hematoma (Table 3). Overall, 3 out of 74 patients (4%) were reoperated due to significant screw backout causing dysphagia, no purchase of the screws being completely in the adjacent disc and screw breakage, respectively. The reoperation rate reported herein is in accordance with the literature where it ranges from 1 to 5% in most studies [16], [18], [19], [32], [36], [48]. Thus far, we have not found it is necessary to operate on the contralateral side, when a patient has a previous anterior cervical approach.

There are certain limitations in the current study. First, as it represents a single institution’s practice over a long period of time, early complications and mishaps that happened at the beginning of the learning curve may have been included. Moreover, complications that were noticed with earlier generation systems are not encountered today.

A review of the recent literature concerning instrumented anterior stabilization of cervical spine injuries revealed that the complication rate is similar among studies, as well as the reoperation rate (Table 4). It seems the technique is in a way permissive, even during the learning curve. Regarding the fact that older generation implants have been included, we would comment that even with earlier systems, the complication rate remained low in large series. In fact, the application of newer semi-constrained plates in cervical spine injuries was associated with an increased reoperation rate and the need for additional posterior stabilization in one recent study [60]. Therefore, although the advances in anterior instrumented fusion have eliminated some of the complications observed with earlier systems, careful application of each implant in the appropriate clinical setting is essential.

Despite the potential complications already described, the advantages of anterior surgery are well documented [32]. The method is very popular among spine surgeons and the trend is increasing compared to posterior procedures [5]. The technologic advances in design and metallurgy allow for an easy and durable instrumentation, thereby earlier complications associated with first generation implants and unrestricted backout systems can now be avoided. As reported in the latest 5-year report of the CSRS, the overall complication rate in elective surgery of the cervical spine can be quoted as 3–5%, with the former number representing the more serious problems [5]. In the same report however, Zeidman et al. reported that trauma accounted for 17% of cases and its neurologic sequelae were not included in the study. Therefore, the complication rate after surgery for cervical spine injuries is expected to be higher.