Introduction

Anterior cervical spine surgery (ACSS) is commonly used for the treatment of numerous cervical disorders, such as traumatic, degenerative and congenital diseases [1, 2]. Dysphagia is one of the most common complications after ACSS that occurs frequently. Previous studies observed that dysphagia easily occurred in the early postoperative stage with reported incidence of up to 88%, and its symptoms were particularly serious in the early phase [1, 3,4,5,6]. Persistent dysphagia could lead to not eating or drinking normally and troubled the patients for a long time [38]. Knowing risk factors and preventative measures of early and persistent dysphagia after ACSS become an important work for the spine surgeons when doing the procedure.

Materials and methods

In March 2017, we searched the database PubMed, Medline, EMBASE, the Cochrane library, Clinical key, Springer link and Wiley Online Library without time restriction using the terms ‘dysphagia’, ‘swallowing disorders’ and ‘anterior cervical spine surgery’ looking for papers published in English that reported risk factors and preventative measures of early or persistent dysphagia after anterior cervical spine surgery.

The search produced a total of 515 published articles (Fig. 1). Abstracts were reviewed and included if dysphagia was a reported patient outcome measure and if the study investigated risk factors and prevention of early or persistent dysphagia after ACSS. Exclusion criteria included case reports, case series, reviews, commentaries, and cadaveric or experimental studies in animals. Fifty-nine articles including 32 prospective and 27 retrospective studies were included in this systematic review (Table 1). Level of evidence ratings were assigned to each article independently by two reviewers (J. L., N. K.) using published guidelines [69].

Fig. 1
figure 1

PRISMA (preferred reporting items for systematic reviews) flow diagram for selection of studies based on inclusion criteria during systematic review

Full size image
Table 1 Studies included in the systematic review
Full size table

Results

We evaluated 59 studies, three of them were level I evidence studies, 29 were level II evidence studies, 22 were level III evidence studies, and 3 were level IV evidence studies. Preventable risk factors and preventative measures are described in Table 2. Unpreventable risk factors are described in Table 3. The evaluated articles are described in the following.

Table 2 Preventable risk factors and preventative measures of early dysphagia after anterior cervical spine surgery
Full size table
Table 3 Unpreventable risk factors of early dysphagia after anterior cervical spine surgery
Full size table

Preventable risk factors and preventative measures

Prolonged operative time

Twelve studies investigated whether prolonged operated time was a risk factor for early or persistent dysphagia. Four studies including two prospective and two retrospective studies observed it as a risk factor, while eight studies including three prospective and five retrospective studies did not find a significant correlation.

Liu et al. [1] found that operative time is the risk factor for dysphagia during the first day (p = 0.009) to the second day (p = 0.017) after ACSS. Kalb et al. [5] found that operative time tended to be longer in those who developed dysphagia (186 vs 169 min). Rihn et al. [7] also found a correlation between operative time and the severity of dysphagia after 12 weeks of one- or two-level anterior cervical decompression and fusion (ACDF) in their prospective study (p = 0.04). Riley et al. [8] reported in a retrospective analysis that the risk of persistent dysphagia increased as the duration of surgery increased after ACDF.

On the other hand, some studies [3, 9,10,11,12,13,14,15] failed to find prolonged operative time to be a significant risk factor.

Use of BMP

There are two forms of BMP: recombinant human BMP-2 (rhBMP-2) and BMP-7 [16]. Three prospective and six retrospective studies relating the use of BMP and dysphagia after ACSS were found.

Cole et al. [17] found that rhBMP use was associated with an increased risk of dysphagia within 30 days postoperatively (OR = 1.3, 95% CI 1.1–1.5). Jain et al. [18] also found that use of rhBMP was significantly associated with the development of dysphagia after ACSS [prevalence, 2.0%; adjusted odds radio (OR), 1.53]. Fineberg et al. [16] also observed a significantly higher rate of dysphagia when BMP was utilized during ACDF (37.2 vs 22.5 per 1000 cases; p < 0.0005). Buttermann et al. [19] concluded that the use of BMP did result in increased dysphagia of patients who had primary one- to three-level ACDF with iliac-crest bone autograft. Lu et al. [20] also found the use of rhBMP-2 significantly increased the severity of dysphagia in patients undergoing 2-level ACDF (p < 0.005). Burkus et al. [21] observed high rates of dysphagia at 24 months after single-level allograft when rhBMP-2 was used (p = 0.001).

Lovasik et al. [12] found no difference in dysphagia incidence between rhBMP-2 or beta-tricalcium phosphate (bTCP) groups of patients who underwent ACDF with polyetheretherketone plastic fusion spacers during a 2-year period in their retrospective study (20 vs 17%, p = 0.5).

Endotracheal tube cuff pressure and cervical retractors

Five prospective studies investigated whether endotracheal tube cuff pressure was a risk factor for early and persistent dysphagia after ACSS. Three of them considered endotracheal tube cuff pressure (ETCP) as a risk factor and suggested decreasing ETCP or using dynamic retractors to prevent dysphagia after ACSS.

Ratnaraj et al. [22] reported that an increased ECTP during neck retraction and a prolonged retraction time were risk factors of postoperative dysphagia at 24 h (p < 0.05, r 2 = 0.61). They suggested that decreasing ETCP to 20 mmHg may be helpful in improving patient comfort following ACSS. Pattavilakom et al. [23] performed a prospective randomized clinical trail to compare the conventional Cloward-style retractor (CRS) with a novel seek retractor system (SRS) after 1- to 2-level ACDF. This study found that average retraction pressure and mean average peak retraction pressure with SRS were all lower than that with CRS (p < 0.001). Mendoza-lattes [24] reported that patients with dysphagia had a significantly higher average intraluminal pressure as well as significantly lower average mucosal perfusion secondary to retraction (p < 0.0001). They suggested that dynamic retraction associates with a lower prevalence of postoperative dysphagia.

Although three level II evidence studies investigated endotracheal tube cuff pressure as a risk factor of early and persistent dysphagia after ACSS, there were still two studies that did not find a correlation between them. Papavero et al. [25] found no correlation between the amount of retraction and early postoperative dysphagia after ACSS in their prospective study. Kowalczyk et al. [26] found decreasing ETCP to 15 mm Hg during surgery had no effect on the prevalence of dysphagia after ACSS (p > 0.05).

Use of plating, plate profile and prominence

Four studies including two prospective and two retrospective studies investigated whether using cervical plate or plate profile and position correlated with early or persistent dysphagia after ACSS. Five studies suggested using Zero-P implant and one study suggested using smaller and smoother profile plate to prevent early and persistent dysphagia.

Zeng et al. [3] observed that use of plate was one of the risk factors of early dysphagia after ACSS (p = 0.012). They also found a bigger protrusion about 1.4 mm in the dysphagia group (p < 0.001). Lee et al. [27] suggested that the use of a smaller and smoother profile plate does reduce the incidence of early and persistent dysphagia after ACSS.

But Jain et al. [18] and Chin et al. [28] failed to find a correlation between plate profile or prominence and postoperative dysphagia.

Miao et al. [29] performed a prospective study to analyze the primary efficacy and safety of a new zero-profile implant (Zero-P) in Chinese population compared with a control group of patients using anterior titanium plate. The authors reported similar clinical and radiographic outcomes of the two groups. But the incidence of dysphagia was lower and the symptom duration was much shorter in the Zero-P group. Three other studies [30,31,32] also found that using Zero-P can reduce the incidence of early dysphagia in their retrospective studies, whereas Vanekn et al. [33] didn’t find significant difference in the incidence of dysphagia between Zero-P interbody spacer and standard interbody spacer.

dC2–C7 angle

The C2–C7 angle was defined as the angle between the lines parallel to the inferior end plate of C2 and C7 vertebral bodies (Cobb’s method). dC2–C7 angle = postoperative C2–C7 angle – preoperative C2–C7 angle [14]. Three studies including two prospective and one retrospective studies investigated dC2–C7 angle as a risk factor of early or persistent dysphagia after ACSS.

Liu et al. [1] found that the dC2–C7 angle significantly associated with early dysphagia from the third to the fifth day after ACSS and when the dC2–C7 angle was greater than 9°, the incidence of dysphagia was significantly increased (p < 0.05). Two other studies [13, 34] also found that when the dC2–C7 angle was greater than 5°, the chance of developing dysphagia was significantly greater. Furthermore, Chen et al. [35] found that dC2–C7 angle was also a risk factor for combined anterior–posterior cervical spine surgery (p = 0.020).

Psychiatric factors

Kang et al. [36] performed a prospective study to identify associations between psychiatric factors and the development of dysphagia 1 year after single-level ACDF. Multivariate logistic regression showed that the psychiatric problem prior to surgery was the risk factor of persistent dysphagia (p = 0.005). They suggested that patients with a psychiatric factor should be counseled before surgery.

But Riley et al. [8] did not find significant relationship between psychiatric factors and dysphagia at 6 and 12 months after ACSS (p > 0.05).

Tobacco usage

Two prospective studies considered tobacco usage as a risk factor of early or persistent dysphagia after ACSS. Siska et al. [37] analyzed that increased degree of dysphagia was associated with tobacco usage 3 weeks after surgery (p = 0.002). Olsson et al. [38] found that the increased prevalence of dysphagia in smokers (38 vs 21%) trended toward statistical significance and the severity of dysphagia was more severe for smokers (p = 0.02).

There are still three other studies [3, 5, 39] that did not find a correlation between tobacco usage and postoperative dysphagia.

Prevertebral soft tissue swelling

Shi et al. [34] reported prevertebral soft tissue swelling (PSTS) change greater than 5 mm was prone to have early postoperative dysphagia (p = 0.000). Riley et al. [8] also found PSTS relating to dysphagia after ACDF in their retrospective study.

But Kang et al. [9] observed no statistically significant difference between PSTS and dysphagia at 1, 3, and 6 months after ACSS (p > 0.05). Kepler et al. [40] found that the degree of dysphagia had no correlation with postoperative dysphagia after 1- to 2-level ACDF (p > 0.05).

Song et al. [41] and Lee et al. [42] found that prevertebral soft tissue swelling as well as postoperative dysphagia reduced significantly with steroid administration during ACSS. But Nam et al. [43] failed to find the correlation between PSTS decrease and steroid application.

SLN or RLN palsy or injury of branches

Two studies [44, 45] considered superior laryngeal nerve (SLN) or recurrent laryngeal nerve (RLN) palsy or injury of branches as risk factor for dysphagia after ACSS. Tervonen et al. [44] suggested RLN should be detected during surgery. Razfar et al. [46] suggested spine surgeons to ensure knowledge of normal and aberrant courses of the SLN and RLN.

Unpreventable risk factors

Age

Seventeen studies including 6 prospective and 11 retrospective studies investigated whether age was a risk factor. Six studies including two prospective studies [12, 15] and four retrospective studies [3, 4, 12, 39] identified old age as a risk factor. Eleven studies including four prospective studies [1, 7, 37, 38] and seven retrospective studies [8, 13, 14, 35, 47,48,49] did not find a significant correlation.

Gender

Eighteen studies including eight prospective and ten retrospective studies investigated whether gender was a risk factor. Six studies including five prospective studies [10, 11, 25, 28, 37] and one retrospective study [3] identified female as a risk factor and one retrospective study [4] identified male as a risk factor. Eleven studies including three prospective studies [1, 27, 39] and eight retrospective studies [8, 12,13,14, 35, 47,48,49] did not find a significant correlation.

Multilevel surgery

Fourteen studies including 3 prospective and 11 retrospective studies investigated whether multilevel surgery was a risk factor. Nine studies including one prospective study [10] and eight retrospective studies [3,4,5, 8, 9, 47,48,49] identified multilevel surgery as a risk factor. Five studies including two prospective studies [14, 27] and three retrospective studies [12, 38, 50] did not find a significant correlation.

Revision surgery

Seven studies including three prospective and four retrospective studies investigated whether revision surgery was a risk factor. Two studies including one prospective study [38] and one retrospective study [50] identified revision surgery as a risk factor. Five studies including two prospective studies [14, 27] and three retrospective studies [11, 13, 47] did not find a significant correlation.

Duration of preexisting pain

Three studies including two prospective studies and one retrospective study investigated whether duration of preexisting pain was a risk factor. Two studies including one prospective study [39] and one retrospective study [8] identified duration of preexisting pain as a risk factor, while one prospective study [38] did not find a significant correlation.

BMI

Nine studies including five prospective and four retrospective studies investigated whether higher BMI was a risk factor. Four studies including two prospective studies [11, 25] and two retrospective studies [5, 50] identified higher BMI as a risk factor. Five studies including three prospective studies [1, 7, 39] and two retrospective studies [13, 37] did not find a significant correlation.

Blood loss

Five studies including one prospective and four retrospective studies investigated whether blood loss was a risk factor. One prospective study [39] identified blood loss as a risk factor. Four retrospective studies [3, 35] did not find a significant correlation.

Upper levels

Fourteen studies including six prospective and eight retrospective studies investigated whether upper levels were a risk factor. Eight studies including two prospective studies [28, 51] and six retrospective studies [5, 9, 12, 35, 47, 49] identified upper level surgeries as a risk factor. Five studies including four prospective studies [10, 11, 37, 40] and one retrospective study [8] did not find a significant correlation.

One retrospective study [52] considered dysphagia a more common clinical problem after ACDF in the mid cervical spine.

Preoperative comorbidities

Two studies [4, 37] investigated preoperative comorbidities such as chronic lung disease and diabetes as a significant risk factor. But they did not give a suggestion whether controlling the preoperative comorbidities could reduce dysphagia after surgery. Two studies [3, 12] failed to find a relationship between preoperative comorbidities and the incidence of dysphagia.

ACDF versus cervical disk replacement

Three studies [1, 4, 53] investigated the incidence of dysphagia between ACDF and cervical disk replacement, they reported lower incidence of dysphagia in the cervical disk replacement group.

Discussion

Several risk factors may lead to early or persistent dysphagia after ACSS, some of them are unpreventable, and some of them are preventable. Although some studies had different conclusions, we do find some preventative measures to reduce postoperative dysphagia.

As we know, if the esophagus and prevertebral soft tissue were retracted for a long time during the surgery, they would be injured and result in severe swelling [40]. The swelling of esophagus and prevertebral soft tissue finally led to patient dysphagia. However, it is likely that multilevel surgeries took longer operative time than single-level surgeries [8]. It still remains controversial that whether prolonged operative time was a risk factor, but avoiding a prolonged operative time and improving surgical techniques do help to prevent dysphagia after ACSS. Avoiding a prolonged operative time could reduce the injury to the prevertebral soft tissue [1]. In complex cervical cases in which a long operative time is predicted, senior surgeons with improved surgical techniques would probably have less dysphagia after surgery than fellows and residents [54].

One of the potential mechanisms leading to dysphagia after BMP use is increasing the incidence of swelling after ACSS [55]. Therefore, spine surgeons should avoid routine use of BMP except for patients who underwent revision surgery or at high risk for pseudarthrosis. For these patients, depomedrol is supposed to be administered locally [12]. Fineberg et al. [16] recommended that if BMP is to be used in ACDF, the patients should be clinically monitored for dysphagia in the early postoperative period. Edwards et al. [56] performed a prospective randomized placebo-controlled double-blind trail and found that patients receiving depomedrol experienced decreased dysphagia incidence and magnitude at all time intervals (p < 0.05). This study provided level I evidence that locally administered depomedrol on a collagen sponge could significantly decrease postoperative dysphagia following ACDF using low-dose BMP-2.

Arts et al. [57] protocoled a double-blind randomized controlled trial to determine whether adjusting ETCP after placement of a retractor during ACSS would prevent postoperative dysphagia after ACSS. Recruitment of patients had started in December 2011 and would be finished at the end of 2013. The results of this level I evidence study may lead to standard adjustment of ETCP in every patient undergoing ACSS to decrease postoperative morbidity, and now we are still waiting for the result of this study. Chen et al. [58] considered that if ETCP was reduced, the surgical field would become narrow and operations would become more difficult to perform. Therefore, they designed a preoperative tracheal traction exercise (TTE) to prevent dysphagia after ACSS. The TTE was performed twice per day, 15 counts each time, for 3 days, starting 4 days before the surgery.

Anterior plating may be responsible for higher rates of early and persistent dysphagia because of the plate’s mass effect on the adjacent esophagus and because contralateral screw placement requires more retraction past the midline than is necessary when a plate is not used [40]. In addition, scar tissue formation on a less smooth plate surface may lead to postoperative dysphagia [27]. Although most studies suggested using smaller and smoother plates or Zero-P implant to prevent early and persistent dysphagia after ACSS, better designed, especially level I, studies are required to clarify this issue.

The overenlargement of cervical lordosis may lead to bulging of the posterior pharyngeal wall, which may reduce pharyngeal space, and influence the normal process of pharyngeal squeeze and laryngeal elevation. Thus, overcorrecting the sagittal alignment of the cervical spine may increase the dysphagia rate [34]. Tian et al. [13] suggested that overenlargement of cervical lordosis should be avoided to reduce the development of postoperative dysphagia.

How psychiatric problem increased the incidence of dysphagia was unexplained, these patients tended to overreport their symptoms of dysphagia [36]. Further studies are required to study the relationship between psychiatric factors and dysphagia after ACSS, and patients with a psychiatric factor should be counseled before surgery as the medical ethics told “to comfort always” [59].

Although two studies observed tobacco usage as a risk factor of early and persistent dysphagia after ACSS, they did not give a suggestion to quit smoking to prevent postoperative dysphagia. Olsson et al. [38] found that there was no difference between those who never smoked and those who formerly smoked of postoperative dysphagia, which may indicate that quitting smoking could help to decrease dysphagia after ACSS. Further studies are required to determine the effect of smoking in postoperative dysphagia.

Suk et al. [60] reported the natural change of PSTS in the early postoperative period. PSTS was increased continuously from the 2nd day and reached a plateau on the 3rd day postoperatively. A gradual decrease was starting at the 4th day after surgery. Whether steroid administration can decrease PSTS remains controversial, but it could prevent and reduce early postoperative dysphagia. Four randomized controlled trails [41, 42, 61, 62], one of which is a double-blinded study [61], found that steroid administration significantly improved swallowing function and reduced dysphagia in the early postoperative period.

Haller et al. [63, 64] explained the clinical anatomy of SLN during ACSS to help us ensure knowledge of them.

Conclusion

Adequate preoperative preparation of the patients including preoperative tracheal traction exercise and quit smoking, properly preventative measures during surgery including maintaining endotracheal tube cuff pressure at 20 mm Hg, avoiding routine use of rhBMP-2, use of zero-profile implant, use of Zephir plate, use of new cervical retractor, steroid application, avoiding prolonged operating time, avoiding overenlargement of cervical lordosis and decreasing surgical levels, doctors ensuring knowledge of anatomy, improving surgical techniques and to comfort always are essential for preventing early and persistent dysphagia after ACSS.