Previous studies have suggested that the mechanism of dysphagia after occipitocervical fusion may be caused by oropharyngeal airway stenosis. The following causes may lead to oropharyngeal stenosis: anatomical abnormalities due to RA, postoperative hematoma, edema of the airway due to intubation, anterior decompression of anterior atlantoaxial subluxation, and improper fixation angle of the upper cervical vertebra [3, 8, 13, 15, 27]. Extensive investigations measuring cervical lateral radiographs before and after surgery found that oropharyngeal stenosis was closely related to the reduced occipitocervical angle after fixation, and some specific occipitocervical angles could be used as imaging indicators to predict postoperative dysphagia [3, 10, 11, 16]. However, there are no clear clinical criteria about which parameter can be used effectively for predicting postoperative dysphagia, especially in patients with C2–3 KFS.
This study reported the influence of C2–3 KFS on preoperative and postoperative occipitocervical measurement parameters while undergoing OCF surgery. Despite the low incidence of KFS, forty patients in this study who underwent OCF surgery had C2–3 KFS. It seems that KFS is a common comorbidity for patients who are required to receive OCF surgery. In our study, we found that C2–3 KFS should not be an independent risk factor for dysphagia; however, C2–3 KFS will significantly interfere with the precise measurement of some common dysphagia-related occipitocervical parameters, such as the O-C2a and O-EAa. Due to the presence of bone deformity of the C2 vertebral body, especially vertebral body fusion in some patients with C2–3 KFS, it may be quite difficult to determine the inferior endplate of the C2 vertebra. There may be inaccurate measurements of the O-C2a and O-EAa.
Miyata et al.  reported that no postoperative dyspnea or dysphagia was observed in patients with positive dOC2a and recommend O-C2a should be equal to or slightly larger than the preoperative O-C2a during OCF. Ota et al.  demonstrated that a decrease in the O-C2 angle by 10° caused a reduction of the nPAS in the neutral position of approximately 37%. It means that the reduction of O-C2a has a significant effect on nPAS. Izeki et al.  revealed that the nPAS and postoperative dyspnea and/or dysphagia did not change over time once the O-C2a had been established after OCF. Hong et al.  reported a patient presented with dysphagia after OCF and the video-fluoroscopic study with barium showed nasopharyngeal regurgitation and delayed aspiration. They noted that hyperflexion of the upper cervical spine decreased the occipitocervical angle, which resulted in narrowing of the oropharyngeal space. The patient’s dysphagia resolved after revision operation. In this study, post-operative nPAS dramatically decreased in dysphagia group. 10 patients still experienced different level of dysphagia at final follow-up and no patients desired revision surgery because of improved neurological function. The cause of dysphagia after OCF was mainly mechanical narrowing of the oropharyngeal airway space regardless of inflammation, edema or neuropraxia, so the possibility of spontaneous recovery was small . In such case, we recommend to perform early revision surgery to resolve dysphagia.
The simulation study combined with case-control study by Meng et al.  indicated that a dO-C2a of − 5° could be the threshold between dysphagia and normal swallowing. Wang et al.  demonstrated that the O-C2a could be critical predictor for postoperative dysphagia in patients undergoing OCF and avoiding O-C2a reduction greater than 5° could effectively avert postoperative dysphagia. Shimizu et al.  explained the risk of dysphagia and hyperlordotic/kyphotic malalignment secondary to excessive O-C2a decrease or increase and suggested that a change in O-C2a of > ± 5° would confer a risk of postoperative dysphagia and distal junctional kyphosis. O-C3a was measured on lateral x-rays in this study due to the difficulty in identifying the inferior endplate of the C2 vertebra caused by C2–3 vertebra fusion. O-C3a in the neutral position ranged from approximately − 16° to 29° in our patients. We discovered that using specific value of O-C3a directly had no predictive effect on postoperative dysphagia, similar to the O-C2a, but the dO-C3a did have a predictive effect on postoperative dysphagia. Therefore, we inferred that, due to the large variation in the cervical vertebra, there was poor accuracy when using only a specific value of the O-C3a to predict dysphagia.
The scatter diagrams showed that the dO-C3a, dO-C2a, dO-Da, and dOc-Axa were all linearly dependent on dnPAS in patients with C2–3 KFS, while there was a higher R2 value between the dO-C3a and dnPAS compared to those of dO-C2a, dO-Da and dOc-Axa. In the multiple regression analysis, it demonstrated that dO-C3a was the only significant parameter that affected the variation of nPAS. It meant that decrease of O-C3a may lead to decrease of nPAS which could pose a threat to the patient’s swallowing function after surgery.
Additionally, dO-C3a was significantly accurate as predictor for postoperative dysphagia in C2–3 KFS patients undergoing OCF surgery and the AUC of dO-C3a was the largest compared to dO-C2a, dO-Da and dOc-Axa. We hypothesized that the measurement error of O-C2a caused by difficulty in selecting the inferior endplate of C2 vertebrae due to C2–3 vertebral body fusion had an effect on the predictive ability of O-C2a for postoperative dysphagia. Therefore, the dO-C3a had a greater ability to predict postoperative dysphagia in patients with C2–3 KFS. Compared to conventional predictive value of postoperative dysphagia of a dO-C2a of − 5°, O-C3a reduction greater than 2° is a risky factor for predicting postoperative dysphagia. Because of more cervical segment involved when measuring O-C3a, a small angular change would result in a remarkable decrease in nPAS, which would further cause postoperative dysphagia. Based on our data and analysis, we recommended that the dO-C3a in OCF should be over − 2° in the neutral position to prevent dysphagia.
Nagashima et al.  proposed the Oc-Axa, a method of measuring the occipitocervical angle, particularly applicable to OCF with abnormalities of the C2 vertebral body. However, previous studies did not explore the strong relationship between the Oc-Axa and postoperative dysphagia after OCF. In addition, when measuring Oc-Axa, we also found the cases of circular external occipital protuberance and there may be positioning errors in the identification of the landmark, as pointed out by Nagashima. In this study, the obstacle to accurately measuring the O-Da and Oc-Axa was the abnormal morphology of the odontoid process in some patients, such as destruction, fracture, and/or deformation. For these patients, to decrease the measuring error, we selected the posterior longitudinal line of the C2 vertebral base to measure the O-Da and Oc-Axa. Despite the use of these methods, the measurement of O-C3a in C2–3 KFS patients could yield more accurate results than the measurement of O-Da and Oc-Axa.
There are several limitations in our present study. First, the study was a retrospective study. Second, when the odontoid process is deformed, the accurate measurement of the above angles will be affected to a certain extent. Third, we relied on patient reported symptoms of dysphagia without formal swallow evaluation like fiberoptic esophagoscopy or Dysphagia Short Questionnaire  pre- and post-operatively. In the end, the cohort in this study was relatively small and from a single patient population. A larger multi-institutional study should be conducted to more accurately predict postoperative dysphagia with appropriate parameters after OCF surgery especially in patients with C2–3 KFS.